Overview: Diacetyl is a powerful flavouring molecule with a distinct ‘buttermilk’ character having a taste threshold in water of around 5.4 ppb. The presence of Diacetyl in fruit juices is extremely undesirable, occurring through microbial activity from poor processing or handling procedures. The off-flavour remains after heat treatment to pasteurise the juice. Lactobacillus and Leuconostoc spp are the most usual causative organisms, although yeasts can also be involved.
Fruit juices provide ideal hosts for the survival and growth of a range of microbes. The AIJN specifies maximum limits for a number of components that can result from microbial contamination of the juice, or from poor plant hygiene during production:-
Volatile Acids as Acetic Acid = 0.4 g/l = 400 mg/l
Ethanol = 3.0 g/l = 3000 mg/l
D/L Lactic Acid = 0.2 g/l = 200 mg/l
While undesirable, these materials generally do not have an adverse effect on flavour, unless at extremely high concentrations. For example, the taste threshold in water for Acetic Acid is 300 ppm, with an odour threshold of 24 ppm. For ethanol, the taste threshold is even higher at around 14,000 ppm and for Lactic Acid it is around 170 ppm.
However, there is no limit given for the presence of a very powerful flavour component, Diacetyl, which has a distinct ‘buttery’ taste with a flavour threshold as low as 0.0054 ppm (5.4 ppb) in water. Diacetyl can be present in juices because of microbiological contamination and is highly undesirable.
Historically, ‘buttermilk’ off-flavours were first detected with the introduction of frozen concentrated orange juice (FCOJ) in Florida in 1946 using low temperature evaporators. The issue became more widespread in the 1950’s with the evolution of the low temperature falling film evaporator, starting with single effect and single pass, then single effect with recirculation and finally multi-effect with recirculation configurations. Some existing tube-in-shell evaporators were linked together to produce a variety of stages and effects, which resulted in long resident times, typically 1 hour for a 40-50 brix concentrate. In turn this allowed both microbial and enzymatic changes to occur. Early evaporators were frequently colonized by Lactobacillus and Leuconostoc spp, which apart from being insanitary, gave rise to the buttery off flavours of diacetyl in the concentrates. Pre-heat treatment was introduced as a means to control both microbial and enzymic activities. The incidence of Lactobacillus and Leuconostoc spp activity in citrus concentrates has largely been overcome with the advent of the modern TASTE (“thermally accelerated short time evaporator”) system.
Diacetyl off-flavour has also been found in orange juice that had been reconstituted from frozen concentrate, left in bulk tanks for 2 -3 days before pasteurisation by hot filling. Although the final product was microbiologically satisfactory, the damage was done through microbiological activity before heat treatment.
There have also been incidents of the ‘buttery’ off-flavour found in clarified juices. For example, in the traditional process for the manufacture of clarified lime juice, unpasteurised liquid from crushed limes was fed to large tanks and held at ambient temperature for 2 – 3 weeks. During storage, a layer of oily pulp formed over the surface protecting the bulk from aerobic microbial activity. The juice was then decanted, filtered and concentrated. The process was satisfactory for many years, but towards the end of the lime season when fruit supplies diminished, it took longer and longer to fill the tanks. The repeated disturbance of the surface potentially allowed microbiological activity to occur. The resulting juice concentrate could be contaminated with Diacetyl. While it is highly unlikely that Lactobacillus or Leuconostoc spp could survive in the low pH environment of lime juice, yeasts are able to do so.
The modern process for lime juice clarification with improved plant cleaning regimes, use of enzyme treatment and ultrafitration has effectively eliminated the occurrence of Diacetyl in the concentrate.
The citrus industry adopted the use of the Vosges – Proskauer (VP) reaction as an analytical tool to estimate the levels of both Diacetyl and AMC (acetyl methyl carbinol) in juices by colourimetric analysis. The method has been used as a rapid QC test as an indicator of both plant hygiene and juice quality.
While AMC, or Acetoin does not have a particularly low taste threshold (around 17 ppm) it is involved in Diacetyl reactions. Acetoin occurs naturally in citrus fruit and increases with maturity when levels of 4ppm have been found in orange juice.
In general, the level of Diacetyl in a juice should be less than 0.2 ppm and preferably less than 0.1 ppm.
A more sophisticated procedure involving liquid/liquid extraction and GC-MS analysis of juices can quantify other marker species such as:-
In the method, any Diacetyl is lost during the extraction stage, but the results give a measure of past microbial activity.
[From our analysis of lime juice samples using the direct VP method, we have established a recommended maximum level of Diacetyl in 450gpl lime juice concentrate as 1.5 ppm, for example].
© Gerald McDonald & Co Ltd. Author: Ray Mears