Sensing surface morphology of biofibers by decorating spider silk and cellulosic filaments with nematic microdroplets

Significance

Biological microfibers are remarkable materials with diverse structural and mechanical properties, such as high wear-resistance, elasticity, and biodegradability. However, with current techniques, there are few robust ways to sense the surface properties of the fibers, which crucially affect the organization of the fibers and their interactions with the surrounding material. In this paper, we show that surfaces of diverse biofibers, including spider silks and cellulosic fibers, can be easily sensed by depositing droplets of a nematic fluid onto the fibers. The droplets reveal the surface properties of the fibers via their optical images, notably showing also the fiber chirality. Further, the droplets are used to study the entanglement of biofibers, as a route toward novel biological and bioinspired materials.

Abstract

Probing the surface morphology of microthin fibers such as naturally occurring biofibers is essential for understanding their structural properties, biological function, and mechanical performance. The state-of-the-art methods for studying the surfaces of biofibers are atomic force microscopy imaging and scanning electron microscopy, which well characterize surface geometry of the fibers but provide little information on the local interaction potential of the fibers with the surrounding material. In contrast, complex nematic fluids respond very well to external fields and change their optical properties upon such stimuli. Here we demonstrate that liquid crystal droplets deposited on microthin biofibers—including spider silk and cellulosic fibers—reveal characteristics of the fibers’ surface, performing as simple but sensitive surface sensors. By combining experiments and numerical modeling, different types of fibers are identified through the fiber-to-nematic droplet interactions, including perpendicular and axial or helicoidal planar molecular alignment. Spider silks align nematic molecules parallel to fibers or perpendicular to them, whereas cellulose aligns the molecules unidirectionally or helicoidally along the fibers, indicating notably different surface interactions. The nematic droplets as sensors thus directly reveal chirality of cellulosic fibers. Different fiber entanglements can be identified by depositing droplets exactly at the fiber crossings. More generally, the presented method can be used as a simple but powerful approach for probing the surface properties of small-size bioobjects, opening a route to their precise characterization.

Natural microfilaments produced by plants, insects, or spiders are fascinating materials not just because of their specific properties such as wear resistance, elasticity, tensile strength, and toughness (1–5) but also because of their microorganization (6–9). Their macroscopic properties can match properties of materials like kevlar but are at the same time biocompatible and biodegradable (10). These fascinating macroscopic properties actually originate from bulk and surface properties of the fibers (1). The chemical composition of the threads combined with their morphology determines the final properties of the material (11–13). The mechanical properties of the spider fibers are determined by the existence of a lyotropic liquid crystalline phase, from which the threads are drawn (14). Such silks are known to include nanoscale networks of defects and cavities that yield surface structures notably dependent on the spider species (3). These differences do not affect much the mechanical performance of the fibers (1, 3, 5). From a technological perspective, many attempts have been made to reproduce these natural bionetworks (15–17). In fact cellulose-based fibers with few micrometers of diameter, produced by electrospinning, can also acquire different morphologies depending upon the processing conditions, giving diverse features of the final threads and mats (18). Therefore, probing the surface structure of the microfibers is crucial for a complete understanding of their individual and interthreaded properties.

From another perspective, nematic complex fluids are materials which are inherently responsive to diverse external stimuli, notably including diverse surface interactions which in the literature are known as the surface anchoring (19). Being effectively elastic materials, the orientational order of nematics responds on long, typically micrometer scales (20–22), which results in a spatially varying birefringence that can be optically detected (23). Recently, it was demonstrated that glass fibers induce numerous defects in a well-aligned nematic liquid crystal cell and thus provide a simple illustration of topological phenomena (24). It is also known that liquid crystal droplets can considerably change their structure by the action of otherwise imperceptibly small external stimuli (21). Pierced nematic and chiral nematic droplets develop defects that can be controlled by the liquid crystal elasticity, chirality, and surface boundary conditions (25, 26) indicating exceptional sensitivity. Therefore, to generalize, putting nematics into contact with diverse surfaces (18, 27) can be used as a simple but very powerful technique to detect the surface properties of microobjects such as biological fibers.

In this paper we demonstrate the surface morphology sensing of biorelevant fibers, including spider silk and cellulosic microfibers, by nematic droplets that are sprayed onto the fibers. Specifically, we explore the chiral and achiral nature of the fiber’s surface and the in-plane or perpendicular alignment fields the fibers impose on the nematic. Droplets with degenerate in-plane and perpendicular alignment of the nematic at their free surfaces are explored, combining experiments and numerical modeling, to allow for tuning of the sensing precision. Further, the entanglement sites of the fiber webs are explored, with the droplets deposited at the sites clearly revealing contact, noncontact, and entangled morphologies.

Surface Morphology of Spider Silk and Cellulosic Fibers

To show how efficient are the sensors based on the elastic response of the nematic molecular orientational order to different fiber surfaces, we select five types of biological fibers: three spider silks and two electrospun cellulose fibers with similar diameters of the order of 1–2.5 µm but different surface characteristics. Two of the chosen spiderweb fibers collected from Araneidae mangora and Latrodectus geometricus have similar morphologies characterized by small granules with diameters of few tens of nanometers randomly distributed along the fiber surface (Fig. 1C and Fig. S1A). The third spider silk fibers collected from Pholcus phalangioides (Fig. S1E) have smoother outer surfaces and, on average, smaller diameters. The electrospun cellulose-based fibers have always a smooth surface morphology but differ in the surface interaction depending on the cellulose derivative and the characteristics of the solutions from which the filaments are drawn (Figs. 1K and 2 A and H and Materials and Methods). The fiber surface is smooth and, when prepared from isotropic cellulose acetate solutions, also of uniform thickness, whereas it has slight width modulations when prepared from chiral nematic hydroxypropylcellulose solutions. All fibers used in our experiments were kept taut with the aid of a network of wires and specially designed electrospinning targets (Materials and Methods and Figs. 1J and 4A).

Fig. 1.

Surface morphology of biological microfibers detected by optical micrographs of nematic droplets. (A–I) Fibers from Araneidae mangora spider silk. (A and B) Spiders from which the fibers were collected and (C) SEM picture of the spider silk. Inset shows the rough surface topography. (D–F) Nematic liquid crystal droplet imposing a homeotropic anchoring at the droplet–air surface coating a silk fiber under (D) crossed polarizers, (E) crossed polarizers with a lambda plate, and (F) parallel polarizers. Note the ellipsoidal fringes, typical for homeotropic anchoring on the fiber surface. (G and H) Numerically simulated transmission micrographs under crossed polarizers and with an additional lambda plate, respectively, and (I) the simulated director profile (droplet radius R, fiber radius r). (J–Q) Cellulose acetate (CA) fibers electrospun from isotropic solution. (K) The SEM picture of the CA fiber is only slightly different from spider silk fiber but produces a completely different (L–N) transmission pattern of the nematic droplet (compare with D–F). Note the defect ring around the fiber. (O and P) Calculated transmission micrographs of the (Q) numerically simulated structure reveal that a fiber enforces tangential anchoring along the fiber axis. Cylinders in I and Q represent the director field distribution in the plane through the center of the fiber, color coded by their orientation. The red ring-like isosurfaces show the reduced nematic degree of order corresponding to −1/2 disclination rings.

Fig. S1.

Spider silk of Latrodectus geometricus and Pholcus phalangioides decorated with nematic liquid crystal droplets. (A) SEM picture of fibers produced by Latrodectus geometricus (also known as the Brown Widow). The image shows a typical morphology of the outer surface of the fiber, with a structure granular on the scale of tens of nanometers, resembling the SEM image in Fig. 1C observed for fibers produced by Araneidae mangora and distinct from structures observed on electrospun cellulose-based fibers (Figs. 1K and 2 A and H). Inset shows a higher magnification of the surface. (B–D) Typical textures observed in a nematic liquid crystal droplet, with homeotropic anchoring at the droplet–air surface, pierced on a fiber produced by Latrodectus geometricus spider between (B) crossed polarizers, (C) crossed polarizers with a lambda plate, and (D) parallel polarizers. As observed for droplets threaded by fibers produced by Araneidae mangora, shown in Fig. 1 D–F, ellipsoidal fringes can also be observed, which can be associated to a homeotropic anchoring on the fiber surface. (E) SEM image of fibers produced by Pholcus phalangioides. The fibers are much thinner on average then those of Latrodectus geometricus. Their surface appears smoother, resembling the morphology of cellulose acetate fibers. (F–H) POM images of a droplet with homeotropic alignment at the liquid crystal air interface pierced through the fiber showing a characteristic texture, which resemble the ones of the droplets threaded by CA fibers that induce planar molecular alignment along their axis: (F) crossed polarizers, (G) crossed polarizers with a lambda plate, and (H) parallel polarizers. (Scale bars in A and E, 1 µm; scale bars in B and F, 10 µm.)

Fig. 2.

Nematic droplets as chiral sensors for biofibers. (A) A cellulose-based electrospun fiber produced from an isotropic solution imposing planar anchoring along its axis is explored by (B–D) homeotropic nematic droplets and (E–G) planar nematic droplets. Note the ring defect encircling the fiber in the transmission micrographs (B) under parallel polarizers, crossed polarizers with lambda plate and crossed polarizers. The transmission patterns are reproduced in numerical simulations with calculated transmission micrographs under (C) crossed polarizers with lambda plate and crossed polarizers, respectively. (D) The fiber piercing through the droplet aligns the ring defect perpendicular to the fiber. (E–G) In planar nematic droplets, the structure is bipolar-like, with symmetric transmission micrographs. (H) A chiral HPC electrospun fibers, imposing chiral in-plane anchoring changes the droplet structure. Note the broken symmetry in homeotropic droplets (I–K) and the much brighter fiber in planar droplets (L–N). The green and blue ribbons on the fiber in D, G, K, and N represent the corresponding surface anchoring direction which is uniform in D and G and chiral in K and N. The viewing direction is from below the droplets.

Fig. 4.

Morphology of the fiber web crossings as revealed by nematic droplets. (A) Electrospun cellulose (HPC) fibers with (B) deposited nematic droplets. (C and D) Two different crossings are probed by nematic droplets and reveal different patterns in transmission micrographs (crossed polarizers, crossed polarizers with a lambda plate, and parallel polarizers). The crossings are clearly different also in the SEM pictures. (E–G) Simulated director fields in diametrical cross sections of droplets with two identical chiral fibers that (E) are fused in the center of the droplet, (F) are in contact, and (G) do not touch. All three configurations are accompanied by different transmission micrographs with crossed polarizers (top micrographs) and crossed polarizers with a lambda plate (bottom micrographs), demonstrating a possibility to identify different crossing morphologies.

In the next step, the differences in the fiber surface properties and the related effect on surrounding media are revealed by decorating fibers with nematic droplets. The droplets with homeotropic anchoring at the droplet–air interface deposited on Araneidae mangora and Latrodectus geometricus silk show ellipsoidal fringes under the polarized light (Fig. 1 D–F and Fig. S1 B–D). Oppositely, the cellulose acetate as well as Pholcus phalangioides fibers induce a totally different pattern, characterized by the formation of a defect ring in the middle of the droplet (Fig. 1 L–N and Fig. S1 F–H). The differences are even more evident when the droplet pierced by the Araneidae mangora fiber is heated from the nematic phase to the isotropic state and then back cooled to the nematic phase (Fig. S2 and Movie S1) and also by using a retardation plate in addition to the crossed polarizers (Fig. 1 E and M). To extract more quantitative fiber surface characteristics, the polarization microscopy analysis is compared with the numerical Landau–de Gennes free energy modeling (Materials and Methods and Supporting Information) to determine the nematic response in the orientational ordering, the direct signature of the fiber surface morphology, and interactions. For the fibers of Araneidae mangora and Latrodectus geometricus, the typical transmission micrographs with ellipsoidal fringes on both sides of the fiber in comparison with simulated polarization micrographs (Fig. 1 G and H) based on predicted director profiles (Fig. 1I) imply homeotropic (perpendicular) fiber surface alignment. The polarization microscope pictures (Fig. 1 O and P) for the cellulose acetate (CA) fibers indicate that axial in-plane surface anchoring is imposed, producing a ring defect encircling the fiber (Fig. 1Q). Similar textures are observed for nematic droplets suspended on fibers collected from Pholcus phalangioides (Fig. S1 F–H). The SEM images of the silks (Fig. S1 A and E) suggest that fibers with a rough surface induce homeotropic anchoring, whereas the smooth fibers induce axial unidirectional in-plane surface anchoring. This suggests the surface topography could play a relevant role in the fiber–nematic interactions.

Fig. S2.

Influence of the temperature on the texture of a nematic droplet pierced with Araneidae mangora fiber. (A) Heating and (B) cooling (snapshots from Movie S1). The sequence of images represents the LC threaded droplet texture evolution from the equilibrium nematic state, also observed at room temperature (T < 33.90 °C) to the isotropic phase (T > 34.10 °C). Snapshots were taken between crossed polarizers. The heating and cooling rates are equal to 0.2 °C/min. (Scale bar, 10 µm.)

For fibers inducing axial nematic alignment (Fig. 2 B–D), the ring is perpendicular to the fiber, even when the droplet is not symmetrically positioned around the fiber, so that the fiber is not piercing through the droplet center. When the anchoring at the droplet–air interface is strong, the ring appears in the middle of the droplet, and when the surface anchoring strength is lowered, the ring can shift away from the center along the fiber. It eventually disappears when the anchoring is very weak and the nematic droplet becomes uniformly aligned along the fiber.

Detecting Chirality of Biofibers

The symmetry and characteristics of the phase from which the biofibers are drawn play a major role in the surface and bulk characteristics of the biological fibers, notably including chirality. As with spiders, which have the ability to produce fibers from lyotropic phases (13, 28), cellulose filaments can be drawn from mesomorphic phases effectively creating fibers of different bulk and surface morphology (29). To explore the sensitivity of our approach to chirality, we prepare chiral and achiral cellulose-based fibers from liquid crystalline hydroxypropylcellulose (HPC) and isotropic CA solutions, respectively. Indeed, droplets with either homeotropic or planar degenerate anchoring at the droplet outer surface can be used to sort and determine the chirality of the biofibers. Droplets with homeotropic anchoring at the droplet–air surface for both achiral and chiral fibers show a ring defect at the middle of the fiber (Fig. 2 B and I). However, the ring is tilted in relation to the fiber axis, when the nematic droplet is pierced by the chiral fiber (Fig. 2I), which offers a simple method for chirality sorting. The chirality is expressed also in droplets with planar surface anchoring providing further sensitivity. Transmission micrographs of droplets with planar anchoring on outer surfaces deposited on achiral fibers are dark and rather uniform (Fig. 2E). On the contrary, chiral fibers in transmission micrographs of such droplets become very bright and the whole texture nonuniform (Fig. 2L).

Numerically modeling homeotropic (Fig. 2 C, D, J, and K) and planar droplets (Fig. 2 F, G, M, and N) on fibers gives good qualitative agreement with the experiments. The chiral in-plane anchoring is modeled by helicoidal surface alignment; see Supporting Information for details and Figs. 2 K and N and 3D, where ribbons illustrate winding of the director on the surface of the fiber. Such surfaces twist the director field and consequently modify light passing through droplets and fibers, making the fiber appear bright under the crossed polarizers. In homeotropic droplets, the ring tilt is found to be the result of coupling between the chiral anchoring at the fiber surface and the noncentral position of the fiber along the viewing direction due to gravity. The offset breaks the cylindrical symmetry of the nematic profile, which in turn forces the nematic to distort differently on the left and the right side of the fiber but only if the nematic anchoring on chiral fiber is helicoidal. The simulations show that the disclination ring tilt is roughly linearly proportional to both the angle of the chiral planar anchoring and the offset of the fiber (Fig. 3 D and E). Note that this effect occurs in the absence of bulk chirality, simply due to the helicoidal surface anchoring, and use of nematics with intrinsic chirality or low twist elastic constant would lead to possibly different structures beyond the scope of this paper.

Fig. 3.

The tilt angle of the ring defect in homeotropic nematic droplets depends approximately linearly on the anchoring chirality and fiber offset. Experimental micrographs (A) under crossed, (B) under parallel, and (C) under crossed polarizers with a lambda plate. (D) Numerically simulated structures reveal a roughly linear dependence of the ring tilt angle on surface chirality strength (anchoring tilt angle). Inset schematically shows the helicoidally winding anchoring direction along the fiber (represented by green-blue strands) characterized by anchoring tilt angle and the tilted disclination loop characterized by the ring tilt angle. (E) The diametrical cross-sections perpendicular to the fiber show director fields in cases of central and offset fiber position. The fiber offset Δx induces a difference in free energy cost on the left and right sides of the fiber, which tilts the ring. The color coding from gray to red marks the out-of-plane component of the director field.

The strength of chirality and handedness of biofibers can be deduced at a more quantitative level by measuring the ring tilt in the transmission micrographs of homeotropic droplets. In Fig. 3 A–C the pierced droplets with homeotropic alignment at the liquid crystal–air interface illustrate a variation of the ring tilt angle along the fiber. By comparing the measured tilt angles with calculated ones (Fig. 3D), a complex chiral surface morphology that is inducing a helicoidally winding molecular anchoring direction along the fiber is unveiled. By depositing droplets of approximately the same size, even the variability of the fiber along its length in principle could be tested, where variations in the surface tilt angle as small as ∼8° can be revealed.

The mechanical properties of electrospun mats and spiderwebs do not depend only on the structure of single filaments or fibers but also on the crossings and the topology of crossings between them. Spiders build complex webs such as the one shown in Fig. 1B, and the morphology of the web is determined by the filaments, which are glued together (30). With electrospun fibers, nonwoven membranes can be produced as well as mats by fusing the fibers with diverse processing conditions (Materials and Methods and Supporting Information), as shown in Fig. 4 A and B. The different crossings can be identified by SEM, but their surface morphology and surface potential can be very efficiently screened by depositing nematic droplets. Notably, the transmission patterns of the droplets are different for intersections of fibers that are (i) fused, (ii) in contact, or (iii) not in contact (Fig. 4 C and D). Numerical simulations validate the obtained results and the three different types of crossings can be identified, each giving different transmission patterns (Fig. 4 E–G).

Summary

We present a sensitive but simple method to distinguish between different surface morphologies of microfibers by piercing nematic droplets with the fibers that are ubiquitous in nature and of high relevance as biomimetic materials. From simple polarization micrographs, supported with numerical modeling of nematic structures and their microscopic images, we could identify specific surface morphology and chirality of fibers and even determine their chiral strength. Two major families of biofibers are used: filaments produced from isotropic or nematic solutions of cellulose and filaments collected from silk-producing spiders. For cellulosic fibers we find that their surfaces induce unidirectional planar alignment, which is revealed to be axial for achiral fibers or helicoidal for chiral fibers produced from liquid crystalline phases. For the considered spider silk from Araneidae mangora, Latrodectus geometricus, and Pholcus phalangioides, we observe that they are achiral and induce homeotropic or in one case axial planar alignment. Liquid crystal droplets (19) seem to be a powerful tool to give insights into artificial cellulose thread structures. Spider silk fibers are more challenging, due to the more diverse surface properties, and deserve more studies to investigate different families of spiders and different silk types from the same species. Further, our method should be complemented by nanoscale electron microscopy and chemical structure analysis to relate the anchoring to topography, nanostructure, and chemistry of the fibers, for instance, to uncover whether the difference in surface anchoring between different spider silks is due to chemical or morphological differences. Nematic microdroplets as sensors also reveal diverse entanglements between the fibers, by simply decorating the contact (entanglement) points with the droplets and analyzing their transmission micrographs.

The presented simple method for surface sensing by nematic droplets inherently allows for several generalizations, which we did not explore in this work but seem rather directly implementable and experimentally accessible. Performing microscopy at different angles with respect to the droplet (beyond only top-view) would allow for further more quantitative mapping of the nematic distortion, in turn providing more information on the fiber surface morphology. Local sensing of biofibers could be performed by applying optical tweezers to the droplets and pulling the droplets along the fibers, giving a length-dependent characterization of the fibers. By varying the density of the surrounding host fluid or substituting air with denser media, the exact piercing height of the fiber relative to the droplet center could be varied, allowing for studies of asymmetric confinement effects. Using liquid crystals with low twist elastic constants or intrinsically chiral materials would lead to numerous additional chiral structures. Applying external fields, such as electric field, could be used as a means for exploring the electrostatic effects at and around the fiber surfaces.

Finally, with this work we try to open a new design route for sensors of biorelevant fiber systems easily revealing their surface properties, including morphology and chirality. Possible further extensions of the method by including multiangle microscopy, tweezing, use of materials with different elastic properties and chiralities, buoyancy variation, and application of external fields are left for future studies.

Materials and Methods

Materials and Experimental Techniques.

The photos in Fig. 1 A, B, and J were taken under solar light with a Canon EOS 450D camera equipped with a 60-mm microlens. The cellulosic microfibers were electrospun from two types of solutions: isotropic and liquid crystalline (details in Supporting Information). To collect the spiderwebs, square meshes (side square about 1 cm) of metallic and polyethylene wires, of 2 mm in diameter, were used. The fibers were collected carefully as suspended straight filaments between the empty spaces of the mesh. The orb-weaver spiderwebs from Araneidae mangora were collected from the garden near Faculty of Science and Technology in Caparica, Portugal. The spiderwebs from Lactrodectus geometricus and Pholcus phalangioides were courtesy of Szymon Przebinda from the itinerant exhibition of National Museum of Natural History and Science in Lisbon.

The morphology of the fibers was observed with a scanning electron microscope CrossBeam Workstation (SEM-FIB; Zeiss Auriga). The SEM images under the in-lens mode have been carried out with an acceleration voltage of 2 kV and aperture size of 30 μm. A thin carbon layer (<20 nm) was deposited on the suspended fibers using a Q150T ES Quorum sputter coater. The nematic liquid crystal 4′-n-pentyl-4-cyanobiphenyl (5CB) was used to produce the microdroplets and the necklaces (Supporting Information). The pierced droplets were observed under a polarized optical microscope (POM) Olympus BX51 coupled to a CCD DP73. Temperature was controlled with a heating stage Mettler FP90.

Modeling of Droplets Pierced by Fibers.

The simulation of the nematic liquid crystal ordering in droplets relies on continuum modeling using Landau–de Gennes (LdG) free energy minimization approach (31). Tensor order parameter Q_ij describes the orientational order of the LC molecules and the tensorial invariants of Q_ij construct the total free energy F (32), combining bulk free energy and surface free energy F_S, which accounts for the anchoring on the droplet surface and the droplet–fiber interface. The anchoring strength and type on the fiber and the droplet surface is different in general. We minimize the free energy numerically by using an explicit Euler relaxation finite difference scheme. Details about the selection of material and interface parameters and modeling of the chiral fiber surface are described in Supporting Information, Numerical Free Energy Modeling.

Transmission polarization micrographs are calculated from the simulated director fields using the Jones $2\times 2$ matrix (Supporting Information, Numerical Free Energy Modeling).

This supporting information gives more in-depth explanations on preparation of fibers, preparation of liquid crystalline necklaces, numerical modeling with free energy calculations, and experimental results.

Preparation of Cellulose and Spider Silk Fibers

The targets used to collect electrospun cellulose fibers were prepared according to similar protocols reported in the reference (26). Rectangular frames, 1 cm wide and 3 cm long, were cut from aluminum foil and used to collect aligned electrospun cellulose-based fibers as shown in Fig. 1J. The fibers were produced from two types of solutions: isotropic and liquid crystalline. The isotropic solutions were prepared from 12% (wt/wt) cellulose acetate (CA) (Aldrich; M_w = 60,000 g/mol, 40% acetyl groups) in a homogeneous mixture of 67% (wt/wt) acetone and 33% (wt/wt) dimethylacetamide (DMac), both from Riedel-de Haen. The liquid crystalline solutions were made from 60% (wt/wt) hydroxypropylcellulose (HPC) (Aldrich; M_w = 100,000 g/mol) in DMac. The solutions were kept in the dark for several days while being stirred. To produce the electrospun filaments the solutions were transferred into a 1-mL syringe (diameter 4.5 mm) attached to a 27-gauge needle (diameter 0.21 mm) and 21-gauge needle (diameter 0.51 mm) for HPC and CA, respectively. The homemade electrospinning apparatus, described earlier in ref. 29, is composed of an infusion syringe pump (KDS100) to control the solution feed rate, a high voltage supply (Glassman EL 30 kV) to create an electric potential difference, and a target holder whose position can be remotely controlled. After applying the electric potential between the syringe tip and the targets previously prepared, the polymeric solutions were constantly fed to the syringe tip at a constant flow rate of 0.04 mL/h for HPC and 0.2 mL/h for CA and accelerated by applied electric field toward the collectors. The optimized operating conditions for the continuous drawing of cellulose derivatives biofibers were a voltage between needle and collector of 15 kV and a distance of 20 cm for HPC, whereas for CA a voltage of 10 kV and a distance of 12 cm were used. During the process the humidity (∼50%) was controlled as well as temperature (23 °C).

The spiderwebs were collected with square meshes (side square about 1 cm) of metallic and polyethylene wires, of 2 mm in diameter, and straight filaments are collected between the empty spaces of the mesh.

The orb-weaver spiderwebs from Araneidae mangora were collected from the garden near Faculty of Science and Technology in Caparica, Portugal.

The webs from Lactrodectus geometricus and Pholcus phalangioides were courtesy of Szymon Przebinda from the itinerant exhibition of National Museum of Natural History and Science in Lisbon, “The Fascinating World of Spiders and Scorpions.”

Spider Latrodectus geometricus is said to be native to Africa and has expanded its range around the world mainly in the last century. It belongs to the family Theridiidae that is of medical interest due to its neurotoxic venom and its effect on human health. This species has an affinity for human structures and residential buildings, where it constructs a web of irregular shape. The web is characterized by the presence of a retreat located at the top with a central disk and gumfoot lines with fluff mass and balls of glue (see, for example, ref. 33).

Pholcus phalangioides fibers were collected in Portugal from inside a house in the Setubal region.

Preparation of Liquid Crystalline Necklaces

The nematic liquid crystal 4′-n-pentyl-4-cyanobiphenyl (5CB) was used. To produce planar anchoring at the outer 5CB droplet, a mixture of glycerol and 5CB [about 10% (wt/wt) of liquid crystal was added to glycerol] was prepared by manually stirring the solution until the desired droplet size was achieved. The liquid crystal droplets were generated at the tip of a metallic copper wire with a 0.1-mm diameter, and the pure 5CB as well as 5CB mixture droplets were deposited on the suspended fibers. The diameter of the droplets was in the range of 2–50 µm.

Numerical Free Energy Modeling

Modeling of the nematic liquid crystal ordering in droplets relies on continuum mean field Landau–de Gennes (LdG) free energy approach (31). Tensor order parameter Q_ij,

$$Q_{ij}=\frac{S}{2}\left(3\hspace{0.17em}{n}_i{n}_j-{\delta }_{ij}\right)+\frac{P}{2}\left(e_i^{\left(1\right)}{e}_j^{\left(1\right)}-e_i^{\left(2\right)}{e}_j^{\left(2\right)}\right),$$

describes the average order and orientation of the LC molecules, where S is the nematic degree of order, n_i is the nematic director, P is the biaxiality, e_i⁽¹⁾ is the secondary director, and ${\mathbf{e}}_2=\mathbf{n}\times {\mathbf{e}}_1$. Order parameter tensor is used to construct total free energy F (32):

$$F=\underset{LC}{{\displaystyle \int }}\left\{\frac{A}{2}{Q}_{ij}{Q}_{ji}+\frac{B}{3}{Q}_{ij}{Q}_{jk}{Q}_{ki}+\frac{C}{4}{\left(Q_{ij}{Q}_{ji}\right)}^2\right\}dV$$

$$+\underset{LC}{{\displaystyle \int }}\left\{\frac{L}{2}\frac{\partial Q_{ij}}{\partial x_k}\frac{\partial Q_{ij}}{\partial x_k}\right\}dV$$

$$+\underset{D_U}{{\displaystyle \int }}\begin{array}{c}\begin{array}{c}\frac{W_U}{2}{\left(Q_{ij}-Q_{ij}^0\right)}^2\end{array}d\mathrm{\Sigma }\end{array}+\underset{D_{PD}}{{\displaystyle \int }}\begin{array}{c}\begin{array}{c}\frac{W_{PD}}{2}\end{array}{\left(\widetilde{Q_{ij}}-\widetilde{Q_{ij}^{\perp }}\right)}^{2}d\mathrm{\Sigma }\end{array}.$$

The LC denotes the integration over the bulk of the liquid crystal and D over the surface of the droplet and the fiber. In our examples, droplet radius to fiber radius ratio R/r is in the range 3.3–10. The first term of F accounts for the variation of the nematic degree of order S, i.e., the possible formation of singular topological defects; A, B, and C are material parameters characterizing nematic phase. The elastic distortions of the uniform nematic state are penalized by the second term where L is the elastic constant. The central interest of this study is the effect of surface anchoring and geometry, so we decided on a single elastic constant approximation of F and on the typical values material parameters: $L=4\times {10}^{-11}\hspace{0.17em}\text{N}$, $A=-0.172\times {10}^6\hspace{0.17em}{\text{J\hspace{0.17em}m}}^{-3}$, $B=-2.12\times {10}^6\hspace{0.17em}{\text{J\hspace{0.17em}m}}^{-3}$, and $C=1.73\times {10}^6\hspace{0.17em}{\text{J\hspace{0.17em}m}}^{-3}$. The use of multiple elastic constants (i.e., to account for different splay, twist, and bend elastic constants) would affect the extent of the areas where the three basic nematic deformation modes were dominant. The final two terms in F account for the LC interactions on the droplet outer and fiber surfaces. The term characterized by W_PD anchoring strength and renormalized tensors $\widetilde{Q_{ij}}$ and $\widetilde{Q_{ij}^{\perp }}$ introduced by Fournier and Galatola (34) is applied only for planar-degenerate surface anchoring realized on outer surfaces of droplets illustrated in Fig. 2 F, G, M, and N, where a rather weak value $W_{PD}={10}^{-5}\hspace{0.17em}{\text{J\hspace{0.17em}m}}^{-2}$ is used. The term characterized by W_U anchoring strength is used for surfaces enforcing local unidirectional anchoring described by Q_ij⁰ surface preferred ordering. In all cases we use $W_U={10}^{-2}\hspace{0.17em}{\text{J\hspace{0.17em}m}}^{-2}$, which corresponds to strong anchoring.

The unidirectional homeotropic anchoring is used for the droplet outer surfaces in simulations for the spider silk and cellulose fibers in Fig. 1 and for both chiral and achiral cellulose fibers in Fig. 2 C, D, J, and K. Unidirectional planar anchoring on achiral and chiral fibers was approximated by enforcing an average local orientation of molecules via Q_ij⁰ characterized by a planar director

$${\mathit{n}}_0=\left\{\cos \phi \sin \alpha ,\hspace{0.17em}\sin \phi \sin \alpha ,\cos \alpha \right\},$$

where $\phi =\arctan \left(\frac{y}{x}\right)+\frac{\pi }{2}$ and the anchoring tilt angle $\alpha$ against the axial direction parameterize the whole range of anchoring orientations, from $\alpha =0$ for achiral fibers (CA fibers with axial unidirectional anchoring in Fig. 2 D and G) to larger tilt angles, used for chiral CA fibers with helically winding anchoring in Figs. 2 J, K, M, and N and 3.

We minimize the free energy numerically by using an explicit Euler relaxation finite difference scheme on a cubic grid (typically 400 × 400 × 400 mesh points) with typical mesh resolution $\mathrm{\Delta }=10\hspace{0.17em}\text{nm}$ to avoid defect pinning (32). The surface is allocated as a thin spherical shell of mesh points with thickness equal to the mesh resolution.

Transmission polarization micrographs are calculated from the simulated director fields using the Jones $2\times 2$ matrix formalism (35) and accounting for a typical birefringence of the nematic liquid crystal.

Experimental Results Illustrated by Additional Figures and Movie

Here we complement the results of spider fibers obtained from Araneidae mangora (Fig. 1) with pictures of droplets pierced by fibers from Latrodectus geometricus (Fig. S1 A–C). Both have similar surface morphologies characterized by roughness on the few tens of nanometers providing a rather weak homeotropic anchoring. Contrarily, the silk fibers from Pholcus phalangioides (Fig. S1 E–H) are very smooth like the cellulose acetate fibers (Figs. 1 K–Q and 2 A–G) and are characterized by a strong planar anchoring along the fiber.

The stability of nematic droplet structures induced by piercing with spider silks from Araneidae mangora that enforce a strong planar anchoring in the axial direction is demonstrated by the reversibility of the structural changes following transitions from the nematic to isotropic and back to the nematic phase (Fig. S2 and Movie S2).