LETTER pubs.acs.org/ac
Optical Sensing Scheme for Carbon Dioxide Using a Solvatochromic Probe Reham Ali, Thomas Lang, Sayed M. Saleh, Robert J. Meier, and Otto S. Wolfbeis* Institute of Analytical Chemistry, Chemo- and Biosensors, University of Regensburg, 93040 Regensburg, Germany
bS Supporting Information ABSTRACT: The novel sensing scheme, unlike previous ones that are based on the use of pH indicator probes, is making use of solvatochromic probe Nile Red (NR). Dissolved in a matrix of ethyl cellulose, it can report the polarity of its microenvironment that is modulated by an additive (a hydrophobic amidine) that is capable of reversibly binding carbon dioxide. The spectra of NR undergo a strong solvatochromic shift both in color (from brick-red to magenta) and in fluorescence (from orange to red) if the respective sensor layer is exposed to gaseous CO2 (gCO2) or dissolved CO2 (dCO2). Both visual and instrumental readouts are possible. The sensor layer responds to gCO2 in the range from 0 to 100% and to dCO2 in the range from 0 to 1 M solutions of bicarbonate (equivalent to a CO2 partial pressure of up to 255 hPa). The detection limits are around 0.23% for gCO2 and 1.53 hPa for dCO2. The response time is in the order of 10 min in the forward direction and 3 min in the reverse direction for gCO2 but up to 25 min in the case of dCO2. The optical response also was quantified using a digital camera by extracting the spectral information using the blue and green color channels (in reflectometry) and the green and red channels (in fluorescence), respectively, and by generating pseudocolor pictures.
T
he quantification of carbon dioxide plays an important role in medicine1,2 including respiration measurements3 and blood gas monitoring,4,5 in environmental sciences6 including monitoring carbon dioxide in the atmosphere7 and in seawater,8 in (packed) food and vegetables,9 and the oil and chemical industry.10,11 Direct optical sensing commonly is accomplished12,13 by the measurement of the near-infrared (NIR) absorption of CO2 between 4200 and 4400 cm1. This method is mainly used to sense gaseous CO2 (gCO2) but hardly works for dissolved CO2 (dCO2) due to the strong interference by water and the complexity of the NIR absorption bands of dCO2. Other optical sensors for CO2 are based on acidbase chemistries by exploiting the fact that CO2 undergoes a reversible reaction with a slightly alkaline buffer system contained in a sensor membrane. The resulting change in pH (due to the consumption of hydroxyl ions by reaction with CO2) can be measured via (a) changes in the absorbance of a pH indicator dye,3,1416 also relative to a reference dye,17,18 or (b) changes in the luminescence of a pH probe.1921 One may further differentiate such indicator based sensors in terms of “wet” and “dry” sensors. In the wet CO2 sensors14,16,22,23 (referred to as the Severinghouse-type sensors,24), a carbonate buffer solution is entrapped in a hydrophobic gas permeable polymer along with a pH indicator dye in its base form. The internal pH depends on the external partial pressure of CO2, and the use of a CO2-permeable but proton impermeable polymer makes the system independent of external pH. The “dry” CO2 sensors (developed by the groups of Ishibashi,25 Walt,26 Mills,1,14,16 and Weigl19) are making use of a pH indicator (in its base form) with a lipophilic quaternary ammonium ion as the counterion. All are immobilized along r 2011 American Chemical Society
with a lipophilic organic base (rather than an aqueous buffer) in a hydrophobic polymer. Both the wet and the dry sensors work well in aqueous systems and therefore have been widely applied to sense CO2, for example, in blood,27 for detecting bacterial activity,28 and the like. The dry systems (also referred to as plastic sensors) have a rather fast response and are easy to make but are more prone to drift and also lack long-term operational stability in addition to their cross-sensitivity to traces of other acids including SO2 and NOx. Therefore, the Severinghouse type of sensor is preferably used in diagnostic instrumentation and in marine research. Jessop’s group2931 has described substituted amidines of fatty acids that possess switchable hydrophilicity. One of them (N,N,N0 -tributylpentanamidine; here referred to as TB-PAM) was reported30 to undergo a large change in hydrophilicity on (reversibly) binding CO2 via an acidbase reaction. TB-PAM is a viscous material that is insoluble in water but becomes completely miscible with water in the presence of CO2. The process of binding CO2 is fully reversible, and we perceived that such a material also may be applied in a sensing scheme for this gas because the process is associated with a large change in the polarity of the solvent. The sensing scheme presented here is based on employing a solvatochromic indicator probe to sense the polarity of the microenvironment of TB-PAM. Two methods for read-out are presented. In the first, the intensities of the absorption or emission of the sensor layers were recorded over time using respective spectrometers and while Received: February 3, 2011 Accepted: March 18, 2011 Published: March 25, 2011 2846
dx.doi.org/10.1021/ac200298j | Anal. Chem. 2011, 83, 2846–2851
Analytical Chemistry exposing them to solutions of defined dCO2 and to gases of defined gCO2. In the second, the responses of the sensor layers to dCO2 were imaged with a digital camera and analyzed by the redgreen-blue (RGB) readout.
LETTER
Scheme 1. Acid/Base Reaction Leading to Reversible Binding of Carbon Dioxide
’ EXPERIMENTAL SECTION Materials. All chemicals were purchased from Sigma Aldrich (www.sigmaaldrich.com) and at the highest purity available; except, sodium hydrogen carbonate was from Merck (www. merck.de). Water free of CO2 was prepared by first boiling doubly distilled water (to remove any dissolved CO2), then bubbling it with nitrogen gas, and storing it in a closed container. The inert and optically transparent polyester support poly(ethylene terephthalate; from Goodfellow; www.goodfellow. com) had a thickness of 125 μm. Argon and carbon dioxide (99.9% purity) were obtained from Linde (www.linde-gase.de). The other reagents and organic solvents were of the highest grade available. Instrumentation. Absorption spectra were acquired on a Cary Win UV spectrophotometer (from Varian Inc.; Victoria, Australia; www.varianinc.com). The emission spectra of the sensor film were acquired with an Aminco AB2 luminescence spectrometer (www.thermo.com). The sensor film was placed in a homemade flow-through cell and photoexcited at 468 nm. Gas mixtures were prepared with a gas-mixing device (MKS; www. mksinst.com). The solutions were transported to the sensor film by a peristaltic pump (Minipuls; from Gilson; www.gilson.com). Preparation of the Sensor Films. These were prepared in a glass vial by mixing 0.5 mg of Nile Red with 150 μL of TB-PAM (whose synthesis is described in the Supporting Information) and 1 mL of a 5% solution of ethyl cellulose (10% ethoxy content) that was prepared by dissolving 500 mg of ethyl cellulose in 10 mL of a toluene/ethanol mixture (80/20; v/v). The resulting cocktail was spread onto the polyester support using a homemade knife coater. The resulting sensor film was airdried at room temperature. From the quantities of materials and solvents employed, the thickness of the dried membrane was calculated to be 12 μm in the case of spectral measurements and 6 μm in the case of time trace measurements. Measurements. Absorption spectra were recorded by fixing the sensor film in a cuvette filled with bicarbonate solutions of defined molarity and, thus, defined pCO2. Fluorescence spectra were acquired with the luminescence spectrometer by coupling the light of its xenon lamp (after it had passed the monochromator) into one branch of a bundle of fiber optic waveguides. The fibers were directed onto the sensor film placed in a self-made flow-through cell. Fluorescence (and scattered light) emitted by the sensor layer is guided back by the other branch of the fiber bundle, passes a monochromator, and is then detected by a photomultiplier tube. The response of the sensor film was studied by passing solutions of defined pCO2 (i.e., defined molarity of bicarbonate) or standard gas mixtures of defined pCO2 (see the Supporting Information) through the flowthrough cell. All data were acquired at room temperature (22 °C). RGB Based Imaging Setup. A digital camera (type EOS 50D; from Canon Ltd.; www.canon.com) was used in the RGB readout techniques. Reflectometric imaging was performed with sensor films placed on a sheet of white paper (of ISO 9001 quality; from Evolve; www.evolve-paper.com) that serves both as a neutral background and as a white reference.
White balance depends on ambient light and has to be set to neutral white. Camera parameters were set as follows: white balance, 2700 K; aperture, 5.6; ISO sensitivity, 100; shutter speed, 0.02 s. Fluorescence images were obtained following photoexcitation with an array of 12, 470 nm LEDs (type L-7113PBC-A; from Kingbright; www.conrad.de). Before hitting the sensors, light was passed through a BG12 color glass filter that, in essence, is permeable to light of wavelengths between 350 and 480 nm. Emitted light was passed through an OG510 long-pass filter (both from Schott, Germany; www.schott.com). Camera parameters were set as follows: white balance, 3000 K; aperture, 5.6; ISO sensitivity, 200; shutter speed, 0.2 s. Split color and ratioed pseudocolor images were obtained in both methods (reflectance and fluorescence) using the ImageJ software (see http://rsbweb.nih.gov/ij/). The sensor films were exposed to different concentrations of dCO2 for 30 min and subsequently photographed. Details on data processing are given in the Supporting Information.
’ RESULTS AND DISCUSSION Choice of Materials. The well-known solvatochromic probe Nile Red (NR) was chosen because of its longwave absorption and emission, commercially availabilty, fair photostability, and solubility in the polymer matrix used. It has a fairly strong molar absorption and previously has been used to sense the polarity of various materials.3234 The colors of both the absorption and emission are solvatochromic but independent of pH in the range between 4 and 9. The polymer matrix of the sensor consists of a solution of the indicator probe and of the switchable amidine TBPAM in a matrix of ethyl cellulose. The latter was chosen because it is permeable to CO2 but impermeable to protons. As a result, the sensor is not at all cross-sensitive to pH. Design of Sensor Layer. The sensor film was designed such that all components dissolve in a single film and form a homogeneous matrix. The components are soluble in organic solvents and then form a cocktail that can be spread on an inert and optically transparent polyester support to form a brick-red film after solvent evaporation. If exposed to dCO2 or humidified gCO2, the amidine TB-PAM reacts with CO2 to produce the hydrophilic bicarbonate salt of the amidine as shown in Scheme 1. This is accompanied by a change in the polarity of the microenvironment and, in turn, with larges changes in both the absorption and emission spectra of NR. The change in color from red to magenta is particularly evident. Similarly, the color of fluorescence changes from orange to red, and fluorescence intensity drops as expected.35 Response to Dissolved CO2. The CO2-sensing capability of the sensor films was first tested with aqueous solutions of defined levels of dissolved CO2 (dCO2). Defined levels of dCO2 can be obtained via solutions of a defined concentration of bicarbonate. The preparation of standard solutions and the calculation of 2847
dx.doi.org/10.1021/ac200298j |Anal. Chem. 2011, 83, 2846–2851
Analytical Chemistry
LETTER
Figure 1. (A) Absorption spectra of the sensor film in water and in solutions containing bicarbonate in the concentrations between 0 and 1 M. (B) Respective emission spectra. (C) Respective calibration plots. (a) Plot obtained using the ratio of the absorbances at 500 and 570 nm. (b) Plot obtained using the ratio of the intensities of the emissions (I) at 590 and 650 nm.
respective pCO2 values are given in the Supporting Information. The absorption spectra of the sensor film were recorded both in CO2-free water and in bicarbonate solution in up to 1 M concentrations (Figure 1A). The absorption maximum of NR undergoes a 20 nm red shift if the sensor film is transferred from water to 1 M bicarbonate solution. The largest changes in absorbance can be observed at 500 and 570 nm. As they occur in the opposite direction, one can perform ratiometric measurements. A calibration plot based on the ratio of the absorbances at 500 nm (the maximal decrease in absorbance) and at 570 nm (the maximal increase in absorbance) is shown in Figure 1C. The spectral changes as a function of the concentration of dCO2 are even larger in fluorescence, as can be seen in Figure 1B. In the absence of CO2, the sensor film displays an emission with a peak at 604 nm. On exposure to dCO2 (i.e., to bicarbonate solutions in 0 to 1 M concentrations), the peaks of the emission maxima shift to maximally 629 nm, thus producing an ∼25 nm red shift. The location of the peak may serve as the analytical information. In addition, the intensity decreases by around 53% at 590 nm but increases by around 33% at 650 nm. Thus, a second kind of analytical information can be obtained using the ratio of the intensities at these two wavelengths. Corresponding calibration curves are shown in Figure 1C. Ratiometric measurements are more robust in general, but for the sake of comparison, we also show the normalized intensity of the emission at a single wavelength (590 nm) as a function of dCO2 (Figure S2 in the Supporting Information). Effects of pH. The sensor film was examined with respect to its cross-sensitivity to pH by recording the spectra of the film in buffer solutions of pH values between 7 and 11 (Figure S3 Supporting Information). No change can be observed in color or fluorescence, which is proof for the specificity of the system for CO2. The emission of NR in aqueous solution is independent of pH in this range anyway. Response to Gaseous CO2. The changes in the fluorescence of the sensor film as a function of % of CO2 in the gas phase (gCO2) in the range from 0 to 100% are shown in Figure 2A. Its intensity decreases by 65% at 585 nm but increases by 18% at 640 nm. In parallel, the peak wavelength shifts by 27 nm (from 593 to 620 nm). The bidirectional changes in emission again enable the establishment of a ratiometric calibration plot. Plots of the change in the emission intensity (the ratio between the intensities at 585 and 640 nm) and the shift in the wavelength of the emission peak versus the concentrations of gCO2 are shown in Figure 2B. The sensor does not respond to completely dry gCO2. This can be explained by looking at the chemical equilibrium given in Scheme 1. Water is a coreactant and usually provided by the
Figure 2. (A) Emission spectra of the sensor film exposed to CO2 in argon (at atmospheric pressure) in concentrations from 0 to 100% CO2. (B) Ratio of the intensities of the emission at 585 and 640 nm. (a) Shift of emission peak (λoλ), where λo and λ are the wavelengths of the emission maxima in absence and presence of carbon dioxide, respectively. (b) Shift of the emission maximum versus the concentration of gCO2.
Table 1. Analytical Figures of Merit Including Forward and Reverse Response Times for the Optical Sensors for dCO2 and gCO2 species gaseous CO2 dissolved CO2
analytical range 0100% 13255 hPa
LOD
forward
reverse
0.23%
∼10 min
3 min
1.5 hPa
∼20 min
∼25 min
solvent (in case of aqueous CO2 solution) or by the humidified gCO2. Depending on the level of CO2, a minimal relative humidity of 610% is required. This requirement is, of course, fulfilled when sensing aqueous samples such as blood. A recent article36 describes a scheme for detection (not continuous sensing) of CO2 in a water/tetrahydrofurane mixture that also requires the presence of substantial quantities of water. It is based on the reaction of liquid dipropylamine to form a carbamate ionic liqid which, in turn, causes a nonfluorescent hexaphenylsilanole to form a fluorescent aggregate. While not specifically indicated in the article, we roughly calculate this method to have a limit of detection of around 0.1% of CO2 in a sample gas. Analytical Figures of Merit. The limits of detection for dCO2 and gCO2, respectively, are 1.5 hPa (equivalent to a 6 mM solution of bicarbonate) and 0.23% of CO2 in argon (see Table 1), assuming that luminescence intensity can be determined with a precision of (1%. The stability and reversibility of the sensor were studied by reading the signal intensities at wavelengths where maximum signal changes do occur. The response (τ90) and recovery times are ∼20 and 25 min, respectively, on switching between CO2-free water and 0.3 M 2848
dx.doi.org/10.1021/ac200298j |Anal. Chem. 2011, 83, 2846–2851
Analytical Chemistry bicarbonate (see Figure S4 in the Supporting Information). The response and recovery times of the sensor for gCO2 on switching between argon and 10% of CO2 in argon are 10 and 3 min, respectively, as can be seen in Figure 3. This is much longer than sensors based on direct (IR) spectroscopy of CO2 and the result of (a) the need for a diffusion of the gas into the sensor phase and (b) the rate of the chemical reaction involved. The response times of sensors based on two phases (sample and sensor) are related to the square of the thickness of sensor films, and we believe that they may can become much faster if thinner films (2 μm rather than 6 μm) are employed, albeit at the expense of signal strength at least in absorbance. The signal changes are fully reversible, but a small drift in the sensor for dCO2 was observed when performing continuous measurements for more than 6 h. We interpret this in terms of both photobleaching and leaching. While Nile Red is (very slightly) soluble in water and leaching will remain negligibly small, the ion pair is likely to be much more soluble and thus may leach out. Much less drift (70%) is observed if the sensor is exposed to a stream of aqueous sample but not illuminated. This points to photodecomposition as a second mechanism causing the drift. One solution to overcome leaching is to cover the sensor layer with an ion-impermeable layer of silicone or poly(tetrafluoroethylene) as shown, for example, in a sensor for halothane.37 The sensor for gCO2, in contrast, is much more stable (data not shown). The sensor films can be stored at room temperature in a desiccator at 0% relative humidity for more than 2 months without alteration in their performance.
Figure 3. Signal changes, reversibility, and response times of the sensor film toward humidified gaseous CO2 by switching between argon and 10% CO2 at 585 nm emission.
LETTER
All kinds of CO2 sensors based on pH-responsive indicators are interfered by other acidic gases including SO2, HCl, and NOx, and this is likely also to be the case here. Similarly, plastic-type sensors are likely to be affected by large concentrations of vapors of hydrocarbons which may cause swelling. Oxygen does not quench the luminescence of NR. All kinds of sensors, and plastictype sensors in particular,15 are affected by temperature, but this has not been studied here. Imaging CO2. Recently, the red-green-blue (RGB) readout of standard digital photographic cameras has been used as a simple imaging technique in different types of sensors.3840 Practically all digital photographic cameras are based on complementary metal oxide semiconductor (CMOS) chips consisting of socalled RGB channels that are sensitive to red, green, and blue parts of the spectrum of visible light, respectively, as shown in Figures S5 and S6 in the Supporting Information. The combination of all three (virtually independent) pictures results in a final color picture. The spectrally separated pictures do contain information that is comparable to a 3-wavelength-only spectrophotometer and allow for ratiometric imaging. Two sensing/ imaging formats are widely used. The first relies on counterclockwise spectral changes of indicators displaying two bands, usually those of the acidic form and its conjugate base, respectively. Alternatively, an indicator dye and an inert reference dye may be used, each matching one channel of the RGB readout. Two ways of RGB imaging are employed here. In the case of absorbance based (more correctly reflectometric) imaging, we benefit from the longwave spectral shift occurring on increasing the concentration of dCO2. The color of the sensor film in the absence of CO2 is red, with an absorption band that fairly well matches the sensitivity of the blue channel. With increasing concentrations of dCO2, the color is shifted toward magenta tone, thus better matching the green channel. The spectral overlap of the absorptions and the RGB channels is represented in Figure S5 in the Supporting Information. Photographic images of the sensor layers were acquired (see Figure 4, left panel), and the data stored in the three color channels were then processed using ImageJ software. The red channel does not contain any information for use in reflectometric pictures and was discarded. Next, the intensity data of the blue channel image were divided by the data of the green channel image, and the ratiometric images were then generated in pseudocolors (Figure 4, left panel, bottom).
Figure 4. Schematic of the imaging scheme. (Left panel) Reflectometric imaging. Row 1 gives the conventional RGB images in the presence of various concentrations of bicarbonate (dCO2). Rows 2 and 3 give the colors of the blue and green channel. The ratiometric images in row 4 reflect the true concentration of dCO2 (in pseudocolors). (Right panel) Fluorescence imaging. Lines 1, 2, and 3 again give RGB images and data of two single channels (here green and red), while the red-to-green ratio is given in row 4 in pseudocolors. 2849
dx.doi.org/10.1021/ac200298j |Anal. Chem. 2011, 83, 2846–2851
Analytical Chemistry The sensing scheme for imaging fluorescence is similar, except that the data are stored in the green and red channel, and the data in the blue channel can be discarded. It is based on the shift that occurs in the fluorescence spectra of the sensor films on increasing the concentration of dCO2. The spectral change and the resulting changes on the overlap with the RGB channels are shown in Figure S6 (Supporting Information). On increasing the level of dCO2, the ratio of the fluorescence signals of the green channel and the red channel increases (Figure 4, left panel). Both methods enable direct and quantitative imaging of dCO2 via an intrinsically referenced readout. They are independent of fluctuations of the light sources, nonuniform illumination, dye concentration, and dye leaching. In the given case (where a single indicator probe is being used), the method is even independent of photobleaching, which is a seriously limiting effect in common photometric, reflectometric, or fluorometric ratiometric imaging schemes based on the use of two dyes. The reflectometric technique (the “poor scientist's imager”) is straightforward and requires a minimum of instrumentation to read the sensor layer, specifically a digital camera and a sheet of white paper. However, it requires the ambient light to have a stable and fairly pure color temperature so to avoid chromatic aberrations which would result in errors in the ratioed intensities. The method for imaging fluorescence emission also is simple but requires an external excitation light source (LEDs, for example) and optical filters. In summary, we are reporting on the first sensor for carbon dioxide that is based on the (local) change in the polarity of the microenvironment of a solvatochromic probe as a result of a reversible reaction between CO2 and a strong base in the presence of water. Both gas and aqueous CO2 show good response. The concentration of CO2 can be directly measured through the change of the emission intensity or the red shift in the wavelength of the probe. In addition to the visual color change from brick-red to magenta, the sensor film has attractive features such as working in both the aqueous and gas phase and being independent of pH and simple in preparation. RGB readout enables digital images to be taken with a standard camera to yield ratiometric pseudocolor pictures after data processing.
’ ASSOCIATED CONTENT
bS
Synthesis of N,N,N0 -tributylpentanamidine (TB-PAM); preparation of standard solutions for sensing dissolved CO2; preparation of gas mixtures for sensing gaseous CO2; spectra of Nile Red (NR), Figure S1; calibration plot for dCO2, Figure S2; effect of solution pH, Figure S3; reversibility of dissolved CO2, Figure S4; spectral sensitivities of the three color channels of the red-green-blue (RGB) camera and overlap with the absorption and the emission spectra of a sensor film, Figure S5 and S6, respectively; data read-out for digital imaging CO2. This material is available free of charge via the Internet at http://pubs.acs.org. Supporting Information.
’ AUTHOR INFORMATION Corresponding Author
*Fax: þ49-941-943-4064. E-mail:
[email protected].
’ ACKNOWLEDGMENT R.A. and S.M.S. thank the Ministry of Higher Education of the Arab Republic of Egypt for financial support.
LETTER
’ REFERENCES (1) Jameison, F.; Sanchez, R. I.; Dong, L. W.; Leland, J. K.; Yost, D.; Martin, M. T. Anal. Chem. 1996, 68, 1298–1302. (2) Fernandez-Sanchez, J. F.; Cannas, R.; Spichiger, S.; Steiger, R.; Spichiger-Keller, U. E. Sens. Actuators, B: Chem. 2007, 128, 145–153. (3) Mills, A.; Lepre, A.; Wild, L. Sens. Actuators, B: Chem. 1997, 39, 419–425. (4) Shin, J. H.; Lee, S. H.; Kim, M. H.; Yoon, H. J.; Lee, J. S.; Cha, G. S.; Nam, H. Clin. Chem. 2000, 46, A60–A61. (5) Prough, D. S.; Stump, D. A.; Roy, R. C.; Gravlee, G. P.; Williams, T.; Mills, S. A.; Hinshelwood, L.; Howard, G. Anesthesiology 1986, 64, 576–581. (6) Wolfbeis, O. S.; Kovacs, B.; Goswami, K.; Klainer, S. M. Mikrochim. Acta 1998, 129, 181–188. (7) Church, N. Trends Food Sci. Technol. 1994, 5, 345–352. (8) Siegenthaler, U.; Sarmiento, J. L. Nature 1993, 365, 119–125. (9) Descoins, C.; Mathlouthi, M.; Le Moual, M.; Hennequin, J. Food Chem. 2006, 95, 541–553. (10) Aresta, M.; Tommasi, I. Energy Convers. Manage. 1997, 38, S373–S378. (11) Blanchard, L. A.; Brennecke, J. F. Ind. Eng. Chem. Res. 2001, 40, 287–292. (12) Wolfbeis, O. S. Anal. Chem. 2006, 78, 3859–3873. (13) Threlkeld, C. H. J Food Sci. 1982, 47, 1222–1225. (14) Mills, A.; Skinner, G. A. Analyst 2010, 135, 1912–1917. (15) Mills, A.; Chang, Q.; McMurray, N. Anal. Chem. 1992, 64, 1383–1389. (16) Oter, O.; Ertekin, K.; Topkaya, D.; Alp, S. Sens. Actuators, B: Chem. 2006, 117, 295–301. (17) Borisov, S. M.; Waldhier, M. C.; Klimant, I.; Wolfbeis, O. S. Chem. Mater. 2007, 19, 6187–6194. (18) Amao, Y.; Nakamura, N. Sens. Actuators, B: Chem. 2004, 100, 347–351. (19) Zhujun, Z.; Seitz, W. R. Anal. Chim. Acta 1984, 160, 305–309. (20) Mills, A.; Chang, Q. Analyst 1993, 118, 839–843. (21) Weigl, B. H.; Wolfbeis, O. S. Sens. Actuators, B: Chem. 1995, 28, 151–156. (22) Tabacco, M. B.; Uttamlal, M.; McAllister, M.; Walt, D. R. Anal. Chem. 1999, 71, 154–161. (23) Cooney, C. G.; Towe, B. C.; Eyster, C. R. Sens. Actuators, B: Chem. 2000, 69, 183–188. (24) Severinghaus, J. W.; Bradley, A. F. J. Appl. Physiol. 1958, 13, 515. (25) Kawabata, Y.; Kamichika, T.; Imasaka, T.; Ishibashi, N. Anal. Chim. Acta 1989, 219, 223–228. (26) Raemer, D. B.; Walt, D. R.; Munkholm C. U.S. Patent 5.005.572, 1991. (27) Gehrich, J. L.; L€ubbers, D. W.; Opitz, N.; Hansmann, D. R.; Miller, W. W.; Tusa, J. K.; Yafuso, M. IEEE Trans. Biomed. Eng. 1986, 33, 117–132. (28) Swenson, F. J. Sens. Actuators, B: Chem. 1993, 11, 315–321. (29) Jessop, P. G.; Heldebrant, D. J.; Li, X. W.; Eckert, C. A.; Liotta, C. L. Nature 2005, 436, 1102–1102. (30) Phan, L.; Jessop, P. G. Green Chem. 2010, 12, 809–814. (31) Mercer, S. M.; Jessop, P. G. ChemSusChem 2010, 3, 467–470. (32) Sackett, D. L.; Wolff, J. Anal. Biochem. 1987, 167, 228–234. (33) Lobnik, A.; Wolfbeis, O. S. J. Sol-Gel Sci. Technol. 2001, 20, 303–311. (34) Chen, S. J.; Chang, H. T. Anal. Chem. 2004, 76, 3727–3734. (35) Greenspan, P.; Mayer, E. P.; Fowler, S. D. J. Cell Biol. 1985, 100, 965–973. (36) Liu, Y.; Tang, Y.; Barashkov, N. N.; Irgibaeva, I. S.; Lam, J. W. Y.; Hu, R.; Birimzhanova, D.; Yu, Y.; Tang, B. Z. J. Am. Chem. Soc. 2010, 132, 13951–13953. (37) Wolfbeis, O. S.; Posch, H. E.; Kroneis, H. Anal. Chem. (Wash.) 1985, 57, 2556–2561. (38) Wang, X. D.; Meier, R. J.; Link, M.; Wolfbeis, O. S. Angew. Chem., Int. Ed. 2010, 49, 4907–4909. 2850
dx.doi.org/10.1021/ac200298j |Anal. Chem. 2011, 83, 2846–2851
Analytical Chemistry
LETTER
(39) Steiner, M. S.; Meier, R. J.; Duerkop, A.; Wolfbeis, O. S. Anal. Chem. 2010, 82, 8402–8405. (40) Stich, M. I. J.; Borisov, S. M.; Henne, U.; Schaeferling, M. Sens. Actuators, B: Chem. 2009, 139, 204–207.
2851
dx.doi.org/10.1021/ac200298j |Anal. Chem. 2011, 83, 2846–2851
