Harnessing molecular machines at the macro-scale

by Dr. M. Ravi Shankar, Professor of Industrial Engineering at University of Pittsburgh 

The natural world excels at utilizing hierarchies ranging from the molecular, macromolecular-scales through to the micro and macrostructural scales to elicit emergent responses critical for sustaining life. Particularly impressive is the ability to harness myofilamental molecular engines in animal muscles to seamlessly assimilate controlled responses. Emulating such structural hierarchies in active matter to harness molecular-scale organization to power active responses has become possible via breakthroughs in molecular switches, self-assembly mediated processing and their incorporation in actuators. As a result, new opportunities emerge for utilizing ambient energy source, including light, heat and chemical potential to direct actuation and manipulation. Fundamentally new classes of soft robotic platforms become possible, which have few parallels in the classical approaches for manipulating and actuating mechanical systems.

Actuation from Order-Disorder Transitions in Soft Matter

A powerful tool for building molecular-scale hierarchies emerges from breakthroughs in the patterning of responsive liquid crystalline polymers 1. Liquid crystals in the traditional sense refer to a phase of matter characterized by orientational order of constituent molecules. Typically, liquid crystalline molecules are characterized by rigid (usually aromatic group-based) segments to which flexible tails (aliphatic/ether-based) are connected. The coupling between the rigid segments leads to the creation of orientationally ordered states. The free energy (F) of this system is defined by the order parameter of the system  equ , where2                 


The liquid crystalline phase is typically stable below a certain nematic temperature, where minimizing the free energy (F) leads to a finite S>0. When the order parameter S becomes 0, the orientational order is destroyed and a disordered (isotropic phase) results. Equation 1 represents the traditional Landau-De Gennes free energy formulation, where  2 , where  3 is a suitable reference temperature. Minimization of free energy with increasing temperature leads to a progressively declining S, which then suddenly drops to 0 to represent the nematic-isotropic transition. Imagine, if the liquid crystals were not constituting a liquid, but were characterized by the flexible ends connected to one another. That is, the flexible spacers, which are connected on one side to the rigid segments, were functionalized to create polymerizable monomers at their other end. An array of functionalizations can be envisaged, including acrylates, -enes, epoxides etc. In such a system, order-disorder transitions can lead to surprisingly potent mechanical responses 3. Figure 1 illustrates the typical modality of strain generation due to the order disorder transition in liquid crystalline polymers (LCP). 

  Strain Generation

Figure 1: Strain generation from Order-Disorder transition in LCP. Adapted from Ref. 3

In the example in Figure 1, in the low-temperature state, the nominal direction of the rigid rod-like molecules is along the vertical axis. This orientation of the molecular axis (also known as the molecular director n) represents the principal direction of the contractile strain generation when the order to disorder transition is driven. Perpendicular to the nematic director, the strains are tensile. Thus, molecular order in the nematic state defines the principal directions of the strain generation in the activated state.

Most importantly, the order-disorder transition is fully reversible, which means that the actuation that results is highly repeatable. Applying an actinic stimulus leads to strain generation, which is fully relaxed when the stimulus is removed. This sets LCP apart from a range of other active or shape memory polymers, which are usually just one-way (e.g. shrink-wrap polymers or Shrinky-dinks). Energizing with a stimulus (say, heat) leads to actuation and strain generation, while removal of the stimulus leads to spontaneous relaxation. Optimizing the material structure in LCP has resulted in fully reversible thermomechanical work-densities of ~1J/cm3 (actuation strain ~100% and actuation stress ~1MPa) 4. Furthermore, LCP material designs report forces >>2000X actuator’s weight in µm and mm-scale formfactors 5-6.

Molecular Architectures for Magnified Actuation Profiles

Consider, a heterogeneous distribution of the molecular directors in a macro-scale flat sample akin to that illustrated in Figure 2A. Here, the nematic director is not aligned uniaxially (e.g. Figure 1), but instead is aligned azimuthally in concentric rings around a core. This corresponds to a +1 topological defect. Actuating this material leads to the shrinking of the material along the nematic director and expansion perpendicular to it. The only way this deformation can be accommodated is by the prior flat sheet spontaneously transforming into a conical geometry. In Figure 2B a 3x3 array of such defects is shown to transform into cones, when heated. During such actuation, the material can be thought to concentrate the work generation at the center of the actuator. As a result, large force generation results in an actuator that can lift ~100’s of times its own weight (Figures 2C and 2D). More noteworthy is that the actuator starts in a flat form-factor and produces a stroke length several 100% its original dimension. This unusual combination of force-displacement characteristics emerges from the combination of the responsiveness of the molecular order to temperature and the organization of the nematic directors in macroscopic objects.



Figure 2: Creating topologically defected (+1 azimuthal) nematic patterns in ~100micrometer films results in unusual actuation profiles with magnified work content. 6

The idea in Figure 2 was further developed by creating multilayered composites of such actuators. The hierarchical composite structure was shown to lift several thousand times its own weight to result in Force-Displacement characteristics, which outstrip competing microactuator technologies by several orders of magnitude 5. Figure 3 illustrates the displacement response of 1 (50 µm), 2 (2 x 50µm) and 4 (4 x 50µm) thick actuator laminates against a superposed blocking force, which is several thousand times the weight of the actuator. 


Figure 3: Force-Displacement Characteristics of multilayered +1 topological defected LCP actuators. Adapted from Ref. 5

Fabricating Molecularly-Ordered Macro-scale Structures

Key to the practical utilization of LCP in soft-robotics, biomedical devices and functional manipulation systems is the ability to fabricate them into desired geometries, while simultaneously imprinting molecular order. The fabrication of flat LCP samples (e.g. Figure 2) has borrowed from ideas in the liquid crystalline display industry. These include, patterning of polymerization cell walls with microtextures or by spin-coating light-responsive dye molecules 1. Utilizing polarized light, pixel-by-pixel, to orient the dye molecules provides the template, which is inherited by the liquid crystalline monomers, when they are introduced into the cell. Here, the ability to self-assemble molecular patterns in the bulk (including highly heterogeneous ones, e.g. Figure 2) from blueprints assembled on the walls of the polymerization cells provides a powerful route to create molecularly-ordered active matter. This eliminates the need to train the material using mechanical deformation, as is the case with conventional shape memory alloys and polymers. While, this approach is restricted to flat films or at best 2.5D geometries, breakthroughs in 3D printing macroscopic free-forms with tailored molecular order presents a remarkably expanded design space. By exploiting a two-step polymerization system, oligomeric inks synthesized from mesogenic liquid crystalline monomers are extruded through a nozzle as illustrated in Figure 4 7. Controlling the raster pattern during the creation of 3D geometries provides the framework for creating macroscopic molecularly-patterned actuators. 


Figure 4: a) Extrusion of oligomeric inks leads to the generation of molecularly ordered strands, which are deposited and cross-linked in situ using light. b) The chemical composition is characterized by liquid crystal monomer, which undergoes a Michel addition reaction with a chain extender. c) Controlling the viscosity and shear rates during extrusion holds the key to modulating the molecular order which is generated. d) Tuning the raster pattern during the material deposition provides the framework for constructing macroscopic structures, where molecular order is defined during the build pattern. e) The optical birefringence due to the molecular order is seen through the polarized optical microscopy images. These molecular-ordered structures are capable of large strain actuation via order-disorder transitions as a result of a stimulus (e.g. heat) 7.

 Light-Powered Machines

An unusual approach for powering actuation is to directly harness photons to power mechanical actuation 8. Light is distinguished as an actinic power source because it can be radiated from large distances, be modulated spatiotemporally in intensity and polarization and can be harvested from ambient sources. Light is also a biologically benign stimulus, which allows for it to be utilized in medical applications, where it is delivered in vivo using optical fibers to power functional manipulation and deployment of devices. The premise of this idea is as follows – consider a light responsive molecular switch, which is incorporated in an LCP matrix as illustrated in Figure 5a. When this molecular switch absorbs a photon of light, it undergoes a photochemical isomerization to transform from a rod-shaped molecule to a bent molecule (Figure 5b). As a result, this perturbs the molecular order in the surrounding matrix leading to a reduction in the molecular order in the overall matrix (Q


Figure 5: a) Photon is absorbed by a photochromic switch embedded in a molecularly-ordered LCP with a prior order parameter Qo. b) Isomerization leads to transformation of the rod-shaped molecule into a bent molecule. c) The bent molecule generates disorder in the LCP: Q

It has been shown that the interplay of such photochemical transformation in molecular order and the geometry of actuator can be harnessed to generate power-dense actuation (~kW/m3) at intensities of irradiation that approach that of laser pointers (~10mW/cm2). In Ref. 9, the mechanical instability observed in the Venus fly-trap was harnessed to drive photomechanical snap-through actuation. This actuation is highly repetitive and can allow for creating novel light-driven actuators and robotic platforms 10. The basic idea is to design, buckled arch-shaped structures using light responsive LCP as illustrated in Figure 6. Then, exploiting the light-driven strain generation and the gradients in the strain due to the graded absorption of light through the thickness, the actuator is advanced to the edge of instability. Subsequently, snap-through results at the speed-of-sound (~10ms) 9


Figure 6: Photomechanical actuation in bistable arches. a) original arch illuminated from the bottom that undergoes gradual deformation to reach b) the primed geometry that spontaneously undergoes snap through to create c) the upward curved arch

This idea was taken further in a range of other multistable geometries, including in bifurcated arches, which can be toggled between discrete states using irradiation with light (Figure 7). The idea is to create a laterally bifurcated geometry, which can be transformed from one state to another by simply changing the position at which light is irradiated. Creating arrays of such actuators offers a framework for creating morphing surfaces, micromirror arrays, programmable topographies, all of which are exclusively controlled using light. Furthermore, these actuators require no other microfabricated elements or composite construction. The material responds directly to light to  generate mechanical actuation.       


Figure 7: Actuation in a laterally bifurcated arch, which is driven to repetitive actuation using light 10. 

Beyond the bistable actuators, it has been shown that light can be used to power microrobotic systems, without encumbering them with any additional mechanisms. Figure 8 illustrates the idea of photomotility where a coiled structure, characterized by a chiral molecular orientation in the LCP spontaneously rolls along a surface in a deterministic fashion 11.


Figure 8: Photomotility in a twisted coil, chiral LCP in response to irradiation from a lamp 11.

If one spiral can move under the influence of light, can multiple spirals be used to create a composite response, where a steerable photon-powered robot can be envisioned? The feasibility of steerable photomotility was illustrated, where a light-driven robotic architecture moved in a programmable on a surface with random topography (Figure 9) 12. Again, no other mechanisms or multimaterial composition was required to achieve such robotic platforms – just the photomotile LCP arranged in a designed configuration was sufficient. 


Figure 9: Composite geometries created from spiral LCP actuators, composed of a chiral molecular order were used to power motility in microrobots using light. a) Curvilinear trajectory. b) travel along the long-axis of the light-driven microbot and c) trajectory transverse to the axis of the microbot 12.


The emergence of molecularly-ordered material systems, which can be blueprinted with response profiles at fine length-scales along with advances in processing them into complex geometries has the potential to enable fundamentally new classes of soft machines. By leveraging biomimetic and engineered designs that exploit mechanical non-linearities opportunities also emerge for creating soft robotic manipulation systems with unprecedented response profiles. These include actuation speeds of several 100mm/s, work-densities approaching J/cm3and power-densities of ~kW/m3. These metrics can outstrip that possible with conventional monolithic active materials, especially at mm and sub-mm length-scales. The maturation of the material systems – e.g. improved photoisomerizable switches with long time-scale bistability, reliable micro and nanofabrication with molecularly-ordered soft matter have the potential to enable fundamentally new classes of light-powered microsystems – Micro-Opto-Mechanical Machines. Current research trajectories also point to the utility of LCP in robotics – macroscopic robotics and human-assistive systems and microrobotic systems for search/rescue and condition monitoring in confined spaces. The ability to build in increasing levels of autonomy on the material-scale itself can enable path-breaking platforms where self-regulation and control during actuation is encoded in the material itself. Thus, rendering obsolete the need for external control electronics, manipulation planning and electronic feedback control in robotic systems.


1.             White, T. J.; Broer, D. J., Programmable and adaptive mechanics with liquid crystal polymer networks and elastomers. Nature materials 2015, 14 (11), 1087.

2.             Prost, J., The physics of liquid crystals. Oxford university press: 1995; Vol. 83.

3.             De Jeu, W. H., Liquid crystal elastomers: materials and applications. Springer: 2012; Vol. 250.

4.             Kim, H.; Boothby, J. M.; Ramachandran, S.; Lee, C. D.; Ware, T. H., Tough, shape-changing materials: crystallized liquid crystal elastomers. Macromolecules 2017, 50 (11), 4267-4275.

5.             Guin, T.; Settle, M. J.; Kowalski, B. A.; Auguste, A. D.; Beblo, R. V.; Reich, G. W.; White, T. J., Layered liquid crystal elastomer actuators. Nature communications 2018, 9 (1), 2531.

6.             Ware, T. H.; McConney, M. E.; Wie, J. J.; Tondiglia, V. P.; White, T. J., Voxelated liquid crystal elastomers. Science 2015, 347 (6225), 982-984.

7.             Ambulo, C. P.; Burroughs, J. J.; Boothby, J. M.; Kim, H.; Shankar, M. R.; Ware, T. H., Four-dimensional printing of liquid crystal elastomers. ACS applied materials & interfaces 2017,9 (42), 37332-37339.

8.             Lee, K. M.; Tabiryan, N. V.; Bunning, T. J.; White, T. J., Photomechanical mechanism and structure-property considerations in the generation of photomechanical work in glassy, azobenzene liquid crystal polymer networks. Journal of Materials Chemistry 2012, 22 (2), 691-698.

9.             Shankar, M. R.; Smith, M. L.; Tondiglia, V. P.; Lee, K. M.; McConney, M. E.; Wang, D. H.; Tan, L.-S.; White, T. J., Contactless, photoinitiated snap-through in azobenzene-functionalized polymers. Proceedings of the National Academy of Sciences 2013, 110 (47), 18792-18797.

10.          Skandani, A. A.; Chatterjee, S.; Smith, M. L.; Baranski, J.; Wang, D. H.; Tan, L.-S.; White, T. J.; Shankar, M. R., Discrete-state photomechanical actuators. Extreme Mechanics Letters 2016,9, 45-54.

11.          Wie, J. J.; Shankar, M. R.; White, T. J., Photomotility of polymers. Nature communications 2016, 7, 13260.

12.          Babaei, M.; Clement, J. A.; Dayal, K.; Shankar, M. R., Steering with light: indexable photomotility in liquid crystalline polymers. RSC Advances 2017, 7 (83), 52510-52516.