Thick mesostructured films via light induced self-assembly

Abstract

As an alternative to solvent evaporation, we show that UV light can serve as a stimulus to promote self-assembly during the formation of surfactant-templated silica films. Despite practical and conceptual advantages, the Light Induced Self-Assembly (LISA) route has been poorly described in comparison with its evaporation induced analogue. In this article, we give an exhaustive overview of the potentiality of this photoinduced process using a range of PEO-b-PPO-b-PEO amphiphilic triblock copolymers, and investigating systematically the structural (²⁹Si MAS NMR, XRD) and textural properties (TEM, N₂ sorption) of the resultant worm-like mesostructured films. Practically, UV irradiation is applied directly to isotropic micrometric films consisting of a nonhydrolyzed alkoxide-amphiphilic copolymer-photoacid generator mixture, obviating the addition of solvent and water. Photoliberation of Brønsted photoacid generates on demand hydrophilic silica species, driving the micellization of the copolymer. The process is single-step and fast, yielding transparent mesostructured films within a few seconds. A target addressed in this feature is to investigate the film structural evolution resulting from the UV irradiation by in situ Fourier transformed infrared analysis (FTIR). Without the interference of solvent, this powerful technique provides a new perspective on the course of the sol–gel process in mesostructured films. For the first time, hydrolysis is assessed kinetically as well as the silica network condensation and the water content of the film.

---

Introduction

Mesoporous inorganic and hybrid films have attracted a growing interest due to their potential in high impact applications such as filtration, catalysis, adsorption and optics.¹ Their properties generally result from the complex interplay between film thickness, porosity, mesopore size and crystalline phase.² By its ability to control simultaneously all these physical and structural characteristics, the Evaporation Induced Self-Assembly^3–5 (EISA) process has rapidly gained in popularity and earned recognition as the reference methodology of preparing mesostructured films. The first step consists of the preparation of a sol solubilizing a surfactant. The latter isotropic solution is then cast and the resultant evaporation of the volatile components causes two major events forming the cornerstones of the EISA approach: the formation of self-assembled surfactant mesophases and the consolidation of the inorganic network through condensation reactions. Nevertheless, the EISA versatility is plagued by the necessity of using high amounts of organic solvent (usually an alcohol), which remains essential for: (1) the sol preparation, (2) the surfactant solubilization; (3) triggering the micellization upon evaporation; and (4) minimizing the siloxane condensation rate to promote the mesophases ordering. Environmental concern but also the importance of an improved control over mesostructuration prompted us to develop an alternative process in which the self-assembly could be controlled by UV light instead, and eliminating the need of solvent.⁶ As far as we are aware, there are only two other examples addressing the use of sol–gel photopolymerization for mesoporous silica film synthesis^7,8 in a protocol substantially different from the one described here. This relative lack of research is somehow surprising, because a Light Induced Self-Assembly (LISA) process is conceptually and practically simpler than the conventional EISA approach.

Our LISA procedure, depicted schematically in Fig. 1, includes non-volatile constituents exclusively and relies on a photoinduced self-assembly catalyzed by photoacid. More specifically, a stable homogeneous solution comprising an alkoxy silica oligomeric precursor, an amphiphilic copolymer and a photoacid generator (PAG) is deposited in the place of a sol, in the form of a film of controlled thickness. Under UV exposition, Brønsted superacids are liberated within the film through PAG photolysis. The low-polymerized hydrophilic silicate species resulting from the photoinduced sol–gel process causes a polarity increase, which eventually drives in situ micellar self-assembly of the copolymer, and ultimately the formation of hybrid mesostructures. In this case, no organic co-solvent was needed, since both the precursor and the amphiphilic copolymer are selected such as to be fully miscible. However, the copolymer is not soluble in the polymerized siloxane network, which is an essential criterion for the surfactant structuration. Water, which is essential for hydrolysis, is not added but provided continuously through atmospheric moisture permeation into the film.⁹ One indirect advantage of using a non diluted medium is the formation of micrometer thick films whereas mesoporous films of only several hundred nanometers are generally reported in the literature.¹⁰ In addition, the catalytic efficiency of the photogenerated acids implies much faster reaction kinetics and well-condensed films with no further need for thermal or chemical post treatment.¹¹

Fig. 1 Schematic pathway for the preparation of mesoporous silica films via a light induced self-assembly process. The starting precursor is polydimethoxysiloxane (PDMOS) a fully methoxy silicate oligomer precursor derived from tetramethoxysilane (TMOS). The UV absorption by the iodonium salt (Φ₂I⁺PF₆⁻) affords a set of homolytic and heterolytic cleavages yielding protonic Brønsted superacids of the structure H⁺PF₆.

Herein, a range of silica-block copolymer mesostructured hybrid films have been synthesized by the LISA-derived method. Such a pathway can be implemented with a wide composition of amphiphilic triblock copolymer templates consisting of a relatively hydrophobic poly(propylene oxide) (PPO) middle block and two hydrophilic poly(ethylene oxide) (PEO) end blocks: PEO_n-b-PPO_m-b-PEO_n. Our approach starts with the extensive structural (XRD, ²⁹Si MAS NMR) and textural (TEM, N₂ sorption) characterization of the photogenerated mesoporous silica films typically exhibiting a regular worm-like structure. As the process takes place in dry conditions, i.e. without the direct addition of water, special attention was given to the role of water to account for the lack of mesoscopic organization. A second target addressed in this feature is to investigate the film structural evolution resulting from the UV irradiation by in situ Fourier transformed infrared analysis (FTIR). Time-resolved FTIR is a valuable technique, rarely reported, to analyze the fast change in chemical composition of the film during the dual micellization and photoinduced sol–gel process.^12–16 Without the interference of solvent prone to saturate the IR signal in EISA, this technique provides a new perspective on the course of the sol–gel process in mesostructured films. For the first time, hydrolysis was assessed kinetically as well as other processes such as the silica network condensation and the water content of the film.

Experimental section

Film preparation

Amphiphilic block copolymers (Pluronic®, HO(CH₂CH₂O)_n(CHCH₃CH₂O)_m(CH₂CH₂O)_nH) were kindly donated by BASF and used as received. In a typical synthesis performed at room temperature, a variable amount of copolymer template, (25 to 75 wt%), was dissolved in polydimethoxysiloxane (PDMOS, ABCR) prior the addition of a PAG (Φ₂I⁺PF₆⁻, 2 wt%, Sigma Aldrich). The homogeneous resultant was found to be stable in the absence of UV light. PDMOS is a non hydrolyzed oligomeric silicate precursor derived from tetramethoxysilane (TMOS); further details regarding its structure have been published elsewhere.⁹ The homogeneous alkoxide-template-PAG solution was deposited onto a glass substrate using an Elcometer 4340 automatic film applicator equipped with a wire wound bar to form a non volatile liquid film with an initial thickness of 10 μm. The UV insulation was carried out at room temperature under a UV conveyor using a microwave powered mercury vapour lamp (H bulb, Fusion). The spectral output of this electrodeless microwave UV lamp is relatively similar to that of a conventional medium-pressure mercury lamp. The belt speed of the conveyor was set at 10 m min⁻¹ and the lamp intensity at 100%. In these conditions, for each pass, the UV exposure time is 0.23 s and the emitted light dose is 1.46 J cm⁻². (UV_A [320–390 nm]: 0.45 J cm⁻², UV_B [280–320 nm]: 0.42 J cm⁻², UV_C [250–260 nm]: 0.09 J cm⁻²and UV_V [395–445 nm]: 0.50 J cm⁻²). The samples were subjected to 10 successive passes under the conveyor to yield transparent hybrid solid films. During UV irradiation, the relative humidity (RH) was carefully monitored to be in the range 27–33%. Film calcination was performed without preliminary hydrothermal treatment at 480 °C for 4 h in air.

For specific time-resolved FTIR experiments, the formulations were dispensed onto a BaF₂ substrate using a bare coater to produce films of similar thickness (10 μm). The in situ IR analysis was performed in transmission configuration (see the characterization section for further details on the spectrophotometric setup), simultaneously with the UV irradiation triggering the polymerization process. In this case, the films were irradiated at a light intensity of 200 mW cm⁻² by the polychromatic light of a mercury-xenon lamp (Hamamatsu, L8251, 200 W) fitted with a 365 nm reflector and coupled with a flexible light-guide. The end of the optical guide was placed at a distance of 3 cm from the film and directed at an incident angle of 90° onto the sample window. The sample was maintained in a horizontal configuration with the same composition and experimental conditions as those photopolymerized under the UV conveyor. During this series of experiments, the RH was carefully maintained between 27 and 33%.

Characterization

XRD patterns of the calcined samples were acquired on a Philips X'pert Pro (PANalytical) diffractometer using Cu-Kα radiation (λ = 0.15418 nm; 0.50 < 2θ < 10°; 0.02°/s). TEM micrographs of the calcined films were taken with a Philips CM200 microscope working at 200 kV. Prior to observation, the powdered film was dispersed into water with ultrasound and a few drops of the suspension were deposited at the surface of a copper observation grid. N₂ adsorption/desorption isotherms were obtained on a Tristar 3000 (Micromeritics). Calcined samples were first degassed at 150 °C for 4 h. Surface areas (S_BET) were determined by the BET method, average pore diameters (D_P) were determined from desorption branch by the BdB method.¹⁷ Pore wall thicknesses (W_thick) were evaluated from the difference between the d-spacing and pore diameter. To obtain quantitatively reliable ²⁹Si solid state NMR spectra, single pulse magic angle spinning (SPE-MAS) experiments were performed on a Bruker Avance II 300 spectrometer operating at B₀ = 7.05 T (Larmor frequency: υ₀(Si) = 59.6 MHz) with a Bruker double channel 7 mm probe. The spectrum was recorded using a pulse angle of π/6, a recycling delay of 80 s respectively, a spinning frequency of 4 kHz and high-power proton decoupling during the acquisition. All the NMR experiments were performed at room temperature and chemical shifts reported thereafter are relative to tetramethylsilane. These recording conditions ensured the quantitative determination of the proportions of the different Qⁿ siloxane species. Deconvolution of the spectrum was performed using Dmfit software.¹⁸

In real-time FTIR experiments, the formulations were simultaneously exposed to UV light and to an IR analytical beam. Such a technique is of interest to assess the sol–gel kinetics throughout the UV irradiation. Infrared spectra obtained in transmission were recorded with a Bruker Vertex 70 spectrophotometer equipped with a liquid-nitrogen-cooled mercury-cadmium telluride (MCT) detector. The spectra were recorded every 0.12 s during the first 18 s then every 5 s from 18 s to 600 s, using a resolution of 4 cm⁻¹. All spectra were baseline corrected prior to integration with the software OPUS 6.5.

Viscosity measurements were performed with a Brookfied DV-II rheometer. Viscosities of pure PDMOS, and hybrid mixture x = 0.25, x = 0.50 and x = 0.75 were respectively of 5, 23, 53 and 96 mPa s.

Results and discussion

I) Photogenerated mesoporous silica film: methodology and characterization

Copolymer solubility in PDMOS. A key element in our procedure is PDMOS,¹⁹ a fully alkoxylated oligomeric precursor derived from tetramethoxysilane (TMOS), serving both as nonvolatile silica source and solvent of the template. In combination with a PAG, the resultant advantage is a film deposition with purely non reactive inorganic species, which contrasts with EISA implying an instable and strongly diluted sol. In this context, it becomes apparent that the deposition step is no longer a crucial step as the surfactant self-assembly is not induced by the evaporation of a volatile phase but through UV irradiation controlling the in situ generation of hydrophilic silicate species. On the other hand, such an approach requires the solubilization of the surfactant in a hydrophobic and weakly polar alkoxy precursor. Taking this fact, it is evident that low molecular weight ionic surfactants are a poor choice and should be replaced by non-ionic amphiphilic copolymers, which feature enhanced solubility in organic solvents.²⁰Table 1 compares systematically the solubility of various symmetric and commercially available triblock PEO_n-b-PPO_m-b-PEO_n copolymers in PDMOS and water, differing from the degree of polymerization of their hydrophilic (n) and hydrophobic (m) blocks. With low Hydrophilic–Lipophilic Balance (HLB) surfactants, PDMOS demonstrates an extended solubilization domain²¹ in comparison with water. As expected, the most hydrophobic surfactants such as L81, L121 and L62 (HLB < 6), featuring short PEO chains and scarcely used in EISA, show an appreciable solubility in PDMOS. Note that these ethylene oxide derivatives do not show an inverse temperature effect on solubility (which is actually the case with water), and may become soluble in PDMOS on moderate heating. However, the most hydrophilic surfactants such as F127 and F108 (HLB ≥ 14) faced solubility problems in PDMOS while their dissolution in water proved to be easy. For the rest of the study, we selected three representative amphiphilic copolymer surfactants: L121, P123 and P105 with a HLB spanning from 2 to 10 and comprising a weight ratio in hydrophobic PPO block of 90, 70 and 50% respectively.

Table 1 Solubility of various amphiphilic triblock copolymers in PDMOS and water

Pluronic	HLB^a	Molecular weight (g mol^−1)	PEOn-b-PPOm-b-PEOn	PPO block (%wt)	Soluble in PDMOS	Soluble in water
a .
L81	2	2750	PEO3PPO39PEO3	90	Yes	No
L121	2	4400	PEO5PPO70PEO5	90	Yes	No
L62	4	2500	PEO6PPO30PEO6	80	Yes	No
L43	6	1850	PEO6PPO21PEO6	70	Yes	Yes
P103	6	4950	PEO17PPO60PEO17	70	Yes	Yes
P123	6	5750	PEO19PPO69PEO19	70	Yes	Yes
P84	8	4200	PEO19PPO43PEO19	60	Yes	Yes
P65	10	3400	PEO20PPO30PEO20	50	Yes	Yes
P85	10	4600	PEO26PPO40PEO26	50	Yes	Yes
P105	10	6500	PEO37PPO56PEO37	50	Yes	Yes
F127	14	12 500	PEO99PPO65PEO99	30	No	Yes
F108	16	14 600	PEO132PPO50PEO132	20	No	Yes

---

Textural and structural characterization of the photogenerated mesoporous films. Independent of the surfactant structure, the present copolymer-PDMOS-PAG films showed a remarkable photolatency over several months. Without volatile components, thick liquid films of 10 μm could be obtained, which gave, after fast UV insulation (2.3 s), solid films of optical quality. The achievement of solid films after photopolymerization appears as a first evidence for an effective condensation of silicate species. In addition, the degree of condensation and the proportion of siloxane substructures (Qⁿ) of the as-synthesized films were also evaluated by ²⁹Si MAS NMR as reported in Table 2. The samples showed high condensation degrees ranging from 76% up to 88%, without further treatment. A greater concentration in surfactant (P123) yielded more condensed structures. Fig. 2 displays the typical diffraction patterns of the calcined samples for surfactant/PDMOS weight ratios (x) ranging from 0.25 to 0.75 at a relative humidity (RH) of 35%, together with the corresponding TEM images. For all three systems (P123, L121 and P105), we find that the copolymer concentration dramatically influences mesostructure development. For a sufficient concentration of template (x ≥ 0.50), every XRD pattern exhibits a single and well-resolved diffraction line at a low 2θ angle, which tends to sharpen and increase in intensity with a higher content of surfactant. These observations are consistent with a disordered or worm-like mesoporous structure^22–24 (similar to that of MSU-X²⁵) arising from randomly arranged elongated micelles. As reported in the literature, the common morphology of PEO-b-PPO-b-PEO micelles in hydrophilic solvents is spherical or cylindrical (rodlike or wormlike).^26,27 The definitive structural assignment was furnished by the TEM pictures confirming for the three selected surfactants the formation of an interconnected (bicontinuous) vermicular structure with relatively well-defined pore size but apparently devoid of long-range order. In this sense, the XRD data cannot actually be indexed to any plane or space group, but rather suggests a repeated pore–pore distance.²⁸ Increasing the surfactant concentration (x = 0.75) has yielded more resolved lines, which can be interpreted as evidence of more uniform mesopores. Remarkably, the initial thickness of the precursor film can be varied in the micrometer range (5–28 μm), without affecting the type of mesostructure formed (see Fig. S1 in ESI†).

Fig. 2 (A) XRD patterns obtained for calcined samples of L121, P123 and P105 at various copolymer/PDMOS wt. ratios: x = 0.25 (dotted line), x = 0.50 (dashed line) and x = 0.75 (plain line). (B) TEM images of the as-calcined films: L121-PDMOS (x = 0.75), P123-PDMOS (x = 0.50), and P105-PDMOS (x = 0.50).

Table 2 Summary of the physicochemical properties of mesoporous silica film prepared by sol–gel photopolymerization with different amphiphilic block copolymer templates (L121, P123 and P105)

	Template/PDMOS (wt%)	Q^2/Q^3/Q^4^a	Condensation degree^b	d spacing (nm)	S BET ^c (m^2 g^−1)	V p (cm^3 g^−1)	D p ^c (nm)	W thick ^c (nm)
a Deconvolution of the spectrum was performed using Dmfit software.^18 b . c Surface areas were determined by the BET method, average pore diameters were calculated by the BdB method from desorption branch.^17 Pore wall thicknesses were evaluated from the difference between d-spacing and pore diameter.
L121	x = 0.25	—	—	—	346	0.21	2.4	—
	x = 0.50	—	—	6.6	500	0.38	3.1	3.5
	x = 0.75	1/46/53	88%	7.9	489	0.46	3.8	4.1
P123	x = 0.25	20/57/23	76%	—	269	0.03	2.2	—
	x = 0.50	5/67/28	81%	8.2	308	0.33	4.2	4.0
	x = 0.75	5/51/44	85%	8.2	374	0.61	6.5	1.7
P105	x = 0.25	—	—	8.4	219	0.14	2.6	5.8
	x = 0.50	3/55/42	85%	8.3	241	0.24	4.0	4.3
	x = 0.75	—	—	8.3	356	0.51	5.7	2.6

---

Nitrogen sorption measurements were also performed, leading to a series of isotherms shown in Fig. 3, obtained for each surfactant at different template/PDMOS ratios. According to IUPAC classification,²⁹ type I isotherm typical of microporous system has been obtained with the P123 surfactant for x = 0.25 ratio whereas others samples exhibited type IV isotherms confirming the mesoporosity. Textural parameters such as the specific surface (S_BET), the pore volume (V_p) and the pore diameter (D_p) were extracted from these isotherms and gathered in Table 2. Whatever the surfactant, an increased surfactant concentration generates a higher number of pores which are also larger in size, resulting in thinner walls. For the nonionic amphiphilic copolymer, pore size is not only controlled by the hydrophobic PPO block length (m) but also by its hydrophilic PEO block (n) as only a portion of the PEO chain interacts with the silica network forming the future walls.^30–33 A very short PEO block (n = 5) of L121 generates the smallest pores (∼2 nm) as the main contribution for the hydrophobic internal zone results only from the PPO block (m = 70) in this case. As expected, larger pores (4–6.5 nm) were obtained with P123 endowed with a comparable PPO block (m = 69) but a longer PEO block (n = 19). Compared with P105 (m = 56, n = 37), the longer PPO block of P123 is likely to reinforce the hydrophobicity of the PEO chains even if these latter are shorter, accounting presumably for the pore diameter expansion with P123.²⁴

Fig. 3 N₂ adsorption/desorption isotherms of calcined films obtained with various surfactants (P123, L121, P105) at three surfactant/PDMOS weight ratios: x = 0.25 (■), x = 0.50 () and x = 0.75 ().

Understanding the worm-like mesostructure. In the light of the previous results, it is surprising that parameters such as the PEO/PPO block ratio and the template/PDMOS ratio, which are known to be determinant in the phase behaviour of poloxamer block copolymers,¹¹ remain ineffective to generate more ordered mesophases. Worm-like structures often result from too highly condensed sols or insufficient surfactant concentration:² some conditions which do not apply to the present system. Rather, we believe that the limited water concentration in the film may be responsible for the systematic formation of disordered mesostructures. Prior to irradiation, the deposited mixture does not include water or a polar solvent. As previously demonstrated,³⁴ the interfacial curvature (micellar interface), which is initially set by the macromolecular architecture, can be further influenced by the ability of the copolymer blocks to swell to different extents with selective solvents. As neat “solvent”, hydrolyzed silica species are probably not sufficiently selective for the PEO block to ensure the formation of ordered mesostructures. Although they presumably occupy the polar domains of the microstructure and promote a micellization, they do not contribute to the swelling of the PEO blocks. In this sense, the resultant (SiO)_4−xSi(OH)_x species should behave similarly to a non aqueous polar solvent in which amphiphilic copolymers are known to form micelles at considerably higher surfactant concentrations than those formed in water.^35–37 However, the addition of a more PEO-resembling solvent, such as water which is a key feature in the access of ordered mesostructured films via EISA^38,39 swells the PEO block and enhances the block segregation, thereby inducing the formation of structures with higher interfacial curvature⁴⁰ (hexagonal, cubic). While the permeation of the atmospheric water into the sample proves to be sufficient to enable the hydrolysis (vide infra), the water uptake is presumably not enough to induce the ordering of the mesophases.

The essence of this reasoning can be captured by the evolution of several textural parameters: first, the fact that the mesostructure only arises at high template concentrations (x ≥ 0.50). Second, the size of the mesopores remains low (<6 nm) compared to other mesoporous systems based on block copolymers⁴¹ (∼7–10 nm with P123), showing that a contracted final mesostructure is preferentially built up. Finally, using the intensity of XRD peaks as a probe for order, we clearly identify that P105, having the longest PEO block, improves substantially the nanosegregation without being able to induce an organization. In addition to its thermodynamical effect, Grosso et al.² suggested an additional role for water: retained in the film, the residual water could confer to the medium fluidity essential to the transition from disordered to ordered micelles. In our water-poor system, the lack of mobility combined with a fast UV-induced gelation is obviously not favourable to the effective conduct of this transition, thus inhibiting the organization of the template in a liquid crystal phase. Consequently, the ability of the block copolymers to exhibit much richer structural polymorphism and to form a great variety of lyotropic phases would require a higher hydration of the silica phase. A high RH is known to promote a higher water concentration inside the film, while a low RH encourages water to evaporate. Further studies using a variety of experimental parameters (RH, light intensity, nature of the surfactant, etc.) are now underway to create the conditions for progressing toward periodically ordered mesostructures.

II) In situ study of sol–gel process in mesostructured hybrid films by real-time FTIR

In situ real-time FTIR experiments were performed to shed light onto the complex and concomitant processes triggered by UV irradiation. There are very few studies reporting the use of this time-resolved technique to study the change in chemical composition of mesostructured silica films.^12–16 Practically, in EISA the acquisition of IR spectra starts just after the casting of a droplet consisting of an alkoxide–alcohol–water–HCl mixture. Such a procedure, inherent in this methodology, hinders the study of the sol–gel process in its entirety: first, the handling of a sol implies a hydrolysis already highly advanced or even complete; second, the strongly diluted medium causes signal saturation during several tens of seconds, before the evaporation of the solvent. Consequently, the dynamic studies obtained by RT-FTIR in the literature have not exploited the excellent resolution time (less than 0.2 s) and have essentially focused on a part of the multiple processes occurring: the solvent evaporation kinetics and the qualitative observation of the silica network consolidation. A notable feature that distinguishes our approach is the solvent-free, non-hydrolyzed and photolatent composition. This permits the UV irradiation to control the sol–gel process and indirectly the mesostructuration to be triggered simultaneously with the analyzing IR beam. The RT-FTIR's interest is therefore substantially enhanced since a complete dynamic survey of the sol–gel polymerization becomes accessible: hydrolysis, condensation and water concentration in the film can be assessed simultaneously throughout the irradiation, so as to build up a reliable pathway. Obviously, in situ FTIR is not really informative about mesostructure and porous texture, but rather translates the temporal change, at microscopic level, in chemical composition of PDMOS.¹ Besides this, we suspect that the surfactant concentration as well as its self-organization ability are two important factors impacting the sol–gel kinetics. Fig. 4 plots a selection of 3-dimensional FTIR absorption spectra acquired in situ during irradiation of a P123-PDMOS-PAG sample, with wavenumber, irradiation time and absorbance as the x, y and z axis, respectively. Our present discussion emphasizes three distinct regions that will be commented on separately.

Fig. 4 In situ time-resolved FTIR spectra on a timescale of 600 s of a PDMOS film containing 50 wt% of P123 under UV irradiation in the range of 2700–3650 cm⁻¹ (A) and 800–1600 cm⁻¹ (B).

2800–3400 cm⁻¹. Prior to irradiation, a broad massif of C–H stretching bands is visible in the region spanning 2800–3000 cm⁻¹, and containing the overlapping symmetric (ν_s) and asymmetric (ν_asym) stretching modes of CH₂ and CH₃ groups. In contrast to methylene moieties which are only part of the copolymer, the methyl can originate both from the PPO block and the methoxy groups of PDMOS. Of high interest is the possibility to distinguish the ν_s(CH₃) and ν_asym(CH₃) from the template and the silicate precursor. In methoxysilyl, the proximity of the oxygen has been shown to cause the red shift of the ν_asym(CH₃) (from 2970 to 2950 cm⁻¹) and the ν_s(CH₃) bands (from 2900 to 2848 cm⁻¹). We note that this latter spectral feature is sufficiently resolved and separated from the other vibration modes to be taken as a marker of the methoxysilyl function hydrolysis. As expected, the UV exposure causes a progressive decrease of the methoxy stretching bands at 2848 cm⁻¹ supporting our attribution and the occurrence of an effective photoacid-catalyzed hydrolysis. Concomitantly, a broad envelope centred at ∼3400 cm⁻¹ is expanding. This band is straightforwardly assigned to the OH–O bonds, reflecting the gradual conversion of the methoxy moieties into silanol groups and the possible adsorption of sol–gel by products (H₂O, methanol). After 600 s of UV irradiation, the sharp peak at 2848 cm⁻¹ has completely vanished, suggesting a complete hydrolysis. Systematic deconvolution and integration of the ν_s(O–CH₃) band gives a unique insight into the hydrolysis kinetics and the influence of the template concentration for P123 and P105-based films, as displayed in Fig. 5. We observe that the presence or absence of surfactant dramatically affects the hydrolysis rate. The addition of 25 wt% template in the formulation initially devoid of P123 implied a brutal slow-down in the hydrolysis kinetics. A straightforward explanation rests on the significant viscosity rise caused by the copolymer dissolution into PDMOS. For the P123 system, we note a 5, 11 and 19-fold increase in viscosity with x = 0.25, 0.50 and 0.75 respectively, compared with a pure alkoxide-based film (x = 0). Many experimental results^9,42 proved that a higher viscosity can affect substantially the water vapour permeation, which is the main mechanism responsible for hydrolysis. Additionally, the differing kinetic profiles between the amorphous sample (25 wt%) and the two mesostructured films (≥50 wt%), strengthens the argument that the in situ micellization may also impact the hydrolysis progress. Both hydrolysis and surfactant self-assembly need water to be effective; therefore a competition between these two processes is likely to take place. In comparison to an amorphous random system, the structuration means a swelling of the PEO block and presumably a nanosegregation of water in the vicinity of the hydrophilic blocks. This confined geometry may impose water limitation for apolar methoxysilyl functions, causing a slow-down of the hydrolysis kinetics.

Fig. 5 Methoxy hydrolysis degree during the sol–gel photopolymerization of films containing various ratios (x) in copolymer template: P123 (A) and P105 (B). x = 0 (, pure PDMOS), x = 0.25 (■), x = 0.50 () and x = 0.75 ().

1000–1250 cm⁻¹. Before irradiation, this region represents the contribution of different vibration bands: the Si–O–Si stretching modes from the oligomeric precursor as well as the C–O stretching associated with the methoxysilyl functions of PDMOS and the PEO block of the template. The high overlapping of these vibrational modes coupled with the consumption of the methoxysilyl functions during the hydrolysis stage (<300 s) makes the observation of the siloxane condensation particularly tricky. However, after the completion of hydrolysis, a qualitative evaluation of the siloxane condensation becomes feasible (Fig. 6). Evidence of progressive polycondensation reactions is reflected by the increase in absorbance of the envelopes of absorption bands assigned to longitudinal optical (LO) and transverse optical (TO) asymmetric stretching modes of siloxane bonds at 1163 cm⁻¹ and 1078 cm⁻¹ respectively.

Fig. 6 Infrared absorption spectra of the P123-PDMOS film (x = 0.50) in the range 800–1500 cm⁻¹, recorded at different irradiation times after the hydrolysis completion.

1640 cm⁻¹ (H₂O bending mode). The presence of molecularly adsorbed water can be evidenced by a distinctive band at 1640 cm⁻¹ attributed to a scissor bending vibration.¹³ Interestingly, this water band has been shown to be a good discriminatory, non-destructive, indicator of water content within the film. Its interest lies in its higher selectivity compared to the molecular stretch vibrations (with a maximum at 3400 cm⁻¹) assigned to OH stretching, and including therefore the contribution of all the hydroxyl-containing compounds such as silanol and methanol groups. Fig. 7 plots the temporal evolution of the integrated absorbance of the 1640 cm⁻¹ band throughout the irradiation for different contents in amphiphilic copolymers (P123 and P105). Whether a surfactant is present or absent, an equivalent concentration in water is dissolved within the film prior to irradiation (although no water was added to the initial formulation). Apparently, the amphiphilic copolymer with hydrophilic PEO chains is unable to confer additional hydrophilicity to the initial film. In contrast to water stretching modes whose position is very sensitive to increased strength of hydrogen bonding, there is no clear shift of the bending mode following the change from an apolar PDMOS environment to a hydration sphere within the siloxane network. However, we note an increased intensity of this band during the irradiation process, quantified in Fig. 7. In all concentrations in P123 or P105, we observe a gradual enrichment in water. Such a trend is consistent with the in situ generation of hydrophilic silica species resulting from hydrolysis, presumably favouring the absorption of atmospheric water. Additionally, water can be released within the film as a byproduct of the condensation reactions. However, we note that this increase is much more moderate in samples prepared with surfactant. Such a low concentration in water observed in copolymer-based films can be correlated with the slow-down of the hydrolysis kinetics (vide supra) compared to pure PDMOS samples. Water migration in polymer membranes is governed by the interplay of two physical processes: water solubility and diffusion. In P123, the presence of a dominant hydrophobic part (70 wt%) in the copolymer presumably affects the solubility coefficient in comparison with a pure hydrophilic silica network. In addition, a macromolecular surfactant may be seen as an obstacle to water mobility, altering this time the diffusion process. P105 shows a similar tendency, except for high concentrations (75 wt%) in which the water uptake is significantly increased, due to the greater hydrophilicity of this surfactant.

Fig. 7 Water content evolution during the sol–gel photo-polymerization of P123-PDMOS (A) and P105-PDMOS (B) films containing various ratios in copolymer: x = 0 (), x = 0.25 (■), x = 0.50 () and x = 0.75 ().

Conclusion

We described an alternative and simple approach to silica/surfactant mesostructured films based on a self-assembly process induced by UV light and no longer by solvent evaporation. The light controllable self-assembly relies on the in situ photogeneration of hydrophilic silica species by the release of photoacids, thus driving the partial solubility and micellization of PEO_n-b-PPO_m-b-PEO_n amphiphilic copolymers. The notable features that characterize our photoinduced approach are the speed process, the absence of volatile components (water, organic solvent) and the deposition of a micrometer film consisting of a photolatent oligomeric alkoxide–template–PAG mixture. Our observation emphasized that the absence of water might alter the self-assembly pathway, with a silicate/surfactant mesophase deviating substantially from the theoretical predictions. As recognized for the EISA process, the achievement of organized mesostructures is expected to depend on a number of experimental parameters that will be thoroughly discussed in a subsequent article.

Our procedure devoid of solvent, and in which the sol–gel process is fully implemented after deposition, has significantly enhanced the interest of in situ FTIR spectroscopy. RT-FTIR permitted a comprehensive study of the sol–gel process taking place, in particular the hydrolysis kinetics which are normally impeded in a conventional EISA process. We contemplate that in the future this technique should offer interesting opportunities to investigate not only the change in chemical composition but also the general topic of self-assembly of surfactant/silica mesophases.

References

1. C. Sanchez, C. Boissière, D. Grosso, C. Laberty and L. Nicole, Chem. Mater., 2008, 20, 682–737.
2. D. Grosso, F. Cagnol, G. J. D. A. A. Soler-Illia, E. L. Crepaldi, H. Amenitsch, A. Brunet-Bruneau, A. Bourgeois and C. Sanchez, Adv. Funct. Mater., 2004, 14, 309–322.
3. C. J. Brinker, Y. Lu, A. Sellinger and H. Fan, Adv. Mater., 1999, 11, 579–585.
4. H. Yang, N. Coombs, I. Sokolov and G. A. Ozin, Nature, 1996, 381, 589–592.
5. H. Yang, A. Kuperman, N. Coombs, S. Mamiche-Afara and G. A. Ozin, Nature, 1996, 379, 703–705.
6. H. De Paz, A. Chemtob, C. Croutxé-Barghorn, S. Rigolet and B. Lebeau, Microporous Mesoporous Mater., 2012, 151, 88–92.
7. D. A. Doshi, N. K. Huesing, M. C. Lu, H. Y. Fan, Y. F. Lu, K. Simmons-Potter, B. G. Potter, A. J. Hurd and C. J. Brinker, Science, 2000, 290, 107–111.
8. S. Nagarajan, J. K. Bosworth, C. Ober, T. P. Russell and J. J. Watkins, Chem. Mater., 2008, 20, 604–606.
9. H. De Paz, A. Chemtob, C. Croutxe-Barghorn, S. Le Nouen and S. Rigolet, J. Phys. Chem. B, 2012, 116, 5260–5268.
10. G. J. A. A. Soler-Illia, C. Sanchez, B. Lebeau and J. Patarin, Chem. Rev., 2002, 102, 4093–4138.
11. Y. Wan and Zhao, Chem. Rev., 2007, 107, 2821–2860.
12. D. Doshi, A. Gibaud, V. Goletto, L. U. Mengcheng, H. Gerung, B. Ocko, S. Han and C. J. Brinker, J. Am. Chem. Soc., 2003, 125, 11646–11655.
13. P. Falcaro, S. Costacurta, G. Mattei, H. Amenitsch, A. Marcelli, M. C. Guidi, M. Piccinini, A. Nucara, L. Malfatti, T. Kidchob and P. Innocenzi, J. Am. Chem. Soc., 2005, 127, 3838–3846.
14. P. Innocenzi, P. Falcaro, D. Grosso and F. Babonneau, J. Phys. Chem. B, 2003, 107, 4711–4717.
15. P. Innocenzi, T. Kidchob, J. M. Bertolo, M. Piccinini, M. C. Guidi and C. Marcelli, J. Phys. Chem. B, 2006, 110, 10837–10841.
16. P. Innocenzi, L. Malfatti, T. Kidchob, P. Falcaro, M. C. Guidi, M. Piccinini and A. Marcelli, Chem. Commun., 2005, 2384–2386.
17. J. C. P. Broekhoff and J. H. de Boer, J. Catal., 1968, 10, 377–390.
18. D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calvé, B. Alonso, J.-O. Durand, B. Bujoli, Z. Gan and G. Hoatson, Magn. Reson. Chem., 2002, 40, 70–76.
19. S. Ceccia, E. A. Turcato, P. L. Maffettone and R. Bongiovanni, Prog. Org. Coat., 2008, 63, 110–115.
20. M. J. Rosen, Surfactants and Interfacial Phenomena, 3rd edn, 2004.
21. G. Malucelli, R. Bongiovanni, M. Sangermano, S. Ronchetti and A. Priola, Polymer, 2007, 48, 7000–7007.
22. S. A. Bagshaw, E. Prouzet and T. J. Pinnavaia, Science, 1995, 269, 1242.
23. S.-S. Kim, T. R. Pauly and T. J. Pinnavaia, Chem. Commun., 2000, 835–836.
24. E. Prouzet and C. Boissiere, C. R. Chim., 2005, 8, 579–596.
25. D. Versace, A. Chemtob, C. Croutxe-Barghorn and S. Rigolet, Macromol. Mater. Eng., 2010, 295, 355–365.
26. P. Alexandridis and J. F. Holzwarth, Macromolecules, 1994, 27, 2414–2425.
27. P. Petrov, J. Yuan, K. Yoncheva, A. H. E. Muller and C. B. Tsvetanov, J. Phys. Chem. B, 2008, 112, 8879–8883.
28. I. Park and T. J. Pinnavaia, Microporous Mesoporous Mater., 2009, 118, 239–244.
29. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquérol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603.
30. D. Zhao, P. Yang, N. Melosh, J. Feng, B. F. Chmelka and G. D. Stucky, Adv. Mater., 1998, 10, 1380–1385.
31. K. Flodström and V. Alfredsson, Microporous Mesoporous Mater., 2003, 59, 167–176.
32. D. Zhao, Q. Huo, J. Feng, B. F. Chmelka and G. D. Stucky, J. Am. Chem. Soc., 1998, 120, 6024–6036.
33. P. Kipkemboi, A. Fogden, V. Alfredsson and K. Flodstrom, Langmuir, 2001, 17, 5398–5402.
34. B. Lindman and P. Alexandridis, Amphiphilic Block Copolymers Self-Assembly and Applications, Elsevier, Amsterdam, the Netherlands, 2000.
35. M. Sjoberg and T. Warnheim, Surf. Sci. Ser., 1997, 67, 179–205.
36. J. Penfold, E. Staples, I. Tucker and P. J. Cummins, J. Colloid Interface Sci., 1997, 185, 424–431.
37. R. Ivanova, B. Lindman and P. Alexandridis, Langmuir, 2000, 16, 3660–3675.
38. G. J. A. A. Soller-Illia, E. L. Crepaldi, D. Grosso, D. Durand and C. Sanchez, Chem. Commun., 2002, 2298–2299.
39. E. L. Crepaldi, G. J. A. A. Soler-Illia, D. Grosso, F. Cagnol, F. Ribot and C. Sanchez, J. Am. Chem. Soc., 2003, 125, 9770–9786.
40. B. Svensson, U. Olsson and P. Alexandridis, Langmuir, 2000, 16, 6839–6846.
41. S. A. El-Safty and J. Evans, J. Mater. Chem., 2002, 12, 117–123.
42. C. Belon, A. Chemtob, C. Croutxé-Barghorn, S. Rigolet, V. Le Houérou and C. Gauthier, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 4150–4158.

---

Footnote

† Electronic Supplementary Information (ESI) available: Figure S1. Effect of the initial film thickness on the XRD patterns of the 75 wt% P123-PDMOS calcined film . See DOI: 10.1039/c2ra21676k

---