ARTICLE
Received 29 Apr 2015 | Accepted 11 Aug 2015 | Published 22 Sep 2015
E. Alonso-Redondo1, M. Schmitt2, Z. Urbach1, C.M. Hui3, R. Sainidou4, P. Rembert4, K. Matyjaszewski3, M.R. Bockstaller2 & G. Fytas1,5
The design and engineering of hybrid materials exhibiting tailored phononic band gaps are fundamentally relevant to innovative material technologies in areas ranging from acoustics to thermo-optic devices. Phononic hybridization gaps, originating from the anti-crossing between local resonant and propagating modes, have attracted particular interest because of their relative robustness to structural disorder and the associated benet to manufacturability. Although hybridization gap materials are well known, their economic fabrication and efcient control of the gap frequency have remained elusive because of the limited property variability and expensive fabrication methodologies. Here we report a new strategy to realize hybridization gap materials by harnessing the anisotropic elasticity across the particlepolymer interface in densely polymer-tethered colloidal particles. Theoretical and Brillouin scattering analysis conrm both the robustness to disorder and the tunability of the resulting hybridization gap and provide guidelines for the economic synthesis of new materials with deliberately controlled gap position and width frequencies.
DOI: 10.1038/ncomms9309 OPEN
A new class of tunable hypersonic phononic crystals based on polymer-tethered colloids
1 Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany. 2 Department of Materials Science and Engineering, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, Pennsylvania 15213, USA. 3 Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, USA. 4 Laboratoire Ondes et Milieux Complexes, UMR CNRS 6294, University of Le Havre, 75 Rue Bellot, 76600 Le Havre, France. 5 Department of Materials Science, FORTH-IESL, PO Box 1527, 71110 Heraklion, Greece. Correspondence and requests for materials should be addressed to G.F. (email: mailto:[email protected]
Web End [email protected] ).
NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 1
& 2015 Macmillan Publishers Limited. All rights reserved.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309
Phononic crystals (PnC), that is, composite materials in which a periodic distribution of elastic parameters facilitates control of the propagation of phonons, hold the
promise to enable transformative material technologies in areas ranging from acoustic and thermal cloaking to thermoelectric devices15. Realizing these opportunities requires strategies to deliberately engineer the phononic band structure of materials in the frequency range of interest. The typical approach involves the exploitation of Bragg-type phononic band gaps (BGs) that result from the destructive interference of waves in periodic media6 (Fig. 1a). Because the central frequency of a BG is directly related to the periodicity of the structure, that is, fBGBc/2a (where a and c denote the lattice parameter and the propagation velocity of elastic waves in the composite, respectively), the control of high-frequency phonons implies the ability to tailor the microstructure of hybrid materials on nanometer scales. Self-assembly processes have attracted particular attention as viable fabrication methodology to realize hybrid materials with appropriate periodicity in the nanoscale. For example, the assembly of colloidal particles into fcc-type crystal structures has been shown to enable the fabrication of hypersonic PnC materials to control phonons in the GHz regime79. However, the sensitivity of BG formation to structural disorder has limited the application of self-assembly methods that are generally susceptible to defect formation. Opportunities to overcome this challenge are provided by the formation of the so-called hybridization-type gaps (HGs), resulting from the avoided crossing of two bands of the same symmetry1014. Figure 1 contrasts the characteristics of BG and HG formation.
Often, at least one of these crossing bands originates from localized resonance modes of the individual particles (Fig. 1b). In this case, the frequency of the resulting HG is determined by the particles geometric and elastic characteristics, that is, fHGBcp/2L
(where L denotes a characteristic length of the particle and cp is
the propagation velocity characterizing the resonance mode). As a consequence of the dependence on local resonances, HGs are robust to disorder10,14, and can be tuned over a wider range of frequencies. Because fHG is related to the length scale of individual particles rather than lattice periods, the anti-crossing mechanism is able to produce phononic band gaps at lower frequencies as compared with BG-analoguesthus providing opportunities to further expand the range of frequencies over which control of phonon propagation can be accomplished. HG formation has been experimentally demonstrated in a range of particulate systems such as polymer spheres in water-like hosts10,13,14 or metal-in-polymer systems, such as millimetre-scale layered
spheres. The latter were rst applied by Sheng and co-workers to realize sonic structures operating in the kHz range15. However, although these structures did prove the concept of HG formation, they are of limited use for PnC applications because of their lack of mechanical stability (liquid suspensions), pronounced optical absorption (in the case of metals) that hampers any type of application based on elasto-optical coupling, as well as fabrication cost and limitation to the kHz frequency regime (micro-fabricated structures). Therefore, a viable methodology for the fabrication of HG materials, which lends itself to the scalable production of robust all-solid PnCs active over a wide high-frequency range, remains an outstanding challenge.
Here we demonstrate a new approach to facilitate HG formation in polymer nanocomposite materials that holds the promise to provide transformative opportunities for the use of self-assembly methods for the scalable fabrication of phononic materials, overcoming the limitations of existing methods to HG formation. This approach harnesses local anisotropy of the elastic parameters across the particle/polymer interface in colloidal hybrid particles rather than the density mismatch between constituents (appearing in metal-in-polymer all-solid composites). In particular, nanocomposite lms prepared by simple solution casting of densely polymer-tethered colloids are shown to exhibit a robust HG in the hypersonic range that is retained in the absence of long-range order. Rigorous theoretical calculations based on multiple-scattering theory (using imperfect boundary conditions, IBCs) reveal the mechanism of HG formation in particulate systems with anisotropic elastic coupling and provide guidelines for the synthesis of polymer-tethered colloids capable of forming HG at deliberately chosen frequencies in the hypersonic range.
ResultsPolymer-tethered colloids and their assemblies. The silica (SiO2)-polystyrene (PS) brush particles, with average radius
Rc 576 nm of the SiO2 core determined by transmission
electron microscopy (TEM), were synthesized using surface-initiated atom transfer radical polymerization as described previously16,17. The polymer grafting density s 0.50.6 chains
per nm2 was determined by elemental analysis and thermogravimetry. The graft characteristics of the particle systems, grafting density s and degree of polymerization N (130980) of the surface tethered chains are listed in Table 1 (sample ID: DPN). The centre-to-centre distance, d 2(Rc h), where h is the
brush height (Fig. 2a), is also listed in Table 1.
Particle brush lms with a thickness of about 50100 mm were prepared by casting from 3 wt% toluene solution (under ambient atmosphere) and subsequent thermal annealing at T 120 C for
24 h. We note that this preparation route is similar to established fabrication methods of colloidal crystal structures, in which weak compressive forces are being harnessed to drive the close-packing of colloidal particles and the annealing of defects to yield highly ordered structures18. The resulting crystal-type particle brush assembly structure is a prerequisite for the application of scattering theory to interpret the mechanism of HG formation. However, in contrast to BG in regular hard-sphere colloidal crystals, the HG in particle brush lms are robust with respect to structural disorder. To demonstrate this important feature, we also explored alternative preparation pathways such as rapid lm casting or the mixing of distinct types of brush particles that result in only short-ranged ordered materials (see also Discussion below).
To assess the microstructure of particle brush assemblies, reference lms of B1 mm thickness (corresponding to stacks of three or four particle brushes) were prepared following analogous
a
b
Flat resonant band
HG
q*</a
fBG ~ c
2a fHG ~
BG
q*=/a
cp 2L
Frequency
0
/a
0
/a
Wavenumber q
Figure 1 | Schematic representation of Bragg and hybridization gaps. (a) Structure-directed, Bragg (BG) gap occurring at a frequencyfBqBZ c/2p at the edge of the Brillouin zone (BZ), qBZ p/a, where c is the
sound velocity in the composite medium and a the lattice constant.(b) Hybridization (HG) gap is originating from an anti-crossing opening up at q*oqBZ and involving a local resonant mode that occurs at a frequency fHG related to the particle resonance characteristics (see the text).
2 NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
& 2015 Macmillan Publishers Limited. All rights reserved.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309 ARTICLE
Table 1 | Samples characteristics and longitudinal sound velocities in the assemblies.
Sample r (nm 2) N d (nm) cPS (m s 1) ceff-th (m s 1) ceff-exp (m s 1) DP100 0.61 130 142 2,820 2,688 2,710 DP400 0.61 400 176 2,562 2,482 2,480 DP600 0.56 630 201 2,526 2,469 2,420 DP1000 0.48 980 214 2,350 2,309 2,360
s, grafting density; N, degree of polymerization; d, distance core to core; c , polystyrene sound velocity; c , computed effective sound velocity; c , measured effective sound velocity.
PS
SiO2
a
50 nm
2h
100 nm
b
d
TEM
B
A
B
C
A
C
Figure 2 | Structure of gradient-interface particle brush colloidal crystal. (a) Bright-eld transmission electron microscopy image of particle brush lm revealing projection of a four particle-stack. In the micrograph, the sixfold intersection of contrast regions is indicative of fcc packing. The inset shows a magnied image of individual brush particles enlightening the inner shell of extended polymer segments (faint ring in the image). The scheme illustrates the conformational transition and interdigitation of tethered polymer chains on silica spheres. (b) Schematic of fcc packing structure of brush particles with interparticle distance d along with an illustration of the projection image that is expected for fcc structure along [111] direction (black arrow in left panel).
procedures and analysed using TEM after lm lift-off. Figure 2a depicts a representative bright-eld TEM of a particle brush (N 1,000) multilayer revealing uniform fcc-type packing of
spheres, identied by the characteristic sixfold contrast pattern of the o1114 projection (Fig. 2b). The result suggests that the driving force for order formation in particle brush systemsat least for the materials and process conditions applied in the present workresemble those of hard sphere systems for which fcc packing is typically observed7,19. Analysis of the particle surface-to-surface distance in particle monolayer structures (see inset of Fig. 2a for N 400) reveals the dependence of brush
height on the degree of polymerization of surface-tethered chains hBN0.80.1 thus conrming the pronounced stretching of surface-grafted chains20.
Recording hypersonic phononic dispersion relations. The experimental Brillouin light scattering (BLS) spectra were recorded for both in-plane and out-of-plane phonon propagation through the selection of the wave vector direction (q|| or q>), as indicated by the two scattering congurations in Fig. 3a,b. In addition, polarized and depolarized BLS spectra were acquired with longitudinal (vv) and transverse (vh) polarizations (see
Methods for details) to reveal the nature of the phonon propagation. In general, for homogeneous media, the transverse phonon is directly observed in the Ivh(o) spectrum, whereas the longitudinal phonon is observed in the isotropic Iiso(o) Ivv(o)-
xIvh(o); for uids21 xB4/3. For the known cases, for example, hard sphere colloidal crystals, Iiso(o) Ivv(o) because of the very
weak intensity of the Ivh(o) spectrum7. The situation is very different in the present particle brush assemblies. Polarized (vv) and depolarized (vh) BLS spectra are shown in the insets of Fig. 3c at one representative q|| value near the observed BG for three particle brush systems. The Ivh(o) contribution to Ivv(o) is unexpectedly signicant and the two spectra have qualitatively different shape. Although the former displays a single but complex peak structure with broad low-frequency wing, Ivv(o)
has a double peak structure that becomes increasingly pronounced with increasing brush thickness as evidenced by the spectra of DP600 (Supplementary Fig. 1).
To account for the features revealed by the experimental spectra, we rst consider the contribution of Iiso(o) that is obtained from
Ivv(o) and Ivh(o) by adjusting the value of x between 0.7 and 4/3. This range of values was theoretically derived for uids and was found to apply for the present spectra at long phonon wavelengths (low q||s). The analysis unequivocally reveals that the unusual asymmetric shape of Ivv(o) at low frequencies is due to the Ivh(o) contribution. Its subtraction unravels the longitudinal phonon polarization character of the Iiso(o) spectra, which are well represented by two Lorentzian lines (red solid lines in inset of Fig. 3c). At low q||, a single Lorentzian t is sufcient to represent the effective medium acoustic phonon (Supplementary Fig. 2). The dispersion plots constructed on the basis of the phonon frequency peaks are shown in Fig. 3c for three distinct particle brush systems. For the two propagation directions, q|| and q>, the dispersion relation is consistent with isotropic mechanical characteristics of lm samples. Note that peak positions of Ivh(o) (open symbols in
Fig. 3c) are both q|| and q> independent, hence identifying the corresponding modes as localized modes in real space. Owing to the asymmetry of the Ivh(o) at the low-frequency side, the open symbols in the dispersion relation of Fig. 3c refer to the peak position of the Ivh(o).
The experimental phononic band diagrams f(q) in Fig. 3c display three pertinent features. (i) In the low q|| range the dispersion is linear (red solid lines); its slope determines the effective medium sound velocity ceff that decreases with increasing PS fraction, as shown in Table 1. (ii) The spectra reveal a single phononic band gap at q*, with gap width Dfg that narrows with PS fraction. (iii) A new at band (q-independent frequency, open circles) is present with almost exclusive transverse polarization. The frequency of the at band ( fat)
decreases with increasing particle size and an anti-crossing with the acoustic branch near the band gap is absent, as evidenced by its evolution in the four systems (see Fig. 3c and Supplementary Fig. 1 for details). The latter implies a different symmetry for each of the two bands (as conrmed by the vh and vv analysis), thus preventing a HG at this band-crossing region, unlike the behaviour shown in Fig. 1b. Furthermore, near the q* region
NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 3
& 2015 Macmillan Publishers Limited. All rights reserved.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309
a Incident light
Scattered light
b
Incident light
[afii9825]
[afii9825] = ([afii9843] [afii9835])/2 [afii9825]
Scattered light
ks
ki
ks
q
ki
q||
[afii9825] = [afii9835]/2
c
0.0205 nm 0.0167 nm 0.0144 nm
f (GHz) f (GHz) f (GHz)
0.2 0.3
0.2
0.1
0.0
Frequency (GHz)
10
5
0 0.00 0.01 0.01
0.02 0.03
20
0.2
vh
vv
vh
Intensity (a.u.)
vv
vh
Intensity (a.u.)
Intensity (a.u.)
vv
0.1
0.1
15
0.0
0.0
5 10
5 10 5 10
DP100 DP400 DP1000
0.02 Wavenumber q (nm1)
0.03 0.02 0.03 0.04
0.00 0.01
0.00
Figure 3 | Recording the phononic band diagram. (a,b) Brillouin light scattering geometries probing phonon propagation along the wave vector q kski, with ks, ki being, respectively, the incident laser and scattered light wave vector. The direction of q is selected either in-plane ((a) transmission geometry) or normal to the plane ((b) reection geometry). In the transmission geometry, the magnitude q is tuned by varying the scattering angle y and is independent of the refractive index of the medium. (c) Experimental dispersion relation (frequency versus wavenumber q) for DP100, DP400 and DP1000 samples, obtained from the corresponding deducted BLS spectra (insets) recorded at a given q (vertical arrows) and tted as a sum of Lorentzian shapes (red lines). The deducted isotropic (grey) spectra is the difference between the intensities recorded in vv (black) and vh (blue) polarizations, Ivv-xIvh, with x being a variable factor between 0.7 and 4/3. In each plot, a clear band gap region (patterned area) and a localized mode (open circles) are observed. The red lines in the low q regime represent the effective medium acoustic mode; the dashed grey lines in the high-frequency branch are guides to the eye, connecting the data acquired with q perpendicular to the substrate plane (blue shaded area) using the reection geometry in a. The frequency of the at mode is indicated by a grey dashed arrow, and wavenumber number of the gap opening is marked with a black solid arrow, with an error of B0.002 nm 1.
there are three modes, in contrast to the experimental band structure of conventional colloidal crystals that display only the lower and upper gap edge frequencies79.
Structure determination and degree of order. To clarify the origin of band gap formation in particle brush lms (HG or BG), it is instructive to correlate the gap formation with the degree of order in these systems. The latter was estimated by evaluating the structural uniformity of particle brush monolayersa summary of the degree of order (measured in terms of the width of the distribution of normalized Voronoi cell areas20) for all particle brush systems is shown in Supplementary Figs 3 and 4. The analysis reveals that gap formation is robust against variations of the degree of order in particle brush systems. To further test the effect of structural disorder, the dispersion relation was determined for mixed binary DP100/DP400 and DP400/DP600 (Supplementary Fig. 5) particle brush lm structures and similarly for DP400/PS (Mw 10 k) with 37.5 wt% PS particle
brush/homopolymer blend systems, in which order is signicantly reduced as compared with the respective uniform analogues (not shown here). In both cases, the phononic band gap is found in the proximity of the corresponding gap of pristine DP400, thus demonstrating its robustness with respect to disorder (Supplementary Fig. 5). The robustness to disorder is indirect evidence for the HG origin of the gap and constitutes a major advantage with respect to fabrication of PnC materials by self-assembly methods. An insight into the origin of HG formation is provided by the analysis of the band diagram of the underlying (idealized) crystal structure.
As revealed by TEM analysis, the colloidal lms correspond to stacks of (111) layers of particles. Their periodic arrangement on a fcc lattice restricts the study of the dispersion relation f(q) only in the rst Brillouin zone (BZ), which in this case is a truncated octahedron (Fig. 4a). Its centre, G, corresponds to wave vectors q 0 and the high-symmetry direction [111] is pointed from G to
the zone hexagonal-face centres L, denoted equivalently as GL. For the scattering geometries of Fig. 3a, all possible experimental q vectors are conned in a plane perpendicular to GL, whose intersection with the BZ forms a hexagon (Fig. 4a). To reproduce theoretically, the experimentally obtained dispersion plots shown in Fig. 3c, one can select the direction of q along GM (Fig. 4a)
corresponding to [112] with M denoting the edge centre of the hexagon.
Modelling of phononic band diagrams. For the theoretical description of an innite fcc crystal of SiO2-core particles embedded in a PS matrix, we use the layered-multiple-scattering formalism22. We rst consider all materials to be homogeneous and isotropic with perfect boundary conditions (PBCs) applied across the SiO2PS interface (see Fig. 4b for details).
We use standard bulk values for PS except otherwise stated (mass density r 1,050 kg m 3, and longitudinal and transverse
elastic velocities cL 2,350 m s 1 and cT 1,210 m s 1,
respectively) and for SiO2 values xed from previous study23 (r 1,850 kg m 3, cL 4,910 m s 1 and cT 3,090 m s 1). In
Fig. 4c, we show the band structure diagram for the DP400 crystal (d 176 nm) along the high-symmetry direction GL, which is
similar to the one along the low-symmetry direction GM. Because
4 NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
& 2015 Macmillan Publishers Limited. All rights reserved.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309 ARTICLE
a
b
Perfect boundary conditions
SiO
PS
kz
L
X
6/d
K
M
M
Imperfect boundary conditions
k
SiO
PS
k
kx
ky
3/2d
L L M
2
Frequency (GHz)
10
8
6
4
2
01.0 0.5 0.0 1.0 0.5 0.0 0.5 1.0 1.5 qdL/ qdL/ qdM/
LHG HG
1
3
LHG
c d e
Figure 4 | Theoretical versus experimental band diagrams. (a) Reciprocal fcc lattice showing a [112] plane (orange hexagon) and the GM direction probed in the experiment. (b) Schematic presentation of PBC and IBC models applied at the SiO2-PS interface. In the IBC model, the springs of respective stiffness kL,T represent the discontinuity of the displacement eld across this interface. The theoretical band diagram of the DP400 is shown for PBC along [111] with the bulk PS sound velocities in c and for IBC with PS velocities (cL 2,560 m s 1, cT 1,320 m s 1) about 9% higher
than bulk PS along [111] (d) and [112] (e). Along GL dark/light solid and dotted blue lines denote non-degenerate (longitudinal), double degenerate (transverse) and deaf computed bands, respectively. In e, all bands are non-degenerate of mixed character; the at band is highlighted (dotted line). Solid and open circles indicate the experimental points (see Fig. 3). Hatched regions denote hybridization gaps (LHG for longitudinal modes; HG for all modes).
of its high symmetry, the choice of GL direction offers, apart from its numerical advantages (rapid calculation and convergence), a more straightforward interpretation of the symmetry24 and thus the polarization of the several eigenmodes: non-degenerate (L1 or L2) and double degenerate (L3), as explained in the Methods (Symmetry Considerations for Band Structure Analysis).
The calculated PBC-based band diagram fails to describe both the at experimental band at about 6 GHz and the higher acoustic branch, although the effective medium slope of the acoustic branch at the long wavelength limit is well reproduced. A small HG for longitudinal modes at about 8 GHz is observed in the vicinity of the BG frequency fBG, away from the experimental one (spanning from 6 to 7.5 GHz). We note that an attempt to mimic the radial evolution of the chains near the silica surface (Fig. 2a), through simple or multiple PS concentric shells with progressively varying elastic parameters, fails as well to describe the experimental behaviour.
To reduce the disparity between the calculated band structure and the experimental results, we introduce IBCs to account for anisotropic behaviour close to the SiO2PS interface (Fig. 4b).
This type of conditions has been recently applied to correctly reproduce the eigenmode frequencies of SiO2PS particle brush
powders in the air23. Although the use of the same stiffness values improves considerably the agreement between theory and experiment, a new readjustment of the PS elastic velocities (about 9% higher than those of bulk PS) and of tangential stiffness value, kT ( 0.0405 GPa nm 1), is needed to capture the
experimental dispersion plot, as shown in Fig. 4d. The effective medium slope is mainly governed by the former, whereas the latter tunes the position of the at (inactive) band. Clearly, nondegenerate bands are BLS active, whereas double degenerate ones are not easily discernible in the BLS spectra. The computed band structure along GM (Fig. 4e) reveals a better agreement between theory and experiment. In this case, because of the lower symmetry along [112], as compared with the high-symmetry [111], all bands are nondegenerate and BLS active to a greater or lesser degree depending on their hybrid polarization (for details, see the Methods section (Symmetry Considerations for Band Structure Analysis)).
To elucidate the origin of the at band, density-of-states (DOS) calculations were performed rst for an individual SiO2 particle (Fig. 5a) and subsequently for an innite (111) array of particles (Fig. 5b) embedded within PS matrix. The IBC model reveals for the case of a single SiO2 particle the existence of a set of triple ( 2l 1, due to spherical symmetry) degenerate dipole (l 1)
modes corresponding to spheroidal (f 4.82 GHz) and torsional
(f 5.20 GHz) polarization. The former exhibits a broad
resonance peak implying a short lifetime and strong leakage outside the sphere, whereas the latter corresponds to rotational modes with the maximum intensity eld occurring at the equator-level (inset of Fig. 5a) and well conned within the sphere (their narrow resonance peak implies a long lifetime). For an array of spheres in a (111) plane with interparticle distance d 201 nm (that is the DP600 case), the lower symmetry of the
system with respect to the single spherical particle leads to a level splitting of each triple state into a double degenerate and a single nondegenerate (Fig. 5d), as conrmed by the corresponding DOS calculation (Fig. 5b). Indeed, we observe three relatively narrow resonance modes, two of them double degenerate (f 5.32 and
6.78 GHz) and one nondegenerate (f 6.24 GHz), as well as one
quasi-bound extremely narrow peak at f 5.24 GHz. We note
that this mode is strictly bound (a delta function in DOS spectra) at normal incidence, and becomes active (a narrow Lorentzian-shaped peak in DOS) at a slightly-off normal incidence.
At resonance frequency, the corresponding displacement eld intensity plot within the unit cell and at a plane at the centres of the spheres (z 0) points out a strong localization inside the
particle and close to its surface (inset of Fig. 5b), decaying away from this plane. The almost exclusive rotational character originates from the torsional polarization of the single sphere mode. Its strongly localized nature leads to a very weak interaction when (111) planes are combined to form the crystal. Consequently, it results in a very at resonance band of L2 symmetry, as can be seen in Fig. 5c, with a frequency approximately equal to the individual particles torsional mode (Fig. 5d). On the contrary, the strong leakage of the rest of the (111) plane modes leads to much broader resonance bands of L1/L3 symmetry, for the fcc crystal along GL direction. They interact with the same-symmetry propagating bands that describe the effective medium assembly; this avoided crossing gives rise to the corresponding HG, which is shown in Fig. 5c. We note that the at band indicates also the low-frequency limit of the rst HG for transverse modes. The observation of the highly localized at band for the rst time at submicrometre scale could nd application as a ag for transverse wave ltering. On the other hand, these highly localized rotational modes could be of importance to future applications where energy harvesting inside the spheres is necessary. The theoretical analysis, apart from its
NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 5
& 2015 Macmillan Publishers Limited. All rights reserved.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309
q
[111]
a
q q
f/f c(%)
20
10
Frequency (GHz)
10
8
6
4
0
z=0
(111)
z
0
0 200 400 600 800 1,000
L L
N
Frequency (GHz)
9
8
7
6
5
4
3
2
1
0
1
HG
200 400 600 800 1,000Degree of polymerization N
HG
3
2
HG
b c
3
0.1
0.08
z
Slope 2.4 Slope 0.7
0.06
0.04
0.02
1
1 )
k T (GPa nm
3
a c
b
0
2
DOS (a.u. )
0 3 0.01 0.00 0.01
4 5
Frequency (GHz) Degree of polymerization N
6 7 8 9 100 500 1,000
One sphere One (111) plane 3D crystal along [111]
(3)
(3)
Frequency (GHz)
7
6
5
4
Figure 6 | Controlling the core-brush assemblies behaviour.(a) Evolution of the frequency gap region (shaded area) for all modes with N, following the theoretical predictions (see the text) along GM. Inset: normalized gap width (fc being the central frequency of the gap). Variation of the tangential stiffness values kT with (b) the frequency of the at band fat calculated in the middle of the band along GM (qdGM 0.5p), and with
(c) the degree of polymerization N of the assemblies considered, showing a power-law behaviour in both cases.
=1Torsionalmode Flat band
Hybridization gaps
(2)
(2)
(1)
(1)
=1 Spheroidal mode
d
Dimensionality
Wavevector q (nm1)
Figure 5 | Physical origin of the dispersion characteristics and their evolutive formation with dimensionality. Calculated DOS for (a) one SiO2 particle in PS (solid/broken lines spheroidal/torsional l 1 modes) and
(b) for one (111) plane of particles of the DP600 lm at slightly off-normal incidence (dark/light blue lines: non-degenerate/double degenerate modes). The mode corresponding to the very sharp peak in DOS plot of (111) plane, is associated to the at band of the crystal (see c). Its eld intensity representation within the unit cell passing at the centre of the sphere (z 0), at f 5.24 GHz, is shown in the inset. White arrows
represent the quasi-pure rotational character of the eld everywhere in the unit cell. (c) The band structure of the corresponding crystal along [111], with non-degenerate (left) and double degenerate (right) calculated bands together with the experimental points (symbols). The notation used is that of Fig. 3. (d) Schematic representation of the evolution of these modes when passing from a single sphere to the whole crystal. Parenthesized numbers denote the number of states per level. In the case of three-dimensional (3D) crystal, the levels indicating the hybridized modes correspond to HG centres.
obvious importance for designing tunable robust hypersonic HGs, unravels these features of potential interest.
DiscussionThe characteristics of elastic wave propagation in particle brush assembly structures are distinctively different from those of known systems with low-density mismatch components (silica in polymer is a typical candidate) where HGs occur in the vicinity of the rst BG for longitudinal modes (Fig. 4c). These densely polymer-tethered particle assemblies exhibitthanks to their local anisotropy at the surface that is modelled here by IBCHGs that can be deliberately tuned by modication of the brush height for a given silica-core size towards lower frequencies (Supplementary Fig. 6). This is remarkable since up to now this effect has only been known to exist in high-density mismatch components (like metal-in-polymer composites). In the present
theoretical framework, the experimental dispersion relations can only be captured by adjustment of sound velocities of PS and the tangential stiffness across the particle/polymer interface. To our knowledge, the control of the HG characteristics (frequency position and width) by introducing interface elastic anisotropy has never been demonstrated before and its features strongly suggest a new class of HG-phononic materials. The rst evidence (shown in Fig. 6a) are the frequencies of the gap region and the decreasing of the gap-width of the HG with increasing degree of polymerization (N) of the grafted chains. This trend reects a dilution effect because of the decrease of the SiO2 core volume fraction.
Second, the assumed value of the longitudinal sound velocity for PS cPS (Table 1) equals the corresponding bulk PS value (cL)
only for the longest PS graft (DP1000). For the shorter grafts, cPS
exceeds cL and the difference increases to 20% for the shortest PS graft (DP100). This systematic trend with graft length (Table 1) is consistent with the increasingly (radially) stretched chain conformation that is expected for shorter brush lengths. Our results indicate that the pronounced stretching of surface-grafted chains leads to strain hardening, resulting from limited chain extensibility that yields the increase in the modulus (and hence cPS) normal to the interface. We note that this observation supports a prior report of stiffening in planar polymer brushes25.
Furthermore, the assumed values of cPS reproduce (within 1%) the effective medium velocity ceff (6th and 7th columns in Table 1). As expected ceff is less than the matrix sound
velocity (cPS) for solid (SiO2) inclusions in solid matrices (PS), which support shear waves26. The same trend is theoretically predicted in the case of all-solid composites involving PBC27, although these models fail to describe quantitatively IBC-type systems.
Third, the quantity kT captures the increase of the at mode frequency (fat) with decreasing degree of polymerization, that is,
6 NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
& 2015 Macmillan Publishers Limited. All rights reserved.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309 ARTICLE
from DP100 to DP1000 (Fig. 3c and Supplementary Fig. 1). This trend is visualized in Fig. 6b,c. In view of the proposed mechanism (Fig. 5d), the coupling of the core eigenmode should be stronger for larger stiffness constant and hence the observed scaling in Fig. 6b should be system dependent. The dependence of kT with N is depicted in Fig. 6c. Interestingly, the dependence of stiffness constant on the degree of polymerization of grafted chains is found to be described by the scaling relation kTBN 0.7.
Although the decrease of kT with increasing N could be anticipated, the particular scaling is an unexpected nding. It reects an almost reciprocal relationship to the brush height (hBN0.80.1) and suggests a direct relation between the chain conformational anisotropy and the apparent interfacial stiffness. Although a more detailed understanding of the origin of the scaling relationship is currently lacking, it is important to point out its role as a simple design guideline for the synthesis of particle brush materials with engineered phononic hybridization gap.
In conclusion, we have demonstrated that the self-assembly of densely polymer-tethered colloidal particles gives rise to phononic materials with HGs that are robust to disorder. Theoretical analysis reveals that the origin of HG formation in particle brush assemblies is the hybridization of local resonant modes controlled by the anisotropic elastic coupling across the particle/brush interface. The analysis further reveals that the phononic properties can deliberately be tuned by variation of particle (and polymer) composition, particle size and the degree of polymerization of tethered chains. In contrast to established metamaterial systems, the new process does not require expensive micro-fabrication but can rather be accomplished by the dense tethering of polymeric chains to the surface of colloidal particles. The ability to synthesize suitable core-shell particles using economic polymerization methods that are already being used on industrial scale in conjunction with the formability of particle brush materials and the demonstrated robustness of HG, renders this new material approach an intriguing candidate for the viable fabrication of phononic hybrid materials promoting the development of a wide range of innovative phononic material technologies.
Methods
Particle brush synthesis. Silica particles (Rc 576 nm) were obtained from
Nissan Chemicals. Styrene (S) was obtained from Aldrich and puried by passing through an alumina-lled column. The synthesis of PS-grafted particle brush systems was performed using surface-initiated atom transfer radical polymerization as described previously16,17. The polymer grafting density s 0.50.6 chains per
nm2 was determined by elemental analysis as well as thermogravimetry. The molecular weight distribution of surface-grafted chains was determined by size exclusion chromatography after dissolution of the particle core in hydrouoric acid. The graft characteristics of the particle systems, grafting density s, degree of polymerization N ( 130980) of the surface tethered chains are listed in Table 1.
Brillouin spectroscopy. Brillouin spectroscopy utilizes the scattering of a probe laser beam from thermally activated phonons along a specied direction dened by the scattering vector q. The magnitude of the scattering vector, dened asq ks-ki, is independent from the refractive index in transmission geometry
(q|| 4p/l sin[y/2], y is the scattering angle and l 532 nm is the wavelength of
the incident light, Fig. 3a), whereas in reection geometry, q>(n) depends on the refractive index (Fig. 3b). The inelastic interaction between the incident photons and thermal phonons is evidenced in the frequency shift f(q) of the BLS spectrum at hypersonic (GHz) frequencies, resolved by a high-resolution tandem FabryPerot interferometer (JRS Instruments). Longitudinal (transverse) displacements have associated a vv (vh) polarization, selected by the input polarizer(v) and output analyser (v or h); v(h) denotes vertically (horizontally) polarized light relatively to the scattered plane.
Theoretical calculations. Multiple-scattering theory is applied to describe the propagation of elastic waves in the colloidal assemblies and calculate the dispersion relation (band structure) for such periodic systems22. The essence of this method consists in the multipole expansion of the elastic eld in each region containing an isotropic and homogeneous material. Discontinuities are treated through interfaces
on which appropriate boundary conditions (BCs) are applied. For instance, in the case of an ensemble of spherical bodies embedded in a host, the elastic eld is expanded into a basis of spherical waves characterized by the angular momentum l, the azimuthal number m and by the polarization P (longitudinal L, and transverse, M or N). For a single sphere, the elastic eigenmodes are organized in two independent (uncoupled) subgroups: the torsional modes (M polarized) and the spheroidal modes (L and N polarized). The application of the appropriate boundary conditions on the surface of the spherical particle leads to the scattered eld by that particle, described by the corresponding transition T-matrix. When bringing together the ensemble of particles, one needs to consider the auto-consistent eld derived by the multiple-scattering process of elastic waves between the particles, in order to describe accurately the scattering by the aggregate; key quantities are the propagation matrix X and the T-matrix of the individual particles. The difference in the DOS of the elastic eld with respect to the innite host matrix (PS), for one particle or an ensemble of such particles, is given by
Dn o
1p @@oImTr ln I T
and Dn o
1p @@oIm Tr ln I T
Tr ln I TO
,
respectively26. The dispersion relation of the elastic modes of a crystal composed by these particles is also derived from these two key matrices, T and I TX. Here, we
apply two different types of boundary conditions on the interface between SiO2 cores and PS host: (i) PBCs assuming continuity of all radial and tangential components of the displacement eld and surface traction, and (ii) IBCs introducing discontinuity of the displacement eld components across the interface (surface traction remains continuous) through a stiffness coefcient k. Note that k is different for the displacement components normal (kL) and tangential (kT) to the spherical surface. This model was originally developed for cylindrical inclusions28 and was recently adapted for spherical particles29. It has been successfully compared with experimental data for individual core-brush particles in the air23.
Symmetry considerations for band structure analysis. Following a group theory analysis24, only three different-symmetry groups of bands are observed: double degenerate active (L3) and nondegenerate active (L1) bands whose modes are excited by a transverse and longitudinal wave incident normally on a nite (111) slab of the crystal (that is, along GL), and nondegenerate inactive (deaf, L2) bands whose modes cannot be excited by any type of elastic waves propagating along GL. The latter become active when longitudinal waves are incident on the structure at a slightly off-normal incidence to the nite (111) slab of the crystal (that is, along a direction slightly different from GL). In a real experiment, all nondegenerate bands will become active (although not to the same degree), as experimentally small angular deviations from a specic direction of propagationhere GLalways exist.
Therefore, one expects that these non-degenerate bands are BLS active in our case. Double degenerate bands are associated to modes coupled to transverse elastic waves incident on the slab. One expects that these double degenerate bands are BLS inactive along GL direction. For directions of lower symmetry, the above scheme is no longer valid. For instance, along GM (which is different from but close to GL) all bands become non-degenerate, but their modes remember the polarization character along GL: they are now hybrid but with a large percentage remaining in the polarization from the corresponding high-symmetry branch.
References
1. Maldovan, M. Sound and heat revolutions in phononics. Nature 503, 209217 (2013).
2. Han, T. et al. Experimental demonstration of a bilayer thermal cloak. Phys. Rev. Lett. 112, 054302 (2014).
3. Xu, H., Shi, X., Gao, F., Sun, H. & Zhang, B. Ultrathin three-dimensional thermal cloak. Phys. Rev. Lett. 112, 054301 (2014).
4. Kraemer, D. et al. High-performance at-panel solar thermoelectric generators with high thermal concentration. Nat. Mater. 10, 532538 (2011).
5. Kirihara, A. et al. Spin-current-driven thermoelectric coating. Nat. Mater. 11,
686689 (2012).
6. Martnez-Sala, R. et al. Sound-attenuation by sculpture. Nature 378, 241241 (1995).
7. Cheng, W., Wang, J. J., Jonas, U., Fytas, G. & Stefanou, N. Observation and tuning of hypersonic bandgaps in colloidal crystals. Nat. Mater. 5, 830836 (2006).
8. Wu, S. et al. Anisotropic lattice expansion of three-dimensional colloidal crystals and its impact on hypersonic phonon band gaps. Phys. Chem. Chem. Phys. 16, 89218926 (2014).
9. Zhu, G. et al. Direct observation of the phonon dispersion of a three-dimensional solid/solid hypersonic colloidal crystal. Phys. Rev. B 88, 144307 (2013).
10. Still, T. et al. Simultaneous occurrence of structure-directed and particle-resonance-induced phononic gaps in colloidal lms. Phys. Rev. Lett. 100, 194301 (2008).
11. Lemoult, F., Kaina, N., Fink, M. & Lerosey, G. Wave propagation control at the deep subwavelength scale in metamaterials. Nat. Phys 9, 5560 (2012).
12. Thomas, E. L. Bubbly but quiet. Nature 462, 990991 (2009).13. Cowan, M. L., Page, J. H. & Sheng, P. Ultrasonic wave transport in a system of disordered resonant scatterers: propagating resonant modes and hybridization gaps. Phys. Rev. B 84, 094305 (2011).
NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 7
& 2015 Macmillan Publishers Limited. All rights reserved.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9309
14. Beltramo, P. J., Schneider, D., Fytas, G. & Furst, E. M. Anisotropic hypersonic phonon propagation in lms of aligned ellipsoids. Phys. Rev. Lett. 113, 205503 (2014).
15. Liu, Z. Y. et al. Locally resonant sonic materials. Science 289, 17341736 (2000).
16. Hui, C. M. et al. Surface-initiated polymerization as an enabling tool for multifunctional (nano-) engineered hybrid materials. Chem. Mater. 26, 745762 (2014).
17. Choi, J., Dong, H., Matyjaszewski, K. & Bockstaller, M. R. Flexible particle array structures by controlling polymer graft architecture. J. Am. Chem. Soc. 132, 1253712539 (2010).
18. Freymann, G., Kitaev, V., Lotsch, B. & Ozin, G. A. Bottom-up assembly of photonic crystals. Chem. Soc. Rev. 42, 25282554 (2013).
19. Woodcock, L. V. Computation of the free energy for alternative crystal structures of hard spheres. Faraday Discuss. 106, 325338 (1997).
20. Choi, J. et al. Effect of polymer-graft modication on the order formation in particle assembly structures. Langmuir 29, 64526459 (2013).21. Berne, B. J. & Pecora, R. Dynamic Light Scattering (Dover, 2000).22. Sainidou, R., Stefanou, N., Psarobas, I. E. & Modinos, A. A layer-multiple-scattering method for phononic crystals and hetero structures of such. Comput. Phys. Commun. 166, 197240 (2005).
23. Schneider, D. et al. Role of polymer graft architecture on the acoustic eigenmode formation in densely polymer-tethered colloidal particles. ACS Macro Lett. 3, 10591063 (2014).
24. Cornwell, J. F. Group Theory and Electronic Energy Bands in Solids (North-Holland, 1969).
25. Urayama, K., Yamamoto, S., Tsujii, Y., Fukuda, T. & Neher, D. Elastic properties of well-dened, high-density poly(methyl methacrylate) brushes studied by electromechanical interferometry. Macromolecules 35, 94599465 (2002).
26. Sainidou, R., Stefanou, N. & Modinos, A. Greens function formalism for phononic crystals. Phys. Rev. B 69, 064301 (2004).
27. Gaunaurd, G. C. & Wertman, W. Comparison of effective medium theories for inhomogeneous continua. J. Acous. Soc. Am. 85, 541554 (1988).
28. Huang, W., Rokhlin, S. I. & Wang, Y. J. Analysis of different boundary condition models for study of wave scattering for ber-matrix interfaces.J. Acous. Soc. Am 101, 20312042 (1997).
29. Sainidou, R. & Rembert, P. in PHONONICS 2015 (3rd International Conference on Phononic Crystals/Metamaterials, Phonon Transport and Phonon Coupling, Book of Abstracts) 318-319 (Paris, France, 2015).
Acknowledgements
M.R.B. and K.M. acknowledge nancial support by the National Science Foundation (via DMR-1410845 and DMR-1501324) as well as the Department of Energy(via DE-EE0006702). The work was supported partially by Aristeia Program-285 (EU, GSST Greece), and P.S. POLY 3333 Project (DFG). We thank Dr Schneider for technical assistance in the BLS experiment.
Author contributions
E.A.-R. and Z.U. contributed to the BLS measurements. C.M.H. and K.M. synthesized the particle brush systems, and M.S. and M.B. fabricated the samples. R.S. and P.R. developed the theoretical description of the dispersion diagrams. G.F. planned the project and together with M.B., R.S. and P.R. wrote the article. All authors have discussed the results and commented on the manuscript.
Additional information
Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications
Web End =http://www.nature.com/ http://www.nature.com/naturecommunications
Web End =naturecommunications
Competing nancial interests: The authors declare no competing nancial interests.
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/
Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/
Web End =reprintsandpermissions/
How to cite this article: Alonso-Redondo, E. et al. A new class of tunable hypersonic phononic crystals based on polymer-tethered colloids. Nat. Commun. 6:8309doi: 10.1038/ncomms9309 (2015).
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
Web End =http://creativecommons.org/licenses/by/4.0/
8 NATURE COMMUNICATIONS | 6:8309 | DOI: 10.1038/ncomms9309 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
& 2015 Macmillan Publishers Limited. All rights reserved.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Copyright Nature Publishing Group Sep 2015
Abstract
The design and engineering of hybrid materials exhibiting tailored phononic band gaps are fundamentally relevant to innovative material technologies in areas ranging from acoustics to thermo-optic devices. Phononic hybridization gaps, originating from the anti-crossing between local resonant and propagating modes, have attracted particular interest because of their relative robustness to structural disorder and the associated benefit to 'manufacturability'. Although hybridization gap materials are well known, their economic fabrication and efficient control of the gap frequency have remained elusive because of the limited property variability and expensive fabrication methodologies. Here we report a new strategy to realize hybridization gap materials by harnessing the 'anisotropic elasticity' across the particle-polymer interface in densely polymer-tethered colloidal particles. Theoretical and Brillouin scattering analysis confirm both the robustness to disorder and the tunability of the resulting hybridization gap and provide guidelines for the economic synthesis of new materials with deliberately controlled gap position and width frequencies.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer