Main

We imparted strong corrugation onto the surface of carboxylated polystyrene microspheres (μPSs) by attaching rigid zinc oxide (ZnO) nanoscale spikes (‘nanospikes’). This involves initial absorption of positively charged ZnO nanoparticles (NPs) onto the negatively charged μPSs and subsequent growth of ZnO nanowires using established protocols6. The resultant hedgehog particles combine micro- and nanoscale structural features (Fig. 1a), and their geometrical and topographical specifications can be adjusted by changing the growth conditions to modify the surface densities, lengths and diameters of nanospikes and the μPS diameters (Fig. 1b–e, Supplementary Information section 1 and Supplementary Figs 2–5).

Figure 1: Hedgehog particles.
figure 1

a, Negatively charged, carboxylate-terminated µPSs are used as core templates (1) on which positively charged ZnO NPs are adsorbed (2). ZnO nanospikes are grown from ZnO nanoparticles (3) to a designed length (4, 6). Hedgehog particles are rendered hydrophobic by exposure to OTMS or PFTS (5). b–e, SEM images of hedgehog particles with different ZnO nanospike lengths: 0.19 µm (b), 0.27 µm (c), 0.4 µm (d), 0.6 µm (e). f, Confocal microscopy of an aqueous dispersion of hydrophilic hedgehog particles with fluorescently labelled μPSs. Inset, SEM image for the same hedgehog particles. g, h, SEM (g) and confocal microscopy (h) of an aqueous dispersion of OTMS-HPs. i, SEM image of particles from the bulk of an aqueous OTMS-HP dispersion collected five days after initial preparation. j, k, Confocal microscopy images of fluorescent OTMS-HPs (green, λmax = 486 nm) with adsorbed hydrophobic CdSe nanoparticles (red, λmax = 655 nm) in an aqueous dispersion (j) and in the dried state (k). l, Photographs of aqueous dispersions of (left to right) hydrophilic hedgehog particles (HPs) with green-dyed μPSs, OTMS-HPs, OTMS-µPSs and OTMS-ZnO nanowires (NWs). m, Photographs of (left to right) ZnO nanoparticles (NPs) in water, ZnO nanoparticles in 1 M NaCl, and OTMS-HPs in 1 M NaCl. n, Photographs of OTMS-HPs in (left to right) 0.1 M NaCl and 0.01 M NaCl.

PowerPoint slide

As-made hedgehog particles, with their polar ZnO surfaces, are highly hydrophilic. They form excellent dispersions in water (Fig. 1f, l) and other hydrophilic solvents. When rendering the hedgehog particles hydrophobic by silanization of the ZnO nanospikes with (7-octen-1-yl) trimethoxysilane (OTMS) or 1H,1H,2H,2H-perfluorooctyltriethoxysilane (PFTS) (Supplementary Methods; spectroscopic evidence of silanization is shown in Supplementary Information section 2 and Supplementary Fig. 9c), they form stable dispersions in heptane and hexane (Supplementary Information section 2 and Supplementary Fig. 6a, b).

Surprisingly, highly corrugated OTMS-modified hydrophobic hedgehog particles (OTMS-HPs) also form dispersions in water (Fig. 1g, h, l), and hydrophilic hedgehog particles disperse in representative hydrophobic solvents such as heptane, hexane and toluene (Fig. 4a–h). This illustrates that surface topography can be used to modulate the interaction between microscale particles and disperse them in phobic solvents.

Figure 4: Dispersion of hydrophilic hedgehog particles in hydrophobic organic solvents.
figure 4

a, Dispersions of hydrophilic hedgehog particles in (left to right) heptane, hexane and toluene. As in the case of dispersion of hydrophobic hedgehog particles in water, after sonication of hydrophilic hedgehog particles in organic solvent there was always a small amount of precipitate in the bottom of the vial. b, Confocal microscopy image of hydrophilic hedgehog particles in heptane. ce, SEM images of hydrophilic hedgehog particles from dispersions in heptane (c), hexane (d) and toluene (e). fh, SEM images of individual hedgehog particles in heptane (f), hexane (g) and toluene (h). Toluene dissolves the µPS core in the hedgehog particles, rendering dispersions of hydrophilic spiky shells. i, Van der Waals interaction potentials VvdW of hydrophilic hedgehog particles (blue) and µPS (red) in heptane. j, Total pair potential VE_DLVO of hydrophilic hedgehog particles in heptane.

PowerPoint slide

Immediately on sonicating various hydrophobic OTMS-HP and PFTS-modified hydrophobic hedgehog particle (PFTS-HP) formulations in water (Supplementary Information section 2 and Supplementary Fig. 7), we observed the formation of a precipitate on the bottom of the vial, floating aggregates on top of the liquid, and a stable opalescent dispersion. Dispersions remain stable and free of aggregation for a subset of particles for at least five days, as verified by scanning electron microscopy (SEM) (Fig. 1i) and dynamic light scattering (Supplementary Information section 2 and Supplementary Table 1). The percentage of hydrophobic hedgehog particle aggregates floating on the surface of the dispersions increased with elongation of the nanospikes (Supplementary Information section 2 and Supplementary Fig. 7a–d), but the colloidal stability of the particles dispersed in water was also enhanced (Fig. 2k–m).

Figure 2: Interface between hydrophobic hedgehog particles and water.
figure 2

ah, Confocal microscopy images of hydrophobic hedgehog particles labelled with hydrophobic CdSe NPs in aqueous dispersions (a, b) and hydrophilic hedgehog particles labelled with hydrophilic CdTe NPs in aqueous dispersions (c, d); hydrophobic hedgehog particles in an aqueous solution containing hydrophilic TGA-stabilized CdTe nanoparticles with green (λmax = 540 nm) emission (e); hydrophilic hedgehog particles in an aqueous solution containing hydrophilic TGA-stabilized CdTe nanoparticles (f); the same sample from image e after five days of storage in dark (g); and the same sample from image f after five days of storage in dark (h). i, SEM image of a hydrophobic hedgehog particle with a self-assembled film of TGA-depleted CdTe nanoparticles between the ZnO nanospikes, indicating the location of the air–water interface. The hydrophobic hedgehog particles were immersed in an aqueous solution of CdTe nanoparticles for 72 h. j, Schematic diagram of the air–water interface, showing the experimental parameters (definitions in Supplementary Information). km, SEM images of aqueous dispersions of hydrophobic hedgehog particles with ZnO nanospike lengths of 0.19 µm (k), 0.40 µm (l) and 0.57 µm (m).

PowerPoint slide

To exclude the possibility that the observed behaviour arises because our samples contain a subpopulation of hydrophilic OTMS-HPs or PFTS-HPs or represent a special case of Janus colloids, that is, colloids consisting of distinct hydrophobic and hydrophilic interfacial sectors7,8,9, we directly probed the hydrophobic nature of the particles after processing them into dried thin films. The filtrate of suspended OTMS-HPs exhibited high water repellency causing the droplets to roll off (the ‘lotus effect’; Supplementary Information section 2, Supplementary Videos 1,2,3 and Supplementary Fig. 8). Further evidence is obtained by injecting hydrophobic cadmium selenide (CdSe) nanoparticles into an aqueous dispersion of OTMS-HPs: confocal and transmission electron microscopy (TEM) images show the expected anchoring of hydrophobic nanoparticles on the spikes (Fig. 1j, k, Supplementary Information section 2 and Supplementary Fig. 9d), thus confirming their hydrophobicity and the uniformity of surface derivatization. The stability of the hydrophobic hedgehog particles in aqueous dispersion did not change on CdSe adsorption.

The wetting of corrugated surfaces4,10,11,12 is often attributed to a Cassie–Baxter wetting mode13,14 and in our case could include formation of an air shell in the vicinity of the μPS core. Such trapped air bubbles15 might provide buoyancy to the hedgehog particles, but are known to be thermodynamically unstable16,17. The presence of trapped air is verified by adding ethanol and observing gas evolving from the dispersion (Supplementary Video 5), and by observing, in high-resolution confocal microscopy images of the particles, concentric shells with markedly different refractive indices as would be expected if an air shell is present (Fig. 2a, b, Supplementary Information section 3 and Supplementary Fig. 10). Hydrophilic hedgehog particles in water have no air shells, and they appear under the same conditions as uniformly lit particles (Fig. 2c, d, Supplementary Information section 3 and Supplementary Fig. 11).

When adding fluorescent cadmium telluride (CdTe) nanoparticles, stabilized with hydrophilic thioglycolic acid (TGA), to an aqueous dispersion of the hydrophobic hedgehog particles, a dark zone devoid of emission around the hedgehog particles confirms the presence of a layer of air. The dimensions of the emission exclusion zones closely match the diameter of hedgehog particles (Fig. 2e). The fact that similar images were obtained after five days of storage in the dark without agitation (Fig. 2g) attests to the long-term stability of dispersions of our hydrophobic hedgehog particles in water, consistent with long-term trapping of air at macroscale corrugated surfaces18 (Supplementary Information section 3 and Supplementary Fig. 13). Hydrophilic hedgehog particles in identical luminescent media appear as bright spots with CdTe nanoparticles localized on and between the nanospikes (Fig. 2f, h).

Strong scattering of photons and electrons by ZnO nanospikes prevents successful optical or cryogenic TEM imaging of the air–water interface within hedgehog particles, but we can locate it by taking advantage of the fact that CdTe nanoparticles can self-assemble into nanowires19 and nanosheets20 at interfaces21: a thin layer of CdTe nanoparticles that assembles more than 200 nm in from the ends of the ZnO nanospikes (Fig. 2i, Supplementary Information section 3 and Supplementary Fig. 14) pinpoints the water meniscus (Fig. 2j). This allows us to calculate an average hedgehog particle density of 0.92 g cm−3 (Supplementary Information section 3), which closely matches the density of water and explains the buoyancy of the particles.

We must also explain why two individual hydrophobic hedgehog particles do not coalesce on collision. To do so, we refer to the extended Derjaguin–Landau–Verwey–Overbeek (E_DLVO) theory, according to which the sum of potentials associated with van der Waals (VvdW), electrical double layer (VDL) and hydrophobic (VHB) interactions approximate the total interaction potential (VE_DLVO) between the hydrophobic hedgehog particles: VE_DLVO = VvdW + VDL + VHB. Evaluating interparticle interactions in different configurations (Fig. 3a–c, Supplementary Information section 4 and Supplementary Fig. 19), we find that hedgehog-particle/hedgehog-particle pair potentials display high repulsive energy barriers of at least 14kBT (kB, Boltzmann’s constant) for the outer contour of spikes (x = 0; Fig. 3d). Penetration of the nanospikes into the interstitial spaces of another hedgehog particle (x < 0) is energetically unfavourable (Fig. 3d).

Figure 3: Interaction potentials of hydrophobic hedgehog particles in aqueous media.
figure 3

ac, Two general configurations, spike-to-spike (S–S; a) and spike-to-gap (S–G; b), are considered, along with the intermediate case in which the ZnO nanospikes face the side walls of opposing particles (ZS–G; c). dg, Interaction potentials between hydrophobic hedgehog particles. d, Pair potentials for hydrophobic hedgehog particles in S–S (VE_DLVO,S–S, black), S–G (VE_DLVO,S–G, orange) and ZS–G (VE_DLVO,ZS–G, green) configurations. The negative values of x correspond to the penetration of ZnO nanospikes into the interstitial spaces of another hedgehog particle; x = 0 corresponds to the outer contour around the spike tips. e, Pair potentials (VE_DLVO,HP) of hydrophobic hedgehog particles in an aqueous dispersion calculated according to the E_DLVO theory for the zeta-potentials at the air–water interface with ζ = −65 mV (black line) and ζ = −35 mV (red line), and for hydrophobic hedgehog particles with short nanospikes from Fig. 2k (green line). f, Hydrophobic interaction potentials of OTMS-HPs (VHB,HP, green) and OTMS-µPSs (VHB,PS, dotted green). g, Van der Waals interaction potentials of OTMS-HPs (VvdW,HP, blue) and OTMS-µPSs (VvdW,PS, dotted blue) and total attractive potentials of OTMS-HPs (VvdW+HB,HP, red) and hydrophobic OTMS-µPS (VvdW+HB,PS, dotted red) in water.

PowerPoint slide

Comparison of the VE_DLVO for hydrophobic hedgehog particles with that for hydrophobic μPSs (Supplementary Information section 4 and Supplementary Fig. 24d) shows that the interfacial corrugations transform the overall attractive potential into a repulsive one. For hedgehog particles with short nanospikes, VE_DLVO reverses such that the interaction becomes attractive (Fig. 3e), matching the experimental results in Fig. 2k. The key reason for the anomalous stability of hedgehog particle dispersions is that VvdW and VHB are greatly decreased for corrugated particles compared with the smooth spheres (Fig. 3f, g), owing to the drastic reduction in the contact area in the former case. The total contour area of tapered spikes represents <3% of the surface area of the smooth particles (Fig. 1a).

The colloidal stability of hydrophobic hedgehog particles in water is also enhanced by the presence of the double electric layer at the air–water interface, increasing VDL. The zeta-potential (ζ) of air bubbles, which affects their electrostatic repulsion, is known to be between −35 mV (ref. 22) and −65 mV (ref. 23). Such high ζ is attributed to autoionization of water24 that may also occur at the hydrophobic interfaces24,25. However, the fact that the hedgehog particle dispersion remains stable in the presence of 0.01–1.0 M NaCl, which leads to screening of electrostatic interactions and coagulation of ‘normal’ dispersions (Fig. 1m, n), shows that any increased electrostatic repulsion has a secondary role and that the anomalous colloidal behaviour of hedgehog particles is dominated by the reduction of attractive interactions between the particles. But limitations of Derjaguin–Landau–Verwey–Overbeek theory for high ionic strengths and nanoscale corrugated surfaces26 may need to be considered for a more complete mechanistic explanation.

If the drastic reduction in attractive components of the pair potential is the reason for the unusual stability of hedgehog particle dispersions, the same effect should occur in dispersions of hydrophilic colloids in hydrophobic solvents. Stable dispersions of hydrophilic hedgehog particles were obtained in heptane, hexane and toluene (Fig. 4a, Supplementary Information section 5 and Supplementary Fig. 26). SEM and confocal microscopy images (Fig. 4b–e) demonstrated non-agglomerated particles in the bulk of these dispersions and physical integrity of hedgehog particles (Fig. 4f–h). The μPS core of the hedgehog particles was dissolved in toluene, thus yielding a dispersion of hydrophilic hedgehog particle shells. As expected, ZnO nanoparticles and ZnO nanowires (Supplementary Information section 5 and Supplementary Fig. 27) do not disperse in the same solvents.

Calculations show that VvdW for this type of dispersion is much reduced compared with smooth spheres, and that the overall pair potential of hydrophilic hedgehog particles in heptane is indeed repulsive with VDLVO,HPs = 1.4kBT at x = 0 nm (Fig. 4i, j, Supplementary Information section 5 and Supplementary Fig. 28). Notably, dispersion in organic solvents lack the air layer between the spikes, and electrostatic interactions in organic solvents are not screened as in aqueous solutions and are therefore longer ranged.

The stability of our surfactant-free hedgehog particles in ‘phobic’ solvents offers a different perspective on scientific and technological problems related to colloidal interactions and might even enable new strategies for processing and dealing with colloids, for developing new drug delivery systems27, and for colloidal self-assembly9,28,29. We also believe that the unusual solvation behaviour of hedgehog particles (contrary to the traditional expectations of particle dispersion stability in hydrophobic/hydrophilic solvents) could be used to develop efficient adsorbers, absorbers, scatterers or catalysts that need to function in both organic and aqueous media.