Effect of Different Oxygen Atmospheres on Color Stability of Modified Atmosphere Packaged Beef Using Non-Invasive Measurement

Abstract

The influence of a 1% oxygen atmosphere on the color stability of modified atmosphere packaged beef was investigated. Beef silverside slices were packed under 1%, 20%, and 70% oxygen atmospheres and stored at 2 °C for 14 days. Color and reflection data were measured non-invasively. The L*a*b* values were analyzed, the color difference ΔE₂₀₀₀, and the levels of myoglobin (Mb), deoxy-(DMb), oxy-(OMb), and metmyoglobin (MMb) were calculated. The 1% oxygen atmosphere resulted in a rapid MMb formation from 0.63 (day 0) to 1.27 (day 8) (p < 0.05). The other samples showed slight increases from 0.65 to 0.80 MMb (20%) and 0.65 to 0.79 MMb (70%). On day 10, the 20% oxygen sample showed an increased MMb formation (1.33 MMb). The 70% atmosphere resulted in a final value of 0.91 MMb after 14 days. These results show that an oxygen content of 1% accelerates the formation of MMb at an early stage. A higher oxygen content in the packaging delays MMb development through OMb formation, which masks MMb creation, to a certain extent. Measuring the packaged meat pieces over a 14-day storage period provides detailed insights into the development of Mb formation and critical points during storage.

1. Introduction

Beef continues to be a popular protein source worldwide, with consumption reaching 70.4 thousand metric tons in 2020 through 2022 and expected to rise to 77.6 thousand metric tons by 2032 [1]. In the meat industry, especially within the beef sector, alterations in color are one of the major causes of food waste [2,3,4,5,6,7,8,9]. The consumers expect the typical cherry-red color, and any deviations will result in a decreased willingness to buy the product, as color is the most important quality indicator in packaged meat products [4,6,10,11,12,13,14]. Reducing discoloration can help prevent food loss, as well as the financial and environmental damage that results from it [4].

In different German slaughterhouses, an unexplainable phenomenon can be observed that leads to an irreversible brown or gray discoloration of vacuum packaged beef during the first weeks of the wet-aging process. In prior experiments, a suitable non-invasive measurement method that leaves the packaging and samples intact has been adapted and altered for the conditions of this study [4]. One of the possible causes of the discoloration could be the formation of a low oxygen atmosphere in the containers, for example, due to incomplete evacuation of the vacuum packages. The relationship between the discoloration of meat and low oxygen atmospheres is described in the literature. Critical oxygen values between 0.1% and 5% oxygen are suggested to favor metmyoglobin (MMb) formation, depending on the source [2,15,16,17,18,19,20,21]. Low oxygen concentrations in the atmosphere result in lower partial oxygen pressure. If the partial pressure does not exceed the critical oxygen partial pressure, the oxygen consumption and the MMb-reducing activity of the meat lead to the deoxygenation of oxymyoglobin (OMb) to form deoxymyoglobin (DMb) and reverse MMb formation [18,21,22,23]. DMb is less resistant to oxidation than is OMb [2,18,19]. Under high vacuum, the presence of DMb is not problematic since oxidative processes are minimized due to the small amount of residual oxygen, which can quickly be reduced to below critical partial oxygen pressure levels by oxygen consumption of the meat [2,4,21]. However, in a low oxygen environment which exceeds the critical oxygen partial pressure, a state in which the MMB formation is promoted and can no longer be reversed by the MMb-reducing activity is reached [15,18,21,24]. This effect results from the increased oxidation rate of Mb, but also from the decreasing MMb-reducing activity, which is higher under anaerobic conditions [21,23].

This study was conducted to investigate whether this mechanism is responsible for the previously described discolorations observed under typical wet-aging conditions. The hypothesis of the study was that a low oxygen atmosphere of 1% leads to an increased MMb formation (discoloration) in the meat samples compared to the results for 20% and 70% oxygen atmospheres, which can be measured non-invasively over a storage time of 14 days. A novelty in this study also includes the non-invasive measurement of myoglobin under these atmospheric conditions over a storage time of 14 days, which is comparable to the typical wet-aging time employed in German slaughterhouses. In particular, the effect of modified atmosphere packaging on beef color should be examined because this type of packaging often serves as sales packaging [25,26]. The aim of the continuous measurement of the myoglobin redox states is to obtain detailed insights into the myoglobin conversion mechanics and critical time points at which certain changes occur.

2. Materials and Methods

2.1. Sample Preparation

Silversides from three different Simmental breed heifers were purchased 24 h after slaughter, non-aged, and unpackaged, from a local slaughterhouse (Co. MEGA, Stuttgart, Germany). The transport to the lab was completed in 20 min by car. After arrival in the lab, the outer layers, fat, and tendons were immediately manually removed using a butcher’s knife. Afterwards the silversides were sliced into 5 × 5 × 1 cm square slices using a semi-automatic slicer (VS 8A, Co. Bizerba, Balingen, Germany). For the slices, meat from the middle of the silversides was used to minimize oxygen exposure of the samples. The cross section of the slices was orthogonal to the muscle fibers in order to minimize scattering effects. Then, the slices were immediately packed using a vacuum chamber machine (C500, Co. Multivac, Wolfertschwenden, Germany) in 220 × 330 mm polyamide/polyethylene side seal bags with a thickness of 90 µm and an oxygen permeability ≤ 56 cm³/m² bar 24 h (Co. Allfo, Waltenhofen, Germany), and different modified atmospheres were applied to the packages. The different standardized modified atmospheres (Co. Westfalen AG, Münster, Germany) contained 1% O₂/99% N₂, 20% O₂/80% N₂, and 70% O₂/30% CO₂. The reason why the high-oxygen atmosphere contained CO₂ in contrast to the other atmospheres was to inhibit microbial growth which at high oxygen levels spoils the meat before the conclusion of the two weeks of storage time. The 70% O₂/30% CO₂ atmosphere is widely applied in the meat industry, making it a suitable reference gas mixture [25,26]. Nine slices were individually packaged per atmosphere (three slices per animal), which resulted in 27 packages. After packaging, the samples were stored at 2 °C in a dark cooling chamber, which resembles typical wet-aging conditions.

2.2. Sample Measurement

To investigate the color and myoglobin redox form development during storage, the L*a*b* and reflection data of the 27 samples were measured non-invasively in the packaging with a spectrophotometer (UltraScan VIS 1091 Diffuse/8° observer angle, Co. HunterLab, Reston, VA, USA), in triplicate. This spectrophotometer comprises a measurement area of 25.4 mm, allowing it to scan a larger area than that of handheld devices. To maximize atmospheric contact with the samples during storage, they were always placed on the same side, ensuring constant atmospheric contact on the opposite side. For reflection measurements, only the side with atmospheric contact was used. To minimize light exposure during the measurements, the samples were stored in a box and only removed for measurement. This procedure was repeated over 14 days. On day 0, 7, and 14, photographs of the packaged samples were captured using an EOS 250D camera (Co. Canon, Tokyo, Japan) and a camera holder, as described by Ruedt et al. [27]. The samples were stored in the dark at 2 °C during the entire experiment, except while being measured.

2.3. Data Evaluation

From the measured L*a*b* data, the color difference ΔE₂₀₀₀ was calculated according to the requirements of the International Organization for Standardization and International Commission on Illumination [28]. A detailed description of the equation is described in our previous work [4].

For the calculation of the relative myoglobin redox state levels, the reflectance data was used. The equations (Equations (1)–(4)) can be found in several sources and are suitable for longer storage experiments, since no reference values are required [6,29,30]. Depending on the source, the myoglobin ratios are accompanied by transformation factors to allow application-dependent display of the data. Since in this experiment, only the changes in the myoglobin forms over time are of importance, the added transformation factors were excluded, and only the K/S ratios of the wavelengths were used. Therefore, the graphs are shown without percentage indication of the myoglobin forms.

$$K/S={\displaystyle \frac{{(1-R)}^2}{ 2 R}}$$

(1)

$$\mathit{DMb}={\displaystyle \frac{K/S(474 \mathrm{nm})}{K/S (525 \mathrm{nm})}}$$

(2)

$$\mathit{OMb}={\displaystyle \frac{K/S (610 \mathrm{nm})}{K/S (525 \mathrm{nm})}}$$

(3)

$$\mathit{MMb}={\displaystyle \frac{K/S (572 \mathrm{nm})}{K/S (525 \mathrm{nm})}}$$

(4)

where: K = absorption coefficient [-]; S = scattering coefficient [-]; R = reflection at respective wavelength [%]; DMb = relative deoxymyoglobin level [-]; OMb = relative oxymyoglobin level [-]; MMb = relative metmyoglobin level [-].

2.4. Statistical Analysis

For the statistical analysis, the mean values, standard deviations, and standard errors were calculated from the raw data using Excel (Co. Microsoft, Redmond, WA, USA). The As an outlier test, the Nalimov test was used to exclude statistically irrelevant values. A total of 7 degrees of freedom and an α-error of 0.05 were used. To determine significant differences (p < 0.05) in the measured values, SIGMAplot v16 (Co. Systat Software GmbH, Erkrath, Germany) was used. The assumptions for ANOVA were checked using a normality test (Shapiro–Wilk) and an equal variance test (Brown–Forsythe) (p < 0.05). In the event that both tests were not passed, a non-parametric one-way analysis of variance on ranks (ANOVA on ranks/Kruskal–Wallis) and an additional post hoc test (Student–Newman–Keuls or Dunn’s test) were executed. The measured data ± standard deviation and the indications for significance can be seen in

Table A1. Lightness (L*) values of the beef samples over time.

L*
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	34.15 Aa ± 2.046	35.40 Aab ± 3.505	37.82 Ab ± 5.202
1	33.14 Aa ± 4.295	36.16 Bb ± 4.014	39.24 Ab ± 3.825
2	34.87 Aa ± 2.548	37.75 Bb ± 3.846	39.36 Ab ± 4.355
6	33.75 Aa ± 4.931	35.79 Ba ± 2.929	37.96 Ab ± 2.951
8	34.62 Aa ± 2.938	35.86 Ba ± 3.723	39.59 Ab ± 3.973
10	34.95 Aa ± 3.507	37.09 Bb ± 2.797	39.94 Ac ± 3.603
13	32.38 Aa ± 4.695	33.82 Ca ± 4.379	37.51 Ab ± 5.211
14	35.16 Aa ± 3.194	36.29 Bab ± 3.624	39.91 Ac ± 4.962

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant differences (p < 0.05) for one atmosphere over time. Different lowercase letters indicate significant differences (p < 0.05) for different atmospheres at the same time.

,

Table A2. Green–red axis (a*) values of the beef samples over time.

a*
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	5.142 ADa ± 1.016	9.515 AEb ± 2.233	13.03 Ac ± 2.492
1	8.566 BCa ± 1.707	16.04 Bb ± 1.727	17.07 Bc ± 1.791
2	7.322 BCa ± 0.822	14.59 BCb ± 2.065	16.34 BCb ± 1.559
6	5.963 ACa ± 0.867	13.51 BCDb ± 0.876	15.36 Cc ± 1.341
8	6.056 ACa ± 0.929	12.66 ACDb ± 1.656	14.01 Ab ± 1.926
10	6.628 Ca ± 0.731	11.07 ADb ± 1.445	13.05 Ac ± 1.643
13	4.663 Da ± 0.985	6.212 Ea ± 2.336	9.965 Db ± 2.450
14	4.371 Da ± 0.828	5.284 Ea ± 1.766	8.394 Eb ± 1.792

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant difference (p < 0.05) for one sample over time. Different lowercase letters indicate significant differences (p < 0.05) for different samples at the same time.

,

Table A3. Blue–yellow axis (b*) values of the beef samples over time.

b*
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	5.587 Aa ± 0.727	8.858 Ab ± 2.274	11.54 Ac ± 2.881
1	8.923 Ba ± 1.865	13.86 Bb ± 2.318	13.29 Bb ± 1.795
2	8.321 Ba ± 1.398	13.42 Cb ± 2.878	14.29 Cb ± 2.437
6	8.887 Ba ± 1.996	12.29 Db ± 1.321	14.16 Cc ± 1.348
8	9.233 Ba ± 2.041	13.00 Eb ± 2.426	13.37 Bb ± 3.099
10	8.667 Ba ± 0.846	11.01 Fb ± 1.401	12.71 Bc ± 1.551
13	9.972 Ca ± 2.582	12.55 Gb ± 3.517	12.29 Bb ± 3.302
14	8.895 Ba ± 2.181	11.07 Fb ± 1.970	11.44 Ab ± 2.211

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant difference (p < 0.05) for one sample over time. Different lowercase letters indicate significant differences (p < 0.05) for different samples at the same time.

,

Table A4. Color difference (ΔE₂₀₀₀) values of the beef samples over time.

ΔE2000
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	0.000 Aa ± 0.000	0.000 Aa ± 0.000	0.000 Aa ± 0.000
1	5.543 Ba ± 1.952	6.788 Bb ± 1.980	4.267 Bc ± 2.088
2	3.971 BCa ± 1.898	6.123 BCb ± 2.587	4.676 Ba ± 2.353
6	4.769 BCa ± 2.135	4.653 BCa ± 1.880	3.859 Ba ± 1.632
8	4.300 BCa ± 1.746	4.957 BCa ± 2.451	4.528 Ba ± 2.532
10	3.602 Ca ± 1.186	4.390 Ca ± 2.226	3.809 Ba ± 2.359
13	4.953 BCa ± 2.172	7.146 BCb ± 3.288	5.714 Bab ± 2.156
14	4.613 BCa ± 1.639	6.625 BCb ± 1.845	5.691 Bab ± 2.670

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant differences (p < 0.05) for one sample over time. Different lowercase letters indicate significant differences (p < 0.05) for different samples at the same time.

,

Table A5. Relative deoxymyoglobin content of the beef samples over time.

DMb
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	1.590 Aa ± 0.071	1.231 Ab ± 0.099	1.123 ABDc ± 0.034
1	1.379 Ba ± 0.129	1.137 Bb ± 0.036	1.118 ABb ± 0.034
2	1.251 Ca ± 0.058	1.087 Cb ± 0.026	1.095 Ab ± 0.034
6	1.263 Ca ± 0.119	1.171 Db ± 0.046	1.147 BDb ± 0.031
8	1.219 Da ± 0.050	1.173 Db ± 0.038	1.144 BCDb ± 0.034
10	1.189 Ea ± 0.068	1.150 Ea ± 0.032	1.115 ACb ± 0.034
13	1.221 Da ± 0.091	1.192 Dab ± 0.087	1.167 Db ± 0.050
14	1.155 Fa ± 0.047	1.174 Eab ± 0.125	1.117 ABb ± 0.034

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant differences (p < 0.05) for one sample over time. Different lowercase letters indicate significant differences (p < 0.05) for different samples at the same time.

,

Table A6. Relative oxymyoglobin content of the beef samples over time.

OMb
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	1.817 Aa ± 0.133	2.273 Ab ± 0.280	2.725 Ac ± 0.473
1	2.445 Ba ± 0.511	3.480 Bb ± 0.692	3.328 Bb ± 0.425
2	2.035 Ca ± 0.182	3.154 BCb ± 0.686	3.245 Bb ± 0.474
6	2.004 Da ± 0.499	3.139 Bb ± 0.345	3.254 Bb ± 0.414
8	1.929 Da ± 0.225	3.143 BCb ± 0.899	2.989 Cb ± 0.686
10	1.955 Da ± 0.212	2.636 ACb ± 0.352	2.759 Ab ± 0.441
13	1.987 Da ± 0.482	2.395 Aab ± 0.844	2.651 Ab ± 0.725
14	1.706 Ea ± 0.179	1.896 Aa ± 0.487	2.221 Db ± 0.461

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant differences (p < 0.05) for one sample over time. Different lowercase letters indicate significant differences (p < 0.05) for different samples at the same time.

and

Table A7. Relative metmyoglobin content of the beef samples over time.

MMb
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0	0.627 Aa ± 0.032	0.645 Aab ± 0.061	0.654 Ab ± 0.101
1	0.654 Ba ± 0.125	0.687 Bab ± 0.171	0.769 Bb ± 0.042
2	0.848 Ca ± 0.071	0.736 Bb ± 0.087	0.727 Cb ± 0.089
6	1.160 Da ± 0.056	0.715 Bb ± 0.086	0.701 Db ± 0.066
8	1.271 Ea ± 0.096	0.801 Cb ± 0.145	0.794 Eb ± 0.108
10	1.303 Ea ± 0.072	0.924 Db ± 0.066	0.827 Fc ± 0.075
13	1.401 Fa ± 0.200	1.248 Ea ± 0.323	0.819 Eb ± 0.157
14	1.282 Ea ± 0.110	1.326 Ea ± 0.307	0.914 Gb ± 0.144

Results listed as mean ± standard deviation of three animals, with three technical replicates per animal, measured in triplicate (27 values per atmosphere) (three means of nine measurements per atmosphere and time). Different uppercase letters indicate significant differences (p < 0.05) for one sample over time. Different lowercase letters indicate significant differences (p < 0.05) for different samples at the same time.

.

The principal component analysis was calculated via Origin Pro 2023 (Co. OriginLab Corporation, Northhampton, MA, USA), using the principal component analysis app. The linear fit equations for the L*a*b* values, color differences, and Mb redox states were calculated using Origin Pro 2023, and the results can be seen in

Table A8. Linear fit equations of the L*, a*, b* values, ΔE₀₀, and the myoglobin redox states versus time.

Linear Fit Equation y = a + b × x
Atmosphere
	1% O2 (-)	20% O2 (-)	70% O2 (-)
L*
a	34.23 ± 0.408	36.32 ± 0.700	38.68 ± 0.629
b	0.026 ± 0.056	−0.026 ± 0.084	0.038 ± 0.080
rP	0.188	−0.127	0.189
a*		*	*
a	6.974 ± 0.690	15.85 ± 1.619	17.26 ± 1.011
b	−0.136 ± 0.077	−0.570 ± 0.205	−0.508 ± 0.126
rP	−0.586	−0.750	−0.855
b*	*
a	6.338 ± 0.541	12.06 ± 1.028	13.81 ± 0.632
b	0.254 ± 0.079	−0.038 ± 0.122	−0.099 ± 0.079
rP	0.797	−0.126	−0.455
ΔE00
a	4.631 ± 0.593	5.723 ± 0.870	3.856 ± 0.501
b	−0.035 ± 0.063	0.002 ± 0.097	0.094 ± 0.059
rP	−0.240	0.010	0.578
OMb
a	1.930 ± 0.078	2.693 ± 0.280	3.248 ± 0.172
b	−0.007 ± 0.010	−0.012 ± 0.037	−0.052 ± 0.022
rP	−0.292	−0.130	−0.700
DMb	*
a	1.404 ± 0.062	1.107 ± 0.021	1.118 ± 0.012
b	−0.019 ± 0.007	0.006 ± 0.003	0.001 ± 0.002
rP	−0.768	0.590	0.314
MMb	*	*
a	0.661 ± 0.044	0.637 ± 0.036	0.728 ± 0.026
b	0.064 ± 0.008	0.028 ± 0.006	0.008 ± 0.004
rP	0.956	0.899	0.581

The equations and significances of the slopes of the linear fitted data with y = respective value, x = time, a = intersection of the y-axis ± standard error, b = slope ± standard error, and r_P = Pearson correlation coefficient. Significant differences (p < 0.05) of the slope to 0 are indicated using *.

.

3. Results and Discussion

3.1. Color Values and Color Difference

The color development of the beef samples over the storage time, expressed as L* a* b* values, can be seen in Figure 1a–c. The use of color values is a suitable method to analyze even nuanced color changes, providing detailed insight into the quality of color change [4]. To quantify the color changes, the color difference ΔE₂₀₀₀ (Figure 2) can be used. Because the color difference does not provide insight into the direction of color change, the combination of color difference and the color values is necessary to recognize the full extent of color change [4]. Additionally, for a better understanding, photos of the samples over the storage time can be seen in

Table 1. Example photographs of the packaged meat samples captured on day 0, 7, and 14 of the experiment.

Photographs
Atmosphere
Time (d)	1% O2 (-)	20% O2 (-)	70% O2 (-)
0
7
14

.

In Figure 1a, the lightness development can be seen. The lightness (L*) values on day 0 increase with the oxygen content of the packaging atmospheres. The samples show a lightness of L* = 34.15 (1%), L* = 35.40 (20%), and L* = 37.82 (70%), respectively. The lightness of the 1% and the 70% samples is significantly different (p < 0.05), with 20% showing no significant difference in either. This trend can be observed over the whole storage time. An atmosphere containing 1% oxygen shows no significant changes over time and reaches values in the range of L* = 32.38 through 35.16. The 20% samples reach slightly higher L* values between L* = 33.82 and 37.75. In this case, there is a significant increase in lightness over the storage time (p < 0.05). An atmosphere of 70% oxygen leads to the highest lightness of L* = 37.51 through 39.94, with no significant changes over time. However, the lightness is significantly higher compared to that of the 1% samples over the entire time period. The differences in lightness for the different atmospheres can be explained by the increased blooming and therefore, increased OMb formation, of the meat [4,31,32]. A slight but not significant increase in the lightness of the samples over time could also be seen in a previous study [4].

In Figure 1b, the a* values for redness over a 14-day period are shown. The a* values which, in the positive range indicate redness, show a pattern similar to that of the L* values on day 0. The lowest a* value of a* = 5.14 can be seen in the 1% atmosphere samples. The 20% samples show a significantly higher value of a* = 9.52. The 70% atmosphere samples show the highest value of a* = 13.03, which is significantly different from that of the other samples. The increase in redness with increasing oxygen content is well known and can be explained by an increasing OMb formation as the oxygen partial pressure in the atmosphere increases [2,4,22]. The increased redness of the samples with an increasing oxygen content in the atmospheres can be seen throughout the experiment. The redness of all samples significantly increased on day 1 to a* = 8.57 (1%), a* = 16.04 (20%), and a* = 17.07 (70%), respectively. The increase is the highest at 20% oxygen content, followed by 70% oxygen and then 1% oxygen. The delayed increase can be explained by the diffusion of oxygen into the deeper parts of the meat tissue, leading to the oxygenation of Mb in deeper layers, contributing to the measured redness [2,33]. The 1% oxygen atmosphere shows a smaller increase due to the low oxygen concentration.

The reason for the smaller difference in redness of the 70% samples between day 0 and 1 compared to the results for the 20% samples could be explained by the fact that the higher partial oxygen pressure leads to an increased OMb formation [2,22]. Therefore, most of the OMb formation process might be finished on day 0 in the 70% samples, while the process is delayed in the 20% samples. From day 1 through 14, all samples show a significant decrease in redness. The 1% samples show results ranging from a* = 7.32 on day 2 to a* = 4.37 on day 14, which is a relatively narrow range, representing a smaller decrease compared to that of the other samples. The 20% samples show a drop from a* = 16.04 on day 2 to a* = 5.28 on day 14. On day 14, there is no significant difference in redness between the 1% and 20% samples. The 70% atmosphere shows a decrease from a* = 17.07 on day 2 to a* = 8.39 on day 14, with significantly higher results than those for the other atmospheres at the end of the experiment. The decrease in the redness of the samples aligns with the decrease in OMb and increase in MMb, which will be discussed in the sections addressing the Mb forms. The 20% and 70% oxygen atmospheres show a lower a* stability compared to that for the 1% group. This is because in the 1% atmosphere, the lack of oxygen inhibits most of the OMb formation. Even though the redness strongly decreases over time, the higher oxygen content of the 70% samples seems to stabilize the red color enough to produce an increased redness, even after 14 days of storage.

The b* values, which describe the yellowness of the samples, can be seen in Figure 1c. Similar to L* and a* development, the general pattern of increasing color values with increasing oxygen content can also be observed here. The initial values on day 0 are b* = 5.59 (1%), b* = 8.86 (20%), and b* = 11.54 (70%), respectively. All values are significantly different (p < 0.05). The relationship between increased redness and yellowness during OMb formation has been reported in the literature [31]. The 1% and 20% samples show a sharp increase in yellowness to b* = 8.92 (1%) and b* = 13.86 (20%), respectively, on day 1. The 70% samples increase at a lower rate until day 2, reaching a value of b* = 14.29. From this point until the end of the experiment, the samples are divided into two groups. The 1% samples show significantly lower b* values than the 20% and 70% atmosphere groups, which do not differ from each other. From day 2, the two groups show different b* trends. The 1% atmosphere group does not change statistically between days 2 and 14 but shows a numerical increase in yellowness, with a final value of b* = 8.90. However, the 20% and 70% samples both show a significant decrease in yellowness over time, reaching values of b* = 11.07 (20%) and b* = 11.44 (70%). As both the a* and b* values increase during the OMb formation in meat, the increase in yellowness over the first few days and the lower b* values of the low oxygen packages align with the results of a previous study [31].

The color difference ΔE₂₀₀₀ is used to determine the extent to which color changes affect human perception. The interpretation of the values is conducted according to the methods of Wieser [34]. In this study, the color differences are categorized into groups of ΔE < 1, imperceptible; 1 < ΔE < 2, very low, almost imperceptible; 2 < ΔE < 3, very low, partially perceptible; 3 < ΔE < 5, perceptible; 5 < ΔE, distinct, strongly perceptible. These values were also used in a previous study [4]. The color difference on day 0 was 0, since that was the reference point for the calculations. The color differences of all modified atmospheres are shown in Figure 2. On day 1, the lowest color difference was found in the 70% oxygen atmosphere group, with ΔE = 4.27 (perceptible), followed by the 1% group, with ΔE = 5.54 (strongly perceptible), and the 20% group, with ΔE = 6.79 (strongly perceptible). All values were significantly different. From day 1 to day 14, all samples showed some fluctuations and reached minimum values of ΔE = 3.60 (1%), ΔE = 3.80 (70%), and ΔE = 4.39 (20%), respectively (all ΔE values are perceptible). On day 14, the color differences were ΔE = 4.61 (1%) (perceptible), ΔE = 5.69 (70%) (strongly perceptible), and ΔE = 6.63 (20%) (strongly perceptible), respectively. The ΔE values were not significantly different from the respective values on day 1. However, the values for 1% and 70% oxygen atmospheres fluctuated enough to require changing the group categorizations according to the ΔE values. The 1% samples decreased from strongly perceptible to perceptible, while the 70% samples increased from perceptible to strongly perceptible.

Figure 2. Color difference development ΔE₂₀₀₀ and standard errors of beef samples packaged and stored under different modified atmospheres for 14 days at 2 °C. (ΔE < 1, imperceptible; 1 < ΔE < 2, very low, almost imperceptible; 2 < ΔE < 3, very low, partially perceptible; 3 < ΔE < 5, perceptible; 5 < ΔE, distinct, strongly perceptible; according to Wieser [34]).

Figure 2. Color difference development ΔE₂₀₀₀ and standard errors of beef samples packaged and stored under different modified atmospheres for 14 days at 2 °C. (ΔE < 1, imperceptible; 1 < ΔE < 2, very low, almost imperceptible; 2 < ΔE < 3, very low, partially perceptible; 3 < ΔE < 5, perceptible; 5 < ΔE, distinct, strongly perceptible; according to Wieser [34]).

On day 14, only the 1% and 20% groups were significantly different (p < 0.05), with the results for the 70% group falling in the middle (p > 0.05). In terms of overall color stability after 14 days, the 20% oxygen packaging resulted in the most color changes, while the 1% packaging achieved the lowest changes. In terms of overall color stability after 14 days, the 20% oxygen packaging resulted in the most color changes, while the 1% packaging achieved the least changes. In terms of overall color stability after 14 days, the 20% oxygen packaging resulted in the most color changes, while the 1% packaging achieved the least changes. In terms of overall color stability after 14 days, the 20% oxygen packaging resulted in the most color changes, while the 1% packaging achieved the least changes. In terms of overall color stability after 14 days, the 20% oxygen packaging resulted in the most color changes, while the 1% packaging achieved the least changes. This behavior could be explained by the lack of oxygen in the packages containing 1% oxygen, which inhibited the OMb formation. The high amount of oxygen in the packages containing 70% oxygen resulted in the increased oxygenation of Mb and therefore, most of the oxygenation was completed when the first measurement was performed; therefore, most of the color changes might not have been detected [22,32].

3.2. Relative Levels of the Redox Forms of Myoglobin

The DMb levels are shown in Figure 3a. On day 0, all samples showed significantly different DMb levels. The highest levels were reached by the 1% oxygen atmosphere sampelswith DMb = 1.59, followed by 20% oxygen (DMb = 1.23), and the 70% oxygen (DMb = 1.12) groups. The increased DMb levels resulting from decreasing oxygen concentrations in the packages were consistent, since the lower concentrations resulted in less oxygen bound to the Mb, triggering higher levels of DMb due to less oxygenation [2,4,6,22]. Over the next two days, the DMb levels of all samples decreased. The 1% oxygen samples showed the most drastic decline, reaching DMb = 1.25, followed by the 20% oxygen samples, with DMb = 1.09 on day 2. Both changes were statistically significant (p < 0.05). The samples containing 70% oxygen did not decrease significantly (DMb = 1.10) on day 2. At this point, there was no significant difference between the 20% and 70% oxygen atmosphere groups. Since the MMb values (Figure 3c) were very narrowly distributed on day 0, the biggest difference could be observed in the OMb and DMb levels. A comparison with Figure 3b shows that the DMb levels are inversed to the OMb levels. A higher DMb level resulted in a lower OMb level and vice versa. Under the atmospheric conditions of the experiment, the main forms of Mb were DMb, OMb, and MMb (with MMb being negligible on day 0). The differences consistently resulted from the oxygenation or lack of oxygenation of Mb, depending on the partial oxygen pressure [2,4,8,22]. The decrease in DMb over the first two days was due to the OMb formation caused by oxygen from the atmosphere reaching deeper tissue levels (stronger effect at 20% and 70%) and the early onset of MMb formation (stronger effect at 1%) [2,35]. From day 2, the 1% atmosphere showed significant decreases until day 14, with final levels reaching DMb = 1.16. The 20% samples showed a significant increase to DMb = 1.17. The 70% oxygen atmosphere samples increased slightly but not significantly to DMb = 1.12. On day 14, the DMb levels of the atmospheres showed a narrower distribution than that at the beginning of the experiment. There was a significant difference between the 20% and 70% oxygen atmosphere groups. The highest oxygen content resulted in the lowest DMb level at the end of the storage time. The results for the 1% and 20% samples were not significantly different.

The development of OMb is shown in Figure 3b. Since OMb is the red meat pigment, it mostly aligns with the a* development (Figure 1b), which was confirmed by previous studies and the literature [2,4,36]. As discussed in the previous paragraph, the curves of the OMb and DMb levels behave inversely because the main color differences resulted from the oxygenation of the Mb at the beginning of the experiment. The lowest level could be observed in the 1% atmosphere samples, with OMb = 1.82, followed by the 20% group (OMb = 2.27), and the 70% group (OMb = 2.73). All values are significantly different (p < 0.05). Until day 1, the OMb levels significantly increased in all atmospheres, with values of 2.46 (1%), 3.48 (20%), and 3.33 (70%), respectively. As previously mentioned regarding the increase in the a* values on day 1, the explanation could be the diffusion of oxygen to the deeper tissue layers of the meat, leading to more OMb formation over time [2,33]. From day 1 through 14, all samples showed a significant decrease in OMb levels; however, the extent was different. The 1% atmosphere showed a decrease from 2.46 to 1.76, stabilizing from day 6 through 13. The 20% oxygen atmosphere samples displayed a decrease from 3.48 to 1.90 OMB. The atmosphere containing 70% oxygen showed declines from OMb = 3.33 to 2.22. Overall, the OMb development reflected the curves of a* values: 1% oxygen showed the lowest values (the differences between the results for the other atmospheres were very high from day 1 through 10) but also the least changes, due to the low oxygen level in the packages. The 70% oxygen atmosphere resulted in the significantly highest final OMb values and greater stabilization than did the 20% oxygen samples, which showed in the highest decrease in OMb.

The MMb levels are shown in Figure 3c. On day 0, the samples were narrowly distributed at 0.63 (1%), 0.65 (20%), and 0.65 MMb (70%), respectively. The results for the 1% and 70% oxygen groups were significantly different (p < 0.05). All samples showed significant increases on day 1, reaching 0.65 (1%), 0.69 (20%), and 0.77 (70%), respectively. From day 1 to day 2, the samples start to behave differently. The results for the 1% atmosphere group increased drastically to 0.85 MMb. The 20% atmosphere results did not significantly increase (0.74 MMb), and the 70% atmosphere results significantly decreased to 0.73. From day 2 to day 13, the results for the 1% atmosphere continually increased until reaching the highest value of 1.40 MMb, which is the highest MMb value measured overall. From day 13 to 14, the MMb value then dropped to 1.29 MMb. Until day 13, the 1% atmosphere group showed significantly higher MMb values compared to those of the other atmospheres. From day 2 to day 8, the 20% and 70% oxygen atmospheres showed the same MMb trends, with no significant differences, rising to values of 0.80 (20%) and 0.79 (70%), respectively, on day 8. On day 10, the 20% atmosphere began to show significantly increasing MMb values, reaching a final value of 1.33 on day 14. The results for the 70% atmosphere, on the other hand, stayed in the same MMb range until day 13 (MMb = 0.82) and only showed a sharp increase in MMb values on day 14 (MMb = 0.91). On day 14, the 1% (1.28) and 20% (1.33) atmospheres showed no significant differences, while the 70% (0.91) atmosphere reached significantly lower MMb levels.

The overall MMb development over time suggests that a 1% atmosphere results in a rapid MMb formation starting on day 0. These results confirm the previous studies that report a promoted Mb oxidation at low oxygen partial pressures [2,15,16,17,18,19,20,21]. At 1% oxygen, the increased Mb oxidation cannot be countered by the MMb-reducing activity, resulting in rapid discoloration [21]. Higher oxygen concentrations seem to delay the MMb formation, to a certain extent. The 20% oxygen atmosphere was able to stabilize the MMb formation until day 8, while the 70% atmosphere was able to stabilize it until day 13. These results lead to the question of whether the higher oxygen content prevents the oxidation of Mb or just masks the MMb formation by forming a thicker OMb layer on the surface, while the MMb formation is unchanged deep within the meat tissue [2]. The answer depends on the oxygen concentration. According to Taylor [37], the depth of the OMb layer in the meat tissue decreased with the oxygen concentration in the surrounding atmosphere. An oxygen concentration of 80% resulted in a penetration depth of 14 mm after 5 days, which is more than that required for the 10 mm samples in the present experiment. The 40% atmosphere, however, only resulted in an 8 mm OMb layer on day 1 and was not measurable after 5 days due to MMb formation. Thus, in terms of the 70% samples, it seems likely that the oxygen permeated the whole sample, stabilizing the Mb, preventing oxidation in all layers [2,4,6,37]. For the 20% sample, a lower oxygen penetration depth is to be expected, which is why a masking effect of the OMb layer covering the MMb layer cannot be excluded. The masking of MMb describes the process of an OMb layer covering the underlying MMb. Under these conditions, it is difficult to measure the underlying MMb [24,38]. According to a recent study by Denzer et al., oxygen diffusion into the meat creates an oxygen gradient, which results in lower oxygen partial pressures below the surface. The lower partial pressures can lead to optimal MMb formation conditions [24]. Therefore, while the surface layers were stabilized by the oxygen in the 20% atmosphere, the deeper areas below the surface were oxidized to MMb, resulting in discoloration. One reason for the increase in MMb content in the 20% atmosphere after 8 days could be the declining MMb-reducing activity [21]. In color-stable meat tissue, oxygen consumption and subsequent MMb reduction prevents discoloration [21]. However, the declining MMb-reducing activity over storage time would change the equilibrium of oxidation and reduction to increased oxidation [21,23]. When comparing the MMb and the OMb values, the stabilizing effect of OMb could be observed, since the 1% atmosphere showed the lowest OMb and the highest MMb levels, while the opposite was true for the 70% atmosphere, displaying the highest OMb levels on most days (especially at day 14), as well as the lowest MMb levels.

3.3. Principal Component Analysis

The principal component analysis (PCA) can be used to show the correlation between the measured Mb, color parameters, and oxygen content of the packaging atmosphere of the samples. To show the changes in correlation over the whole storage time, three PCAs were performed on day 0 (Figure 4), 6 (Figure 5), and 14 (Figure 6).

Figure 4 shows the PCA on day 0. More detailed PCA data are listed in

Table A9. Loadings (vector coordinates) of the principal component analysis and cumulative variances.

PCA Loadings
	Day 0			Day 6			Day 14
	PC1 61.8%	PC2 29.8%	PC3 5.5%	PC1 58.8%	PC2 23.4%	PC3 11.8%	PC1 38.1%	PC2 26.2%	PC3 17.3%
L*	1.139	4.669	3.292	0.886	−2.646	1.009	−0.338	2.690	−2.570
a*	2.014	−0.363	−0.983	2.088	0.498	−0.195	2.961	0.691	−0.262
b*	1.940	−1.054	2.947	1.930	0.191	0.182	1.352	1.526	3.190
ΔE00	−1.456 × 10−19	1.212 × 10−18	3.925 × 10−17	−0.231	1.829	2.819	−1.107	2.079	0.904
OMb	1.722	−3.249	−0.636	1.873	1.400	−0.551	3.019	−0.130	0.970
DMb	−1.906	−0.303	2.801	−1.546	1.807	−0.822	0.403	−2.642	−0.428
MMb	0.433	5.721	−2.418	−1.983	−0.771	0.215	−2.292	−0.244	2.689

in the Appendix A. The total cumulative variance of 97.1% consists of PC1 (61.8%), PC2 (29.8%), and PC3 (5.5%). This means that the PCA can explain 97.1% of the variance, while 2.9% is declared random or unexplained [39]. A comparison with the results in the literature shows that this PCA displays a high cumulative variance, resulting in a precise description of the data [39,40,41,42]. The first important insight is the fact that OMb, DMb, and MMb form a triangle. That is because the dominant positive components are OMb (PC1), DMb (PC3), and MMb (PC2), meaning that an increase in one form results in the decrease in the other two. These results align with the Mb data from this study and the literature [2,4,8]. The OMb and a* vectors both show a similar component pattern, with PC1 as the major positive component. This connection also makes sense, since OMb is the red muscle pigment, and this result aligns with the findings of previous experiments and the literature [2,4,8]. Regarding the oxygen concentration of the samples, it is interesting to note that the 1% samples are scattered in the direction of DMb, while an increasing oxygen concentration leads to a shift away from DMb. These findings also align with the fact that increased oxygen concentration leads to increased oxygenation of Mb [22].

Figure 5 shows the PCA on day 6. The total cumulative variance of 93.7% consists of PC1 (58.5%), PC2 (23.4%), and PC3 (11.8%). With 6.3% of unexplained variance, the PCA is less accurate than it was on day 0, but the results are still higher than for the PCAs found in literature [39,40,41,42]. OMb, DMb, and MMb still form a triangle, for the previously mentioned reasons; however, the vector directions changed (see Table A9). The OMb values still show high similarities to the a* values; however, the b* values now also align with both. These connections have also been observed in the literature [31]. Regarding the oxygen concentration of the packages, a change can be seen compared to the levels on day 0. The 1% oxygen samples show an increased alignment with MMb and a shifting away from DMb, which is due to the high MMb formation (compare Figure 3c). Increased oxygen concentrations in the packages still lead to shifts in the OMb direction, as observed on day 0.

Figure 6 shows the PCA on day 0. The total cumulative variance of 81.6% consists of PC1 (38.1%), PC2 (26.2%), and PC3 (17.3%). On day 14 the highest percentage of unexplained variance (18.4%) can be seen. However, even in this case, the cumulative variance is still very high [39,40,41,42]. On day 14, as in the other PCAs, the Mb forms remain in a triangular direction, for the same reasons as previously mentioned. The directions shifted again (see Table A9). The connection of OMb and a* is still observable, even though the association decreased, to some extent. The oxygen concentration pattern differentiates significantly on day 14. The 1% oxygen atmosphere still shows a strong tendency towards MMb but also shifted towards DMb. The results for the 20% atmosphere are widely scattered but overall, they show a strong shift to MMb. The results for the 70% atmosphere also shifted in the MMb direction, with OMb playing the more dominant role.

Overall, the PCA shows the key findings of this study in a simple way. On day 0, the main difference between the samples is due to the oxygenation of Mb, which results in more OMb and less DMb with increased oxygen concentration. On day 6, the 1% samples are associated with MMb formation, while the other samples still show OMb association, depending on the oxygen concentration. This shows that a 1% oxygen atmosphere leads to faster MMb formation compared to that of the other atmospheres. On day 14, all samples shifted in the MMb direction, with the 20% samples showing the greatest effect and the 70% samples the least. Therefore, the 70% atmosphere can stabilize the color for a longer period compared to the 20% oxygen atmosphere, which also shows high MMb formation after 14 days.

4. Conclusions

An increased oxygen concentration in the packaging atmosphere leads to increased OMb formation and therefore, increased redness. A 1% oxygen atmosphere results in rapid MMb formation. The MMb formation can be delayed, to a certain extent, with increased oxygen concentrations in the packaging leading to the stabilization of the Mb or masking of the MMb layer by covering it with OMb, depending on the penetration depth of the oxygen. The hypothesis that a 1% low oxygen atmosphere leads to an increased MMb formation (discoloration) in the meat samples compared to that for the 20% and 70% oxygen atmospheres over a storage time of 14 days can be confirmed. The non-invasive measurement over the storage time provided deeper insight into the myoglobin interconversion mechanisms.

In future studies, the question of whether a 1% oxygen atmosphere leads to the highest MMb formation will be investigated by employing oxygen concentrations ranging from of 0% to 5%. Subsequently, trials employing partial vacuum packaging will be carried out to determine whether or not these conditions can be reproduced using non-ideal packaging containing residual oxygen in order to identify the causes of quality defects such as discoloration.