Fast response photonic crystal pH sensor based ... - Semantic Scholar
Sensors and Actuators B 150 (2010) 183–190
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Fast response photonic crystal pH sensor based on templated photo-polymerized hydrogel inverse opal Jinsub Shin a , Paul V. Braun b,∗∗ , Wonmok Lee a,∗ a
Department of Chemistry, Sejong University, 98 Gunja-Dong, Gwngjin-gu, Seoul 143-747, Republic of Korea Department of Materials Science and Engineering, Materials Research Laboratory, Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA b
a r t i c l e
i n f o
Article history: Received 10 March 2010 Received in revised form 17 June 2010 Accepted 10 July 2010 Available online 16 July 2010 Keywords: Fast response Photonic crystal pH sensor Hydrogel Inverse opal
a b s t r a c t Polymer hydrogels can exhibit large reversible volume changes in response to external stimuli, and thus are regarded as excellent materials for chemical sensors. In this report, we demonstrate a mechanically robust and fast response photonic crystal pH sensor fabricated by templated photo-polymerization of hydrogel monomers within the interstitial space of a self-assembled colloidal photonic crystal. Throughout a rigorous optimization of the photo-polymerization, pH sensors showing a response time of less than 10 s upon a pH change were fabricated. Repeated pH changes revealed that the sensor has a long lifetime (>6 months) without degradation of the response time or reproducibility in pH-driven color change. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Stimuli-responsive chemical sensors are of great importance for environmental, food, and biological applications. For sensing chemical stimuli, complicated electrical or optical detection strategies are often used which require sophisticated and expensive detection systems. Recently developed photonic crystal hydrogel sensor strategies which rely on the diffraction induced color change of the sensor are now however enabling the detection of the analyte molecules simply with the naked eye [1]. A common detection mechanism is based on a volume change of hydrogel exerted by Donnan potential between the analyte ions and their receptor which is covalently bound to the hydrogel building block. In order to compensate the local ion inhomogeneity, there is influx of water molecules to the hydrogel. The optical readout of such hydrogel structures is because they are imprinted with a periodic structure, forming a so-called photonic crystal, and the volume change in the periodic structure results in the color change of the sensor. Such color is often called ‘structural color’ due to the origin of color. Important advantages of the hydrogel photonic crystal sensor are that relatively simple optical detection techniques are
∗ Corresponding author. Tel.: +82 2 3408 3212; fax: +82 2 462 9954. ∗∗ Corresponding author. E-mail addresses: [email protected] (P.V. Braun), [email protected] (W. Lee). 0925-4005/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2010.07.018
possible, and that a wide variety of analytes including protons [2–5], metal ions [6], glucose [7–9], and volatile organic compounds (VOC) [10], to name a few have been successfully detected using such sensors. Much of the pioneering work on hydrogel photonic crystal sensor was first reported by Holtz and Asher [1]. They developed the polymerized crystalline colloidal array (PCCA) concept, which relies on highly charged polystyrene (PS) colloids to form a crystalline array within a cross-linkable aqueous medium. The sensing moiety, which induces volume change upon exposure to the analyte, is polymerized into the hydrogel [3]. Asher also used highly charged silica particles as the PCCA template, and compared the pH and ethanol sensing capability before and after etching silica particles [5]. In this study, the detailed sensing mechanism was investigated from thermodynamic standpoint, in which hydrogel swelling was well described by thermodynamic interaction between polymer/solvent and Donnan equilibrium of the bound charge as well. Lyon has reported on hydrogel microspheres made from N-isopropylacrylamide (NIPAM) and acrylic acid (AA), which form a temperature and pH responsive fluidic colloidal crystalline array in water [2]. Another novel approach is the inverse opal-type hydrogel photonic crystal sensor. Takeoka and coworkers demonstrated various inverse opal sensors fabricated from the silica colloidal crystal template. After polymerization of hydrogel within the interstitial space of the colloidal crystal, silica template was removed by hydrofluoric acid, and the remaining inverse opal hydrogel exerted a variety of sensing phenomena depending on the chemical equilibrium between the analyte and the bound receptor
184
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
[11,12]. Braun et al. reported on the fabrication of pH and glucose responsive inverse opal hydrogel sensors using a polystyrene based colloidal crystal template [4,7,13]. In spite of relative simplicity and versatility of hydrogel photonic crystal sensors, the slow response time due to a hindered diffusion of the analyte molecule through the hydrogel often limits their practical applications [14,15]. We previously showed that the use of inverse opal hydrogel structure under an optimized fabrication strategy can significantly improve the response time of a hydrogel based pH sensor [4]. Asher recently reported an improved response kinetics of glucose sensor [9]. Here, we apply a polymerization method for forming the hydrogel; photo-polymerization is advantageous in that it enables fine tuning of crosslinking density and void spaces within the inverse opal hydrogel which will consequently affect response time. In this report, we present a fabrication of inverse opal hydrogel pH sensor based on template photo-polymerization of pH-sensitive hydrogel with exceptionally rapid kinetics and long-term stability. Optimization of fabrication processes including photo-polymerization of hydrogel, and the response time upon repeated pH variation is rigorously investigated. 2. Materials and methods 2.1. Materials Three different Polystyrene (PS) microspheres (PS-220, PS-240, PS-260) with diameters of 220, 240, and 260 nm were used for preparation of the colloidal crystal template structures. PS-220 was purchased from Alfa Aesar, and PS-240 and PS-260 were synthesized by emulsion polymerization following a published procedure [16]. Styrene monomer (≥99%, Aldrich) was used after purification through activated aluminum oxide (basic, Acros). Potassium persulfate (99.99%, Aldrich) (initiator) and sodium dodecyl sulfate (>99.8%, Aldrich) (surfactant) were used as received. For the templated photo-polymerization of the pH-responsive hydrogel, 2-hydroxyethyl methacrylate (HEMA) (96%, Junsei), acrylic acid (AA) (99%, Junsei), and ethylene glycol dimethacrylate (EGDM) (98%, Aldrich) were used as monomers without further purification. Irgacure-651 (Ciba Specialty Chemicals) was used as the photoinitiator. 2.2. Syntheses of PS microspheres Emulsion polymerization of PS-240 was performed following a published procedure [16]. Briefly, 5 mg of sodium dodecyl sulfate (SDS) and 50 mg potassium persulfate were mixed with 15 mL of deionized (DI) water in a 25-mL round bottomed flask and degassed by nitrogen bubbling. The flask containing the mixture was placed in a water bath at 70 ◦ C, and further degassed for 30 min. 3 ml of styrene monomer was injected using syringe, and reaction was run with vigorous stirring for 4 h. The crude emulsion was filtered through a glass filter and kept in a semi-permeable Membrane (Cellu Sep T4, Membrane Filtration Products) which was soaked in 5 L of DI water for removal of unreacted monomer and surfactant. DI water was continuously replenished until resistivity reaches 12 M. Finally, the ∼10 wt% of PS emulsion was kept in a bottle, containing a small amount of ion exchange resin (AG501-X8, BioRad) added for maintaining the purity of water. The procedure for emulsion polymerization of PS-260 was the same except 3 mg of SDS was used. 2.3. Fabrication of colloidal crystal template Colloidal crystal templates were fabricated by modifying the flow cell method reported by Park et al. [17,18]. A standard
microscope slide glass was cut to 25 × 38 mm2 , and used as a bottom substrate. For a top substrate a hole (diameter = 3 mm) was made at the center of a slide glass using high speed mini bench drill (Seogwang T&M). All substrates were treated with Piranha solution (H2 SO4 :H2 O2 = 3:1 by volume), and rinsed with DI-water several times. The top substrate was made hydrophobic by soaking it in 1 mM trichlorooctadecylsilane (TCI) dissolved in 2,2,4-trimethylpentane (Jusei) for 30 min. Then a glass tube (length = 30 mm, diameter = 5 mm) was bonded to a hole of top substrate using epoxy resin (Hardex). At this stage, the whole top surface was evenly coated with the same epoxy resin to have ∼3 mm thickness. As a spacer, a square cut Mylar film (thickness: 25 m, Cheil Industry) having an open space of 10 × 15 mm2 area at the center was pre-cleaned with ethanol, and the channels were made using razor blade for discharging water during the colloidal crystallization. A Mylar spacer was placed between the bottom and the top substrate, and tightly clipped using six metal clips. The emulsion of PS microsphere was diluted to 0.4 wt%, and ∼0.5 mL aliquot was introduced through the glass tube on the top substrate which was subsequently pressurized by rubber bulb. The complete cell was placed in ultrasonicator (JAC Ultrasonic 2010, KODO) by 30◦ , and colloidal crystallization was performed for several days. The smaller microspheres required a longer crystallization period. Once the colloidal crystal was formed, the cell was air-dried, and subsequently annealed in a vacuum oven (LK-Lab) at 60 ◦ C for 4 h [17,18]. 2.4. Preparation of inverse opal hydrogel pH sensor 2.5 g 2-hydroxyethyl methacrylate (HEMA, Junsei), 35 mg Acrylic acid (AA, Junsei), 25 mg Ethylene glycol dimethacrylate (EGDM, Aldrich), 75 mg Irgacure-651 (Ciba specialty Chemicals) and, 0.625 g DI-water were mixed a vial, and ultra-sonicated for 10 min using a bath sonicator (SD-80H, Seong Dong). The polymerization mixture was infiltrated within the colloidal crystal template though the glass tube of a flow-cell, and air pressure was exerted to facilitate infiltration. After remaining liquid was removed, photopolymerization was performed using a high intensity UV-lamp (SB-100P/F, Spectronics Corporation) through a neutral density filter (Edmund optics) or a home-made epoxy filter (Hardex) for 2 h. Upon completion of photo-polymerization, the top substrate and spacer were removed from the flow-cell, and the templated hydrogel was dipped in chloroform (Duksan pure chemicals) for 24 h to remove PS colloidal template. Subsequent rinsing with chloroform and dipping in acetonitrile (ACN, Duksan pure chemicals) produced an iridescence which reflects a successful polymerization of the inverse opal hydrogel structure. The hydrogel was finally soaked in pH 1.5 phosphate buffer solution via DI water. A typical procedure for the preparation of hydrogel sensor is illustrated in Fig. 1. The phosphate buffer solutions of various pH were prepared by mixing different volumes of 0.1 M KH2 PO4 (aq.) (Duksan pure chemicals), 0.1 M HCl (aq.) (Samchun chemicals), 0.1 M NaOH (aq.) (Samchun chemicals). pH measurement of the hydrogel sensor was performed using a pH meter(SP-701, Suntex) at ambient temperature. 2.5. Characterizations Structural characterizations of the colloidal crystal and the resultant inverse opal hydrogel were performed using Hitachi S-4700 field-emission scanning electron microscope (SEM). The inverse opal samples were dried in vacuum oven (LK-Lab) at ambient temperature overnight, and subsequently freeze-fractured using liquid N2 . The reflectance spectra from the colloidal crystals were obtained using a fiber optic UV–vis spectrometer (AvaSpec, Avantes) coupled with a reflected light microscope (L2003A,
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
185
Fig. 1. Schematic illustration of experimental procedures.
Bimeince) through a objective lens (20×/0.30 NA). For calculating reflectance, a silver mirror (Edmund optics) was assumed to have a 100% reflectance. pH dependent reflectance measurements were performed on the hydrogel sensor placed in a glass petri dish filled with a buffer solution of pH of interest. The photograph of the hydrogel sensors were taken using a digital camera (IXUS 75, Canon) under illumination of white light from Fiber Optic halogen Illuminator (Model 190, Dolan-Jenner industries).
3. Results and discussion Upon completion of flow-induced colloidal crystallization, a colloidal crystal template using a monodisperse PS microsphere exhibited a bright color by selective Bragg reflection of visible light as shown in Fig. 2(a). As the method is known to form face centered cubic (FCC) structure with the [1 1 1] crystal plane normal to the substrate, the color is owing to the the interference of multiple reflections at every [1 1 1] layer. Formation of the FCC closed-packed structure was confirmed by SEM images of the colloidal crystal (Fig. 2(b)). A modified Bragg equation describes the relation between the wavelength of maximum reflection (reflection maxima, max ) and the particle diameter (dPS ) of the colloidal crystal assuming FCC structure. max =
8 1/2 d 3
m
(f1 n21 + f2 n22 − sin2 )
= 1.633dneff ;
1/2
=1.633d(f1 n21 + f2 n22 )
assuming 1st order diffraction at =0
1/2
(1)
where, d is the diameter of the PS colloidal particles, m is the order of the Bragg diffraction, n1 , n2 , f1 , f2 are the refractive index and filling factor of the sphere and interstice respectively, is the angle measured from the normal to the plane of colloidal crystal. For a FCC PS colloidal crystal, the material 1, 2 are PS and air, thus one can describe the equation using nPS = 1.5, nair = 1.0, fPS = 0.74, fair = 0.26 as following: max = 1.633dPS × 1.39
(2)
Since multiple experimental steps are involved in the preparation of inverse opal hydrogel pH sensor, the quality of final sensor strongly depended upon the optimization of experimental techniques. In the mixture of hydrogel precursor, three monomers have their respective roles. HEMA is a primary building block of the hydrogel, EGDM is a crosslinker which provides elasticity to the hydrogel, enhancing reproducibility of swelling and deswelling, and AA is the pH sensing moiety owing to its carboxyl pendant group. Once exposed to UV light, Irgacure 651, the photoinitiator in this study initiates the radical copolymerization of three monomers. Throughout rigorous optimization of mixing ratio for hydrogel precursor, we found that 19.2 mmol of HEMA, 0.48 mmol of AA, 0.29 mmol of Irgacure usually resulted in a mechanically robust, yet fast responding hydrogel sensor. Fig. 2(c) shows a typical inverse opal hydrogel in aqueous buffer exhibiting strong reflected color which implies the original FCC structure of PS colloidal crystal was well transformed to an inverse FCC structure by photo-polymerization of precursor mixture followed by subsequent removal of PS template. The inverse FCC structure was confirmed by SEM image of the freeze-fractured hydrogel as shown in Fig. 2(d). The change of structural color between Fig. 2(a) and (c) is attributed to a decreased neff due to inversion of polymer-filled region which results in a blue shift of the Bragg reflection. Along with the composition of the precursor, the polymerization time and exposure wavelength also affected the hydrogel formation. In general, a short polymerization time (3 h) resulted in frequent delamination of hydrogel, perhaps due to excessive crosslinking. 2 h of UV exposure appeared to be the optimum polymerization time for forming a mechanically stable hydrogel. We have also investigated the effect of the optical filtering to regulate the intensity of the exposure and its wavelength at the same time. The photoinitiator, Irgacure-651, is known to have the maximum absorbance at the wavelength close to that of the UV-lamp (365 nm) used in this study. The original and the attenuated spectra of the UV-lamp are shown in Fig. 3. However, a direct exposure of the intense UV-light for 2 h resulted in a delamination or disruption of the polymerized structure. Therefore, we tested
186
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
Fig. 2. Colloidal crystal template and hydrogel inverse opal. (a) As-prepared colloidal crystal template in the flow cell; (b) SEM image of colloidal crystal (PS-220); (c) Hydrogel inverse opal in pH 1.5 buffer; and (d) SEM image of a dried hydrogel inverse opal.
neutral density filters (NDF, FS-ND-REF, Edmund optics) with optical density (OD) of 0.5, 1.0, 2.0 and 3.0 respectively for reduction of the light intensity. In addition, a homemade optical filter using a 2-mm thick layer of epoxy resin (Hardex, absorption spectrum shown in Fig. 3(a)) coated to a slide glass was also tested. As shown in Fig. 3(b), the transmittance spectrum of the UV light through the epoxy filter shows a strong attenuation similar to that of the NDF3.0. The epoxy filter also exhibited the maximum transmittance at 389 nm which is 23 nm red-shifted from that of original spectrum
of UV lamp. The maximum intensity of incident light through epoxy filter was located at 389 nm since the light below that wavelength has been cut-off. Thus, homemade epoxy filter exhibits attenuation effect and wavelength shift simultaneously. The optical filters were tested for photo-polymerization by placing each of them 5 cm below UV lamp and 5 cm above precursor infiltrated colloidal template. Upon a 2 h exposure in the presence of the optical filter, the color of each inverse opal was monitored. The strongest iridescence was reproducibly observed from the inverse opal hydrogel when
Fig. 3. (a) Absorption spectrum of the epoxy filter used in this study; (b) Transmitted light intensity of the UV lamp for photo-polymerization through NDF-3.0 and epoxy filter.
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
187
Fig. 4. Reflectance spectra of as-prepared inverse opal pH sensors (blue curve) measured at pH 1.5 fabricated from the different sacrificial colloidal templates using (a) PS 220 nm, (b) PS 240 nm, and (c) PS 260 nm. For comparison, reflectance spectrum of the original colloidal crystal template before infiltration (red curve) is shown in each figure.
the epoxy filter or NDF-3.0 was used, while stronger light exposure did not result in an iridescent inverse opal. The poor iridescence upon direct (unfiltered) or NDF-attenuated light may be due to unhomogeneous polymerization through the thickness of the sample. The photoinitiator, Irgacure-651 absorbs much more strongly below 375 nm than above 375 nm [19]. Both direct and NDF attenuated light probably results in a gradient of polymerization rate (fast at the top, and slow lower down), while the epoxy filtered light, which is at a longer wavelength, probably results in a nearly homogeneously polymerized sample. We examined whether the intermediate solvent used between the chloroform colloid etch and the water-based measurements affects the final quality of HEMA-based hydrogel inverse opal structure. Ethanol and acetonitrile (ACN) were both tested as intermediate solvents since they are miscible with both chloroform and water. Although ethanol generates partially iridescent hydrogel, it tends to disrupt inverse opal structure by substantially swelling the HEMA gel. ACN was found to be the best intermediate solvent, and it results in samples which are highly iridescent in water, and have homogeneous color changes during the sensor operation over large area. Based on the optimized photo-polymerization condition as above, inverse opal hydrogels using PS microspheres of different sizes were fabricated. Fig. 4 shows the reflectance spectra (blue curve) of three inverse opal hydrogels in aqueous buffer at pH 1.5 obtained from the sacrificial colloidal crystals of PS 220 nm, PS 240 nm, and PS 260 nm respectively. Reflectance spectra of the
original PS colloidal crystals (red curve) are shown together for comparison with the corresponding spectra of hydrogel inverse opal. At pH 1.5, the hydrogel is in the collapsed state since all the carboxyl groups in the polymer are protonated. The stop bands of all three systems exhibited full width at half maximum (FWHM) as narrow as 40 nm respectively, producing strong structural colors (refer supplementary material). As described by the Bragg relationship (Eq. (1)), a larger unit lattice size (distance between pores) within the inverse opal structure resulted in more red-shifted stop band peak [4]. Fig. 5(a) shows a typical color change of the inverse opal pH sensor as a function of pH measured during either pH increase or decrease (between 1.26 and 6.99), and the corresponding shifts in max . As shown in Fig. 5, the clear color changes of the inverse opal hydrogel were observed over a large area with pH variation, and the color spectrum covered the visible wavelength between a pH of 1.26 and 6.99 as shown as solid curves in Fig. 5(a). As pH increased above 6.99, the red shift in max continued slightly, but the corresponding color change was not recognizable since red color is located at the long wavelength edge of the visible spectrum. Since most of AA units are deprotonated by a pH of 8, the pH sensitivity (dmax /dpH) in basic solution is small. The acrylate monomer containing polyprotic acids pendent group, which have broad ionization pH ranges, would be a good choice to widen pH sensing window of the photonic crystal hydrogel sensor. The peak shifts during pH decrease are plotted (dashed curves in Fig. 5(a))
Fig. 5. (a) pH-dependent color change and respective reflectance spectra of hydrogel pH sensor fabricated from PS-220 measured between pH of 1.26 and 6.99. A series of the spectra are obtained during the pH increase (solid curves), while the other series (dashed curves) are those during the pH decrease. (b) Plot of max changes upon pH increase and decrease showing a weak hystersis.
188
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
Fig. 6. pH-dependent change of reflectance maxima (max ) under various buffers of (a) KH2 PO4 (filled square), K(CH3 COO) (filled circle), K2 CO3 (filled triangle) for screening anion effect and (b) KH2 PO4 (open square), Na2 HPO4 (open triangle) for cation effect respectively. The sensor was fabricated from PS-240. The initial concentration of every buffer is 0.1 M, which is subsequently mixed with aq. HCl or NaOH for pH adjustment.
along with those during pH increase, which shows a slight hysteresis (max vs. pH shown in Fig. 5(b)) as previously reported [4,5]. Recently, Asher and coworkers extensively investigated the swelling mechanism of polyHEMA-based hydrogel sensor, in which they attributed the major reason for the hysteresis to the low ionic strength to fully re-equilibriate the Donnan potential during pH decrease. The slow diffusion of ions through the hydrogel was also accounted in their report. Later in the current report, the detailed pH response kinetics of our hydrogel sensor will be discussed. Although it is assumed that the hydrogel sensor selectively responds to the pH variation, there exist complex ion equilibria in the mixed buffer system which could affect Donnan equilibrium exerted by proton. Thus, possible matrix effects need to be screened to assure the precision of measurement. For most of the sensing experiments in this study, 0.1 M potassium phosphate (KH2 PO4 ) buffer mixed with HCl and/or NaOH was used since its buffering pH range decently covers the most sensitive pH range of our hydrogel system (around pH 5–6) Nonetheless, other buffer systems of weak acids and conjugate bases can also be used instead of phosphate. In Fig. 6(a) the dependences of the sensor response (max vs. pH increase) on different anions in buffer such as potassium acetate (K(CH3 COO)) and potassium carbonate (K2 CO3 ) as well as potassium phosphate are shown. Likewise, the cation effects on sensor responses were also tested using sodium phosphate instead of potassium phosphate as shown in Fig. 6(b). All those buffers were prepared at the concentration of 0.1 M. Overall, the variation of buffer systems did not cause significant changes in max at pH ranges from 2 to 8 as shown in Fig. 6(a) and (b), although carbonate buffer showed a slight red-shift effect which is indicated by filled triangle in Fig. 6(a). From Fig. 6, it is obviously shown that the most sensitive pH range to change hydrogel volume is pH
5–6 region regardless of buffers. At pH below 4 or above 7, our sensor exhibited poor sensitivity. Although not shown, the buffer concentration effect was also tested in this study (supplementary materials). When highly concentrated buffer solution (∼1 M) was used, a common “salting-out” effect was observed which shrinks the hydrogel consequently blue-shifting max , while those below 0.1 M did not show significant effect on max . A significantly distinguishing feature of the inverse opal pH sensor fabricated in our study is a fast response time upon pH variation. The response kinetics are demonstrated in Fig. 7. pH variation experiments indicated that the highest sensitivity (max /pH) of our hydrogel sensor was observed over a pH range of 5–6 which is evidently shown in Fig. 6. Although the pKa (where Ka is acid dissociation constant) of AA is 4.25 at 25 ◦ C, the actual pKa of AA within copolymeric hydrogel is reported to be slightly higher due to a electrostatic interaction of adjacent AA units [8]. Thus we measured the response time during the pH increase (5 → 6) and pH decrease (6 → 5) as well. For the pH increase experiment, the sensor in the pH 5 buffer was transferred to a petri dish containing fresh pH 6 buffer. The pH decrease experiment was done in the same manner by transferring the sensor to a fresh pH 5 buffer. The time dependent max change for pH increase is shown in Fig. 7(a). It is evident that max rapidly increases without the diffusion-limited slow max shift at the initial stage of pH change as reported previously [4,5]. In order to define the sensor response time, max vs. time plot during pH variation was fit to a single exponential function (y = max,t=0 + A exp(−t/)) with a r2 = 0.974 (Fig. 7(a)). The sensor response time was defined to be of the exponential function. When the initial pH 5 was changed to pH 6, the response time was measured to be 12.1 s, more than 30 times faster than a similar hydrogel sensor reported earlier [4]. In
Fig. 7. (a) Response kinetics of pH sensor made from PS-240 colloidal crystal template during pH increase from 5 to 6. max vs. time is fitted by a single exponential function with = 12.1 s as drawn by solid curve. (b) Response kinetics during pH decrease from 6 to 5. max vs. time is fitted by a double exponential function with 1 = 8.1 and 2 = 208 s as drawn by solid curve. (c) Response kinetics with respect to different range of pH variation.
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
addition to the pH increase experiment, the response time for the pH decrease (6–5) was also monitored. In this case, single exponential function did not accurately fit the experimental results. A closer look at Fig. 7(b) reveals that after a rapid blue shift of max from 625 to 580 nm within 20 s, a slow restoration of max back to 540 nm takes place. A double exponential fit was performed, and the function (y = max,t=0 + A1 exp(−t/ 1 ) + A2 exp(−t/ 2 )) with 1 = 8.1 s and 2 = 208 s resulted in a good fit (r2 = 0.995) (Fig. 7(b)), implying there might be two distinct deswelling mechanisms (the fitted values of other parameters are max,t=0 = 553, A1 = 40, A2 = 31 respectively). The initial hydrogel thickness is 25 m, and there is a ∼10% longitudinal swelling at a pH of 5 (refer to Fig. 5(b)). Assuming proton diffusion is the same in the inverse opal structure as in an aqueous medium, and that ion exchange at the carboxylic acid group is a diffusion-limited process, = l2 /DH+ ∼ = 7.6 s (l = 27.5 m, DH+ = 8.0 × 10−5 cm2 /s) which coincides very well with experimentally obtained 1 (=8.1 s). However, proton diffusion through the ionic hydrogel network should be retarded compared to that through the free water [15]. For a diffusion limited process, Tanaka successfully formulated the time required for ion (proton) diffusion through the ionic hydrogel using the hydrogel volume before and after swelling (V/V0 ) as shown in the following relationship, which can be further simplified using the ratio of diffraction maxima (/0 ∼ 625/580 = 1.08) assuming that the swelling occurs only along the thickness direction [4,5]. =
(V/V0 )
2/3
l0 1 + ([AA]/Ka ) Dion
∼ =
2/3 l0 1 + [AA]/Ka 0
Dion
l0 [AA], Ka , and Dion respectively denote hydrogel thickness (25 m), the concentration acrylic acid (0.16 M) within the hydrogel, acid dissociation constant (5.62 × 10−5 M), and the diffusion constant of proton (8.0 × 10−5 cm2 /s). The response time assuming our system is a solid hydrogel was calculated to be 219 s, which appears in accordance with the measured 2 (207 s). Since our inverse opal hydrogel primarily consists of void spaces interconnected by holes (Fig. 8), when the voids are connected, the diffusion of proton can take place both through the void space or through the hydrogel network. When the sensor is in its initial form (at low pH), the inverse opal will contain ∼26% (or less) of hydrogel and the remainder of volume will be void space (free water) as shown in Fig. 8. Since the void space is dominant and highly interconnected, the most of ionic species are expected to diffuse rapidly through the voids, and then into the hydrogel, resulting in an aqueous diffusion limited response time (∼8 s). Once the hydrogel is swollen at high pH the hydrogel will occupy much of the void space as shown as the right figure in Fig. 8 (the hydrogel is pinned to the substrate, and thus lateral swelling is not allowed), which has been experimentally confirmed in our previous investigation [13]. Since the proton is now diffusing through a swollen gel, diffusion is much slower, and such hindered diffusion will be reflected by 2 , which is convoluted with the fast diffusion ( 1 ) since some void spaces still exist in the hydrogel [13]. The two exponential decay function as shown in Fig. 7(b) may be attributed to such a dual diffusion mechanism. By applying lower pH buffer than 5 during the pH decrease, max can be further blue-shifted to exhibit more explicit color change of the hydrogel. Larger max changes upon larger pH gradients are clearly shown in Fig. 7(c). A dual diffusion mechanism appears to explain the observed deswelling kinetics quite well, however in an attempt to optimize the response, it is necessary to compare the hydrogel sensors with different crosslinking densities. We compared the response time of two pH sensors in which only the exposure condition for the photo-polymerization is different. As previously discussed, photo-polymerization without an epoxy filter resulted in a high crosslinking density and often delamination of the hydrogel from
189
Fig. 8. Schematic representation of a [111] layer of FCC inverse opal hydrogel at low and high pH respectively. The hydrogel holds the as-prepared inverse opal structure at low pH, and the voids become filled with swollen hydrogel at an increased pH assuming that the lateral swelling is minimal.
the substrate. However, the inverse opal sensor could be obtained by short exposure (20 min) of unfiltered (strong) UV light. Fig. 9 compares the pH sensing kinetics for the sensors obtained either by normal exposure condition (2 h photo-polymerization using epoxy filter) or strong exposure condition (20 min without epoxy filter). It is obvious that the hydrogel obtained by strong UV exposure is less swollen in the same pH buffer exhibiting lower max due to a high crosslinking density. Since the same sacrificial PS particles were used for both cases, the crosslinking density is the only reason for the apparent max differences at the same pH buffer. More importantly, a longer pH response time ( 1 ) was observed for highly crosslinked hydrogel sensor. A single exponential function could fit both max changes fairly well, giving 1 for an optimized sensor and a highly crosslinked sensor to be 12.8 and 101 s respectively. 101.0 s is still fast compared to our previous report, in which the response time for pH change between 5 and 6 was as slow as 1200 s [4]. The slower response of the prior system could be attributed to a higher crosslinking density, smaller interconnecting holes, larger hydrogel content, and higher AA concentration than our sensor (5 vs. 2.5%). Our sensor has a different physical structures compared with that reported by Asher and coworkers due to the distinctly different fabrication strategies. Their sensor had a much higher hydrogel volume fraction, and the thickness is thicker, the calculated and the observed response time were in the order of several hours [5]. Thus, we concluded that the optimization of the photopolymerization conditions such as exposure time, intensity, and wavelength of UV-lamp are crucial in order to achieve fast response
Fig. 9. Response kinetics obtained from two different hydrogel sensors. Two sensors are fabricated using the same monomer mixture and the PS template (PS240), only differing in the UV exposure conditions for photo-polymerization. The upper response kinetics is obtained from an optimized exposure condition (using epoxy filter), and the lower kinetics is from a strong UV exposure without optical filter. Both data are fitted by single exponential fit to give 1 of 11.8 and 101 s respectively.
190
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
Fig. 10. Repeatability and reproducibility of pH sensing experiment on a sensor made from PS-240. (a) Repeated pH-dependent reflectance maxima shifts, (b) Repeated response time of PS-240 sensor measured over 3 months.
chemical sensor, since they determine the crosslinking density of hydrogel and the size of holes between voids, the most important control factors for the response time. In order to demonstrate repeatability and reproducibility of the hydrogel sensor more rigorously, fifty pH variations between 5 and 6 were performed over 3 months. Fig. 10(a) shows sensing kinetics measurement plots for randomly selected data, and Fig. 10(b) is the plot of response time vs. number of experiments. The sensor was still highly responsive 5 months after the fabrication without visible degradation of iridescence or mechanical robustness. We expect actual lifetime of the sensor to be longer than 1 year unless it is completely dried. 4. Conclusions Through a rigorous optimization of fabrication procedures including the monomer ratio, polymerization time, optical filtering of UV light, and colloid, a fast responding hydrogel inverse opal pH sensor was successfully formed. The pH sensor exhibited response time of ∼12 s for pH variations between 5 and 6, which was ∼100 times improved compared to the previously reported inverse opal pH sensors [4]. In addition to improved response time, sensor lifetime was found to be longer than 5 month without loss of reproducibility or iridescent color. We expect that the distinguished features of the hydrogel sensor developed in this study such as fast response, and mechanical robustness will find promise a variety of chemical sensor applications. In particular, fast response enables a real time sensing of chemical species in continuous processes such as chromatography. By incorporation of novel vinyl monomers containing various sensing moieties in the pendent group, investigations on a various chemical sensors are underway. Acknowledgments This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2008-331-D00153). PVB thanks the US DOE ‘Light-Material Interactions in Energy Conversion’ Energy Frontier Research Center under grant DE-SC0001293 for support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.snb.2010.07.018.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
J.H. Holtz, S.A. Asher, Nature 389 (1997) 829–832. J.D. Debord, L.A. Lyon, J. Phys. Chem. B 104 (2000) 6327–6331. K. Lee, S.A. Asher, J. Am. Chem. Soc. 122 (2000) 9534–9537. Y.J. Lee, P.V. Braun, Adv. Mater. 15 (2003) 563–566. X. Xu, A.V. Goponenko, S.A. Asher, J. Am. Chem. Soc. 130 (2008) 3113. A.V. Goponenko, S.A. Asher, J. Am. Chem. Soc. 127 (2005) 10753. Y.J. Lee, S.A. Pruzinsky, P.V. Braun, Langmuir 20 (2004) 3096–3106. M. Men-Moshe, V. Alexeev, S.A. Asher, Anal. Chem. 78 (2006) 5149–5157. M.M.W. Muscatello, L.E. Stunja, S.A. Asher, Anal. Chem. 81 (2009) 4978. T. Endo, Y. Yanagida, T. Hatsuzawa, Sens. Actuators B 125 (2007) 589– 595. Y. Takeoka, M. Watanade, Langmuir 18 (2002) 5977–5980. M. Honda, T. Seki, Y. Takeoka, Adv. Mater. 21 (2009) 1801–1804. Y.J. Lee, C.E. Heitzman, W.R. Frei, H.T. Johnson, P.V. Braun, J. Phys. Chem. B 110 (2006) 19300–19306. J. Wang, Y. Cao, Y. Feng, F. Yin, J. Gao, Adv. Mater. 19 (2007) 3865–3871. S. Mafe, J.A. Manzanares, A.E. English, T. Tanaka, Phys. Rev. Lett. 79 (1997) 3086. M. Egen, R. Zentel, Chem. Mater. 14 (2002) 2176–2183. S.H. Park, B. Gates, Y.N. Xia, Adv. Mater. 11 (1999) 462–466. B. Gates, D. Qin, Y.N. Xia, Adv. Mater. 11 (1999) 466–469. Data Sheet for “Irgacure-651” provided by Ciba Specialty Chemicals Inc., 2001.
Biographies Jinsub Shin received his BS degree from Sejong University, Korea in 2009. He is currently a graduate student in Sejong University under an advice of Prof. Wonmok Lee at the Department of Chemistry. He is working on the fabrication of photonic crystal devices such as hydrogel sensors, and micro-lasers based on colloidal selfassembly techniques. Paul V. Braun is a Professor of Materials Science and Engineering, and an affiliate of the Frederick Seitz Materials Research Laboratory, the Beckman Institute for Advanced Science and Technology, and the Department of Chemistry at the University of Illinois at Urbana-Champaign. Prof. Braun’s research focuses on the synthesis and properties of 3D architectures with a focus on materials with unique optical, electrochemical, thermal, and mechanical properties. Prof. Braun received his BS degree with distinction from Cornell University in 1993, and his PhD in Materials Science and Engineering from Illinois in 1998, both in Materials Science and Engineering. Following a postdoctoral appointment at Bell Labs, Lucent Technologies, he joined the faculty at Illinois in 1999. Wonmok Lee is an Assistant Professor of the Department of Chemistry at Sejong University. He received his BS degree in 1995 from Pohang University of Science and Technology, MS degree in 1997 and PhD degree in 2001 from the same university. After postdoctoral periods in University of Illinois at Urbana-Champaign (UIUC) and Massachusetts Institute of Technology (MIT), Prof. Lee developed an industrial career at Samsung Advanced Institute of Technology (SAIT). He joined the faculty at Sejong University since 2007. Prof. Lee’s research area covers the polymer electrolyte membranes, photonic bandgap materials, sensors, and other optical devices using self-assembled polymers.
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Fast response photonic crystal pH sensor based on templated photo-polymerized hydrogel inverse opal Jinsub Shin a , Paul V. Braun b,∗∗ , Wonmok Lee a,∗ a
Department of Chemistry, Sejong University, 98 Gunja-Dong, Gwngjin-gu, Seoul 143-747, Republic of Korea Department of Materials Science and Engineering, Materials Research Laboratory, Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA b
a r t i c l e
i n f o
Article history: Received 10 March 2010 Received in revised form 17 June 2010 Accepted 10 July 2010 Available online 16 July 2010 Keywords: Fast response Photonic crystal pH sensor Hydrogel Inverse opal
a b s t r a c t Polymer hydrogels can exhibit large reversible volume changes in response to external stimuli, and thus are regarded as excellent materials for chemical sensors. In this report, we demonstrate a mechanically robust and fast response photonic crystal pH sensor fabricated by templated photo-polymerization of hydrogel monomers within the interstitial space of a self-assembled colloidal photonic crystal. Throughout a rigorous optimization of the photo-polymerization, pH sensors showing a response time of less than 10 s upon a pH change were fabricated. Repeated pH changes revealed that the sensor has a long lifetime (>6 months) without degradation of the response time or reproducibility in pH-driven color change. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Stimuli-responsive chemical sensors are of great importance for environmental, food, and biological applications. For sensing chemical stimuli, complicated electrical or optical detection strategies are often used which require sophisticated and expensive detection systems. Recently developed photonic crystal hydrogel sensor strategies which rely on the diffraction induced color change of the sensor are now however enabling the detection of the analyte molecules simply with the naked eye [1]. A common detection mechanism is based on a volume change of hydrogel exerted by Donnan potential between the analyte ions and their receptor which is covalently bound to the hydrogel building block. In order to compensate the local ion inhomogeneity, there is influx of water molecules to the hydrogel. The optical readout of such hydrogel structures is because they are imprinted with a periodic structure, forming a so-called photonic crystal, and the volume change in the periodic structure results in the color change of the sensor. Such color is often called ‘structural color’ due to the origin of color. Important advantages of the hydrogel photonic crystal sensor are that relatively simple optical detection techniques are
∗ Corresponding author. Tel.: +82 2 3408 3212; fax: +82 2 462 9954. ∗∗ Corresponding author. E-mail addresses: [email protected] (P.V. Braun), [email protected] (W. Lee). 0925-4005/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2010.07.018
possible, and that a wide variety of analytes including protons [2–5], metal ions [6], glucose [7–9], and volatile organic compounds (VOC) [10], to name a few have been successfully detected using such sensors. Much of the pioneering work on hydrogel photonic crystal sensor was first reported by Holtz and Asher [1]. They developed the polymerized crystalline colloidal array (PCCA) concept, which relies on highly charged polystyrene (PS) colloids to form a crystalline array within a cross-linkable aqueous medium. The sensing moiety, which induces volume change upon exposure to the analyte, is polymerized into the hydrogel [3]. Asher also used highly charged silica particles as the PCCA template, and compared the pH and ethanol sensing capability before and after etching silica particles [5]. In this study, the detailed sensing mechanism was investigated from thermodynamic standpoint, in which hydrogel swelling was well described by thermodynamic interaction between polymer/solvent and Donnan equilibrium of the bound charge as well. Lyon has reported on hydrogel microspheres made from N-isopropylacrylamide (NIPAM) and acrylic acid (AA), which form a temperature and pH responsive fluidic colloidal crystalline array in water [2]. Another novel approach is the inverse opal-type hydrogel photonic crystal sensor. Takeoka and coworkers demonstrated various inverse opal sensors fabricated from the silica colloidal crystal template. After polymerization of hydrogel within the interstitial space of the colloidal crystal, silica template was removed by hydrofluoric acid, and the remaining inverse opal hydrogel exerted a variety of sensing phenomena depending on the chemical equilibrium between the analyte and the bound receptor
184
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
[11,12]. Braun et al. reported on the fabrication of pH and glucose responsive inverse opal hydrogel sensors using a polystyrene based colloidal crystal template [4,7,13]. In spite of relative simplicity and versatility of hydrogel photonic crystal sensors, the slow response time due to a hindered diffusion of the analyte molecule through the hydrogel often limits their practical applications [14,15]. We previously showed that the use of inverse opal hydrogel structure under an optimized fabrication strategy can significantly improve the response time of a hydrogel based pH sensor [4]. Asher recently reported an improved response kinetics of glucose sensor [9]. Here, we apply a polymerization method for forming the hydrogel; photo-polymerization is advantageous in that it enables fine tuning of crosslinking density and void spaces within the inverse opal hydrogel which will consequently affect response time. In this report, we present a fabrication of inverse opal hydrogel pH sensor based on template photo-polymerization of pH-sensitive hydrogel with exceptionally rapid kinetics and long-term stability. Optimization of fabrication processes including photo-polymerization of hydrogel, and the response time upon repeated pH variation is rigorously investigated. 2. Materials and methods 2.1. Materials Three different Polystyrene (PS) microspheres (PS-220, PS-240, PS-260) with diameters of 220, 240, and 260 nm were used for preparation of the colloidal crystal template structures. PS-220 was purchased from Alfa Aesar, and PS-240 and PS-260 were synthesized by emulsion polymerization following a published procedure [16]. Styrene monomer (≥99%, Aldrich) was used after purification through activated aluminum oxide (basic, Acros). Potassium persulfate (99.99%, Aldrich) (initiator) and sodium dodecyl sulfate (>99.8%, Aldrich) (surfactant) were used as received. For the templated photo-polymerization of the pH-responsive hydrogel, 2-hydroxyethyl methacrylate (HEMA) (96%, Junsei), acrylic acid (AA) (99%, Junsei), and ethylene glycol dimethacrylate (EGDM) (98%, Aldrich) were used as monomers without further purification. Irgacure-651 (Ciba Specialty Chemicals) was used as the photoinitiator. 2.2. Syntheses of PS microspheres Emulsion polymerization of PS-240 was performed following a published procedure [16]. Briefly, 5 mg of sodium dodecyl sulfate (SDS) and 50 mg potassium persulfate were mixed with 15 mL of deionized (DI) water in a 25-mL round bottomed flask and degassed by nitrogen bubbling. The flask containing the mixture was placed in a water bath at 70 ◦ C, and further degassed for 30 min. 3 ml of styrene monomer was injected using syringe, and reaction was run with vigorous stirring for 4 h. The crude emulsion was filtered through a glass filter and kept in a semi-permeable Membrane (Cellu Sep T4, Membrane Filtration Products) which was soaked in 5 L of DI water for removal of unreacted monomer and surfactant. DI water was continuously replenished until resistivity reaches 12 M. Finally, the ∼10 wt% of PS emulsion was kept in a bottle, containing a small amount of ion exchange resin (AG501-X8, BioRad) added for maintaining the purity of water. The procedure for emulsion polymerization of PS-260 was the same except 3 mg of SDS was used. 2.3. Fabrication of colloidal crystal template Colloidal crystal templates were fabricated by modifying the flow cell method reported by Park et al. [17,18]. A standard
microscope slide glass was cut to 25 × 38 mm2 , and used as a bottom substrate. For a top substrate a hole (diameter = 3 mm) was made at the center of a slide glass using high speed mini bench drill (Seogwang T&M). All substrates were treated with Piranha solution (H2 SO4 :H2 O2 = 3:1 by volume), and rinsed with DI-water several times. The top substrate was made hydrophobic by soaking it in 1 mM trichlorooctadecylsilane (TCI) dissolved in 2,2,4-trimethylpentane (Jusei) for 30 min. Then a glass tube (length = 30 mm, diameter = 5 mm) was bonded to a hole of top substrate using epoxy resin (Hardex). At this stage, the whole top surface was evenly coated with the same epoxy resin to have ∼3 mm thickness. As a spacer, a square cut Mylar film (thickness: 25 m, Cheil Industry) having an open space of 10 × 15 mm2 area at the center was pre-cleaned with ethanol, and the channels were made using razor blade for discharging water during the colloidal crystallization. A Mylar spacer was placed between the bottom and the top substrate, and tightly clipped using six metal clips. The emulsion of PS microsphere was diluted to 0.4 wt%, and ∼0.5 mL aliquot was introduced through the glass tube on the top substrate which was subsequently pressurized by rubber bulb. The complete cell was placed in ultrasonicator (JAC Ultrasonic 2010, KODO) by 30◦ , and colloidal crystallization was performed for several days. The smaller microspheres required a longer crystallization period. Once the colloidal crystal was formed, the cell was air-dried, and subsequently annealed in a vacuum oven (LK-Lab) at 60 ◦ C for 4 h [17,18]. 2.4. Preparation of inverse opal hydrogel pH sensor 2.5 g 2-hydroxyethyl methacrylate (HEMA, Junsei), 35 mg Acrylic acid (AA, Junsei), 25 mg Ethylene glycol dimethacrylate (EGDM, Aldrich), 75 mg Irgacure-651 (Ciba specialty Chemicals) and, 0.625 g DI-water were mixed a vial, and ultra-sonicated for 10 min using a bath sonicator (SD-80H, Seong Dong). The polymerization mixture was infiltrated within the colloidal crystal template though the glass tube of a flow-cell, and air pressure was exerted to facilitate infiltration. After remaining liquid was removed, photopolymerization was performed using a high intensity UV-lamp (SB-100P/F, Spectronics Corporation) through a neutral density filter (Edmund optics) or a home-made epoxy filter (Hardex) for 2 h. Upon completion of photo-polymerization, the top substrate and spacer were removed from the flow-cell, and the templated hydrogel was dipped in chloroform (Duksan pure chemicals) for 24 h to remove PS colloidal template. Subsequent rinsing with chloroform and dipping in acetonitrile (ACN, Duksan pure chemicals) produced an iridescence which reflects a successful polymerization of the inverse opal hydrogel structure. The hydrogel was finally soaked in pH 1.5 phosphate buffer solution via DI water. A typical procedure for the preparation of hydrogel sensor is illustrated in Fig. 1. The phosphate buffer solutions of various pH were prepared by mixing different volumes of 0.1 M KH2 PO4 (aq.) (Duksan pure chemicals), 0.1 M HCl (aq.) (Samchun chemicals), 0.1 M NaOH (aq.) (Samchun chemicals). pH measurement of the hydrogel sensor was performed using a pH meter(SP-701, Suntex) at ambient temperature. 2.5. Characterizations Structural characterizations of the colloidal crystal and the resultant inverse opal hydrogel were performed using Hitachi S-4700 field-emission scanning electron microscope (SEM). The inverse opal samples were dried in vacuum oven (LK-Lab) at ambient temperature overnight, and subsequently freeze-fractured using liquid N2 . The reflectance spectra from the colloidal crystals were obtained using a fiber optic UV–vis spectrometer (AvaSpec, Avantes) coupled with a reflected light microscope (L2003A,
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
185
Fig. 1. Schematic illustration of experimental procedures.
Bimeince) through a objective lens (20×/0.30 NA). For calculating reflectance, a silver mirror (Edmund optics) was assumed to have a 100% reflectance. pH dependent reflectance measurements were performed on the hydrogel sensor placed in a glass petri dish filled with a buffer solution of pH of interest. The photograph of the hydrogel sensors were taken using a digital camera (IXUS 75, Canon) under illumination of white light from Fiber Optic halogen Illuminator (Model 190, Dolan-Jenner industries).
3. Results and discussion Upon completion of flow-induced colloidal crystallization, a colloidal crystal template using a monodisperse PS microsphere exhibited a bright color by selective Bragg reflection of visible light as shown in Fig. 2(a). As the method is known to form face centered cubic (FCC) structure with the [1 1 1] crystal plane normal to the substrate, the color is owing to the the interference of multiple reflections at every [1 1 1] layer. Formation of the FCC closed-packed structure was confirmed by SEM images of the colloidal crystal (Fig. 2(b)). A modified Bragg equation describes the relation between the wavelength of maximum reflection (reflection maxima, max ) and the particle diameter (dPS ) of the colloidal crystal assuming FCC structure. max =
8 1/2 d 3
m
(f1 n21 + f2 n22 − sin2 )
= 1.633dneff ;
1/2
=1.633d(f1 n21 + f2 n22 )
assuming 1st order diffraction at =0
1/2
(1)
where, d is the diameter of the PS colloidal particles, m is the order of the Bragg diffraction, n1 , n2 , f1 , f2 are the refractive index and filling factor of the sphere and interstice respectively, is the angle measured from the normal to the plane of colloidal crystal. For a FCC PS colloidal crystal, the material 1, 2 are PS and air, thus one can describe the equation using nPS = 1.5, nair = 1.0, fPS = 0.74, fair = 0.26 as following: max = 1.633dPS × 1.39
(2)
Since multiple experimental steps are involved in the preparation of inverse opal hydrogel pH sensor, the quality of final sensor strongly depended upon the optimization of experimental techniques. In the mixture of hydrogel precursor, three monomers have their respective roles. HEMA is a primary building block of the hydrogel, EGDM is a crosslinker which provides elasticity to the hydrogel, enhancing reproducibility of swelling and deswelling, and AA is the pH sensing moiety owing to its carboxyl pendant group. Once exposed to UV light, Irgacure 651, the photoinitiator in this study initiates the radical copolymerization of three monomers. Throughout rigorous optimization of mixing ratio for hydrogel precursor, we found that 19.2 mmol of HEMA, 0.48 mmol of AA, 0.29 mmol of Irgacure usually resulted in a mechanically robust, yet fast responding hydrogel sensor. Fig. 2(c) shows a typical inverse opal hydrogel in aqueous buffer exhibiting strong reflected color which implies the original FCC structure of PS colloidal crystal was well transformed to an inverse FCC structure by photo-polymerization of precursor mixture followed by subsequent removal of PS template. The inverse FCC structure was confirmed by SEM image of the freeze-fractured hydrogel as shown in Fig. 2(d). The change of structural color between Fig. 2(a) and (c) is attributed to a decreased neff due to inversion of polymer-filled region which results in a blue shift of the Bragg reflection. Along with the composition of the precursor, the polymerization time and exposure wavelength also affected the hydrogel formation. In general, a short polymerization time (3 h) resulted in frequent delamination of hydrogel, perhaps due to excessive crosslinking. 2 h of UV exposure appeared to be the optimum polymerization time for forming a mechanically stable hydrogel. We have also investigated the effect of the optical filtering to regulate the intensity of the exposure and its wavelength at the same time. The photoinitiator, Irgacure-651, is known to have the maximum absorbance at the wavelength close to that of the UV-lamp (365 nm) used in this study. The original and the attenuated spectra of the UV-lamp are shown in Fig. 3. However, a direct exposure of the intense UV-light for 2 h resulted in a delamination or disruption of the polymerized structure. Therefore, we tested
186
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
Fig. 2. Colloidal crystal template and hydrogel inverse opal. (a) As-prepared colloidal crystal template in the flow cell; (b) SEM image of colloidal crystal (PS-220); (c) Hydrogel inverse opal in pH 1.5 buffer; and (d) SEM image of a dried hydrogel inverse opal.
neutral density filters (NDF, FS-ND-REF, Edmund optics) with optical density (OD) of 0.5, 1.0, 2.0 and 3.0 respectively for reduction of the light intensity. In addition, a homemade optical filter using a 2-mm thick layer of epoxy resin (Hardex, absorption spectrum shown in Fig. 3(a)) coated to a slide glass was also tested. As shown in Fig. 3(b), the transmittance spectrum of the UV light through the epoxy filter shows a strong attenuation similar to that of the NDF3.0. The epoxy filter also exhibited the maximum transmittance at 389 nm which is 23 nm red-shifted from that of original spectrum
of UV lamp. The maximum intensity of incident light through epoxy filter was located at 389 nm since the light below that wavelength has been cut-off. Thus, homemade epoxy filter exhibits attenuation effect and wavelength shift simultaneously. The optical filters were tested for photo-polymerization by placing each of them 5 cm below UV lamp and 5 cm above precursor infiltrated colloidal template. Upon a 2 h exposure in the presence of the optical filter, the color of each inverse opal was monitored. The strongest iridescence was reproducibly observed from the inverse opal hydrogel when
Fig. 3. (a) Absorption spectrum of the epoxy filter used in this study; (b) Transmitted light intensity of the UV lamp for photo-polymerization through NDF-3.0 and epoxy filter.
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
187
Fig. 4. Reflectance spectra of as-prepared inverse opal pH sensors (blue curve) measured at pH 1.5 fabricated from the different sacrificial colloidal templates using (a) PS 220 nm, (b) PS 240 nm, and (c) PS 260 nm. For comparison, reflectance spectrum of the original colloidal crystal template before infiltration (red curve) is shown in each figure.
the epoxy filter or NDF-3.0 was used, while stronger light exposure did not result in an iridescent inverse opal. The poor iridescence upon direct (unfiltered) or NDF-attenuated light may be due to unhomogeneous polymerization through the thickness of the sample. The photoinitiator, Irgacure-651 absorbs much more strongly below 375 nm than above 375 nm [19]. Both direct and NDF attenuated light probably results in a gradient of polymerization rate (fast at the top, and slow lower down), while the epoxy filtered light, which is at a longer wavelength, probably results in a nearly homogeneously polymerized sample. We examined whether the intermediate solvent used between the chloroform colloid etch and the water-based measurements affects the final quality of HEMA-based hydrogel inverse opal structure. Ethanol and acetonitrile (ACN) were both tested as intermediate solvents since they are miscible with both chloroform and water. Although ethanol generates partially iridescent hydrogel, it tends to disrupt inverse opal structure by substantially swelling the HEMA gel. ACN was found to be the best intermediate solvent, and it results in samples which are highly iridescent in water, and have homogeneous color changes during the sensor operation over large area. Based on the optimized photo-polymerization condition as above, inverse opal hydrogels using PS microspheres of different sizes were fabricated. Fig. 4 shows the reflectance spectra (blue curve) of three inverse opal hydrogels in aqueous buffer at pH 1.5 obtained from the sacrificial colloidal crystals of PS 220 nm, PS 240 nm, and PS 260 nm respectively. Reflectance spectra of the
original PS colloidal crystals (red curve) are shown together for comparison with the corresponding spectra of hydrogel inverse opal. At pH 1.5, the hydrogel is in the collapsed state since all the carboxyl groups in the polymer are protonated. The stop bands of all three systems exhibited full width at half maximum (FWHM) as narrow as 40 nm respectively, producing strong structural colors (refer supplementary material). As described by the Bragg relationship (Eq. (1)), a larger unit lattice size (distance between pores) within the inverse opal structure resulted in more red-shifted stop band peak [4]. Fig. 5(a) shows a typical color change of the inverse opal pH sensor as a function of pH measured during either pH increase or decrease (between 1.26 and 6.99), and the corresponding shifts in max . As shown in Fig. 5, the clear color changes of the inverse opal hydrogel were observed over a large area with pH variation, and the color spectrum covered the visible wavelength between a pH of 1.26 and 6.99 as shown as solid curves in Fig. 5(a). As pH increased above 6.99, the red shift in max continued slightly, but the corresponding color change was not recognizable since red color is located at the long wavelength edge of the visible spectrum. Since most of AA units are deprotonated by a pH of 8, the pH sensitivity (dmax /dpH) in basic solution is small. The acrylate monomer containing polyprotic acids pendent group, which have broad ionization pH ranges, would be a good choice to widen pH sensing window of the photonic crystal hydrogel sensor. The peak shifts during pH decrease are plotted (dashed curves in Fig. 5(a))
Fig. 5. (a) pH-dependent color change and respective reflectance spectra of hydrogel pH sensor fabricated from PS-220 measured between pH of 1.26 and 6.99. A series of the spectra are obtained during the pH increase (solid curves), while the other series (dashed curves) are those during the pH decrease. (b) Plot of max changes upon pH increase and decrease showing a weak hystersis.
188
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
Fig. 6. pH-dependent change of reflectance maxima (max ) under various buffers of (a) KH2 PO4 (filled square), K(CH3 COO) (filled circle), K2 CO3 (filled triangle) for screening anion effect and (b) KH2 PO4 (open square), Na2 HPO4 (open triangle) for cation effect respectively. The sensor was fabricated from PS-240. The initial concentration of every buffer is 0.1 M, which is subsequently mixed with aq. HCl or NaOH for pH adjustment.
along with those during pH increase, which shows a slight hysteresis (max vs. pH shown in Fig. 5(b)) as previously reported [4,5]. Recently, Asher and coworkers extensively investigated the swelling mechanism of polyHEMA-based hydrogel sensor, in which they attributed the major reason for the hysteresis to the low ionic strength to fully re-equilibriate the Donnan potential during pH decrease. The slow diffusion of ions through the hydrogel was also accounted in their report. Later in the current report, the detailed pH response kinetics of our hydrogel sensor will be discussed. Although it is assumed that the hydrogel sensor selectively responds to the pH variation, there exist complex ion equilibria in the mixed buffer system which could affect Donnan equilibrium exerted by proton. Thus, possible matrix effects need to be screened to assure the precision of measurement. For most of the sensing experiments in this study, 0.1 M potassium phosphate (KH2 PO4 ) buffer mixed with HCl and/or NaOH was used since its buffering pH range decently covers the most sensitive pH range of our hydrogel system (around pH 5–6) Nonetheless, other buffer systems of weak acids and conjugate bases can also be used instead of phosphate. In Fig. 6(a) the dependences of the sensor response (max vs. pH increase) on different anions in buffer such as potassium acetate (K(CH3 COO)) and potassium carbonate (K2 CO3 ) as well as potassium phosphate are shown. Likewise, the cation effects on sensor responses were also tested using sodium phosphate instead of potassium phosphate as shown in Fig. 6(b). All those buffers were prepared at the concentration of 0.1 M. Overall, the variation of buffer systems did not cause significant changes in max at pH ranges from 2 to 8 as shown in Fig. 6(a) and (b), although carbonate buffer showed a slight red-shift effect which is indicated by filled triangle in Fig. 6(a). From Fig. 6, it is obviously shown that the most sensitive pH range to change hydrogel volume is pH
5–6 region regardless of buffers. At pH below 4 or above 7, our sensor exhibited poor sensitivity. Although not shown, the buffer concentration effect was also tested in this study (supplementary materials). When highly concentrated buffer solution (∼1 M) was used, a common “salting-out” effect was observed which shrinks the hydrogel consequently blue-shifting max , while those below 0.1 M did not show significant effect on max . A significantly distinguishing feature of the inverse opal pH sensor fabricated in our study is a fast response time upon pH variation. The response kinetics are demonstrated in Fig. 7. pH variation experiments indicated that the highest sensitivity (max /pH) of our hydrogel sensor was observed over a pH range of 5–6 which is evidently shown in Fig. 6. Although the pKa (where Ka is acid dissociation constant) of AA is 4.25 at 25 ◦ C, the actual pKa of AA within copolymeric hydrogel is reported to be slightly higher due to a electrostatic interaction of adjacent AA units [8]. Thus we measured the response time during the pH increase (5 → 6) and pH decrease (6 → 5) as well. For the pH increase experiment, the sensor in the pH 5 buffer was transferred to a petri dish containing fresh pH 6 buffer. The pH decrease experiment was done in the same manner by transferring the sensor to a fresh pH 5 buffer. The time dependent max change for pH increase is shown in Fig. 7(a). It is evident that max rapidly increases without the diffusion-limited slow max shift at the initial stage of pH change as reported previously [4,5]. In order to define the sensor response time, max vs. time plot during pH variation was fit to a single exponential function (y = max,t=0 + A exp(−t/)) with a r2 = 0.974 (Fig. 7(a)). The sensor response time was defined to be of the exponential function. When the initial pH 5 was changed to pH 6, the response time was measured to be 12.1 s, more than 30 times faster than a similar hydrogel sensor reported earlier [4]. In
Fig. 7. (a) Response kinetics of pH sensor made from PS-240 colloidal crystal template during pH increase from 5 to 6. max vs. time is fitted by a single exponential function with = 12.1 s as drawn by solid curve. (b) Response kinetics during pH decrease from 6 to 5. max vs. time is fitted by a double exponential function with 1 = 8.1 and 2 = 208 s as drawn by solid curve. (c) Response kinetics with respect to different range of pH variation.
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
addition to the pH increase experiment, the response time for the pH decrease (6–5) was also monitored. In this case, single exponential function did not accurately fit the experimental results. A closer look at Fig. 7(b) reveals that after a rapid blue shift of max from 625 to 580 nm within 20 s, a slow restoration of max back to 540 nm takes place. A double exponential fit was performed, and the function (y = max,t=0 + A1 exp(−t/ 1 ) + A2 exp(−t/ 2 )) with 1 = 8.1 s and 2 = 208 s resulted in a good fit (r2 = 0.995) (Fig. 7(b)), implying there might be two distinct deswelling mechanisms (the fitted values of other parameters are max,t=0 = 553, A1 = 40, A2 = 31 respectively). The initial hydrogel thickness is 25 m, and there is a ∼10% longitudinal swelling at a pH of 5 (refer to Fig. 5(b)). Assuming proton diffusion is the same in the inverse opal structure as in an aqueous medium, and that ion exchange at the carboxylic acid group is a diffusion-limited process, = l2 /DH+ ∼ = 7.6 s (l = 27.5 m, DH+ = 8.0 × 10−5 cm2 /s) which coincides very well with experimentally obtained 1 (=8.1 s). However, proton diffusion through the ionic hydrogel network should be retarded compared to that through the free water [15]. For a diffusion limited process, Tanaka successfully formulated the time required for ion (proton) diffusion through the ionic hydrogel using the hydrogel volume before and after swelling (V/V0 ) as shown in the following relationship, which can be further simplified using the ratio of diffraction maxima (/0 ∼ 625/580 = 1.08) assuming that the swelling occurs only along the thickness direction [4,5]. =
(V/V0 )
2/3
l0 1 + ([AA]/Ka ) Dion
∼ =
2/3 l0 1 + [AA]/Ka 0
Dion
l0 [AA], Ka , and Dion respectively denote hydrogel thickness (25 m), the concentration acrylic acid (0.16 M) within the hydrogel, acid dissociation constant (5.62 × 10−5 M), and the diffusion constant of proton (8.0 × 10−5 cm2 /s). The response time assuming our system is a solid hydrogel was calculated to be 219 s, which appears in accordance with the measured 2 (207 s). Since our inverse opal hydrogel primarily consists of void spaces interconnected by holes (Fig. 8), when the voids are connected, the diffusion of proton can take place both through the void space or through the hydrogel network. When the sensor is in its initial form (at low pH), the inverse opal will contain ∼26% (or less) of hydrogel and the remainder of volume will be void space (free water) as shown in Fig. 8. Since the void space is dominant and highly interconnected, the most of ionic species are expected to diffuse rapidly through the voids, and then into the hydrogel, resulting in an aqueous diffusion limited response time (∼8 s). Once the hydrogel is swollen at high pH the hydrogel will occupy much of the void space as shown as the right figure in Fig. 8 (the hydrogel is pinned to the substrate, and thus lateral swelling is not allowed), which has been experimentally confirmed in our previous investigation [13]. Since the proton is now diffusing through a swollen gel, diffusion is much slower, and such hindered diffusion will be reflected by 2 , which is convoluted with the fast diffusion ( 1 ) since some void spaces still exist in the hydrogel [13]. The two exponential decay function as shown in Fig. 7(b) may be attributed to such a dual diffusion mechanism. By applying lower pH buffer than 5 during the pH decrease, max can be further blue-shifted to exhibit more explicit color change of the hydrogel. Larger max changes upon larger pH gradients are clearly shown in Fig. 7(c). A dual diffusion mechanism appears to explain the observed deswelling kinetics quite well, however in an attempt to optimize the response, it is necessary to compare the hydrogel sensors with different crosslinking densities. We compared the response time of two pH sensors in which only the exposure condition for the photo-polymerization is different. As previously discussed, photo-polymerization without an epoxy filter resulted in a high crosslinking density and often delamination of the hydrogel from
189
Fig. 8. Schematic representation of a [111] layer of FCC inverse opal hydrogel at low and high pH respectively. The hydrogel holds the as-prepared inverse opal structure at low pH, and the voids become filled with swollen hydrogel at an increased pH assuming that the lateral swelling is minimal.
the substrate. However, the inverse opal sensor could be obtained by short exposure (20 min) of unfiltered (strong) UV light. Fig. 9 compares the pH sensing kinetics for the sensors obtained either by normal exposure condition (2 h photo-polymerization using epoxy filter) or strong exposure condition (20 min without epoxy filter). It is obvious that the hydrogel obtained by strong UV exposure is less swollen in the same pH buffer exhibiting lower max due to a high crosslinking density. Since the same sacrificial PS particles were used for both cases, the crosslinking density is the only reason for the apparent max differences at the same pH buffer. More importantly, a longer pH response time ( 1 ) was observed for highly crosslinked hydrogel sensor. A single exponential function could fit both max changes fairly well, giving 1 for an optimized sensor and a highly crosslinked sensor to be 12.8 and 101 s respectively. 101.0 s is still fast compared to our previous report, in which the response time for pH change between 5 and 6 was as slow as 1200 s [4]. The slower response of the prior system could be attributed to a higher crosslinking density, smaller interconnecting holes, larger hydrogel content, and higher AA concentration than our sensor (5 vs. 2.5%). Our sensor has a different physical structures compared with that reported by Asher and coworkers due to the distinctly different fabrication strategies. Their sensor had a much higher hydrogel volume fraction, and the thickness is thicker, the calculated and the observed response time were in the order of several hours [5]. Thus, we concluded that the optimization of the photopolymerization conditions such as exposure time, intensity, and wavelength of UV-lamp are crucial in order to achieve fast response
Fig. 9. Response kinetics obtained from two different hydrogel sensors. Two sensors are fabricated using the same monomer mixture and the PS template (PS240), only differing in the UV exposure conditions for photo-polymerization. The upper response kinetics is obtained from an optimized exposure condition (using epoxy filter), and the lower kinetics is from a strong UV exposure without optical filter. Both data are fitted by single exponential fit to give 1 of 11.8 and 101 s respectively.
190
J. Shin et al. / Sensors and Actuators B 150 (2010) 183–190
Fig. 10. Repeatability and reproducibility of pH sensing experiment on a sensor made from PS-240. (a) Repeated pH-dependent reflectance maxima shifts, (b) Repeated response time of PS-240 sensor measured over 3 months.
chemical sensor, since they determine the crosslinking density of hydrogel and the size of holes between voids, the most important control factors for the response time. In order to demonstrate repeatability and reproducibility of the hydrogel sensor more rigorously, fifty pH variations between 5 and 6 were performed over 3 months. Fig. 10(a) shows sensing kinetics measurement plots for randomly selected data, and Fig. 10(b) is the plot of response time vs. number of experiments. The sensor was still highly responsive 5 months after the fabrication without visible degradation of iridescence or mechanical robustness. We expect actual lifetime of the sensor to be longer than 1 year unless it is completely dried. 4. Conclusions Through a rigorous optimization of fabrication procedures including the monomer ratio, polymerization time, optical filtering of UV light, and colloid, a fast responding hydrogel inverse opal pH sensor was successfully formed. The pH sensor exhibited response time of ∼12 s for pH variations between 5 and 6, which was ∼100 times improved compared to the previously reported inverse opal pH sensors [4]. In addition to improved response time, sensor lifetime was found to be longer than 5 month without loss of reproducibility or iridescent color. We expect that the distinguished features of the hydrogel sensor developed in this study such as fast response, and mechanical robustness will find promise a variety of chemical sensor applications. In particular, fast response enables a real time sensing of chemical species in continuous processes such as chromatography. By incorporation of novel vinyl monomers containing various sensing moieties in the pendent group, investigations on a various chemical sensors are underway. Acknowledgments This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2008-331-D00153). PVB thanks the US DOE ‘Light-Material Interactions in Energy Conversion’ Energy Frontier Research Center under grant DE-SC0001293 for support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.snb.2010.07.018.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
J.H. Holtz, S.A. Asher, Nature 389 (1997) 829–832. J.D. Debord, L.A. Lyon, J. Phys. Chem. B 104 (2000) 6327–6331. K. Lee, S.A. Asher, J. Am. Chem. Soc. 122 (2000) 9534–9537. Y.J. Lee, P.V. Braun, Adv. Mater. 15 (2003) 563–566. X. Xu, A.V. Goponenko, S.A. Asher, J. Am. Chem. Soc. 130 (2008) 3113. A.V. Goponenko, S.A. Asher, J. Am. Chem. Soc. 127 (2005) 10753. Y.J. Lee, S.A. Pruzinsky, P.V. Braun, Langmuir 20 (2004) 3096–3106. M. Men-Moshe, V. Alexeev, S.A. Asher, Anal. Chem. 78 (2006) 5149–5157. M.M.W. Muscatello, L.E. Stunja, S.A. Asher, Anal. Chem. 81 (2009) 4978. T. Endo, Y. Yanagida, T. Hatsuzawa, Sens. Actuators B 125 (2007) 589– 595. Y. Takeoka, M. Watanade, Langmuir 18 (2002) 5977–5980. M. Honda, T. Seki, Y. Takeoka, Adv. Mater. 21 (2009) 1801–1804. Y.J. Lee, C.E. Heitzman, W.R. Frei, H.T. Johnson, P.V. Braun, J. Phys. Chem. B 110 (2006) 19300–19306. J. Wang, Y. Cao, Y. Feng, F. Yin, J. Gao, Adv. Mater. 19 (2007) 3865–3871. S. Mafe, J.A. Manzanares, A.E. English, T. Tanaka, Phys. Rev. Lett. 79 (1997) 3086. M. Egen, R. Zentel, Chem. Mater. 14 (2002) 2176–2183. S.H. Park, B. Gates, Y.N. Xia, Adv. Mater. 11 (1999) 462–466. B. Gates, D. Qin, Y.N. Xia, Adv. Mater. 11 (1999) 466–469. Data Sheet for “Irgacure-651” provided by Ciba Specialty Chemicals Inc., 2001.
Biographies Jinsub Shin received his BS degree from Sejong University, Korea in 2009. He is currently a graduate student in Sejong University under an advice of Prof. Wonmok Lee at the Department of Chemistry. He is working on the fabrication of photonic crystal devices such as hydrogel sensors, and micro-lasers based on colloidal selfassembly techniques. Paul V. Braun is a Professor of Materials Science and Engineering, and an affiliate of the Frederick Seitz Materials Research Laboratory, the Beckman Institute for Advanced Science and Technology, and the Department of Chemistry at the University of Illinois at Urbana-Champaign. Prof. Braun’s research focuses on the synthesis and properties of 3D architectures with a focus on materials with unique optical, electrochemical, thermal, and mechanical properties. Prof. Braun received his BS degree with distinction from Cornell University in 1993, and his PhD in Materials Science and Engineering from Illinois in 1998, both in Materials Science and Engineering. Following a postdoctoral appointment at Bell Labs, Lucent Technologies, he joined the faculty at Illinois in 1999. Wonmok Lee is an Assistant Professor of the Department of Chemistry at Sejong University. He received his BS degree in 1995 from Pohang University of Science and Technology, MS degree in 1997 and PhD degree in 2001 from the same university. After postdoctoral periods in University of Illinois at Urbana-Champaign (UIUC) and Massachusetts Institute of Technology (MIT), Prof. Lee developed an industrial career at Samsung Advanced Institute of Technology (SAIT). He joined the faculty at Sejong University since 2007. Prof. Lee’s research area covers the polymer electrolyte membranes, photonic bandgap materials, sensors, and other optical devices using self-assembled polymers.
