Reactive & Functional Polymers 79 (2014) 68–76

Contents lists available at ScienceDirect

Reactive & Functional Polymers

journal homepage: www.elsevier .com/ locate / react

Hydrogel films engineered in a mesoscopically ordered structure andresponsive to ethanol vapors

http://dx.doi.org/10.1016/j.reactfunctpolym.2014.03.0161381-5148/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +39 09123863710; fax: +39 09123860840.E-mail address: clelia.dispenza@unipa.it (C. Dispenza).

Clelia Dispenza a,⇑, Maria Antonietta Sabatino a, Sabina Alessi a, Giuseppe Spadaro a, Leonardo D’Acquisto a,Riccardo Pernice b, Gabriele Adamo b, Salvatore Stivala b, Antonino Parisi b, Patrizia Livreri b,Alessandro C. Busacca b

a Dipartimento di Ingegneria Chimica, Gestionale, Informatica, Meccanica (DICGIM), Università of Palermo, Viale delle Scienze, Edificio 6, 90128 Palermo, Italyb Dipartimento di Energia, ingegneria dell’Informazione, e modelli Matematici (DEIM), Università degli Studi di Palermo, Viale delle Scienze, Ed. 9, 90128 Palermo, Italy

a r t i c l e i n f o

Article history:Received 6 November 2013Accepted 29 March 2014Available online 12 April 2014

Keywords:Responsive-hydrogelsSwellingPhotonic crystalsEthanol vapor sensing

a b s t r a c t

Responsive hydrogels filling the interstitial spaces of photonic crystals can form mesoscopically struc-tured materials, which exhibit reversible shifts in the Bragg diffracted light as a response of environmen-tal changes. These materials can be used to generate chemical or biochemical sensors. The present workreports on the synthesis and characterization of ethanol responsive hydrogels that can be used in thedesign of novel breathalyzers. The dynamic mechanical behavior of the macroscopic hydrogels and theirswelling features in the presence of different liquids or vapors have been investigated to orientate thechoice of the best responsive material and curing process. The swelling behavior of a selected hydrogelembedding the photonic crystal made of polystyrene nanoparticles as function of the concentration ofethanol vapor was studied through UV–Vis optical transmission spectroscopy and compared to thebehavior of the macrogel analogue.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Stimuli responsive hydrogels have been extensively studied asactive materials for a variety of chemical and biochemical sensors,for their ability to undergo reliable, robust and often reversible‘‘phase transitions’’ in response to small local environmental mod-ifications, which bring them away from their equilibrium states[1–6].

Different hydrogel systems, in terms of chemical composition,crosslinking degree and density can be designed to confer specificsensitivity to a stimulus or a combination of stimuli, such as tem-perature, electrical field, light, pH, solvent composition and specificions, and to control the extent of material modification, generallyin terms of its stiffness, shape, volume or mass [7–9].

Hydrogel-based sensors can be miniaturized and integratedinto microfluidic systems and microsystems [10–13].

At the basis of the hydrogel responsiveness there are modifica-tions of polymer–solvent interactions, as a result of changes of theenvironmental conditions, which may cause further solvent uptakeand swelling or solvent outflow and deswelling. Therefore, mainlythermodynamic criteria and network mechanics govern hydrogels

responsiveness. Synchronized, cooperative motions of severalpolymer segments of the network are required for these systemsto achieve a local or global free energy minimum in a new equilib-rium position. As a result, the time-dependence of these responsesmay exceed the temporal switching dynamics required for sensingapplications. Reducing the mass or the volume of the switchingsystems, and simultaneously increasing the interfacial area withthe environment, yet preserving the robustness of the signalderived by the collective response of the material, should facilitatemass/heat transport and enable higher responsivity and/or fasterresponse.

Stimuli-responsive hydrogel thin films, with a periodicallyordered structure at the mesoscale, represent a promising platformfor obtaining increased stimulus-sensitivity and reduced response-time optical sensors. As demonstrated by the pioneering work ofAsher and coauthors [14], stimuli-responsive hydrogels filling theinterstitial spaces of either silica or polymeric nanoparticles orga-nized in a crystalline array can diffract the light at visible wave-lengths determined by the lattice spacing, d, which gives rise toan intense color in the visible range. Hydrogel swelling or shrink-ing, in response to an external stimulus, changes the meanseparation between the colloidal spheres and induces a diffractedlight Bragg peak shift to either longer or shorter wavelengths.Removal of nanoparticles template leads to the formation of

Table 1Composition of the monomers in the feed.

System I (mol) II (mol)

HEMA 0.35 0.27PEG200DMA 0.0013 0.0013AA 0.009 0.09Water 0.634 0.634

C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76 69

hydrogel inverse photonic crystals (HIPCs) with a periodic distribu-tion of voids.

The hydrogel-infiltrated photonic crystal, depending on itschemical composition, can exhibit significant reversible visiblelight diffraction shifts in response to a variety of stimuli, such astemperature, pH, presence of specific molecules or biomoleculesand mechanical forces [15,16,14,17,18]. The stimuli-responsive-ness must be accompanied by an adequate elasticity and chemicalstability of the hydrogel network. As a result the mesostructuredfilm will be able to withstand repeated swelling/de-swellingcycles, when in use, as well as erosion due to prolonged exposureto the swelling medium. Moreover, it will survive to the action ofacid etching or organic solvents extraction required by the colloi-dal crystal template removal process for the preparation of theinverse photonic crystal [19].

For the development of ethanol vapors sensors, ethanol-respon-sive hydrogels can be considered as active materials. Yet, it shouldbe considered that ethanol can be only one of the components of avapor-phase in equilibrium with as aqueous liquid mixture of dif-ferent composition, generally being water the most abundant com-ponent. Therefore, regardless the number and variety of othercomponents that can be present in the vapor phase, the presenceof water imposes either the recourse to technological solutionsfor dehydration, with relative costs implications, or, more prefera-bly, to properly designed materials which are able respond to vari-ations of ethanol concentration also in the presence of substantialamounts of water.

There are interesting studies which report Bragg diffractionshifts in a wide region of the visible spectral region when e.g., acrosslinked 2-hydroxyethyl methacrylate (HEMA) hydrogel isexposed either to pure liquid water or to concentrated ethanol/water liquid solutions [20].

However, at the best of our knowledge, there are no equivalentstudies that report on the ability of hydrogel inverse opals to spe-cifically respond to ethanol vapors after saturation with water. Inparticular, the hydrogel network should be designed so that itcan uptake and retain water, when exposed to water vapor-richatmospheres, and swell further when the atmosphere is progres-sively concentrated of ethanol vapors. For this purpose, 2-hydroxy-ethyl methacrylate (HEMA) was used as the main building block ofthe network for its known favorable Flory–Huggins mixing param-eter with ethanol. Acrylic acid (AA) at two different ratios was alsoconsidered as co-monomer for its affinity toward water and itscontribution to hydrogel network mechanical properties, due tothe establishment of further crosslinking through strong secondaryinteractions. Finally, poly(ethylene glycol) dimethacrylate (PEG200-

DMA) was used as crosslinking agent. The polymerization processcombined a ‘‘cold’’ UV-photocrosslinking step and a thermal post-cure. FTIR spectroscopy and dynamic mechanical thermal analysiswere carried out on the macrogel analogues (macrogels), i.e., wall-to-wall hydrogels polymerized from the same monomer mixtureand with the same cure and post-cure cycles as for the mesoscop-ically ordered hydrogel films. That was performed in order to seekfor confirmation of the success of polymerization and to withdrawuseful information on the complex shear moduli of the hydrogelssynthetized, respectively. Preliminary swelling studies were car-ried out exposing macrogels to liquid water, ethanol, methanoland acetone and to any of these organic solvents after equilibrationwith water. The most promising of the two formulations, selectedfrom the preliminary screening, was used to infiltrate a polysty-rene (PS) colloidal crystal generated onto pre-etched silica micro-scope slides through self-assembly from aqueous dispersions ofthe nanoparticles [21]. The periodically ordered hydrogel filmwas then evaluated as active component of an ethanol vapor opti-cal sensor by means of UV–Vis transmission measurements at thevariance of ethanol vapor concentration. Differences in the swell-

ing behavior of the periodically ordered PS/hydrogel thin filmand the corresponding macrogel will be discussed.

2. Experimental

2.1. Preparation of HEMA/AA hydrogels

Hydrogels are composed by 2-hydroxyethyl methacrylate(HEMA), acrylic acid (AA) and poly(ethylene glycol)200 dimethacry-late (PEG200DMA), all from Sigma Aldrich. HEMA was distilledunder vacuum prior to the use, while all the other monomers wereused as received. Azobis-isobutyronitrile (AIBN) was used as initi-ator at 0.15% wt/vol. Two formulations were investigated, with twodifferent amounts of AA, as reported in Table 1. The monomer andsolvent (water) mixture was cured within silicon molds of 12 mmdiameter closed with a quartz cover slip to prevent excessive waterloss during photo-polymerization. UV-irradiation was carried outfor 2 h with a UV irradiator from Helios Italquarz ‘‘Polymer 125UV’’, equipped with a high-pressure Hg lamp ‘‘Zp-type’’ (2 mW/cm2). The temperature inside the chamber was maintained at25 ± 1 �C. After photocuring, a thermal post-curing treatment inan oven at 60 �C for further 2 h was carried out. Hydrogel circulardisks (12 mm diameter and 2 mm thickness) were then subjectedto prolonged extraction (72 h) with water at room temperatureto remove unreacted monomers.

2.2. Solubility tests and FTIR analysis of poly(HEMA-co-AA) hydrogels

Gel fractions in water, defined as GF = Wg/W0, where W0 is thetotal amount of solids in the sample before extraction and Wg isthe weight of the water-insoluble network, were measured bySoxhlet extraction. Reported data are the average of minimumthree independent measurements. Variability on the datum iswithin the unity.

Structural confirmation of the HEMA–AA copolymerization andcrosslinking upon photo-irradiation and thermal-post curing wassought via FTIR analysis carried out on the water insoluble frac-tions. In particular, FTIR spectra were recorded in the range4000–400 cm�1 using a Perkin–Elmer 1720 Fourier TransformSpectrophotometer with a resolution of 1 cm�1, each spectrumrecorded after 100 scan. Solid samples were dispersed in KBr andpressed into disks.

2.3. Dynamic mechanical thermal analysis

Dynamic mechanical properties of the hydrogels were investi-gated by Dynamic Mechanical Thermal Analysis, using a Rheomet-rics DMTA V instrument. DMTA analysis was performed onhydrogel pellets in their equilibrium swelling conditions in waterat 25 �C. Tests were performed in compression, both as functionof the strain, at the fixed frequency of 1 Hz, and as function ofthe frequency in the 0.1–10 Hz range, at a fixed strain of 0.5%(within the LVR). Temperature was maintained at the constanttemperature of 25 ± 1 �C. Storage modulus (E0) and loss modulus(E00) versus either strain or frequency were recorded. Highly repro-ducible mechanical spectra were obtained.

70 C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76

2.4. Swelling properties of poly(HEMA-co-AA) hydrogels

Swelling properties of hydrogels by liquid media were deter-mined on the insoluble portions of the hydrogels, accordingly toestablished procedures [22].

Hydrogel samples, previously air dried at 50 �C overnight, wereimmersed at room temperature in different liquids, namely water,ethanol, methanol and acetone, and in any of these organic sol-vents after their equilibration in water. Swelling ratio values, SR,defined as Ws/Wd, where Ws and Wd are the measured weight ofthe hydrogel in the swollen state at fixed time intervals and inthe initial dry state, respectively, were recorded as function ofthe time. Reported values are the average of minimum eight sam-ples from minimum two independent preparations.

Swelling of hydrogels by water or water/ethanol vapor mixtureswere carried out on hydrogel circular disks previously equilibratedin air at room temperature (24 �C) and relative humidity (RH) of45%. Samples were placed on the measuring plate of a precisionbalance (Mettler Toledo XS 205) in a sealed chamber where alsoa reservoir for liquids was placed. The reservoir was initially filledwith liquid water. Weight increase of samples exposed to a water-saturated air atmosphere was followed over time until a plateauwas reached, then given volumes of ethanol were added to waterto obtain a water/ethanol vapor mixtures of known composition(see Table 2). Variation of sample weight, after the addition ofliquid ethanol to water, was recorded. The inspection of samples(Fig. 1) subjected to the swelling experiment under a light opticalmicroscope shows roughness on the surface when the hydrogel isexposed to water and water/ethanol mixtures. This behavior isreversible.

2.5. Refractive index measurements

Refractive index measurements were carried out on the macro-gels analogues, both in ‘‘dry’’ state (i.e., in equilibrium with air at atemperature of 23.8 �C and relative humidity of 45%), and in wateror ethanol saturated conditions. A fully automated refractometer(Metricon Model 2010/M Prism Coupler), providing high accuracymeasurements of refractive index (Index accuracy ±0.001, Indexresolution: ±0.0005), was employed. Measurements were per-formed around the stop band wavelength (532.0 nm).

2.6. Preparation of HEMA/AA hydrogel infiltrated photonic crystals

Self-assembly of a crystalline colloidal array of monodispersepolystyrene nanoparticles was carried out using a commercialpolystyrene nanoparticle dispersion (Polysciences, 0.3 wt% solidcontent) with an average diameter of 220 nm (±5%). The 3Dclose-packed array of polymer nanoparticles was grown onto Pira-nha solution pre-etched glass slides through a solvent evaporation-induced vertical deposition method, maintaining the substrate infixed vertical position [23].

Direct opals (DO) produced using this technique showed a fairlyuniform iridescent surface (Fig. 2a) and a regular 3D periodicarrangement of the polystyrene colloid (Fig 2b–c) as observed by

Table 2Composition of the liquid and vapor phases at equilibrium (24 �C and RH of 45%) usedfor the swelling experiments both with macrogels and infiltrated photonic crystals.

Liquid phaseEtOH:H2O(vol:vol)

EtOH in the vaporphase (molefraction)

H2O in the vaporphase (molefraction)

EtOH in thevapor phase(ppm)

1:9 9.61 � 10�3 0.030 15,2202:9 16.3 � 10�3 0.0295 26,0003:9 21 � 10�3 0.0289 33,3604:9 24.7 � 10�3 0.0284 39,033

scanning electron microscopy analysis carried out by a Philips505 Quanta 200 ESEM FEG. The quality of template was alsoassessed by optical transmission measurements carried out at dif-ferent distances from the top edge of the deposit [21].

Small droplets (approx. 10 ll) of monomers aqueous solution(system II) were dropped on the top-edge of the direct photoniccrystal deposit, which was inclined at a fixed angle of 15� withrespect to the horizontal plane. The infiltration process was fol-lowed by the color change of the deposit from white to colorless.When infiltration was completed, a quartz slide was placed ontop of the deposit to avoid uncontrolled accumulation of materialon the surface and limit water loss during photo-polymerization.Cure and post-cure were carried out as described above for themacrogel analogue. After removal of the top-cover, samples wereequilibrated in bidistilled water at room temperature for 24 h, thenstored in air at room temperature.

Preservation of the periodic structure was also confirmed bySEM microscopy analysis on the hydrogel-infiltrated opals afterremoval of the PS template through immersion in chloroform over-night, rinsing with acetonitrile and water and drying in air. Panelsa and b of Fig. 3 refer to the surface of the film, while panels c and dto the interiors in a fractured portion of the same. The surfaceshows disordered zones, due to the presence of incomplete crystalplanes in the direct template, portions of the surface where theporosity is occluded by a continuous layer of polymer, and portionswith an open and regularly laid out porosity. Conversely, through-out the thickness of sample the porosity at the mesoscale looksfairly ordered (see Fig. 3 c and d).

2.7. Optical characterization of HEMA-co-AA hydrogel infiltratedphotonic crystals

In order to perform the optical characterization of the samples,we used the set-up described in detail elsewhere [21]. Transmis-sion measurements were performed in the 400–700 nm range witha wavelength step of 1 nm, dividing the transmitted power by theincident one. The light beam was spectrally filtered by a mono-chromator and focused onto the opal through a 50 mm lens. Forthe detection, a Silicon Photomultiplier (SiPM) [24] was used,while a lock-in amplifier was employed to reduce noise. In orderto reduce the influence of external light, the lens and the SiPMwere placed into a black box. The sample was enclosed into a100 cm3 sealed polyethylene box, possessing two fused silica win-dows to allow the passage of light and a small hole connected tothe outside of the black box through a hose to introduce liquids.

The sample was exposed to a controlled vapor atmosphere cre-ated by a layer of liquid filling the bottom of the container. Partic-ular care was taken to avoid any contact between the sample andthe liquid phase. The temperature was monitored during all themeasurements.

Distilled water was first inserted into the box through the hosefor a first series of measurements; subsequently, known volumesof liquid ethanol were added to the water phase.

A dwell time of around 220–240 min was considered for theliquid–vapor equilibrium to be established before the transmissioncurves were acquired at regular time intervals (10 min). When thetransmission curves were invariant with time, more liquid ethanolwas added to the container. Ethanol vapors concentration is thesame as for the swelling experiments with macrogels (see Table 2).

3. Results and discussion

3.1. Structural, physico-chemical and dynamic-mechanical properties

Synthesis of macro HEMA-co-AA hydrogels with two differentHEMA/AA ratios was carried out by photo-induced polymerization

Fig. 1. Visual inspection of the macrogel disks under the light optical microscope: as prepared (a) after 5 min of exposure to water (b), after 15 min of exposure to water and(c).

Fig. 2. Direct photonic crystal film made of polystyrene nanoparticles onto glass: bright green iridescent surface (a) and SEM micrograph showing its microstructurecomprised of close packed microspheres (b and c).

C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76 71

at room temperature for the evaluation of network chemical stabil-ity, elasticity and responsiveness to ethanol and other organic mol-ecules, such as acetone and methanol, often present in mixturewith ethanol, e.g., in alveolar air exiting while breathing.

The obtainment of a chemical, permanent gel is confirmed bysolubility tests carried out with a variety of solvents, includingwater, acetonitrile, N-methyl pyrrolidone and chloroform. The gelfraction in water, as measured by the Soxhlet apparatus straightafter photo-polymerization, resulted to be 93 ± 1% for both sys-tems. No significant weight changes were observed when hydrogeldisks were kept immersed for several days in chloroform, which isthe solvent generally employed for the removal of the polystyrenenanoparticles used as template. Resistance to other organic sol-vents, like ethanol, methanol and acetone will be discussed inthe following.

In Fig. 4(a and b) FTIR spectra of the two formulations in the4000–2000 cm�1 and 2000–450 cm�1 regions are reported. Spectraof the two main monomers, HEMA and AA, are also reported forcomparison. The hydroxyl group band for hydrogels (Fig. 4a)stretches significantly toward the lower wavenumbers for the con-tribution of vibrations from polymeric and hydrogen bondedhydroxyls. The corresponding bending vibrations in the 1500–1250 cm�1 region are shifted to 1284 and 1238 cm�1 from theirusual position at 1322 and 1300 cm�1 in the HEMA monomer. Abroad band in the range 3200–2500 cm�1, also present in the AAmonomer, is clearly evident for both systems and more pro-nounced for the system richer in acrylic acid. This band can beattributed to hydrogen bonded carboxylic groups. Both hydrogelsshow a distinct shift of the carbonyl vibration band toward thehigher frequencies (1730 cm�1), with respect to the HEMA

monomer (1720 cm�1) as a result of formation HEMA-co-AAcopolymer segments. The carbonyl band for the monomer isshifted to a higher position with respect to the usual wavenumber,for the presence of the conjugated unsaturation (see Fig. 4b). Noevidence of unsaturated groups is present in both hydrogels, asrevealed by the disappearance of the band at 1635 cm�1, charac-teristic of C@C stretching vibration, and bands at about 980–950 cm�1 characteristic of d CH (out of plane) vibrations for themonomers.

Dynamic mechanical strain sweep and frequency sweep curvesare reported in Fig. 5 (a and b). They refer to samples that havebeen previously immersed in water until constant weight wasattained.

The rheological behavior of the hydrogels obtained from formu-lations I and II revealed similar characteristics. Both systems showa fairly wide linearity range in the storage (E0) and loss (E00) modulicurves of the strain sweep tests (Fig. 5a), with E0 always lying aboveE00. The frequency sweep tests show an E0 plot which is almostinvariant with frequency within two decades. This behavior ischaracteristic of chemically crosslinked networks. The higher elas-tic modulus and the lower loss modulus for the formulation with ahigher AA content suggest a higher crosslinking density for thissystem, probably due to a reinforcing effect from hydrogen bondsinvolving carboxyl groups.

The swelling behavior of the two hydrogels was investigated byimmersing air-dried samples in liquid water first and measuringthe water uptake as function of the time until a plateau wasreached. Samples were then immersed in pure ethanol. In a parallelexperiment, similar air-dried samples were directly immersed inpure ethanol. Swelling ratio as function of the time is reported in

Fig. 3. Hydrogel inverse photonic crystal made by polymerizing the hydrogel in the interstitial voids of the polystyrene.

72 C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76

Fig. 6(a and b). For the system I (Fig. 6a) the initial rate of solventuptake is higher for water than for ethanol. When swelling is car-ried out in water, a plateau at about SR = 1.4 is reached after about8 h, corresponding to the ‘‘equilibrium’’ swelling condition in thissolvent. When samples are then transferred to ethanol, the swell-ing ratio further increases until a second plateau is reached atabout SR = 1.95. When samples were swelling directly in pure eth-anol starting from the dry state, the experiments were alwaysinterrupted before the samples could reach the ‘‘equilibrium’’ con-dition, because they were starting to loose their integrity. This wasnot the case of ethanol exposure after swelling in water. A similarbehavior was observed for the formulation II (Fig. 6b). In general,for this system higher swelling values were attained in water(SR = 1.45), in ethanol after water (SR = 2.10) and in ethanol beforeerosion (SR = 2.20). These results point out to higher solvent resis-tance and higher swelling ability for the system with the highercontent of AA. These two properties, together with the observedslightly higher storage modulus and lower loss modulus, suggestthe formation of a more homogeneously crosslinked network,which can better rehydrate from the dry state. Lack of homogene-ity in the network density, with denser and looser portions of thematerial, may lead to irreversible collapse of the loose parts upondrying and limited swelling ability of the more densely crosslinkedportions [25].

The higher specific affinity for ethanol of these HEMA basedhydrogels was clearly demonstrated by the comparison betweenthe swelling curves of samples equilibrated in water and immersedeither in ethanol or in methanol afterwards. Higher SR values wereattained with ethanol, as it is shown in Fig. 6(a) for formulation I,while direct exposure to methanol caused faster erosion (datanot shown).

Acetone showed a markedly different behavior. It caused con-traction of the water-swollen hydrogel, e.g., SR from 1.45 to 1.25

for the hydrogel II. Samples that were dried and weighted afterdwelling in acetone did not show any significant dry weight loss,thus suggesting deswelling due to water outflow, rather than net-work erosion. When dry samples were directly immersed in ace-tone, they did not show significant solvent uptake.

The large differential swelling in ethanol after water, muchmore pronounced for ethanol than for methanol, and deswellingin acetone, all together, encourage to pursue on the use of thesesystems, and particularly of formulation II, as active material forethanol vapor sensors.

In a second series of swelling experiments, dry samples of for-mulation II were placed on the plate of an analytical balance andexposed to a water vapor saturated atmosphere. The swelling ratiowas measured as function of the time at 24 �C, as shown in Fig. 7.After reaching a plateau in the swelling curve the influence ofincreasing content of ethanol vapor in the atmosphere was inves-tigated. In particular, three different ethanol/water volume ratiosin the liquid phase were considered, as reported in Table 2. It isinteresting to observe that, although the absolute values of SRwhen the dry hydrogel disks were exposed to vapor saturatedatmospheres are lower and swelling kinetics slower than in thecase of immersion in liquids, the behavior is qualitatively similar,with samples reaching a plateau when exposed to water vaporsand further increasing their SR when the vapor phase was progres-sively enriched of ethanol vapors, up to a concentration of16.3 � 10�3 expressed as molar fraction of ethanol in air. Thisincrease in SR is not longer evident when the ethanol content isincreased to 21 � 10�3.

3.2. Optical characterization of the macrogel

Refractive index measurements on the macrogel II, reported inFig. 8, were performed at the wavelength of 532.0 nm and in three

Fig. 4. FTIR spectra of HEMA-co-AA macrogels and main constituent monomers inthe 4000–2000 cm�1 (a) and 2000–400 cm�1 (b) spectral ranges: acrylic acid (1); 2-hydroxyl ethyl methacrylate (2); formulation I (3); formulation II (4).

C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76 73

different conditions: (i) at the initial or ‘‘reference’’ state for thehydrogel, i.e., when the sample is in equilibrium with air at a tem-perature of 23.8 �C and relative humidity of 45% (grey dashed line);(ii) for a water-saturated sample that desorbs water when it is keptin air (open circles); and (iii) for the ethanol-saturated samplewhile it desorbs ethanol (open squares). It can be observed thatthe refractive index decreases when the hydrogel absorbs eitherliquid water (point A) or liquid ethanol (point B). In particular,the air-saturated sample showed a refractive index drop from1.513 to 1.468 when it is equilibrated in water, to regain 60% ofits original value after 60 min and its ‘‘reference’’ value after about

Fig. 5. Storage modulus (open symbol) and loss modulus (filled symbol) in strain sweep m(triangles).

500 min. This behavior points out to a good reversibility of thewater swelling/deswelling behavior of the hydrogel. A similartrend could be observed when the sample is immersed into etha-nol. In this case, the initial drop is from 1.513 to 1.435. Then, therefractive index increases again, more slowly than for water. Itregains about 68% of its original value after 360 min and its ‘‘refer-ence’’ value after about 2500 min. Indeed, the swelling process inethanol is also fully reversible and, more interestingly, the affinityof the network for ethanol is higher than for water, since the eth-anol desorption process is significantly slower.

3.3. Optical characterization of the hydrogel-infiltrated photoniccrystal

When the responsive hydrogel fills the gaps of a photonic crys-tal, the optical readout of the Bragg diffraction peak shift, or trans-mitted light intensity at a given wavelength, as a result of vaporsuptake can be used to determine the swelling degree of the meso-structured hydrogel thin film.

It was preliminary observed that the infiltration process of thehydrogel in the interstitial spaces of the PS colloidal crystal didnot alter its periodic structure and the Bragg resonance peak atnormal incidence was shifted from 510 nm, for the bare direct PSopal, to 543 nm as a result of the increase in spacing after hydrogelinfiltration.

The refractive index measurements described in the previoussubsection were used to calculate the filling factor of the infiltratedopal. In particular, a filling factor value (f) for the infiltrated opal off = 0.225 was estimated from the calculation of the effective refrac-tive index of the infiltrated photonic crystal, accordingly to Eqs. (1)and (2) [28]:

neff ;i ¼kB

2ffiffi23

qD

ð1Þ

and

n2eff ;i � 0:74n2

PS þ fn2hydr þ ð0:26� f Þn2

air ð2Þ

where kB is the Bragg resonance peak wavelength, D is the diameterof the polystyrene microspheres, nPS is the polystyrene refractiveindex, nhydr is the hydrogel refractive index, nair is the air refractiveindex. For the refractive index of the hydrogel equilibrated in air avalue of nhydr = 1.513 at 532 nm, experimentally measured on themacrogel analogue, was used.

Fig. 9a shows the transmission spectra as function of the timewhen the film, previously equilibrated in air is exposed to a water

ode (a) and frequency sweep mode (b), for formulation I (circles) and formulation II

Fig. 6. Swelling ratio as function of time for the macro HEMA-co-AA hydrogels immersed in different liquids: formulation I (a); formulation II (b). Curves that are prolongedrefer to systems that were equilibrated in pure water first and immersed in a different organic liquids afterwards. Spline lines are only a visual guide for the eye.

Fig. 7. Swelling ratio as function of time for the macro HEMA-co-AA hydrogel IIexposed to atmospheres containing a different concentration of ethanol vapors.

Fig. 8. Refractive index versus time (min). Open circles represent the water-saturated sample (point A) while drying out in air. Open squares represent theethanol case. Grey dashed line represents the ‘‘reference’’ state, i.e., the sample inequilibrium with air at a temperature of 23.8 �C and relative humidity of 45%.

74 C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76

vapor saturated atmosphere; Fig. 8b–e shows the main character-istics of the optical stop band as function of the time; panel f ofFig. 8 displays the swelling ratio Ds/D0 of the infiltrated hydrogelcalculated accordingly to equation [28]:

kB ¼ 2

ffiffiffi23

rD

Ds

D0

ffiffiffiffiffiffiffin2

eff

qkB ¼ 2

ffiffiffi23

rD

DS

D0neff ð3Þ

The transmission spectra of the films exposed to a water vaporsaturated atmosphere show a remarkable simultaneous progres-sive reduction of the transmission curve and peak red shift, up toa point when the spectra no longer change. The invariance of themeasured optical property, which corresponds to the swellingequilibrium for the infiltrated hydrogel, is attained now in the timescale of a few hours (250 min) rather than in days as for the mac-rogel analogue. In particular, the optical transmittance decreaseswhile the Bragg diffraction peak shifts from 543 nm to 576 nm.The two characteristics of the optical response pertain to two dif-ferent phenomena. The reduction of transmittance in Fig. 9(a)can be associated to a modification of the surface operated bythe water, inducing surface roughness and thereby scattering.The surface modification has been shown in paragraph 2.4 andmajor details can be found in [21]. The increase of Bragg diffractionpeak corresponds to the increase of the crystal lattice distance ofpolystyrene microparticles as a result of hydrogel swelling.

The steady state condition corresponds to a swelling degree of1.06, in good agreement to the values measured for the macrogelanalogue (see Fig. 9b). It is worth noticing that the refractive indexvariation of the macrogel, described in the previous subsection,would cause a blue-shift in the diffracted Bragg peak of only about4 nm. Moreover, such a shift would be in the opposite directionwith respect to the experimentally observed one. Therefore, themeasured red-shift is only ascribed to swelling.

As already seen in the previous subsection, the refractive indexof the macrogel decreases when it absorbs water (in this casewater vapor). This leads to a dielectric contrast increase for thephotonic crystal (being the refractive index of the polystyreneindependent on the water concentration). As known [26,27], boththe full width at half-maximum (FWHM) of the stop band andthe peak depth increase with higher dielectric contrasts, thusexplaining the behavior of the curves in Fig. 9(a).

The effect of progressive concentration of the atmosphere in eth-anol vapors on the optical transmission spectra was investigated ina range of high partial pressure values. In particular, in Fig. 10(a andb), the spectra and the calculated swelling ratio curve are reported.The presence of ethanol vapors in the atmosphere induces a redshift of the Bragg peak wavelength, due to further swelling of thehydrogel. A marked initial drop of the whole transmission curve,with respect to its position in the ‘‘equilibrium’’ conditions attainedin the presence of water vapors only, is observed. Transmittance isalmost entirely recovered with time. Concomitantly, there is a sig-nificant increase in band gap depth. The reduction in transmittance,also in this case, is due to an increase of the roughness of the thinhydrogel layer present on the opal surface, when the sample is

Fig. 9. (a and b) Transmission spectra (a), and swelling degree values (b) as function of the time for the hydrogel-infiltrated photonic crystal exposed to water vapor saturatedair at 24 �C. The ‘‘0 min’’ curve or data point refers to the sample equilibrated in air at 24 �C and 45% RH.

Fig. 10. (a and b) Transmission spectra (a) and swelling degree values (b) as function of the time for the hydrogel-infiltrated photonic crystal exposed to air in equilibriumwith a 1:9 vol/vol ethanol/water mixture at 24 �C. The ref curve in plot (a) or the grey line in plots (b) correspond to the steady state condition for the system at 24 �C and100% RH.

Fig. 11. (a–c) Swelling degree values as function of the exposure time for different compositions of the vapor/air atmosphere. Values reported in the legend refer to the molarfraction of ethanol in air (a). Transmission spectra (b) and swelling degree values (c) for steady-state conditions at different concentrations of ethanol vapor in theatmosphere.

C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76 75

exposed to ethanol vapor. The main difference, if compared to thecase of water vapor exposition, is that the roughness is rapidlyattenuated to disappear in a short time. The swelling degree furtherincreases with respect to the equilibrium condition in water toreach a plateau at 1.125. The same observations concerning thedielectric contrast increase can also be applied for the sampleexposed to ethanol vapors. It is worth noticing that, although thedielectric contrast augments more than in the previous case, thisincrease is not enough to cause an overall blue-shift, being the con-tribution from swelling (red-shift) still dominant.

In Fig. 11(a), the swelling degrees as function of the time for thefour investigated concentrations of vapors in the atmosphere areshown. While the perturbation of the concentration in the vaporphase is appreciated as a sudden and significant modification ofthe transmission spectra, the steady state conditions establishwithin few tens of minutes, the smaller is the relative variationin concentration the shorter is the time to reach ‘‘equilibrium’’.In the follow up of this research, the film exposed to the atmo-sphere will be the one where the polystyrene template has beenremoved. A significant reduction of the response time is expected

76 C. Dispenza et al. / Reactive & Functional Polymers 79 (2014) 68–76

as a result of the increase of surface to bulk ratio in the responsivematerial and related increase in polymer chains mobility.

In Fig. 11(b and c) the optical transmission curves and swellingratio values at the different ethanol mole fractions in vapor phasein the steady state conditions are shown. A linear response both inthe Bragg peak wavelength and in the swelling ratio, as a functionof the mole fraction of ethanol present in the atmosphere, isobserved. The linearity in the swelling degree suggests that thehydrogel, although it has equilibrated to the atmosphere which isexposed to, is still far from fully saturated conditions and it shouldfurther swell if exposed to a higher concentration of ethanolvapors. This result is quite remarkable, since the macrogel ana-logue was no longer swelling for a molar ratio of ethanol above21 � 10�3. The linearity of the shift of Bragg peak wavelength ofthe diffracted light is directly related to the linearity of the swellingdegree modification in this concentration range, since the swellingprocess changes the mean separation between the colloidal parti-cles in the infiltrated opal and, thus, the Bragg diffraction peakposition.

The Bragg shifts associated to the hydrogel swelling are revers-ible. The time required by the mesoscopically structured film toregain its structural color (green) after exposure to ethanol vaporsresulted to be less than 1 min.

4. Conclusions

Networks of 2-hydroxyethyl methacrylate-co-acrylic acid(HEMA-co-AA) were synthetized to obtain hydrogels that are ableto uptake water but that show significant incremental swellingwhen exposed to ethanol, after being equilibrated with water.The swelling behavior in the presence of other organic solvents,such as methanol or acetone, was significantly different. Acetone,in particular, caused deswelling of the hydrogel, thus suggestinga potential selectivity of these materials for the detection of etha-nol. When HEMA-co-AA hydrogels were polymerized within theinterstitial spaces of a polystyrene colloidal crystal, mesoscopicallystructured, periodic films were obtained, which Bragg diffractedlight in the visible portion of the spectrum. An optical characteriza-tion of the hydrogel infiltrated photonic crystals, when exposed towater vapor phases and at different water/ethanol concentrations,showed significant Bragg diffraction peak shifts. The film exposedto ethanol vapor changes color from iridescent green to red andit reverts to green when no longer exposed to ethanol. Theresponse time of the mesoscopically structured thin film is signif-icantly shorter than for the macrogel analogue (minutes as oppo-site to days) and it can be likely reduced by removal of the

polystyrene template, as it will be shown in the follow up of thisresearch.

Acknowledgements

The activity is partially funded by National Ministry of Univer-sity with the Project PRIN 2008 ‘‘Studio di fattibilità e sviluppo pro-totipale di sensori elettro-ottici per la misura di BrAC’’, prot.2008F7P4MX.

References

[1] R. Langer, D.A. Tirrell, Nature 428 (2004) 487–492.[2] I. Tokarev, S. Minko, Soft Matter. 5 (2009) 511–524.[3] H. Kim, R.E. Cohen, P.T. Hammond, D.J. Irvine, Adv. Funct. Mater. 16 (2006)

1313–1323.[4] I. Yoshimura, Y. Miyahara, N. Kasagi, H. Yamane, A. Ojida, I. Hamachi, J. Am.

Chem. Soc. 126 (39) (2004) 12204–12205.[5] C. Ruan, K. Zeng, C.A. Grimes, Anal. Chim. Acta 497 (2003) 123–131.[6] F. Lamberti, S. Giulitti, M. Giomo, N. Elvassore, J. Mater. Chem. B 1 (38) (2013)

5083–5091.[7] L.A. Haines, K. Rajagopal, B. Ozbas, D.A. Salick, D.J. Pochan, J.P. Schneider, J. Am.

Chem. Soc. 127 (2005) 17025–17029.[8] T. Miyata, T. Uragami, K. Nakamae, Adv. Drug Deliver. Rev. 54 (2002) 79–98.[9] D. Klinger, K. Landfester, Polymer 53 (23) (2012) 5209–5231.

[10] B. Ziaiea, A. Baldi, M. Leia, Y. Guc, R.A. Siegel, Adv. Drug Deliver. Rev. 56 (2)(2004) 145–172.

[11] D. Kuckling, Colloid Polym. Sci. 287 (2009) 881–891.[12] J. Shin, P.V. Braun, W. Lee, Sens. Actuators B 150 (2010) 183–190.[13] Y. Nishijima, K. Ueno, S. Juodkazis, V. Mizeikis, H. Misawa, T. Tanimura, K.

Maeda, Opt. Express 15 (2007) 12979–12988.[14] J.H. Holtz, S.A. Asher, Nature 389 (1997) 829–832.[15] J.M. Weissman, H.B. Sunkara, A.S. Tse, S.A. Asher, Science 274 (1996) 959–960.[16] M.M.W. Muscatello, S.A. Asher, Adv. Funct. Mater. 18 (2008) 1–8.[17] J.T. Zhang, X. Chao, X. Liu, S.A. Asher, Chem. Commun. 49 (2013) 6337–6339.[18] Y. Hu, X. Jiang, L. Zhang, J. Fan, W. Wu, Biosens. Bioelectron. 48 (2013) 94–99.[19] J. Wang, Y. Cao, Y. Feng, F. Yin, J. Gao, Adv. Mater. 19 (22) (2007) 3865–3871.[20] X. Xu, A.V. Goponenko, S.A. Asher, J. Am. Chem. Soc. 130 (10) (2008) 3113–

3119.[21] R. Pernice, G. Adamo, S. Stivala, A. Parisi, A.C. Busacca, D. Spigolon, M.A.

Sabatino, L. D’Acquisto, C. Dispenza, Opt. Mater. Express 3 (11) (2013) 1820–1833.

[22] C. Dispenza, M. Ricca, C. Lopresti, G. Battaglia, M. La Valle, D. Giacomazza, D.Bulone, Polym. Chem. 2 (1) (2011) 192–202.

[23] F. Marlow, P. Sharifi Muldarisnur, R. Brinkmann, C. Mendive, Angew. Chem. Int.Ed. 48 (2009) 6212–6233.

[24] G. Adamo, D. Agrò, S. Stivala, A. Parisi, C. Giaconia, A. Busacca, M.C. Mazzillo, D.Sanfilippo, P.G. Fallica, Proc. SPIE 8629 (2013) 86291A.

[25] M. Ricca, V. Foderà, D. Giacomazza, M. Leone, G. Spadaro, C. Dispenza, Colloid.Polym. Sci. 288 (2010) 969–980.

[26] V. Morandi, F. Marabelli, V. Amendola, M. Meneghetti, D. Comoretto, Adv.Funct. Mater. 17 (15) (2007) 2779–2786.

[27] D.M. Mittleman, J.F. Bertone, P. Jiang, K.S. Hwang, V.L. Colvin, J. Chem. Phys.111 (1) (1999) 345–354.

[28] S.N. Kasarova, N.G. Sultanova, C.D. Ivanov, I.D. Nikolov, Opt. Mater. 29 (11)(2007) 1481–1490.