3.4.1. A look at yoghurt aroma compounds
The typical flavours of fermented milk are mainly due to a blend of the following compounds: lactic acid, carbon compounds such as acetaldehyde, acetone, acetate and diacetyl, non-volatile acids such as pyruvic, oxalic and succinic acids, volatile acids such as acetic, propionic and formic acids and products from the thermal degradation of proteins, lipids or lactose.
Ott et al. identified 91 aroma compounds (GC-olfactometry) in yoghurt among which 21 were detected more frequently and would thus have a major impact on flavour. Acetaldehyde is found in significant quantities and is responsible for the characteristic smell of yoghurt. Diacetyl, pentane-2,3-dione, and dimethyl sulphide also have a major impact on yoghurtflavour [36, 37].
Acetaldehyde was firstly reported by Pette et al.  as the main aromatic compound in yoghurt. During manufacture, production of this compound is only highlighted when a certain level of acidification is reached (pH 5.0). Concentrations found in the final product are 0.7 to 15.9 mg.kg-1.The maximum amount is obtained at pH 4.2 and stabilizes at pH 4.0. The production of acetaldehyde and other flavour compounds by S. thermophilus and Lb. bulgaricus occurs during yoghurt fermentation and the final amount is dependent on specific enzymes which are able to catalyse the formation of carbon compounds from the various milk constituents.
Three metabolic pathways producing acetaldehyde were identified and some pathways may take place simultaneously :
From glucose in the glycolytic pathway,
From the degradation of DNA,
From L-threonine with threonine aldolase.
However, 90% of acetaldehyde produced by Lb. bulgaricus comes from glucose and 100% in the case ofS. thermophilus .
Diacetyl and pentane-2,3-dione also have a significant impact on the final aroma of yoghurt: 1 mg of diacetyl and 0.1 mg of pentane-2, 3-dione per kg of yoghurt are produced by lactic acid bacteria during fermentation.These diketones are produced by decarboxylation of their precursors, 2-acetolactate and 2-aceto-hydroxybutyrate . These compounds are thermally unstable and in the presence of oxygen are converted into their corresponding diketones[40, 41]. Moreover, during storage at 4 °C, the concentration of the two diketones increases slightly  due to the basal metabolic activity of the bacteria.
Agitating a mixed culture of Lactococcus and Leuconostoc promotes diacetyl production by allowing oxidative decarboxylation of 2-acetolactate [42, 43]. In unstirred cultures, the redox potential of the medium decreases rapidly at the start of fermentation. Only acetoin and 2-acetolactate areproduced. The authors also showed that controlled oxygenation of the Lactococcus lactis ssp. lactis culture medium favoured diacetyl production by increasing the activity of diacetyl synthase .
Neijssel et al.  showed that the distribution of carbon flux from pyruvate depended on the NADH / NAD+ ratio, intracellular redox potential or the concentration of metabolites and particularly that of pyruvate.Finally, the authors suggested adding air or oxygen to milk in order to increase the amount of diacetyl in cheese .
References  and  are the only articles that mention dimethyl sulphide as a compound having a significant impact on the flavour of yoghurt. The metabolic pathways involved in the synthesis of sulphur compounds are not well-known in yoghurt. However, the literature mentions these synthetic pathways in the development of cheese flavour.
In general, the majority of sulphur aromatic compounds come from methionine . Methanethiol is easily oxidized to dimethyl disulphide and dimethyl trisulphide . The appearance of these compounds is the direct result of the methanethiol content and is modulated by the low redox potential in Cheddar. Dimethyl sulphide is produced by a metabolic pathway that does not involve methanethiol, but that is different to that of dimethyl sulphide and trimethyl disulphide from methionine .
Studieshave also shown that when the redox potential decreases, methanethiol and hydrogen sulphide concentrations increase . Moreover, the cheeses to which reducing compounds (dithiothreitol or glutathione) were added contained higher amounts of sulphur compounds and had better qualitative and quantitative flavour performances . It therefore seems that a reducing environment is essential for the production of aroma compounds by bacteria. If a cheese is exposed to air, the redox increases and this leads to the oxidation of sulphur compounds, resulting in lower quality aromatics.
3.4.2. Impact on aroma biosynthesis by lactic acid bacteria
Studies on aroma biosynthesis by LAB usually take into account environmental factors such as pH and temperature. However, the Eh of the medium has not yet been considered, although it is supposed to affect bacterial metabolism [49, 50]. Martin et al. determined to what extent Eh can affect the metabolic pathways involved in the production of aroma compounds in Lb. bulgaricus and S. thermophilus.Four aroma compounds (acetaldehyde, dimethyl sulphide, diacetyl and pentane-2,3-dione) were chosen as metabolic tracers of lactic acid bacteria metabolism. The same gaseous conditions as in the study ofthe effect of Eh on model acid skim milk gels and non-fat yoghurt were chosen. The amounts of each of the four aroma compounds extracted using a headspace solid-phase micro-extraction technique (HS-SPME) and analysed using gas chromatography coupled with mass spectrometry (GC-MS) during 28 days of storage are reported in Table 3.
Firstly, the authors focused on the impact of these different Eh conditions on the biosynthesis of these four aromas by bacteria after one day of storage . In the standard yoghurt (made in ambient air), diacetylwas observed in the highest concentrations,and acetaldehyde the lowest. This result is contrary to the literature where the lowest concentrationswere reported for dimethyl sulphide (0.013-0.070 mg.kg-1; measured using dynamic and trapped headspace GC [37, 41]). In the same way, published concentrations were generally higher for acetaldehyde (0.7-15.9 mg.kg-1) than in our standard yoghurt (0.18 mg.kg-1). In the literature, the concentrations of diacetyl (0.31-17.3 mg.kg-1) and 2,3-pentanedione (0.02-4.5 mg.kg-1) were lower than in our standard yoghurt (162 mg.kg-1 and 115 mg.kg-1 respectively).An explanation for these differences can be put forward: the quantification technique used by Ott et al. and Imhof et al. was dynamic and trapped headspace GC. This technique requires Tenax® traps which may be saturated, as we showed in a preliminary experiment. Furthermore, in our study,to enable a more complete extractionof the aroma compounds, a saturated solution of NaCl was added to the yoghurt. Finally, we did not use the same speciesof LAB as Ott and Imhof, which may have resulted in different quantities of the various aroma compounds.
Yoghurts made with air bubbling had significantly higher concentrations of acetaldehyde and diacetyl compared to standard yoghurts. The concentration of dimethyl sulphide was significantly lower and that of pentane-2,3-dione was the same.
With N2bubbling, the concentration of acetaldehyde was similar to that in yoghurts made with air bubbling, whereas the concentration of dimethyl sulphide was lower. The concentration of diacetyl was the same as in standard yoghurts and the concentration of pentane-2,3-dione was not significantly different from that in yoghurts made in air (bubbling or not).
The authors also demonstrated that oxidative Eh conditions clearly increased the production of aroma compounds . These results are consistent with the bibliography. Oxidative conditions stimulated the production of volatile sulphur compounds such as dimethyl sulphide, and aldehydes such as acetaldehyde . In the presence of oxygen, the oxidative decarboxylation of 2-acetolactate and 2-aceto-hydroxybutyrate to diacetyl and pentane-2,3-dione respectively was also favoured[40, 42, 44, 52]. For diacetyl, our result can be explained by the fact that in anaerobic conditions lactic acid bacteria dehydrogenate the NADH produced during glycolysis via lactate dehydrogenase (LDH) activity. Boumerdassi et al. confirmed that oxygen increases NADH oxidase activity , which causes NADH re-oxidation at the expense of LDH, butanediol dehydrogenase and acetoin dehydrogenase activity. Then, excess pyruvate is partially eliminated through acetolactate production, which increases diacetylproduction .
Finally, bubbling with N2 – H2 (reducing conditions), the concentration of acetaldehyde and pentane-2,3-dione was the same as in standard yoghurts. The concentration of dimethyl sulphide was the same as in yoghurts made without oxygen and the concentration of diacetyl was significantly lower than under the other three Eh conditions.
Then,  kept the yoghurts in Hungate tubes at 4 °C for 28 days in order to prevent exposure of the contents to oxygen, and the gaseous conditions applied to the milk arethus assumed to be constant during storage.
|Gaseous conditions and storage period (days)||Aroma compound (mg.kg)|
|1||0.18a ± 0.02||10.16a ± 0.59||162.08a ± 13.49||115.25a ± 33.70|
|7||0.13ab ± 0.02||10.16a ± 0.99||115.55cb ± 5.13||84.33ab ± 1.86|
|14||0.10b ± 0.00||9.52a ± 0.42||91.63c ± 8.02||71.56ab ± 4.49|
|21||0.18a ± 0.00||13.06b ± 1.21||141.29ab ± 11.77||65.19b ± 5.49|
|28||0.16ab ± 0.01||11.82ab ± 0.48||112.66c ± 7.56||51.79b ± 2.95|
|1||0.28a ± 0.00||5.27a ± 0.53||299.90a ± 18.37||123.47a ± 3.23|
|7||0.14b ± 0.01||6.34ab ± 0.30||127.58bc ± 3.93||83.98b ± 2.20|
|14||0.11b ± 0.01||6.85b ± 0.32||104.23b ± 4.66||78.10b ± 3.15|
|21||0.24a ± 0.02||9.33c ± 0.55||143.84c ± 6.69||67.89c ± 0.80|
|28||0.13b ± 0.02||6.78b ± 0.30||101.91b ± 3.03||54.36d ± 4.19|
|1||0.35a ± 0.04||3.72a ± 0.22||147.79a ± 9.91||110.87a ±4 .49|
|7||0.22bc ± 0.02||5.68ab ± 0.34||103.13bc ± 6.50||75.67b ± 1.98|
|14||0.18c ± 0.01||6.16b ± 0.17||78.80c ± 4.60||58.16c ± 0.48|
|21||0.28ab ± 0.04||10.34c ± 1.37||122.27ab ± 8.94||51.95c ± 2.29|
|28||0.18c ± 0.00||7.16b ± 0.77||84.76c ± 9.92||38.72d ± 3.03|
|N2 – H2 bubbling|
|1||0.17ab ± 0.00||2.71a ± 0.57||102.73a ± 9.10||76.99a ± 5.90|
|7||0.13bc ± 0.02||5.49b ± 0.53||89.61ab ± 7.70||66.10ac ± 2.29|
|14||0.11c ± 0.01||6.07bc ± 0.43||74.35b ± 5.42||57.60bc ± 0.86|
|21||0.19a ± 0.01||7.48c ± 0.53||112.09a ± 7.82||52.10bd ± 0.59|
|28||0.12c ± 0.01||7.22c ± 0.63||78.36b ± 9.30||40.60d ± 7.93|
Evolution of average amounts of aroma compounds (mg.kg-1) quantified in headspace of yoghurts made under different Eh conditions (ambient air, bubbling with air, bubbling with N2 and bubbling with N2 – H2) during 28 days of storage. ACH: Acetaldehyde (A); DMS: Dimethyl sulphide (B); DY: Diacetyl (C); PTD: pentane-2,3-dione (D). Values are means of experiments carried out in triplicate. Values in the same column should be compared.
During the 28 days of storage, for the standard yoghurt, the quantities of acetaldehyde and dimethyl sulphide produced were relatively stable, while diketone concentrations significantly decreased.
Foryoghurts made with air bubbling, the aroma profiles remained almost constant. During storage, the concentration of acetaldehyde decreased slightly whereas that of dimethyl sulphide increased slightly. The diketone concentration significantly decreased.
For yoghurts made without oxygen (bubbling with N2), the quantities of acetaldehyde, diacetyl and pentane-2,3-dione decreased during storage while that of dimethyl sulphide increased.
For yoghurts produced under reducing conditions (bubbling with N2 – H2), the aroma profiles during storage were the same as those made without oxygen. The concentration of acetaldehyde, diacetyl and pentane-2,3-dione decreased while that of dimethyl sulphide increased.
Furthermore, during storage, different profiles were observed for the four aromas depending on the Eh conditions . Under oxidizing conditions (+170 to +245 mV), the concentration of acetaldehyde was relatively stable during storage, which is in accordance with the literature [9, 55, 56] and the concentration of dimethyl sulphide was also stable.On the contrary, under reducing conditions (-300 to -349 mV), the concentration of acetaldehyde decreased and that of dimethyl sulphide increased. The metabolic pathways involved in the biosynthesis of sulphur compounds are still unclear. Under reducing conditions, it seems that another pathway promotes the production of dimethyl sulphide and that acetaldehyde may be reduced to ethanol. For diketones, whatever the Eh conditions, the concentration decreasedduring storage. Diacetyl and pentane-2,3-dione can be reduced respectively to acetoin and pentane-2,3-diol .