Biological, physical and morphological factors for the programming of a novel microbial hygromorphic material

To optimise growth conditions, B. subtilis strain 168 vegetative cells were grown, subsequently sporulated, and evaluated for final spore concentration, culture volume yield, and magnitude of deflection (as a measure of hygromorphic response) under a range of protocols. This included the impact of nutrient availability, aeration, Fe²⁺ availability, washing protocol and culture volume.

This led to the development of an optimised sporulation protocol. During a two-phase growth period, the vegetative cells were provided with abundant aeration and essential cations such as Fe²⁺ to support development of peptidoglycan and the complex chemical morphology of some specific coat proteins responsible for the mechanics of the spore's adhesion properties. After incubating in 10 ml Luria-Bertani (LB) medium (Sigma) at 37 °C in a shaking incubator at 200 rpm for 12 h, the culture reached early exponential growth (achieving an OD_600nm of 0.8 (Abhyankar et al 2011)), and was transferred to a larger volume (90 ml to make 100 ml culture) of LB culture medium with the same chemical composition. This secondary, 2–3 h incubation in the larger volume culture promoted a high yield of actively dividing bacteria without limiting cell division through nutrient limitation. The cultures were then washed with phosphate buffered saline (Sigma) to remove any remaining LB media through a centrifuge–discard–resuspend method. A third and final 800 ml culture was used to initiate sporulation through restriction of available nutrients within the specialist sporulation media (Difco sporulation media adjusted to a pH of 7.6) (Sigma). The sporulation process was monitored through Schaeffer–Fulton staining (Sigma) every day from day 3 by heat fixing a 1 ml sample to a microscope slide and dyeing with Malachite green (step one of Shaeffer–Fulton staining method) slowly over 15 min. The slide was then counter stained with saffron (step two of Shaeffer–Fulton staining method) and light microscopy was utilised to visualise this sporulation progress.

Production of bacterial spore HBCs provides opportunities to 'design in' control parameters to programme computational responses from the material beyond that achieved by programming the biological factors through optimising sporulation protocols. A previous study (Birch et al 2021) demonstrated programmability of deflection magnitude through increasing and decreasing the spore concentration used to produce the HBCs. To build from this within this paper the potential for programming through physical factors was investigated:

• Inert substrate: the impact of substrate type on magnitude and repeatability of deflection.
• Material thickness: the potential to preprogramme a deflection magnitude through substrate material thickness.
• Etching: the potential to control deflection form (etching direction and effect of localised etching investigated).

These physical factors were investigated by producing a material pallet of bacterial spore HBCs, each exploring a different design factor (table 1).

All substrates were cut with a standard laser cutter (with a cutting resolution of 0.2 mm). The exposed edges were dabbed with blotting paper to remove any excess molten material and cleaned thoroughly with ethanol (Sigma). Once clean, the effect of adding a biological glue to aid spore adhesion was investigated by treating some samples with 8 μl cm⁻² of poly-L-lysine (Sigma) and drying for 3 min until the glue layer became tacky before the spore layer was applied. The impact of substrate etching (transverse, longitudinal and diagonal etch lines, and the effect of localised etching) on deflection programming was also incorporated into the HBC material pallet.

The introduction of etching has enabled us to investigate the potential to produce fold versus bend deflections. In the first study two identical hexagonal stars were laser cut from 0.4 mm latex, with one being fabricated with 16.6 μl spore solution spread uniformly across each of the six star triangles (1 cm isosceles) and the second hexagonal star being laser etched across each of six fold regions. The same spore volume was applied discretely along the etching line. In the second study a simple 2D cube net (square face, 2 × 2 cm) was laser cut from 0.4 mm latex, etched at each of the 'fold' vertices and 16.6 μl of spore solution applied discretely along the etching lines.

The effect of spore concentration on deflection was also included in the range of samples prepared for the material pallet (volumes ranged from 4.2 to 33.2 μl cm⁻²). Samples were placed on a rocking shaker to ensure no pooling occurred during the drying process, after which, samples were incubated at 99% RH for 24 h to equilibrate the deflection response programmed into the material through its fabrication protocol (figure 2).

Zoom In Zoom Out Reset image size

A new methodology for measuring deflection angle was also developed to allow post analysis of sample photographs. A small, single sample rig was developed to ensure exact placement of each sample to produce photographs, all with the same set reference point (taken using Canon EOS 1300D, with electro focus short (EFS) 18–55 mm and Canon compact macro lens electro focus (EF) 50 mm, 1:2.5 with image stabiliser control camera, set on a fixed tripod) for analysis. Using Adobe Illustrator these results photographs were post analysed and used to map the deflection arc of each side of the HBC sample by producing a circle. These circles' sizes were determined by overlaying onto the photograph, following the deflection arc of the HBC as closely as possible. The centre point of these circles identified the centre of the defection arc. The start point of the deflection was defined as the edge of the spore deposition area on the latex and the end point of deflection was defined as the point of the end of the actuator. The angle between these two lines produced a segment from which the deflection angle could be measured. This geometric approach to data collection provided an accurate deflection angle (taken from segment produced) for each side of the HBCs. The relationship between substrate thickness and deflection angle achieved at 42% RH was presented graphically using Excel and analysed statistically using linear regression.

Exploration of the physical factors produced a material palette which provided the data to select fabrication combinations required to produce HBCs which would deform at a given rate and a given magnitude across a given range of environmental RH. This provided the platform for the next stage of development which focused on programming morphological factors. Morphology here conveys a meaning of 'functional shape change', with deflection (shape change) of the HBCs creating an opening aperture which functions to increase the exchange area for air flow as a passively responsive ventilation panel. A 3D printed (Ender 3 Pro) modular casement system was designed within Rhino® to allow clamping of six HBC actuators within one module. This modular system used a stud and socket connection design, producing a 'Lego' style aggregation to harness and scale the HBC palette into programmable humidity responsive apertures. This framework could house a range of different samples for testing (figure 3).

Zoom In Zoom Out Reset image size

To measure the open area of each aperture they were incubated at 99% RH for 30 min before being removed from the humidity chamber and exposed to room humidity (42%) and filmed (using Canon EOS 1300D, with EFS 18–55 mm and Canon compact macro lens EF 50 mm, 1:2.5 with image stabiliser control camera on a tripod) for a further 30 min. Stills were taken every 3 min from this recording and analysed. The 'Image Trace' tool within Adobe Illustrator was used to isolate the open area within each aperture, then using the 'Dynamic Measure' plug-in a quantifiable value of aperture area at a range of RH could be obtained. This method was repeated three times to provide a triplicate data set. This provided understanding of each apertures' open area range across its environmental humidity range.

Performance testing of these apertures focused on assessing the HBC's potential for application within the built environment as a passive trickle ventilation membrane. To explore this, a new casement system was developed to house one type of HBC, producing a benchtop prototype panel. These benchtop prototype panels consisted of a 3D printed polylactic acid (PLA) honeycomb lattice casement system, hosting 27 discrete apertures, formed from six actuators of 0.4 mm latex and 8.3 μl cm⁻² of B. subtilis spore solution per aperture. This HBC type was selected from the materials pallet due to having the greatest range in open area and speed of active shape change.

This panel of 27 apertures was subjected to the same methodology as with the individual aperture investigations, however, to capture the aggregated response all of apertures open area range were recorded and combined to give a maximum and minimum open area for the panel as a whole system.

The humidity test chamber (figure 4) was designed which facilitated the repeatable creation of artificial environments to explore the HBC benchtop prototype behaviour and performance at regulating airflow between two contrasting RH environments. This was achieved by subdividing a glass container to provide two adjacent controlled humidity environments. They were separated by the test panel which was airtight excepting airflow through the open aperture areas. The temperature and humidity of each side could be monitored using separate surface mounted humidity monitors (EEEkit digital hygrometer). The humidity of the air entering the first chamber was controlled through use of an air flow pump connected to an air stone within a water bath providing a >99% RH airstream. The second adjacent chamber contained dry air which was achieved by using anhydrous calcium sulphate as a desiccant (Drierite) to provide a contrasting minimal RH environment.

Zoom In Zoom Out Reset image size

Firstly, the biological factors at the cellular and ultracellular level were explored through the optimisation of growth and sporulation conditions to biologically preprogramme optimum spore hygromorphic response and yield. Previous growth and sporulation methodologies to produce B. subtilis spores for use as hygromorphs (Chen and Mahadevan 2013, Yao et al 2015, Ramirez-Figueroa et al 2016) have utilised a 'standard' LB culture methodology which does not target optimisation of hygromorphic capabilities of the spores. Within this paper, optimised protocol was developed and hygromorphic capabilities of enhanced spores was investigated (figure 6). The results show that the optimisation of the B. subtilis vegetative cell growth and sporulation protocols with 3× the hygromorphic response (70.7° compared to 220° for the optimised protocol) observed through HBC deflection (for comparable HBC fabrication) (figure 7) and produced a 40 fold increase in spore yield per batch (8 ml) in comparison to spores produced in the previous studies (0.2 ml) using a standard LB-based media culture and Difco sporulation media (DSM) sporulation protocol (Birch et al 2021). The authors are not aware of other studies where magnitude of spore specific hygromorphic response has been evaluated. This increase in hygromorphic response is likely to reflect an increase in peptidoglycan content of the cortex layer of the spores in optimised conditions as proxy studies have suggested the hygromorphic expansion force is generated by binding of molecular water in the porous structure of this macromolecule (Piotrowska et al 2021). The optimised sporulation protocol provided an increase in aeration and nutrient supply which also resulted in an increase in spore yield. This reflects similar findings by Elisashvili et al, which were attributed to enhanced reproduction rate of vegetative bacteria with higher nutrient availability, and a higher sporulation % in aerobic conditions (2019).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

After the optimisation of biological elements, material explorations were followed, to investigate how physical factors influence biomechanics, to preprogramme shape changes and response complexity through fabrication and inert substrate interactions, and to produce a palette of HBCs.

3.2.1. Substrate material selection

Three flexible materials were investigated as potential substrates to produce HBC actuators: latex, silicone, and thermoplastic polyurethane (TPU). Latex has previously been used with poly-L-lysine to successfully to produce Bacillus HBCs (Birch et al 2021). This study has further demonstrated that without poly-L-lysine there is delamination between the spore layer and latex after ten deflection cycles (figures 8(d) and (e)). However, latex deteriorates when exposed to UV light and has limited potential for 3D fabrication so explorations into alternatives were explored. Silicon is a synthetic alternative to latex, and offers options for more complex fabrication methods, while TPU has the advantage of being readily fabricated using advanced 3D-printing technologies. Both alternatives were investigated with the use of poly-L-lysine. The results demonstrated that HBCs fabricated with TPU showed no deflection (figure 8(c)), while those fabricated with silicon showed a deflection of a similar magnitude to that of latex initially (figure 8(a)), but the bio-composite was not as resilient and failed after the third hydration–dehydration cycle (figure 8(b)). Recent work has suggested some success with a silicone adhesive which may offer opportunities to work with silicone in future studies (Wang et al 2022).

Zoom In Zoom Out Reset image size

3.2.2. Substrate thickness

Building upon the spore concentration findings from the previous work (Birch et al 2021), this study has determined that increasing the thickness of latex used to form a HBC decreases the deflection angle achieved (figure 9, represented graphically in figure 10) and through linear regression it has been demonstrated there is a direct correlation between the two variables with R² = 0.9968 and as such, a desired deflection angle may be preprogramed into the material at fabrication through this parameter. It also promotes the concept that the magnitude of deflection is programmed by the actuation stresses generated by the differential expansion of the two layers of the bilayer, since, an earlier study (Birch et al 2021) demonstrated that a higher concentration of spores in the bilayer fabrication resulted in an increase in magnitude of deformation.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.2.3. Material etching

Etching investigations yielded promising results with repeatable deflection magnitude and direction achieved. Etching the top surface of the latex prior to coating with biological glue and B. subtilis spore solution resulted in the sample deflecting perpendicularly to the etch direction and consequently a twist, curl and roll deflection may consistently be preprogramed during fabrication (figure 11).

Zoom In Zoom Out Reset image size

This finding supports recent studies (Zuo et al 2023), where they were investigating directional deformation with hydrogel bilayers as soft actuators. They also adopted a laser etching methodology to create patterned surface markings, which resulted in similar twist, curl, and roll deflections. These findings together suggest laser etching reduces bending stiffness of the latex in one direction resulting in effective preferential bending, thereby programming the deformation. Localised etching and spore deposition resulted in a fold rather than curve deflection (figure 12).

Zoom In Zoom Out Reset image size

This folding response was also demonstrated in the simple self-assembly folding pilot study. It was observed that a 3D open-topped cube could be successfully self-assembled from a 2D net at room humidity and returned to its two-dimensional form in a high humidity environment (figure 13) mimicking some of the much more refined hydrogel self-assembly demonstrations by Tibbits (2014). The extent of the fold angle was approximately 90°–100° and lower than that seen with the hexagon star study where folding reached a 180° angle. However, the folding of the faces of the self-assembly cube were restricted because they overlapped with other faces.

Zoom In Zoom Out Reset image size

Latex thickness showed great potential to preprogramme an environment specific response within the aggregated system. Aperture size increased at a given RH as latex thickness used to fabricate the HBC decreased. These panel apertures were monitored over a 30 min period ranging from 99% RH to 42% and reached their maximum response within a 15 min active zone (figure 14). This fast response time of these spore-based HBC prototypes appears faster than the highly developed wood-based 3D printed apertures which respond within 35 min (Tahouni et al 2023). It is likely the increase in response speed is due to a short diffusion distance into uniform microscopic spores.

Zoom In Zoom Out Reset image size

Image trace methods provided accurate mapping and evaluation of aperture kinematics, providing a detailed understanding of range of motion and activation points. It was observed that an 85% increase in aperture open area can be achieved when transferred between 99% and 42% RH. When interpreting these results, there is notable consistency between the 27 replicate apertures that combine to form the prototype panel, and repeatability of kinematic response over three repeated hydration–dehydration cycles showed low variation as seen from the error bars in figure 15.

Zoom In Zoom Out Reset image size

Initial performance testing of these benchtop prototypes within the experimental test chamber highlighted a dual action response of the HBC apertures which contribute to their impact on RH of the chambers. Firstly, as expected, the apertures close when exposed to high humidity environments, reducing the open area and rate of air diffusion across between the chambers. This maintained the humidity level in the chamber for longer within chamber B than the control because movement of humid air was reduced. Secondly, this study highlighted the bacterial spores' capabilities as a humidity sink. It was observed that as humid air passed through the HBC apertures, moisture was removed from the air. While spore water content was not measured directly, it is likely this moisture was retained by the spores. A humidity sink response was proposed with a simulated model using hydrogel (Aviv et al 2020) and this could suggest B. subtilis spores have a similar humidity sink capacity and are able to reduce RH of a high humidity environment. Conversely, this stored moisture could be re-released into the environment at low RH increasing the humidity. The concept of a hygromorphic material acting as a humidity sink and demonstrating the impact of material–environment-coupling on humidity gradients has also been demonstrated (Treml et al 2018).