Determination of CO 2 Content in the Headspace of Spoiled Yogurt Packages

e CO2 formation during food storage can often be correlated with the increase in yeast population. Yogurt and other dairy products are susceptible to yeast contamination. Accumulation of CO2 in the headspace of yogurt packages can lead to the eventual blowing o of the package. erefore, determination of CO2 in the yogurt packages can indicate eventual unsafety of the product. e aim of this paper was to determine CO2 concentration in the headspace of dierent yogurt containers contaminated with yeast at the levels of 1 and 5 CFU/ml. Yeast Candida kefyr, previously isolated from spoiled yogurt, was used for contamination. Contaminated and control samples of yogurt were incubated at 30°C. A device based on tunable diode laser absorption spectroscopy was used for the measurement of CO2 concentration. e CO2 content in all analysed samples changed in a similar manner with slow increase to the value of 6% during the rst 30 h and, after that, rapid accumulation to 17–20%. e initial level of yeast contamination did not have signi cant inuence to the CO2 content trend.e increase in the number of yeast was observed after 10 h of incubation, and the nal value of 6-7 log·CFU/cm was reached after 40 h of incubation. e signi cant increase in the yeast number can be correlated with the CO2 content in a way that CO2 concentration of 6% can be considered as critical for microbial spoilage. Since the TDLAS technique is simple and nondestructive, it can be a promising possibility for detection of the microbial spoilage in food.


Introduction
Food safety has become one of the main global concerns due to the rapid economic growth and industry development.
is required the use of di erent monitoring techniques in the eld of food quality and safety [1].e parameters that usually re ect the food quality, shelf life, and safety are temperature, pH, freshness, presence of indicator microorganisms, and leakage or content of gases (CO 2 and O 2 ).Conventional techniques of food safety determination are mostly time consuming, laborious, and involve the destruction of the sample.Also, pretreatment of the sample is required which leads to di culties in real-time and online monitoring [2].Development of high-performance, inexpensive, nondestructive method for food safety analysis has become one of the most important problems in food research.
e use of di erent imaging techniques in food quality monitoring gained a lot of interest in the past few years.Many in-line and o -line applications of Raman spectroscopy technique, hyper spectral imaging, magnetic resonance imaging, soft X-ray imaging, ultrasound imaging, thermal imaging, uorescence imaging, and odour imaging in food analysis have been reported [2,3].Here, we focus on tunable diode laser absorption spectroscopy (TDLAS) that is inherently a noninvasive method as it measures gas content in closed containers by passing through the headspace inside the package.Due to simplicity, high selectivity, and sensitivity, TDLAS can be applied in di erent industrial processes, gas pipelines, and power generation plants, as well as in beverage, agro-food, and pharmaceutical sealing processes [4,5].
TDLAS is based on the fact that each molecule has a unique light absorption characteristic that absorb the laser beam at a very specific wavelength.e concentration, temperature, pressure, and in some cases even velocity and mass flux of gases in a gaseous mixture can be determined by measuring the absorption profile through a diode laser beam that is wavelength tuned to scan along the whole line profile.At wavelengths slightly different from these narrow absorption lines, there is no absorption from the particular gas.TDLAS enables noninvasive measurement through a wide range of closed containers, from bags to trays and cups, and a wide range of materials, from transparent to almost dark films [6,7].In food industry, TDLAS technique has been already applied for checking closure tightness and leak detection, for in-line measurement and monitoring of fermentation process in beer, sparkling wine, and soft drink industry, and for analyzing packages in modified atmosphere [8][9][10][11][12][13].
Yogurt is a firm or liquid fermented milk product to be consumed up to a few weeks after the production before a reduction in the sensory characteristics occurs [14].Despite the low pH value (3.8-4.5),yogurt can be a selective environment for the growth of contaminants, especially fungi which can cause significant economic losses due to offflavour and changes in visual appearance [15].e spoilage of yogurt by yeasts and consequential CO 2 accumulation can induce swelling and, in the bottom line, blowing off of the container.erefore, indirect measurement of CO 2 content in a headspace of the yogurt packages can refer to yeast contamination.
is paper reports the determination of the CO 2 content in the headspace of yogurt cups and bottles during storage.In order to examine the process, the samples were contaminated with the yeast Candida kefyr.CO 2 content was continuously measured by a TDLAS device specially constructed for these measurements.

CO 2 Measurement.
e content of gaseous CO 2 in the headspace of closed yogurt containers is measured by a device specially realized for these measurements and based on TDLAS.
e device is based on a VCSEL source operating on the 2 µm CO 2 absorption band and is meant to measure the content of CO 2 in the headspace of axially symmetrical containers such as bottles and cups.e sample to be measured is placed on a rotating table with adjustable height, in order to transmit light from the source through a diameter of the sample volume within its headspace and to collect emerging light with an extended InGaAs detector placed on the opposite side (Figure 1).Sample rotation during the measurement is required to smooth out irregularities in the sample itself as well as to reduce optical fringing due to light interference.
Wavelength modulation spectroscopy technique is used to obtain maximum sensitivity [16]. is is done by overlapping a high-frequency sinewave carrier over the linear wavelength scanning signal and doing a demodulation at the receiver side that is synchronous with the carrier frequency and its harmonics.is kind of processing is implemented on a fully software-defined basis by a single data acquisition card, which synchronously controls the laser source while acquiring photodetector signal.An embedded computer provides control on the signals and measurements, while performing lock-in demodulation and signal processing of the recorded absorption profiles to evaluate the gas concentration.
e optical front end of the instrument and sample management frame were designed to be easily integrated in an incubator to allow continuous measurement on thermally controlled samples.Additionally, a calibration protocol was identified and validated so that the instrument could be easily calibrated on the field and applied to a wide range of available samples.To calibrate the device, dummy packages were placed in the instrument and flushed with known, calibrated gas mixtures.is procedure also provides a direct assessment of the output linearity.Measurements during the initial validation process reported linearity up to 25% CO 2 volume, a minimum detectable threshold of 60 ppm (in a controlled environment), and a resolution of 160 ppm over a 5% CO 2 volume reference level [17].As in scattering samples, such as the majority of food containers, the actual optical path length may differ significantly from the geometrical source-todetector distance, and it is important to note that calibration is container dependent, and thus, calibration has to be taken on every considered type of package.
e accuracy of the measurement was validated by measuring CO 2 content in the containers flushed with a gas mixture containing 6% CO 2 and 94% N 2 , custom made from Messer gas factory (Pancevo, Serbia).

Experimental Conditions.
A total number of 50 samples (9 samples of cups of 180 g, 21 samples of bottles of 500 g, and 20 samples of bottles of 1000 g) were used for the experiment.e experiments were performed in triplicate.e samples were supplied directly from the production line and aseptically inoculated with the isolated yeast strain at a level of 1 CFU/cm 3 for bottles of 500 g and 1000 g, and 5 CFU/cm 3 for cups and bottles of 500 and 1000 g.Samples were inoculated with a suspension of yeast for which turbidity was adjusted to be in accordance with 0.5 McFarland standard.A few uncontaminated cups and bottles were used as control samples.e samples were incubated and CO 2 measurements were performed during the incubation process, while the temperature was kept at 30 °C.
Every sample which showed a signi cant change in the CO 2 content in the headspace was analysed for a number of yeasts, lactic acid bacteria (by serial dilution method), and pH value (by the pH meter "Hanna HI 92240").

Statistical Analysis.
e error bars were calculated in Excel (Microsoft, USA) as standard deviation on experimental data.Prediction area ellipses were determined based on correlation between the number of yeast cells and CO 2 content using Statistica 7.0 (StatSoft Inc., USA) for Windows.e meaning of the ellipses is that the probability of a new pair of measurement values falling within the marked area is estimated to be 90%.

Results and Discussion
Along with the mould contamination, contamination with yeasts is the largest microbiological problem in fermented milk products.Yeasts are generally well recognized as a typical type of spoilage microorganisms in yogurt and yogurt-based drinks [18].
e yeast used in this study, isolated from spoiled yogurt and identi ed as Candida kefyr, is well recognized as a spoilage microorganism in dairy products [19].
e changes in the CO 2 content in the headspace of the di erent packages (PVC cups: 180 g; PET bottles: 500 g and 1000 g) and the number of yeasts in contaminated yogurt ("Vedro, 2.8% milk fat") are presented in Figure 2. e yogurt samples were contaminated with the yeast Candida kefyr at two levels of initial cell concentrations: 1 CFU/cm 3 (bottles) and 5 CFU/cm 3 (cups and bottles).In order to discuss and compare changes in the CO 2 content in the headspaces of two types of yogurt, the results of the previous experiment [20] with the bottles (500 g) contaminated at the initial level of 1 CFU/cm 3 are also included.
Regardless of the type of container and initial yeast concentration, the CO 2 contents in the headspaces of the samples changed in a similar manner in the rst 30 hours of monitoring, that is, CO 2 content slightly and linearly increased from the initial value of 1-3% to about 6%.From that moment, the CO 2 content increased faster reaching the level of 17-20% CO 2 at the end of the monitoring.Slightly slower increase in CO 2 content was noticed only in the 500 g bottles contaminated at the initial yeast level of 1 CFU/cm 3 , while in all other samples, regardless of the type of container or initial yeast number, the changes in the content of CO 2 are similar (Figure 2(a)).e number of yeast changed similarly regardless of the initial number and the type of the container (Figure 2(b)).Linear increase in the yeast number was noticed during the rst 40 h of incubation from about 2 log•CFU/cm 3 (after 10 h of incubation) to about 6-7 log•CFU/cm 3 after 40 h and then remained nearly on the same level till the end of monitoring.Higher yeast number was detected in cups in the period 18-28 h, but after that the number is in accordance with the rest of samples (Figure 2(b)).e maximum yeast number determined in all samples was in accordance with the previous investigation which reported increase in the number of psychotropic yeasts to 6 log•CFU/g after 2, 10, and 20 days for the samples stored at 25, 15, and 5 °C, respectively [21].
During the period of incubation of control samples under the same conditions, the CO 2 content slightly and linearly increased to about 5-6% (Table 1) and the containers were not bloated.e yeast could not be detected by the applied method in the control yogurt samples at the beginning and after 10 h of incubation, but after certain time of incubation, the yeast concentration increased to the detectable level due to increased temperature of the incubation.At the end of monitoring period, the detected yeast concentration was in the range of 1.9-2.9log•CFU/cm 3 (Table 1).
e LAB microbiota seemed not to be in uenced by the presence of yeast since the number of LAB was not signi cantly changed during the incubation of contaminated and not-contaminated yogurt (Table 1).e LAB number during the conduction of the experiment was in the range of 5.5-7.2log•CFU/cm 3 in all samples which is in accordance with number of lactic acid bacteria normally present during the storage of the yogurt [22].Also, the pH values in the samples were generally unchanged during the period of the experiment, ranking 4.2-4.3which is in compliance with the literature [22].e concentration of CO 2 in the headspace of a package at the beginning of measurement was 1.9-3.6%(Table 1), which is much higher than the atmospheric concentration of 0.03% CO 2 probably because of the establishing of equilibrium with the product itself.Dissolved CO 2 depends on the partial pressure of the CO 2 and the available amount of CO 2 .e CO 2 amount in the headspace is in relation to the headspace volume and the CO 2 concentration [23].Compared to the quantity of yogurt (180 g, 500 g, and 1000 g), the volume of headspace in the packaging is small (about 25 cm 3 of air).
Based on the averaged values of the data presented in Figure 2, the relations between CO 2 content in the headspace of the packages (cups and bottles of di erent volume) and the number of yeasts (regardless of the initial level of contamination) can be observed.e relations could be divided into two areas (Figure 3).e presence of yeasts at the level of up to 5 log•CFU/cm 3 induced the accumulation of CO 2 in the range of 3-7% in the headspace of the yogurt containers (open ellipse on Figure 3) regardless of the duration of the incubation.On the contrary, at yeast counts higher than about 5 log•CFU/cm 3 , there was a signi cantly greater accumulation of CO 2 content above yogurt (grey ellipse in Figure 3).Since the contamination was done with the yeast isolated from spoiled yogurt, we presumed that the same strain caused the accumulation of CO 2 in contaminated and not-contaminated samples.e similar level of yeast (10 5 -10 6 CFU/g) has been reported to be the cause of yeasty and fermented off-flavours and gassy appearance in yogurt [18].According to the results, the CO 2 content above 6% can be considered critical for microbial spoilage since at this point the level of detected number of yeast is significant.is confirmed our previous state [20,24] that this content of CO 2 in the headspace of fermented milk products is an indicator of possible contamination.

Conclusions
Contamination with yeasts, which induces the increase in the CO 2 content, often creates the largest microbiological problem in dairy products.erefore, CO 2 measurement could be used as an indication of dairy products spoilage.Scarce literature data for the determination of the CO 2 content in dairy products increase the novelty of the conducted investigation with the use of TDLAS technology for measuring CO 2 content.Also, data indicated the correlation between the content of developed and accumulated CO 2 in the headspace of yogurt packages and the presence of contamination by yeasts.Since measuring of CO 2 content is much easier and more precise than yeast number determination, this efficient and contactless technique for monitoring the CO 2 content represents a simple nondestructive method for the detection of yogurt yeast contamination.e determined CO 2 content out of predefined ranges could refer to inadequate control or storage conditions during yogurt production and distribution.So, the TDLAS technique shows to be a promising novel approach in designing instruments for the CO 2 content measuring in the headspace of yogurt packages.

Figure 2 :
Figure 2: Changes in the CO 2 content (a) and the number of yeasts (b) during the incubation (30 °C) of yogurt ("Vedro," 2.8% milk fat) packed in PVC cups (180 g) and the PVP bottles (500 g, 1000 g) contaminated with the yeast Candida kefyr with initial concentration of 1 CFU/cm 3 (open symbols) and 5 CFU/cm 3 (black symbols).e error bars represent standard deviation on experimental data.

Table 1 :
Values (±SD) of pH, CO 2 content, and number of lactic acid bacteria (LAB) and yeast in yogurt not-contaminated (Ø) and contaminated with the yeast Candida kefyr (initial cell concentrations 1 and 5 CFU/cm 3 ), and packed in cups and bottles at the beginning (0) and the end of monitoring (end).Determined on MRS agar, * * determined on SMA agar with chloramphenicol. *