User login

You are here

Journal Club for March 2018: Colloidal Self-Assembly of Architected Nanomaterials

xwgu's picture

Structural hierarchy through self-assembly

Architected materials with nano and microscale features can have extraordinarily high strength, stiffness, and toughness per weight. Two well-known biological materials with these characteristics are nacre (mother of pearl)1 and the claw of the mantis shrimp.2 These materials are made of up periodic arrangements of hard domains interspersed with soft matter (Fig. 1). Structural hierarchy exists at different length scales, and varies spatially across the material.

Figure 1. Hierarchically architected biological composites. A) “Brick and mortar” arrangement of calcium carbonate platelets in nacre (image from ref. 1). B) Oriented chitin fibrils in stomatopod dactyl club (image from ref. 2).

It has been challenging to reproduce this structural complexity and the resulting properties in engineering materials. Conventional fabrication techniques are not sufficient, especially at the nanoscale. One promising approach for the fabrication of architected nanomaterials with superior mechanical properties is self-assembly of hybrid organic-inorganic colloidal nanoparticles. Self-assembly is the process of forming an organized structure from disordered parts without the application of external forces.3 This concept is borrowed from nature – biology routinely uses self-assembly to build structures at the molecular to the cellular level - but self-assembly can even be used to construct furniture.4 The self-assembly of colloidal nanoparticles utilizes interparticle forces (e.g. Van der Waals, electrostatic, entropic) to form solid structures while immersed in solution or at fluid interfaces. Nanoparticle solids have many levels of structural hierarchy, many analogous to those in atomic crystals, spanning the molecular scale to the scale of the overall structure (see Figure 2). The fundamental building block is the inorganic nanoparticle, which is surrounded by a layer of organic ligands that stabilizes the nanoparticle surface. Organic ligands on neighboring nanoparticles are in contact in self-assembled nanoparticle solids, but interactions between ligands and inorganic nanoparticles, and between inorganic cores also play a part in self-assembly and the resulting mechanical properties.


Figure 2. Structural hierarchy at the level of individual ligands, close-packed arrays of nanoparticles (superlattice), and crystalline grains formed from nanoparticles (supercrystal). This structural richness leads to highly tunable mechanical properties, and strength, stiffness and hardness. Multifunctional nanoparticle solids can be fabricated by incorporating nanoparticles with novel optical, electrical, magnetic and catalytic properties.


Strong and flexible membranes

Nanoparticle solids can be thin films or 3D supercrystals. The mechanical properties of a nanoparticle monolayer film were first reported by Heinrich Jaeger and coworkers.5 Au nanoparticles covered in dodecanethiol ligands were assembled into monolayers and suspended on a holey silicon nitride support to form strong and flexible freestanding membranes. Atomic force microscopy was used to probe the elastic properties and breaking force of the suspended nanoparticle membranes (Figure 3). The films were found to have an elastic modulus of ~6 GPa (revised to 4 GPa in a subsequent publication6), which is surprisingly high considering the lack of specific interactions between ligands (e.g. hydrogen bonding, electrostatic) and the low intrinsic modulus of dodecanethiol, which is a liquid in bulk form. The high elastic modulus was attributed to Van der Waals forces between densely interdigitated ligands on neighboring particles,5,7 as the ligands are too short for entanglement. AFM deflection on membranes made with CoO-oleic acid and Fe/Fe3O4-oleylamine nanoparticles revealed that ligand-nanoparticle bonding and the packing density of ligands on the nanoparticle surface also contribute to the mechanical properties.6 Nanocrystal monolayers formed from DNA-ligated Au nanoparticles were also found to be surprisingly strong and flexible.8


Figure 3. A,B) SEM images of self-assembled nanoparticle membranes suspended on a SiN support. C) Membrane deflection force-displacement curve for different membrane diameters (images from ref. 5).

Properties of 3D supercrystals

Nanoindentation into a thick film made of CdSe nanocrystals resulted in a rate-dependent elastic modulus of around 8 GPa.9 When ligands on the nanocrystals are removed using solvent, the elastic modulus is reduced to ~2 GPa and the supracrystal densifies like a granular material. Podsiadlo et al. performed indentation on crystalline supercrystals and disordered films made of nanoparticles of various compositions (PbS, CdSe, CoPt3), sizes, and capping ligands.10 Elastic modulus and hardness was higher in the supercrystals than the films, and for particles of larger size because of stronger particle-particle interactions. Diamond anvil cells have been used to hydrostatically compress supercrystals made with PbS-oleic acid nanoparticles, while observing structural changes with small angle X-ray scattering and X-ray diffraction (Figure 4).11 Bulk modulus was found to be ~51 GPa, which corresponds to a high ligand elastic modulus of ~2.5 GPa. The supercrystal maintained its lattice structure up to 55 GPa, even after the nanocrystals undergo a pressure-induced phase transformation.


Figure 4. A) Supercrystals formed from PbS nanoparticles. Scalebar on inset is 30 nm. B) Optical image of supercrystal and ruby powder (internal pressure gauge) in diamond anvil cell chamber. Images from ref. 11.


Polymer-grafted nanoparticle superlattices reveal structure-property relationships

The nanoparticle monolayers discussed above have similar mechanical properties despite differences in ligand chemistry and structure. It is unclear how the structure of the nanoparticle arrays, (e.g. nanoparticle vacancies and interstitials, close-packed vs. disordered arrays) and ligand conformation affect mechanical properties. Our group explored these questions using polymer-grafted nanoparticles.12 Thiolated polystyrene with a molecular weight of up to 20 kg/mol (50 kg/mol is the threshold for entanglement) was attached to Au nanoparticles using ligand exchange. These nanoparticles were self-assembled at an air-liquid interface at various rates by controlling temperature and partial pressure (closed and open vessels). Superlattices with defects were formed under fast self-assembly conditions (Figure 5). A buckling based method was used to measure the bending and elastic modulus of few-layered superlattice films. It was found that increased lattice disorder results in an elevated elastic modulus, up to ~19 GPa, which is far higher than the modulus predicted by the Halpin-Tsai model (~3 GPa). Elastic modulus increased with increasing polystyrene molecular weight in few-layered films, but this relationship was not observed in thick superlattice films measured using nanoindentation. These results indicate that the polystyrene ligands take on non-equilibrium configurations under fast self-assembly conditions, and polymer conformation governs mechanical properties rather than the spatial arrangement of the inorganic nanoparticles.


Figure 5. Superlattice structure and mechanical behavior for 3-kg/mol PS-Au superlattices self-assembled at different drying rates. Representative (A–D) TEM images and (E–H) corresponding statistical analysis of the distribution of interparticle spacing for superlattice dried at (A and E) room temperature while covered with a glass slide (RT-covered); (B and F) RT-uncovered; (C and G) 43 °C-uncovered; and (D and H) 58 °C- uncovered. E–H are plotted as a function of particle center-to-center distance, D, over particle diameter, d. (Scale bars, 20 nm.) (I) Bending modulus of 3-kg/mol PS-Au superlattices and calculated bending modulus for bulk Au and 3-kg/mol PS based on literature values (37). (J) Elastic modulus of 3-kg/mol PS-Au super- lattices with thicknesses equivalent to two and three superlattice layers. Figure and caption are reproduced from ref. 12.

Nanocomposites have also been self-assembled under equilibrium conditions from “hairy nanoparticles” that are synthesized by growing polymers on the surface of inorganic nanoparticles.13,14 These studies use longer polymers and lower grafting densities, which leads to more compliant and ductile composites.


Chemical modification of self-assembled superlattices

The structure and properties of close-packed superlattices can be modified after assembly. Dreyer et al. assembled Fe3O4-oleic acid nanoparticles into 3D superlattices which were subsequently heated to induce crosslinking of oleic acid molecules on neighboring particles.15 The resulting material exhibits high strength (~500 MPa), hardness (~4 GPa) and elastic modulus (up to ~80 GPa) which rivals that of other synthetic nanocomposites.16 The dense interdigitation of oleic acid ligands leads to the formation of a stiff array of short alkyl chains after crosslinking, which gives rise to the observed mechanical properties. In another recent study, a mixture of Au and Zn0.2Fe0.8O nanocrystals were formed into ~115 nm rods.17 A chemical treatment was used to remove the organic ligands from the surface of the nanocrystals, which induced sintering of Au nanocrystals to form a porous structure filled with Zn0.2Fe0.8O nanocrystals. The resulting nanorods have an elastoplastic stress-strain response, and yield strength and ductility slightly below that of pure Au nanorods (Fig. 6). In addition, these nanorods combined the magnetic properties of the Zn0.2Fe0.8Onanocrystals and the plasmonic properties of the Au nanocrystals, such that a magnetic field can control the transmission of infrared light through nanorods.


Figure 6. Au-Zn0.2Fe0.8O hybrid nanorods. A) Nanorods are fabricated by depositing nanoparticles into a nanoimprinted template, removing the template, and then removing the native ligands on the nanoparticles. B) SEM image of array of nanorods. C) Tensile stress-strain responses of pure Au nanorod and hybrid nanorod. Images from ref. 17.


Complex architectures: beyond close-packed spheres

Superlattices made with spherical nanoparticles are limited to close-packed structures (e.g. face-centered cubic, hexagonal closed packed), and disordered assemblies. Complex, 3D superlattices can be self-assembled from anisotropic nanoparticles, or mixtures of different types and sizes of spherical nanoparticles (Figure 7).18–20


Figure 7. A) Herringbone-like tiling of TbF3 hexagonal nanoplates (from ref. 18). B) 3D superlattice of truncated octahedral Ag nanocrystals (from ref. 20). C) Ternary nanocrystal superlattice film made from 16.5 nm Fe3O4, 7 nm Fe3O4, and 5 nm FePt nanocrystals (from ref. 19).

It remains extremely challenging to fabricate complex, 3D superlattices that are large enough for the measurement of mechanical properties (>1 μm). The self-assembly of anisotropic nanoparticles occurs under a narrow range of processing conditions (e.g. concentration, particle size distribution, ligand coverage, temperature, pressure) that is not well understood. To my knowledge, there is only one study on the mechanical properties of a complex, 3D superlattice. Superlattices formed from close-packed CdSe/CdS octapods were tested in compression, and found to have a stress-strain response typical of a cellular foam (Figure 8).21 Elastic modulus was found be around 6 GPa, but yield strength and the dependence of mechanical properties on relative density and beam dimensions were not reported.


Figure 8. A) TEM image of CdS/CdSe octapod. B-D) SEM images of assembled structure. E) Schematic of assembly process. F) Load-displacement curve (from ref. 21).


Future directions

Further progress on the mechanical properties of self-assembled nanomaterials requires an improved understanding of the self-assembly process, which would allow the fabrication of large superlattices with a variety of lattice structures. Self-assembly could also be combined with other fabrication techniques to form larger structures. For instance, “bricks” made from self-assembled colloidal nanoparticles could be embedded in a matrix, or 3D printed as standalone structures or coatings. The mechanical properties of nanoparticle solids are mostly due to the organic ligands, so it is of interest to develop stimulus responsive or self-healing ligands to expand the existing abilities of self-assembled nanomaterials.



(1) Barthelat, F. J. Mech. Phys. Solids 2014, 73, 22–37.

(2) Weaver, J. C.; Milliron, G. W.; Miserez, A.; Evans-Lutterodt, K.; Herrera, S.; Gallana, I.; Mershon, W. J.; Swanson, B.; Zavattieri, P.; DiMasi, E.; Kisailus, D. Science 2012, 336 (6086), 1275–1280.

(3) Bishop, K. J. M.; Wilmer, C. E.; Soh, S.; Grzybowski, B. A. Small 2009, 5 (14), 1600–1630.

(4) Tibbits, S. Fluid-Assembly Chair

(5) Mueggenburg, K. E.; Lin, X.-M.; Goldsmith, R. H.; Jaeger, H. M. Nat. Mater. 2007, 6 (9), 656–660.

(6) He, J.; Kanjanaboos, P.; Frazer, N. L.; Weis, A.; Lin, X.-M.; Jaeger, H. M. Small 2010, 6 (13), 1449–1456.

(7) Landman, U.; Luedtke, W. D. Faraday Discuss. 2004, 125 (0), 1.

(8) Cheng, W.; Campolongo, M. J.; Cha, J. J.; Tan, S. J.; Umbach, C. C.; Muller, D. A.; Luo, D. Nat. Mater. 2009, 8 (6), 519–525.

(9) Lee, D.; Jia, S.; Banerjee, S.; Bevk, J.; Herman, I. P.; Kysar, J. W. Phys. Rev. Lett. 2007, 98 (2), 26103.

(10) Podsiadlo, P.; Krylova, G.; Lee, B.; Critchley, K.; Gosztola, D. J.; Talapin, D. V.; Ashby, P. D.; Shevchenko, E. V. J. Am. Chem. Soc. 2010, 132 (26), 8953–8960.

(11) Podsiadlo, P.; Lee, B.; Prakapenka, V. B.; Krylova, G. V.; Schaller, R. D.; Demortière, A.; Shevchenko, E. V. Nano Lett. 2011, 11 (2), 579–588.

(12) Gu, X. W.; Ye, X.; Koshy, D. M.; Vachhani, S.; Hosemann, P.; Alivisatos, A. P. Proc. Natl. Acad. Sci. U. S. A. 2017, 114 (11), 2836–2841.

(13) Schmitt, M.; Hui, C. M.; Urbach, Z.; Yan, J.; Matyjaszewski, K.; Bockstaller, M. R. Faraday Discuss. 2016, 186 (0), 17–30.

(14) Koerner, H.; Drummy, L. F.; Benicewicz, B.; Li, Y.; Vaia, R. A. ACS Macro Lett. 2013, 2 (8), 670–676.

(15) Dreyer, A.; Feld, A.; Kornowski, A.; Yilmaz, E. D.; Noei, H.; Meyer, A.; Krekeler, T.; Jiao, C.; Stierle, A.; Abetz, V.; Weller, H.; Schneider, G. A. Nat. Mater. 2016, 15 (5), 522–528.

(16) Podsiadlo, P.; Kaushik, A. K.; Arruda, E. M.; Waas, A. M.; Shim, B. S.; Xu, J.; Nandivada, H.; Pumplin, B. G.; Lahann, J.; Ramamoorthy, A.; Kotov, N. A. Science 2007, 318 (5847), 80–83.

(17) Zhang, M.; Magagnosc, D. J.; Liberal, I.; Yu, Y.; Yun, H.; Yang, H.; Wu, Y.; Guo, J.; Chen, W.; Shin, Y. J.; Stein, A.; Kikkawa, J. M.; Engheta, N.; Gianola, D. S.; Murray, C. B.; Kagan, C. R. Nat. Nanotechnol. 2017, 12 (3), 228–232.

(18) Ye, X.; Chen, J.; Engel, M.; Millan, J. A.; Li, W.; Qi, L.; Xing, G.; Collins, J. E.; Kagan, C. R.; Li, J.; Glotzer, S. C.; Murray, C. B. Nat. Chem. 2013, 5 (6), 466–473.

(19) Dong, A.; Ye, X.; Chen, J.; Murray, C. B. Nano Lett. 2011, 11 (4), 1804–1809.

(20) Henzie, J.; Grünwald, M.; Widmer-Cooper, A.; Geissler, P. L.; Yang, P. Nat. Mater. 2011, 11 (2), 131–137.

(21) Ceseracciu, L.; Miszta, K.; De Angelis, F.; Marras, S.; Prato, M.; Brescia, R.; Scarpellini, A.; Manna, L. Nanoscale 2013, 5 (2), 681–686.

Figure11.42 MB
Figure2.png283.42 KB
Figure3.png1.27 MB
Figure4.png568.17 KB
Figure6.png539.85 KB
Figure7.png1.27 MB
Figure8.png834.31 KB
Figure5jpg.jpg3.84 MB


Teng zhang's picture

Hi Wendy,

Thanks a lot for your very nice review. I am particularly interested in the discussion of "Strong and flexible membranes". This is a new filed to me, and I have a few questions

1. For a membrane formed by nanoparticles with polymer or DNA ligands, do you observe asymmetric elatic modulus during tension and compression?

2. What is the typical fracture toughness of such membranes? And is the fracture in membranes brittle or ductile? 

3. For nanoscale thin film structures, topological defects can cause substantial out-of deformation. Examples include nanoscale Cu film [1] and graphene [2-3]. Do you see similar effects in these nanocrystal monolayers/membranes?

4. Working with Prof. Huajian Gao and Prof. Xiaoyan Li, we showed that topological defects can be utilized to design 3D morphlogy of graphene and furture tailor the mechaical properties of graphene [4-5]. I was wondering whether defects can also induce 3D shapes of these membranes?


1. Zhang, Xiaopu, Jian Han, John J. Plombon, Adrian P. Sutton, David J. Srolovitz, and John J. Boland. "Nanocrystalline copper films are never flat." Science 357, no. 6349 (2017): 397-400.

2. Lehtinen, O., S. Kurasch, A. V. Krasheninnikov, and U. Kaiser. "Atomic scale study of the life cycle of a dislocation in graphene from birth to annihilation." Nature communications 4 (2013): 2098.

3. Warner, Jamie H., Ye Fan, Alex W. Robertson, Kuang He, Euijoon Yoon, and Gun Do Lee. "Rippling graphene at the nanoscale through dislocation addition." Nano letters 13, no. 10 (2013): 4937-4944.

4. Zhang, Teng, Xiaoyan Li, and Huajian Gao. "Defects controlled wrinkling and topological design in graphene." Journal of the Mechanics and Physics of Solids 67 (2014): 2-13.

5. Zhang, Teng, Xiaoyan Li, and Huajian Gao. "Designing graphene structures with controlled distributions of topological defects: A case study of toughness enhancement in graphene ruga." Extreme Mechanics Letters 1 (2014): 3-8.

xwgu's picture

Hi Teng,

Thanks for the questions!

1) No one has looked at a nanoparticle array with the same composition and geometry in both tension and compression. It's tricky to try to learn about tension/compression asymmetry by comparing pillar compression and nanoindentation into thick films, to membrane deflection (tensile deformation) on monolayer or few-layered membranes because of the differences in pre-strain, and even ligand conformation in the compression and tension samples because the different processing steps to make the samples.

2) Fracture be brittle or ductile depending on the ligands used. The short chain polystyrene ligands that I used in my study are brittle at room temperature so fracture of the nanoparticle thin films is brittle as well. Dodecanethiol capped Au nanoparticle membranes are also brittle (Wang et al., Faraday Discussions (2015)) because the ligands are too short for entanglement. Fracture strengths for this system ranged from 11-15 MPa, and depended on membrane thickness and size of the inorganic nanoparticle (fracture toughness has not been measured). Superlattices formed with longer, less dense polystyrene ligands (polymer brushes) can be brittle or ductile, with fracture toughness that is 0.2-0.9 of the toughness of the bulk polymer (Choi et al., Soft Matter (2012)). 

3) Defects of individual nanoparticles (like vacancies and interstitials) may cause dips or bumps at the surface of film, but I did not see any evidence of this using AFM. The polymer ligands near the defect may stretch or compress to minimize the defect. Ripples are not seen in freestanding nanoparticle membranes because the membranes are under tension. I imagine that ripples could form if the membrane could be made very big, or tension free. 

4. The defects in nanoparticle superlattices definitely affect their mechanical properties. I'm not sure if defects could control topological features in a freestanding structure. One difference between the nanoparticle membranes and a 2D material like graphene is that the "bonding" between nanoparticles is not directional.

Teng zhang's picture

Hi Wendy,

Thanks for your detailed explanation.



Xiaoyan Li's picture

Hi Wendy,

Thank you very much for your sharing this comprehensive and inspiring review! It is amazing for colloidal nanoparticles to spontaneously assemble into various architectured nanostructures and nanomaterials with unique mechanical properties!

I am very interested in your impressive work about polymer-grafted nanoparticle supper-lattices. I have two questions: (1) It was observed that some defects (such as vacancies, interstitials) formed during self-assembly of nanoparticles. What are the main factors to control and determine the formation and distribution of these defects? (2) During self-assembly of nanoparticles, can these nanoparticles form the polycrystalline configurations with grain boundaries? Thank you very much in advance!

xwgu's picture

Hi Xiaoyan!

Thanks for the encouraging comments and interesting questions!

1) The main factor for controlling defect formation and distribution was the rate of self-assembly of the nanoparticles on the fluid interface. During the self-assembly process, nanoparticles are deposited in their native solvent (hexane or toluene) onto a droplet of an immiscible fluid (ethylene glycol). As the solvent dries, the nanoparticles are confined to a smaller and smaller space above the immiscible fluid and nanoparticles try to find an energetically favorable position relative to other nanoparticles. If drying occurs slowly, the nanoparticles are likely to form into a close-packed conformation. If drying occurs quickly, the nanoparticles are frozen into energetically unfavorable positions which results in defects. These two limits are analogous to slow cooling to form an atomic crystal and rapid quenching to form a glass. 


2) Yes, these nanoparticles can form polycrystals with grain boundaries. The size of crystalline domains depends on nucleation vs growth much like in atomic crystals. 

Xiaoyan Li's picture

 Hi Wendy,

Thank you very much for your clear explanations, which are very helpful for my understanding the self-assembly of nanoparticles.

I have one more questions: as you exemplified in this review, some nanoparticles can assemble into 2D membrane structures, while the other nanoparticles can form 3D super-lattice structures. What is the main difference in self-assembly of nanoparticles during formation between 2D and 3D structures? Thank you very much in advance!

xwgu's picture

Hi Xiaoyan,

Nanoparticle self-assembly follows the same principles as (atomic) crystal growth. Superlattice growth begins when a solid aggregate of nanoparticles nucleate from the liquid phase, and gets bigger when additional nanoparticles are added to the initial solid cluster. A 2D structure will form when growth is confined to the lateral directions. This can occur when the particles are placed on an immiscible fluid interface, or a solid substrate that is well wetted by the nanoparticle solution. 3D nanoparticle crystals by allowing growth in 3-dimensions, such as by gently destabilizing nanoparticles in solution by adding an anti-solvent (i.e. adding acetone to a hexane-nanoparticle solution) to induce the particles to crystallize. 

Thank you for the question!


Xiaoyan Li's picture

Hi Wendy,

Thank you very much for your explanation which is very helpful for my understanding of superlattices self-assembled by nanoparticles.


ahmedettaf's picture

Hi Wendy, 

Thank you for the very interesting and well written article. I have a couple of questions:

 1- These assembled structures show some kind of periodicity. Has there been any studies on their dynamic response and their wave mitigation properties? Periodic structures may exhibit frequency band gaps and may have extreme properties (e.g. auxetic response, negative refraction,..etc). Existence of defects may lead to breakage in Parity time symmetry, unidirectional wave propagation, and localized modes. Are these phenomena of interest to the community of researchers working on that scale? 2-I wonder what the effect of the substrate curvature on the self-assembly process and the mechanical response of the resulting structures is. In particular, what happens if we introduce corrugation in the substrate and assemble the nanoparticles on the resulting topography. There should be some effect on stiffness and toughness of the resulting membrane. It is also possible that there will be some residual stresses or inelastic deformation locked in due to synthesis on curvilinear geometry. I am interested in designing tunable and adaptive materials system and thus I am curious about such effects on the nanoscale. Thanks again for the intriguing discussion and good luck in your future endeavors!

xwgu's picture

Hi Ahmed,

Thank you for the thought provoking questions!

1) I am not aware of any studies on the dynamic response of these structures, but the type of phenomena that you bring up should be of interest.  2) It's possible that the mechanical properties of the self-assembled structures would depend on the substrate. A corrugated surface could affect the mechanical properties of the resulting structures if the curvature of the corrugations is on the order of the nanoparticle spacing (~10 nm). I would imagine that the amount and direction of residual strain in the superlattice is affected if the polymeric ligands conform to a curved surface in some non-equilibrium configuration. This could be interesting! 

mohamedlamine's picture

Dear Xwgu,

Very interesting analysis.

What is the behavior of these new nanomaterials under cyclic loadings. It is shown in figures 1.A and 1.B that the materials may have a somewhat brittleness. Are they very elastic like you have cited the elastic modulus' values or do they have dynamic displacements similarly to sandwitch laminates i.e cyclic and linear stresses ?

xwgu's picture

Hi Mohamedlamine, 

Thanks for the question! 

The ductility/brittleness of the self-assembled materials largely depends on the properties of the organic ligands attached to the particles. In my work, this ligand was polystyrene (brittle at room temperature) so the self-assembled structures were brittle (but with high elastic modulus). Detailed cyclic testing has not been performed (a few cycles of deformation has been carried out in Heinrich Jaeger's work). I would not expect these materials to have the properties of sandwich laminates because of the large differences in their structures. 


Subscribe to Comments for "Journal Club for March 2018: Colloidal Self-Assembly of Architected Nanomaterials"

Recent comments

More comments


Subscribe to Syndicate