Top Banner
Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörs a,1 , Diana Courty b , Rafael Libanori a , and André R. Studart a,1 a Complex Materials, Department of Materials, ETH Zurich, 8093 Zurich, Switzerland; and b Laboratory for Nanometallurgy, Department of Materials, ETH Zurich, 8093 Zurich, Switzerland Edited by David A. Weitz, Harvard University, Cambridge, MA, and approved March 9, 2016 (received for review December 15, 2015) Living organisms often combine soft and hard anisotropic building blocks to fabricate composite materials with complex microstruc- tures and outstanding mechanical properties. An optimum design and assembly of the anisotropic components reinforces the material in specific directions and sites to best accommodate multidirectional external loads. Here, we fabricate composite films with periodic modulation of the softhard microstructure by simultaneously using electric and magnetic fields. We exploit forefront directed-assembly approaches to realize highly demanded material microstructural de- signs and showcase a unique example of how one can bridge colloi- dal sciences and composite technology to fabricate next-generation advanced structural materials. In the proof-of-concept experiments, electric fields are used to dictate the position of the anisotropic par- ticles through dielectrophoresis, whereas a rotating magnetic field is used to control the orientation of the particles. By using such unprec- edented control over the colloidal assembly process, we managed to fabricate ordered composite microstructures with up to 2.3-fold en- hancement in wear resistance and unusual site-specific hardness that can be locally modulated by a factor of up to 2.5. colloids | self-assembly | wear resistance | composites | dielectrophoresis L ightweight composites offer an attractive alternative toward minimization of the energy demand of mobility systems from automobiles and airplanes to trains and ships. Continuous stiff fibers or discontinuous anisotropic particles have often been used as reinforcing phase to enhance the specific mechanical properties of such lightweight polymer-based composites. Al- though the fabrication costs of long-fiber reinforced composites remain prohibitive for the large-scale production of high-per- formance materials, the processing of polymer-based composites reinforced with discontinuous stiff elements can often be adapted to the automated manufacturing processes commonly used in the polymer industry, such as injection molding. However, these fabrication processes usually offer a limited control over the alignment and spatial distribution of the reinforcing particles, preventing a full optimization of the microstructural design in composite materials. Thus, the development of new strategies to assemble microstructured composites with tailored reinforcing architectures is crucial for the design and fabrication of even lighter and cheaper structural materials. Control over the spatial distribution and orientation of parti- cles in composites can be achieved using colloidal assembly tools that guide/template the organization of the reinforcing particles while the material is still in a fluid state. Recently, we have shown that control of particle orientation using an external magnetic field can lead to a significant increase in the materialsresistance against deformation, wear, and crack initiation, widening their range of applications (1, 2). Independent of orientation control, tuning the spatial distribution of particles through magnetic fields, centrifugal forces, or film welding techniques has also been demonstrated to enhance the functional properties of the resulting composite (1, 3, 4). Despite such successful examples of microstructured polymer-based composites, the use of colloidal assembly tools to control the reinforcement architecture of lightweight materials remains largely unexplored. Because well- defined periodic variations of local mechanical properties are expected to strongly affect the global mechanical performance of composites, assembly tools to enable simultaneous spatial and orientation control of reinforcing elements are highly demanded. Structures with periodically modulated and graded local me- chanical properties occur frequently in biological materials, as in the spicules of sea sponges (5, 6), mussel threads (7), and dental enamel (8). Despite their repeated occurrence in many strong and tough biological materials, such locally varying mechanical properties have been until recently overseen as an important crack-arresting mechanism in biological materials (6). A prime example of biological composites exhibiting intricate micro- structure with modulated mechanical properties is the enamel of human tooth, which is the hardest tissue present in the human body and is subjected to considerable abrasive stresses during mastication. The remarkable mechanical resilience of natural enamel stems from its unique reinforcement architecture. Dental enamel possesses oriented inorganic prisms, which form highly mineralized hard domains aligned perpendicular to the tooth surface (9, 10). Prisms contain higher concentration of hydroxy- apatite crystals compared with the softer and more proteinaceous interfacial domain between prisms, called sheaths (11, 12). Such periodic variations in composition, microstructure, and properties within the different domains of the enamel layer (8) are likely to be key in providing wear resistance to the material. In this archi- tecture, locally concentrated and vertically oriented reinforcing particles increase the surface hardness of the material, whereas the soft domains are expected to arrest local microcracks and to decrease the global elastic modulus of the surface. This complex softhard architecture eventually increases the fracture toughness of the material. The combination of high hardness, high fracture toughness, and low elastic modulus is known to enhance the wear resistance of materials (13). Due to the complex nature of the microstructural design of these natural materials, similar control over alignment and spatial distribution of reinforcing elements has not yet been achieved in synthetic materials. The development of directed colloidal assembly routes that enable the fabrication of such intricate periodic architectures would not only lead to Significance Combined magnetic and electric fields offer unprecedented control over the position and orientation of particles, enabling the formation of programmed colloidal assemblies and com- posite films with locally modulated mechanical properties. The amplitude of this modulation can be controlled to form poly- mer-based bioinspired films with reinforced islands. Strikingly, such mechanical modulation can be designed to improve con- siderably the wear resistance of composite films. Author contributions: A.F.D. and A.R.S. designed research; A.F.D. and D.C. performed research; R.L. contributed new reagents/analytic tools; A.F.D. and D.C. analyzed data; A.F.D., R.L., and A.R.S. wrote the paper; and all authors contributed to writing the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1 To whom correspondence may be addressed. Email: [email protected] or [email protected]. This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1524736113/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1524736113 PNAS | April 26, 2016 | vol. 113 | no. 17 | 46234628 APPLIED PHYSICAL SCIENCES
6

Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

Oct 15, 2018

Download

Documents

vuongliem
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

Periodically microstructured composite films made byelectric- and magnetic-directed colloidal assemblyAhmet Faik Demirörsa,1, Diana Courtyb, Rafael Libanoria, and André R. Studarta,1

aComplex Materials, Department of Materials, ETH Zurich, 8093 Zurich, Switzerland; and bLaboratory for Nanometallurgy, Department of Materials, ETHZurich, 8093 Zurich, Switzerland

Edited by David A. Weitz, Harvard University, Cambridge, MA, and approved March 9, 2016 (received for review December 15, 2015)

Living organisms often combine soft and hard anisotropic buildingblocks to fabricate composite materials with complex microstruc-tures and outstanding mechanical properties. An optimum designand assembly of the anisotropic components reinforces the materialin specific directions and sites to best accommodate multidirectionalexternal loads. Here, we fabricate composite films with periodicmodulation of the soft–hard microstructure by simultaneously usingelectric and magnetic fields. We exploit forefront directed-assemblyapproaches to realize highly demanded material microstructural de-signs and showcase a unique example of how one can bridge colloi-dal sciences and composite technology to fabricate next-generationadvanced structural materials. In the proof-of-concept experiments,electric fields are used to dictate the position of the anisotropic par-ticles through dielectrophoresis, whereas a rotating magnetic field isused to control the orientation of the particles. By using such unprec-edented control over the colloidal assembly process, we managed tofabricate ordered composite microstructures with up to 2.3-fold en-hancement in wear resistance and unusual site-specific hardness thatcan be locally modulated by a factor of up to 2.5.

colloids | self-assembly | wear resistance | composites | dielectrophoresis

Lightweight composites offer an attractive alternative towardminimization of the energy demand of mobility systems from

automobiles and airplanes to trains and ships. Continuous stifffibers or discontinuous anisotropic particles have often beenused as reinforcing phase to enhance the specific mechanicalproperties of such lightweight polymer-based composites. Al-though the fabrication costs of long-fiber reinforced compositesremain prohibitive for the large-scale production of high-per-formance materials, the processing of polymer-based compositesreinforced with discontinuous stiff elements can often be adaptedto the automated manufacturing processes commonly used inthe polymer industry, such as injection molding. However, thesefabrication processes usually offer a limited control over thealignment and spatial distribution of the reinforcing particles,preventing a full optimization of the microstructural design incomposite materials. Thus, the development of new strategies toassemble microstructured composites with tailored reinforcingarchitectures is crucial for the design and fabrication of evenlighter and cheaper structural materials.Control over the spatial distribution and orientation of parti-

cles in composites can be achieved using colloidal assembly toolsthat guide/template the organization of the reinforcing particleswhile the material is still in a fluid state. Recently, we have shownthat control of particle orientation using an external magneticfield can lead to a significant increase in the materials’ resistanceagainst deformation, wear, and crack initiation, widening theirrange of applications (1, 2). Independent of orientation control,tuning the spatial distribution of particles through magneticfields, centrifugal forces, or film welding techniques has alsobeen demonstrated to enhance the functional properties of theresulting composite (1, 3, 4). Despite such successful examples ofmicrostructured polymer-based composites, the use of colloidalassembly tools to control the reinforcement architecture oflightweight materials remains largely unexplored. Because well-defined periodic variations of local mechanical properties are

expected to strongly affect the global mechanical performance ofcomposites, assembly tools to enable simultaneous spatial andorientation control of reinforcing elements are highly demanded.Structures with periodically modulated and graded local me-

chanical properties occur frequently in biological materials, as inthe spicules of sea sponges (5, 6), mussel threads (7), and dentalenamel (8). Despite their repeated occurrence in many strongand tough biological materials, such locally varying mechanicalproperties have been until recently overseen as an importantcrack-arresting mechanism in biological materials (6). A primeexample of biological composites exhibiting intricate micro-structure with modulated mechanical properties is the enamel ofhuman tooth, which is the hardest tissue present in the humanbody and is subjected to considerable abrasive stresses duringmastication. The remarkable mechanical resilience of naturalenamel stems from its unique reinforcement architecture. Dentalenamel possesses oriented inorganic prisms, which form highlymineralized hard domains aligned perpendicular to the toothsurface (9, 10). Prisms contain higher concentration of hydroxy-apatite crystals compared with the softer and more proteinaceousinterfacial domain between prisms, called sheaths (11, 12). Suchperiodic variations in composition, microstructure, and propertieswithin the different domains of the enamel layer (8) are likely tobe key in providing wear resistance to the material. In this archi-tecture, locally concentrated and vertically oriented reinforcingparticles increase the surface hardness of the material, whereasthe soft domains are expected to arrest local microcracks and todecrease the global elastic modulus of the surface. This complexsoft–hard architecture eventually increases the fracture toughnessof the material. The combination of high hardness, high fracturetoughness, and low elastic modulus is known to enhance the wearresistance of materials (13). Due to the complex nature of themicrostructural design of these natural materials, similar controlover alignment and spatial distribution of reinforcing elements hasnot yet been achieved in synthetic materials. The development ofdirected colloidal assembly routes that enable the fabrication ofsuch intricate periodic architectures would not only lead to

Significance

Combined magnetic and electric fields offer unprecedentedcontrol over the position and orientation of particles, enablingthe formation of programmed colloidal assemblies and com-posite films with locally modulated mechanical properties. Theamplitude of this modulation can be controlled to form poly-mer-based bioinspired films with reinforced islands. Strikingly,such mechanical modulation can be designed to improve con-siderably the wear resistance of composite films.

Author contributions: A.F.D. and A.R.S. designed research; A.F.D. and D.C. performedresearch; R.L. contributed new reagents/analytic tools; A.F.D. and D.C. analyzed data;A.F.D., R.L., and A.R.S. wrote the paper; and all authors contributed to writing the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence may be addressed. Email: [email protected] [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1524736113/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1524736113 PNAS | April 26, 2016 | vol. 113 | no. 17 | 4623–4628

APP

LIED

PHYS

ICAL

SCIENCE

S

Page 2: Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

synthetic materials with outstanding mechanical performancebut also provide model systems for a deeper understanding ofstructure–property relationships in biological materials that ex-hibit this recurring structural feature.

Results and DiscussionHere, we combine external electric and magnetic fields to directthe assembly of anisotropic particles into polymer-based com-posite films that exhibit simultaneous spatial and orientationcontrol of reinforcing elements. By replicating the periodicmodulation of local mechanical properties found in many naturalsystems, the proposed colloidal assembly route leads to complexfilm architectures with significantly enhanced wear resistancecompared with unstructured counterparts. We first investigatethe assembly of spherical micrometer-sized colloidal particlesusing dielectrophoresis (14), as this method provides an easy andaccessible tool to control the spatial distribution of colloidalparticles guided by an external electric field. Next, we combinethe dielectrophoresis process with magnetically assisted particlealignment to fabricate composite materials with tunable modu-lation of mechanical properties and enhanced wear properties.A key requirement to enable simultaneous electric and mag-

netic assembly of particles is to create a colloidal system that hassufficient contrast in dielectric and magnetic properties betweenthe dispersed and continuous phases. On the basis of our previouswork (1, 15), magnetic contrast can be easily created by coatingthe surface of anisotropic reinforcing microplatelets with super-paramagnetic iron oxide nanoparticles (SPIONs). Because theyare only weakly affected by Brownian motion, the coated micro-platelets exhibit ultrahigh magnetic response (UHMR) (1). Thisallows for platelet alignment in magnetic fields as low as 1 mTwith minor SPIONs concentrations in the range of 0.01–0.5 vol %.For an aspect ratio larger than 10 (b/a > 10), the magnetostaticpotential energy of platelets, in an external magnetic field of H,can be approximated by the following formula: UmðplateletÞ=ð2=3Þπ½ða+ΔÞðb+ΔÞ2 − ab2�μo½χ2p=ðχp + 1Þ� H2sin2ψ , where 2b isthe diameter, 2a is the thickness, χp is the magnetic susceptibilityof the platelet, μo is the magnetic permeability of the vacuum, Δ isthe thickness of the iron oxide nanoparticle coating, and ψ is theangle between the applied magnetic field and the long axis of theellipsoidal shell. Such potential energy is derived assuming an el-lipsoid geometry for the platelets and the magnetic coating.According to this equation, the potential will exhibit a maximumat ψ = 90o and a minimum at ψ = 0o, which will be the favorablestate of the platelets under a magnetic field. A mismatch of themagnetic susceptibility of the platelets with the suspending liquidwill polarize the platelets, and eventually align them along themagnetic field direction, as it is most favorable for an induceddipole to be along the field. Ultrahigh magnetic response of ourplatelets is observed for a very narrow range of platelet diameterand aspect ratio, due to a compromise between magnetic, gravi-tational, and thermal energies (1). When suspended in a liquidphase with mismatched dielectric constant, the UHMR plateletscan be also spatially controlled with the help of tunable electricalfield gradients while simultaneously exposed to a rotating magnetto program their orientation.For spatial control of suspended colloidal particles we used a

process known as dielectrophoresis (DEP). DEP is definedas the translational motion of particles caused by an inhomo-geneous electric field (14) and has been previously used forparticle manipulation (16–20), particle/cell separation (21–26),and particle assembly/trapping. (17, 19, 27–38) Electricallydriven particle motion depends on the dielectric constant contrastbetween the solvent and the particle. A particle with a dielectricconstant («p) different from the suspending medium («m) acquiresa dipole moment under the influence of an electric field and iseither attracted toward (if «p > «m, “positive” DEP), or repelledfrom (if «p < «m, “negative” DEP) the regions with the strongestelectric field. In an inhomogeneous electric field E(r) (39), parti-cles experience a DEP force Fdep, Fdep =−ð1=2Þυp«eff«0∇jEj2ðrÞ,where «0 is the permittivity of vacuum, vp is the volume of the

particle, and «eff = 3β«m/(1−βφ)2 is the effective dielectric con-stant of the particle, with φ as the particle volume fraction, β(ω) =Re{[«pp(ω)−«mp(ω)]/[«pp(ω)+2«mp(ω)]}, ω is the frequency, and«pp(ω) and «mp(ω) are the complex permittivity of the particlesand of the suspending medium, respectively. As the formula states,Fdep scales with volume of the particle and reaches values of thesame order of kT, random thermal forces, when the particle size isbelow 200 nm for dielectric particles and 30 nm for metal particles(40). The dependence of the DEP force on the square of theelectric field makes it unrelated to the field direction. Because ofthis insensitivity to the field direction, we used a sinusoidal fieldwith a frequency of 1 MHz to inhibit polarization of the doublelayer and undesirable electrohydrodynamic effects. As shown inthe relation above, the force depends on the gradient of E2, whichin turn is directly affected by the magnitude of the applied electricfield. Here, we use the magnitude of the electric field as the maintuning parameter to affect the DEP forces and thus the spatialdistribution of colloidal particles.To manipulate particle positions, we fabricated electrodes

with micrometer-sized features. An electrode possessing micro-fabricated voids modulates the electric field strength between thetwo opposing electrodes, which acts as a virtual template for theassembly of the colloids in 3D (Fig. 1). Fig. 1A shows the sketchof the sample cell with the microfabricated electrode and thecounter electrode. Fig. 1 B and D shows the modulations ofelectric field strength between these electrodes simulated byusing finite-element calculations with COMSOL Multiphysics5.2. Electric field strength modulations cause field gradients lo-calized at the micrometer features of the electrode, which exert aDEP force to particles according to the equation shown above.Such forces localize the particles over the micrometer features ofthe electrode. For particles with a negative dielectric mismatchwith respect to the solvent, the force field looks as in Fig. 1C, wherethe particles are pushed toward the voids of the micropatterned

A

B

C

ED

E Electric Field Strength [V

/m]

Fig. 1. Sample cell design and electric field strength simulations. (A) Samplecell design with microfabricated electrodes. (B) Sketch of the one-sidedmicrofabricated electrode cell and the corresponding micrograph of theelectric field strength variations within the cell. (C) Force field diagramdepicting the negative DEP force applied to particles at a distance 0.30× thevoid size over a patterned electrode. (Inset) Sketch of the micropatternedelectrode. The blue ring highlights the location of highest field gradient,where the force exerted onto the particles is largest. (D) Plot of the electricfield strength along the electrode at an elevation of 20% of the pitch size.(E) Electric field strength oscillations given at different elevations to dem-onstrate the height dependence of the DEP forces. The indicated elevationrefers to the height from the electrode surface relative to the void size of themicropattern.

4624 | www.pnas.org/cgi/doi/10.1073/pnas.1524736113 Demirörs et al.

Page 3: Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

electrode. This is the case when silica particles with «p = 2.4 aresuspended in DMSO with «m = 47. Finite-element calculationsshowed that the electric field strength modulations depend onthe distance from the patterned electrode surface as seen in Fig.1E. Note that DEP forces decrease as a function of elevationfrom the electrode surface.Fluorescently labeled 1-μm silica particles were used to ex-

perimentally demonstrate the DEP assembly with confocal mi-croscopy. The sample cell was filled with a 5 vol % dispersion ofparticles in a DMSO:water mixture (41). To match the refractiveindex of the particles with that of the solution, 11.6% water and88.4% DMSO were mixed by weight. While the field was off,colloids were spread all over the sample cell. When an electricfield of 1 Vrms/μm was applied colloids moved to the low-electric-field regions, namely the voids of the patterned electrode. Fig. 2A and B shows examples of assembly and disassembly for anelectrode with a square array of round voids. Fluorescent par-ticles are assembled in the low-field regions when an externalfield is applied (Fig. 2B). As the field is turned off, this assemblyis partially destroyed by the Brownian motion of the particlesand local differences in osmotic pressure. Fig. 2A shows an im-age 1 min after the field was turned off. Although the concen-tration of particles is higher in the round voids, particles areobserved all over the electrode. Note that the fluorescence in-tensity contrast in the voids and rest of the electrode originatesonly due to the transparency difference between the voids andthe gold-coated pattern.An electric field of 0.8 Vrms/μm or higher is usually enough to

push all of the particles toward the field wells to form a 3D as-sembly within these wells, when the particle concentration issufficiently high (Supporting Information and Movie S1). Fig. 2Cshows a confocal microscopy volume scan of a 3D assemblybetween the two electrodes. Assembled particles are shaped intotapered pillars due to the simultaneous action of gravity and thegradient in electrical field strength between the electrode andcounter electrode (Fig. 1E and Figs. S1 and S2). These resultsare in agreement with the finite-element analysis (Fig. 1) and

show that we can obtain full positional control of spherical col-loids in both 2D and 3D.As the assembly is field dependent and it is partially destroyed

after the field is turned off, we evaluated possible routes to fixthe assembly in place before thermal randomization. To fix theassembly such that it is permanent after the field is turned off, weadded a total of 10 vol % ethoxylated trimethylolpropane tri-acrylate (ETPTA, MW 912, Sigma-Aldrich) monomer mixedwith the photoinitiator 2-hydroxy-2-methyl-1-phenyl-1-propanoneinto the DMSO:water mixture. The assembly was then consoli-dated by UV illumination over the sample for 10 min, while thefield was kept on. SEM images of such permanent assembliesafter drying are shown in Fig. 2 D and E for round and squarevoids, respectively.The assemblies of particles are strictly dictated by the design

of the electrodes, which allows us to shape the particle-richpatterns freely. Fig. 3 demonstrates possible assemblies pro-duced using various electrode designs. As the field gradientscompress the particles into a confined volume, the local volumefraction of colloids is increased and particles start to crystallize(Fig. S3).DEP forces within the sample cell were high enough to also

allow for manipulation of non-Brownian particles. Using silica-coated alumina platelets with average size of 8.3 μm, we ob-served particle motion toward the electric field wells generatedby the patterned electrodes. Surprisingly, this was possible evenafter sedimentation of the alumina particles. Because silica dis-plays the same surface charge as the glass sample cells, thecoating improves the manipulation of the alumina platelets bypreventing them from sticking to the cell walls. Moreover, thesilica coating eases manipulation by increasing the dielectricmismatch between the solvent and the particle (see the Sup-porting Information for details).To control not only the spatial distribution but also the ori-

entation of the particles, we used alumina platelets that alsocontained SPIONs within the thin silica shell (15). This surfacemodification makes the particles responsive to magnetic fields,allowing us to control their biaxial orientation by applying arotating magnetic field (1, 42, 43). Eventually, a combination of arotating magnetic field and electric field microgradients can leadto a simultaneous control of the orientation and position of theparticles, respectively. Whereas the magnetic field was used toalign the platelets vertical or horizontal to the gravity direction,DEP was used to localize them in the electric field wells. Fig. 4Ashows the setup used to control simultaneously the alignmentand spatial distribution of the magnetized platelets. After theassembly was fixed by photocuring the polymer matrix, sampleswere dried and imaged by SEM. Fig. 4B shows an example ofplatelets oriented vertically while assembled in a square arrayof round voids. In-plane alignment can also be obtained bychanging the rotation plane of the magnetic field (Fig. 4C). The

Fig. 3. Various designs of electrodes resulting in complex 3D self-assembliesof colloidal particles. (A) C-shaped electrodes assemble particles into a 3DGreek theater shape. (Inset) The 3D structure constructed from a stack ofconfocal images. (B) Another assembly on a star-shaped electrode design.(Insets) Zoomed image (Top) of a single star, which shows the crystallineorder of the particles in the assembly, and (Bottom) the 3D reconstruction ofthe confocal images. (C) A cross electrode design results in an assembly of 3Dcross shapes. (Insets) Zoomed image (Top) and a 3D construction image ofthe assembly (Bottom). Note that the bottom shape dictated by the elec-trode design evolves to a rounder shape as it grows in height toward thepattern-free electrode. (Scale bars, 50 μm.)

A

D E

B C

Fig. 2. Directed self-assembly of fluorescently labeled spherical colloidsusing negative DEP. (A) Fluorescent colloids are found to spread all over thesample cell 1 min after the field is turned off. (Inset) Zoomed-in image of theedge of the void. (B) When an electric field of 1 Vrms/μm is applied particlesmove to the low-field regions, corresponding to the voids of the electrode.(C) The assembly spans the volume between the two electrodes and forms a3D pillar that becomes thinner toward the unpatterned electrode. The as-pect ratio, h/d, of the pillar is ∼10, with a height h of 100 μm and a diameterd of 10 μm (here “g” denotes the direction of the gravity). (D and E) SEMimages of assemblies fixed within a photosensitive polymer matrix while thefield was on. Structures shown in D and E were obtained using micro-patterned electrodes with round (D) and square (E) voids, respectively. As-sembly is consolidated within the photocurable polymer matrix. (All scalebars are 20 μm except the one in the inset E, which is 10 μm.)

Demirörs et al. PNAS | April 26, 2016 | vol. 113 | no. 17 | 4625

APP

LIED

PHYS

ICAL

SCIENCE

S

Page 4: Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

smaller panels in Fig. 4 B and C are zoomed images on a singleisland showing the respective orientation of the magnetizedplatelets. Note that B in Fig. 4 B and C indicates the plane ofalignment of the magnetic field, which is perpendicular to therotation axis.Composites with controlled spatial distribution and orientation

of platelets exhibit tunable site-specific mechanical reinforcement.The region of the matrix with the platelets is preferentially rein-forced against external loads whereas the platelet-free regions aremore prone to deformation. We demonstrate this by performingVickers hardness measurements on the composite surface facingthe patterned electrode. Drying the sample lying on the patternedelectrode leads to a flat topography with the reinforcing plateletswithin 10 to a maximum of 30 μm from the surface. Vickers in-dents were made at different xy positions on this surface (Fig. S5)as indicated in the sketch of the composite with reinforcingplatelet islands shown in Fig. 5A. Indents were created on thematrix without platelets (orange) and on sites reinforced withplatelets (purple).Fig. 5A shows an SEM image of a composite after a series of

microindentations using the same load on and off the platelet

islands. The composite with horizontally aligned platelets nicelydisplays spatially controlled reinforcement on the surface. Diamond-shaped frames showing the outer periphery of the small andlarger Vickers indents were drawn in the SEM image as guidesto the eye. Indents made on sites locally reinforced with plate-lets are clearly smaller compared with those formed in non-reinforced areas. As shown in Fig. 5B (purple) the hardness ofregions reinforced with horizontally aligned platelets is about1.5× higher than that exhibited by the matrix alone. Remarkably,an assembly with vertically aligned platelets presents hardnessvalues that are about 2.5× higher (Fig. 5B, green) on the rein-forced regions compared with the matrix without platelets. Thisprovides a way to tune the strength of the local reinforcementwhile retaining the spatial control. Tuning the reinforcement by

Fig. 4. Combination of magnetic and electric fields for the self-assembly ofanisotropic colloidal particles. A combination of electric and magnetic fieldswas used to direct the position and the orientation of silica-coated aluminaplatelets, respectively. (A) A rotating magnet, as sketched, orients themagnetized platelets in the plane of rotation, whereas the electric fieldpositions the particles due to DEP forces. (B) SEM images of a compositewhere platelets are assembled in a square array of round islands with ver-tical alignment. (C) A composite film with horizontally aligned platelets in asquare array of square islands. (Scale bar is 100 μm in B and C but 50 μm insmall panels of C.) Top image in C has the same alignment as bottom (C). Thesquare depicts the plane of alignment (horizontal/in-plane alignment)whereas oblique square indicates that the plane of alignment is perpen-dicular to the image (vertical/out-of-plane alignment).

I

II IIIII

IV

Beam Voltage 5kV Beam Voltage 30kV

I

II

III

IV

V

IVV

IIIII

I

A

C

D

F

E

B

Fig. 5. Vickers hardness and wear resistance of composites with spatiallyand orientationally controlled platelets. (A, Top) Sketch indicating the dis-tribution of platelet-containing islands within the matrix and the spotswhere Vickers indentation was performed. (A, Bottom) SEM image of asample with islands of horizontally aligned platelets after being locallyprobed with several Vickers indents. (B) Local Vickers hardness of islandscontaining horizontally and vertically aligned platelets compared with thenonreinforced surrounding matrix. (C) Two SEM images of the same regionof the sample taken with different beam voltages to identify the location ofthe islands containing vertically aligned platelets underneath the compositesurface (see also Fig. S4). (D and E) Wear behavior of the compositesdepending on (D) the alignment of platelets within the compartments and(E) the compartmentalization of the reinforcing elements. (F) Wear volumesgiven for randomly oriented (V), noncompartmentalized (IV), and com-partmentalized ordered (I–III) structures after 300 wear ball rotations.

4626 | www.pnas.org/cgi/doi/10.1073/pnas.1524736113 Demirörs et al.

Page 5: Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

controlling the orientation was demonstrated before by Libanoriet al. (44) but here we add spatial control to that and providelocal reinforcement control. Note that the local concentration ofthe particles within the compartments in our experiments isaround 40 vol %. Our hardness results are in line with what onewould expect by extrapolating the data given by Libanori et al.(44) to higher platelet volume fractions.To identify the positions of the platelets under SEM, we took

images of the same area with low (3–5 kV) and high beam voltages(30 kV). The low voltage only shows the outermost surface,whereas high voltages enable deeper penetration and provide in-formation about the actual position of the platelets (Fig. 5C).Nanoindentation measurements on these samples agree with

the Vickers tests. Horizontally aligned sample exhibited hardnessof 134 ± 3 MPa on the regions reinforced with platelets and 89 ±10 MPa for the platelet-free matrix. For the vertically alignedsample the hardness on reinforced regions increased to 395 ±54 MPa whereas the hardness for the matrix was measured to be120 ± 11 MPa. Nanoindentation tests performed on these samplesrevealed similar trends for the elastic moduli (Fig. S6). Elasticmoduli for the horizontally aligned sample were 967 ± 30 and 746 ±77 MPa on top of the platelet island and on the platelet-free re-gions, respectively (45). For the vertically aligned sample weobtained an elastic modulus of 6,687 ± 760 MPa on top of platelets,and 2,168 ± 329 MPa for the platelet-free regions (Fig. S6).To investigate the effect of the locally reinforced islands on

the abrasion behavior of the composite surfaces, we tested thewear resistance of these composites by measuring the displace-ment of a rotating wear ball into the material while subjected to anormal load of 0.197 N. The results show that the wear resistanceis strongly influenced by the orientation of the platelets (Fig.5D). The horizontal alignment is prone to delamination and thusleads to stronger wear than in the vertically aligned composites.Surprisingly, we observed a significant difference between thewear resistances of vertically aligned composites depending onthe direction of the ball rotation compared with that of platelets.When the ball rotation direction was perpendicular to the align-ment of the platelets, the composite was more wear-resistant (Fig.5D, orange). We attribute this to the fact that vertical plateletsaligned perpendicular to the rotation direction provide a largerarea of stiff material that can oppose the abrasion action of thewearing ball. Another insightful result is the effect of compart-mentalization on the wear properties of the composites. Com-partmentalization here refers to the placement of platelets intoseparate islands, while keeping the same global particle volumefraction. Remarkably, we observed that the compartmentalizedsamples were more wear-resistant compared with the non-compartmentalized composites (Fig. 5 E and F and Fig. S7). Weattribute this to the locally increased particle concentration withinthe islands. This allows particles to be in closer proximity to oneanother in a jammed-like state, thus preventing larger displace-ments within the matrix. Based on the size of the islands and thepitch of the pattern, the local volume fraction of reinforcementincreases by up to ∼40 vol % within the platelet islands. Note thatwear enhancement is a collective effect of particle assembliesinteracting with the wear ball, which can be appreciated from thetypical crater size (formed after wear), that lies around 2 ± 0.3 mmand covers a couple of tens of particle islands (Fig. S7). Earlierliterature reports suggest that compartmentalization should alsodecrease the friction on the surface (46, 47).The compartmentalized architecture of our most wear-

resistant synthetic composites shares common features with that ofabrasion-resistant heterogeneous structures found in natural mate-rials. In dental enamel, local mechanical heterogeneity arises fromthe different density and strength of mineral crystallites within theprisms and sheaths of the structure (8). Therefore, our resultssuggest that such periodic variation of the local mechanical re-inforcement in enamel may be one of the underlying mechanismsthat provides the tooth with a remarkable resistance against wearduring mastication loads.

ConclusionMagnetic and electric fields can be combined to fabricate high-performance composites with unprecedented control over theposition and orientation of particles, enabling the formation ofcomplex and programmed reinforcing architectures. The for-mation of islands of concentrated platelets and the control overtheir orientation provide an ample design space for locallymodulating the mechanical properties of composite films. Theamplitude of this modulation depends on the concentration andorientation of platelets within islands. As an example, islandscontaining vertically aligned platelets are ∼2.5× harder than thesurrounding polymer matrix. Such mechanical modulation con-siderably improves the wear resistance of the composite films.Islands containing vertically aligned platelets were found to bethe most resistant against wear. Strikingly, we found that suchcomposites also exhibit anisotropic wear properties, with stron-ger abrasion resistance in the direction orthogonal to that of thealignment of vertical platelets. The periodic variation in localmechanical properties engineered in these synthetic systems isconceptually similar to that found in biological materials like thetooth enamel. Hence, our results suggest that periodic mechan-ical modulation might be an important design principle used innature to increase the wear resistance of biological materialsexposed to highly abrasive environments. Beyond wear proper-ties, the mechanically modulated architectures proposed heremay be further explored to provide physical cues to guide thebehavior of living cells on biomaterial surfaces.

Materials and MethodsColloids and Fixed Assemblies. The spherical colloids were core–shell silicaparticles, displaying a fluorescent-labeled core of diameter d ∼ 250 nm,surrounded by a nonfluorescent shell with a thickness of 300 nm. Suchparticles were synthesized following a protocol similar to that proposed byvan Blaaderen and Vrij (48) Model suspensions were prepared by dispers-ing these core–shell particles in a refractive-index–matching mixture of wa-ter (11.6 wt%) and DMSO (88.4 wt%). Anisotropic, 8.3 ± 4-μm-long and 403 ±139-nm-thick alumina platelets were purchased from Merck KGaA (WhiteSapphire grade). Three hundred platelets were analyzed for the de-termination of the reported thickness and diameter. Platelets were magne-tized as described by Erb et al. (1) and coated with a silica shell following aprocedure similar to that shown by Libanori et al. (15) To obtain compositesfrom the assembled colloids, a method adapted from Jiang andMcFarland (49)was used. A total of 10 vol % ETPTA (MW 912, Sigma-Aldrich) monomer wasmixed with the photoinitiator 2-hydroxy-2-methyl-1-phenyl-1-propanone intothe DMSO:water mixture, which contained 10 wt % alumina particles. Themonomer:initiator volume ratio for ETPTA and photoinitiator was 100:1. Theresulting suspension was kept in the dark until further processing of the fluidto obtain composites with different reinforcing architectures. Particle assem-blies were fixed by UV exposure over 1-cm distance for 10 min, while the fieldwas on. An 8-W UV lamp (Camag) with a wavelength of 366 nm was used.

Electrode Fabrication. Gold grids were fabricated by standard photolithog-raphy. This involved dissolution of the exposed photoresist regions, sub-sequent e-beam evaporation of 4-nm titanium and 40 nmgold layers, and lift-off of the masking photoresist by sonication. The fabrication was performedat the ETH clean room facility FIRST.

Magnetic Alignment. Alignment of the magnetically modified aluminaplatelets was performed by using a neodymium magnet (Supermagnete,Webcraft AG) positioned ∼15 mm above the sample cell and rotated at afrequency of ∼1.5 Hz (Fig. 4A). The permanent magnet used (15 × 15 ×8 mm3) generates a magnetic field of ∼0.4 T on the surface.

Preparation of Composite Samples. Fluid cells for the directed assembly ofparticles were made by stacking two gold electrodes using 80–100-μm cov-erslips as spacers. One of the electrodes contained micropatterns to modu-late the local electrical field. Electric fields were applied to assemble theparticles into the desired compartmentalized structures. Magnetic fieldswere used in some of the experimental series to also gain control over theplatelet orientation. The photosensitive monomer added to the fluid wasthen polymerized to fix the assembled structures. After fixation, the samplecell was opened and the film was dried under vacuum, while lying on themicropatterned electrode, before further characterization. Electrodes with

Demirörs et al. PNAS | April 26, 2016 | vol. 113 | no. 17 | 4627

APP

LIED

PHYS

ICAL

SCIENCE

S

Page 6: Periodically microstructured composite films made … · Periodically microstructured composite films made by electric- and magnetic-directed colloidal assembly Ahmet Faik Demirörsa,1,

100- or 50-μm square voids were used for the preparation of samples formechanical testing. Before the mechanical tests, an epoxy layer was de-posited over the film to help detachment from the patterned electrode andserve as a substrate for the tests (Fig. S5). In all experiments, we used amicropatterned bottom electrode and an unpatterned top electrode.

Vickers Indentation. A Wolpert Microhardness tester MXT-α was used tomeasure Vickers hardness of the samples. Measurements were carried outusing an HV 0.01 (0.0981-N load) indenter and a dwell time of 10 s.

Wear Resistance Measurements. The setup used for wear tests is similar to theone developed by Tervoort et al. (50). Wear is induced by rolling a 25.4-cmhard zirconia ball on the surface of the sample. A slurry of silicon carbideparticles (∼5-μm size, F1200 C5, Washington Mills Ltd.) was used as abrasivemedium. Such suspension was dripped over the testing site at a feed rate of0.4 cm3/min and also works to rinse away wear residues. For the wearmeasurements, we used at least three different samples and ran a total of atleast seven tests for each reported data point. The plots show averages ofthese measurements. The indicated error bars represent the variation ofresults obtained from different runs and samples.

Nanoindentation. A TriboIndenter (Hysitron Inc.) equipped with a Berkovichtip was used to carry out nanoindentation measurements. A minimum of 10

measurements were performed on and off the platelet regions. Tests wereconducted under load-controlled mode with load and unload segments of 5 seach. The maximum force was set to 7,000 μN. The obtained load-displace-ment curves were analyzed using the Oliver–Pharr method (51).

Finite-Element Analyses. Finite-element analyses were conducted by usingComsol 5.2 to estimate the electric field strength landscape imposed to theparticles. Such simulations are semiquantitative and do not include in-formation about the colloid and its response to the frequency. Although notquantitative, this simplified approach provides useful information on howthe field strength changes in 3D for a given potential applied between themicropatterned and unpatterned electrode plates (10 V in our case). Providedwith the information whether particles have a negative or positive mismatchwith the dispersing medium, such analyses predict the precise location ofparticles in the assembly (Fig. S8).

ACKNOWLEDGMENTS. We are thankful to Tobias Schwarz (ScopeMmicroscopy center at ETH), Dr. Kirill Feldman (Soft Materials, ETH),Hortense Le Ferrand, Dimitri Kokkinis (Complex Materials, ETH), and Prof.Ralph Spolenak (Laboratory for Nanometallurgy, ETH) for help withinstruments and discussions. A.F.D. and A.R.S. are grateful for the financialsupport from the Swiss National Science Foundation (Grants 200021_126646and PZ00P2_148040).

1. Erb RM, Libanori R, Rothfuchs N, Studart AR (2012) Composites reinforced in threedimensions by using low magnetic fields. Science 335(6065):199–204.

2. Libanori R, Münch FHL, Montenegro DM, Studart AR (2012) Hierarchical re-inforcement of polyurethane-based composites with inorganic micro- and nano-platelets. Compos Sci Technol 72(3):435–445.

3. Libanori R, et al. (2012) Stretchable heterogeneous composites with extreme me-chanical gradients. Nat Commun 3:1265.

4. Chand N, Dwivedi UK, Sharma MK (2007) Development and tribological behaviour ofUHMWPE filled epoxy gradient composites. Wear 262(1–2):184–190.

5. Miserez A, et al. (2008) Effects of laminate architecture on fracture resistance ofsponge biosilica: Lessons from nature. Adv Funct Mater 18(8):1241–1248.

6. Kolednik O, Predan J, Fischer FD, Fratzl P (2011) Bioinspired design criteria for dam-age-resistant materials with periodically varying microstructure. Adv Funct Mater21(19):3634–3641.

7. Harrington MJ, Masic A, Holten-Andersen N, Waite JH, Fratzl P (2010) Iron-clad fibers: Ametal-based biological strategy for hard flexible coatings. Science 328(5975):216–220.

8. Ge J, Cui FZ, Wang XM, Feng HL (2005) Property variations in the prism and the or-ganic sheath within enamel by nanoindentation. Biomaterials 26(16):3333–3339.

9. Warshawsky H (1989) Organization of crystals in enamel. Anat Rec 224(2):242–262.10. Meckel AH, Griebstein WJ, Neal RJ (1965) Structure of mature human dental enamel

as observed by electron microscopy. Arch Oral Biol 10(5):775–783.11. Mann S (1997) The biomimetics of enamel: A paradigm for organized biomaterials

synthesis. Ciba Found Symp 205:261–269, discussion 269–274.12. Boyde A (1997) Microstructure of enamel. Ciba Found Symp 205:18–27, discussion 27–31.13. Leyland A, Matthews A (2004) Design criteria for wear-resistant nanostructured and

glassy-metal coatings. Surf Coat Tech 177–178:317–324.14. Pohl HA (1978) Dielectrophoresis: The Behavior of Neutral Matter in Nonuniform

Electric Fields (Cambridge Univ Press, Cambridge, UK).15. Libanori R, Reusch FB, Erb RM, Studart AR (2013) Ultrahigh magnetically responsive

microplatelets with tunable fluorescence emission. Langmuir 29(47):14674–14680.16. Bartlett AP, Agarwal AK, Yethiraj A (2011) Dynamic templating of colloidal patterns

in three dimensions with nonuniform electric fields. Langmuir 27(8):4313–4318.17. Beltramo PJ, Furst EM (2013) Predicting the disorder-order transition of dielectropho-

retic colloidal assembly with dielectric spectroscopy. Electrophoresis 34(7):1000–1007.18. Blaire G, et al. (2013) Hybrid Bio-Mag-MEMS combining magnetophoresis and di-

electrophoresis. Eur Phys J B 86(4):1–6.19. Frusawa H, Inoue M (2012) Scanning dielectrophoresis for the directed-assembly and

non-contact manipulation of colloidal particles. Intell Autom Soft Comput 18(2):187–199.

20. Green NG, Morgan H (1999) Dielectrophoresis of submicrometer latex spheres. 1.Experimental results. J Phys Chem B 103(1):41–50.

21. Gascoyne PRC, Vykoukal J (2002) Particle separation by dielectrophoresis.Electrophoresis 23(13):1973–1983.

22. Markx GH, Pethig R (1995) Dielectrophoretic separation of cells: Continuous separa-tion. Biotechnol Bioeng 45(4):337–343.

23. Morgan H, Hughes MP, Green NG (1999) Separation of submicron bioparticles bydielectrophoresis. Biophys J 77(1):516–525.

24. Zhou H, White LR, Tilton RD (2005) Lateral separation of colloids or cells by dielec-trophoresis augmented by AC electroosmosis. J Colloid Interface Sci 285(1):179–191.

25. Green NG, Morgan H (1998) Separation of submicrometre particles using a combinationof dielectrophoretic and electrohydrodynamic forces. J Phys Appl Phys 31(7):L25.

26. Krishnan JN, et al. (2009) Rapid microfluidic separation of magnetic beads throughdielectrophoresis and magnetophoresis. Electrophoresis 30(9):1457–1463.

27. Brown AP, Milner KR, Allsopp DWE, Betts WB (1998) Characteristic dielectrophoreticcollection of 20 nm latex beads measured by an impedance technique. Electron Lett34(20):1934–1936.

28. Freer EM, Grachev O, Duan X, Martin S, Stumbo DP (2010) High-yield self-limitingsingle-nanowire assembly with dielectrophoresis. Nat Nanotechnol 5(7):525–530.

29. Hermanson KD, Lumsdon SO, Williams JP, Kaler EW, Velev OD (2001) Dielectropho-retic assembly of electrically functional microwires from nanoparticle suspensions.Science 294(5544):1082–1086.

30. Hoffman PD, Sarangapani PS, Zhu Y (2008) Dielectrophoresis and AC-induced as-sembly in binary colloidal suspensions. Langmuir 24(21):12164–12171.

31. Hölzel R, Calander N, Chiragwandi Z, Willander M, Bier FF (2005) Trapping singlemolecules by dielectrophoresis. Phys Rev Lett 95(12):128102.

32. Hughes MP, Morgan H (1998) Dielectrophoretic trapping of single sub-micrometrescale bioparticles. J Phys Appl Phys 31(17):2205–2210.

33. Matsue T, Matsumoto N, Uchida I (1997) Rapid micropatterning of living cells by re-pulsive dielectrophoretic force. Electrochim Acta 42(20–22):3251–3256.

34. Raychaudhuri S, Dayeh SA, Wang D, Yu ET (2009) Precise semiconductor nanowire

placement through dielectrophoresis. Nano Lett 9(6):2260–2266.35. Jain S, Gupta S (2013) Dielectrophoretic coassembly of binary colloidal mixtures in AC

electric fields. Langmuir 29(52):16105–16112.36. Kim J, Lee S, Francis Suh J-K, Ho Park J, Shin H-J (2013) Active control of dielec-

trophoretic force at nanowire electrode for ultrahigh single nanoparticle manipula-

tion yield. Appl Phys Lett 102(6):063105–063105–4.37. McMullan JM, Wagner NJ (2012) Directed self-assembly of colloidal crystals by di-

electrophoretic ordering. Langmuir 28(9):4123–4130.38. Liu Y, Chung J-H, Liu WK, Ruoff RS (2006) Dielectrophoretic assembly of nanowires.

J Phys Chem B 110(29):14098–14106.39. Sullivan MT, et al. (2006) An electric bottle for colloids. Phys Rev Lett 96(1):015703.40. Grzelczak M, Vermant J, Furst EM, Liz-Marzán LM (2010) Directed self-assembly of

nanoparticles. ACS Nano 4(7):3591–3605.41. Yethiraj A, Thijssen JHJ, Wouterse A, van Blaaderen A (2004) Large-area electric-field-

induced colloidal single crystals for photonic applications. Adv Mater 16(7):596–600.42. Erb RM, Segmehl J, Schaffner M, Studart AR (2012) Temporal response of magneti-

cally labeled platelets under dynamic magnetic fields. Soft Matter 9(2):498–505.43. Erb RM, Segmehl J, Charilaou M, Löffler JF, Studart AR (2012) Non-linear alignment

dynamics in suspensions of platelets under rotating magnetic fields. Soft Matter

8(29):7604–7609.44. Libanori R, Erb RM, Studart AR (2013) Mechanics of platelet-reinforced composites

assembled using mechanical and magnetic stimuli. ACS Appl Mater Interfaces 5(21):10794–10805.

45. Stampfl J, et al. (2008) Photopolymers with tunable mechanical properties processed by

laser-based high-resolution stereolithography. J Micromech Microeng 18(12):125014.46. Yoon E-S, et al. (2006) Tribological properties of bio-mimetic nano-patterned poly-

meric surfaces on silicon wafer. Tribol Lett 21(1):31–37.47. Singh A, Suh K-Y (2013) Biomimetic patterned surfaces for controllable friction in

micro- and nanoscale devices. Micro Nano Syst Lett 1(1):6.48. van Blaaderen A, Vrij A (1992) Synthesis and characterization of colloidal dispersions

of fluorescent, monodisperse silica spheres. Langmuir 8(12):2921–2931.49. Jiang P, McFarlandMJ (2005) Wafer-scale periodic nanohole arrays templated from two-

dimensional nonclose-packed colloidal crystals. J Am Chem Soc 127(11):3710–3711.50. Tervoort TA, Visjager J, Smith P (2002) On abrasive wear of polyethylene. Macromolecules

35(22):8467–8471.51. Oliver WC, Pharr GM (1992) An improved technique for determining hardness and

elastic modulus using load and displacement sensing indentation experiments.J Mater Res 7(6):1564–1583.

52. van Blaaderen A, Ruel R, Wiltzius P (1997) Template-directed colloidal crystallization.Nature 385(6614):321–324.

4628 | www.pnas.org/cgi/doi/10.1073/pnas.1524736113 Demirörs et al.