How to Dip-Coat and Spin-Coat Nanoporous Double-Gyroid Silica

pubs.acs.org/Langmuir © 2009 American Chemical Society

How to Dip-Coat and Spin-Coat Nanoporous Double-Gyroid Silica Films with EO19-PO43-EO19 Surfactant (Pluronic P84) and Know it Using a Powder X-ray Diffractometer Michael P. Tate, Vikrant N. Urade, Steven J. Gaik, Christopher P. Muzzillo, and Hugh W. Hillhouse* Received September 11, 2009. Revised Manuscript Received November 9, 2009 Previously, the synthesis of highly oriented pure double-gyroid nanoporous silica films has been demonstrated using evaporation-induced self-assembly (EISA) and dip-coating with a specialty triblock surfactant (PEO-PPO-alkyl) as the template. For these films, grazing-incidence small-angle X-ray scattering (GISAXS) was used to determine orientation and structure. However, GISAXS is not widely available, and we have observed significant batch-to-batch variability in the PEO-PPO-alkyl surfactants used. Here, we show for the first time: (1) synthesis of highly oriented pure doublegyroid nanoporous silica films using freely available EO19-PO43-EO19 surfactant (Pluronic-P84) as the nanostructuredirecting agent, (2) the use of spin-coating and dip-coating EISA to fabricate the double-gyroid films, and (3) the use of θ-θ X-ray diffractometers (commonly available and typically used for powder X-ray diffraction, PXRD) to identify the double-gyroid phase. Processing diagrams for P84 using dip-coating and spin-coating are shown in order to map the dependency of the nanostructure on solution composition, relative humidity, and solution aging time. In addition, an effect of the rate of evaporation during EISA is observed via dependence on the angular velocity in spin-coating. Also, through quantitative comparison of the GISAXS patterns with corresponding PXRD patterns, previously unexplained diffraction peaks in the PXRD patterns are shown to result from diffraction from crystallographic planes that are not parallel to the substrate (typically not observed in PXRD) due to the small angles involved and the nonzero acceptance angle of the PXRD Soller slits. These peaks provide a means to distinctly identify the double-gyroid phase using PXRD. The same trends relating aging-time-before-coating to the phase that forms via EISA are observed with EO19-PO43-EO19 as was the case in previous studies using EO17-PO14-C12. This shows the generality of use of aging time to synthesize nanoporous silica films with nonionic surfactants. Finally, a list of “tips and tricks” is provided to facilitate easy reproducible synthesis of double-gyroid nanoporous silica thin films in other laboratories.

Introduction The discovery1,2 of highly ordered periodic nanoporous silica powders formed by surfactant self-assembly generated significant interest and excitement due to the near-endless possibilities to make new nanostructured materials for gas adsorption, catalysis, and membrane separations, and for new nanostructured optical, electronic, or magnetic devices. Very soon after this discovery, researchers developed methods to form continuous films of similar materials with spin-coating3,4 and dip-coating techniques.5 This synthesis method has been commonly known as evaporation-induced self-assembly (EISA),6 since the nanostructure forms as solvent evaporates from a thin liquid coating, concentrating the nonvolatile components on the substrate. Since these first reports of EISA, the understanding of the roles of the coating solution composition (ratio of nonvolatiles, such as the *To whom correspondence should be addressed. E-mail: hugh@ purdue.edu. (1) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834–10843. (2) Schuth, F. In Mesoporous Crystals and Related Nano-Structured Materials; Terasaki, O., Ed.; Elsevier: Amsterdam, 2004; Vol. 148, pp 1-13. (3) Ogawa, M. J. Am. Chem. Soc. 1994, 116, 7941–7942. (4) Ogawa, M. Chem. Commun. 1996, 1, 1149–1150. (5) Lu, Y. F.; Ganguli, R.; Drewien, C. A.; Anderson, M. T.; Brinker, C. J.; Gong, W. L.; Guo, Y. X.; Soyez, H.; Dunn, B.; Huang, M. H.; Zink, J. I. Nature 1997, 389, 364–368. (6) Brinker, C. J.; Lu, Y. F.; Sellinger, A.; Fan, H. Y. Adv. Mater. 1999, 11, 579– 585. (7) Alberius, P. C. A.; Frindell, K. L.; Hayward, R. C.; Kramer, E. J.; Stucky, G. D.; Chmelka, B. F. Chem. Mater. 2002, 14, 3284–3294. (8) Cagnol, F.; Grosso, D.; Soler-Illia, G.; Crepaldi, E. L.; Babonneau, F.; Amenitsch, H.; Sanchez, C. J. Mater. Chem. 2003, 13, 61–66.

Langmuir 2010, 26(6), 4357–4367

inorganic to surfactant ratio)7 and the humidity during and after coating8 have evolved. By varying these two parameters, a processing diagram can be created showing regions where a specific phase can be formed. For most systems, many different phases can be synthesized from a single surfactant by changing only these two parameters. However, as was recently demonstrated in a series of two papers,9,10 there is a third dimension to this phase space that is the aging time of the coating solution prior to dip- or spin-coating. Processing diagrams can then be created that reveal nanostructured film phases that could not be synthesized without aging. This approach was used to synthesize purephase double-gyroid silica films using a poly(ethylene oxide)-bpoly(propylene oxide)-b-alkyl surfactant with the nominal formula EO17-PO12-C14.9-11 For a recent review of self-assembly and the generation of order in EISA of nanostructured films, see Innocenzi et al.12 Here, we report the synthesis of double-gyroid nanoporous silica films using freely available Pluronic P84 surfactant. We utilize the method of aging the coating solution prior to dipcoating in order to control which nanostructure self-assembles and illustrate that the behavior is the same as that of the PEO-PPO-alkyl/silica system. In addition, we report here the first synthesis of double-gyroid nanoporous silica films using (9) Bollmann, L.; Urade, V. N.; Hillhouse, H. W. Langmuir 2007, 23, 4257–4267. (10) Urade, V. N.; Bollmann, L.; Kowalski, J. D.; Tate, M. P.; Hillhouse, H. W. Langmuir 2007, 23, 4268–4278. (11) Urade, V. N.; Wei, T.-C.; Tate, M. P.; Kowalski, J. D.; Hillhouse, H. W. Chem. Mater. 2007, 19, 768–777. (12) Innocenzi, P.; Malfatti, L.; Kldchob, T.; Falcaro, P. Chem. Mater. 2009, 21, 2555–2564.

Published on Web 12/03/2009

DOI: 10.1021/la903443p

4357

Article

Tate et al.

spin-coating with EISA, which from a material utilization standpoint is advantageous compared with dip-coating. Also, since the films are nearly perfectly oriented with their (211) planes parallel to the substrate, determination of the structure using a θ-θ diffractometer (or a θ-2θ diffractometer, both are typical for “powder” X-ray diffraction instruments) with a 1D detector is difficult since only the (211) and it is higher order reflections (which are not independent) are typically observable. As a result, X-ray characterization has required 2D GISAXS which can be used to easily observe the many Bragg peaks that appear at angles away from the specular plane. However, here we show how powder X-ray diffraction (PXRD) patterns may be used to identify well-ordered double-gyroid films. Through quantitative comparison with grazing-incidence small-angle X-ray scattering (GISAXS), we show how small additional peaks present in the PXRD patterns occur due to off-specular diffraction peaks reaching the 1D PXRD detector via the finite width receiving slits. The ability to observe these additional peaks in the PXRD patterns can allow for identification of the double-gyroid nanostructure without the need for GISAXS if the lattice constants of the film are small enough and the Soller slit spacing is large enough. However, transmission electron microscopy (TEM) should always be used a check to verify the phase topology. In addition, there are many commonly overlooked details and several loose ends that prevent many from effectively reproducing EISA syntheses. Just a few factors include the potential importance of evaporation rate,13 substrate type,14 and thermal treatment temperature.14 Here, we provide an extensive list of “tips and tricks” to help researchers in other laboratories reproducibly synthesize nanoporous silica films in general and the doublegyroid phase in particular.

Experimental Methods and Discussion Calculation of the Coating Solution Composition. In addition to reporting conditions and experimental details necessary to synthesize double-gyroid films from freely available Pluronic P84, this report serves to illustrate the generality of the method of controlling curvature reported by Urade and coworkers9,10 previously for EISA films templated by EO17-PO12C14. Thus, we describe here the rational for how the coating compositions were selected. The coating solutions were prepared from tetraethyl orthosilicate (TEOS), Pluronic P84 surfactant, water, HCl, and ethanol. After evaporation of the solvent, water, and HCl, only the hydrolyzed and partially condensed silica oligomers and surfactant remain on the substrate. The ratio of these two nonvolatile components is a key parameter that controls the phase that self-assembles and was estimated from the binary surfactant/water phase diagram reported by Alexandridis et al.15 For the P84/water system, the double-gyroid phase occurs from 62 to 65 wt % P84, and assuming the excess volume of mixing is zero, the volume fraction corresponding to this region is given by ΦP84

wP84 =FP84 ¼ wP84 =FP84 þ wH2 O =FH2 O

ð1Þ

where FP84 is the density of Pluronic P84 (1.03 g/cm3), wP84 is the mass fraction of P84, and FH2O is the density of water (1.00 g/cm3). (13) Gibaud, A.; Grosso, D.; Smarsly, B.; Baptiste, A.; Bardeau, J. F.; Babonneau, F.; Doshi, D. A.; Chen, Z.; Brinker, C. J.; Sanchez, C. J. Phys. Chem. B 2003, 107, 6114–6118. (14) Klotz, M.; Albouy, P. A.; Ayral, A.; Menager, C.; Grosso, D.; Van der Lee, A.; Cabuil, V.; Babonneau, F.; Guizard, C. Chem. Mater. 2000, 12, 1721–1728. (15) Alexandridis, P.; Olsson, U.; Lindman, B. Langmuir 1998, 14, 2627–2638.

4358 DOI: 10.1021/la903443p

Thus, the double-gyroid region corresponds to the range of volume fraction of 0.61-0.64. The volume fraction of surfactant in the surfactant/water system is then equated to the volume fraction of the surfactant in the surfactant/silica system. ΦP84 ¼

VP84 VP84 þ Vinorg

ð2Þ

The unknown inorganic volume fraction (Vinorg) can then be expressed10 as a function of the extent of condensation of the silica under the assumption that the volume occupied by the silica clusters in the film may be estimated by the sum of the volume of dense silica (with equal number of moles of silicon atoms) and a volume of water (with equal number of moles as the latent water of condensation contained in the partially condensed silica clusters). Vinorg ¼

mTEOS MSiO2 ð4 -xÞ mTEOS MH2 O þ 2 MTEOS FH2 O MTEOS FSiO2

ð3Þ

where mTEOS is the mass of TEOS, MTEOS is the molecular weight of TEOS, FSiO2 is the density of fully condensed silica, SiO2 (assumed to be 2.2 g/cm3), MH2O is the molecular weight of water, FH2O is the density of water (assumed to be 1.0 g/cm3), and x is the average number of silicon-oxygen-silicon bridges per silicon atom condensed at the aging time just prior to coating. Note, that mTEOS/MTEOS is equal to the number of moles of silicon in the system. This calculation assumes what may be called “perfect mixing” between the silica and the ethylene oxide block of the surfactant, which entails that the silica clusters fill all the volume that water occupies in the binary surfactant/water system. However, it is expected that this assumption is not fully met in reality, and that there may be some volume of silica that does not completely mix with the ethylene oxide blocks. There are several factors that may contribute to imperfect mixing. (1) Enthalpic: Due to silica condensation prior to coating, there are increasingly fewer hydroxyl groups per unit volume. This decreases the number of Lewis acid sites (the hydrogens from tSi-OH) that may interact with the Lewis base sites on the surfactant (the oxygens in (-O-CH2-CH2-). This decreases the enthalpic component that drives mixing. (2) Entropic: As the silicic acid monomers, Si(OH)4, condense to Si(O)x(OH)y species, the entropy of mixing between the larger silica oligomer and the ethylene oxide block becomes less favorable when compared to the entropy of mixing that would be obtained by mixing the monomer with the ethylene oxide block (just as the entropy of mixing of polymers is reduced by a factor of 1/N, where N is the number of segments in the polymer). This “less than perfect” mixing could cause nanostructures with lower interfacial curvature to form than that predicted by these equations. Given that caveat, with ΦP84 from eq 1 and Vinorg from eq 3, VP84 in eq 2 may be calculated below as VP84 ¼

ΦP84 Vinorg ð1 -ΦP84 Þ

ð4Þ

Finally, the ratio of the number of silicon atoms in solution to the number of ethylene oxide groups in solution is a useful way to compare between different surfactant/silica systems, as the silica is associated with hydrophilic EO segment of the surfactant. For Pluronic P84 surfactant, this ratio is calculated as follows: Si=EO ¼

mTEOS MP84 yVP84 FP84 MTEOS

ð5Þ

Langmuir 2010, 26(6), 4357–4367

Tate et al.

where MP84 is the molecular weight of P84 (4200 g/mol) and y is the number of EO blocks per P84 molecule (which is equal to 38). Preparation of the Coating Solution. In a typical synthesis, we hold the mass of silica constant in order to maintain a constant hydrolysis ratio and to maintain constant silica condensation kinetics. Thus, we vary the amount of surfactant to change the volume fraction of surfactant in eq 2. To determine the Si/EO ratios used in the coating solutions, the average number of Si-O-Si bridges per Si (i.e., x in eq 3) must be calculated. In previous work under identical conditions (except for the surfactant), we examined the distribution of the condensing silicates species within the coating solution as a function of aging time using 29Si NMR spectroscopy. As an estimate, we use the same distribution of Si-O-Si bonds as a function of time in this work. Thus, for a given mass of TEOS, mTEOS, the volume of the inorganic phase as a function of aging time can be estimated using eq 3, and for constant ΦP84 the Si/EO ratio at which we expect to see the double-gyroid phase is obtained from eq 5. Using this technique, the phase boundaries for the Pluronic P84/water system can be translated into a processing diagram for highly ordered nanostructured films made by EISA as a function of aging time. The reduction in the volume fraction of the inorganic phase with progressive silica condensation results in a decrease in the interfacial curvature with increasing aging of coating solutions. After 3 days of aging, the average number of Si-O-Si bridges in solution changes slowly, and as a result the predicted nanostructure symmetries also change slowly with additional aging time. However, with the previously used EO17-PO12-C14/ silica system, we observed systematically slightly higher curvatures than calculations predicted.10 This is likely due to the presence of water that stays in the film due to hydrogen bonding with EO and the silica oligomers. Thus, lower Si/EO ratios, corresponding to the predicted lamellar phase, were chosen as a starting point to self-assemble double-gyroid phase Pluronic-P84/ silica films. The dip-coating solutions were prepared by mixing a silica precursor solution with a surfactant precursor solution. The surfactant precursor solution was prepared by mixing 28.00 g of ethanol with the calculated quantity of P84. For Si/EO ratios of 1.00, 1.10, 1.20, 1.30, 1.40, and 1.50, 12.92 g, 11.75 g, 10.78 g, 9.95 g, 9.23 g, and 8.62 g of P84 was used, respectively. The precursor solutions were sealed in high density polyethylene (HDPE) bottles and gently shaken at room temperature (21
C) for 24 h. The silica solutions were prepared by adding 25.72 g of ethanol, 12.70 g of a dilute 0.0177 M hydrochloric acid in water solution, and 24.39 g of TEOS (in that order) to HDPE bottles. The 0.0177 M hydrochloric acid solution was prepared by adding reverse osmosis deionized water to ACS reagent grade 37 wt % HCl. The HDPE bottles of the silica precursor solution were sealed and stirred with Teflon stir bars at room temperature for exactly 20 min. The surfactant precursor solutions were then added to the silica solutions to create the coating solutions, which resulted in final molar ratios of 1.00 TEOS/(0.0263, 0.0239, 0.0219, 0.0202, 0.0188, or 0.0175 P84)/0.00192 HC/6.02 H2O/ 9.96 EtOH. The coating solutions were sealed and left to age at room temperature in the dark without agitation. The P84 was received from BASF (for free samples, call 800.669.BASF), the ethanol from EM Sciences, the TEOS from Aldrich, and the HCl from Aldrich. All chemicals were ACS reagent grade and used as received. Spin-coating solutions were prepared similarly, using 5.08, 5.45, and 5.85 g of P84 each in 9 g of ethanol to create Si/EO ratios of 1.28, 1.19, and 1.10, respectively. The final molar ratios were 1 TEOS/(0.021, 0.022, or 0.024 P84)/0.00188 HCl/6.01 H2O/8.11 EtOH. Langmuir 2010, 26(6), 4357–4367

Article

Substrate Cleaning. Rectangular fluorine doped tin oxide substrates (20 mm  50 mm  2.2 mm) with 8 Ω/sq sheet resistance were cleaned by immersion in a boiling 2 wt % Alconox laboratory detergent solution for 5 min and then rinsed with copious quantities of DI water. The substrates were then immersed in a freshly prepared “aqua regia” solution consisting of a 1:3 ratio (by weight) of 70 wt % nitric acid and 37 wt % hydrochloric acid reagents for 20 min at room temperature and then rinsed with copious amounts of DI water. The substrates were then dried using compressed air and used immediately for dip-coating or spin-coating. Cautionary safety note: aqua regia fumes upon mixing, releasing nitrosyl chloride, chlorine, and nitric oxide, and thus, the cleaning procedure should be conducted in a fume hood. In addition, the release of these vapors also causes an explosion hazard if placed in a sealed container (like a waste jar). Finally, mixing aqua regia with organic solvents can be very dangerous due to the strong oxidizing nature of the solution. Aqua regia also reacts violently with metals, so this cleaning procedure should not be used with metal or metal-coated substrates. More sensitivity to substrate cleanliness has been noticed for spin-coating, and less aggressive cleaning procedures may be used for dip-coating. However, when in doubt, this procedure should be used. Dip-Coating and EISA of Nanostructured Silica Thin Films. Dip-coated films were prepared using a KSV Instruments model LM dip-coater mounted on top of a polycarbonate chamber of dimensions 1.80 cm  1.20 cm  1.80 cm. The dipping arm extends down through a hole in the chamber. No seals are used, but the hole is wide enough such that the dipping arm can rise and lower without friction. A Control Company model 35519-050 relative humidity (RH) probe extends down through another hole in the chamber to monitor the RH. The RH probe is regularly calibrated using three saturated salt solutions of LiCl, K2CO3, and KCl yielding standard relative humidity values of 11.3%, 43.2%, and 84.3%, respectively, at 20
C. The RH was controlled in the chamber by flowing a constant RH room temperature air stream into the chamber at approximately 30 SCFH prepared by proportionally mixing a dry air stream and an air stream that has been saturated with water vapor. The substrate is mounted on the end of the dipping arm inside the chamber. A 25 mL glass beaker is filled to within 2 mm of the top rim with coating solution and placed under the substrate. The humidified gas inlet tube is directed away from the beaker to reduce any convective mass transport before or during coating. The chamber is then closed, and the RH is monitored. The flow rates of the wet and dry air streams are adjusted until the desired RH is obtained in the chamber and is stable. Substrates were then immersed and withdrawn from the coating solution both at 1 mm/s. After substrate withdrawal, the substrates were left in the dip-coating chamber for 15 min under controlled RH to ensure that structures had stabilized. Substrates were then transferred quickly to another controlled RH chamber at 40% RH to dry and age overnight. The “as-synthesized” films were analyzed at this stage and then calcined at 400
C in air for 4 h to remove the surfactant template. The ramping rates for heating and cooling during calcination were both 1
C/min, except for the 2D hexagonal film shown in Figure 3, for which the heating and cooling ramps were 0.5
C/min. The coating solutions were reused for all films at the given Si/EO ratio. When not in use, the coating solutions were returned to the HDPE bottles and kept sealed. The 2D hexagonal films shown in Figures 2 and 3 were dip-coated at an aging time of 2 h, using a Si/EO ratio of 1.50. Lamellar films shown in Figures 2 and 3 were dip-coated at an aging time of 3 days, using a Si/EO ratio of 1.10. Double-gyroid films shown in Figure 2 were coated DOI: 10.1021/la903443p

4359

Article

Tate et al.

Figure 1. “Top-view” of the powder X-ray diffractometer geometry employed for the film characterization and in estimating the observable diffraction peaks in the “off-specular” 2θf direction. Note that the center axis of the diffractometer is a vertical line on the page that passes through the center of the sample.

at 5 days aging, using a Si/EO ratio of 1.30, and the double-gyroid films shown in Figure 3 were coated at 3 days aging with a Si/EO ratio of 1.20. Spin-Coating and EISA of Nanostructured Silica Thin Films. Spin-coated films were prepared using a Laurell model WS-400A-6NPP/LITE spin-coater with the top lid closed. The spin-coating chamber is 21.5 cm in diameter and 3.5 cm from substrate to the top lid. RH was controlled during spin-coating by feeding the humidified air stream (prepared as described above) directly into the spin-coating enclosure through a port in the side. Aliquots of coating solution were added to the center of the substrate through a small hole in the top lid of the spin-coater using a pipet while the substrate was rotated at 200 rpm. Immediately after liquid addition, the angular velocity was rapidly increased to a given angular velocity set point. All films were spun for 15 min at controlled RH to ensure solidification of the silica network under uniform conditions. Films were then transferred quickly to an enclosed container maintained at a constant 40% RH overnight. Spin-coated films were calcined as above. Characterization of the Nanostructured Thin Films. GISAXS patterns were collected using a three-pinhole SAXS camera (Molecular Metrology) with a microfocus Cu KR X-ray source (λ = 1.54 A˚), an Osmic MaxFlux confocal X-ray optic, and a gas-filled two-dimensional multiwire detector at a camera length of 1600 mm. The details and geometry are described elsewhere.11,16 Using NANOCELL (a GISAXS pattern fitting code that employs the distorted wave Born approximation),16 the nanostructure orientation, symmetry, and lattice constants were determined. PXRD patterns were recorded on a Scintag X2 instrument with a ribbon shaped beam from a sealed tube Cu anode source and a Cu Kβ and Cu KR2 energy filtered detector with a radius of 250 mm. The source slit height was 0.02 mm, and both the “scattering slit” and “receiving slit” that sit between the sample and detector were set to 0.02 mm. The widths of the scattering and receiving slits (WS and WR) are 12 mm, which does allow some scattering in the direction parallel to the plane of the substrate (i.e., “off-specular” or 2θf direction) to reach the detector. The instrument slit geometry is shown in Figure 1 to illustrate the limitations on the maximum observable off-specular angles due to each slit. (16) Tate, M. P.; Urade, V. N.; Kowalski, J. D.; Wei, T. C.; Hamilton, B. D.; Eggiman, B. W.; Hillhouse, H. W. J. Phys. Chem. B 2006, 110, 9882–9892.

4360 DOI: 10.1021/la903443p

Using simple trigonometry and the conservative assumption that all diffraction from the sample comes from the center point of the sample, the maximum scattering angle in the 2θf direction is given by eqs 6 and 7 for the scatter and receiving slits, respectively. An approximation for the maximum axial divergence due to the Soller slits (the maximum 2θf) is given by eq 8. However, manufacturers typically list the maximum 2θf for their Soller slits explicitly, referring to it as the maximum “acceptance angle.” For our Soller slits, this is 1.1
. The distance to the scatter slit (R) and to the receiving (P) are 200 and 242 mm, respectively. The maximum observable axial divergence of the instrument is then the smallest value obtained from eqs 6-8.   Ws ¼ 1:7
ð6Þ 2θmaxðscatter SlitÞ ¼ tan -1 2R

2θmaxðreceiving slitÞ ¼ tan -1



WR 2P

 ¼ 1:4


ð7Þ

2θmaxðSoller slitsÞ ¼ tan -1 ðWSS =MÞ ¼ 1:1


ð8Þ

Therefore, the limiting component in our PXRD geometry is the Soller slits, which should be true for all XRD diffractometers since the Soller slits function is to decrease the axial divergence, and 2θfmax is 1.1
. Thus, the information in a 2D window of the diffraction pattern can actually be observed in “1D” PXRD. This 2D window is bounded by the minimum 2θ possible (set by the divergence, scattering, and receiving slit heights) and plus/minus 2θfmax (set by the Soller slit spacing). The window that is observable in PXRD is shown graphically on the 2D GISAXS patterns in Figures 2, 3, and 5 by dotted white lines. It should be noted, however, that the diffracted beam intensity that is registered on the PXRD detector drops as 2θf increases and drops to zero at the angle 2θfmax.

Results and Discussion Synthesis and Characterization of Double Gyroid Nanostructures. Thin films dip-coated from a solution with a Si/EO ratio of 1.20 after 72 h of aging produce (211) oriented, highly ordered double-gyroid symmetry nanostructures. Like previous reports, the double-gyroid nanostructures synthesized here contract along the Langmuir 2010, 26(6), 4357–4367

Tate et al.

Article

Figure 2. PXRD and 2D GISAXS patterns (collected at an angle of incidence, Ri, = 0.20
) for as-synthesized dip-coated films of different morphologies. GISAXS images are shown both with the specular beam stop (using a high intensity beam) and without the specular beam stop (using a beam filter) to remove ambiguity of diffraction spots in the specular plane. Overlaid circles and squares on the GISAXS patterns represent intensities due to transmitted and reflected Bragg spots, respectively, as calculated using NANOCELL.16 The dotted white box represents the region observable in the accompanying PXRD pattern.

[211] direction (i.e., normal to the substrate), which results in a distortion of the original cubic Ia3d nanostructure. Here, the lattice constants for the as-synthesized dip-coated double-gyroid films (shown in Figure 2b) are a = 17.8 nm, b = c = 18.2 nm, R = 90.8
, β = γ = 91.6
, which results from a 4% contraction of the cubic structure. After calcination, as shown in Figure 3b, the lattice constants become a = 14.4 nm, b = c = 18.2 nm, R = 97.1
, β = γ = 108.7
, which is the result of a 43% contraction along the [211] direction. A contraction of 40-43% was consistently obtained with all double-gyroid samples synthesized in this manner and calcined at 400
C. This contraction can be reduced by calcining at lower temperatures. The change in nanostructure between as-synthesized and calcined nanostructures results from template removal and additional silica condensation during calcination. Finally, we compared transmission electron microscopy (TEM) micrographs from the [111], [211], and [311] projections (see Figure 4) with simulated electron density projections to confirm the presence of the wellordered double-gyroid nanostructure. The electron density projections are simulated using knowledge of lattice constants obtained from GISAXS, and estimates of silica wall thicknesses. The details of the image simulations can be found elsewhere.11 The good match between the observed and simulated TEM micrographs in Figure 4, in addition to the GISAXS calculations, confirms the formation of the double-gyroid nanostructure. Langmuir 2010, 26(6), 4357–4367

Characterization of Double Gyroid Nanoporous Silica By PXRD. In the PXRD pattern for the calcined double-gyroid nanoporous silica film made by spin-coating (see Figure 5c), four Bragg diffraction peaks are clearly identifiable. Due to the fact that the films are oriented with the (211) parallel to the substrate in the idealized PXRD geometry, only the (211) and (422) peaks should be observed. However, due to the small angles involved and the spacing between Soller slits, the detector may capture diffracted beams not in the specular plane. The peaks observed at angles less than 1.88
2θ in Figure 5c (and the peaks observed at angles less than 1.92
2θ in Figure 3b) are from just this and are identified as originating from the (112) and (2-11) planes. As can be seen in the GISAXS patterns (Figure 5d), these diffraction peaks are not in the specular plane due to the oriented nature of the films, and yet may be observed in the PXRD patterns due to the width of the PXRD detector and the configuration of the slits between the sample and the detector. A box drawn with dotted white lines is shown in the 2D GISAXS patterns (see Figure 5d) and indicates the region of space from where the PXRD will collect some diffracted intensity. Note that this area may be different for different diffractometers, Soller slits, and scattering/ receiving slit widths. To confirm that the extra diffraction peaks are indeed the (112) and (2-11), the peak positions were quantitatively fit using the DOI: 10.1021/la903443p

4361

Article

Tate et al.

Figure 3. PXRD and GISAXS patterns (collected at an angle of incidence, Ri, = 0.20
) for calcined dip-coated films of different morphologies. GISAXS images are shown both with the specular beam stop (using a high intensity beam) and without the specular beam stop (using a beam filter) to remove ambiguity of diffraction spots in the specular plane. Overlaid circles and squares on the GISAXS patterns represent intensities due to transmitted and reflected Bragg spots, respectively, as calculated using NANOCELL.16 The dotted white box represents the region observed in the accompanying PXRD pattern.

same lattice constants already determined for the nanostructure from the GISAXS patterns. However, precise interpretation of these off-specular Bragg peaks PXRD pattern is complicated because of the deviation from Bragg’s law due to refraction. Bragg’s law does not account for refraction at the air/film interface and thus needs to be modified to properly predict the location of diffraction peaks from these films.17 The modified Bragg’s law is shown below. dapp ¼

2sin

cos -1 ðcos

λ
θPXRD =cos Rc Þ

ð9Þ

where dapp is the apparent d-spacing of the film, λ is the wavelength of the X-ray radiation, θPXRD is half the experimental 2θ observed in the PXRD spectrum, and Rc is the critical angle of the film. If the Bragg peaks are in the specular plane, the apparent d-spacing given by eq 9 is equal to the actual d-spacing of the material. For off specular peaks, only the component of the scattering vector normal to the substrate, sz, contributes to the apparent d-spacing observed in the PXRD. As a result, the apparent d-spacing is calculated from (17) Tanaka, S.; Tate, M. P.; Nishiyama, N.; Ueyama, K.; Hillhouse, H. W. Chem. Mater. 2006, 18, 5461–5466.

4362 DOI: 10.1021/la903443p

sz rather than the magnitude of s. dapp ¼

1 sz

ð10Þ

Thus, using a list of d-spacings as determined from fitted GISAXS patterns using NANOCELL, a list of sz values may be determined. In addition, the Rc needed to solve eq 9 was also calculated from the fitted GISAXS pattern. With this information and the given λ, a list of 2θPXRD is calculated (as shown in Table 1). Comparison of the calculated 2θPXRD with the experimental 2θPXRD reveals a small nearly constant offset in all peaks (and in each sample) of 0.10
2θPXRD. This small difference is attributed to error in the sample height, which is more pronounced at small angles.17 The value of the error in the calculated 2θPXRD suggests that the error in the sample height is less than 175 μm. The comparison between GISAXS and PXRD was carried out for as-synthesized 2D hexagonal, double-gyroid, and lamellar samples to create fingerprint identification of the as-synthesized phases (see Figures 2 and 3). The as-synthesized 2D hexagonal phase (Figure 2a) exhibited c2mm plane group symmetry, oriented with the [01] direction perpendicular to the substrate, and lattice parameters of a = 10.2 nm, b = 15.3 nm, R = β = 90
. The as-synthesized lamellar phase (Figure 2c) had lattice parameters Langmuir 2010, 26(6), 4357–4367

Tate et al.

Article

Figure 5. GISAXS patterns for spin-coated as-synthesized (a, b) and calcined (c, d) double-gyroid nanoporous silica films. The circles and squares are positions of transmitted and reflected Bragg spots calculated using NANOCELL16 with a (b) 2% and (d) 40% contraction of a (211) oriented Ia3d symmetry nanostructure with a = 21 nm. The dotted white line represents the limits observed by the PXRD. (c) Corresponding PXRD pattern for the calcined film.

Figure 4. Comparison of observed and simulated TEM Images. (a-c) Unprocessed transmission electron micrographs of calcined double-gyroid nanoporous silica films looking down the (a) [111], (b) [211], and (c) [311]. (d-f) Simulations of the [111], [211], and [311] projections, respectively, using an Ia3d (a = 21 nm) nanostructure contracted 40% along the [211]. The lattice constants of the contracted nanostructure are a = 15.9 nm, b = c = 19.9 nm, R = 96.9
, β = γ = 107.3
.

of a = b = ¥, c = 8.5 nm, R = β = γ = 90
, and collapsed completely upon calcination (Figure 3c), as expected. The calcined 2D hexagonal film had lattice parameters of a = 10.0 nm, b = 9.9 nm, R = β = 90
, which represents a contraction of 46% in the direction perpendicular to the substrate when compared to true 2D hexagonal symmetry. Good agreement was achieved between the calculated 2θPXRD as determined from GISAXS information and the observed 2θPXRD measured experimentally by PXRD, as can be seen in Table 1. The fingerprint patterns (additional small peaks, but also the relative peak width and peak position of the first high intensity PXRD peak) may be used to identify the double-gyroid phase from the hexagonal or lamellar phases, particularly after calcination. Care should be taken in discriminating the (02) peak from a 2D hexagonal phase, the (211) of the double-gyroid phase, and the (001) from the lamellar, and confirmation of the doublegyroid phase should always be done by comparing TEM images to the “railroad track” pattern shown in Figure 4b. However, the characteristic PXRD patterns can be used to map out a processing diagram for any given synthesis without having to conduct TEM for every sample. Thus, this procedure opens the door to the synthesis of the double-gyroid phase to research groups without access to GISAXS. However, caution must be used when attempting to observe peak positions at very small 2θPXRD values (typically less than 1
2θPXRD). From the geometrical calculations for the observable Bragg peaks (as seen by the dotted white box in Figure 2b), the (121) and (21-1), should be seen in the PXRD. The 2θPXRD angles which these planes correspond to are 1.012
and 0.93
. At these angles, the intensity from the specular beam overwhelms the intensity of diffraction peaks, making it difficult to separate the peaks from the noise. Processing Diagrams. Using the six Si/EO ratios, 1.00, 1.10, 1.20, 1.30, 1.40, and 1.50, we developed processing diagrams for Langmuir 2010, 26(6), 4357–4367

films dip-coated at 40% RH as a function of aging time (see Figure 6). Based on the volume fraction arguments presented above in the experimental methods, we should observe the double-gyroid with Si/EO ratios between 1.10 and 1.30 immediately after mixing the surfactant solution with the silica solution. However, the actual phase observed for an aging time less than 2 h for all three Si/EO ratios is 2D hexagonal, not double-gyroid. This observation mirrors the trends we have observed and reported for the EO17-PO12-C14/silica system.11 Upon further aging (120 h), the double-gyroid nanostructure is observed in the predicted region. While the silica cluster size increases as the solution ages, the volume occupied per silicon atom decreases with aging time since water is condensed out of the oligomers as the coating solution ages prior to dip-coating. This translates into a decrease in the volume of the nonvolatile hydrophilic component (and a decrease in the interfacial curvature) with increasing aging and results in an evolution from 2D hexagonal to double-gyroid nanostructures. Spin-coating processing diagrams are shown in Figure 7 as a function of RH and angular velocity. In addition to changing the angular velocity and relative humidity, the solution composition was also varied to examine the coupling of all three variables (see Figure 7). The solution aging time was fixed at a point (12 days) where all phases were observed in the dip-coated films. At a Si/EO ratio of 1.28, the only phase to appear is a high curvature 2D hexagonal (plane group c2mm) independent of the RH or angular velocity. Similarly, a solution with a Si/EO ratio of 1.10 results in only lamellar or mixed phases for all RH at all angular velocities. For a Si/EO ratio of 1.19, the double-gyroid phase appears at high RH and low angular velocity; however, the double-gyroid phase does not persist as the humidity is dropped or the angular velocity is increased. In both cases, the doublegyroid transitions through a mixed phase to a low curvature lamellar phase.

Tips and Tricks The reliable and reproducible synthesis of double-gyroid (DG) or other nanostructured films by EISA requires paying attention to many experimental details. The narrower the window is in phase space, the more precisely the experimental conditions must be controlled. Some details are intuitive, but others are much less so. In order help others fabricate films, DOI: 10.1021/la903443p

4363

Article

Tate et al. Table 1. Quantitative Comparison of GISAXS and PXRD Data

hkl

d-spacing (A˚)a

Rf (
)

2θf (
)

sz (A˚-1)

apparent d-spacing (A˚)b

calculated 2θPXRD (
)c

corrected 2θPXRD (
)d

experimental 2θPXRD (
)

0.46 0.71 1.00 1.30 1.61 1.92 1.92 2.24 3.90

0.96 1.33 1.61 1.92 1.92

0.96 1.80 3.61

1.80

calcined double gyroid 112 211 121 211 121 211 220 321 422e

76.72 71.55 64.39 57.23 50.93 45.38 43.08 34.98 22.69

0.24 0.83 0.84 1.16 1.48 2.12 1.80 2.12 ;

1.11 1.05 0.97 0.84 0.61 0.00 0.65 1.11 ;

0.004 0.040 0.016 0.014 0.018 0.022 0.022 0.025 0.044

278.19 139.10 92.73 69.54 55.63 46.36 46.35 39.71 0.04

0.480 0.731 1.020 1.322 1.631 1.943 1.943 2.257 3.918

3.73

calcined 2D hexagonal 11 02 04e

70.62 49.59 24.80

0.77 1.65 3.43

0.88 0.00 0.00

0.008 0.020 0.040

101.36 50.69 24.80

0.944 1.782 3.59

as-synthesized double gyroid 112 211 121 211 022 121 211 220 321 231 321 400 422e 633e

74.74 74.46 74.00 73.37 63.85 72.58 71.65 62.70 47.74 47.38 46.97 44.50 35.82 23.88

0.18 0.30 0.48 1.01 0.67 1.22 1.42 1.08 1.28 1.83 2.03 1.48 ; ;

1.16 1.11 1.02 0.88 1.12 0.65 0.00 0.68 1.17 0.88 0.34 1.12 ; ;

0.002 0.005 0.007 0.009 0.009 0.011 0.014 0.014 0.016 0.018 0.020 0.018 0.028 0.042

439.2 219.6 146.4 109.8 109.8 87.86 73.24 73.21 62.74 54.89 48.79 54.89 35.82 23.88

0.413 0.540 0.703 0.883 0.882 1.070 1.261 1.261 1.456 1.652 1.849 1.652 2.496 3.723

0.46 0.59 0.75 0.93 0.93 1.12 1.31 1.31 1.51 1.70 1.90 1.70 2.55 3.77

0.94 1.13 1.31 1.31

as-synthesized 2D hexagonal 02 04

76.56 38.28

1.11 2.26

0.00 0.00

0.013 0.026

78.26 39.13

1.187 2.290

1.17 2.27

1.17

as-synthesized lamellar 1 2 3e 4e

85.00 0.97 0.00 0.012 86.9 1.080 1.18 1.17 42.50 2.00 0.00 0.023 43.4 2.069 2.17 2.20 28.33 ; ; 0.035 28.3 3.144 3.24 3.24 21.25 ; ; 0.047 21.25 4.181 4.28 a Determined from NANOCELL using GISAXS patterns. b Apparent d-spacing = 1/sz. c Calculated 2θPXRD which accounts for the effects of refraction in the film using Rc = 0.18
. d Corrects 2θPXRD for sample height misalignment in the PXRD at small angles. e Higher order peaks calculated from first order peaks rather than observed intensity in the GISAXS detector.

Figure 6. Observed nanostructured phases for Pluronic P84/silica system for six different Si/EO ratios as a function of solution aging time dip-coated at 1 mm/s under a 40% RH atmosphere. Letters indicate data points where 2D hexagonal (H), double-gyroid (G), and lamellar (L) phases were observed. 4364 DOI: 10.1021/la903443p

we have listed below some practical considerations, tips, and tricks. 1. Substrate cleaning is very important and can affect the phase that forms. For glass substrates, we clean with fresh aqua regia as noted in the experimental procedure above. Alternatively, one can skip the aqua regia treatment and boil the substrates in Alconox laboratory detergent, rinse very very thoroughly, and dry under flowing nitrogen. This aquaregia-free cleaning method is typically sufficient, but we have noticed more sensitivity to substrate cleanliness with spin-coating. 2. Ethanol purity/dryness is also important. Pure ethanol is hygroscopic, and repeated use of a reagent bottle or exposure to humid conditions will allow the ethanol to take up water. Water in the ethanol can significantly affect the aging time and thus the interfacial curvature. Fresh ACS Reagent grade 200 proof ethanol should be used in the experiments, and care should be taken to keep the ethanol waterfree. Langmuir 2010, 26(6), 4357–4367

Tate et al.

Article

4.

5.

6.

7. Figure 7. Processing diagrams for spin-coated P84/silica nanostructured films as a function of RH and angular velocity from solutions aged 12 days with Si/EO ratios of (a) 1.28, (b) 1.19, and (c) 1.10. Letters indicate data points where 2D hexagonal (H), double-gyroid (G), lamellar (L), and mixed (M) phases were observed.

3.

The quality of the silicon alkoxide (TEOS) is also important and should be fresh. TEOS is hygroscopic and will take up water from the atmosphere and initiate the hydrolysis-condensation process. This will affect the evolution of the silica clusters, and it introduces uncertainty into the aging time. So, fresh, dry TEOS should be used, and again care should be taken to minimize the exposure of TEOS to humid air. We have found that a freshly opened bottle of TEOS in laboratory air will last approximately 1 month before changes are observed in the processing diagram, even if it is closed after each use. Thus, we purchase small bottles and use them quickly after opening. Use of a drybox or air-free techniques could also be

Langmuir 2010, 26(6), 4357–4367

8.

9.

employed, but such steps are unnecessary with fresh reagents. ACS reagent grade ethanol sold by Aldrich is listed as nominally 37 wt % HCl. However, this may vary from 36.5% to 38.0%. We use the nominal value for calculations. Also, care should be taken not to leave the HCl bottle open, as the HCl is quite volatile and slowly escapes to the atmosphere. You can watch the pH of a concentrated HCl solution in an open beaker increase with time. This has two effects: the HCl may be weaker than expected, and more water will unknowingly be added to the coating solution, affecting the condensation and hydrolysis kinetics. One can attempt to measure the pH of the diluted HCl solution, but the error in that measurement can be significant. We have found that simply limiting the exposure of the HCl to the atmosphere and weighing out the ACS reagent to be the most reliable. If any coating solution actually gels during the first month, it is certainly due to contamination or too high of pH (weak acid). Properly made coating solutions typically take months to gel. We have noticed batch-to-batch variations in many polyethylene oxide based surfactants. In some cases, there is a high molecular weight impurity. For P84, this can be identified by dissolving the surfactant in ethanol to 50 wt %. After equilibrating, any undissolved residue is not EO19-PO43-EO19 and can be centrifuged and discarded. The quantity of surfactant remaining in solution should then be recalculated based on the original quantity minus the undissolved dry mass that was discarded. This solution can then be diluted to obtain the surfactant solution that is mixed with the silica solution. It should go without saying that care needs to be taken to ensure that all reagents, solvents, water, and surfactants should be measured carefully and accurately. Many syntheses that have gone awry in our own lab can be traced to such errors. Measuring out precise amounts of gooey or viscous surfactants requires care and patience. The window of aging time that yields the DG phase is sensitive to the hydrolysis ratio (H2O/TEOS), pH, the temperature at which the solution is aged before coating, and exposure to UV light. In our standard synthesis, we typically use a ratio of 6:1 H2O/TEOS, a target calculated pH of 1.76 assuming 100% dissociation of HCl, and age at our laboratory temperature of 21
C in an undisturbed and dark location. Coating solutions can be refrigerated to partially arrest the effects of aging. A useful procedure is to age at room temperature until the phase one desires is formed during EISA and then put the coating solution into a freezer. Aliquots may then be removed to use for coating. The rate of mass transport of water and ethanol away from the drying film is important and can affect the phase that self-assembles. As a result of this effect, in the case of spin-coating, the angular velocity and the substrate size, shape, and thickness can affect the self-assembly since they each affect the gas phase flow pattern and thus affect the rate of evaporation. Note that the substrate thickness only matters when the substrate is not circular. As the aspect ratio of a DOI: 10.1021/la903443p

4365

Article

rectangular substrate increases, it becomes more like a helicopter blade. The turbulence associated with the leading edge can dramatically affect the mass transport away from the film. In the case of dipcoating, stray convective currents or stagnant pockets must be considered. Thus, the size of the dipcoating chamber and how it is vented/humidified can matter. In addition, the coating solution should be filled to the rim of the beaker such that there is not a region of stagnant (humidified and ethanol rich) air just above the solution. Dip-coating from a halffilled beaker or a bottle can yield a different phase (or disordered material) as opposed to the desired phase. 10. If your laboratory relative humidity is close to 40%, as is typical in many buildings, the DG phase should be able to be fabricated by dip-coating without humidity control. 11. If your lab humidity is too dry, too wet, or simply too variable, the relative humidity can be controlled in a small chamber several ways without too much expense. Two of the easiest methods are: (1) FixedPoint Methods Employing Salt Solutions: One can use a closed chamber with a saturated salt solution to control humidity at specific discrete values from 11% with lithium chloride to 84.3% with potassium chloride.18 The salt solution acts by maintain three-phase equilibrium between the undissolved solid salt, the saturated salt solution, and the humidity of the air. Extra moisture is wicked out of the air by more salt dissolving, and moisture is given to the air by more salt precipitating out of solution. Salt creep can be messy, but a setup designed to avoid salt creep has been described elsewhere19 for use in museum display cases. The issue with using this approach for EISA is the long time needed to reach the target relative humidity after closing the chamber and the buildup of ethanol or other solvents in the vapor phase of a closed chamber which can significantly affect the rate of evaporation in the EISA process. The time response can be improved by adding a fan, which should be turned off prior to coating to prevent from affecting the evaporation rate. The effect of solvent buildup can be reduced by limiting the exposure of the coating solution to the vapor inside the chamber and by using activated carbon or other adsorbents to remove the solvent from the vapor. (2) Divided Flow Humidity Generation: One can mix metered or controlled portions of a 0% RH dry air stream with a saturated 100% RH air stream to achieve any desired relative humidity. We typically use the divided flow technique and force controlled humidity air through the dip-coating or spin-coating chamber, as it has the advantage of continuously replacing the vapor phase with a known humidity and simultaneously removing ethanol or other solvents. The response time is fast, and the flow can be turned off to avoid stray convection currents. It can be implemented in a low cost manner with a manual humidity probe in the chamber, a few rotameters with manual needle valves, a couple of pressure regulators (attached to (18) Greenspan, L. J. Res. Natl. Bur. Stand., Sect. A 1977, 81, 89–96. (19) Creahan, J. West. Assoc. Art Conserv. 1991, 13, 17–18.

4366 DOI: 10.1021/la903443p

Tate et al.

12.

13.

14.

15.

a high pressure air line to generate the dry air stream and the feed stream to be humidified), large bottles to hold clean water, glass frits/diffusers to generate small bubbles, and some plastic tubing. All together, this can be set up for less than $200 excluding the pressure regulators. Another option to generate the saturated stream is to set up a countercurrent packed column with water flowing down and air flowing up to saturate the wet gas stream with water. Other methods: Department store humidifiers can be used to manually increase the humidity of a chamber. When used with a manual relative humidity probe, this method can be easy and effective. Temperature can also be used as a means to control humidity. The humidity of any air stream (saturated or just lab air) can be reduced by cooling it to a given temperature, allowing extra moisture to condense out, and then reheating it to room temperature. Relative humidity probes drift and may be degraded due to acidic vapors from the EISA process, and thus, probes should be regularly and carefully calibrated using the vapor that is in equilibrium with a saturated salt solution. Other molecules in the vapor phase (such as ethanol or other solvents) can dramatically affect the interfacial curvature between hydrophobic and hydrophilic domains and hence affect the processing diagram. So, if we are performing coatings in any enclosed space (i.e., a humidity controlled chamber), either we ensure that a high flow rate of an inert sweep gas (humidified clean air) is active or we monitor the volatile species concentration with a “volatile organic carbon” (VOC) meter and ensure that it is less than 10 ppm. Note that such “cosurfactants” in the vapor phase could be used to purposely control interfacial curvature if the concentration of the species in the vapor is controlled. Regarding PXRD characterization, to observe the low angle peaks (at lower angles than the (211)) without interference from the incident beam, one must make sure the divergence, scattering, and receiving slit heights are set to very small values. We typically use values of 0.02 mm for each and just increase the data collection time to increase the signal-to-noise ratio to resolve small peaks. Even though one can use these PXRD fingerprints to identify which films are DG and which are 2D hexagonal or lamellar in a large data set, one should always use TEM as a means of verification on representative samples. Films can be scraped from any substrate onto a TEM grid and imaged even with a modestly capable transmission electron microscopes. A magnification of only about 40 000 is needed. We typically use calcination at 400
C in air to remove the P84 surfactant template. However, calcination at temperatures as low as 200
C are possible, although there may be some residual carbon left behind.20 The higher the calcination temperature, the larger the contraction perpendicular to the substrate, and

(20) Kleitz, F.; Schmidt, W.; Schuth, F. Microporous Mesoporous Mater. 2003, 65, 1–29.

Langmuir 2010, 26(6), 4357–4367

Tate et al.

Article

the easier it is to identify phases with PXRD. Ozone treatment (particularly with an ozone generator, not using UV) is also effective. Also, solvent extraction of the template is possible with acetone, toluene, or other solvents after a mild heat treatment at about 110
C to enhance the extent of condensation of the silica. 16. Finally, despite the fact that EISA of nanostructured silica films is sensitive to reagent purity/freshness, batch-to-batch variation in surfactant template, and mass transport, it must be stressed that one can usually turn the three “curvature control knobs” of inorganic/surfactant ratio, relative humidity, and aging time to tune the synthesis to yield the desired phase. By generating a processing diagram, we have always been able to find the region of parameters that yield double-gyroid films.

Conclusions In conclusion, this report describes the formation of doublegyroid nanoporous films by dip-coating and spin-coating. The room temperature synthesis process is simple, is highly reproducible, and uses chemicals that are easily available to the nanoporous materials research community (in the case of P84, it is free). In addition, after characterization of P84 templated films by GISAXS and TEM, we identified a method to characterize the films using only PXRD by quantitative comparison of GISAXS and PXRD patterns. The quantitative comparison revealed that the Bragg diffraction peak locations in the PXRD may be shifted to account for refraction at the air/film interface and the fact that some of the diffraction peaks included in the PXRD patterns are not in the specular plane, which then allows for accurate deter-

Langmuir 2010, 26(6), 4357–4367

mination of the lattice constants. Moreover, we show that the inclusion of off-specular diffraction peaks in the PXRD patterns produce fingerprint-like patterns, which may be used to identify quickly and easily the double gyroid from other P84/silica phases. Finally, we showed that the angular velocity during spin-coating affects the formation of ordered nanostructured phases for solutions with a Si/EO ratio of 1.19 and aged 12 days. At low angular velocity (1500 rpm), we synthesized a double-gyroid phase, while at higher angular velocities (6100 rpm) only lamellar phases were formed. Identification of this additional control parameter is crucial to the successful synthesis of the doublegyroid nanoporous silica phase by spin-coating. These advancements in the synthesis and characterization of the double-gyroid nanoporous silica phase will broaden the availability of this material and allow researchers to develop films for new applications. Further, this report serves as an example of the generality of our approach to controlling the phase that selfassembles by combining the three curvature control knobs of composition, humidity, and aging time. It is likely that this method can be used for any ethylene oxide (or similar) headgroup surfactant. It should also apply to precursors (other than TEOS) that can be made to slowly condense. And finally, it may be a route to controlling curvature in other metal oxide sol-gel systems. Acknowledgment. We acknowledge partial funding from the NINE program at Sandia National Laboratory (NNEDC Award Number 689415) and partial funding from the National Science Foundation under the CAREER award (0134255-CTS). Also, NSF-funded facility for In-situ X-ray Scattering from Nanomaterials and Catalysts (MRI program award 0321118-CTS) was used to collect GISAXS data. We also thank BASF for providing Pluronic P84.

DOI: 10.1021/la903443p

4367