Molecularly Imprinted Polymer Nanoparticles for ... - Semantic Scholar
sensors Article
Molecularly Imprinted Polymer Nanoparticles for Formaldehyde Sensing with QCM Munawar Hussain, Kira Kotova and Peter A. Lieberzeit * Faculty for Chemistry, Department of Physical Chemistry, University of Vienna, Währinger Strasse 42, A-1090 Vienna, Austria; [email protected] (M.H.); [email protected] (K.K.) * Correspondence: [email protected]; Tel.: +43-1-4277-52341 Academic Editor: Franz L. Dickert Received: 23 March 2016; Accepted: 22 June 2016; Published: 30 June 2016
Abstract: Herein, we report on molecularly imprinted polymers (MIPs) for detecting formaldehyde vapors in air streams. A copolymer thin film consisting of styrene, methacrylic acid, and ethylene glycol dimethacrylate on quartz crystal microbalance (QCM) yielded a detection limit of 500 ppb formaldehyde in dry air. Surprisingly, these MIPs showed specific behavior when tested against a range of volatile organic compounds (VOCs), such as acetaldehyde, methanol, formic acid, and dichloromethane. Despite thus being a suitable receptor in principle, the MIPs were not useful for measurements at 50% humidity due to surface saturation by water. This was overcome by introducing primary amino groups into the polymer via allyl amine and by changing the coating morphology from thin film to nanoparticles. This led to the same limit of detection (500 ppb) and selectivity as before, but at the real-life conditions of 50% relative humidity. Keywords: molecularly imprinted polymer; quartz crystal microbalance; formaldehyde detection; MIP nanoparticles
1. Introduction Formaldehyde, a pungent-smelling gas, is regarded one of the main toxic indoor pollutants [1–3] and increasingly draws attention to itself also in outdoor environments [4]. It is carcinogenic and causes a range of conditions including central nervous system damage, immune system disorders, respiratory diseases, and so-called sick building syndrome [5]. Household products and building materials [1,2] are the main indoor sources of formaldehyde. All developed countries have laws to limit the maximum concentration of formaldehyde in building materials. Levels in indoor air are typically in the range of 1–100 ppb (or even higher) depending on formaldehyde source and ventilation [6]. In industry, formaldehyde serves in many processes, e.g., as wood fixative, in paints, insulation foams, dry cleaning solutions, detergents, cosmetics, and pharmaceuticals [7,8]. The World Health Organization (WHO) and the National Institute for Occupational Safety and Health (NIOSH) have set a permissible long-term exposure limit of 0.08- and 1-ppm formaldehyde, respectively. Given the need to monitor formaldehyde concentration in air, portable, cost-effective, robust detectors are highly desirable and have attracted substantial scientific interest, as summarized in a recent review [9]. Many of those sensors are based on inorganic oxidic materials with optimized affinity, but optical approaches have also been published [10]. For introducing shape selectivity, molecularly imprinted polymers (MIPs) [11,12] have attracted increasing attention. In MIP synthesis, the potential analyte acts as a template. Self-assembly of the monomers around this template leads to adapted cavities in the polymer matrix. MIPs have been developed for a wide range of species including small molecules/VOC [13], proteins [14–16], and their aggregates [17] up to entire microorganisms [18]. Inherently, there is hardly a size limit for templates. Hence, metal ions are the smallest species for which MIPs have been reported [19,20]. Despite of course being much larger than a metal ion, the formaldehyde Sensors 2016, 16, 1011; doi:10.3390/s16071011
www.mdpi.com/journal/sensors
Sensors 2016, 16, 1011
2 of 9
a metal ion, the formaldehyde molecule is very small compared to other VOC. This is also reflected Sensors 2016, 16, 1011 2 of 9 in the fact that few articles have been published on formaldehyde imprinting so far: The earliest paper on formaldehyde MIPs combined with quartz crystal microbalance (QCM) detection was published in molecule 2005 [21]. is It very reports appreciable a static measuring asthat wellfew as selectivity small compareddetection to other limits VOC. in This is also reflected chamber in the fact articles factors of about four on against larger aldehydes. Another on pre-concentrating have been published formaldehyde imprinting so far: The paper earliestreported paper on formaldehyde MIPs combined with crystal microbalance (QCM) detection was published in 2005 [21]. It reports formaldehyde on quartz MIPs for electrophoretic separation [22]. Furthermore, fluoral–polyaniline double appreciable detection in a static measuring as well as selectivity factorsisofnot about fouron layers have been used limits for formaldehyde sensingchamber [23], even though this approach based against larger aldehydes. Another paper reported on pre-concentrating formaldehyde on MIPs for molecular imprinting. Two more recent publications in the field of MIP-based formaldehyde sensing electrophoretic separation [22]. Furthermore, fluoral–polyaniline double layers have been used for rely on more complex materials for recognition: One is based on a MIP-Ag/LaFeO 3 nanoparticle formaldehyde sensing [23], even though this approach is not based on molecular imprinting. Two more composite [24], the other one on layer-by-layer assembly of gold nanoparticles and PMAA structures recent in the field MIP-based formaldehyde on moretocomplex materials [25]. Thepublications former article does notofstate any selectivity, whichsensing makesrely it difficult assess the effect of for recognition: One is based on a MIP-Ag/LaFeO nanoparticle composite [24], the other one onon 3 imprinting. The latter indeed reports appreciable imprinting effects. Nonetheless, both systems rely layer-by-layer assembly of gold nanoparticles and PMAA structures [25]. The former article does not polymer–nanoparticle composites. Those pre-concentrate formaldehyde on the surfaces of affinity state any selectivity, which makes it difficult to assess the effect of imprinting. The latter indeed reports materials, which is an approach that has proven useful also for other volatile pollutants [26]. Herein, appreciable imprinting effects. Nonetheless, both systems rely on polymer–nanoparticle composites. we report the design of single-phase recognition materials based on MIPs for assessing formaldehyde Those pre-concentrate formaldehyde on the surfaces of affinity materials, which is an approach that vapor concentrations in flow systems for direct gas phase detection of this analyte. has proven useful also for other volatile pollutants [26]. Herein, we report the design of single-phase recognition materials based on MIPs for assessing formaldehyde vapor concentrations in flow systems 2. Materials and Methods for direct gas phase detection of this analyte.
We purchased all chemicals from Sigma-Aldrich and VWR, respectively, in the highest available
2. Materials Methods purity and keptand them in storage according to manufacturer recommendations.
We purchased all chemicals from Sigma-Aldrich and VWR, respectively, in the highest available
2.1.purity MIP and Synthesis kept them in storage according to manufacturer recommendations. structural 2.1.The MIP Synthesisformulae of all monomers and reagents are summarized in Figure 1. To synthesize
MIPs, we dissolved 13 mg (0.12 mmol) styrene, 13 mg (0.15 mmol) methacrylic acid, and 60 mg (0.31 structural formulae of all monomers and reagents are summarized synthesize mmol) The ethylene glycol dimethacrylate (EGDMA) in a reaction tube using in 500Figure µL of1.aTo binary solvent MIPs, we dissolved 13 mg (0.12 mmol) styrene, 13 mg (0.15 mmol) methacrylic acid, and 60 mg mixture comprised of 200 µL of methanol and 300 µL of dimethyl formamide (DMF) and added 15 (0.31 mmol) ethylene glycol dimethacrylate (EGDMA) in a reaction tube using 500 µL of a binary solvent µL of an aqueous formaldehyde solution (37%–38% w/w, stabilized with 10% methanol). If the mixture comprised of 200 µL of methanol and 300 µL of dimethyl formamide (DMF) and added 15 µL polymer contained primary amino groups for recognition (see Section 3.2), the reaction mixture also of an aqueous formaldehyde solution (37%–38% w/w, stabilized with 10% methanol). If the polymer contained 13 mg (0.26 mmol) allyl amine before adding the template. After obtaining homogeneous, contained primary amino groups for recognition (see Section 3.2), the reaction mixture also contained transparent solutions, we added 6 mg of azobisisobutyronitrile (AIBN) as the radical initiator. 13 mg (0.26 mmol) allyl amine before adding the template. After obtaining homogeneous, transparent Polymerization took place by UV irradiation (λmax = 360 nm and 210 W) for 80 minutes until reaching solutions, we added 6 mg of azobisisobutyronitrile (AIBN) as the radical initiator. Polymerization took theplace gel by point of the mixture. by point the same UV irradiation (λmax =Non-imprinted 360 nm and 210polymers W) for 80 (NIPs) minuteswere until synthesized reaching the gel of procedure, replacing the formaldehyde solution with 9 µL of distilled water. Hence, the solvent the mixture. Non-imprinted polymers (NIPs) were synthesized by the same procedure, replacing the mixtures of bothsolution MIPs and NIPs the same of water, 2% (9and µL NIPs of ~500 formaldehyde with 9 µLcontain of distilled water.amount Hence, the solventnamely, mixturesroughly of both MIPs µL). For sensor measurements, 15namely, µL of this oligomer mixture spin-coated onto one electrode contain the same amount of water, roughly 2% (9 µL of ~500were µL). For sensor measurements, 15 µL pair of a dual-electrode QCM at 4000 rpm on a custom-made spin-coater, while the second pair was of this oligomer mixture were spin-coated onto one electrode pair of a dual-electrode QCM at 4000 rpm spin-coated with NIPs at the same speed. on a custom-made spin-coater, while the second pair was spin-coated with NIPs at the same speed.
Figure reagentsused. used. Figure1. 1. Monomers Monomers and reagents
Sensors 2016, 16, 1011
3 of 9
SensorsFor 2016,generating 16, 1011
3 ofof9 MIP NPs, 500 µL of a MIP oligomer solution was pipetted into 10 mL acetonitrile (AcCN) under vigorous stirring and was then kept on a magnetic stirrer for 12 h. The resulting homogenous NP 500 suspensions were centrifuged rpminto for 10ten The For generating MIP NPs, µL of a MIP oligomer solutionat was4000 pipetted mLminutes. of acetonitrile supernatant was removed by pipetting. Then, we re-suspended the particles in 500 µL of AcCN and (AcCN) under vigorous stirring and was then kept on a magnetic stirrer for 12 h. The resulting used this setup to suspensions coat devices.were For that purpose,at 5 µL of rpm the NP was spin-coated at 4000 homogenous NP centrifuged 4000 for suspension ten minutes. The supernatant was rpm onto the respective QCM electrode. The spin-coated QCM were kept overnight for drying prior removed by pipetting. Then, we re-suspended the particles in 500 µL of AcCN and used this setup to sensitive For measurements and5atomic microscopy (AFM) It was necessary to to mass coat devices. that purpose, µL of force the NP suspension was studies. spin-coated atnot 4000 rpm onto further cross-link stabilizeThe thespin-coated NP layers,QCM because proved stable during gas-phase the respective QCMorelectrode. were they kept overnight for drying prior to mass measurements. sensitive measurements and atomic force microscopy (AFM) studies. It was not necessary to further Surface morphology of NP layers was recorded withstable a VEECO Instruments Nanoscope IVa in cross-link or stabilize the NP layers, because they proved during gas-phase measurements. contact mode with silicon nitride tips. suspensions directly spin-coated onto IVa QCM Surface morphology of NP layers wasNP recorded with a were VEECO Instruments Nanoscope in electrodes as described above and assessed after drying overnight. contact mode with silicon nitride tips. NP suspensions were directly spin-coated onto QCM electrodes
as described above and assessed after drying overnight. 2.2. QCM Manufacturing and Measuring Setup 2.2. QCM andstructures Measuring(fSetup QCMManufacturing dual electrode 0 = 10 MHz) were screen printed with brilliant gold paste containing gold colloid (GGP 2093126, GmbH, Hanau, Germany) onto AT-cut QCM 12.5% dual electrode structures (f0 = 10Heraeus MHz) were screen printed with brilliant goldquartz paste substrates (13.8 mm in diameter, 168-µm-thick, purchased from Great Microtama Industries, containing 12.5% gold colloid (GGP 2093126, Heraeus GmbH, Hanau, Germany) onto AT-cut quartz Surabaya, Indonesia). After that, we burned purchased the structures 400 Microtama °C for 4 hIndustries, to removeSurabaya, organic substrates (13.8 mm in diameter, 168-µm-thick, from at Great ˝ components and reveal the blank gold electrodes. We then spin-coated the respective recognition Indonesia). After that, we burned the structures at 400 C for 4 h to remove organic components material onto electrode as described The QCM the wasrespective then mounted in a custom-made and reveal the an blank gold electrodes. We above. then spin-coated recognition material ontogas an cell (see Figure 2B) and connected to was a custom-made circuit. An Agilent HP5313A electrode as described above. The QCM then mountedoscillating in a custom-made gas cell (see Figure 2B) frequency counter out the resonance frequency of the oscillator. Frequency as acounter functionread of time and connected to aread custom-made oscillating circuit. An Agilent HP5313A frequency out was recorded through an Agilent GPIB/USB adapter by a custom-made LabView routine. the resonance frequency of the oscillator. Frequency as a function of time was recorded throughGas an samplesGPIB/USB were produced bybya agas mixing apparatus of mass flow valves (type by RS-485) Agilent adapter custom-made LabViewconsisting routine. Gas samples were produced a gas addressed by a digital controller (Brooks Instruments SLA5850 S, Hatfield, PA, USA). Defined gas mixing apparatus consisting of mass flow valves (type RS-485) addressed by a digital controller (Brooks streams were bubbled through PA, gasUSA). washing flasks de-ionized water and analyte, Instruments SLA5850 S, Hatfield, Defined gascontaining streams were bubbled through gas washing respectively. The entire setup can be seen in Figure 2A. flasks containing de-ionized water and analyte, respectively. The entire setup can be seen in Figure 2A.
Figure 2. (A) airair streams with defined humidity andand gasgas concentration; (B) (A)Setup Setupused usedfor forgenerating generating streams with defined humidity concentration; Quartz crystal microbalance (QCM) mounted in measuring cell.cell. (B) Quartz crystal microbalance (QCM) mounted in measuring
3. 3. Results Resultsand andDiscussion Discussion 3.1. Development of MIP Thin Films Previous work has shown that acrylate-based acrylate-based polymers polymers are are suitable suitable for for formaldehyde formaldehyde MIPs MIPs[21]. [21]. However, formaldehydeis is inherently lipophilic. Hence, our was ideatowas to reduce of the However, formaldehyde inherently lipophilic. Hence, our idea reduce polaritypolarity of the polymer polymer system, but increase polarizability to achieve better interaction between polymer and system, but increase polarizability to achieve better interaction between polymer and analyte. For that analyte. For that purpose, we introduced styrene as a3 shows monomer. Figure 3 shows thefor QCM sensor purpose, we introduced styrene as a monomer. Figure the QCM sensor results a MIP thin film and the respective NIP when exposed to different vapor concentrations of formaldehyde. Evidently,
Sensors 2016, 1011 Sensors 2016, 16,16, 1011
4 of 4 of 99
Sensors 2016, 16, 1011
4 of 9
resultsfor fora aMIP MIPthin thinfilm filmand andthe therespective respectiveNIP NIPwhen whenexposed exposedtotodifferent differentvapor vaporconcentrations concentrationsofof results formaldehyde.Evidently, Evidently,the theMIP MIPyield yieldsubstantially substantiallyhigher higherfrequency frequencyshifts shiftsatatallallconcentrations concentrations formaldehyde. the MIP yield substantially higher frequency shifts at all concentrations than the corresponding NIP. than the corresponding NIP. The resulting imprinting effect—i.e., the ratio between MIPs andNIPs— NIPs— than the corresponding NIP. The resulting imprinting effect—i.e., the ratio between MIPs and The resulting imprinting effect—i.e., the ratio between MIPs and NIPs—reaches a factor of four, which is reachesa afactor factorofoffour, four,which whichisisappreciably appreciablylarge largegiven giventhe thesmall smallsize sizeofofthe theformaldehyde formaldehydemolecule. molecule. reaches appreciably large given the small size of the formaldehyde molecule. Additionally, the magnitude of Additionally, the magnitude the frequency shiftdepends depends formaldehyde concentration. Overall, Additionally, the magnitude ofofthe frequency shift ononformaldehyde concentration. Overall, the frequency shift depends on formaldehyde concentration. Overall, the setup thus leads to a sensor the setup thus leads to a sensor characteristic with a lower limit of detection (LoD) of 1 ppm (S/N the setup thus leads to a sensor characteristic with a lower limit of detection (LoD) of 1 ppm (S/N characteristic with a lowerall limit of detection (LoD) ofare 1are ppm (S/N ratio ěwithin 3). Furthermore, alloffrequency ratio≥ ≥3).3).Furthermore, Furthermore, allfrequency frequency responses fully reversible within fewtens tens ofseconds. seconds. ratio responses fully reversible a afew responses areinfully reversible within a few tens ofthe seconds. Especially, in QCM gas phase measurements, Especially, inQCM QCM gasphase phase measurements, thedifference difference signal between MIPs andNIPs NIPscan can Especially, gas measurements, ininsignal between MIPs and the difference in signal between MIPs and NIPs can clearly be attributed to mass effects caused by the clearlybebeattributed attributedtotomass masseffects effectscaused causedbybythe theincorporation incorporationofofformaldehyde formaldehydemolecules moleculesinto into clearly incorporation of formaldehyde molecules into recognition sites within the polymer matrix. This again isthe recognitionsites sites withinthe thepolymer polymermatrix. matrix. Thisagain again remarkable giventhe thesmall small size recognition within This isisremarkable given size ofofthe remarkable given the small size of the formaldehyde molecule. formaldehyde molecule. formaldehyde molecule.
Figure 3.QCM QCM sensor responses molecularly imprinted polymer (MIP) andnon-imprinted non-imprinted Figure sensor responses ofofmolecularly imprinted polymer (MIP) and Figure 3.3.QCM sensor responses of molecularly imprinted polymer (MIP) and non-imprinted polymer polymer (NIP), respectively, pulses containing different vapor concentrations formaldehyde. (NIP), respectively, to pulses containing different vapor concentrations of formaldehyde. polymer (NIP), respectively, toto pulses containing different vapor concentrations ofof formaldehyde.
Despite these appreciable results, thin film sensors were notdeveloped developed further after turned Despite these appreciable results, thin film sensors were further it itturned Despite these appreciable results, thin film sensors were notnot developed further afterafter it turned out out that they did not yield useful results in humid air, which makes them useless for operation out that they did not yield useful results in humid air, which makes them useless for operation inin that they did not yield useful results in humid air, which makes them useless for operation in real-life real-life conditions. One of the possible reasons of such a response is indicated in Figure 4: switching real-life conditions. One of the possible reasons of such a response is indicated in Figure 4: switching conditions. One of the possible reasons of such a response is indicated in Figure 4: switching from from50% 50%rH rHtotodry dryairairchanges changesthe thefrequency frequencyofofthe theMIP-coated MIP-coatedchannel channel byalmost almost1400 1400Hz. Hz.It Itisis from 50% rH to dry air changes the frequency of the MIP-coated channel by almostby 1400 Hz. It is therefore thereforehighly highlyprobable probablethat thatwater watermolecules moleculesblock blockaccess accesstotoformaldehyde formaldehydebinding bindingsites siteswithin withinthe the therefore highly probable that water molecules block access to formaldehyde binding sites within the MIP thin MIP thin film. Nonetheless, the MIP thin films serve as proof of the principle that appreciable MIP Nonetheless, thin film. Nonetheless, thefilms MIPserve thin as films serve asprinciple proof ofthat the appreciable principle that appreciable film. the MIP thin proof of the formaldehyde formaldehydesensor sensorresponses responsescan canbebeachieved achievedbybythis thisstrategy. strategy. formaldehyde sensor responses can be achieved by this strategy.
Figure 4.Sensor Sensor responses ofMIP MIP and NIP, respectively, when switching from stream with Figure 4.4.Sensor responses ofofMIP and NIP, respectively, when switching fromfrom an air stream with with 50% Figure responses and NIP, respectively, when switching anan airairstream 50% rH to dry air. rH torH dry 50% toair. dry air.
Sensors 2016, 16, 1011 Sensors 2016, 16, 1011
5 of 9 5 of 9
3.2. Generating 3.2. Generating MIP MIP Nanoparticles Nanoparticles Using MIP MIP NPs NPs seems seems aa rational rational approach approach to to address address these these challenges. challenges. MIP have drawn drawn Using MIP NPs NPs have increasing attention, attention, because because they they often often yield increasing yield higher higher sensitivity sensitivity and and selectivity selectivity [27,28] [27,28] than than thin thin films, films, partly due to shorter average diffusion pathways. Even at higher humidity, formaldehyde molecules partly due to shorter average diffusion pathways. Even at higher humidity, formaldehyde molecules should be be able abletotoreach reachbinding bindingsites sites within polymer matrix. purpose, we synthesized should within thethe polymer matrix. ForFor thatthat purpose, we synthesized MIP MIPbyNP by precipitation, because this approach rather straightforward does not require NP precipitation, because this approach is ratherisstraightforward and does and not require additional additionalsuch reagents, such as surfactants. Figure 5 shows an AFM image ofMIP a resulting NP layer reagents, as surfactants. Figure 5 shows an AFM image of a resulting NP layerMIP deposited on deposited on a glass substrate. a glass substrate.
Figure 5. Atomic force microscopy (AFM) image of formaldehyde MIP nanoparticles. Figure 5. Atomic force microscopy (AFM) image of formaldehyde MIP nanoparticles.
Evidently, the surface is homogeneously covered with NPs having diameters in the range of 80–150 nm. Precipitation hence was successful and led to the desired particulate particulate material. material. However, However, such particles did not yield appreciable responses at elevated humidity. Obviously, the non-specific adsorption molecules ontoonto the surface and selective incorporation of formaldehyde molecules adsorption ofofwater water molecules the surface and selective incorporation of formaldehyde compete with one another. Hence, greater affinity between thebetween polymerthe andpolymer formaldehyde is necessary, molecules compete with one another. Hence, greater affinity and formaldehyde which can be which achieved themodifying polymer [29]. is well-known from medical and biological is necessary, canby bemodifying achieved by the It polymer [29]. It is well-known from medical sciences that formaldehyde strongly interacts strongly with primary amino groups (this phenomenon is used and biological sciences that formaldehyde interacts with primary amino groups (this for stabilizing is tissue mortemtissue for further investigation) by further forminginvestigation) imines (Schiff by bases) with phenomenon usedetc. forpost stabilizing etc. post mortem for forming primary amino bases) groupswith of proteins. the reaction is driven to the end product, this coursetoleads to a imines (Schiff primaryIf amino groups of proteins. If the reaction is of driven the end covalent bond between C of the(double) carbonylbond groupbetween and the the N ofCthe group. group Hence,and we product, (double) this of course leads to the a covalent of amino the carbonyl decided primary amino our material via allyl amine, eveninto though sensing the N of to theintroduce amino group. Hence, wegroups decidedinto to introduce primary amino groups our material of aims ateven reversible Ab-initio ofinteractions. Hall and Smith [30] revealed that viacourse allyl amine, thoughinteractions. sensing of course aimscalculations at reversible Ab-initio calculations the energy between thethat amine-aldehyde adduct forming in the the first step of the reaction of Hall anddifference Smith [30] revealed the energy difference between amine-aldehyde adduct ´1 for methyl amine and formaldehyde in the absence and the imine is ´46 of water. However, forming in the first kJ¨mol step of the reaction and the imine is −46 kJ∙mol−1 for methyl amine and ´ 1 −1 the energy of the transition state is +112.3 kJ¨mol theabove theofenergy of the amine-aldehyde adduct. formaldehyde in the absence of water. However, energy the transition state is +112.3 kJ∙mol Such barrier minimizes the adduct. probability actual iminebarrier formation, especially in the gas abovehigh-energy the energy of the amine-aldehyde Suchofhigh-energy minimizes the probability phase and imine considering the low amountsin of the formaldehyde expected inthe indoor Thus, the of actual formation, especially gas phase vapors and considering lowair. amounts of interactions between and the respective should still be reversible. formaldehyde vaporsformaldehyde expected in indoor air. Thus, theMIP interactions between formaldehyde and the Adding allyl amine did change morphology of the nanoparticles: AFM characterization of respective MIP should still benot reversible. such Adding NP layers showed in the rangeof of the around 100–150 nm, which is the same size allyl amineparticle did notdiameters change morphology nanoparticles: AFM characterization of as forNP polymers withoutparticle allyl amine. However, additional monomer lead which to highly appreciable such layers showed diameters in the the range of around 100–150 nm, is the same size sensor effects, as without can be seen Figure However, 6. Two aspects are immediately noticeable: material as for polymers allylinamine. the additional monomer lead to Firstly, highly the appreciable indeed leads toassensitive frequency responses towardare formaldehyde despite the fact that the sensor effects, can be seen in Figure 6. Two aspects immediatelyvapors noticeable: Firstly, the material indeed leads to sensitive frequency responses toward formaldehyde vapors despite the fact that the corresponding MIP thin films (data not shown) did not give rise to measureable effects. Secondly,
Sensors 2016, 16, 1011 Sensors 2016, 16, 1011
6 of 9 6 of 9
sensors based on MIP NPs respond faster analyte pulses compared with thin films. corresponding MIP thin films (data substantially not shown) did nottogive rise to measureable effects. Secondly, According to the data shown in Figure 6, equilibrium is reached within roughly half a minute for sensors based on MIP NPs respond substantially faster to analyte pulses compared with thin films. each vaporto pulse. In theshown case ofinMIP thin6,films (see Figure 3), the same process lastshalf for around three According the data Figure equilibrium is reached within roughly a minute for minutes for each pulse. Again, this strongly suggests that accessibility of the respective binding sites each vapor pulse. In the case of MIP thin films (see Figure 3), the same process lasts for around three in the material a seminal rolestrongly in sensing. In thethat case of nanoparticles with a roughly 100-nm minutes for eachplays pulse. Again, this suggests accessibility of the respective binding sites diameter, possible diffusion pathways are in the range of at least several tens of nanometers. Thin in the material plays a seminal role in sensing. In the case of nanoparticles with a roughly 100-nm films, however, arediffusion typically pathways 100–200 nm high on each (asseveral determined bynanometers. network analyzer diameter, possible are in the rangeelectrode of at least tens of Thin measurements) and have a diameter of roughly 6 mm (electrode diameter after screen printing is 5 films, however, are typically 100–200 nm high on each electrode (as determined by network analyzer mm, and the layers are somewhat Thus, both the accessible surface is much lower is than for measurements) and have a diameter wider). of roughly 6 mm (electrode diameter after screen printing 5 mm, NP, and diffusion pathways within the polymer can be expected to be much longer. The MIP NP and the layers are somewhat wider). Thus, both the accessible surface is much lower than for NP, and sensor reaches LoDwithin = 0.5 ppm formaldehyde vapor intohumid air,longer. whichThe is the same as thatreaches for the diffusion pathways the polymer can be expected be much MIP NP sensor thin film at dry conditions. During all our measurements (same quartz over several days), LoD = 0.5 ppm formaldehyde vapor in humid air, which is the same as that for the thin film sensor at dry signals returned toall their baselines, strongly indicating reversible—i.e., non-covalent— conditions. During our respective measurements (same quartz over several days), sensor signals returned to binding. their respective baselines, strongly indicating reversible—i.e., non-covalent—binding.
Figure allyl amine towards formaldehyde at Figure 6. 6. QCM QCM responses responsesof ofMIP MIPnanoparticles nanoparticles(NPs) (NPs)containing containing allyl amine towards formaldehyde 50% rH.rH. at 50%
3.3. Selectivity Of course, sensitivity alone is not sufficient to claim successful imprinting. For that, selectivity must also be assessed. Table 1 summarizes the outcome of these experiments for both thin films and nanoparticle layers toward 100-ppm pulses of a range of compounds each. Table 1. Sensor responses of MIP thin films and NPs (containing allyl amine in the monomer mixture) at 0% rH and 50% rH, respectively, toward 100-ppm formaldehyde (HCHO), dichloromethane (CH2Cl2), methanol (MeOH), formic acid, acetone ethanol (EtOH), acetaldehyde (AcCHO), and acetonitrile (AcCN). Sensor
HCHO
CH2 Cl2
MeOH
Formic Acid
Acetone
EtOH
AcCHO
AcCN
Thin film (0% rH)
´65 ˘ 2 Hz
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
NP (50% rH)
´24 ˘ 1 Hz
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
Selectivity patterns of MIPsomewhat thin films and MIP NP signals. All Figure those 7. compounds are either similar inlayers size showing or are relative chemically related to formaldehyde: Methanol and formic acid are the reduction and oxidation product of formaldehyde, 3.3. SelectivityAcetaldehyde also contains the CHO group and acetone at least the carbonyl functionality. respectively. The dichloromethane molecule larger than formaldehyde. Surprisingly, none of these Of course, sensitivity aloneisisjust notslightly sufficient to claim successful imprinting. For that, selectivity compounds gave rise to any sensor response revealing specific behavior of the sensor, as all frequency must also be assessed. Table 1 summarizes the outcome of these experiments for both thin films and shifts stayed layers belowtoward the noise level pulses of 1 Hz the oscillator. For each. better visualization, Figure 7 nanoparticle 100-ppm of on a range of compounds summarizes thecompounds normalized are sensor responses for these twoinsystems: shifts in this All those either somewhat similar size or the arefrequency chemically related to
formaldehyde: Methanol and formic acid are the reduction and oxidation product of formaldehyde,
Sensors 2016, 16, 1011
7 of 9
case are related to the sensor response toward the 100-ppm formaldehyde. Error bars for formaldehyde responses correspond to standard deviations obtained for three parallel measurements. In the case Figure 6. QCM of MIP nanoparticles (NPs) containing allyl amine towards formaldehyde of frequency shiftsresponses that are not statistically significant, the noise level of the oscillator was used to at 50% rH. determine the errors.
Figure 7. 7. Selectivity Selectivitypatterns patterns of of MIP MIP thin thin films films and and MIP MIP NP NP layers layers showing showing relative relative signals. signals. Figure
3.3. Selectivity As can be seen, selectivity is the same in both cases, i.e., for thin films containing no allyl amine as well Of as nanoparticles containing amine. The data also successful represents imprinting. results obtained at two different course, sensitivity alone allyl is not sufficient to claim For that, selectivity humidity levels. Although of course not directly comparable, one can clearly see that the sensor must also be assessed. Table 1 summarizes the outcome of these experiments for both thin films and surprisingly reacted specifically towards formaldehyde: no statistically significant signal could be nanoparticle layers toward 100-ppm pulses of a range of compounds each. observed for anycompounds other VOC except formaldehyde during triplicate measurements of those compounds. All those are either somewhat similar in size or are chemically related to The comparably small size of the formaldehyde molecule combined with its very distinct functionality formaldehyde: Methanol and formic acid are the reduction and oxidation product of formaldehyde, may be the reason for such unusually high selectivity, which we previously only observed once in non-covalent MIPs [31]. Both sensitivity and selectivity results very strongly corroborate the outstanding abilities of the imprinting procedure: the MIPs can distinguish between molecules that are rather similar in size, such as formaldehyde and methanol. They also discriminate according to size: acetaldehyde hardly yields any effects on the MIP layer despite containing the same functional group as formaldehyde. 4. Conclusions We herein present a gas sensor for formaldehyde detection based on MIP, relying on both thin films and nanoparticles, respectively. It can detect vapor concentrations down to 500 ppb with outstandingly high selectivity. Optimal results are obtained with MIP NP. The latter point especially reaches beyond immediate science: MIP NP can be introduced into production processes of sensors fairly straightforwardly and thus, in principle, open up a way for commercializing such systems. Acknowledgments: This work has been funded by the European Commission within the seventh research framework EU-FP7, project number EU-FP7-NMP-2010-LARGE-4-263382 “PHOTOSENS,” which we gratefully acknowledge. Author Contributions: M.H. developed the polymer system and carried out thin film and initial nanoparticle experiments. Furthermore, he wrote the first draft of the manuscript. K.K. optimized MIP NP synthesis and carried out additional experiments with that matrix. P.A.L. supervised the work of both M.H. and K.K., gave feedback and ideas. He also guided M.H. in preparing the first draft of the manuscript and wrote the final version. Conflicts of Interest: The authors declare no conflict of interest.
Sensors 2016, 16, 1011
8 of 9
Abbreviations The following abbreviations are used in this manuscript: AcCN AFM AIBN DMF EGDMA MIP NIP NP QCM rH VOC
acetonitrile atomic force microscopy azoisobutyronitrile dimethyl formamide ethylene glycol dimethacrylate molecularly imprinted polymer non-imprinted polymer nanoparticles quartz crystal microbalance relative humidity volatile organic compound
References 1. 2.
3.
4. 5. 6.
7. 8. 9. 10.
11. 12. 13. 14. 15.
16. 17.
Pickrell, J.A.; Mokler, B.V.; Griffis, L.C.; Hobbs, C.H.; Bathija, A. Formaldehyde release rate coefficients from selected consumer products. Environ. Sci. Technol. 1983, 17, 753–757. [CrossRef] [PubMed] Pickrell, J.A.; Griffis, L.C.; Mokler, B.V.; Kanapilly, G.M.; Hobbs, C.H. Formaldehyde release from selected consumer products—Influence of chamber loading, multiple products, relative-humidity, and temperature. Environ. Sci. Technol. 1984, 18, 682–686. [CrossRef] Gunschera, J.; Mentese, S.; Salthammer, T.; Andersen, J.R. Impact of building materials on indoor formaldehyde levels: Effect of ceiling tiles, mineral fiber insulation and gypsum board. Build. Environ. 2013, 64, 138–145. [CrossRef] Salthammer, T. Formaldehyde in the ambient atmosphere: From an indoor pollutant to an outdoor pollutant? Angew. Chem. Int. Ed. 2013, 52, 3320–3327. [CrossRef] [PubMed] Salthammer, T.; Mentese, S.; Marutzky, R. Formaldehyde in the indoor environment. Chem. Rev. 2010, 110, 2536–2572. [CrossRef] [PubMed] Descamps, M.N.; Bordy, T.; Hue, J.; Mariano, S.; Nonglaton, G.; Schultz, E.; Tran-Thi, T.H.; Vignoud-Despond, S. Real-time detection of formaldehyde by a sensor. Sens. Actuators B Chem. 2012, 170, 104–108. [CrossRef] Wang, J.; Liu, L.; Cong, S.Y.; Qi, J.Q.; Xu, B.K. An enrichment method to detect low concentration formaldehyde. Sens. Actuators B Chem. 2008, 134, 1010–1015. [CrossRef] Wang, J.; Zhang, P.; Qi, J.Q.; Yao, P.J. Silicon-based micro-gas sensors for detecting formaldehyde. Sens. Actuators B Chem. 2009, 136, 399–404. [CrossRef] Chung, P.R.; Tzeng, C.T.; Ke, M.T.; Lee, C.Y. Formaldehyde gas sensors: A review. Sensors 2013, 13, 4468–4484. [CrossRef] [PubMed] Boersma, A.; van Ee, R.J.; Stevens, R.S.A.; Saalmink, M.; Charlton, M.D.B.; Pollard, M.E.; Chen, R.; Kontturi, V.; Karioja, P.; Alajoki, T. Detection of low concentration formaldehyde gas by photonic crystal sensor fabricated by nanoimprint process in polymer material. Proc. SPIE 2014, 9141. [CrossRef] Haupt, K.; Linares, A.V.; Bompart, M.; Bernadette, T.S.B. Molecularly imprinted polymers. Top. Curr. Chem. 2012, 325, 1–28. Li, S., Ge, Y., Piletsky, S.A., Lunec, J., Eds.; MIP-based sensors. In Molecularly Imprinted Sensors: Overview and Applications; Elsevier: Amsterdam, The Netherlands, 2012; pp. 339–354. Latif, U.; Rohrer, A.; Lieberzeit, P.A.; Dickert, F.L. Qcm gas phase detection with ceramic materials-VOCs and oil vapors. Anal. Bioanal. Chem. 2011, 400, 2457–2462. [CrossRef] [PubMed] Schirhagl, R.; Podlipna, D.; Lieberzeit, P.A.; Dickert, F.L. Comparing biomimetic and biological receptors for insulin sensing. Chem. Commun. 2010, 46, 3128–3130. [CrossRef] [PubMed] Wangchareansak, T.; Sangma, C.; Choowongkomon, K.; Dickert, F.; Lieberzeit, P. Surface molecular imprints of WGA lectin as artificial receptors for mass-sensitive binding studies. Anal. Bioanal. Chem. 2011, 400, 2499–2506. [CrossRef] [PubMed] Phan, N.V.H.; Sussitz, H.F.; Lieberzeit, P.A. Polymerization parameters influencing the QCM response characteristics of BSA MIP. Biosensors 2014, 4, 161–171. [CrossRef] [PubMed] Chunta, S.; Suedee, R.; Lieberzeit, P.A. Low-density lipoprotein sensor based on molecularly imprinted polymer. Anal. Chem. 2016, 88, 1419–1425. [CrossRef] [PubMed]
Sensors 2016, 16, 1011
18. 19. 20. 21. 22.
23.
24.
25.
26. 27. 28. 29. 30. 31.
9 of 9
Hayden, O.; Lieberzeit, P.A.; Blaas, D.; Dickert, F.L. Artificial antibodies for bioanalyte detection—Sensing viruses and proteins. Adv. Funct. Mater. 2006, 16, 1269–1278. [CrossRef] Branger, C.; Meouche, W.; Margaillan, A. Recent advances on ion-imprinted polymers. React. Funct. Polym. 2013, 73, 859–875. [CrossRef] Bajwa, S.Z.; Dumler, R.; Lieberzeit, P.A. Molecularly imprinted polymers for conductance sensing of Cu2+ in aqueous solutions. Sens. Actuators B Chem. 2014, 192, 522–528. [CrossRef] Feng, L.; Liu, Y.J.; Zhou, X.D.; Hu, J.M. The fabrication and characterization of a formaldehyde odor sensor using molecularly imprinted polymers. J. Colloid. Interf. Sci. 2005, 284, 378–382. [CrossRef] [PubMed] Sun, H.; Lai, J.P.; Fung, Y.S. Simultaneous determination of gaseous and particulate carbonyls in air by coupling micellar electrokinetic capillary chromatography with molecular imprinting solid-phase extraction. J. Chromatogr. A 2014, 1358, 303–308. [CrossRef] [PubMed] Antwi-Boampong, S.; Peng, J.S.; Carlan, J.; BelBruno, J.J. A molecularly imprinted fluoral-p/polyaniline double layer sensor system for selective sensing of formaldehyde. IEEE Sens. J. 2014, 14, 1490–1498. [CrossRef] Zhang, Y.M.; Liu, Q.J.; Zhang, J.; Zhu, Q.; Zhu, Z.Q. A highly sensitive and selective formaldehyde gas sensor using a molecular imprinting technique based on Ag-LaFeO3 . J. Mater. Chem. C 2014, 2, 10067–10072. [CrossRef] Iqbal, N.; Afzal, A.; Mujahid, A. Layer-by-layer assembly of low-temperature-imprinted poly(methacrylic acid)/gold nanoparticle hybrids for gaseous formaldehyde mass sensing. RSC Adv. 2014, 4, 43121–43130. [CrossRef] Mustafa, G.; Lieberzeit, P.A. Molecularly imprinted polymer-Ag2 S nanoparticle composites for sensing volatile organics. RSC Adv. 2014, 4, 12723–12728. [CrossRef] Wackerlig, J.; Lieberzeit, P.A. Molecularly imprinted polymer nanoparticles in chemical sensing—Synthesis, characterisation and application. Sens. Actuators B Chem. 2015, 207, 144–157. [CrossRef] Poma, A.; Turner, A.P.F.; Piletsky, S.A. Advances in the manufacture of mip nanoparticles. Trends Biotechnol. 2010, 28, 629–637. [CrossRef] [PubMed] Kotova, K.; Hussain, M.; Mustafa, G.; Lieberzeit, P.A. MIP sensors on the way to biotech applications: Targeting selectivity. Sens. Actuators B Chem. 2013, 189, 199–202. [CrossRef] Hall, N.E.; Smith, B.J. High-level ab initio molecular orbital calculations of imine formation. J. Phys. Chem. A 1998, 102, 4930–4938. [CrossRef] Hussain, M.; Iqbal, N.; Lieberzeit, P.A. Acidic and basic polymers for molecularly imprinted folic acid sensors—QCM studies with thin films and nanoparticles. Sens. Actuators B Chem. 2013, 176, 1090–1095. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
Molecularly Imprinted Polymer Nanoparticles for Formaldehyde Sensing with QCM Munawar Hussain, Kira Kotova and Peter A. Lieberzeit * Faculty for Chemistry, Department of Physical Chemistry, University of Vienna, Währinger Strasse 42, A-1090 Vienna, Austria; [email protected] (M.H.); [email protected] (K.K.) * Correspondence: [email protected]; Tel.: +43-1-4277-52341 Academic Editor: Franz L. Dickert Received: 23 March 2016; Accepted: 22 June 2016; Published: 30 June 2016
Abstract: Herein, we report on molecularly imprinted polymers (MIPs) for detecting formaldehyde vapors in air streams. A copolymer thin film consisting of styrene, methacrylic acid, and ethylene glycol dimethacrylate on quartz crystal microbalance (QCM) yielded a detection limit of 500 ppb formaldehyde in dry air. Surprisingly, these MIPs showed specific behavior when tested against a range of volatile organic compounds (VOCs), such as acetaldehyde, methanol, formic acid, and dichloromethane. Despite thus being a suitable receptor in principle, the MIPs were not useful for measurements at 50% humidity due to surface saturation by water. This was overcome by introducing primary amino groups into the polymer via allyl amine and by changing the coating morphology from thin film to nanoparticles. This led to the same limit of detection (500 ppb) and selectivity as before, but at the real-life conditions of 50% relative humidity. Keywords: molecularly imprinted polymer; quartz crystal microbalance; formaldehyde detection; MIP nanoparticles
1. Introduction Formaldehyde, a pungent-smelling gas, is regarded one of the main toxic indoor pollutants [1–3] and increasingly draws attention to itself also in outdoor environments [4]. It is carcinogenic and causes a range of conditions including central nervous system damage, immune system disorders, respiratory diseases, and so-called sick building syndrome [5]. Household products and building materials [1,2] are the main indoor sources of formaldehyde. All developed countries have laws to limit the maximum concentration of formaldehyde in building materials. Levels in indoor air are typically in the range of 1–100 ppb (or even higher) depending on formaldehyde source and ventilation [6]. In industry, formaldehyde serves in many processes, e.g., as wood fixative, in paints, insulation foams, dry cleaning solutions, detergents, cosmetics, and pharmaceuticals [7,8]. The World Health Organization (WHO) and the National Institute for Occupational Safety and Health (NIOSH) have set a permissible long-term exposure limit of 0.08- and 1-ppm formaldehyde, respectively. Given the need to monitor formaldehyde concentration in air, portable, cost-effective, robust detectors are highly desirable and have attracted substantial scientific interest, as summarized in a recent review [9]. Many of those sensors are based on inorganic oxidic materials with optimized affinity, but optical approaches have also been published [10]. For introducing shape selectivity, molecularly imprinted polymers (MIPs) [11,12] have attracted increasing attention. In MIP synthesis, the potential analyte acts as a template. Self-assembly of the monomers around this template leads to adapted cavities in the polymer matrix. MIPs have been developed for a wide range of species including small molecules/VOC [13], proteins [14–16], and their aggregates [17] up to entire microorganisms [18]. Inherently, there is hardly a size limit for templates. Hence, metal ions are the smallest species for which MIPs have been reported [19,20]. Despite of course being much larger than a metal ion, the formaldehyde Sensors 2016, 16, 1011; doi:10.3390/s16071011
www.mdpi.com/journal/sensors
Sensors 2016, 16, 1011
2 of 9
a metal ion, the formaldehyde molecule is very small compared to other VOC. This is also reflected Sensors 2016, 16, 1011 2 of 9 in the fact that few articles have been published on formaldehyde imprinting so far: The earliest paper on formaldehyde MIPs combined with quartz crystal microbalance (QCM) detection was published in molecule 2005 [21]. is It very reports appreciable a static measuring asthat wellfew as selectivity small compareddetection to other limits VOC. in This is also reflected chamber in the fact articles factors of about four on against larger aldehydes. Another on pre-concentrating have been published formaldehyde imprinting so far: The paper earliestreported paper on formaldehyde MIPs combined with crystal microbalance (QCM) detection was published in 2005 [21]. It reports formaldehyde on quartz MIPs for electrophoretic separation [22]. Furthermore, fluoral–polyaniline double appreciable detection in a static measuring as well as selectivity factorsisofnot about fouron layers have been used limits for formaldehyde sensingchamber [23], even though this approach based against larger aldehydes. Another paper reported on pre-concentrating formaldehyde on MIPs for molecular imprinting. Two more recent publications in the field of MIP-based formaldehyde sensing electrophoretic separation [22]. Furthermore, fluoral–polyaniline double layers have been used for rely on more complex materials for recognition: One is based on a MIP-Ag/LaFeO 3 nanoparticle formaldehyde sensing [23], even though this approach is not based on molecular imprinting. Two more composite [24], the other one on layer-by-layer assembly of gold nanoparticles and PMAA structures recent in the field MIP-based formaldehyde on moretocomplex materials [25]. Thepublications former article does notofstate any selectivity, whichsensing makesrely it difficult assess the effect of for recognition: One is based on a MIP-Ag/LaFeO nanoparticle composite [24], the other one onon 3 imprinting. The latter indeed reports appreciable imprinting effects. Nonetheless, both systems rely layer-by-layer assembly of gold nanoparticles and PMAA structures [25]. The former article does not polymer–nanoparticle composites. Those pre-concentrate formaldehyde on the surfaces of affinity state any selectivity, which makes it difficult to assess the effect of imprinting. The latter indeed reports materials, which is an approach that has proven useful also for other volatile pollutants [26]. Herein, appreciable imprinting effects. Nonetheless, both systems rely on polymer–nanoparticle composites. we report the design of single-phase recognition materials based on MIPs for assessing formaldehyde Those pre-concentrate formaldehyde on the surfaces of affinity materials, which is an approach that vapor concentrations in flow systems for direct gas phase detection of this analyte. has proven useful also for other volatile pollutants [26]. Herein, we report the design of single-phase recognition materials based on MIPs for assessing formaldehyde vapor concentrations in flow systems 2. Materials and Methods for direct gas phase detection of this analyte.
We purchased all chemicals from Sigma-Aldrich and VWR, respectively, in the highest available
2. Materials Methods purity and keptand them in storage according to manufacturer recommendations.
We purchased all chemicals from Sigma-Aldrich and VWR, respectively, in the highest available
2.1.purity MIP and Synthesis kept them in storage according to manufacturer recommendations. structural 2.1.The MIP Synthesisformulae of all monomers and reagents are summarized in Figure 1. To synthesize
MIPs, we dissolved 13 mg (0.12 mmol) styrene, 13 mg (0.15 mmol) methacrylic acid, and 60 mg (0.31 structural formulae of all monomers and reagents are summarized synthesize mmol) The ethylene glycol dimethacrylate (EGDMA) in a reaction tube using in 500Figure µL of1.aTo binary solvent MIPs, we dissolved 13 mg (0.12 mmol) styrene, 13 mg (0.15 mmol) methacrylic acid, and 60 mg mixture comprised of 200 µL of methanol and 300 µL of dimethyl formamide (DMF) and added 15 (0.31 mmol) ethylene glycol dimethacrylate (EGDMA) in a reaction tube using 500 µL of a binary solvent µL of an aqueous formaldehyde solution (37%–38% w/w, stabilized with 10% methanol). If the mixture comprised of 200 µL of methanol and 300 µL of dimethyl formamide (DMF) and added 15 µL polymer contained primary amino groups for recognition (see Section 3.2), the reaction mixture also of an aqueous formaldehyde solution (37%–38% w/w, stabilized with 10% methanol). If the polymer contained 13 mg (0.26 mmol) allyl amine before adding the template. After obtaining homogeneous, contained primary amino groups for recognition (see Section 3.2), the reaction mixture also contained transparent solutions, we added 6 mg of azobisisobutyronitrile (AIBN) as the radical initiator. 13 mg (0.26 mmol) allyl amine before adding the template. After obtaining homogeneous, transparent Polymerization took place by UV irradiation (λmax = 360 nm and 210 W) for 80 minutes until reaching solutions, we added 6 mg of azobisisobutyronitrile (AIBN) as the radical initiator. Polymerization took theplace gel by point of the mixture. by point the same UV irradiation (λmax =Non-imprinted 360 nm and 210polymers W) for 80 (NIPs) minuteswere until synthesized reaching the gel of procedure, replacing the formaldehyde solution with 9 µL of distilled water. Hence, the solvent the mixture. Non-imprinted polymers (NIPs) were synthesized by the same procedure, replacing the mixtures of bothsolution MIPs and NIPs the same of water, 2% (9and µL NIPs of ~500 formaldehyde with 9 µLcontain of distilled water.amount Hence, the solventnamely, mixturesroughly of both MIPs µL). For sensor measurements, 15namely, µL of this oligomer mixture spin-coated onto one electrode contain the same amount of water, roughly 2% (9 µL of ~500were µL). For sensor measurements, 15 µL pair of a dual-electrode QCM at 4000 rpm on a custom-made spin-coater, while the second pair was of this oligomer mixture were spin-coated onto one electrode pair of a dual-electrode QCM at 4000 rpm spin-coated with NIPs at the same speed. on a custom-made spin-coater, while the second pair was spin-coated with NIPs at the same speed.
Figure reagentsused. used. Figure1. 1. Monomers Monomers and reagents
Sensors 2016, 16, 1011
3 of 9
SensorsFor 2016,generating 16, 1011
3 ofof9 MIP NPs, 500 µL of a MIP oligomer solution was pipetted into 10 mL acetonitrile (AcCN) under vigorous stirring and was then kept on a magnetic stirrer for 12 h. The resulting homogenous NP 500 suspensions were centrifuged rpminto for 10ten The For generating MIP NPs, µL of a MIP oligomer solutionat was4000 pipetted mLminutes. of acetonitrile supernatant was removed by pipetting. Then, we re-suspended the particles in 500 µL of AcCN and (AcCN) under vigorous stirring and was then kept on a magnetic stirrer for 12 h. The resulting used this setup to suspensions coat devices.were For that purpose,at 5 µL of rpm the NP was spin-coated at 4000 homogenous NP centrifuged 4000 for suspension ten minutes. The supernatant was rpm onto the respective QCM electrode. The spin-coated QCM were kept overnight for drying prior removed by pipetting. Then, we re-suspended the particles in 500 µL of AcCN and used this setup to sensitive For measurements and5atomic microscopy (AFM) It was necessary to to mass coat devices. that purpose, µL of force the NP suspension was studies. spin-coated atnot 4000 rpm onto further cross-link stabilizeThe thespin-coated NP layers,QCM because proved stable during gas-phase the respective QCMorelectrode. were they kept overnight for drying prior to mass measurements. sensitive measurements and atomic force microscopy (AFM) studies. It was not necessary to further Surface morphology of NP layers was recorded withstable a VEECO Instruments Nanoscope IVa in cross-link or stabilize the NP layers, because they proved during gas-phase measurements. contact mode with silicon nitride tips. suspensions directly spin-coated onto IVa QCM Surface morphology of NP layers wasNP recorded with a were VEECO Instruments Nanoscope in electrodes as described above and assessed after drying overnight. contact mode with silicon nitride tips. NP suspensions were directly spin-coated onto QCM electrodes
as described above and assessed after drying overnight. 2.2. QCM Manufacturing and Measuring Setup 2.2. QCM andstructures Measuring(fSetup QCMManufacturing dual electrode 0 = 10 MHz) were screen printed with brilliant gold paste containing gold colloid (GGP 2093126, GmbH, Hanau, Germany) onto AT-cut QCM 12.5% dual electrode structures (f0 = 10Heraeus MHz) were screen printed with brilliant goldquartz paste substrates (13.8 mm in diameter, 168-µm-thick, purchased from Great Microtama Industries, containing 12.5% gold colloid (GGP 2093126, Heraeus GmbH, Hanau, Germany) onto AT-cut quartz Surabaya, Indonesia). After that, we burned purchased the structures 400 Microtama °C for 4 hIndustries, to removeSurabaya, organic substrates (13.8 mm in diameter, 168-µm-thick, from at Great ˝ components and reveal the blank gold electrodes. We then spin-coated the respective recognition Indonesia). After that, we burned the structures at 400 C for 4 h to remove organic components material onto electrode as described The QCM the wasrespective then mounted in a custom-made and reveal the an blank gold electrodes. We above. then spin-coated recognition material ontogas an cell (see Figure 2B) and connected to was a custom-made circuit. An Agilent HP5313A electrode as described above. The QCM then mountedoscillating in a custom-made gas cell (see Figure 2B) frequency counter out the resonance frequency of the oscillator. Frequency as acounter functionread of time and connected to aread custom-made oscillating circuit. An Agilent HP5313A frequency out was recorded through an Agilent GPIB/USB adapter by a custom-made LabView routine. the resonance frequency of the oscillator. Frequency as a function of time was recorded throughGas an samplesGPIB/USB were produced bybya agas mixing apparatus of mass flow valves (type by RS-485) Agilent adapter custom-made LabViewconsisting routine. Gas samples were produced a gas addressed by a digital controller (Brooks Instruments SLA5850 S, Hatfield, PA, USA). Defined gas mixing apparatus consisting of mass flow valves (type RS-485) addressed by a digital controller (Brooks streams were bubbled through PA, gasUSA). washing flasks de-ionized water and analyte, Instruments SLA5850 S, Hatfield, Defined gascontaining streams were bubbled through gas washing respectively. The entire setup can be seen in Figure 2A. flasks containing de-ionized water and analyte, respectively. The entire setup can be seen in Figure 2A.
Figure 2. (A) airair streams with defined humidity andand gasgas concentration; (B) (A)Setup Setupused usedfor forgenerating generating streams with defined humidity concentration; Quartz crystal microbalance (QCM) mounted in measuring cell.cell. (B) Quartz crystal microbalance (QCM) mounted in measuring
3. 3. Results Resultsand andDiscussion Discussion 3.1. Development of MIP Thin Films Previous work has shown that acrylate-based acrylate-based polymers polymers are are suitable suitable for for formaldehyde formaldehyde MIPs MIPs[21]. [21]. However, formaldehydeis is inherently lipophilic. Hence, our was ideatowas to reduce of the However, formaldehyde inherently lipophilic. Hence, our idea reduce polaritypolarity of the polymer polymer system, but increase polarizability to achieve better interaction between polymer and system, but increase polarizability to achieve better interaction between polymer and analyte. For that analyte. For that purpose, we introduced styrene as a3 shows monomer. Figure 3 shows thefor QCM sensor purpose, we introduced styrene as a monomer. Figure the QCM sensor results a MIP thin film and the respective NIP when exposed to different vapor concentrations of formaldehyde. Evidently,
Sensors 2016, 1011 Sensors 2016, 16,16, 1011
4 of 4 of 99
Sensors 2016, 16, 1011
4 of 9
resultsfor fora aMIP MIPthin thinfilm filmand andthe therespective respectiveNIP NIPwhen whenexposed exposedtotodifferent differentvapor vaporconcentrations concentrationsofof results formaldehyde.Evidently, Evidently,the theMIP MIPyield yieldsubstantially substantiallyhigher higherfrequency frequencyshifts shiftsatatallallconcentrations concentrations formaldehyde. the MIP yield substantially higher frequency shifts at all concentrations than the corresponding NIP. than the corresponding NIP. The resulting imprinting effect—i.e., the ratio between MIPs andNIPs— NIPs— than the corresponding NIP. The resulting imprinting effect—i.e., the ratio between MIPs and The resulting imprinting effect—i.e., the ratio between MIPs and NIPs—reaches a factor of four, which is reachesa afactor factorofoffour, four,which whichisisappreciably appreciablylarge largegiven giventhe thesmall smallsize sizeofofthe theformaldehyde formaldehydemolecule. molecule. reaches appreciably large given the small size of the formaldehyde molecule. Additionally, the magnitude of Additionally, the magnitude the frequency shiftdepends depends formaldehyde concentration. Overall, Additionally, the magnitude ofofthe frequency shift ononformaldehyde concentration. Overall, the frequency shift depends on formaldehyde concentration. Overall, the setup thus leads to a sensor the setup thus leads to a sensor characteristic with a lower limit of detection (LoD) of 1 ppm (S/N the setup thus leads to a sensor characteristic with a lower limit of detection (LoD) of 1 ppm (S/N characteristic with a lowerall limit of detection (LoD) ofare 1are ppm (S/N ratio ěwithin 3). Furthermore, alloffrequency ratio≥ ≥3).3).Furthermore, Furthermore, allfrequency frequency responses fully reversible within fewtens tens ofseconds. seconds. ratio responses fully reversible a afew responses areinfully reversible within a few tens ofthe seconds. Especially, in QCM gas phase measurements, Especially, inQCM QCM gasphase phase measurements, thedifference difference signal between MIPs andNIPs NIPscan can Especially, gas measurements, ininsignal between MIPs and the difference in signal between MIPs and NIPs can clearly be attributed to mass effects caused by the clearlybebeattributed attributedtotomass masseffects effectscaused causedbybythe theincorporation incorporationofofformaldehyde formaldehydemolecules moleculesinto into clearly incorporation of formaldehyde molecules into recognition sites within the polymer matrix. This again isthe recognitionsites sites withinthe thepolymer polymermatrix. matrix. Thisagain again remarkable giventhe thesmall small size recognition within This isisremarkable given size ofofthe remarkable given the small size of the formaldehyde molecule. formaldehyde molecule. formaldehyde molecule.
Figure 3.QCM QCM sensor responses molecularly imprinted polymer (MIP) andnon-imprinted non-imprinted Figure sensor responses ofofmolecularly imprinted polymer (MIP) and Figure 3.3.QCM sensor responses of molecularly imprinted polymer (MIP) and non-imprinted polymer polymer (NIP), respectively, pulses containing different vapor concentrations formaldehyde. (NIP), respectively, to pulses containing different vapor concentrations of formaldehyde. polymer (NIP), respectively, toto pulses containing different vapor concentrations ofof formaldehyde.
Despite these appreciable results, thin film sensors were notdeveloped developed further after turned Despite these appreciable results, thin film sensors were further it itturned Despite these appreciable results, thin film sensors were notnot developed further afterafter it turned out out that they did not yield useful results in humid air, which makes them useless for operation out that they did not yield useful results in humid air, which makes them useless for operation inin that they did not yield useful results in humid air, which makes them useless for operation in real-life real-life conditions. One of the possible reasons of such a response is indicated in Figure 4: switching real-life conditions. One of the possible reasons of such a response is indicated in Figure 4: switching conditions. One of the possible reasons of such a response is indicated in Figure 4: switching from from50% 50%rH rHtotodry dryairairchanges changesthe thefrequency frequencyofofthe theMIP-coated MIP-coatedchannel channel byalmost almost1400 1400Hz. Hz.It Itisis from 50% rH to dry air changes the frequency of the MIP-coated channel by almostby 1400 Hz. It is therefore thereforehighly highlyprobable probablethat thatwater watermolecules moleculesblock blockaccess accesstotoformaldehyde formaldehydebinding bindingsites siteswithin withinthe the therefore highly probable that water molecules block access to formaldehyde binding sites within the MIP thin MIP thin film. Nonetheless, the MIP thin films serve as proof of the principle that appreciable MIP Nonetheless, thin film. Nonetheless, thefilms MIPserve thin as films serve asprinciple proof ofthat the appreciable principle that appreciable film. the MIP thin proof of the formaldehyde formaldehydesensor sensorresponses responsescan canbebeachieved achievedbybythis thisstrategy. strategy. formaldehyde sensor responses can be achieved by this strategy.
Figure 4.Sensor Sensor responses ofMIP MIP and NIP, respectively, when switching from stream with Figure 4.4.Sensor responses ofofMIP and NIP, respectively, when switching fromfrom an air stream with with 50% Figure responses and NIP, respectively, when switching anan airairstream 50% rH to dry air. rH torH dry 50% toair. dry air.
Sensors 2016, 16, 1011 Sensors 2016, 16, 1011
5 of 9 5 of 9
3.2. Generating 3.2. Generating MIP MIP Nanoparticles Nanoparticles Using MIP MIP NPs NPs seems seems aa rational rational approach approach to to address address these these challenges. challenges. MIP have drawn drawn Using MIP NPs NPs have increasing attention, attention, because because they they often often yield increasing yield higher higher sensitivity sensitivity and and selectivity selectivity [27,28] [27,28] than than thin thin films, films, partly due to shorter average diffusion pathways. Even at higher humidity, formaldehyde molecules partly due to shorter average diffusion pathways. Even at higher humidity, formaldehyde molecules should be be able abletotoreach reachbinding bindingsites sites within polymer matrix. purpose, we synthesized should within thethe polymer matrix. ForFor thatthat purpose, we synthesized MIP MIPbyNP by precipitation, because this approach rather straightforward does not require NP precipitation, because this approach is ratherisstraightforward and does and not require additional additionalsuch reagents, such as surfactants. Figure 5 shows an AFM image ofMIP a resulting NP layer reagents, as surfactants. Figure 5 shows an AFM image of a resulting NP layerMIP deposited on deposited on a glass substrate. a glass substrate.
Figure 5. Atomic force microscopy (AFM) image of formaldehyde MIP nanoparticles. Figure 5. Atomic force microscopy (AFM) image of formaldehyde MIP nanoparticles.
Evidently, the surface is homogeneously covered with NPs having diameters in the range of 80–150 nm. Precipitation hence was successful and led to the desired particulate particulate material. material. However, However, such particles did not yield appreciable responses at elevated humidity. Obviously, the non-specific adsorption molecules ontoonto the surface and selective incorporation of formaldehyde molecules adsorption ofofwater water molecules the surface and selective incorporation of formaldehyde compete with one another. Hence, greater affinity between thebetween polymerthe andpolymer formaldehyde is necessary, molecules compete with one another. Hence, greater affinity and formaldehyde which can be which achieved themodifying polymer [29]. is well-known from medical and biological is necessary, canby bemodifying achieved by the It polymer [29]. It is well-known from medical sciences that formaldehyde strongly interacts strongly with primary amino groups (this phenomenon is used and biological sciences that formaldehyde interacts with primary amino groups (this for stabilizing is tissue mortemtissue for further investigation) by further forminginvestigation) imines (Schiff by bases) with phenomenon usedetc. forpost stabilizing etc. post mortem for forming primary amino bases) groupswith of proteins. the reaction is driven to the end product, this coursetoleads to a imines (Schiff primaryIf amino groups of proteins. If the reaction is of driven the end covalent bond between C of the(double) carbonylbond groupbetween and the the N ofCthe group. group Hence,and we product, (double) this of course leads to the a covalent of amino the carbonyl decided primary amino our material via allyl amine, eveninto though sensing the N of to theintroduce amino group. Hence, wegroups decidedinto to introduce primary amino groups our material of aims ateven reversible Ab-initio ofinteractions. Hall and Smith [30] revealed that viacourse allyl amine, thoughinteractions. sensing of course aimscalculations at reversible Ab-initio calculations the energy between thethat amine-aldehyde adduct forming in the the first step of the reaction of Hall anddifference Smith [30] revealed the energy difference between amine-aldehyde adduct ´1 for methyl amine and formaldehyde in the absence and the imine is ´46 of water. However, forming in the first kJ¨mol step of the reaction and the imine is −46 kJ∙mol−1 for methyl amine and ´ 1 −1 the energy of the transition state is +112.3 kJ¨mol theabove theofenergy of the amine-aldehyde adduct. formaldehyde in the absence of water. However, energy the transition state is +112.3 kJ∙mol Such barrier minimizes the adduct. probability actual iminebarrier formation, especially in the gas abovehigh-energy the energy of the amine-aldehyde Suchofhigh-energy minimizes the probability phase and imine considering the low amountsin of the formaldehyde expected inthe indoor Thus, the of actual formation, especially gas phase vapors and considering lowair. amounts of interactions between and the respective should still be reversible. formaldehyde vaporsformaldehyde expected in indoor air. Thus, theMIP interactions between formaldehyde and the Adding allyl amine did change morphology of the nanoparticles: AFM characterization of respective MIP should still benot reversible. such Adding NP layers showed in the rangeof of the around 100–150 nm, which is the same size allyl amineparticle did notdiameters change morphology nanoparticles: AFM characterization of as forNP polymers withoutparticle allyl amine. However, additional monomer lead which to highly appreciable such layers showed diameters in the the range of around 100–150 nm, is the same size sensor effects, as without can be seen Figure However, 6. Two aspects are immediately noticeable: material as for polymers allylinamine. the additional monomer lead to Firstly, highly the appreciable indeed leads toassensitive frequency responses towardare formaldehyde despite the fact that the sensor effects, can be seen in Figure 6. Two aspects immediatelyvapors noticeable: Firstly, the material indeed leads to sensitive frequency responses toward formaldehyde vapors despite the fact that the corresponding MIP thin films (data not shown) did not give rise to measureable effects. Secondly,
Sensors 2016, 16, 1011 Sensors 2016, 16, 1011
6 of 9 6 of 9
sensors based on MIP NPs respond faster analyte pulses compared with thin films. corresponding MIP thin films (data substantially not shown) did nottogive rise to measureable effects. Secondly, According to the data shown in Figure 6, equilibrium is reached within roughly half a minute for sensors based on MIP NPs respond substantially faster to analyte pulses compared with thin films. each vaporto pulse. In theshown case ofinMIP thin6,films (see Figure 3), the same process lastshalf for around three According the data Figure equilibrium is reached within roughly a minute for minutes for each pulse. Again, this strongly suggests that accessibility of the respective binding sites each vapor pulse. In the case of MIP thin films (see Figure 3), the same process lasts for around three in the material a seminal rolestrongly in sensing. In thethat case of nanoparticles with a roughly 100-nm minutes for eachplays pulse. Again, this suggests accessibility of the respective binding sites diameter, possible diffusion pathways are in the range of at least several tens of nanometers. Thin in the material plays a seminal role in sensing. In the case of nanoparticles with a roughly 100-nm films, however, arediffusion typically pathways 100–200 nm high on each (asseveral determined bynanometers. network analyzer diameter, possible are in the rangeelectrode of at least tens of Thin measurements) and have a diameter of roughly 6 mm (electrode diameter after screen printing is 5 films, however, are typically 100–200 nm high on each electrode (as determined by network analyzer mm, and the layers are somewhat Thus, both the accessible surface is much lower is than for measurements) and have a diameter wider). of roughly 6 mm (electrode diameter after screen printing 5 mm, NP, and diffusion pathways within the polymer can be expected to be much longer. The MIP NP and the layers are somewhat wider). Thus, both the accessible surface is much lower than for NP, and sensor reaches LoDwithin = 0.5 ppm formaldehyde vapor intohumid air,longer. whichThe is the same as thatreaches for the diffusion pathways the polymer can be expected be much MIP NP sensor thin film at dry conditions. During all our measurements (same quartz over several days), LoD = 0.5 ppm formaldehyde vapor in humid air, which is the same as that for the thin film sensor at dry signals returned toall their baselines, strongly indicating reversible—i.e., non-covalent— conditions. During our respective measurements (same quartz over several days), sensor signals returned to binding. their respective baselines, strongly indicating reversible—i.e., non-covalent—binding.
Figure allyl amine towards formaldehyde at Figure 6. 6. QCM QCM responses responsesof ofMIP MIPnanoparticles nanoparticles(NPs) (NPs)containing containing allyl amine towards formaldehyde 50% rH.rH. at 50%
3.3. Selectivity Of course, sensitivity alone is not sufficient to claim successful imprinting. For that, selectivity must also be assessed. Table 1 summarizes the outcome of these experiments for both thin films and nanoparticle layers toward 100-ppm pulses of a range of compounds each. Table 1. Sensor responses of MIP thin films and NPs (containing allyl amine in the monomer mixture) at 0% rH and 50% rH, respectively, toward 100-ppm formaldehyde (HCHO), dichloromethane (CH2Cl2), methanol (MeOH), formic acid, acetone ethanol (EtOH), acetaldehyde (AcCHO), and acetonitrile (AcCN). Sensor
HCHO
CH2 Cl2
MeOH
Formic Acid
Acetone
EtOH
AcCHO
AcCN
Thin film (0% rH)
´65 ˘ 2 Hz
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
NP (50% rH)
´24 ˘ 1 Hz
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
Below noise
Selectivity patterns of MIPsomewhat thin films and MIP NP signals. All Figure those 7. compounds are either similar inlayers size showing or are relative chemically related to formaldehyde: Methanol and formic acid are the reduction and oxidation product of formaldehyde, 3.3. SelectivityAcetaldehyde also contains the CHO group and acetone at least the carbonyl functionality. respectively. The dichloromethane molecule larger than formaldehyde. Surprisingly, none of these Of course, sensitivity aloneisisjust notslightly sufficient to claim successful imprinting. For that, selectivity compounds gave rise to any sensor response revealing specific behavior of the sensor, as all frequency must also be assessed. Table 1 summarizes the outcome of these experiments for both thin films and shifts stayed layers belowtoward the noise level pulses of 1 Hz the oscillator. For each. better visualization, Figure 7 nanoparticle 100-ppm of on a range of compounds summarizes thecompounds normalized are sensor responses for these twoinsystems: shifts in this All those either somewhat similar size or the arefrequency chemically related to
formaldehyde: Methanol and formic acid are the reduction and oxidation product of formaldehyde,
Sensors 2016, 16, 1011
7 of 9
case are related to the sensor response toward the 100-ppm formaldehyde. Error bars for formaldehyde responses correspond to standard deviations obtained for three parallel measurements. In the case Figure 6. QCM of MIP nanoparticles (NPs) containing allyl amine towards formaldehyde of frequency shiftsresponses that are not statistically significant, the noise level of the oscillator was used to at 50% rH. determine the errors.
Figure 7. 7. Selectivity Selectivitypatterns patterns of of MIP MIP thin thin films films and and MIP MIP NP NP layers layers showing showing relative relative signals. signals. Figure
3.3. Selectivity As can be seen, selectivity is the same in both cases, i.e., for thin films containing no allyl amine as well Of as nanoparticles containing amine. The data also successful represents imprinting. results obtained at two different course, sensitivity alone allyl is not sufficient to claim For that, selectivity humidity levels. Although of course not directly comparable, one can clearly see that the sensor must also be assessed. Table 1 summarizes the outcome of these experiments for both thin films and surprisingly reacted specifically towards formaldehyde: no statistically significant signal could be nanoparticle layers toward 100-ppm pulses of a range of compounds each. observed for anycompounds other VOC except formaldehyde during triplicate measurements of those compounds. All those are either somewhat similar in size or are chemically related to The comparably small size of the formaldehyde molecule combined with its very distinct functionality formaldehyde: Methanol and formic acid are the reduction and oxidation product of formaldehyde, may be the reason for such unusually high selectivity, which we previously only observed once in non-covalent MIPs [31]. Both sensitivity and selectivity results very strongly corroborate the outstanding abilities of the imprinting procedure: the MIPs can distinguish between molecules that are rather similar in size, such as formaldehyde and methanol. They also discriminate according to size: acetaldehyde hardly yields any effects on the MIP layer despite containing the same functional group as formaldehyde. 4. Conclusions We herein present a gas sensor for formaldehyde detection based on MIP, relying on both thin films and nanoparticles, respectively. It can detect vapor concentrations down to 500 ppb with outstandingly high selectivity. Optimal results are obtained with MIP NP. The latter point especially reaches beyond immediate science: MIP NP can be introduced into production processes of sensors fairly straightforwardly and thus, in principle, open up a way for commercializing such systems. Acknowledgments: This work has been funded by the European Commission within the seventh research framework EU-FP7, project number EU-FP7-NMP-2010-LARGE-4-263382 “PHOTOSENS,” which we gratefully acknowledge. Author Contributions: M.H. developed the polymer system and carried out thin film and initial nanoparticle experiments. Furthermore, he wrote the first draft of the manuscript. K.K. optimized MIP NP synthesis and carried out additional experiments with that matrix. P.A.L. supervised the work of both M.H. and K.K., gave feedback and ideas. He also guided M.H. in preparing the first draft of the manuscript and wrote the final version. Conflicts of Interest: The authors declare no conflict of interest.
Sensors 2016, 16, 1011
8 of 9
Abbreviations The following abbreviations are used in this manuscript: AcCN AFM AIBN DMF EGDMA MIP NIP NP QCM rH VOC
acetonitrile atomic force microscopy azoisobutyronitrile dimethyl formamide ethylene glycol dimethacrylate molecularly imprinted polymer non-imprinted polymer nanoparticles quartz crystal microbalance relative humidity volatile organic compound
References 1. 2.
3.
4. 5. 6.
7. 8. 9. 10.
11. 12. 13. 14. 15.
16. 17.
Pickrell, J.A.; Mokler, B.V.; Griffis, L.C.; Hobbs, C.H.; Bathija, A. Formaldehyde release rate coefficients from selected consumer products. Environ. Sci. Technol. 1983, 17, 753–757. [CrossRef] [PubMed] Pickrell, J.A.; Griffis, L.C.; Mokler, B.V.; Kanapilly, G.M.; Hobbs, C.H. Formaldehyde release from selected consumer products—Influence of chamber loading, multiple products, relative-humidity, and temperature. Environ. Sci. Technol. 1984, 18, 682–686. [CrossRef] Gunschera, J.; Mentese, S.; Salthammer, T.; Andersen, J.R. Impact of building materials on indoor formaldehyde levels: Effect of ceiling tiles, mineral fiber insulation and gypsum board. Build. Environ. 2013, 64, 138–145. [CrossRef] Salthammer, T. Formaldehyde in the ambient atmosphere: From an indoor pollutant to an outdoor pollutant? Angew. Chem. Int. Ed. 2013, 52, 3320–3327. [CrossRef] [PubMed] Salthammer, T.; Mentese, S.; Marutzky, R. Formaldehyde in the indoor environment. Chem. Rev. 2010, 110, 2536–2572. [CrossRef] [PubMed] Descamps, M.N.; Bordy, T.; Hue, J.; Mariano, S.; Nonglaton, G.; Schultz, E.; Tran-Thi, T.H.; Vignoud-Despond, S. Real-time detection of formaldehyde by a sensor. Sens. Actuators B Chem. 2012, 170, 104–108. [CrossRef] Wang, J.; Liu, L.; Cong, S.Y.; Qi, J.Q.; Xu, B.K. An enrichment method to detect low concentration formaldehyde. Sens. Actuators B Chem. 2008, 134, 1010–1015. [CrossRef] Wang, J.; Zhang, P.; Qi, J.Q.; Yao, P.J. Silicon-based micro-gas sensors for detecting formaldehyde. Sens. Actuators B Chem. 2009, 136, 399–404. [CrossRef] Chung, P.R.; Tzeng, C.T.; Ke, M.T.; Lee, C.Y. Formaldehyde gas sensors: A review. Sensors 2013, 13, 4468–4484. [CrossRef] [PubMed] Boersma, A.; van Ee, R.J.; Stevens, R.S.A.; Saalmink, M.; Charlton, M.D.B.; Pollard, M.E.; Chen, R.; Kontturi, V.; Karioja, P.; Alajoki, T. Detection of low concentration formaldehyde gas by photonic crystal sensor fabricated by nanoimprint process in polymer material. Proc. SPIE 2014, 9141. [CrossRef] Haupt, K.; Linares, A.V.; Bompart, M.; Bernadette, T.S.B. Molecularly imprinted polymers. Top. Curr. Chem. 2012, 325, 1–28. Li, S., Ge, Y., Piletsky, S.A., Lunec, J., Eds.; MIP-based sensors. In Molecularly Imprinted Sensors: Overview and Applications; Elsevier: Amsterdam, The Netherlands, 2012; pp. 339–354. Latif, U.; Rohrer, A.; Lieberzeit, P.A.; Dickert, F.L. Qcm gas phase detection with ceramic materials-VOCs and oil vapors. Anal. Bioanal. Chem. 2011, 400, 2457–2462. [CrossRef] [PubMed] Schirhagl, R.; Podlipna, D.; Lieberzeit, P.A.; Dickert, F.L. Comparing biomimetic and biological receptors for insulin sensing. Chem. Commun. 2010, 46, 3128–3130. [CrossRef] [PubMed] Wangchareansak, T.; Sangma, C.; Choowongkomon, K.; Dickert, F.; Lieberzeit, P. Surface molecular imprints of WGA lectin as artificial receptors for mass-sensitive binding studies. Anal. Bioanal. Chem. 2011, 400, 2499–2506. [CrossRef] [PubMed] Phan, N.V.H.; Sussitz, H.F.; Lieberzeit, P.A. Polymerization parameters influencing the QCM response characteristics of BSA MIP. Biosensors 2014, 4, 161–171. [CrossRef] [PubMed] Chunta, S.; Suedee, R.; Lieberzeit, P.A. Low-density lipoprotein sensor based on molecularly imprinted polymer. Anal. Chem. 2016, 88, 1419–1425. [CrossRef] [PubMed]
Sensors 2016, 16, 1011
18. 19. 20. 21. 22.
23.
24.
25.
26. 27. 28. 29. 30. 31.
9 of 9
Hayden, O.; Lieberzeit, P.A.; Blaas, D.; Dickert, F.L. Artificial antibodies for bioanalyte detection—Sensing viruses and proteins. Adv. Funct. Mater. 2006, 16, 1269–1278. [CrossRef] Branger, C.; Meouche, W.; Margaillan, A. Recent advances on ion-imprinted polymers. React. Funct. Polym. 2013, 73, 859–875. [CrossRef] Bajwa, S.Z.; Dumler, R.; Lieberzeit, P.A. Molecularly imprinted polymers for conductance sensing of Cu2+ in aqueous solutions. Sens. Actuators B Chem. 2014, 192, 522–528. [CrossRef] Feng, L.; Liu, Y.J.; Zhou, X.D.; Hu, J.M. The fabrication and characterization of a formaldehyde odor sensor using molecularly imprinted polymers. J. Colloid. Interf. Sci. 2005, 284, 378–382. [CrossRef] [PubMed] Sun, H.; Lai, J.P.; Fung, Y.S. Simultaneous determination of gaseous and particulate carbonyls in air by coupling micellar electrokinetic capillary chromatography with molecular imprinting solid-phase extraction. J. Chromatogr. A 2014, 1358, 303–308. [CrossRef] [PubMed] Antwi-Boampong, S.; Peng, J.S.; Carlan, J.; BelBruno, J.J. A molecularly imprinted fluoral-p/polyaniline double layer sensor system for selective sensing of formaldehyde. IEEE Sens. J. 2014, 14, 1490–1498. [CrossRef] Zhang, Y.M.; Liu, Q.J.; Zhang, J.; Zhu, Q.; Zhu, Z.Q. A highly sensitive and selective formaldehyde gas sensor using a molecular imprinting technique based on Ag-LaFeO3 . J. Mater. Chem. C 2014, 2, 10067–10072. [CrossRef] Iqbal, N.; Afzal, A.; Mujahid, A. Layer-by-layer assembly of low-temperature-imprinted poly(methacrylic acid)/gold nanoparticle hybrids for gaseous formaldehyde mass sensing. RSC Adv. 2014, 4, 43121–43130. [CrossRef] Mustafa, G.; Lieberzeit, P.A. Molecularly imprinted polymer-Ag2 S nanoparticle composites for sensing volatile organics. RSC Adv. 2014, 4, 12723–12728. [CrossRef] Wackerlig, J.; Lieberzeit, P.A. Molecularly imprinted polymer nanoparticles in chemical sensing—Synthesis, characterisation and application. Sens. Actuators B Chem. 2015, 207, 144–157. [CrossRef] Poma, A.; Turner, A.P.F.; Piletsky, S.A. Advances in the manufacture of mip nanoparticles. Trends Biotechnol. 2010, 28, 629–637. [CrossRef] [PubMed] Kotova, K.; Hussain, M.; Mustafa, G.; Lieberzeit, P.A. MIP sensors on the way to biotech applications: Targeting selectivity. Sens. Actuators B Chem. 2013, 189, 199–202. [CrossRef] Hall, N.E.; Smith, B.J. High-level ab initio molecular orbital calculations of imine formation. J. Phys. Chem. A 1998, 102, 4930–4938. [CrossRef] Hussain, M.; Iqbal, N.; Lieberzeit, P.A. Acidic and basic polymers for molecularly imprinted folic acid sensors—QCM studies with thin films and nanoparticles. Sens. Actuators B Chem. 2013, 176, 1090–1095. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
