Polymer Composite Film for

The solvatochromic dye Nile Red dispersed in selected hydrogen bond .... (15) Rose-Pehrsson, S. L.; Grate, J. W.; Ballantine, D. S.; Jurs, P. C. Anal...
0 downloads 0 Views 135KB Size
Anal. Chem. 2001, 73, 3441-3448

Rational Design of a Nile Red/Polymer Composite Film for Fluorescence Sensing of Organophosphonate Vapors Using Hydrogen Bond Acidic Polymers Igor Levitsky* and Sergei G. Krivoshlykov

ALTAIR Center, LLC, 1 Chartwell Circle, Shrewsbury, Massachusetts 01545 Jay W. Grate*

Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, P. O. Box 999, Richland, Washington 99352

The solvatochromic dye Nile Red dispersed in selected hydrogen bond acidic polymer matrixes demonstrated strong fluorescence enhancement at the presence of dimethyl methylphosphonate (DMMP) vapors. Two hydrogen bond acidic polymers were examined as dye matrixes, one with fluorinated alcohol groups on a polystyrene backbone (PSFA) and the other with fluorinated bisphenol groups alternating with oligo(dimethylsiloxane) segments (BSP3). The combination of hydrogen bond acidic polymer (a strong sorbent for DMMP) with the solvatochromic dye led to initial depression of the dye fluorescence and a significant red shift in the absorbance and fluorescence spectra. DMMP sorption changed the dye environment and dramatically altered the fluorescence spectrum and intensity, resulting in a strong fluorescence enhancement. It is proposed that this fluorescence enhancement is due to the competition set up between the dye and the sorbed vapor for polymeric hydrogen-bonding sites. The highest responses were obtained with BSP3. DMMP detection has been demonstrated at sub-ppm DMMP concentrations, indicating very low detection limits compared to previous Nile Red/ polymer matrix fluorescence vapor sensors. Nile Red/ poly(methyl methacrylate) films prepared for comparisons exhibited substantially lower response to DMMP. Rational selection of polymers providing high sorption for DMMP and competition for hydrogen-bonding interactions with Nile Red yielded flourescent films with high sensitivity. Polymers serve many roles in chemical sensor structures, and in many cases they collect and concentrate analyte molecules on sensor surfaces.1 They play this role in a variety of sensor array and electronic nose configurations for gas-phase vapor sensing, including those based on acoustic wave devices, optical sensors, and chemiresistor sensors.2 As such, the sorbent polymer can play (1) Harsanyi, G. e. Polymer Films in Sensor Applications; Technomic Publishing Co.: Lancaster, PA, 1995. 10.1021/ac0014966 CCC: $20.00 Published on Web 06/01/2001

© 2001 American Chemical Society

an important role in the sensitivity and selectivity of a sensor, and its interactions with sorbed vapor molecules influence the analytical performance. Accordingly, vapor/polymer interactions have been investigated as a means to rationally understand and design sensors and sensor arrays, using linear solvation energy relationships to systematically explore the role of solubility properties and fundamental interactions in selectivity and diversity.3-6 This approach has been exploited mainly in acoustic wave sensor arrays,3 although one early paper described how the sorbent polymer could be used to improve the performance of chemiresistor sensors for a specific class of vapors.7 Thus far, the selection of polymers for sensors based on fluorescent dyes in polymer matrixes has been largely empirical. In the first full paper on dye/polymer fluorescent sensing arrays, Walt and co-workers described 14 polymers tested in combination with Nile Red (NR).8 A later account listed 16 polymers.9 Still another paper described 19 coated fibers,10 and it has been found empirically that copolymers of varying compositions exhibit different selectivities.11 One article mentioned screening over 100 candidate polymer matrixes.12 Thus far, none of these reports has described the use of a strongly hydrogen bond acidic polymer as the matrix for a dye/polymer fluorescent sensor. There exists one (2) Albert, K. J.; Lewis, N. S.; Schauer, C. L.; Sotzing, G. A.; Stitzel, S. E.; Vaid, T. P.; Walt, D. R. Chem. Rev. 2000, 100, 2595-2626. (3) Grate, J. W. Chem. Rev. 2000, 100, 2627-2648. (4) Grate, J. W.; Abraham, M. H.; McGill, R. A. In Handbook of Biosensors: Medicine, Food, and the Environment; Kress-Rogers, E., Nicklin, S., Eds.; CRC Press: Boca Raton, FL, 1996; pp 593-612. (5) McGill, R. A.; Abraham, M. H.; Grate, J. W. CHEMTECH 1994, 24 (9), 27-37. (6) Grate, J. W.; Abraham, M. H. Sens. Actuators B 1991, 3, 85-111. (7) Grate, J. W.; Klusty, M.; Barger, W. R.; Snow, A. W. Anal. Chem. 1990, 62, 1927-1924. (8) White, J.; Kauer, J. S.; Dickinson, T. A.; Walt, D. R. Anal. Chem. 1996, 68, 2191-2202. (9) Walt, D. R. Acc. Chem. Res. 1998, 31, 267-278. (10) Johnson, S. R.; Sutter, J. M.; Engelhardt, H. L.; Jurs, P. C.; White, J.; Kauer, J. S.; Dickinson, T. A.; Walt, D. R. Anal. Chem. 1997, 69, 4641-4648. (11) Dickinson, T. A.; Walt, D. R.; White, J.; Kauer, J. S. Anal. Chem. 1997, 69, 3413-3418. (12) Walt, D. R.; Dickinson, T.; White, J.; Kauer, J.; Johnson, S.; Engelhardt, H.; Sutter, J.; Jurs, P. Biosens. Bioelectron. 1998, 13, 695-699.

Analytical Chemistry, Vol. 73, No. 14, July 15, 2001 3441

Figure 1. Structural repeat units of PSFA, BSP3, and PMMA polymers.

report of a hydrogen bond acidic polymer/dye composite for optical sensing in an absorbance format, but not fluorescence.13 Strongly hydrogen bond acidic polymers are not commercially available but have been utilized on surface acoustic wave (SAW) sensors since the first reports on SAW sensor arrays.14-16 Subsequent work has described a number of such polymers,17-20 with recent work describing a polymer containing fluorinated bisphenol segments alternating with oligo(dimethylsiloxane) segments.19,20 These bisphenol-containing polymers are referred to as BSP polymers; the structural repeat unit of BSP3 is shown in Figure 1. Strongly hydrogen bond acidic materials are desirable because they are strong sorbents for hydogen bond basic vapor analytes, a result that is expected from fundamental principles and has been demonstrated experimentally.6,17-24 Accordingly, they have been used in the sensing of basic organophosphorus vapors. Hydrogen bond formation between organophosphorus compounds and hydrogen bond acidic polymers has been established using infrared spectroscopy.17 While acoustic wave sensors respond directly to the amount of vapor sorbed by a surface thin film, other sorbent polymerbased sensors incorporate additional components that play a role in signal transduction. Chemiresistors can be prepared by dispersing conducting particles in sorbent insulating polymer matrixes, as is currently under investigation using carbon particles to carry current through the film.25-27 Recent work has shown that the (13) Krech, J. H.; Rose-Pehrsson, S. L. Anal. Chim. Acta 1997, 341, 53-62. (14) Ballantine, D. S.; Rose, S. L.; Grate, J. W.; Wohltjen, H. Anal. Chem. 1986, 58, 3058-3066. (15) Rose-Pehrsson, S. L.; Grate, J. W.; Ballantine, D. S.; Jurs, P. C. Anal. Chem. 1988, 60, 2801-2811. (16) Grate, J. W.; Rose-Pehrsson, S. L.; Venezky, D. L.; Klusty, M.; Wohltjen, H. Anal. Chem. 1993, 65, 1868-1881. (17) Snow, A. W.; Sprague, L. G.; Soulen, R. L.; Grate, J. W.; Wohltjen, H. J. Appl. Polym. Sci. 1991, 43, 1659-1671. (18) Abraham, M. H.; Andonian-Haftvan, J.; Du, C. M.; Diart, V.; Whiting, G.; Grate, J. W.; McGill, R. A. J. Chem. Soc., Perkin Trans. 2 1995, 369-378. (19) Grate, J. W.; Kaganove, S. N.; Patrash, S. J.; Craig, R.; Bliss, M. Chem. Mater. 1997, 9, 1201-1207. (20) Grate, J. W.; Kaganove, S. N.; Patrash, S. J. Anal. Chem. 1999, 71, 10331040. (21) Chang, Y.; Noriyan, J.; Lloyd, D. R.; Barlow, J. W. Polym. Eng. Sci. 1987, 27, 693. (22) Barlow, J. W.; Cassidy, P. E.; Lloyd, D. R.; You, C. J.; Chang, Y.; Wong, P. C.; Noriyan, J. Polym. Eng. Sci. 1987, 27, 703-715. (23) Grate, J. W.; Snow, A.; Ballantine, D. S.; Wohltjen, H.; Abraham, M. H.; McGill, R. A.; Sasson, P. Anal. Chem. 1988, 60, 869-875. (24) Abraham, M. H.; Hamerton, I.; Rose, J. B.; Grate, J. W. J. Chem. Soc., Perkin Trans. 2 1991, 1417-1423. (25) Doleman, B. J.; Lonergan, M. c.; Severin, E. J.; Vaid, T. P.; Lewis, N. S. Anal. Chem. 1998, 70, 4177-4190.

3442

Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

sensitivity is based entirely on the sorption of vapors by the insulating sorbent polymer, with the carbon particles playing a passive role with regard to selectivity.28 In work reported in 1990, semiconducting phthalocyanine nanoparticles were incorporated into two-component Langmuir-Blodgett (LB) films for chemiresistor sensors.7 A hydrogen bond acidic polymer, fluoropolyol, served as the second component in an approach designed to maximize the sorption of organophosphorus vapors. This rational approach was successful in obtaining very sensitive chemiresistor sensors, with selectivity governed largely by the sorbent properties of the fluoropolyol matrix. It was noted that “Deliberate selection of the second film component to optimize sorption of the vapor to be detected should favorably influence the sensitivity and selectivity of the sensor” and “An array of chemiresistor sensors for use in a chemical sensor array detector using pattern recognition could be logically designed by varying the sorbent characteristics of the LB films on those sensors via the second film component.”7 This array design concept is currently exploited in array-based sensing using carbon particle/polymer composite chemiresistors with various polymeric film components. Use of hydrogen bond acidic polymers in dye/polymer fluorescent sensors may also offer performance benefits. First, use of these polymers will enhance the sorption of basic analytes of interest into the sensing film. This is always desirable, since only sorbed molecules are detected. Second, hydrogen bond acidic polymers would add diversity to arrays, which is desirable for discrimination using pattern recognition. Third, specific interactions of the polymer with the dye may set up a system for enhanced sensing of basic vapors. Most solvatochromic dyes have basic functional groups that may hydrogen bond with a hydrogen bond acidic polymer matrix. Sorption of basic vapors may compete with the dye and displace it from hydrogen-bonding sites, altering its local environment and altering its fluorescent properties. Although an early paper on Nile Red in poly(dimethylsiloxane) for fluorescent sensing demonstrated detection limits down to 10 ppm,29 the sensitivity of dye/polymer fluorescent sensors in an array format was rather low in the initial array studies, with detection limits in the one thousand to several thousand parts per million (ppm) range.8 In one paper, the lowest tested concentrations were one-third of the saturation vapor pressure.10 Recent work has focused on improving sensitivity by using dyes adsorbed on particle surfaces rather than dispersed in polymers, with detection limits for organic vapors down to 1 ppm when the signals from hundreds of sensor elements are summed to reduce noise.30 Detection limits for explosive-like vapors in the low ppb have been reported using dyes adsorbed on derivatized silica particles.31 Fluorescent conjugated polymer films for TNT detection have been reported.32,33 In the present paper, we examine the fluorescence and sensing properties of Nile Red dye in two hydrogen bond acidic polymers. (26) Lonergan, M. c.; Severin, E. J.; Doleman, B. J.; Beaber, S. A.; Grubbs, R. H.; Lewis, N. S. Chem. Mater. 1996, 8, 2298-2312. (27) Severin, E. J.; Doleman, B. J.; Lewis, N. S. Anal. Chem. 2000, 72, 658668. (28) Severin, E. J.; Lewis, N. S. Anal. Chem. 2000, 72, 2008-2015. (29) Barnard, S. M.; Walt, D. R. Environ. Sci. Technol. 1991, 25, 1301-4. (30) Dickinson, T. A.; Michael, K. L.; Kauer, J. S.; Walt, D. R. Anal. Chem. 1999, 71, 2192-2198. (31) Albert, K. J.; Walt, D. R. Anal. Chem. 2000, 72, 1947-1955. (32) Yang, J.-S.; Swager, T. M. J. Am. Chem. Soc. 1998, 120, 5321-5322. (33) Yang, J.-S.; Swager, T. M. J. Am. Chem. Soc. 1998, 120, 11864-11873.

Figure 2. Structure of Nile Red.

One is the BSP3 polymer mentioned above. The second is a hexafluoro-2-propanol-modified polystryrene,17 denoted as PSFA in the present paper. (This polymer has been referred to as P4V in other papers.) A polymer that is not hydrogen bond acidic, poly(methyl methacrylate) (PMMA), was also examined for comparison. PMMA and methyl methacrylate copolymers have been used previously in Nile Red/polymer fluorescent sensors.8-11,34 Structural repeat units of these polymers are compared in Figure 1. The structure of Nile Red is shown in Figure 2. Films of Nile Red in BSP3 and PSFA demonstrate exceptional fluorescence properties and high sensitivity for dimethyl methylphosphonate (DMMP) vapors. EXPERIMENTAL SECTION Materials. All dyes, PMMA, spectroscopic grade chloroform, ethanol, and other organic solvents used in our experiments were commercial products purchased from Aldrich. They were used without further purification. BSP3 polymer was prepared at Pacific Northwest National Laboratory (PNNL) according to published procedures.19 PSFA polymer was synthesized by Polymer Source, Inc. following the procedure of Snow et al.17 Film Preparation. Solutions for spin casting were prepared in spectroscopic grade chloroform. The polymer concentration was 0.05 M in repeat units for all polymers. Spin-cast films were prepared on glass substrates at a spinning rate of 1000 rpm. Thickness measurements were carried out by ellipsometry, giving measured thicknesses (error ( 20%) of 550, 540, and 420 Å for BSP3, PSFA, and PMMA polymers, respectively. Spectral Measurements. UV-visible spectra were measured using a Perkin-Elmer Lambda spectrometer. Uncorrected fluorescence spectra were recorded with a SLM 8001 fluorometer. All solution measurements were made in spectroscopic grade chloroform. All films were positioned into a quartz cuvette at an angle of 23° to the excitation beam. All spectral measurements were made at the ALTAIR Center. Vapor Exposures. Fluorescence measurements under exposure to saturated vapors of analyte were carried out by adding a small amount (50 µL) of the analyte solvent directly in the bottom of cuvette. After the cuvette was capped, the fluorescence signal was recorded as a function of time. Generally, final fluorescence spectra were recorded under equilibrium conditions 10-15 min after the solvent was added. All results described in this paper that involve exposures to saturated vapors were conducted using this liquid solvent injection method. Fluorescence responses to DMMP were also carried out by delivering gas-phase DMMP vapors into the curvette. Saturated DMMP vapor was collected in a 60-mL syringe from the headspace above the liquid solvent in a closed container. The vapor (34) Dickinson, T. A.; White, J.; Kauer, J. S.; Walt, D. R. Nature (London) 1996, 382, 697-700.

was diluted by pulling additional clean gas into the syringe or injecting vapor sample into a 100-mL dilution chamber containing clean gas. When repetitive dilutions were carried out, a fresh syringe was used for each step and the dilution chamber was flushed between steps. The gas flow inlet to the dilution chamber was connected to a tank of compressed dry nitrogen fitted with a needle valve and a Gilmont calibrated rotameter. The gas flow outlet from the dilution chamber was connected to a valve. Beyond the valve, the outlet was either left free for flushing the chamber or was connected to the cuvette for delivery of diluted DMMP. A separate flow path from the dry nitrogen tank could be used to flush the cuvette without going through the dilution chamber. Diluted DMMP was delivered to the cuvette by two methods. In the first method, the dilution chamber was connected to the cuvette and diluted vapor was displaced into the cuvette by flowing nitrogen into the dilution chamber inlet. Since this method dilutes the vapor as it is delivered to the cuvette, actual delivered concentrations are less than the reported concentration prepared in the dilution chamber. (If perfect mixing occurs in the dilution chamber during dilution, the exit concentration follows a exponential law.35) Using valves, the input gas to the cuvettte can be switched from clean nitrogen to diluted DMMP and back. In the second method, 60 mL of diluted DMMP vapor in the syringe was diplaced directly into the cuvette by depressing the syringe plunger over a ∼90-s period. This approach does not dilute the vapor as in the first method. Therefore, fluorescent responses to diluted vapor obtained by this method were greater than those observed when an equivalent vapor dilution was prepared in the dilution chamber and delivered by displacement with flowing nitrogen. The concentration of DMMP vapor was calculated from published data.5 Dilution by 60 times yielded 20 ppm, for example, by collecting 10 mL of saturated DMMP vapor, pulling an additional 50 mL of clean gas into the syringe, and injecting 10 mL of this mixture into the 100-mL dilution chamber. Dilutions to 100-200 ppb concentrations involved using a clean syringe to collect diluted DMMP from the dilution chamber for an additional similar dilution step. RESULTS AND DISCUSSION Survey of Polymer/Dye Combinations. Three polymers were considered as the matrix environment for fluorescent dyes (see Figure 1). Two of these, BSP3 and PSFA, were hydrogen bond acidic dyes selected in order to test their usefulness in detecting organophosphonate compounds in a fluorescent sensing mode. PMMA provided a control material that lacks hydrogen bond donating groups. Several dyes were considered and these are listed in Table 1. These dyes were selected on the basis of their high quantum yields in solution,36 their solvatochromism due to internal charge transfer in the excited state, and different emission range from blue to red. Most of these highly emissive compounds demonstrate the solvatochromic effect37 (excluding AO and AM), and their absorption band should be sensitive to changes in polarity, polarizability, acidity, or basicity of their solvation environment. (35) Barrat, R. S. Analyst 1981, 106, 817-849. (36) Berlman, I. B. Handbook of Fluorescence Spectra of Aromatic Molecules; Academic Press: New York, 1971. (37) Reichardt, C. Chem. Rev. 1994, 94, 2319-2358.

Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

3443

Table 1. Fluorescence Responses of Polymer/Dye Composite Films to Saturated DMMPa BSP3 ET λex ) 325 nm ANS λex ) 370 nm RB λex ) 515 nm AO λex ) 470 nm AM λex ) 400 nm NR λex ) 550 nm

PSFA

PMMA

λmax, nm

∆λmax

∆I/I

λmax, nm

∆λmax

∆I/I

381 458 558 498 500 626

+3 +10 +7 +11 +20 -32

2.8 3.5 0.25 0.22 3.2 4.3 (18)b

378 458 558

+2 +15

3.5 0.12

507 630

+14 -20

3.4 3.7

λmax, nm

∆λmax

∆I/I

378 444 562 506 518 598

0 +10

0.8 2.8

+7 +5 +2

0.95 2.1 1.0

a The initial position of the fluorescence maximum (λ max,), the maximum shift (∆λmax, signs “+” and “-” mean red and blue shifts), and the intensity change (∆I/I) of dye/polymer films after 100-s exposure under saturated DMMP vapors. ET, 7-ethoxycoumarin; ANS, 1,8 anilinonaphthalenesulfonic acid; RB, rhodamine B; AO, acridine orange; AM, 3-aminofluoranten; NR, Nile Red. The fluorescence intensity during exposure was monitored at λmax,. The dye concentrations were 2.5 × 10-4 M (excluding AO and AM with concentrations 5 × 10-5 and 2.5 × 10-2 M, respectively) and polymer concentration was 5 × 10-2 M (repeat units) in chloroform solution before spin casting. b The value in parentheses corresponds to excitation wavelength λex ) 530 nm and monitoring wavelength λ ) 590 nm

For example, Nile Red is a well-known solvatochromic dye that has been used as a fluorescence probe in biological applications38,39 and polarity studies in solvents and polymers.40,41 As noted in the introduction, NR has been intensively employed in the creation of polymer/dye fluorescence sensors in fiber-optic arrays.8,34 The selected dyes in BSP3, PSFA, and PMMA matrixes were surveyed by determining their responses (spectral shift and intensity changes) after 100 s of exposure to saturated DMMP vapor obtained by the addition of 50 µL of liquid DMMP to the cuvette as describd in the Experimental Section. Each polymer/ dye film was excited at the absorption maximum of the dye dispersed in the polymer before exposure to DMMP. The results are given in Table 1, where the spectral shift of the fluorescence maximum under DMMP exposure is listed as ∆λmax and the response intensity changes are given as ∆I/I. All these dyes in the strong hydrogen bond acidic polymer (BSP3, PSFA) matrixes exhibited responses to DMMP that were equal to or better than the same dyes in the PMMA matrix. It was noteworthy that even nonsolvatochromic dyes AO and AM in BSP3 and PSFA matrixes demonstrated a relatively good response to DMMP vapors. NR in BSP3 and PSFA matrixes were the only cases where the DMMP vapor caused a blue shift. In addition, the NR in BSP3 gave the strongest responses of all combinations considered, exhibiting both a dramatic blue shift and a strong fluorescence enhancement. On this basis, NR was selected for further detailed study. Fluorescence Properties of Nile Red in BSP3, PSFA, and PMMA Polymer Matrixes. The spectra of NR in BSP3 and PSFA films show a significant red shift relative to NR in either PMMA or chloroform solution. In addition, the NR fluorescence intensity in BSP3 and PSFA matrixes is considerably lower than that in the PMMA matrix. Normalized absorption and fluorescence spectra of the NR in BSP3, PSFA, and PMMA films and chloroform solution are compared in Figure 3. The observed red shift and fluorescence quenching can be explained by the influence of polarity and hydrogen bond acidity (38) Ruvinov, S. B.; Xiang-Jiao, Y.; Parris, K. D.; Banik, U.; Ahmed, S. A.; Miles, E. W.; Sackett, D. L. J. Biol. Chem. 1995, 270, 6357-6369. (39) Daban, J. R.; Montserrat, S.; Bartolome, S. Anal. Biochem. 1991, 199, 162-168. (40) Vauthey, E. Chem. Phys. Lett. 1993, 216, 530-536. (41) Dutta, A. K.; Kamada, K.; Ohta, K. J. Photochem. Photobiol., A 1996, 93, 57-64.

3444 Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

on charge transfer in the NR π-electron system. Polar environments typically induce a fluorescence red shift and the intramolecular hydrogen-bonding interaction leads to fluorescence quenching.42 Both BSP3 and PSFA polymers have high dipolarity and hydrogen bond acidity and can form hydrogen bonds with NR. Thus, initially (before DMMP exposure) NR demonstrates low quantum yields in BSP3 and PSFA films. The concentration of the dye in the polymer matrix is another important consideration for luminescence sensor optimization. Additional quenching of NR fluorescence in spin-cast BSP3 films was observed when the NR concentration in the chloroform casting solution exceeded 2.5 × 10-4 M. Similar results were observed for PSFA. Most likely, the reason for such quenching is dye aggregation leading to an exciton coupling effect,43,44 which is typical for dyes having planar conjugated chromophores. Confirmation of aggregation can be found in the dependence of the UV-visible spectra on NR concentration, shown for BSP3 films in Figure 4. Increasing the NR concentration leads to the appearance of a new blue-shifted band associated with formation of the collective excited states. It follows from the exciton model for H-aggregates43 that the highest oscillator strength should be concentrated in the excited states with high energy (blue-shifted band in UV-visible spectra). Aggregate excited states with low energy have low oscillator strength. Thus, after fast termalization preceding an emission, low-energy states with a weak emission will be populated. This effect leads to fluorescence quenching (or self-quenching) in molecular aggregates. To avoid the effects of aggregates on fluorescence performance, sensing experiments were conducted using films prepared from chloroform solution with NR concentration 2.5 × 10-4 M. Influence of DMMP on Nile Red Properties in Polymer Matrixes. Figure 5 demonstrates the pronounced NR fluorescence enhancement and the spectral blue shift in BSP3 and PSFA matrixes under exposure to DMMP-saturated vapor and rather small changes in the PMMA matrix. These results are equilibrium spectra obtained after injecting 50 µL of liquid DMMP into the cuvette, as described in the Experimental Sections. (42) de Silva, A. P.; Gunarate, T. G.; Huxley, A. J. M.; McCoy, C. P. Chem. Rev. 1997, 97, 1515-1566. (43) Kasha, M. Spectroscopy of the Excited State; Plenum Press: New York, 1976. (44) Levitsky, I. A.; Kishikawa, K.; Eichhorn, S. H.; Swager, T. M. J. Am. Chem. Soc. 2000, 122, 2474-2479.

Figure 3. Normalized absorption (a) and fluorescence (b) spectra of NR in solution and polymer matrixes: (1) chloroform solution (10-6 M); (2) PMMA; (3) BSP3; (4) PSFA polymers. Polymer films were spin cast onto glass substrate (1000 rpm) from chloroform solution with NR and polymer concentrations of 2.5 × 10-4 and 5 × 10-2 M (repeat units), respectively.

Figure 4. Nile Red UV-visible spectra for different dye concentrations in BSP3 spin-cast films. NR concentration values are given for solution used for film preparation.

The observed strong changes in the dye emission and spectral shifts are consistent with a competition between NR and DMMP for interaction with hydrogen-bonding sites in the BSP3/PSFA polymer. As mentioned above, most likely NR forms hydrogen bonds with these polymers resulting in the low quantum yield and the red shift. NR has four basic sites (two nitrogen and two oxygen atoms) with the potential to hydrogen bond to the matrix

Figure 5. Nile Red fluorescence spectra before (1) and after (2) DMMP exposure in BSP3 (a), PSFA (b), and PMMA (c) spin-cast films. Curves 3 are curves 1 at ×5 and ×10 for PSFA and BSP3, respectively.

polymer. The solute DMMP is a strong hydrogen bond basic vapor that also seeks to hydrogen bond with polymeric hydroxyl groups. In general, phosphonates are stronger hydrogen bond bases than amines and much stronger hydrogen bond bases than ketones or ethers. The relative hydrogen bond basicities of various Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

3445

Table 2. Hydrogen Bond Basicity vapor name

ΣβH 2

vapor name

ΣβH 2

n-hexane acetonitrile aniline diethyl ether methyl ethyl ketone

0 0.32 0.41 0.45 0.51

pyridine 2-propanol dimethylamine N,N-dimethylformamide dimethyl methylphosphonate

0.52 0.56 0.67 0.74 1.05

functional groups are compared in Table 2 using Abraham’s ΣβΗ 2 parameter.4-6 Thus, DMMP molecules having higher binding affinity to BSP3/PSFA sites than NR can break NR hydrogen bonds “releasing” the dye to a less hydrogen-bonded state. The fluorescence spectrum then more closely resembles the spectrum of NR in a PMMA matrix. The dye has more intensive emission, and the spectrum is blue shifted relative to the spectrum of the dye in BSP3/PSFA in the absence of DMMP. In PMMA, the dye fluorescence is not so strongly red shifted or quenched by interaction with the polymer prior to DMMP exposure. As a result, DMMP exposure then results in relatively small NR fluorescence changes. Figure 6a compares the fluorescent responses of the three polymer/NR composite films at 635 nm, using a common excitation wavelength of 535 nm. The latter represents approximately the mean absorbance maximum of NR in the various matrixes (see Figure 3). The detection wavelength is the fluorescence maximum in the BSP3 and PSFA polymers. The BSP3 matrix provides a higher fluorescence response signal than the PSFA matrix (Figure 6a). The fluorescence change for PMMA polymer is almost negligible with respect to BSP3 and PSFA polymers. Optimization of Excitation and Detection Wavelengths. The sensitivity for particular dye/polymer/vapor combinations can be improved by selecting excitation and detection wavelengths that are optimal for the particular system. Figure 7 shows the visible absorption spectra for NR in the BSP3 matrix before and after DMMP exposure. The absorbance maximum is significantly blue shifted, indicating that performance may be improved by choosing a shorter excitation wavelength (530 nm) that matches the absorbance peak under vapor exposure. Similarly, the detec-

tion wavelength can be optimized at the peak fluorescence wavelength under DMMP exposure (590 nm) as shown in Figure 5. Fluorescent responses using these wavelengths are compared with the previous wavelength selection in Figure 6b. It is apparent that this improves the flourescent response sigificantly, a result also indicated by the ∆I/I values in Table 1 for NR/BSP3. The optimal λex and λdet values are at shorter wavelengths with respect to the position of maximum of visible absorption and fluorescence bands before DMMP exposure since NR interaction with DMMP leads to the blue shift of these spectra. Thus, the spectral shift can provide the effective amplification of the response signal. Diluted DMMP Exposures. Initial experiments at diluted DMMP vapor concentrations have demonstrated detection of DMMP at ppm and sub-ppm levels. Vapor dilutions were performed as described in the Experimental Section. Figure 8 shows a typical time trace for NR emission in the BSP3 matrix on switching between flowing nitrogen gas and flowing diluted DMMP vapor. The sensing material demonstrates rapid responses to the analyte vapor with rapid, nearly complete recovery. In this experiment, a 20 ppm sample of DMMP was prepared in a dilution chamber located upstream from the cuvette before each exposure and delivered by displacement with flowing nitrogen, as described in the Experimental Section. There is clearly plenty of signal in response to DMMP in the low-ppm range. Experiments with direct injection of 20 ppm DMMP samples from a syringe into the cuvette yielded fluorescence intensities that were 30-40% greater than those seen in Figure 8. Additional experiments performed by direct injection of DMMP diluted to 100-200 ppb levels demonstrated 20-30% increases in the fluorescence intensity compared to clean gas. Although more rigorous experiments will be required to establish the detection limits, actual detection at sub-ppm levels indicates promising sensitivity for these fluorescent films. Interferences. The NR/BSP3 films were also tested against common organic vapors and water as potential interferences, all at saturated concentration. These results are presented in Figure 9 in two ways. In Figure 9a, relative responses are presented with each vapor at saturated vapor concentration; i.e., all vapors are at

Figure 6. Time traces of NR fluorescence in BSP3, PSFA, and PMMA films at λex ) 550 nm and λdet ) 635 nm in response to saturated DMMP vapors (a). Time traces of NR fluorescence in BSP3 film at different excitation and monitoring wavelengths (b). The slit was closed while DMMP liquid was added to the cuvette between the collection of the baseline and the recording of the response. All initial intensities prior to vapor exposure were normalized to that of the NR/BSP3 composite film. 3446 Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

Figure 7. Visible absorption spectra of NR in BSP3 film before and after saturated DMMP exposure.

Figure 9. Relative responses of a NR/BSP3 film (λex ) 530 nm, λdet ) 590 nm) to saturated vapor concentrations without (a) and with (b) normalization on the concentration. Signals were determined after 300 s of vapor exposure.

Figure 8. Time trace of the NR fluorescence signal in a BSP3 polymer matrix (λex ) 530 nm, λdet ) 590 nm) under alternating exposures to diluted DMMP vapors (see text) and clean nitrogen gas.

P/Psat ) 1, where P is the vapor pressure and Psat is the saturated vapor pressure. The greatest responses are observed for DMMP, and these are significantly higher than those to water vapor, a ubiquitous potential interference. In Figure 9b, the sensitivities were compared by normalizing to vapor concentration in order to compare all vapors at equal concentration (in ppm). The sensitivity to DMMP exceeds that to water, benzene, ethanol, and chloroform by 58, 750, 37, and 1000 times, respectively, after normalization. Discussion. The interaction of Nile Red with the strongly hydrogen bond acidic polymers greatly reduced the fluorescent intensity. This reduction in fluorescent intensity set up the dye/ polymer system for emission increases upon DMMP sorption, leading to sensing films with lower detection limits than those reported previously for Nile Red/polymer fluorescent films. We attribute the high sensitivity to the use of a polymer designed to sorb the analyte of interest and the competition set up between the dye and the analyte for polymer interaction sites. Thus, the use of a polymer designed to favorably sorb the analyte of interest and the design for competing dye/polymer and analyte/ polymer interactions resulted in sensing films with superior sensitivity.

The sensitivity achieved to date represents loading the BSP3 polymer with the maximum dye allowable while avoiding dye aggregates. We have not optimized the ratio of dye molecules relative to polymer repeat units, so the potential exists for even better sensitivity. The dye polymer/repeat unit ratio can be altered either by adjusting the dye concentration in the polymer or by adjusting the synthesis of the polymer to increase the length of the oligo(dimethylsiloxane) segments, thus decreasing the density of interactive bisphenol units in the material. The synthetic method for BSP3 is versatile in that it allows this type of synthetic adjustment to adapt the polymer to the requirements of different sensing formats.19,45,46 Our approach to fluorescent sensing is in some ways analogous to the dye/matrix sensing films described previously by Dickert that were used in an absorbance sensing mode.47,48 Crystal violet lactone dispersed in matrixes including bisphenol A yields a deep blue film. In this acidic matrix, an oxygen carbon bond of the covalent form of the dye is broken to yield a highly colored triphenylmethane carbocation. These films respond to basic vapors that disrupt the dye/matrix hydrogen-bonding network and reduce the concentration of dye in the triphenyl carbocation form with a decrease in absorbance. A detection limit of 10 ppm for gas-phase acetone has been reported for these absorbance-based sensors. Though not demonstrated in the present paper, it is anticipated that NR/BSP3 films will be useful in sensor arrays. These films will offer improved sensitivity to vapors such as organophosphorus compounds and increase the chemical diversity in the array. (45) Grate, J. W.; Kaganove, S. N. Polym. News 1999, 24, 149-155. (46) Grate, J. W.; Kaganove, S. N.; Nelson, D. A. Chem. Innovations 2000, 30 (11) 29-37. (47) Dickert, F. L.; Lehmann, E. H.; Schreiner, S. K.; Kimmel, H.; Mages, G. R. Anal. Chem. 1988, 60, 1377-1380. (48) Dickert, F. L.; Schreiner, S. K.; Mages, G. R.; Kimmel, H. Anal. Chem. 1989, 61, 2306-2309.

Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

3447

Hydrogen bond acidic polymers have not been used in such arrays previously, and the emission spectrum shifts in response to DMMP are in opposite direction to those observed for most NR/ polymer films in response to organic vapors. The NR/BSP3 and NR/PSFA films provided spectral shifts in the presence of DMMP that were in the opposite direction to the NR/PMMA film studied in this paper. The direction is also the opposite to NR in poly(dimethylsiloxane) exposed to benzene29 and NR in polycaprolactone exposed to organic solvents.9 Thus, we have established an experimental basis for expecting the designed dye/polymer films to yield arrays with improved performance for the detection of organophosphorus vapors in particular and for distinguishing organic vapors in general.

3448

Analytical Chemistry, Vol. 73, No. 14, July 15, 2001

ACKNOWLEDGMENT The authors acknowledge funding from the U.S. Departement of Energy (DOE). I.L. and S.G.K. are grateful for funding by DOE Grant DE-FG02-99ER82737. J.W.G. is grateful for funding from the DOE Office of Nonproliferation and National Security, NN-20. The Pacific Northwest National Laboratory is a multiprogram national laboratory operated for the Department of Energy by Battelle Memorial Institute. Received for review December 19, 2000. Accepted April 16, 2001. AC0014966