[Paper Review] The dry history of liquid computers

Andrew Adamatzky|arXiv (Cornell University)|Nov 25, 2018

Advanced Data Storage Technologies23 references4 citations

TL;DR

This paper presents a comprehensive historical and technical overview of liquid-based computing devices spanning over a century, from hydraulic calculators to modern reaction-diffusion and droplet-based systems. It demonstrates how liquids serve as versatile substrates for computation through fluid dynamics, chemical reactions, and physical interactions, highlighting their potential for unconventional, biomimetic, and energy-efficient computing architectures.

ABSTRACT

A liquid can be used to represent signals, actuate mechanical computing devices and to modify signals via chemical reactions. We give a brief overview of liquid based computing devices developed over hundreds of years. These include hydraulic calculators, fluidic computers, micro-fluidic devices, droplets, liquid marbles and reaction-diffusion chemical computers.

Motivation & Objective

• To document and analyze the evolution of liquid-based computing devices from the 19th century to the present.
• To demonstrate the computational versatility of liquids through fluid dynamics, chemical reactions, and physical interactions.
• To highlight underappreciated historical prototypes and modern microfluidic systems as viable platforms for unconventional computing.
• To inspire future research by showcasing the potential of liquid substrates in creating adaptive, energy-efficient, and biologically inspired computing systems.

Proposed method

• The paper surveys historical and contemporary liquid computing devices using a chronological, device-based classification system.
• It analyzes hydraulic algebraic machines that use water displacement to solve algebraic equations via geometric immersion.
• It examines fluidic logic systems where pressure and flow represent Boolean logic states in pneumatic or hydraulic channels.
• It details reaction-diffusion computers that use propagating wave fronts in excitable chemical media like the Belousov-Zhabotinsky reaction to perform logic operations.
• It explores droplet and liquid marble systems where physical collisions or chemical diffusion between droplets implement logic gates.
• It reviews microfluidic and emulsified systems, including BZ vesicles and droplets, where signal transmission occurs via diffusing reagents or wave propagation.

Experimental results

Research questions

• RQ1 How have liquid-based systems historically solved mathematical and computational problems before the advent of electronic computers?
• RQ2 What are the physical and chemical mechanisms by which liquids can represent, transmit, and process information?
• RQ3 How do modern microfluidic and reaction-diffusion systems achieve reliable logic operations using liquid substrates?
• RQ4 What are the advantages and limitations of liquid-based computing compared to electronic or mechanical systems?
• RQ5 Can liquid-based systems be scaled into complex, reconfigurable, and reversible computing architectures?

Key findings

• A hydraulic algebraic machine from 1901 used water displacement in a paraboloid-shaped body to compute the nth root of a number by equating displaced water weight to the desired power.
• Lukyanov’s 1936 hydraulic integrator and Phillips’ 1949 computer used fluid flow and pressure to solve differential equations and simulate dynamic systems.
• Fluid mappers from 1949 used fluid flow to explore and map templates, solving mazes by visualizing flow paths.
• Belousov-Zhabotinsky (BZ) reaction-diffusion systems demonstrated the ability to compute Voronoi diagrams and skeletons of shapes through wave-front interactions.
• Droplet-based logic systems, including chemotactic droplets and pressure-driven droplets, solved mazes and implemented logic gates via controlled collisions and flow.
• BZ droplets and vesicles in emulsions demonstrated signal transmission, memory functions, and implementation of Boolean gates such as XOR and NOR through diffusive coupling and wave propagation.