ARTICLE pubs.acs.org/Langmuir
Inverse Opals of Molecularly Imprinted Hydrogels for the Detection of Bisphenol A and pH Sensing Nebewia Griffete,† Hugo Frederich,‡ Agnes Maître,‡ Serge Ravaine,§ Mohamed M. Chehimi,† and Claire Mangeney*,† †
ITODYS, Universite Paris Diderot (UMR 7086), 15 rue Jean de Baïf, 75013 Paris, France INSP, UPMC (UMR 7588), 140 rue de Lourmel, 75015 Paris, France § CRPP, Universite Bordeaux 1 (UPR 8641), 115 Avenue Schweitzer, 33600 Pessac, France ‡
ABSTRACT: Inverse opal films of molecularly imprinted polymers (MIP) were elaborated using the colloidal crystal template method. The colloidal crystals of silica particles were built by the LangmuirBlodgett technique, allowing a perfect control of the film thickness. Polymerization in the interspaces of the colloidal crystal in the presence of bisphenol A (BPA) and removal of the used template provides 3D-ordered macroporous methacrylic acid-based hydrogel films in which nanocavities derived from bisphenol A are distributed within the thin walls of the inverse opal hydrogel. The equilibrium swelling properties of the nonimprinted (NIPs) and molecularly imprinted polymers (MIPs) were studied as a function of pH and bisphenol A concentration, while the molecular structures of the bulk hydrogels were analyzed using a cross-linked network structure theory. This study showed an increase in nanopore (mesh) size in the MIPs after BPA extraction as compared to NIPs, in agreement with the presence of nanocavities left by the molecular imprints of the template molecule. The resulting inverse opals were found to display large responses to external stimuli (pH or BPA) with Bragg diffraction peak shifts depending upon the hydrogel film thickness. The film thickness was therefore shown to be a critical parameter for improving the sensing capacities of inverse opal hydrogel films deposited on a substrate.
’ INTRODUCTION The increased global concern about manmade chemicals such as endocrine disruptors (EDCs) stimulates the development of selective and sensitive analytical methods to detect trace amounts in environmental, biological, and foodstuff samples.1 Most of the methods to date are based on high performance liquid chromatography (HPLC),2 gas chromatography (GC),3 and capillary electrophoresis (CE).4 The use of MIPs as selective sorbent materials allows for a customized sample treatment step prior to the final determination. This is of special interest when the sample is complex and the presence of interferences could prevent final quantification via typical chromatographic techniques coupled to common detectors.5 Nevertheless, these current approaches involve not only expensive instruments, but also a large number of separate analytical procedures, resulting in a complex, time-consuming, and laborious screening procedure. Thereby, it is highly desirable and useful to develop novel approaches for easy and rapid drug detection without costly instruments and prolix processes, but that exhibit high sensitivity and specificity. Recently, the group of Li proposed an original approach combining molecular imprinting and colloidal crystal to prepare imprinted photonic polymers with 3D, highly ordered, macroporous structures and specific binding nanocavities for a rapid and selfreporting assay with high sensitivity and specificity.68 Because of their periodic porous structure, such materials (inverse opals) r 2011 American Chemical Society
exhibit fascinating optical properties (Bragg diffraction) and bright structural colors. Particularly, if these highly ordered macroporous materials are made from responsive polymer hydrogels, they are able to swell or shrink in aqueous solution upon molecular recognition or environmental conditions to change the periodic spacing, leading to a change in optical properties. Combining this approach with the LangmuirBlodgett (LB) technique, our group recently elaborated inverse opal hydrogel (IOH) films deposited on a transparent PMMA plate, with thicknesses of 2.8 μm and containing a planar defect layer.9 The defect layer was shown to enhance the sensing capacities of the optical sensor to detect bisphenol A (BPA), an endocrine disruptor (EDC).10 However, many factors can affect the swelling behavior of thin hydrogel films deposited on a substrate and thereby their sensing properties. Among these factors, the adhesion of the film is a critical parameter, which hampers the swelling along the plane of the substrate and favors vertical swelling in the vicinity of the surface, as was already described by Braun et al.11 One could expect that this constraint on swelling should become weaker for thick hydrogel films, far from the substrate. Therefore, the film thickness should be another important factor influencing the Received: July 22, 2011 Revised: November 14, 2011 Published: November 16, 2011 1005
dx.doi.org/10.1021/la202840y | Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
Scheme 1. Schematic Illustration of the Procedure Used for the Preparation of a Molecularly Imprinted Inverse Opal Hydrogel (IOH) Film
swelling properties of the hydrogels. Nevertheless, this parameter has never been investigated so far, probably due to the difficulty to prepare inverse opal hydrogel films with controlled thicknesses. In this article, we address this issue by taking advantage of the LB technique to perfectly control the number of silica particle layers constituting the colloidal crystal template1214 and therefore the thickness of the resulting IOH (see Scheme 1). By studying the pH and BPA response of IOH films with various thicknesses, we could bring new insights on their swelling behavior and optimize their detection limit, which is a crucial parameter for sensing applications. The hydrogel was made of a molecularly imprinted poly(methacrylic acid) reticulated using ethyleneglycol dimethacrylate (EGDMA) as a comonomer and BPA as the template molecule. We first examined the swelling properties of the bulk hydrogels of nonimprinted and molecularly imprinted polymers especially in response to pH or various BPA concentration changes. Using a cross-linked network structure theory,1518 we could obtain insights on the molecular structures of the bulk hydrogels and on their nanopore (mesh) sizes. The responses of the corresponding inverse opal hydrogels to external stimuli (pH and BPA) were then investigated as a function of film thickness, evidencing the importance of this parameter for improving their sensing capacities.
’ EXPERIMENTAL SECTION Materials. Methacrylic Acid (MAA), ethylene glycol dimethacrylate (EGDMA), and BPA were purchased from Aldrich. 2,20 -Azobis (isobutyronitrile) (AIBN) was obtained from Fluka. Tetraethoxysilane (TEOS, Fluka), ammonia (29% in water, J. T. Baker), styrene (Aldrich), azodiisobutyramidine dihydrochloride (Aldrich), and aminopropyltriethoxysilane (Aldrich) were used without purification. Cyclohexane (J. T. Baker, France, 99%) and dibutyltin dilaurate (DBTDL, Sigma-Aldrich, 98%) were used as received. Tolylene-2,4-diisocyanate (TDI,μ SigmaAldrich) was purified by distillation under vacuum. Ethylene glycol (EG, Sigma-Aldrich, 99%) was used as received. PMMA slides with 40 30 3 mm3 for photonic film supports were cleaned with anhydrous ethanol. Hydrofluoric acid was purchased from Acros. Preparation of Nonimprinted and Molecularly Imprinted Hydrogels. The typical preparation of hydrogel of MAA-co-EGDMA was as follows: 10 g of MAA, 0.05 g of EGDMA, and 10 mL of ethanol were mixed thoroughly. To this solution was added 0.1 g of AIBN. After being mixed completely, the obtained clear solution was deoxygenated and kept for at least 30 min. The hydrogel precursor solution was irradiated in a photochemical reactor operating at 365 nm. After polymerization, the hydrogels were cut into cylinders with a diameter of 20 mm and a thickness of 10 mm. Hydrogels were kept in distilled water for at least 24 h, replacing the water daily to remove any impurities, such as monomers,
initiator, etc. For the MIP preparation, 0.1 g of BPA was added. BPA was removed from the polymer thanks to a 2.0 M acetic acid solution during 1 h. Preparation of Hydrogel Inverse Opals. The method employed for the synthesis of 280 and 265 nm silica particles and corresponding colloidal crystals was similar to the one described in a previous works.1214 Prior to fabrication of a photonic imprinted film, a homogeneous monomer mixture of MAA (0.4 g), EGDMA (0.02 g), AIBN (0.04 g), and BPA (0.04 g) in ethanol (0.4 g) was degassed under nitrogen for 3 min and dropped on a silica colloidal crystal. Glass slides with a colloidal-crystal film were coated with a PMMA slide (40 30 3 mm3) and held together to retain the abovementioned precursor mixture. Once the colloidal crystal of the formed sandwich structure became transparent, a successful infiltration process was completed. After the removal of excess precursors, photopolymerization was performed under UV light at 365 nm for 2 h. The sandwich structure was immersed in 4% hydrofluoric acid solution for 3 h to separate double slides and completely etch the silica colloids. The polymer film on the PMMA substrate was soaked in a 2.0 M acetic acid solution for 2 h and rinsed with deionized water to remove the BPA. The photonic nonimprinted polymer was prepared using the same procedure but in the absence of BPA. Recognitive Studies. Batch adsorption experiments were carried out by allowing a weighed amount of polymer to reach equilibrium with BPA solution of known concentration. In typical experiments, MIPs and NIPs (35 mg) were placed in 20 mL of 0.1 mol/L bisphenol A solution in distilled water. After 24 h incubation, the residual concentration of BPA in the aqueous solution was determined by UVvis spectroscopy. Instrumentation. SEM micrographs were obtained with a Cambridge 120 apparatus fitted with a zirconated tungsten filament. The acceleration voltage was set at 20 kV. All specimens were coated with gold prior to analysis to minimize static charging effects. UVvis analysis for the determination of BPA concentration in solution was carried out using a Cary Win UV spectrophototometer. The inverse opal hydrogels were optically characterized by Cary Win 500 spectrophotometer in specular geometry. The samples were deposited in a glassware where the different solutions were added. The samples were illuminated with a fibered and collimated halogen source covering the whole 350800 nm spectral range, and the reflected light was collected by a second optical fiber symmetric to the first one. The fibers were mounted on rotating stages allowing a precise selection of the incident and collection angles. The light spot on the opal had a size of 4 mm2 (at 20° incidence), and the distance between the sample and the optical fiber was 10 cm. Specular reflection spectra were normalized to the incident light spectrum. After each spectrum, the solution was extracted and replaced by another one. All of the experiments were done at room temperature. Determination of Equilibrium Swelling Ratio. Swelling experiments were performed on nonimprinted and imprinted hydrogels in phosphate buffer at three different pH values (2, 4, and 6) at 25 °C. 1006
dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
The % swelling ratio (% SR) and the % equilibrium swelling (% ES) were calculated as: mt md 100 ð1Þ %SR ¼ md %ES ¼
ms md 100 md
ð2Þ
where mt and md are weights of swollen hydrogels for a given time and dried hydrogels, respectively. Here, ms is the weight of the swollen hydrogel at equilibrium. Initial weight of the hydrogel, after drying, was taken over a single-pan digital microbalance (sensitive to (0.01 mg). The polymer volume fraction in the gel immediately after preparation (relaxed state), ν2r, and the polymer volume fraction of the swollen gels (swollen state), ν2s, were determined using eqs 3 and 4: 0 11 mr 1 F2 C B B C md C ð3Þ ν2r ¼ B1 þ @ A F1 11 ms 1 F2 C B B C md C ¼ B1 þ @ A F1 0
ν2s
fi ν2s V̅ r
ð9Þ
where fi is the mol fraction of the ionic unit in the gel system and V r is the average molar volume of polymer repeat units. Combining eqs 8 and 9, we obtain: ! i2 ν2s 2 f i 2 πion ¼ RT ð10Þ 4I V̅ 2r
Ka ¼
ð6Þ
where R is the gas constant and T is the temperature. The most important feature of cross-linked polymers that differs from the un-cross-linked polymer solutions of the same chemical nature may be a memory effect of the initial condition. Once a gel is formed, its structure is more or less fixed depending on the initial condition, resulting in an emergence of the frozen structure. Thus, one should also take into account the parameters at the initial condition for the study of gels.20 To describe the elastic contribution πel to the swelling pressure, we use here the phantom network model:21 ! 2=3 1=3 ð1 2=jÞF2 ν2r ν2s πel ¼ RT ð7Þ Mc where ϕ is the number of branches originating from a cross-linking site. The ionic contribution πion to the swelling pressure is caused by the concentration difference of counterions between the hydrogel
½Hþ ½A ½HA
ð11Þ
where [HA] is the concentration of undissociated carboxylic acid groups, [A] is the concentration of dissociated carboxylic acid groups, and [H+] is the concentration of hydrogen ions. i is defined as follows: ½A ½A ½AH ¼ i¼ ½A ½AH þ ½A 1 þ ½AH
ð12Þ
With eq 12, eq 13 may be rewritten as follows: i¼
ð5Þ
The mixing term is satisfactorily represented by FloryHuggins type expression of the form:19 RT ½lnð1 ν2s Þ þ ν2s þ χν2s 2 ν1
c2 ¼
ð4Þ
’ THEORETICAL BASIS The pH-responsive swelling behavior of the hydrogels was analyzed within the framework of the FloryRehner theory of swelling.15 According to the FloryRehner theory, the osmotic pressure π of a hydrogel during swelling is given as the sum of the pressures due to polymersolvent mixing (mix), due to deformation of network chains to a more elongated state (el), and due to the nonuniform distribution of mobile counterions between the hydrogel and the external solution (ion):
πmix ¼
In the above equation, i is the degree of ionization, I is the ionic strength of the swelling medium, and c2 is the concentration of the ionizable polymer (mol cm3). c2 can be written in terms of polymer structural parameters for copolymeric hydrogels as:
For the hydrogels with monoprotic acid moieties, there is only one equilibrium:
where mr is the weight of the hydrogel after preparation, and F1 and F2 are the densities of the solvent and polymer network, respectively.
π ¼ πmix þ πel þ πion
and the outer solution. The ideal Donnan theory gives πion as the pressure difference of mobile ions inside and outside the hydrogel: ! i2 c22 πion ¼ RT ð8Þ 4I
K a =½Hþ Ka Ka ¼ þ ¼ pH 1 þ K a =½Hþ ½H þ K a þ Ka 10
ð13Þ
If we substitute in eq 13, the following equation can be obtained for πion: !2 Ka RTν2s 2 f i 2 πion ¼ ð14Þ 4I V̅ r 2 10pH þ K a The complete equilibrium expression, which accounts for the mixing, elastic-retractive, and ionic contributions to the osmotic pressure of monoprotic polymeric networks, is given below: !2 Ka V 21 f i 2 1 2 RT 2 ν2s lnð1 ν2s Þ ν2s 10pH þ K a 4I V̅ r 2=3 5=3
¼χ þ
ð1 2=jÞV 1 Fν2r ν2s Mc
ð15Þ
’ RESULTS AND DISCUSSION Swelling Studies. Swelling in Deionized Water. Characterization of the cross-linked structures of the MIPs and NIPs samples 1007
dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
Figure 1. % Swelling ratio of (a) nonimprinted hydrogel and (b) imprinted hydrogel in distilled water () and in buffer solutions at pH 2.0 (- - - -) and 6.0 ( 3 3 3 3 ), at fixed I = 0.10 M. Insets show photographs of the BPA-free imprinted hydrogels after exposure to pH 2 (bottom inset) and pH 6 (upper inset) phosphate buffers.
was achieved by equilibrium swelling studies in distilled water. Typically, dry hydrogel discs were immersed in distilled water for 4 days leading to a progressive swelling of the gels (see typical hydrogel discs in the inset of Figure 1). The swelling experiments for MIPs were performed after extraction of BPA. The swelling at equilibrium was observed to be significantly higher for the BPA-free MIP gels with SR = 378% as compared to the 256% measured for the NIPs (see Figure 1). This indicates that the presence of BPA nanocavities in the polymer network after extraction of the template molecules leads to the formation of a more nanoporous structure. To obtain a better insight on the structural properties of the hydrogels, the equilibrium swelling data were used to evaluate their crosslinked structure. Typically, the number average molecular weight between cross-links, Mc was calculated.22 This parameter is an indication of the cross-linked nature of the hydrogel, as high values of Mc imply loosely cross-linked hydrogel. The results of this study are presented in Table 1.23 They indicate that the Mc value of the nonimprinted polymer is significantly lower (780 g mol1 corresponding to ∼9 methacrylic acid units) than that of the BPA-free imprinted polymer (2100 g mol1 corresponding to ∼24 methacrylic acid units), resulting in a lower cross-link density (νe = F2/Mc)24 for the later polymer. These differences in cross-link densities between the MIPs and NIPs are probably due to the inclusion of the template molecule during polymerization, which creates additional free space or vacuoles of which the polymeric network must form around. The higher penetrant uptake of recognitive polymers as compared to control polymers can therefore be attributed to more porous networks. Mesh Size and Micro- and Nanoporous Structural Analysis. The free spaces inside the hydrogel networks are often regarded as the “pores”. Depending upon the size of these pores, hydrogels can be macroporous, microporous, or nonporous (also often named nanoporous) (see Scheme 2). A structural parameter that can be used to describe the size of the recognitive pores is the correlation length, ξ, which is defined as the linear distance between two adjacent cross-links and can be calculated using the following equation: ξ¼
2 1=2
αðr o Þ
ð16Þ
Table 1. Structural Parameters of the MIP and NIP Hydrogels hydrogels
ν2s
Mc (g mol1)
Na
ξ (Å)
νe 104 (g cm3)
NIPs
0.21
780
9
42
16.6
BPA-free MIPs
0.16
2100
24
75
6.3
a
N corresponds to the number of methacrylic acid units between cross-links.
Scheme 2. Schematic Representation of the Mesh Size in a Molecularly Imprinted Hydrogel
Here, α is the elongation ratio of the polymer chains in any direction, and (ro2)1/2 is the root-mean-square, unperturbed, end-to-end distance of the polymer chains between two neighboring cross-links. For isotropically swollen hydrogels, the elongation ratio, α, can be related to the swollen polymer volume fraction, ν2s, using eq 17. α ¼ ν2s 1=3
ð17Þ
The unperturbed end-to-end distance of the polymer chain between two adjacent cross-links can be calculated using eq 18, where Cn is the Flory characteristic ratio, l is the length of the bond along 1008
dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
the polymer backbone (for methacrylates polymers, 1.54 Å), and N is the number of links per chain that can be calculated by eq 19. ðr o 2 Þ1=2 ¼ l ðCn NÞ1=2 N ¼
ð18Þ
2Mc Mr
ð19Þ
In eq 19, Mr is the molecular weight of the repeating units from which the polymer chain is composed. Finally, when one combines eqs 1619, the correlation distance between two adjacent cross-links in a swollen hydrogel can be obtained: !1=2 2C M n c l ð20Þ ξ ¼ ν2s 1=3 Mr The network mesh size, ξ, was calculated using eq 20. In this expression, the Flory characteristic ratio Cn was CnPMAA = 14.6.25 The results of these calculations are presented in Table 1 and are quite revealing of the modifications of porosity induced by the imprinting process, with ξ values noticeably higher for BPA-free MIPs as compared to NIPs. Swelling in Buffer Solutions at Various pH’s. The pH-responsive properties of the hydrogels were studied by dynamic swelling experiments in PBS solutions with pH values varying from 2 to 6 (at a fixed ionic strength of 0.1M). Figure 1 shows that the equilibrium swelling percents of both MIPs and NIPs become higher with increasing pH, due to the ionization of the carboxylic groups above their pKa. Typical numerical photos of the swollen gels at pH 2 and 6 are shown in the inset of Figure 1. To obtain new insights on the evolution of the hydrogel structures in response to pH changes, the effects of the external pH on the mesh size of MIPs and NIPs were determined, as summarized in Table 2. This study shows that the mesh sizes increase with pH on both sides of the pKa value of PMAA (pKaPMAA = 5.65).26 This behavior is due to the swelling of the hydrogels when the carboxylic acid groups of the PMAA chains become ionized and generate an electrostatic repulsion between these charged groups.27 It is to note that the relative increase in mesh size with pH is enhanced for the BPA-free MIPs compared to the NIPs, in agreement with a lower cross-link density. Therefore, these results demonstrate that, under the same polymerization conditions, the presence of Table 2. Influence of pH on the Mesh Size of the MIP and NIP Hydrogels pH 2
pH 4
pH 6
ν2s
ξ (Å)
ν2s
ξ (Å)
ν2s
ξ (Å)
NIP
0.23
41
0.21
42
0.17
45
BPA-free MIP
0.20
70
0.18
73
0.12
83
hydrogels
the template molecule inside the MIPs induces strong modification of the hydrogel network structure, leaving larger pores after extraction of the template, as compared to NIPs. Although usual specificity tests classically compare the capacities of MIPs and NIPs to bind a template molecule, it is important to note that these differences in polymer porosity existing between imprinted and their corresponding nonimprinted control structures could overshadow the contribution of enhanced affinity for the template. Rebinding of BPA and Evaluation of the Polymers. The swelling behavior and recognitive properties of the hydrogels in response to the presence of the BPA template molecule or a structurally similar one (bisphenol F (BPF)) were then investigated. The bulk MIPs and NIPs were incubated with BPA and BPF (0.1 mol L1). After 24 h of incubation, the relative weight gains ((msBPA m0)/m0) were measured, where msBPA corresponds to the weight of the gel after incubation with BPA (msBPA) or BPF (msBPF), and m0 is the weight of the gel just before incubation with the target molecule. The values of relative weight gains are reported in Table 3. They show that the highest weight gain occurs when the BPA-imprinted hydrogel is incubated with BPA, while its incubation with BPF or the incubation of the control samples (NIP) with these molecules lead to smaller weight gain values. As these values correspond to both the uptake of water by the hydrogel and the adsorption of the target molecule, it could not be used to calculate the imprinting efficiency of the MIPs. Instead, the recognitive capacity (RC) defined as the mass of adsorbate per unit mass of adsorbent was calculated from the following equation: RC ¼
ðCi Ce ÞV t md
ð21Þ
where RC is in mg adsorbate/g dry gel adsorbent, Ci and Ce are initial and equilibrium concentration of adsorbate in solution in mg mL1, Vt is the volume of solution in mL, and md is the mass of dry gel adsorbent in g. This parameter permitted us to determine the imprinting efficiency, IE, defined by eq 22. IE ¼
RCMIP RCNIP
ð22Þ
The results presented in Table 3 evidence a superior recognitive capacity of MIPs for BPA as compared to NIPs with an imprinting efficiency of 3.5, while it falls to 1.8 for the structurally similar BPF molecule. Therefore, this study highlights the affinity and higher capacity of the imprinted polymer toward the template molecule and gives convincing proof of its specificity and selectivity to BPA. Inverse Opals of NIPs and MIPs. The three-step approach employed for the construction of BPA-imprinted photonic polymer hydrogels includes the preparation of a colloidal-crystal
Table 3. Relative Weight Gains, Recognitive Capacity, and Imprinting Efficiency of MIPs and NIPs in the Presence of BPA or BPF BPA hydrogels
(msBPA m0)/m0 %
MIPs
15
NIPs
6
1
(msBPF m0)/m0 %
BPF
RC (mg g )
IE
5
2.1
3.5
9
0.6 1009
1
RC (mg g )
IE
0.9
1.8
0.5 dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
Figure 2. SEM images of (a) the colloidal-crystal template elaborated from 280 nm silica particles and (b) the resulting BPA-imprinted photonic polymer film.
Figure 3. (a) Optical response of the inverse opal hydrogel films after BPA extraction upon soaking in buffer solutions at different pH values. (b) Bragg peak variation (Δλ) versus pH for IOH films with various thicknesses.
template, the polymerization of the preordered complex of BPA with functional monomers in the interspaces of the colloidal crystal, and the removal of the used templates (colloid particles and BPA molecules). The silica colloidal-crystal templates deposited on glass substrates were obtained from monodispersed silica spheres using the LangmuirBlodgett technique. The template molecule (BPA), functional monomer (MAA), and cross-linking agents (EGDMA) were then mixed to generate a prepolymerization solution. The mixture was filled into the void spaces of the colloidal-crystal array by capillary force by using a sandwich structure (see the Experimental Section). Upon polymerization, the structure was frozen in a 3D network of polymers. The removal of silica particles and the embedded BPA molecules from the imprinted polymer matrix affords 3D highly ordered and interconnected macroporous arrays with specific nanocavities that could specifically interact with the BPA molecule through noncovalent interactions. Figure 2 shows the typical SEM images of the used colloidalcrystal template and the resulting hydrogel inverse opal film. One observes that the close-packed structured photonic crystal film is converted to air spheres surrounded by hydrogel after polymerization and etching of the silica particles. The persistence of the periodic structure under vacuum and the moderate shrinkage suggest that the IOH is mechanically robust.
Because of the periodicity of the inverse opal structure, the optical properties of these materials are those of a photonic crystal. However, it is noteworthy that opals obtained from the LangmuirBlodgett technique are a combination of random cfc and hc structures, so that they can be analyzed as multilayer films made of 2D hexagonal structured particle monolayers. Therefore, the reflection peaks recorded by the spectrophotometer in specular geometry can be interpreted as interferences between light reflection on dense (111) planes and can be analyzed with Bragg’s law, as follows: λmax ¼ 1:633ðd=mÞðD=D0 Þðna 2 sin θ2 Þ1=2
ð23Þ
where λmax is the wavelength at maximum intensity of the Bragg diffraction peak, d is the sphere diameter of the silica colloidal particle, m is the order of Bragg diffraction, D/D0 is the degree of swelling of the gel (D and D0 denote the diameters of the gel in the equilibrium state at a certain condition and in the reference state, respectively), na is the average refractive index of the porous gel at a certain condition, and θ is the angle of incidence. According to this equation, if the molecular recognition process could cause the swelling or shrinkage of the prepared hydrogel, the readable optical signal may be detectable. Nevertheless, it is to note that the strong adhesion of the IOH to the substrates prevents the hydrogel films from macroscopically swelling in 3-D without 1010
dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
Figure 4. (a) Optical response of the IOH films after BPA extraction ( ) and upon soaking in solutions with BPA 1014 M ( 3 3 3 3 ), 1013 M ( 3 3 3 3 ), 1012 M ( 3 3 ), 109 M (), and 105 M (- - - -). (b) Plot of the Bragg peak shifts observed for the BPA-imprinted films (b) and the nonimprinted films (2) in response to different BPA concentrations. The response of the BPA-imprinted films upon soaking in solutions with different concentrations of BPF is quasi the same as the NIP, and consequently points are superimposed.
Scheme 3. Molecular Structure of Bisphenol A (BPA), Bisphenol F (BPF), and 2-Phenylphenol (2PP)
Table 4. Effect of the IOH Film Thickness on the BPA Detection Limit and on the Bragg Peak Variation at 107 M IOH film thickness (μm) Δλ/λ at [BPA] = 107 M detection limit of BPA (mol/L)
delaminating from the substrate. Therefore, for thin IOH films, the swelling/shrinkage process occurs mainly in a vertical direction with regard to the substrate, in agreement with previously reported work by Braun and co-workers.9 Nevertheless, for thicker films, the constraint due to the adhesion of the hydrogel to the substrate should be weakened far from the surface. Therefore, to evaluate the effect of the IOH film thickness upon the reversible swelling process, a study of the pH response of nonimprinted films of various thicknesses was performed. The control of thickness was possible thanks to the use of the LangmuirBlodgett technique, which allowed one to perfectly control the number of layers of silica particles deposited to build the colloidal crystal template. pH Response of IOH. As shown in Figure 3a, the Bragg peak shifted markedly as a function of the pH from λmax = 505 nm at pH 2 to 687 nm at pH 9, in agreement with the pH-responsive character of the IOH. Indeed, the pH dependence of diffraction results from an ionization of the IOH carboxyl groups, which leads to immobilization of counterions inside the gel. This results in an osmotic pressure, which swells the gel against its restoring elastic constant. Thus, an increased pH increases the ionization; the gel swells and the diffraction red shifts. Interestingly, the diffraction peak shifts much more as the hydrogel thickness increases, as shown in Figure 3b. These results underline the key role of the hydrogel/substrate interface for restraining the swelling capacities of the hydrogel films and evidence the importance of film thickness to provide greater degrees of freedom to the deposited hydrogels. Recognitive and Sensing Studies of MIP Inverse Opals. Recognitive studies toward detection of BPA were then performed on MIP inverse opals. Figure 4 shows the optical response of a MIP inverse opal film soaked in a series of various concentrations of BPA. The Bragg diffraction peak shifts regularly with increasing concentrations to the longer wavelength region until
2.8 8 109
4.2 30 1012
6.6 80 1013
reaching a plateau value at around 107 M, which implies that the IOH film is saturated to adsorbed analyte. To further elucidate the molecular recognition properties of the imprinted materials, the nonimprinted 3D-ordered macroporous hydrogels (NIP) as control samples were also synthesized under the same preparation conditions as the imprinted photonic polymer, but in the absence of BPA templates. Figure 4b displays the evolution of the diffraction peak of the nonimprinted photonic films upon exposure to solutions containing various concentrations of BPA. Probably due to some nonspecific adsorption, only a slight fluctuation in the Bragg peak was observed in the case of the nonimprinted hydrogel film, which is quite different from the recognition behavior exhibited by the prepared MIP. These results tend to indicate that the microenvironments created by molecular imprinting are responsible for the observed results of BPA-imprinted photonic polymers. The selectivity tests of the BPA-imprinted films were carried out by using two molecules (BPF and 2-phenylphenol), which are structurally similar to BPA, as the reference compounds (see Scheme 3). As compared to BPA, these molecules can only induce slight optical shifts under the same measurement conditions (see Figure 4b). The fact that the prepared MIP can delicately discriminate between these molecules suggests that the cooperative effect of shape, size, and interaction sites of the formed binding sites plays a critical role in the high-selectivity molecular recognition process of MIP: only BPA rather than other pollutant molecules can specifically occupy the imprinted nanocavities within the MIP film and cause the volume change of the hydrogel film, thereby inducing the shift of the Bragg diffraction peak. Influence of the Film Thickness on the Detection Limit. Three colloidal crystals of various numbers of layers were used as templates for the preparation of the corresponding molecularly imprinted polymer inverse opals with thicknesses of (i) 2.8, 1011
dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
Langmuir
ARTICLE
Table 5. Macropore Diameters (Dpores in nm) of IOH Estimated by Fitting the Reflectivity Spectra with Bragg’s Law (Eq 23)
a
Dpores BPA 1014 Dpores BPA 1013 Dpores BPA 1012 Dpores BPA 109 Dpores BPA 107
thickness of the MIP inverse
Dpores inverse
Dpores after
opal
opal
extraction
M
M
M
M
M
d280a = 2.8 μm
278
258
258
258
259
260
261
d280 a = 4.2 μm
271
252
252
252
254
257
264
d265 a = 6.6 μm
208
191
192
193
194
201
207
The indices of d correspond to the diameters of the silica particles used to build the colloidal crystal template.
(ii) 4.2, and (iii) 6.6 μm. Upon exposure to BPA solutions in phosphate buffer, the amplitude of the diffraction peak shift was shown to increase with film thickness. Concomitantly, the detection limit was improved with increased film thickness, as summarized in Table 4. To elucidate the structural modifications occurring within these films of MIP inverse opals upon BPA detection, the spectra were fitted with Bragg’s law (from eq 1) to estimate the evolution of their macropore diameters (Dpores). As shown in Table 5, just after silica etching by HF, Dpores was found to be slightly smaller than the diameter of the silica particles used in the template, indicating a contraction of the macropores upon expulsion of water, during the drying process. Treatment of the IOH by acetic acid results in a further decrease of Dpores evidencing a collapse of the hydrogel structure due to the release of BPA molecules from their nanocavities. In the presence of BPA, the pore diameter appears to increase progressively. Interestingly, the amplitude of these changes increases with the thickness of the IOH film (from a ΔDpores = 3 nm for 2.8 μm film thickness to ΔDpores = 16 nm for 6.6 μm thickness). Similarly, a detectable Dpores variation (g2 nm) is observed at higher BPA concentrations for thin 2.8 μm films ([BPA] = 109 M) as compared to thicker 6.6 μm films ([BPA] = 1013 M). These results further highlight the importance of film thickness as a critical parameter for improving the swelling capacity and the sensing properties of IOH films.
’ CONCLUSION The combination of molecularly imprinted hydrogels and photonic crystals is a promising strategy for the development of BPA optical sensors. Indeed, it was shown that the introduction of nanocavities left by the BPA template after extraction from the MIPs leads to an increase of the mesh size in the hydrogel network and enhances its swelling capacities. Furthermore, it retains a high affinity and recognitive capacity toward the template BPA molecule as well as a reasonable specificity. Concomitantly, the three-dimensional, highly ordered macroporous arrays of molecularly imprinted hydrogels formed around the colloidal crystal templates display Bragg diffraction peaks that are sensitive to the lattice changes of the structure upon pH modifications or BPA recognition. Interestingly, by using the LangmuirBlodgett technique to elaborate the colloidal crystal template, we could achieve a perfect control of the IOH film thicknesses and obtain original results on the influence of this parameter upon their swelling properties and sensing capacities. It was shown that thick films generate (i) an improved pH response and (ii) a lower detection limit for BPA sensing (down to 1013 mol L1), as compared to thinner ones. ’ AUTHOR INFORMATION Corresponding Author
*Phone: 33-01-57276878. Fax: 33-01-572772. E-mail: mangeney@ univ-paris-diderot.fr.
’ REFERENCES (1) Mattiessen, P.; Johnson, I. Environ. Pollut. 2006, 146, 9. (2) Li, X. J.; Zeng, Z. R.; Gao, S. Z.; Li, H. B. J. Chromatogr., A 2004, 1023, 15. (3) Spyridaki, M. H.; Kiousi, P.; Vonaparti, A.; Valavani, P.; Zonaras, V.; Zahariou, M.; Sianos, E.; Tsoupras, G.; Georgakopoulos, C. Anal. Chim. Acta 2006, 242, 573. (4) Macchia, M.; Manetto, G.; Mori, C.; Papi, C.; Pietro, N. D.; Salotti, V.; Bortolotti, F.; Tgliaro, F. J. Chromatogr., A 2001, 924, 499. (5) Feng, Q.; Zhao, Li.; Yan, W.; Lin, J.; Zheng, Z. J. Hazard. Mater. 2009, 167, 282. (6) Wu, Z.; Tao, C.; Lin, C.; Shen, D.; Li, G. Chem.-Eur. J. 2008, 14, 1358. (7) Hu, X.; Huang, J.; Zhang, W.; Li, M.; Tao, C.; Li, G. Adv. Mater. 2008, 20, 4074. (8) Hu, X.; Li, G.; Li, M.; Huang, J.; Li, Y.; Gao, Y.; Zhang, Y. Adv. Funct. Mater. 2008, 18, 575. (9) (a) Griffete, N; Frederich, H.; Maitre, A.; Schwob, C.; Chehimi, M. M.; Mangeney, C. J. Colloid Interface Sci. 2011, 364, 18. (b) Griffete, N.; Frederich, H.; Maitre, A.; Ravaine, S.; Chehimi, M. M.; Mangeney, C. J. Mater. Chem. 2011, 21, 13052. (10) Lang, I. A.; Galloway, T. S.; Scarlett, A.; Henley, W. E.; Depledge, M.; Wallace, R. B.; Melzer, D. J. Am. Med. Assoc. 2008, 300, 1303. (11) Lee, Y.; Pruzinsky, S. A.; Braun, P. V. Langmuir 2004, 20, 3096. (12) Masse, P.; Reculusa, S.; Clays, K.; Ravaine, S. Chem. Phys. Lett. 2006, 422, 251. (13) Masse, P.; Pouclet, G.; Ravaine, S. Adv. Mater. 2008, 20, 584. (14) Masse, P.; Vallee, R. A. L.; Dechezelles, J.-F.; Rosselgong, J.; Cloutet, E.; Cramail, H.; Zhao, X. S.; Ravaine, S. J. Phys. Chem. C 2009, 113, 14487. (15) Flory, P. J.; Rehner, J. J. Chem. Phys. 1943, 11, 521. (16) Brannon-Peppas, L.; Peppas, N. A. Chem. Eng. Sci. 1991, 46, 715. (17) Peppas, N. A.; Merrill, E. W. J. Polym. Sci., Part A: Polym. Chem. 1976, 14, 441. (18) Aykara, T. C.; Bozkaya, U.; Lu, O. J. Polym. Sci., Part B: Polym. Phys. 2003, 41, 1656. (19) Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953. (20) Norisuye, T.; Masui, N.; Kida, Y.; Ikuta, D.; Kokufuta, E.; Ito, S.; Panyukov, S.; Shibayama, M. Polymer 2002, 43, 5289. (21) James, H. M.; Guth, E. J. Polym. Sci. 1949, 4, 153. (22) As the swelling experiments were performed in distilled water, with pH values below the pKa of PMAA, the experimental values of Mc were calculated using the following equation for nonionic phantom networks: Mc ¼
ð1 2=jÞV 1 Fν2r 2=3 ν2s 5=3 lnð1 ν2s Þ þ ν2s þ χν2s 2
(23) The relevant experimental parameters that were used are as follows: ionic strength, I = 1 104 mol cm3; molar volume of the solvent, V1 = 18 cm3 mol1; the number branches originating from a cross-linking site, Φ = 3; Ka = 2.2 106, F1 = 1, and F2 = 1.33, and the Flory interaction parameter, χ = 0.54. (24) Martens, P.; Anseth, K. S. Polymer 2000, 41, 7715. (25) Spizzirri, U. G.; Peppas, N. A. Chem. Mater. 2005, 17, 6719. (26) Greenwald, H. L.; Luskin, L. S. Handbook of Water-Soluble Gums and Resins; Davidson, R. L., Ed.; 1980. (27) Kim, B.; Peppas, N. A. Macromolecules 2002, 35, 9545.
1012
dx.doi.org/10.1021/la202840y |Langmuir 2012, 28, 1005–1012
