pith. sign in

arxiv: 1907.11385 · v1 · pith:BPBGLLIFnew · submitted 2019-07-26 · 💻 cs.ET · cond-mat.soft

Liquid metal solves maze

Andrew Adamatzky , Alessandro Chiolerio , Konrad Szaci{\l}owski This is my paper

Pith reviewed 2026-05-24 15:35 UTC · model grok-4.3

classification 💻 cs.ET cond-mat.soft
keywords galliumliquid metalmaze solvingelectrical currentsodium hydroxideunconventional computingpath findingdroplet navigation
0
0 comments X

The pith

A gallium droplet solves a physical maze by moving along paths of highest electrical current density when DC voltage is applied between start and finish.

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

The paper shows that a compact droplet of liquid gallium placed in a maze filled with sodium hydroxide solution will travel from a start point to a destination when direct current flows between those points. The motion occurs because the droplet's high electrical conductivity pulls it toward regions of strongest current while its surface tension keeps it from breaking apart and its conformability lets it turn corners. A reader would care because the demonstration replaces programmed search algorithms with the droplet's own physical response to the electric field. The approach is presented as a route to liquid-state computing devices that operate without solid electronics.

Core claim

A maze filled with sodium hydroxide solution is solved by a gallium droplet when direct current is applied between start and destination loci. During the maze solving the droplet stays compact due to its large surface tension, navigates along lines of the highest electrical current density due its high electrical conductivity, and goes around corners of the maze's corridors due to its high conformability. The droplet maze solver has a long life-time due to the negligible vapour tension of liquid gallium and its corrosion resistance and its operation enables computational schemes based on liquid state devices.

What carries the argument

The gallium droplet, which remains intact and follows gradients of electrical current density inside the NaOH-filled maze.

If this is right

  • Maze solving becomes a direct physical process driven by current density rather than an algorithmic search.
  • Liquid-state devices gain a demonstrated path-finding capability that can be reused across multiple mazes.
  • The negligible vapor pressure and corrosion resistance of gallium allow repeated or prolonged operation of the solver.
  • The same material properties support construction of other liquid-based computational elements that interact with electric fields.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same current-following behavior could be tested in non-planar or branched channel networks to check whether the droplet still selects the shortest or lowest-resistance route.
  • Replacing the fixed electrodes with movable contacts might allow the droplet to solve dynamically changing mazes without resetting the setup.
  • Because the droplet's motion depends on conductivity contrast, mixtures of different electrolytes could be used to encode weighted paths inside a single maze.

Load-bearing premise

The droplet remains compact and follows lines of highest electrical current density without becoming stuck, dissolving, or deviating due to surface tension, conductivity, and conformability of gallium in the NaOH environment.

What would settle it

A controlled run in which the droplet either dissolves, fragments, or fails to reach the destination electrode while current is applied would falsify the claim that the droplet reliably solves the maze.

Figures

Figures reproduced from arXiv: 1907.11385 by Alessandro Chiolerio, Andrew Adamatzky, Konrad Szaci{\l}owski.

Figure 1
Figure 1. Figure 1: Topology of mazes used in experiments, scaling is 50% of the original Figure 1: Topology of mazes used in experiments, scaling is 50% of the original i [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Snapshots of the gallium droplet travelling in the maze Figure 2: Snapshots of the gallium droplet travelling in the maze M1 [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Dynamic of velocity of the gallium droplet travelling in (a) maze (Fi2) d (b) M(Fi4) Figure 3: Dynamic of velocity of the gallium droplet travelling in (a) maze [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Snapshots of the gallium droplet travelling in the maze Figure 4: Snapshots of the gallium droplet travelling in the maze [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Illustration of the (a) experiment and (b) its scheme when the droplet Figure 5: Illustration of the (a) experiment and (b) its scheme when the droplet [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Visualisation of the highest current density in the maze [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 6
Figure 6. Figure 6: Visualisation of the highest current density in the maze Figure 6: Visualisation of the highest current density in the maze M2. P ldild ihl hbd ildih [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FEM results showing a comparison between Figure 7: FEM results showing a comparison between [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
read the original abstract

A room temperature liquid metal features a melting point around room temperature. We use liquid metal gallium due to its non-toxicity. A physical maze is a connected set of Euclidean domains separated by impassable walls. We demonstrate that a maze filled with sodium hydroxide solution is solved by a gallium droplet when direct current is applied between start and destination loci. During the maze solving the droplet stays compact due to its large surface tension, navigates along lines of the highest electrical current density due its high electrical conductivity, and goes around corners of the maze's corridors due to its high conformability. The droplet maze solver has a long life-time due to the negligible vapour tension of liquid gallium and its corrosion resistance and its operation enables computational schemes based on liquid state devices.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript claims to demonstrate experimentally that a gallium liquid-metal droplet solves a physical maze filled with NaOH solution when DC current is applied between start and destination points. The droplet is asserted to remain compact due to surface tension, follow lines of highest current density due to high conductivity, navigate corners due to conformability, and operate with long lifetime due to negligible vapor tension and corrosion resistance, enabling liquid-state computational schemes.

Significance. If the result holds with supporting data, the work offers a direct experimental observation of a physical system performing maze-solving via intrinsic material properties (conductivity, surface tension, conformability) without fitted parameters or derived equations. This could contribute to explorations of unconventional or liquid-state computing, though the absence of quantitative metrics currently limits evaluability and broader impact.

major comments (2)
  1. [Abstract] Abstract: The central claim of a successful demonstration is unsupported by any quantitative data, success rates, timing measurements, controls (e.g., no-current or alternative paths), or error analysis, preventing evaluation of whether the droplet reliably follows current-density lines or remains compact as asserted.
  2. [Abstract] Abstract: The assertion of 'corrosion resistance' and 'long life-time' enabling extended operation is load-bearing for the compactness and longevity claims, yet no reaction-rate data, lifetime measurements under DC bias, or references addressing the known amphoteric reaction of Ga with NaOH (producing soluble gallate and H2) are provided to counter expected dissolution or bubble disruption.
minor comments (1)
  1. [Abstract] Abstract: The final sentence on 'computational schemes based on liquid state devices' is stated without any concrete example, architecture, or reference, leaving the broader implication unclear.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed review and valuable feedback on our manuscript. We have carefully considered the comments and provide point-by-point responses below. We believe the revisions will address the concerns and improve the clarity and rigor of the work.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim of a successful demonstration is unsupported by any quantitative data, success rates, timing measurements, controls (e.g., no-current or alternative paths), or error analysis, preventing evaluation of whether the droplet reliably follows current-density lines or remains compact as asserted.

    Authors: We agree with the referee that additional quantitative support would enhance the manuscript. The demonstration in the original submission was primarily visual through experimental footage. In the revised version, we have included quantitative data from repeated experiments, such as success rates across multiple trials, average solving times, and comparisons with control experiments (no applied current and alternative electrode placements). Error bars and statistical analysis have been added to support the claims of reliability and path following. revision: yes

  2. Referee: [Abstract] Abstract: The assertion of 'corrosion resistance' and 'long life-time' enabling extended operation is load-bearing for the compactness and longevity claims, yet no reaction-rate data, lifetime measurements under DC bias, or references addressing the known amphoteric reaction of Ga with NaOH (producing soluble gallate and H2) are provided to counter expected dissolution or bubble disruption.

    Authors: The referee raises an important point regarding the chemical stability of gallium in NaOH. While the manuscript highlighted corrosion resistance based on the short experimental timescales, we acknowledge the need for supporting evidence. We have added references to studies on gallium-alkali interactions and included new experimental data on the droplet's lifetime under DC bias, demonstrating minimal mass loss and no disruptive bubbling over the operational period. The claim has been revised to specify the conditions under which long lifetime is observed. revision: yes

Circularity Check

0 steps flagged

No circularity: purely experimental observation with no derivation chain

full rationale

The manuscript reports a physical experiment in which a gallium droplet is observed to traverse a NaOH-filled maze under applied DC voltage. No equations, fitted parameters, predictions, or mathematical derivations are present. The central claim is a direct empirical result rather than a quantity computed from prior inputs or self-citations. No load-bearing steps reduce to self-definition, fitted inputs, or imported uniqueness theorems. The paper is therefore self-contained as an experimental demonstration.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are introduced; the work relies on established physical properties of gallium and electric fields.

pith-pipeline@v0.9.0 · 5652 in / 997 out tokens · 26109 ms · 2026-05-24T15:35:47.331559+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · 1 internal anchor

  1. [1]

    Hot ice computer.Physics Letters A, 374(2):264–271, 2009

    Andrew Adamatzky. Hot ice computer.Physics Letters A, 374(2):264–271, 2009

    work page 2009

  2. [2]

    Slime mold solves maze in one pass, assisted by gradient of chemo-attractants.IEEE transactions on nanobioscience, 11(2):131–134, 2012

    Andrew Adamatzky. Slime mold solves maze in one pass, assisted by gradient of chemo-attractants.IEEE transactions on nanobioscience, 11(2):131–134, 2012. Adamatzky, Chiolerio, Szaciłowski: Liquid metal solves maze 12

    work page 2012

  3. [3]

    Physical maze solvers

    Andrew Adamatzky. Physical maze solvers. all twelve prototypes implement 1961 lee algorithm. InEmergent Computation, pages 489–504. Springer, 2017

    work page 1961

  4. [4]

    A brief history of liquid computers.Philosophical Transactions of the Royal Society B, 374(1774):20180372, 2019

    Andrew Adamatzky. A brief history of liquid computers.Philosophical Transactions of the Royal Society B, 374(1774):20180372, 2019

    work page 2019

  5. [5]

    Supplementary materials

    Andrew Adamatzky, Alessandro Chiolerio, and Konrad Szaciłowski. Supplementary materials. Liquid metal solves maze. http://doi.org/10.5281/zenodo.3344993, 2019

    work page doi:10.5281/zenodo.3344993 2019

  6. [6]

    Collision-free path plan- ning in the belousov-zhabotinsky medium assisted by a cellular automaton

    Andrew Adamatzky and Benjamin de Lacy Costello. Collision-free path plan- ning in the belousov-zhabotinsky medium assisted by a cellular automaton. Naturwissenschaften, 89(10):474–478, 2002

    work page 2002

  7. [7]

    Finding the optimal path with the aid of chemical wave.Physica D: Nonlinear Phenomena, 106(3-4):247–254, 1997

    K Agladze, N Magome, R Aliev, T Yamaguchi, and K Yoshikawa. Finding the optimal path with the aid of chemical wave.Physica D: Nonlinear Phenomena, 106(3-4):247–254, 1997

    work page 1997

  8. [8]

    Electric current solves mazes.Physics Education, 49(4):443, 2014

    Simon Ayrinhac. Electric current solves mazes.Physics Education, 49(4):443, 2014

    work page 2014

  9. [9]

    The electron in the maze

    Simon Ayrinhac. The electron in the maze. In Andrew Adamatzky, editor, Shortest Path Solvers. From Software to Wetware, pages 409–420. Springer, 2018

    work page 2018

  10. [10]

    Dynamics of microfluidic droplets

    Charles N Baroud, Francois Gallaire, and Rémi Dangla. Dynamics of microfluidic droplets. Lab on a Chip, 10(16):2032–2045, 2010

    work page 2032

  11. [11]

    Dynamics of chemotactic droplets in salt concentration gradients.Langmuir, 30(40):11937–11944, 2014

    Jitka Cejkova, Matej Novak, Frantisek Stepanek, and Martin M Hanczyc. Dynamics of chemotactic droplets in salt concentration gradients.Langmuir, 30(40):11937–11944, 2014

    work page 2014

  12. [12]

    The pH sensitivity of solvated tectomer electronics

    Alessandro Chiolerio, Carsten Jost, Thomas C. Draper, and Andrew Adamatzky. The ph sensitivity of solvated tectomer electronics.arXiv preprint arXiv:1901.05519v1, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1901

  13. [13]

    Smart fluid systems: the advent of autonomous liquid robotics

    Alessandro Chiolerio and Marco B Quadrelli. Smart fluid systems: the advent of autonomous liquid robotics. Advanced Science, 4(7):1700036, 2017

    work page 2017

  14. [14]

    Quadrelli

    Alessandro Chiolerio and Marco B. Quadrelli. Colloidal energetic systems. Energy Technology, 1800580:1–10, 2019

    work page 2019

  15. [15]

    Path planning using Laplace’s equation

    Christopher I Connolly, JB Burns, and R Weiss. Path planning using Laplace’s equation. InRobotics and Automation, 1990. Proceedings., 1990 IEEE International Conference on, pages 2102–2106. IEEE, 1990

    work page 1990

  16. [16]

    Glow discharge based device for solving mazes

    Alexander E Dubinov, Artem N Maksimov, Maxim S Mironenko, Nikolay A Pylayev, and Victor D Selemir. Glow discharge based device for solving mazes. Physics of Plasmas, 21(9):093503, 2014. Adamatzky, Chiolerio, Szaciłowski: Liquid metal solves maze 13

    work page 2014

  17. [17]

    Solving mazes using microfluidic networks.Langmuir, 19(11):4714–4722, 2003

    Michael J Fuerstman, Pascal Deschatelets, Ravi Kane, Alexander Schwartz, Paul JA Kenis, John M Deutch, and George M Whitesides. Solving mazes using microfluidic networks.Langmuir, 19(11):4714–4722, 2003

    work page 2003

  18. [18]

    A potential field approach to path planning

    Yong K Hwang and Narendra Ahuja. A potential field approach to path planning. Robotics and Automation, IEEE Transactions on, 8(1):23–32, 1992

    work page 1992

  19. [19]

    Maze solving by chemotactic droplets.Journal of the American Chemical Society, 132(4):1198–1199, 2010

    István Lagzi, Siowling Soh, Paul J Wesson, Kevin P Browne, and Bartosz A Grzybowski. Maze solving by chemotactic droplets.Journal of the American Chemical Society, 132(4):1198–1199, 2010

    work page 2010

  20. [20]

    An algorithm for path connections and its applications

    Chin Yang Lee. An algorithm for path connections and its applications. IRE transactions on electronic computers, 3:346–365, 1961

    work page 1961

  21. [21]

    mercury beating heart

    Shu-Wai Lin, Joel Keizer, Peter A. Rock, and Herbert Stenschke. On the mechanism of oscillations in the "mercury beating heart".Proc. Natl. Acad. Sci., 71(11):4477–4481, 1974

    work page 1974

  22. [22]

    Springer, 2019

    Jing Liu, Lei Sheng, and Zhi-Zhu He, editors.Liquid metal soft machines. Springer, 2019

    work page 2019

  23. [23]

    pulsierendes quecksilberherz

    Hartwig Moellenkamp, Bolko Flintjer, and Jansen Walter. 200 jahre "pulsierendes quecksilberherz". CHEMKON, 1(2):117–125, 1995

    work page 1995

  24. [24]

    Maze solving automatons for self-healing of open interconnects: Modular add-on for circuit boards.Applied Physics Letters, 106(12):123103, 2015

    Aswathi Nair, Karthik Raghunandan, Vaddi Yaswant, Sreelal S Pillai, and Sanjiv Sambandan. Maze solving automatons for self-healing of open interconnects: Modular add-on for circuit boards.Applied Physics Letters, 106(12):123103, 2015

    work page 2015

  25. [25]

    Petrusevski, and Sani Demiri

    Netodija Najdolski, Valentin Mirceski, Vladimir M. Petrusevski, and Sani Demiri. Mercury beating heart: Modifications to the classical demonstration. J. Chem. Educ.N, 84(8):1292–1295, 2007

    work page 2007

  26. [26]

    Superspreading driven by marangoni flow

    Alex D Nikolov, Darsh T Wasan, Anoop Chengara, Kalman Koczo, George A Policello, and Istvan Kolossvary. Superspreading driven by marangoni flow. Advances in colloid and interface science, 96(1-3):325–338, 2002

    work page 2002

  27. [27]

    The method of potential functions for coding constraints of the external space in an intelligent mobile robot.Soviet Automatic Control, 17(6):45–51, 1984

    VV Pavlov and AN Voronin. The method of potential functions for coding constraints of the external space in an intelligent mobile robot.Soviet Automatic Control, 17(6):45–51, 1984

    work page 1984

  28. [28]

    Glow discharge in microfluidic chips for visible analog computing

    Darwin R Reyes, Moustafa M Ghanem, George M Whitesides, and Andreas Manz. Glow discharge in microfluidic chips for visible analog computing. Lab on a Chip, 2(2):113–116, 2002

    work page 2002

  29. [29]

    The lee path connection algorithm.IEEE Transactions on computers, 100(9):907–914, 1974

    Frank Rubin. The lee path connection algorithm.IEEE Transactions on computers, 100(9):907–914, 1974. Adamatzky, Chiolerio, Szaciłowski: Liquid metal solves maze 14

    work page 1974

  30. [30]

    Epithelial cell guidance by self-generated egf gradients.Integrative Biology, 4(3):259–269, 2012

    Cally Scherber, Alexander J Aranyosi, Birte Kulemann, Sarah P Thayer, Mehmet Toner, Othon Iliopoulos, and Daniel Irimia. Epithelial cell guidance by self-generated egf gradients.Integrative Biology, 4(3):259–269, 2012

    work page 2012

  31. [31]

    Presentation of a maze-solving machine

    Claude E Shannon. Presentation of a maze-solving machine. In8th Conf. of the Josiah Macy Jr. Found.(Cybernetics), pages 173–180, 1951

    work page 1951

  32. [32]

    Analogue computation of collision-free paths

    L Tarassenko and A Blake. Analogue computation of collision-free paths. In Robotics and Automation, 1991. Proceedings., 1991 IEEE International Conference on, pages 540–545. IEEE, 1991

    work page 1991

  33. [33]

    Parallel analogue computation for real-time path planning

    Lionel Tarassenko, Gillian Marshall, Felipe Gomez-Castaneda, and Alan Murray. Parallel analogue computation for real-time path planning. In VLSI for Artificial Intelligence and Neural Networks, pages 93–99. Springer, 1991

    work page 1991

  34. [34]

    Self-propelled oil droplets consuming “fuel” surfactant.Journal of the American Chemical Society, 131(14):5012–5013, 2009

    Taro Toyota, Naoto Maru, Martin M Hanczyc, Takashi Ikegami, and Tadashi Sugawara. Self-propelled oil droplets consuming “fuel” surfactant.Journal of the American Chemical Society, 131(14):5012–5013, 2009

    work page 2009

  35. [35]

    Scott J. Turner. Homeostasis as a fundamental principle for a coherent theory of brains. Philosophical Transactions R. Soc. B, 374(20180373), 2019

    work page 2019

  36. [36]

    Vining, Fernando Esponda, Melanie E

    William F. Vining, Fernando Esponda, Melanie E. Moses, and Stephanie Forrest. How does mobility help distributed systems compute?Philosophical Transactions R. Soc. B, 374(20180375), 2019

    work page 2019

  37. [37]

    The maze solving computer

    Richard A Wallace. The maze solving computer. InProceedings of the 1952 ACM national meeting (Pittsburgh), pages 119–125. ACM, 1952

    work page 1952

  38. [38]

    Lei Wang and Jing Liu. Electromagnetic rotation of a liquid metal sphere or pool within a solution.Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 471(2178):20150177, 2015

    work page 2015

  39. [39]

    A potential maze solving algorithm for a micromouse robot

    L Wyard-Scott and Q-HM Meng. A potential maze solving algorithm for a micromouse robot. InCommunications, Computers, and Signal Processing,

    work page

  40. [40]

    IEEE, 1995

    Proceedings., IEEE Pacific Rim Conference on, pages 614–618. IEEE, 1995

    work page 1995

  41. [41]

    Yun, David Cortie, Lei Jiang, and Xiaolin Wang

    Zhenwei Yu, Frank W. Yun, David Cortie, Lei Jiang, and Xiaolin Wang. Discovery of a voltage-stimulated heartbeat effect in droplets of liquid gallium. Phys. rev. Lett., 121:024302, 2018

    work page 2018

  42. [42]

    Synthetically chemical-electrical mecha- nism for controlling large scale reversible deformation of liquid metal objects

    Jie Zhang, Lei Sheng, and Jing Liu. Synthetically chemical-electrical mecha- nism for controlling large scale reversible deformation of liquid metal objects. Scientific reports, 4:7116, 2014

    work page 2014

  43. [43]

    Self-fueled biomimetic liquid metal mollusk.Advanced Materials, 27(16):2648–2655, 2015

    Jie Zhang, Youyou Yao, Lei Sheng, and Jing Liu. Self-fueled biomimetic liquid metal mollusk.Advanced Materials, 27(16):2648–2655, 2015

    work page 2015