Development and validation of a ph sensitive colorimetric sensing label for monitoring fish freshness

With the increasing consumers' awareness of the importance of food quality, new packaging technologies emerged to cater to their needs. Intelligent packaging is a packaging technology that allows consumers to observe -in real-time- the quality of food products. The main component of intelligent packaging is a pH-sensitive, color-changing dye that responds to certain metabolites produced during the storage period, due to either microbial or chemical deterioration of the product. Thus, upon accumulation of a certain amount of the target metabolite, the color of the intelligent label changes, informing the consumers that the food product is not edible. This promising technology is not only expected to have a huge impact on the consumers' satisfaction, but it will also contribute to an overall reduction in food waste, which will lead to decreasing both economic loss and preserving resources. Real-time quality observation will eliminate the need to keep track of the best-before date, which means that no food product will be disposed of before actually reaching the spoilage threshold. Even though lots of research related to intelligent packaging is present, the implementation of this technology in real food systems and supply chains is still limited. However, in the coming years the application of intelligent packaging is expected to increase tremendously. In this thesis, the aim was to fabricate and validate an intelligent label intended to be used in fish packaging. The labels produced have a three-layer structure: a color-changing layer that consists of a pH-sensitive dye and a binder, upper layer with hydrophobic properties to prevent moisture from the outer atmosphere entering the label, and lower layer that allows moisture and other metabolites to reach the color-changing layer. For the color-changing layer, bromothymol blue (BTB) was used as the pH-sensitive dye, while both methyl cellulose (MC) and cellulose acetate (CA) were used as binders. For the outer layer, polyethylene terephthalate (PET) film and for the inner layer polyethylene (PE) were used. Two dye solutions were prepared: bromothymol blue and methyl cellulose (water base), and bromothymol blue and cellulose acetate (solvent based). To produce labels sensitive to different metabolites (i.e., CO2 & total volatile basic amines (TVB-Ns)), either HCl or NaOH were added to decrease or increase the initial pH of the labels according to the target metabolite (HCl for TVB-N sensitive labels and NaOH for CO2 sensitive labels). The four different solutions (MC-based with either HCl or NaOH, and CA-based with either HCl or NaOH) were cast in circular shapes over the PET film and dried at 40°C overnight for the MC-based labels, and at room temperature for the CA-based labels. The labels were then cut and laminated with PE polymer. The labels were divided into two groups: CO2 sensitive labels (BTB-MC-NaOH and BTB-CA-NaOH) and TVB-N sensitive labels (BTB-MC-HCl and BTB-CA-HCl). Accordingly, two different simulations were conducted: CO2 simulation with CO2 concentrations of 0%, 5%, 10%, 15%, 20%, 25%, and 30%, and trimethyl amine (TMA) as a representative of TVB-Ns, with concentrations of 0%, 5%, 10%, 15%, 20%, 25%, and 30%. The CO2 was flushed into polyethylene/polyamide (PE/PA) packages containing the CO2 sensitive intelligent labels using a modified atmospheric packaging (MAP) machine. The packages containing the TVB-N sensitive intelligent labels TMA solutions of different concentrations were placed in closed vials that were reopened after flushing the packages with N2 gas prevent the interference of other gases. The two groups were then stored at 4°C for 10 days and the color change was observed and measured using a portable colorimeter every 2 storage days to determine the total color change (ΔE) by measuring L*, a*, b* values. To validate the intelligent labels in a real food system, atlantic bonito (Sarda sarda) fish was used. The fish were placed in PE/PA packages containing all four intelligent labels (BTB-MC-NaOH, BTB-CA-NaOH, BTB-MC-HCl and BTB-CA-HCl). To further imitate the commercially available fish packaging, two different MAP compositions usually used in fish packaging were used: 100% N2 and 70% N2 30% CO2. The packages were then stored at 4°C for 10 days. In addition to monitoring the total color change of the labels, quality of fish was monitored by measuring headspace gas composition, total mesophilic aerobic bacteria, Pseudomonas spp., pH, TVB-N, TMA concentration using gas chromatography (GC), and sensory evaluation. In CO2 simulation, BTB-MC-NaOH labels changed from blue to green at 15% CO2, while BTB-CA-NaOH labels changed from green to yellow at 25% CO2. The color change in both labels was visible to the naked eye within 24 hours and stayed constant through the storage period. For the TMA simulation, BTB-MC-HCl and BTB-CA-HCl labels changed color completely at 15% TMA (BTB-MC-HCl from yellow to navy and BTB-CA-HCl from yellow to green). For the CO2 sensitive labels, the color change was also visible to the naked eye within 24 hours and stayed constant through the storage period. These results were promising since the color-changing range of the TVB-N sensitive labels is also considered the spoilage threshold of fish. In the food validation trial, the headspace gas composition (CO2 and O2 concentrations) did not change significantly during the storage period (p<0.05), however, trace amounts of O2 were found in both MAP groups' packages despite not having O2 in the MAP composition, probably due to the release of O2 from the pores of fish during the storage period. Both total mesophilic aerobic bacteria and Pseudomonas spp. in both MAP groups increased significantly (p<0.05). In 70% N2 30% CO2 group, total mesophilic aerobic bacteria increased from 5.22 log cfu/g to 7.07 log cfu/g and Pseudomonas spp. increased from 3.74 log cfu/g to 6.80 log cfu/g by the 10th storage day, while in 100% N2 group, total mesophilic aerobic bacteria increased from 5.22 log cfu/g to 7.75 log cfu/g and Pseudomonas spp. increased from 3.74 log cfu/g to 7.94 log cfu/g by the 10th storage day. Since the spoilage threshold for fish under MAP is 7 log cfu/g, 70% N2 30% CO2 group reached upper acceptable limit (6 log cfu/g) after 6th day (approximately on 7th day) and microbial spoilage threshold (7 log cfu/g) after the 8th day while 100% N2 group reached upper acceptable limit after the 4th day (approximately on 5th day) and microbiological spoilage threshold at the 6th day. The difference between the two groups is thought to be the effect of CO2 in the 70% N2 30% CO2 group, which may have contributed to the slightly longer shelf-life, since CO2 is thought to have an inhibitory effect on bacterial growth. Throughout the storage period (10 days at 4C) no significant difference in pH values occurred in fish packaged under the two different MAP compositions (p>0.05). As part of the TVB-Ns, the TMA concentration in the headspace also showed a similar gradual increase over the storage period (p<0.05), although the highest value did not exceed 6 ppm in both MAP groups. The TMA level in 100% N2 group was slightly higher than 70% N2 30% CO2 group, even though the final TVB-N levels are similar. This could be the result of CO2 presence in 70% N2 30% CO2 group, which has a more inhibitory effect on the specific spoilage organisms that contribute to the production of TMA. Considering the score 3 as the acceptability level for all sensory attributes, both MAP groups (30% CO2 70% N2 and 100% N2) are considered unacceptable by the 6th day of storage for all the sensory attributes (color, odor, texture, and overall acceptability). The CO2 sensitive labels in 70% N2 30% CO2 MAP group, changed color from blue to green in BTB-MC-NaOH labels and form green to yellow in BTB-CA-NaOH labels by the second day of storage since 30% CO2 was used in the MAP composition, which color change is the same as the color change observed in these labels in the CO2 simulation results. The total color change values (ΔE) for these labels stayed almost constant with no significant difference (p>0.05) in both labels during the storage period (10 days at 4 °C). In 100% N2 group, the CO2 sensitive labels (BTB-MC-NaOH and BTB-CA-NaOH) did not change color, since the CO2 level in the packages stayed in the range 0-5% during the storage period. The total color change values (ΔE) of these labels stayed almost constant for both labels with no significant difference throughout the storage period (p>0.05). The results in both groups can be considered a further validation of the CO2 simulation results. TVB-N sensitive labels (BTB-MC-HCl and BTB-CA-HCl) did not show any color change during the storage period of fish in both MAP groups. The total color change values (ΔE) of these labels showed no significant differences, whether within the same group during the storage period or within the two different groups within the storage period (p>0.05). In the TMA simulation, the color change in both TVB-N sensitive labels (BTB-MC-HCl and BTB-CA-HCl) changed color when the TMA level reached 10 ppm. Since the TMA values obtained for both MAP groups did not exceed 6 ppm, the result is in parallel with the TMA simulation results. In conclusion, the results of this research show that the fabricated intelligent labels -regardless of the binder used- could successfully monitor the different concentrations of spoilage metabolites (i.e., CO2 and TMA) under simulated conditions. However, since fish is usually packaged under MAP (usually without oxygen in the MAP gas compositions), it has a different spoilage mechanism than fish packaged under atmospheric conditions. Hence, these labels are not suitable to appropriately monitor the freshness of fish under MAP conditions, which is a factor often dismissed when conducting a validation experiment for intelligent labels.