A hybrid computer uses rotating mechanical beams for memory and electrical contacts for logic, performing reprogrammable computation through physical motion in environments where conventional electronics struggle.
(Nanowerk Spotlight) A lock resistant to brute-force attacks because its combination exists not as a number but as a precise sequence of physical rotations. A robot controller that processes commands without a microchip, its logic encoded in the geometry of snapping beams. Engineers have now demonstrated both functionalities within a single device, showcasing a new class of mechanical computing that processes information through physical deformations rather than electronic signals.
The concept predates silicon by centuries. Charles Babbage’s analytical engine, nineteenth-century automated looms, and ancient astronomical calculators all processed information mechanically. Contemporary research pursues something more ambitious: materials that sense, remember, and respond autonomously, functioning in environments where conventional electronics fail. High-radiation zones, extreme temperatures, and applications demanding radical simplicity all call for alternatives to silicon.
The key enabling structure is the bistable mechanism, a device that snaps between two stable positions like a light switch and stays put without consuming power. Each position can represent a binary digit. Chain several together, and you have mechanical memory. Engineer them to interact during transitions, and logic operations emerge.
A fundamental obstacle has stalled progress. Systems optimized for logic operations typically lack persistent memory. Those with robust memory handle only simple computations. Unifying both capabilities within a single architecture capable of sequential decision-making, where outputs depend on the full history of past states, has remained elusive. Most designs also rely on linear actuation, pushing and pulling along straight paths. This imposes stroke-length limits and geometric constraints that cap achievable complexity.
A research team from McGill University in Montreal, Canada, has now addressed these challenges. Publishing in Advanced Science (“Rotary Electromechanical System Integrating Non‐Reciprocal Memory and Combinational Logic”), the researchers present a rotary electromechanical system that integrates non-reciprocal mechanical memory with combinational electrical logic.
The device operates as a finite-state machine, a computing construct consisting of discrete states and defined transition rules. A vending machine that tracks coin insertions and product selections before dispensing an item follows this same logic pattern. Here, the finite-state machine runs on rotational torque rather than linear motion.
Characterization of the rotary bistable module with switchable global polarization states. ( 𝐴) Design of the rotary bistable module. ( 𝐵) Experimental demonstration of the 90 ◦rotation-driven module. ( 𝐶) FEA and experimental mechanical characterization of distinct modules with varying in-plane thickness. ( 𝐷) Deformation mode comparison between the initial point and the maximum torque point of the pre-shaped beam. ( 𝐸) Von Mises stress distribution of six representative deformation modes of the pre-shaped beam under an applied torque. ( 𝐹) Contour plot of equilibrium angle, 𝜃∗ , in the design space defined by non-dimensional geometry parameters, 𝐻∕𝐿and 𝑅∕𝐿. ( 𝐺) Contour plot of bistability index, 𝐸𝑚𝑖𝑛 ∕𝐸𝑚𝑎𝑥 , in the design space defined by non-dimensional geometry parameters, 𝐻∕𝐿and 𝑅∕𝐿. ( 𝐻) Thickness-independent equilibrium angle and bistable index properties of the module. (Image: Reproduced from DOI:10.1002/advs.202522133, CC BY) (click on image to enlarge)
The core component is a bistable module built from three linked outer rings and one inner ring connected by four compliant beams. These beams are arranged with fourfold symmetry, meaning four identical elements spaced 90° apart around the center. Twisting the outer rings while holding the inner ring fixed causes snap-through buckling at approximately 90° of rotation. The module flips between horizontal and vertical orientations, encoding binary 0 and 1. Key dimensions include a beam length of 16 mm, inner ring radius of 6 mm, and beam height of 5.4 mm.
Careful parametric design ensured that each transition spans exactly 90°. Beam thickness controls the snap-through force without altering this equilibrium angle, allowing independent tuning of mechanical stiffness and geometric behavior. Finite element simulations and torsional experiments on specimens with beam thicknesses ranging from 0.53 to 0.74 mm confirmed these properties.
Electrical functionality arises from conductive copper tape wrapped around the module and elastic contact ports fixed at orthogonal positions. In the horizontal state, the module touches contacts aligned along the x-direction. Rotated 90°, it engages contacts along the y-direction, with a reliable contact range spanning 84° to 102°. This switchable pathway creates a reconfigurable circuit. One orientation yields output 1 along x and 0 along y; the other reverses these values. A single mechanical bit thereby implements two logic gates simultaneously: a NOT gate that inverts the input and a Buffer gate that passes it unchanged. LED indicators signal outputs instantly.
Stacking two modules with different beam thicknesses produces a two-bit system exhibiting deterministic transition sequences. Under clockwise torque, the softer module snaps first. Reversing the stacking order yields a different sequence, creating non-reciprocal memory: forward and backward rotations traverse distinct state pathways. This history dependence is the hallmark of sequential logic.
Wiring contact ports in series or parallel extends the system to additional logic gates. Series connections require both inputs to be active for current to flow, producing AND logic. Parallel connections allow current to flow if either input is active, producing OR logic. Combining these building blocks with NOT gates generates NAND, NOR, XNOR, and XOR functions. The architecture matches the input-output behavior of standard digital logic while retaining embedded memory.
The researchers then doubled the module count, scaling from two bits to four. This four-bit finite-state machine includes 16 possible configurations, with transition rules governed by stacking order and deformation sequences. The team demonstrated three applications. Digital combination locking emerged from the complex state transitions: the system reaches an unlocked state only after receiving a precise rotational sequence, making it significantly more difficult to breach without knowledge of the correct configuration pathway. A controller for a Mecanum-wheeled car, whose specially angled rollers permit omnidirectional movement, transmitted four-bit outputs wirelessly to steer along programmed trajectories. A half-adder performed binary addition by pairing modules of identical thickness in parallel, then stacking two such pairs in series.
Reconfiguring the system required no hardware changes. Altering the stacking order or rewiring contacts modified transition rules directly. The car traced entirely different paths simply by adjusting the control unit’s electromechanical configuration.
The authors acknowledge practical constraints. Bit counts above four strain the design because beam thicknesses must fall within a range ensuring robust bistability while maintaining sufficient stiffness gradients across stacked layers.
Yet the demonstrated principles point toward broader applications. Tunable equilibrium angles spanning 40° to 120° suggest that higher information density remains achievable with refined geometries. By exploiting rotational input, a modality largely unexplored in prior mechanical computing, the design escapes the stroke-length constraints of linear systems. Integration with gear clusters could enable distributed or parallel electromechanical computing. The authors note potential applications in microelectromechanical systems, where information storage and processing could merge within compact devices.
The core achievements, combining sequential state evolution with reprogrammable transition rules and concurrent logic, establish a foundation for intelligent mechanical systems capable of adapting to changing conditions, bringing embedded computation to settings that demand alternatives to conventional electronics.
For authors and communications departmentsclick to open
Lay summary
Prefilled posts
Nanowerk Newsletter
Get our Nanotechnology Spotlight updates to your inbox!
Thank you!
You have successfully joined our subscriber list.
Become a Spotlight guest author! Join our large and growing group of guest contributors. Have you just published a scientific paper or have other exciting developments to share with the nanotechnology community? Here is how to publish on nanowerk.com.
