A CO₂ sensor based on Pt-porphyrin dye and FRET scheme for food packaging applications

Abstract

An optochemical CO₂ sensor is described which uses a phosphorescent reporter dye PtTFPP and a colourimetric pH indicator α-naphtholphthalein incorporated in plastic matrix together with a phase transfer agent tetraoctyl- or cetyltrimethylammonium hydroxide. Good spectral overlap of PtTFPP emission and α-naphtholphthalein absorption at around 650 nm provides efficient energy transfer from PtTFPP via FRET mechanism, thus reducing its phosphorescence lifetime. Experiments were carried out to optimize the composition and working characteristics of such a sensor for the measurement of headspace CO₂ in foods packaged under modified atmosphere. The resulting thin film sensor coatings showed robust optical responses to CO₂, response (1 min, 99.9%) and recovery times (4 min, 99.9%) from 0 to 100% CO₂. As expected, the CO₂ sensor showed significant cross-sensitivity to O₂ quenching, and this was accounted for by a tandem O₂ sensor having similar phosphorescent characteristics but no CO₂ sensitivity. Upon storage in an open container, the sensors were stable for at least 14 days at 4 °C and 50 days at −20 °C, however at room temperature they were seen to deteriorate within 7 days loosing colour and sensitivity to CO₂. At the same time, in food and modified atmosphere environment the sensor retained its sensitivity to CO₂ for 21 days at 4 °C which is sufficient for many packaged products. Migration of sensor components examined with a standard set of standard food stimulants was only detectable in 95% ethanol and in olive oil for PtTFPP (0.06–0.13 μg/ml) and in 50% and 95% ethanol for α-naphtholphthalein (9.86–21.65 μg/ml).

Introduction

Detection of CO₂ is important for many areas, particularly in blood gas analyzers, biological and medical research [1], [2], food packaging and shelf life studies [3], [4], marine and environmental monitoring [5], [6]. The main techniques for CO₂ measurement include Severinghaus type electrode, infrared (IR) spectroscopy, gas chromatography (GC), mass spectrometry (MS) and optochemical sensors. Severinghaus CO₂ sensor consists of a glass electrode immersed in bicarbonate buffer and covered with a hydrophobic gas permeable membrane, which detects pH changes [7]. Its limitations are the use of liquid reagents, indirect detection or ionic form of CO₂, interference by basic or acidic gases, slow response times and high maintenance costs. IR absorption spectroscopy allows precise and direct CO₂ detection via absorbance at 2.6 and 4.3 μm, however it suffers from strong interference by water vapour and enclosure materials (plastics) and requires rather sophisticated equipment and fixed measurement geometry [8], [9]. GC and MS techniques are also destructive, slow (∼20 min), have limited throughput and require sampling and calibration [10].

Optochemical CO₂ sensors have high application potential. Initially such systems relied on the principles of Severinghaus electrode using pH-optode instead of the electrode. They demonstrated simplicity, portability, low cost, fast response and flexibility, but possessed weaknesses similar to the electrodes [11]. This was overcome in the approach proposed by A. Mills [1], [12], [13], in which the pH/CO₂ sensitive dye was incorporated in a hydrophobic polymeric membrane together with a hydrophobic phase transfer agent (PTA) such as tetraoctylammonium hydroxide (TOA-OH). The PTA acts as an ion pair for the indicator preventing its leaching and also retains some water which is necessary for system operation [14]. Specific requirements of different applications are not easy to satisfy due to the lack of suitable fluorescent dyes and sensor materials. Thus, for packaging applications rather low sensitivity is required to cover the range 0–100% or 0–100 kPa of CO₂ [15]. This can be realized using indicator dyes with relatively high pK_a values. Whereas for environmental applications and process control high sensitivity to CO₂ is usually required. To facilitate diffusion of CO₂ and reduce response time, a plasticizer can be added to the polymeric membrane [16].

With respect to signal readout from a CO₂ sensor, basic qualitative and semi-quantitative systems can use simple visual detection via colour change. However for accurate quantitative detection instrumental readout is usually required and in this case photoluminescence based sensors offer a significant potential. Classical approaches rely on fluorescence intensity measurements [11], but these are affected by drifts in optoelectronic system, dye photobleaching, sample properties and measurement geometry. This can be circumvented by the schemes with internal referencing [17], [18], [19]. Thus, in ratiometric luminescence intensity scheme signals at two different wavelengths are measured, one is analyte-sensitive while the other is analyte-insensitive, and related to each other. This improves system performance and stability but still cannot fully compensate for light scattering, reflection and differential sample absorbance influencing the measurement.

Another system with internal referencing uses the DLR scheme (Dual lifetime referencing), whereby the short-lived analyte-sensitive fluorophore is co-immobilized with an analyte-insensitive reference luminophore having long lifetime and similar spectral characteristics, and then measured by phase fluorometry. This DLR scheme has been applied to the quantitative determination CO₂, using fluorescent pH indicator 1-hydroxypyrene-3,6,8-trisulfonate (HPTS) together with Ru(dpp)₃²⁺ dye co-immobilized in a hydrophobic organically modified silica matrix [3]. By changing the PTA TOA-OH to CTA-OH it was possible to extend the range of CO₂ with a phase shift change of 13.5 degrees between 0 and 100% CO₂ [3].

CO₂ sensors which use inner filter quenching effect were also described. One such system consisted of a phosphorescent platinum octaetylporphyrin (PtOEP) and a pH-sensitive NP dye in poly(vinylidene chloride-co-vinyl chloride-ethyl cellulose) (PVCD-EC) thin films or microparticles with low permeability to oxygen [20]. These sensors showed fast response times (<9 s), reproducibility and shelf life of >4 months. A low-cost handheld optoelectronic device with a paired emitter–detector diode arrangement acts as a colorimetric detector for these sensors [21], [22].

Förster Resonance Energy Transfer (FRET) scheme is a further system described as efficient for CO₂ detection. One such sensor for measurement dissolved CO₂ employed tetraphenylporphyrin (TPP) and α-naphtholphthalein (NP) in a poly(isobutyl methacrylate) (P(IBM)) matrix measuring changes in fluorescence intensity [23]. Its response and recovery times were <6 s, and no hysteresis was observed during the measurement [24], [25]. But instead of intensity alteration the use of long-decay phosphorescent indicator dyes and analyte-dependent changes in luminescence lifetime (LT) of the sensor [4], [11] can provide accurate CO₂ quantification with relatively simple instrumentation similar to the one developed for O₂ measurement. Such systems are demanding for many industrial applications, particularly food packaging [26]. They require sensor materials with microsecond lifetimes which are influenced by CO₂ content in the sample, such as a long-decay Ru(II) complex as fluorescent donor (CO₂ insensitive) and a pH/CO₂-sensitive acceptor co-immobilized in a host matrix together with a PTA [4], [11]. To produce optimal FRET, the two chromophores should be in close proximity and have good overlap of acceptor excitation and donor emission spectra. In this case, LT based detection with a single excitation source and photodetector can be realized [14]. Thus far, such systems, accomplished with Ru(II) complexes, have relatively short lifetimes [11], [27].

In this work, we present a polymeric solid-state CO₂ sensor which uses phosphorescent Pt-porphyrin (PtTFPP) reporter, pH-sensitive NP acceptor, solution FRET scheme and LT measurements. The sensor material was optimized for food packaging applications and underwent detailed characterization with respect to its CO₂ sensitivity, response and recovery times, stability, cross-sensitivity to oxygen and temperature. Sensor behaviour upon storage and operational stability in packaged foods was evaluated, and migration of sensor compounds into food was examined using standard panel of food simulants.