Atomic force microscopy (AFM) of microfibres fabricated from the self-assembly and fracturing of silica nanoparticles reveals mesoporous structure with hcp packing. Pore size distribution for (20 – 30) nm sized particles are calculated to lie within rtet ~(2.2 – 3.3) nm and roct ~(4.2 – 6.2) nm for the octahedral and tetrahedral sites. The experimentally measured distribution, using N2 adsorption, is r ~(2 - 6) nm, in excellent agreement suggesting a highly controllable and periodic porosity using these structures. The potential for a number of material and device applications is discussed.
© 2013 Optical Society of America
The recent report on self-assembled optical microwire waveguides from silica nanoparticles using evaporative self-assembly and fracturing [1,2] introduces a new approach to fabricating specialty microfibres and sensor filaments with applications spanning photonic waveguides and devices to chemical detection and filtering. The unique combination of convective, microfluidic flow during evaporation of a pinned drop and attractive intermolecular forces leads to highly ordered “spontaneous” packing - this self-organisation occurs even on a relatively rough, random surface of an amorphous material such as glass, negating the need for templating or the use of ordered, crystalline substrates. But with the attractive forces consolidated, the alignment is not precisely in the direction of the receding drop nor can the packing diminish in intensity with relative drop size reductions leading to the accumulation of very high stresses with evaporation. Eventually, these stresses lead to fracturing along this receding direction, which is opposite to the convective flow depositing the particles at the pinned drop edges. It is the control of these fractures that give rise to the slab microfibres, avoiding previous reliance on surfactant based chemical self-assembly such as hydrogen bond-assisted self-assembly of mesoporous silica fibres precipitated and polymerised on the water side , and even spinning of high viscosity sol gels to produce solid fibres . By replacing the dependence of solvent and related gel processes, and indeed removing solvent-solvent interfaces, the opportunity to integrate complex mixtures and demonstrate mixed nanoparticle self-assembly is greatly simplified and made possible. In this way, for example, the first integration of single photon sources were demonstrated using the nitrogen vacancy containing diamond nanoparticles in silica glass , a step forward in integrating quantum communications into existing material platforms. In the spherical caplet limit, the fractures lead to tapered slithers of self-assembled glass. With some asymmetry introduced in the drop, convective flow can be tuned and this fracturing can likewise be tailored to generate very uniform slab waveguides instead; laser assisted directional evaporative self-assembly (“LADESA”), potentially with sub-micron precision, has so far offered an unprecedented level of finesse in control .
One of the attractive assumptions is that the formation of these microfibres implies a high degree of ordering in these structures, essentially a porous crystalline packing of spheres containing periodic arrays of both nanoparticles and interstitial voids. The potential of significant uniform and periodic porosity and volume offers a different approach to those used producing conventional porosity glass [5,6]. This includes a three dimensional alternative to chemically grown mesoporous silica and zeolite with ordered pore structure prepared using various supramolecular templates and interfacial silica-surfactant self-assembly processes [3,7–11] as well as sol-gel dip coating [12,13]. Electrically assisted deposition and self-assembly of surfactant templated mesoporous silica films , and even femtosecond laser generated mesoporous silica [15,16], has also been reported.
Such silica porosity has tremendous value. For example, Vycor glass with 28% porosity introduced during the conventional quenching phase of glass fabrication is used for chromatography . By contrast, bottom up self-assembly is significant for biophotonics applications since the spheres have readily hydrated surfaces that are often hydrophilic offering, for example, potential entrapment of biological and chemical species during the fabrication stage . As mesoporous larger particles, it is increasingly attracting attention for tunable drug delivery [18,19]. Control of the distribution of pores, particularly periodically, is also of interest for potential filtering and molecular sieve applications , heavy metal detection [8,9], gas detection, inline conventional and optical chromatography, ion exchange, catalysis and photovoltaics, biodiagnostics and biostorage, and other applications and novel techniques. Specifically the combination with novel photonics is especially promising: integrating uniformly spaced, mostly transparent, materials on the nanoscale can improve nonlinear systems, produce web-like plasmonic structures and open a novel path to metamaterials by combining metals and glass. Further, a multiple number of identical and periodically distributed nano-volumes can enable nanoreactions, the basis of both biological and chemical diagnostics, to be scaled in number, or even type, without compromise to enhance detection and for other applications. All these are nanoscale equivalents to what has been proposed within structured optical “lab-in-a-fibre” technologies  where the equivalent “micro or nano chambers” were determined by the width of the channels drawn into the fibre in various structured and photonic crystal fibre configurations. The smallest such channels were < 10 nm produced in tapered form  but feasible in optical fibre form with repeated drawing. Considering these are now on the spatial scale of existing nanocavities, any further reductions will be complicated by the rising strain contribution from local tetrahedral ordering . Bottom-up self-assembly offers a solution around this. For our proposed analogous “lab-in-a-microfibre”, a very high level of control over these nano chambers is required.
In this work, the nano-porous structure of these optical microfibres is characterised in two ways. Surface features are profiled with both scanning electron microscopy (SEM) and atomic force microscopy (AFM) – the measured uniformity and height distribution provides some indirect measure of the volume distribution beneath. The surface is shown to be highly ordered and mesoporous, consistent with that expected from the volume lattice formed from densest packing. With gas adsorption, this porosity is shown to be made up of a pore distribution that correlates with simple calculations consistent with densest packing as the overwhelming driver. The combination of results, including relative mechanical robustness over 11 cm, also indicates an overwhelmingly ordered structure.
1. Microfibre fabrication and characterisation
Microfibres were fabricated by a novel gravity assisted directional evaporative self-assembly (“GADESA”). A colloidal suspension of ~40 wt% silica nanoparticles obtained from Sigma Aldrich (particle size ϕ ~20 – 24 nm; trace quantities of NH+ stabilized the suspension) was diluted to ~5 wt% using deionized water (ϕ ~20 to 30 nm measured in previous work  by DLS – Fig. 1(a)). Approximately 5 mL was then placed using a standard burette at one end of a round glass mirror substrate (ϕ = 25 cm) tilted at ~1°. The drop, under the influence of gravity, was slightly elongated towards the lower end of the surface. It was also not spherical and clearly the surface had a discontinuity and flat top under gravity. Under this gravitational influence, preferential direction of convective flow allowed longer wires to form within the flat region. Characteristic of these wires is the presence of bifurcations in the propagating cracks, from the outside in running parallel to the initial fractures, generating their uniform properties. An image of such a bifurcation is shown near the asymmetric “coffee stain effect”  edge in Fig. 1 (b). The wires ran parallel from top to bottom (Fig. 1(c)). The bulk of the wires were measured to be ~11.5 cm with a cross section A ~40 x 10 μm. Under optical microscopy at each end, the wires were measured to have a taper aspect ratio , TAR = 1. In contrast to the high precision control of microfluidic flow using lasers , this approach enables the quick production of larger quantities of longer wires. For microscopic examination purposes the wires were broken up into smaller pieces.
1.1 AFM characterisation
AFM measurements were undertaken directly on the wires. The uniform long wires were broken into small pieces and placed onto a metal puck for atomic force microscopy (AFM) analysis at room temperature (22 °C). The closed loop measurements were taken with 2 nm XY step resolution and a Z sensitivity of less than 50 pm in AC mode using an Asylum Research Cypher AFM and a Tap300Al-G Budget Sensors cantilever.
Figures 2(a)-2(c) shows typical scans. From the profiles, despite some variation in height on the surface (the standard deviation lies within the AFM resolution indicating a very uniform surface profile, noting that the AFM probe is probably unable to resolve sharply the small gaps between two individual particles side by side. The statistical analysis therefore suggests that the surface is made up of a single layer with variations almost certainly arising from particle size variations. It is evident that the bulk surface of the structures have an extremely uniformly distributed hcp-like packing configuration, although in parts there is ABC packing of ccp (fcc) configuration and there is some physical distortion along the planes (Fig. 2(c)). Both hcp and fcc packing are observed for hard spheres [24–28]. Given that both these configurations have similar lowest free energies for packing and are essentially the same except where layers are offset by 60 degrees in packing position on the hexagonal unit of the layer before, it is interesting that the surface data suggests principally hcp over fcc, perhaps reflecting the 2-D planar nature of the microfluidic formation process within the drops as particles flow inside out during evaporation. Either way, the packing volume density will be ~74% in both cases, leaving ~26% free volume within the interstitial regions for an ideal structure . It is worth noting that such a percentage of porous volume from a self-assembled silica microwire is comparable with specially prepared porous glasses such as “thirsty” Vycor glass which has porosity near 28% and an average internal pore diameter between 4 and 7 nm . The pore sizes expected from the unit cell of the ideal lattice can be readily estimated.
1.2 Calculated pore volume and distribution
In a perfect hcp or fcc lattice structure with identical nanoparticles packed densest, there are two sizes of interstitial regions, occupying 26% of the total volume . Of these, there are twice as many tetrahedral interstitial regions compared to octahedral (8 versus 4 within an fcc/hcp unit cell) although the volume is six times smaller. The size of a structure that can be supported by each interstitial region is defined by the radius of a sphere, r, that can fit into the gap over the radius of the nanoparticle, R ( = ϕ/2). Assuming the size (ϕ) of most of our nanoparticles lies between 20 and 30 nm, for an octahedral interstice roct = 0.41R ~(4.1 - 6.2) nm and for a tetrahedral interstice this is rtet = 0.22R ~(2.2 - 3.3) nm – these sizes are significantly smaller than the smallest channels thus far produced using top down drawing processes . Occasionally, there are some wires in which there exists what appears to be a higher free energy packing bcc-like configuration - in this case these values become roct = 0.155R ~(1.6 – 2.6) nm and rtet = 0.291R ~(2.9 – 4.4) nm respectively. Figure 3 shows the trend in individual interstice volume calculated from the unit cell properties of hcp (or fcc). Clearly, there is considerable scope to tune the volume size using different sized nanoparticles or adjusting the crystal topology. The sizes produced lie just inside the International Union of Pure & Applied Chemistry (IUPAC) definition for mesoporous structures, between microporous (< 2 nm) and macroporous (>50 nm) materials [26, 27].
1.2 Gas adsorption
Whilst there are some visible distortions in the structure arising in part from a size distribution of nanoparticles ~(20 - 30) nm and from the AFM resolution (~2nm), there is a very clear crystal packing configuration throughout. Given the visibility of some surface sites where a particle is missing, the extent of this packing and whether there are significant voids (defects) internally needs to be resolved. This can be addressed in part by gas adsorption to directly measure pore sizes and compare with those expected for ideal packing.
N2 adsorption studies can quantify internal surface area, total interstitial volume and pore size and area distributions. Mesoporous silica has been well studied and data obtained correlated with other methods such as H NMR studies [31, 32].
Gas adsorption was undertaken using nitrogen at liquid nitrogen temperatures (77 K) and obtaining the pressure isotherms for adsorption and desorption using Quantachrome Instruments' Autosorb® iQ-C system. The total volume was determined from the amount of N2 adsorbed at a partial pressure P/Po ~0.995, close to unity assuming pores are completely filled with liquid N2 (Vliq) and there is a monolayer present. During desorption with pressure change, hysteresis is observed in the adsorption/desorption isotherms:33] which works well for mesoporous structures. As the pressure is dropped, the calculation of surface area is made using Brunauer, Emmett and Teller (BET)  analysis:
From Fig. 4(a) the total interstitial pore volume is Vi ~1.7 mL/g. Similarly, the total interstitial surface area is SA ~101 m2/g. The distribution in pore radius is measured to be r = (2 - 6) nm, shown in Fig. 4(b). This size distribution is in agreement with the calculated interstices, r ~(2.2 – 6.2) nm, assuming most of our nanoparticles lie within 20 and 30 nm.
3. Results and discussion
The agreement between the calculated data and that obtained through gas adsorption is a strong indication of just how ordered the bulk of the material is, suggesting the surface variations are primarily due to erosion during fabrication rather than any packing limitation. Such a correlation for the volume may also be suggestive of some particle size selectivity during self-assembly where the particles on the measured DLS fringes (outside of 20 - 30 nm) shown in Fig. 1 make little contribution to deviation of pore size. This is in of itself potentially interesting given the relationship between packing lattice and particle size. Although observed to be mostly hcp, the presence of pockets of more fcc local structures is not unexpected. The ease of fabricating such tiny pore sizes, commensurate in scale with existing nanocavities in conventional glass, is a striking difference between top down and bottom-up fabrication processes. Further, it raises fundamentally intriguing questions about the nature of the glass structure within the nanoparticles themselves and whether internal nanoparticle strain contributes to other phenomena, possibly eventually covalent bonding between particles. As these approaches are integrated increasingly into practical systems, research into these differences is predicted to grow both in fundamental and applied importance.
The overall assembled structure may be described as crystalline though it may actually be better represented by a distributed local crystalline order given the uneven presence of fcc and some observable line defects. AFM data suggests extremely uniform surface packing with a preference of hcp over fcc. Unfortunately, the field of view with the AFM (like many diagnostic tools with nanoscale spatial resolution) is clearly small, making it difficult to determine the extent of a distribution of crystalline zones relative to that of a total single crystal structure. The system may therefore be an interesting larger scale analog to the debate of glassy systems generally where x-ray distribution profiles suggest highly ordered cristobalite-like very local structure (< 2 nm) [35,36] made up of tessellated non-additive covalent bonding of tetrahedral units, disappearing over intermediate length scales. On the other hand, there are interesting key differences - in the first instance the nanoparticle can be considered closer to a classical hard sphere atom (still used to describe crystal structures). But more significantly, the dispersion forces holding the nanoparticles together are additive in nature and therefore have a long range impact compared to covalent bonds, one that might lead to longer range order and other interesting size-dependent effects over simple tessellation. It is noted that whilst x-ray diffraction is a useful tool to provide structural information at or near the surface, it too applies best to a largely surface or thin film process and is a timely procedure to assess overall the volume structure.
Gas adsorption has provided confidence, albeit indirectly through deductive reasoning, that the system is, consistent with the remarkable AFM data, overall uniform and the interstitial pores likely to be periodically distributed. Remarkable agreement with an idealised crystalline packing is found. These results overall show that evaporative self-assembly of nanoparticles within a drop on a hydrophilic surface leads to extremely ordered packing through self-organisation driven by microfluidic convective flow and intermolecular attraction. The fracturing process that gives rise to the slab microfibres may be expected to follow the planes of this packing. Bottom-up self-assembly is a powerful tool for controlling and distributing mesoporous structure within materials such as glass, particularly in a periodic fashion that is not available to other fabrication methods. An estimated pore volume of 26% (with some variation from this ideal single sized sphere lattice) can presumably be increased by using larger spheres or mixing spheres to create other crystal interstitial sites. Although studies were undertaken in microfibre form, the technology is applicable in extended versions in two and three dimensions such as self-assembled nanoparticle films on waveguide substrates (or optical fibres), potentially far more amenable to device and sensing applications, particularly photonics. Comparable to VYCOR in total pore volume and other porous glasses, this self-assembled silica offers a novel and unique approach to chromatography and other diffusive systems, such as molecular filters and traps. Good optical transparency in microfibre form might lead to novel optical chromatography. The two different pore sizes potentially allow very selective filtering to be undertaken within the structure (in addition to that at the surface) so that it can be analysed directly, potentially using optical means [1,2] as well as other existing techniques. Given the uniformity of the process, it is possible to envisage a stage where individual nanoreactions or processes can be undertaken in different pores in a controlled manner, potentially feeding into each other – characteristics of future lab-in-a-wire, or lab-in-a-microfibre, technologies. Multiple-site simultaneous repetition of processes that are close to identical offers the unique potential of scaling up nanoreaction products, improving the signal-to-noise of various detection means. This may also reduce significantly both sensor detection and diagnostic tolerances required for processes such as single molecule detection. Finally the nanoscale integration of materials into silica, once considered almost impossible, opens an alternative approach to the fabrication of metamaterials and other novel composites.
The authors acknowledge support from the Australian Research Council (ARC) through grants ARC FT110100116 and FT110100225. M. Ma acknowledges an iPL summer scholarship.
References and links
1. M. Naqshbandi, J. Canning, B. C. Gibson, M. M. Nash, and M. J. Crossley, “Room temperature self-assembly of mixed nanoparticles into photonic structures,” Nat Commun 3, 1188 (2012). [CrossRef] [PubMed]
2. J. Canning, H. Weil, M. Naqshbandi, K. Cook, and M. Lancry, “Laser tailoring surface interactions, contact angles, drop topologies and the self-assembly of optical microwires,” Opt. Mater. Express 3(2), 284–294 (2013). [CrossRef]
3. Q. Huo, D. Zhao, J. Fang, K. Weston, S. K. Buratto, G. D. Stucky, S. Schacht, and F. Schuth, “Room temperature growth of mesoporous silica fibers: A new high-surface-area optical waveguide,” Adv. Mater. 9(121), 974–978 (1997). [CrossRef]
4. K. Matsuzaki, D. Arai, N. Taneda, T. Mukaaiyama, and M. Ikemura, “Continuous silica glass fiber produced by sol-gel process,” J. Non-Cryst. Sol. 112(1–3), 437–441 (1989).
5. W. Haller, “Chromatography on Glass of Controlled Pore Size,” Nature 206(4985), 693–696 (1965). [CrossRef]
6. R. Schnabel and P. Langar, “Controlled-pore glass as a stationary phase in chromatography,” J. Chromatogr. A 544, 137–146 (1991). [CrossRef]
7. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, and J. S. Beck, “Ordered mesoporous molecular sieves synthesized by a liquid-crystal template mechanism,” Nature 359(6397), 710–712 (1992). [CrossRef]
8. K. Moller and T. Bein, “Inclusion chemistry in periodic mesoporous hosts,” Chem. Mater. 10(10), 2950–2963 (1998). [CrossRef]
9. G. E. Fryxell, Y. Lin, S. Fiskum, J. C. Birnbaum, H. Wu, K. Kemner, and S. Kelly, “Actinide Sequestration Using Self-Assembled Monolayers on Mesoporous Supports,” Environ. Sci. Technol. 39(5), 1324–1331 (2005). [CrossRef] [PubMed]
10. Q. Huo, D. I. Margolese, and G. D. Stucky, “Surfactant Control of Phases in the Synthesis of Mesoporous Silica-Based Materials,” Chem. Mater. 8(5), 1147–1160 (1996). [CrossRef]
12. Y. Lu, R. Ganguli, C. A. Drewien, M. T. Anderson, C. J. Brinker, W. Gong, Y. Guo, H. Soyez, B. Dunn, M. H. Huang, and J. I. Zink, “Continuous formation of supported cubic and hexagonal mesoporous films by sol–gel dip-coating,” Nature 389(6649), 364–368 (1997). [CrossRef]
13. D. Zhao, P. Yang, N. Melosh, J. Feng, B. F. Chmelka, and G. D. Stucky, “Continuous mesoporous silica films with highly ordered large pore structure,” Adv. Mater. 10(16), 1380–1385 (1998). [CrossRef]
14. G. Herzog, E. Sibottier, M. Etienne, and A. Walcarius, “Electrochemically assisted self-assembly of ordered and functionalized mesoporous silica films: impact of the electrode geometry and size on film formation and properties,” Faraday Discuss. (2013), doi:. [CrossRef]
15. M. Lancry, B. Poumellec, J. Canning, K. Cook, and F. Brisset, “Ultrafast nanoporous silica formation driven by femtosecond laser irradiation ,” Laser and Photon. Rev, 1-10, (2013) doi . [CrossRef]
16. J. Canning, M. Lancry, K. Cook, and B. Poumellec, “Zeosil formation by femtosecond laser irradiation,” Bragg Gratings, Photosensitivity & Poling in Glass Waveguides (BGPP), OSA’s Advanced Photonics Congress, Colorado United States, (Optical Society of America, June 2012).
17. P. Levitz, G. Ehret, S. K. Sinha, and J. M. Drake, “Porous Vycor glass: the microstructure as probed by electron microscopy, direct energy transfer, small angle scattering, and molecular adsorption,” J. Chem. Phys. 95(8), 6151 (1991). [CrossRef]
18. I. I. Slowing, J. L. Vivero-Escoto, C.-W. Wu, and V. S.-Y. Lin, “Mesoporous silica nanoparticles as controlled release drug delivery and gene transfection carriers,” Adv. Drug Deliv. Rev. 60(11), 1278–1288 (2008). [CrossRef] [PubMed]
19. S. Gai, P. Yang, P. Ma, L. Wang, C. Li, M. Zhang, and L. Jun, “Uniform and size-tunable mesoporous silica with fibrous morphology for drug delivery,” Dalton Trans. 41(15), 4511–4516 (2012). [CrossRef] [PubMed]
20. J. Canning, “New Trends in Structured Optical Fibres for Telecommunications and Sensing,” 5th Int. Conf. on Optical Commun. & Networks and 2nd Int. Symp. on Advances & Trends in Fiber Optics & Applications (ICOCN/ATFO 2006), Chengdu, China, (July 2006).
21. C. M. Rollinson, S. T. Huntington, B. C. Gibson, S. Rubanov, and J. Canning, “Characterization of nanoscale features in tapered fractal and photonic crystal fibers,” Opt. Express 19(3), 1860–1865 (2011). [CrossRef] [PubMed]
22. J. Canning, “Structured Optical Fibres and the Application of their Linear and Non-Linear Properties,” in Selected Topics in Metamaterials and Photonic Crystals Ed. Antonello Andreone, Andrea Cusano, Antonello Cutolo, Vincenzo Galdi, University of Sannio, Italy (World Scientific, 2011) ISBN 978–981–4355–18–6
23. R. D. Deegan, O. Bakajin, T. F. Dupont, G. Huber, S. R. Nagel, and T. A. Witten, “Capillary flow as the cause of ring stains from dried liquid drops,” Nature 389(6653), 827–829 (1997). [CrossRef]
24. L. Cademartiri, K. J. M. Bishop, P. W. Snyder, and G. A. Ozin, “Using shape for self-assembly,” Phil. Tran. R. Soc. A: 28, 370, 2824–2847 (2012). [CrossRef]
25. C. B. Murray, C. R. Kagan, and M. G. Bawendi, “Self-organization of CdSe nanocrystallites into three-dimensional quantum dot superlattices,” Science 270(5240), 1335–1338 (1995). [CrossRef]
27. P. Jiang, J. F. Bertone, K. S. Hwang, and V. L. Colvin, “Single-crystal colloidal multilayers of controlled thickness,” Chem. Mater. 11(8), 2132–2140 (1999). [CrossRef]
28. D. M. Adams, Inorganic Solids, John Wiley & Sons, Great Britain, (1974).
29. L. B. McCusker, F. Liebau, and G. Engelhardt, “Nomenclature of structural and compositional characteristics of ordered microporous and mesoporous materials with inorganic hosts (IUPAC Recommendations 2001,” Pure Appl. Chem. 73(2), 381–394 (2001). [CrossRef]
30. J. Rouquérol, D. Avnir, C. W. Fairbridge, D. H. Everett, J. H. Haynes, N. Pericone, J. D. F. Ramsay, K. S. W. Sing, and K. K. Unger, “Recommendations for the characterization of porous solids,” Pure Appl. Chem. 66(8), 1739–1758 (1994). [CrossRef]
31. P. J. Branton, P. G. Hall, K. S. W. Sing, H. Reichert, F. Schüth, and K. K. Unger, “Physisorption of argon, nitrogen and oxygen by MCM-41, a model mesoporous adsorbent,” Faraday Trans. 90, 2965–2967 (1994). [CrossRef]
32. R. Schmidt, E. W. Hansen, M. Sticker, D. Akporiaye, and O. H. Ellestad, “Pore Size Determination of MCM-41 Mesoporous Materials by means of 'H NMR Spectroscopy, N2 adsorption, and HREM. A Preliminary Study,” J. Am. Chem. Soc. 117(14), 4049–4056 (1995). [CrossRef]
33. E. P. Barrett, L. G. Joyner, and P. P. Halenda, “The determination of pore volume and area distributions in porous substances. I. computations from nitrogen isotherms,” J. Am. Chem. Soc. 73(1), 373–380 (1951). [CrossRef]
34. S. Brunauer, P. H. Emmett, and E. Teller, “Adsorption of gases in multimolecular layers,” J. Am. Chem. Soc. 60(2), 309–319 (1938). [CrossRef]
35. J. T. Randall, H. P. Rooksby, and B. S. Cooper, “Structure of glasses: the evidence of X-ray diffraction,” J. Soc. Glass Technol. 14, 219 (1930).
36. E. Bourova, S. C. Parker, and P. Richet, “Atomistic simulation of cristobalite at high temperature,” Phys. Rev. B 62(18), 12052–12061 (2000). [CrossRef]