Marimo machines: oscillators, biosensors and actuators

Background The green algae balls (Aegagropila linnaei), known as Marimo, are large spherical colonies of live photosynthetic filaments, formed by rolling water currents in freshwater lakes. Photosynthesis therein produces gas bubbles that can attach to the Marimo, consequently changing its buoyancy. This property allows them to float in the presence of light and sink in its absence. Results We demonstrate that this ability can be harnessed to make actuators, biosensors and bioprocessors (oscillator, logic gates). Factors affecting Marimo movement have been studied to enable the design, construction and testing of working prototypes. Conclusions A novel actuator design is reported, incorporating an enhanced bubble retention system and the design and optimisation of a bio-oscillator is demonstrated. A range of logic gates (or, and, nor, nand, xor) implementable with Marimo have been proposed. Electronic supplementary material The online version of this article (10.1186/s13036-019-0200-5) contains supplementary material, which is available to authorized users.


S2 Rotor balance rig
A bespoke rig was built to balance the rotor of Marimo motor. The dimensions of the (collapsible) frame were: base 250 mm wide x 750 mm length and support arms 400 mm tall with adjustable width 0 to 500 mm. The frame was constructed from anodised aluminium section (40 mm × 40 mm) with integrated 'T' slots. The sections were bolted together. Rubber feet were added to four (folding) legs. Four neodymium magnets (grade N35) were stacked on both sides forming magnets of 30 mm diameter and 40 mm thickness. One magnet was mounted on a screw thread to allow precise adjustment of the gap between the magnets, see Fig. S2. Figure S2: Rotor balance rig The motor frame could not support the rotor out of the water and balancing in the water was impractical. Consequently, a balancing and operating rig were required. The balance rig required stronger magnets than the motor frame as rotor weight was not reduced by positive buoyance (in water). The rotor was balanced by manually adding weight to the acrylic disc until the out-of-balance was less than 50 g mm.

S3 Marimo Motor power output calculations
When illuminated, the Marimo photosynthesise and produce oxygen bubbles. These bubbles, due to their density being lower than that of water's, rise and become trapped against the inside surface of the individual cylinder containing the Marimo. When the cylinder reaches to the top of its rotation the bubbles are able to escape, permitting rotation of the motor to continue. During rotation, the bubbles in each cylinder rise by 250 mm.
The average volume of gas generated by a Marimo is approximately 3.5 cm 3 d −1 . With a rotational speed of 0.2 rev h −1 , half a rotation takes 2.5 h and 0.36 cm 3 of gas is generated. To calculate the lift, the mass of the (rising) gas bubbles is subtracted from the mass of the water 'sinking' (becoming displaced): (1.00 g cm −3 × 0.36 cm 3 ) − (0.001 g cm −3 × 0.36 cm 3 ) = 0.36 g. (1) Potential energy (PE) can be expressed as PE = mgh, where m is mass, g is acceleration due to gravity, and h is the distance displaced (or height). Therefore, PE = mgh = 3.6 × 10 −4 kg × 9.81 m s −2 × 0.25 m = 0.9 mJ.
The resulting value is based on a single Marimo completing one revolution, therefore the energy can be more properly displayed as 0.9 mJ Marimo −1 rev −1 , when rotating at 0.2 rev h −1 .
S2 Table S1: The weights of floats tested for use with Marimo, both before and after drilling and before and after submersion in water for 72 h. The diameter of all spheres was 25 mm. High density polyethylene (HDPE) and polypropylene (PP) were tested. All measurements were taken on an analytical balance.

S6 Illumination levels
A range of lamps were used during experimentation. Illumination against distance for each type of lamp is shown in table S3. Photosynthetically Active Radiation (PAR) light sensor (Campbell Scientific Ltd, model SQ-120, measurement repeatability: < 1 %, linear response: 5 µmol m −2 s −1 mV −1 ) was used to measure illumination levels. The output voltage of the sensor was measured with a Fluke 8846A precision multimeter. Photosynthetic Photon Flux Density (PPFD) rather than lux was recorded, as the PAR sensor measures how many photons (within the portion of the spectrum useful for photosynthesis) are striking the surface. Every photon is counted the same. By contrast, a lux sensor measures light in the visible spectrum and weights the light best seen by human eyes. Lux sensors tend to measure lower than PAR/PPFD sensors. Conversion of PPFD (µmol m −2 s −1 ) to lux is achieved using the manufacturers instructions.
The Marimo oscillator was tested using a Martin Rush PAR 1 70W RGBW Cree LED Lamp. Beam angle 20 degrees. Independent & variable power on primary colours (red, blue & green): 0 to 255 (0 to 100 %). The logic testing used a bespoke lamp. 10 W LED array (660 nm) was mounted on fan cooled heatsink, see Fig. S3(a). The optics were adapted from a LEDSPOT-3WW spot lamp (Pulse Audio Ltd) and configured to work with LED array and heatsink. A beam focus of 4 degrees was achieved, see Fig. S3(b). Motor testing used variable power, 300W, full spectrum, plant growth lamp, brand name LAPUTA, model number '60LEDS grow light'. The power output of all lamps can be seen in table S3.