Enhanced emission based optical carbon dioxide sensing in presence of perfluorochemicals (PFCs)

Sensors and Actuators B 115 (2006) 672–677

Enhanced emission based optical carbon dioxide sensing in presence of perfluorochemicals (PFCs) Kadriye Ertekin ∗ , Serap Alp University of Dokuz Eylul, Faculty of Arts and Sciences, Department of Chemistry, 35160 Tinaztepe, Izmir, Turkey Received 16 July 2005; received in revised form 23 October 2005; accepted 25 October 2005 Available online 6 December 2005

Abstract An emission based simple and fast method has been proposed for the determination of gaseous and dissolved CO2 . Perfluorochemicals (PFCs), also known as medical gas carriers have been used for the first time together with newly synthesized fluorophore; 4-[(p-N,N-dimethylamino)benzylidene]2-phenyloxazole-5-one (DPO), in ethyl cellulose matrix. In the first stage of the study precise determination of the acidity constant (pKa ) of the DPO has been performed in the employed polymer matrix. In second stage, the response of the sensor composition to gaseous and dissolved CO2 has been evaluated in the absence and presence of the PFC. It should be noted that the solubility of CO2 in fluorocarbons is about 10–20 times as that observed in the parent hydrocarbons or in water, respectively, and once doped into the sensing film, considerably enhance the response of the sensing agent. © 2005 Elsevier B.V. All rights reserved. Keywords: Perfluorochemicals (PFCs); Optical carbon dioxide sensor; Emission based CO2 sensing

1. Introduction Monitoring of carbon dioxide levels has of great importance in environmental and biomedical analysis as well as industrial processes. Probably, in near future, the monitoring of pCO2 levels of atmosphere and dissolved CO2 levels of seawater will be compulsory. For this reason investigations on CO2 sensor development gained more and more importance. Most of the optical CO2 sensor designs utilize indicator dyes with pKa values between 7.4 and 10.0, which are doped into the polymer matrices. The number of the fluorescent pH indicators for carbon dioxide sensing is rather limited. The fluorescent pH indicator 1-hydroxypyrene-3,6,8-trisulfonate (HPTS) has been one of the most frequently used indicator dyes in optical CO2 sensor design by chemists [1–6]. HPTS is an inexpensive, highly water-soluble pH indicator with a pKa of ∼7.3 in aqueous buffers. However, in most of the sensor designs, the indicator dye must be introduced into lipophilic matrices in water-nonsoluble ion pair form. Weigl and Wolfbeis proposed the method

∗

Corresponding author. Tel.: +90 232 412 8689; fax: +90 232 453 4188/2153. E-mail addresses: [email protected], [email protected] (K. Ertekin). 0925-4005/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2005.10.036

of lipophilic extraction of ion pair of HPTS with a quaternary ammonium salt into the membrane phase and this approach has been applied many times [1–6]. Some investigations on pCO2 sensor design with HPTS resulted with the commercialization of products [7–10]. Tabacco et al. used the pH-sensitive fluorescent dye carboxy-SNAFL-1 in immobilized form for CO2 sensing purpose in surface seawater [11]. Amao and Nakamura used fluorescent dye-tetraphenylporphyrin (TPP) together with pH indicator-alpha-naphtholphthalein [12]. Nakamura and Amao offered new optical CO2 sensors based on the luminescence intensity changes of europium (III) complex; tris(thenoyltrifluoroacetonato)europium dihydrate depending on the absorption change of pH indicator dyes, thymol blue, phenol red or cresol red, in presence of CO2 [13,14]. Mills et al. developed an alternative optical sensor to commercially available capnometers based on infrared sensing of CO2 , widely used for clinical analysis [15]. In this article, an emission based, simple and fast method is proposed for the determination of gaseous and dissolved CO2 . A newly synthesized fluorophore, 4-[(p-N,Ndimethylamino)benzylidene]-2-phenyloxazole-5-one has been used for carbon dioxide sensing together with “carbon dioxide carrier” perfluorochemical (PFC) in ethyl cellulose matrix.

K. Ertekin, S. Alp / Sensors and Actuators B 115 (2006) 672–677

As far as we know, PFCs are used for the first time as carbon dioxide carriers to enhance the solubility of CO2 . PFCs are clinically important organic compounds in which all hydrogen atoms have replaced with fluorine atoms, and therefore they can dissolve large volumes of oxygen and carbon dioxide. Gas molecules occupy so-called ‘molecular cavities’ within liquid PFCs, but no chemical reactions are involved. This solubility is related to the molecular volume of the dissolving gas and decreases in the order of CO2  O2 > CO > N2 [16]. Therefore, a PFC doped ethyl cellulose matrix is expected to exhibit better carbon dioxide solubility. In this paper we describe the formulation and characterization of a PFC doped film sensor with response and recovery characteristics, which appear to be suitable for both gaseous and dissolved CO2 analysis.

2. Experimental 2.1. Materials The fluorescent pH indicator 4-[(p-N,N-dimethylamino)benzylidene]-2-phenyloxazole-5-one (DPO) was synthesized and purified according to the literature [17]. Dichloromethane, ethanol and toluene were obtained from Aldrich and used without further purification. The perfluorochemical [nonadecafluorodecanoic acid; CF3 (CF2 )8 CO2 H] was purchased from Fluka. Ethyl cellulose (with an ethoxy content of 46%) was purchased from Aldrich (Steinheim, Germany). Teflon AF (type 1600) was purchased from Du Pont polymers (Switzerland). Hexafluorobenzene and tetrabutylammonium hydroxide (TBAOH, in ispropanol) were obtained from Merck (Germany). The polyester support (Mylar type) was provided from DuPont, Turkey. Acid solutions and buffers used for investigating the pH effects were prepared with high quality pure water. Carbon dioxide and nitrogen of 99.9% purity were obtained from Gunes Company, Izmir, Turkey. Gas flows were measured with a highly sensitive flow meter and mixed properly in a homemade gas mixing chamber. All of the experiments were performed at barometric pressure and room temperature of 25 ± 2 ◦ C. Carbon dioxide sensing was carried out using different partial pressures of CO2 and nitrogen gases in the range of 0–100% entering the home made gas mixing chamber by controlling the gas flow rates with sensitive flow meters. The final flow rate of the gas mixture was 500 mL/min.

673

2.2. Instruments The absorption spectra of the solutions and sensor films were recorded by using a Shimadzu UV-1601 UV–vis spectrophotometer. Steady state fluorescence and excitation spectra were measured using Varian Cary Eclipse Spectrofluorometer with a Xenon flash lamp as the light source. The excitation and emission slits were set to 5 nm during CO2 sensing studies. The detector voltage was kept at 600 V. 2.2.1. Polymer film preparation The optode polymer films were prepared by mixing 500 mg of ethyl cellulose in 10 mL of a toluene/ethanol mixture (80/20, v/v). Then 66 ␮L of the TBAOH and 0.3 mmol of DPO were added into the mixture and mixed in a glass vial. Later, 0.3 mmol/120 ␮L ethanolic solution of PFC was added into the sensor composition. The resulting mixture was saturated with 100% CO2 gas for 2 min. Thereby, the DPO was converted into the acidic (faintly yellow) form. In the next step, it was bubbled with nitrogen for 30 min and completely converted to the basic form (orange color) again. The resulting mixtures were spread onto a 125 ␮m polyester support (Mylar TM type) with a spreading device. The dried films were covered with the non-viscous Teflon solution by spreading technique. The Teflon membrane is permeable for gaseous but impermeable for ionic species. Since the Teflon membrane would reject the acidic protons, some of the films are separated without Teflon coating for pKa determination studies. All of the thin films were kept in a desiccator to avoid the damage from ambient air of the laboratory. The polyester support was optically fully transparent, ion impermeable and exhibited good adhesion to ethyl cellulose (EC). The most important function of the polyester was to act as a mechanical support because the thin EC films were difficult to handle. Each sensing film was cut to 1.2 cm width, fixed diagonally into the sample cuvette, and the absorption or emission spectra were recorded. 3. Results and discussion 3.1. Choice of indicator dye DPO was chosen due to its promising characteristics such as long wavelength absorption (λabs max = 465 nm), excitation and emission maximum (λexc = 480 nm, λem max max = 532 nm) in visible region, high molar extinction coefficient (εmax = 177,000 L mol−1 cm−1 ) and, excellent solubility in PFC doped EC matrix. Fluorescence emission spectra of DPO in the presence and absence of PFC and quaternary ammonium base TBAOH was shown in Fig. 1. On the other hand, the knowledge of dissociation constant (pKa ) of DPO in PFC doped EC matrix is important in order to provide information on chemical reactivity range of the indicator dye. Indicators with pKa values between 6.8 and 10.0 are essential for sensitive pCO2 optodes. In order to evaluate the availability of DPO for CO2 sensing, determination of the acidity constant has been performed in the EC matrix. pH induced emission spectra of the Teflon coating-free sensor membranes

674

K. Ertekin, S. Alp / Sensors and Actuators B 115 (2006) 672–677

linear fitting algorithm of Gauss–Newton–Marquardt method,
pKa = pH + log

(Ix − Ib ) (Ia − Ix )



where Ia and Ib are the fluorescence intensities of acidic and basic forms and Ix is the intensity at a pH near to the pKa . The calculated pKa value reveals that the DPO dye can be used as a fluorescent CO2 indicator when doped into EC matrix in presence of PFC. 3.2. Sensing scheme

Fig. 1. Fluorescence emission spectra of DPO in presence and absence of perfluorochemical (PFC) and quaternary ammonium base TBAOH. (a) DPO dye + PFC, (b) DPO dye + PFC + TBAOH and (c) DPO dye.

The pCO2 sensor matrix contains the fluorescent pH indicator DPO, the CO2 carrier PFC, and a quaternary ammonium base (TBAOH) in ethyl cellulose. TBAOH is added as a lipophilic counter ion to stabilize the DPO in the matrix. The presence of TBAOH tunes the sensitivity of the sensor and enhances the stability since it also acts as a sink for acidic species. The sensing scheme is based on two processes, the first being the diffusion of humidified CO2 through the Teflon membrane into the sensor slide, and the second being the reaction with highly fluorescent DPO. The indicator is functional owing to the very limited water content of the matrix, which causes the formation of carbonic acid with CO2 . DPO embedded in EC undergoes a completely reversible protonation–deprotonation equilibrium when exposed to carbonic acid and becomes protonated. Sensor slides respond to the proton resulting from conjugation change throughout the molecule by a decrease in emission intensity, which has been used as the analytical signal. 3.3. Effect of presence of PFCs on the sensor response

Fig. 2. pH induced emission spectra of the Teflon free sensor membrane in the pH range of 9.0–4.0.

were recorded in the pH range of 9.0–4.0 (see Fig. 2). During measurements a relative signal change of approximately 100% was attained. Fig. 3 shows the plot of normalized fluorescence intensity (Ix /Ib ) versus measured pH values in buffered acid solutions. pKa value was found as 6.9 from the inflection point of the plot. The pKa value was also calculated as 6.97 by using non-

Effect of the presence of PFCs in the matrix material on sensor performance can be seen from Fig. 4. The spectral data in Fig. 4I and II were acquired in the absence and presence of the PFCs and illustrate the effect of varying concentrations of CO2 in the partial pressure range of 0.0–5.0%. The given range covers the clinically important levels of CO2 (0.1–5.0%). The two sensor membranes shown in Fig. 4I and II contain same amount of dye, however, the PFC containing one exhibits a brighter fluorescent intensity and enhanced performance on breath exhalation test (5.0% or 50,000 ppm CO2 ). The relative signal change caused by exhalation was 21.4 and 53.6% in the absence and presence of PFC, respectively. This means a 2.5-fold enhancement in relative signal change in solid phase, and consequently in sensor performance. In a reservoir-type sensor design, amount of the PFC in sensor composition, and consequently the sensor performance can be increased, but in our case, a decrease in the adhesion level is observed between the matrix material and Mylar polyester support due to the presence of high levels of PFCs. 3.4. Spectral response of the sensor to gaseous and dissolved CO2

Fig. 3. Plot of normalized fluorescence intensities (Ix /Ib ) vs. measured pH values of eflon-free sensor membrane in buffered acid solutions.

The carbon dioxide sensitive cocktail composition yields an emission peak at λmax = 532 nm which responds to CO2

K. Ertekin, S. Alp / Sensors and Actuators B 115 (2006) 672–677

675

Fig. 6. The non-linearized calibration plot of the Teflon coated sensor slide (I0 /I) vs. pCO2 (a) in presence and (b) absence of the PFC in the partial pressure range of 0–80% pCO2 .

commercial gas phase sensors work up to or around 5% pCO2 . Fig. 6 illustrates the non-linearized calibration plot of sensor composition (I0 –I) versus pCO2 in the absence and presence of PFC. As for Fig. 6, it can be concluded that presence of PFC increases the limit of linearity of the sensor from 20 to 40% pCO2 in the gas phase. 3.5. Dissolved CO2 sensing studies Fig. 4. Effect of the presence of PFC in the matrix material on sensor performance. Sensor performance after exposure to human breath (I) in the absence and (II) in the presence of PFC.

by a decrease in emission intensity that can be used as the analytical signal (Fig. 5). Upon exposure to partial CO2 pressures from 0 to 80%, the membrane exhibits a 66% relative signal change in emission intensity. According to the results of the preliminary emission based measurements, the efficient dynamic working range of the sensor was found between 0 and 40% CO2 . Human breath (after exhalation) contains approximately 5% (50,000 ppm) of CO2 and most of the offered

Sensor performance was also characterized for low concentrations of dissolved CO2 . Standard solutions were prepared freshly from 1 M NaHCO3 stock solution prior to the measurement. Dilute solutions of sodium hydrogencarbonate were used to form CO2 calibration graphs. CO2 -free standard solutions were prepared with bidistilled water after boiling, bubbling with nitrogen, and kept in closed containers. Concentrations were calculated using the following equations and constants [18,19]. CO2(g) ↔ CO2(aq) ,

log K1 = −0.47

CO2(aq) + H2 O ↔ H2 CO3 ,

log K2 = −1.41

H2 CO3 ↔ H+ + HCO3 − ,

log K3 = −6.38

HCO3 − ↔ H+ + CO3 −2 ,

log K4 = −10.38

In a HCO3 − solution the relationship between the partial pressure of dissolved CO2(g) (pCO2 ) and proton concentration is as follows: [H+ ] + [[H+ ] [Na+ ] − Kw [H+ ] K3 ([H+ ] + 2K4 ) 3

αPCO2 = [H2 CO3 ] =

Fig. 5. Response of the Teflon coated sensor membrane upon exposure to partial CO2 pressures from 0 to 80%. (a) 0%, (b) 3%, (c) 10%, (d) 15%, (e) 20%, (f) 30%, (g) 40%, (h) 60% and (i) 80% of pCO2 .

2

where α = K1 K2 [H2 O], Kw is the water dissociation constant and [Na+ ] is the concentration of sodium ions present. Calculated H2 CO3 concentrations (the sum of hydrated CO2(aq) and real H2 CO3 ), and corresponding pCO2 values in solutions made up from NaHCO3 are as follows.

676

K. Ertekin, S. Alp / Sensors and Actuators B 115 (2006) 672–677

Total NaHCO3 (mol L−1 )

[H2 CO3 ] (mol L−1 )

pCO2 in RTILs (atm)

4 × 10−6

3.90 × 10−7

10.04 × 10−6 1.79 × 10−5 7.17 × 10−5 5.98 × 10−4 4.78 × 10−3

2 × 10−5 2 × 10−4 2 × 10−3 2 × 10−2

6.97 × 10−7 2.79 × 10−6 2.33 × 10−5 1.86 × 10−4

Twenty microlitres portions of standard solutions of NaHCO3 were added into the sensor containing cuvette. The change in fluorescence intensity due to the addition of different concentrations of dissolved CO2 was measured. The time when 90% equilibrium was reached (τ 90 ) is recorded as the response time. All the experiments were carried out at room temperature of 25 ± 1 ◦ C. The dynamic range for the detection of H2 CO3 (CO2 + H2 O) was between 3.96 × 10−6 and 2.0 × 10−2 mol L−1 . The corresponding calibration graph exhibited good linearity. It should be noted that the X-axis (concentration) of the calibration graph is logarithmic (y = 28.687x + 307.49 and R2 = 0.9944) and covers a large scale. However, in solution phase studies attained relative signal chance was not as high as the gas phase studies. This can be attributed to the presence of additional Teflon membrane. The linearized calibration graph of the sensor membrane is shown in Fig. 7.

Fig. 8. Response characteristics of the sensor film in an alternating atmosphere of 80% CO2 and 100% N2 .

than response. In the present film, enhanced response (larger dynamic range) and reproducibility is due to the formulation. However, the upper limit of response time can be improved. As well as the dye took place in the formulation, presence of the PFC and film thickness has great importance in determining the rate of response of the sensor. In our case, film thickness is approximately 15 ␮m, which can be tuned and shorter response times can be reached. The reproducibility of the optical responses was assessed by repeatedly introducing of 80% CO2 and 100% N2 , and was found to be fully reversible. The level of reproducibility was characterized with a R.S.D. of 5.8% (n = 7 cycle).

3.6. Response and reproducibility

4. Conclusion

Fig. 8 illustrates the response characteristics of the sensor film in an alternating atmosphere of 80% CO2 and 100% N2 . Response time is defined as the time taken for the film to attain 90% of its original signal intensity when the gas is changed (τ 90 ). The presence of PFC’s in matrix composition has a considerable effect on the response and recovery times of the films. In the absence of the PFC’s, the response and recovery times (τ 90 ) of oxygen sensing membranes, after exposure to 80 and 0% pCO2 , were found to be 96 and 126 s, respectively. In presence of PFC, response time is between 15 and 16 s and recovery time is about 25–30 s (see Fig. 8). The slightly hyperbolic recovery lines also support that recovery from a CO2 atmosphere is slower

In this study, the PFC for the first time has been used as carbon dioxide carrier to enhance the solubility of CO2 in solid matrix. In addition, due to the compatibility of the sensor composition with the solid-state optical components (in particular with LED’s and fiber optics) and fast response of the membrane to gaseous and dissolved CO2 , this composition looks like a promising CO2 sensor material. Acknowledgement Funding this research was provided by the projects (04 kb Fen 104, 105 and 019), Scientific Research Funds of Dokuz Eylul University and The Scientific and Technological Research Council of Turkey (TUBITAK). References

Fig. 7. The linearized calibration graph of the sensor membrane after exposure to the NaHCO3 solutions in the concentration range of 3.96 × 10−6 –2.0 × 10−2 mol L−1 .

[1] B.H. Weigl, O.S. Wolfbeis, New hydrophobic materials for optical carbon dioxide sensors based on ion pairing, Anal. Chim. Acta 302 (2–3) (1995) 249. [2] B. M¨uller, P.C. Hauser, A fluorescence optical sensor for low concentrations of dissolved carbon dioxide, Analyst 121 (3) (1996) 339. [3] O.S. Wolfbeis, B. Kovacs, K. Goswami, S.M. Klainer, Fiber optic fluorescence carbon dioxide sensor for environmental monitoring, Microchim. Acta 129 (3–4) (1998) 181. [4] C. von Bueltzingsloewen, A.K. McEnvoy, C. McDonagh, B.D. MacCraith, I. Klimant, C. Krause, O.S. Wolfbeis, Sol–gel based optical carbon dioxide sensor employing dual luminophore referencing for application in food packaging technology, Analyst 127 (2002) 1478. [5] G. Neurauter, I. Klimant, O.S. Wolfbeis, Fiber-optic microsensor for high-resolution pCO2 sensing in marine environment, Fresenius’ J. Anal. Chem. 366 (2000) 481.

K. Ertekin, S. Alp / Sensors and Actuators B 115 (2006) 672–677 [6] K. Ertekin, I. Klimant, G. Neurauter, O.S. Wolfbeis, Chracterization of a reservoir-type capillary optical micro sensor for pCO2 measurements, Talanta 59 (2) (2003) 261. [7] M. Yafuso, J.K. Suzuki, Gas sensors, US Patent 4,824,789 (1989). [8] J.E. Alderete, A.D. Olstein, S.C. Furlong, Optical carbon dioxide sensor, and associated methods of manufacture, US Patent 5,714,121 (1998). [9] W. Adrian, B.S. Mark, Optical carbon dioxide sensors, Novartis AG, US Patent 6,338,822 (2002). [10] S.C. Furlong, Simultaneous dual excitation/single emission fluorescent sensing method for pH and pCO2 , US Patent 5,672,515 (1997). [11] M.B. Tabacco, M. Uttamlal, M. McAllister, D.R. Walt, An autonomous sensor and telemetry system for low-level pCO2 measurements in seawater, Anal. Chem. 71 (1) (1999) 154. [12] Y. Amao, N. Nakamura, An optical sensor for CO2 using thymol blue and europium (III), Sens. Actuators B Chem. 100 (3) (2004) 347. [13] N. Nakamura, Y. Amao, Optical CO2 sensor with the combination of colorimetric change of pH indicator and internal reference luminescent dye, Bull. Chem. Soc. Jpn. 76 (7) (2003) 1459. [14] N. Nakamura, Y. Amao, Optical sensor for carbon dioxide combining colorimetric change of a pH indicator and a reference luminescent dye, Anal. Bioanal. Chem. 376 (5) (2003) 642. [15] A. Mills, A. Lepre, L. Wild, Breath-by-breath measurement of carbon dioxide using a plastic film optical sensor, Sens. Actuators B (1997) 38–39.

677

[16] K.C. Lowe, Fluorinated blood substitutes and oxygen carriers, J. Fluorine Chem. 118 (2002) 19. [17] S. Icli, H. Icil, S. Alp, H. Koc, A. McKillop, NMR, absorption and fluorescence parameters of azlactones, Spectrosc. Lett. 27 (9) (1994) 1115. [18] W. Stumm, J.J. Morgan, Aquatic Chemistry, An Introduction Emphasizing Chemical Equilibria in Natural Waters, second ed., Wiley, New York, 1981, p. 780. [19] A. Mills, Q. Chang, N. McMurray, Equilibrium study on colorimetric plastic film sensors for carbon dioxide, Anal. Chem. 64 (1992) 1383.

Biographies Kadriye Ertekin, has a B.Sc. in chemistry, an M.Sc. and Ph.D. in analytical chemistry from Ege University, Izmir, Turkey. She works as an assistant professor in Dokuz Eylul University. Her current research interests include photo-characterization of newly synthesized dyes, optical sensors for pH, CO2 , O2 and cations. Serap Alp, has a B.Sc. in chemistry, an M.Sc. and Ph.D. in organic chemistry from Ege University, Izmir, Turkey. She works as an associated professor in Dokuz Eylul University. Her current research interests include synthesis and photo-characterization of fluorescent dyes, ionis liquids and development of optical sensors for pH, CO2 , O2 and cations.