ARTICLE pubs.acs.org/ac
Relative Humidity Sensors Based on an Environment-Sensitive Fluorophore in Hydrogel Films John C. Tellis, Christopher A. Strulson, Matthew M. Myers, and Kristi A. Kneas* Department of Chemistry and Biochemistry, Elizabethtown College, One Alpha Drive, Elizabethtown, Pennsylvania 17022, United States
bS Supporting Information ABSTRACT: A fluorescence-based sensing scheme exploiting an environment-sensitive fluorophore embedded in a hydrogel has been developed for measurement of relative humidity (RH). The fluorophore, dapoxyl sulfonic acid (DSA), is incorporated into two different hydrogel films, agarose and a copolymer of acrylamide and 2-(dimethylamino)ethyl methacrylate (DMAEM) cross-linked with N,N0 -methylenebisacrylamide. The swelling and contracting of the hydrogels in response to relative humidity alters the polarity of the environment of DSA, stimulating a shift in the emission wavelength. From 0 to 100% RH, acrylamide-DMAEM sensors exhibited a 40 and 15 nm wavelength shift in still air and flowing gas, respectively. Agarose sensors showed a 40 nm wavelength shift from 0 to 100% RH in still air and a 30 nm shift from 0 to 70% RH in flowing gas. Response times for both sensors were 15 min in still air and less than 5 min in flowing gas. The sensing approach is straightforward and cost-effective, yields sensors with characteristics suitable for commercial measurement of RH (i.e., sensitivity, response times, reproducibility), and allows ease of adaptability to specific RH measurement requirements. The results support the potential extension of the method to a wide variety of analytes in the vapor phase and aqueous solution by incorporation of functionalized “smart” hydrogels.
M
easurement and control of relative humidity (RH) continue to be of great importance in diverse areas and industries. For example, the monitoring of humidity is commercially important in museums and libraries for the protection of precious paintings, sculptures, and documents. Industrial applications include the control of humidity in the production of semiconductors, textiles, and food products.1,2 However, current RH sensors can be hindered by a number of factors including slow response times (e.g., >40 min in still air and as high as 3 min in moving air), hysteresis, susceptibility to electromagnetic interference, and high cost.1-3 To address one or more of these shortcomings, a number of optical RH sensing approaches have been reported in the literature. In one example, Shi et al. monitored changes in the refractive index of polymeric photonic crystals upon variation of RH. Though sensitivity was limited, response times in moving air (∼1.5-30 s) were far superior to those of commercial electrochemical sensors. Responses were nonlinear over the reported range of 40-95% RH, and hysteresis (i.e., directional dependence of response) was observed.4 Matsushima et al. observed a color change in ionic dyes incorporated into sugar gels. Not fully calibrated, the observed response was attributed to the switching of aggregate states of the dyes when exposed to water vapor.5 In a reflectance-based approach reported by Dacres and Narayanaswamy, a sensor composed of crystal violet in a Nafion film exhibited exceptional parts per million detection limits but a very small dynamic range (0-0.25%) that limits its usefulness in most commercial applications.6 Itagaki et al. measured the change in reflectance of a porphyrin immobilized in Nafion films. Sensor r 2010 American Chemical Society
performance was varied by altering the ratio of porphyrin to Nafion, but an optimal composition which allowed for both a large dynamic range and good sensitivity was not found.7 A number of luminescence-based RH sensors have been developed, but most rely on intensity-based measurements which are prone to calibration errors.8 For example, humidityinduced changes in the luminescence of ruthenium complexes embedded in polymer supports have been studied by Xu et al., Takato et al., and Bedoya et al. In the last, the change in excited state lifetime of the complex was monitored with varying humidity.9-11 In an approach that is similar to that reported here, Otsuki and Adachi embedded polarity-sensitive fluorophores into hydroxypropylcellulose and monitored wavelength and frequency shifts upon adsorption of water. A nonlinear, 30-40 nm wavelength shift was observed between 0% and 87% humidity. The sensor films were not calibrated in still air, but response times even in flowing air were greater than 10 min, thus limiting potential for use in most commercial applications.12,13 In the work presented here, the approach of Otsuki and Adachi was modified to make use of an environment-sensitive fluorophore incorporated into hydrogels, which exhibit extensive swelling upon absorption of water. Specifically, dapoxyl sulfonic acid (DSA) was chosen because of its high quantum yield, large Stokes shift, solubility in water and ease of incorporation during polymerization, and affinity for aqueous interfaces within the Received: October 4, 2010 Accepted: December 3, 2010 Published: December 22, 2010 928
dx.doi.org/10.1021/ac102616w | Anal. Chem. 2011, 83, 928–932
Analytical Chemistry
ARTICLE
hydrogel. Most importantly, the emission of DSA exhibits exceptional environment sensitivity, a result of the solvent-sensitive intramolecular charge transfer occurring in the excited state.14 The potential use of hydrogels in optical RH sensing has been documented previously.15 In the current work, the hydrogel swells and contracts in response to changing RH and amount of absorbed water within its pores. DSA responds to the resultant change in polarity, with longer emission wavelengths at higher RH and polarity and shorter emission wavelengths at lower RH and polarity. A wavelength-based calibration eliminates a number of errors associated with intensity-based measurements, including those resulting from changes in source intensity, detector sensitivity, and sensor film position, as well as photodegradation of the luminescent compound.8 A significant benefit of this sensing method is derived from its functional simplicity and potential for wide applicability. The engineering of analyte-sensitive luminophores proves to be somewhat challenging, and sensor design is further complicated by the limited predictability of luminescence behavior in polymeric supports.16 The development of analyte-sensitive “smart” hydrogels is a rapidly growing area of research that provides a viable alternative to use of an analyte-sensitive luminophore.17-19 The current work demonstrates the potential of the generalized sensing strategy in which analyte sensitivity is imparted by the hydrogel, and the environment-sensitive luminophore is utilized to monitor swelling behavior as a function of analyte concentration.
persulfate initiator was added via delivery of a saturated solution. At 2 min, small volumes of the monomer solution (400 μL) were delivered via a microliter pipet onto prescored glass microscope slides as previously described for agarose sensors. Upon polymerization and drying at room temperature for 4 days, the resultant films were dried in a laboratory oven at 55 °C. Calibration of Sensors. Relative humidity (RH) was measured at ambient temperature using a Vernier RH-BTA relative humidity probe connected to a Vernier LabQuest data acquisition device. The humidity probe is reported to provide an accuracy of (5% RH with a resolution of 0.16% RH from 0 to 95% RH. Relative humidity was systematically varied in one of two ways. For flowing gas calibrations, a previous approach, which makes use of two ultrapure nitrogen gas cylinders and a mixing chamber, was modified for use.12,13 One nitrogen gas flow, sent through a deionized water bubbler, was varied relative to a second, dry nitrogen flow to obtain the desired relative humidity. The airtight mixing chamber was fitted with an RH probe and an outlet tube directed into the cuvette containing the sensor film. For calibrations in still air, saturated salt solutions of desired relative humidity were prepared and allowed to equilibrate in a closed container at room temperature: 0% RH, Drierite; 6% RH, LiBr; 11% RH, LiCl; 21% RH, KCH3CO2; 33% RH, MgCl2; 54% RH, Mg(NO3)2 3 6H2O; 57% RH, NaBr; 75% RH, NaCl; 85% RH, KCl; 98% RH, K2SO4.20 The sensor film was placed into a cuvette in the sealed container for 75 min, at which time the cuvette was capped and the emission of the sensor film was measured. Fluorescence spectra were obtained on a Perkin-Elmer LS50-B spectrofluorometer using 365 nm excitation and a 1 cm triangular quartz cuvette (NSG Precision Cells, type 81) to obtain front face measurements in the standard right-angle cell holder. Scattered radiation was minimized by use of horizontal and vertical polarizing filters in the excitation and emission path, respectively. When the emission intensity exceeded the narrow dynamic range of the instrument, neutral density filters (OD of 0.2, 0.3, or 0.5; ThorLabs) were used in the emission path. To reduce error in the determination of the maximum emission wavelength, spectra were fit to the sum of two Gaussian distributions. The χ2 values for the normalized spectral data did not exceed 0.25 (300 points). Reported emission maxima were determined from the best-fit maximum intensities and plotted as a function of RH. A single Gaussian distribution did not accurately fit the peaks, which exhibited some asymmetry in the form of a subtle shoulder toward the red edge. Response and recovery times were determined by introducing a step change in RH and obtaining spectra rapidly at various time intervals until both emission intensities and wavelengths remained stable. A step change was achieved by subjecting a capped cuvette containing an equilibrated sensor film to a different RH via a flowing gas stream or by exposure to ambient RH in still air. The response times (i.e., low to high RH) and recovery times (i.e., high to low RH) were determined by calculating the time at which the emission wavelength change reached 90% of its final value with respect to zero time.
’ EXPERIMENTAL SECTION Materials. Acrylamide (99þ%), 2-(dimethylamino)ethyl methacrylate (DMAEM, 99%), and N,N0 -methylenebisacrylamide (BIS, 96%) were obtained from Acros Organics. Potassium persulfate (99þ%) was purchased from Sigma-Aldrich, agarose powder (laboratory grade) was purchased from Sigma, and dapoxyl sulfonic acid monosodium salt (DSA) was purchased from Invitrogen. High purity gases (>99.5%) were obtained from Roberts Oxygen Company. Drierite desiccant (8 mesh), LiBr (99þ%), and Mg(NO3)2 3 6H2O (98%) were obtained from Acros. LiCl (Certified ACS), KCH3CO2 (Certified ACS), MgCl2 (Certified ACS), NaBr (Certified ACS), NaCl (Certified ACS), KCl (Certified ACS), and K2SO4 (Certified ACS) were obtained from Fisher. All solid and liquid reagents were handled with gloves and used as received. Standard safety precautions and waste treatment procedures were followed. Preparation of Agarose Sensors. Agarose gels were prepared by dissolving agarose powder (1 wt %) in deionized water and heating with stirring to 70 °C. Upon cooling at 50 °C, DSA was added to give a concentration of 150 μM. Small volumes (1-2 mL) were cast onto prescored glass microscope slides (Corning, 7.6 cm 2.5 cm 1 mm), spread uniformly over the surface, and allowed to dry at room temperature for at least 2 days. Each glass slide resulted in six 2.5 cm 0.8 cm thin films which were left on the glass for measurement. Preparation of Acrylamide/DMAEM Sensors. To an aqueous solution containing acrylamide (57 wt %), DMAEM (40 wt %), and BIS cross-linker (3 wt %), DSA was added to give a concentration of 50 μM. Alternatively, monomer ratios were varied to obtain 60 or 80 wt % DMAEM and 1.5 or 6 wt % of BIS cross-linker. The solution was purged with ultrapure nitrogen and stirred for 20 min, after which 0.2 wt % of potassium
’ RESULTS AND DISCUSSION Agarose Sensors. Results obtained for sensor films prepared from agarose are consistent with previous work, which suggested that agarose hydrogels would be well suited for use 929
dx.doi.org/10.1021/ac102616w |Anal. Chem. 2011, 83, 928–932
Analytical Chemistry
ARTICLE
Figure 2. Still air calibration of agarose sensor film ([) with error bars representing standard error. Trend lines for flowing gas calibrations with increasing RH (solid line) and decreasing RH (dashed line) are included for comparison. Recovery time of sensor upon step change from 100% to 64% RH (inset).
Figure 1. Calibration of agarose sensor film on day 1 (4) and day 9 ([) in flowing gas with increasing RH (solid line) and decreasing RH (dashed line). Recovery time of sensor upon step change from 96% to 3% RH (inset).
in polymer-based RH sensors due to their hydrophilicity and ease of preparation.15 Figure 1 shows calibration results obtained at ambient temperature for an agarose sensor film measured on two different days in flowing gas of varying RH, while Figure S-1, provided in the Supporting Information, shows the fluorescence spectra obtained for an agarose sensor with varying RH. There is an obvious shift toward longer emission wavelengths with increasing RH, a result of the swelling of the hydrogel and increased polarity experienced by the solvatochromic DSA upon absorption of water. The sensor exhibited reversible swelling and a smooth but nonlinear calibration with greater than a 30 nm shift in wavelength on going from 0 to 100% RH. The two different calibrations shown in Figure 1, obtained more than a week apart, demonstrate reproducibility of the results for the same sensing film and stability of the sensing film with time. Further evidence of day-day reproducibility of agarose sensor films as well as batch-batch reproducibility in terms of observed emission wavelength and shift of emission wavelength as a function of RH are apparent in Table S-1, provided in the Supporting Information, which includes a summary of emission wavelengths observed at 0% and 100% RH for a series of agarose sensing films. Hysteresis was observed as apparent by the differing results for increasing and decreasing RH in Figure 1. A possible explanation for the observed difference would be if results were obtained prior to complete recovery of the sensing film in response to decreased RH. In previous work, the response and recovery times of agarose films were reported to be roughly 4 and 30-55 s, respectively, when exposed to human breathing.15,21 The agarose sensor reported here exhibited in flowing gas a comparable response time (∼10-15 s) but a much greater recovery time that exceeded 20 min (inset of Figure 1). Because measurements reported in Figure 1 were recorded at times that far exceeded the recovery time (>30 min), the hysteresis observed in Figure 1 cannot be explained as an artifact resulting from a difference in response and recovery of the sensing film. Instead, a more likely explanation is related to cavity theory. As RH is decreased, water likely is entrapped in the porous films because of more rapid shrinkage of the film surface relative to the void volume within the hydrogel.22 The calibrations in Figure 1 exhibit the same
trend as reported by other authors offering a similar explanation for observed hysteresis.23 Given that most commercial applications necessitate measurement of RH in areas of limited air flow, such as in homes, museums, and hospitals, it was important to assess the usefulness of the agarose sensor film in still air.1 Figure 2 shows composite results for the calibration of a single agarose sensing film at room temperature. Replicate measurements were obtained 10 days apart, and the error bars represent the standard error for measurements obtained at each RH. Reproducibility was poorer than that obtained for flowing gas calibrations, making it difficult to assess the extent of hysteresis in still air. As expected, the response and recovery times, 15 and 30 min, respectively, were lengthier than those measured in flowing air, but they are markedly improved over response times reported for commercial sensors when used in still air.3 Between 0 and 70% RH, the apparent trend and sensitivity are comparable to those obtained for flowing gas calibrations with increasing RH, but the observed emission wavelengths are further red-shifted. In still air, the sensor film gives a wider dynamic range with enhanced sensitivity at high humidity. The observed red-shift in emission wavelength and enhanced response at high RHs for still air calibrations relative to flowing gas calibrations may be attributed to condensation within the sensor films, which is less likely to occur in flowing gas. Unfortunately, very few examples exist of RH sensing systems that have been calibrated in still and flowing air, so it is uncertain whether this phenomenon is typical for polymer supported RH sensors. The measured differences in response between still and flowing gas are reproducible, thus necessitating independent calibrations for each set of conditions (i.e., still air and flowing gas) in which sensor films are to be used. Acrylamide/DMAEM Sensors. Upon use of an acrylamide/ DMAEM copolymer in the development of fluorescence-based pH sensors, it was observed that the emission color under UV light changed markedly when exposed to high RH. This observation led to further characterization of these hydrogels for measurement of RH. Figure 3 shows results for acrylamide/ DMAEM sensors with 40%, 60%, and 80% DMAEM and 3% BIS measured at room temperature, and Table S-1, provided in the 930
dx.doi.org/10.1021/ac102616w |Anal. Chem. 2011, 83, 928–932
Analytical Chemistry
ARTICLE
Figure 3. Flowing gas calibrations of (gray )) 40%, (9) 60%, and (4) 80% DMAEM sensor films. Recovery time of sensor upon step change from 97% to 3% RH (inset).
Figure 4. Still air (gray 2, [) and flowing gas (4,]) calibrations of two different acrylamide/40% DMAEM sensor films (1, 4; 2, ]).
Supporting Information, includes a summary of the emission wavelengths observed at 0% and 100% RH for the series of acrylamide/DMAEM sensing films that were investigated. Again, there is an observed red shift with increasing RH, a result of the swelling of the hydrogel and increased polarity experienced by DSA. Consistent with a decrease in hydrophilicity, the measured wavelengths for the acrylamide/DMAEM sensors are shorter than those reported for the agarose sensors. The acrylamide/ DMAEM sensor films with 40%, 60%, and 80% DMAEM exhibited good linearity (R2 = 0.996, 0.995, and 0.998, respectively) and moderate sensitivity (0.170 ( 0.003, 0.160 ( 0.006, and 0.130 ( 0.002 nm/%RH, respectively) over the full 0-100% RH range, with no observed hysteresis (Figure S-2 in the Supporting Information). At most, the uncertainty in experimentally determined RH values from the calibration curves is (1.5-3.5% as determined by propagation of error and taking into account the fact that the uncertainties in the measured RH of standards are greater than those for the measured emission wavelengths.24 The response time of the acrylamide/DMAEM sensors to an increase in RH was determined to be 20 min), and hysteresis within the dynamic range limit their usefulness in commercial applications requiring absolute measurement of RH. Acrylamide/DMAEM sensors addressed some of the limitations of agarose sensors, with excellent linearity in flowing gas, no hysteresis, and more comparable response and recovery times. The more limited sensitivity in flowing gas and different calibration trends in flowing gas and still air, however, are less desirable. Both sensing systems displayed good day-day reproducibility as well as excellent photostability and sensor film lifetimes in excess of 6 months. For applications requiring relative rather than absolute RH measurements, the dramatic emission color change, which is visible to the naked eye under ultraviolet excitation with an inexpensive black light, allows for the potential use of the sensor films without instrumentation, thus drastically reducing the cost of sensor production and use. The different performance characteristics (i.e., sensitivity, dynamic range, response/recovery time) of the two polymer systems reported here provide evidence of the ability to develop sensors that meet the design criteria for particular applications. The RH sensors reported here suggest the broader potential of a low cost and versatile inductive sensing method that exploits an environment-sensitive luminophore embedded in an analyte specific “smart” hydrogel. Since analyte selectivity can be derived from the hydrogel rather than the luminophore, a wider variety of synthetic approaches is available to target analytes of interest and to increase selectivity by modifying pendant groups on the monomers. In addition, the ease of fabrication of sensor films allows them to be adapted readily for use in remote, fiber optic sensing. ’ ASSOCIATED CONTENT
bS
Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. 932
dx.doi.org/10.1021/ac102616w |Anal. Chem. 2011, 83, 928–932
