Using Isothermal Microcalorimetry to Understand Sourdough Starters

The calScreener is an isothermal microcalorimeter which can assess the heat produced by living organisms to give reliable data about the metabolic activity within the sample. The 48-vial plate sits on top of a heat-flux detecting sensor within the instrument, which can measure the heat produced at an interval of one measurement/second. This data is retrieved by Symcel’s software to produce an individual, real-time heat flow profile for each vial on the plate.

This technology was suitable to investigate the activity and affecting paraments on a sourdough starter. Sourdough starters provide the natural leavening agent when making doughs and batters and they simply consist of water, flour, and salt. This mixture is left to ferment for around a week, which allows the bacteria present in the water, flour, and the air to feed off the sugars within the flour to produce carbon dioxide, acids, and alcohol. The starter is continually fed more flour and water to ensure microbial succession, reaching a point where the only organisms still present are wild yeasts and good bacteria like those that produce lactic acid and give sourdough its distinctive taste.

The calScreener was used to investigate a variety of parameters affecting the sourdough starter such as finding the metabolic activity of different concentrations of sourdough starter, the effect of different flours on metabolic activity and the effect of different salt levels on metabolic activity.

Experiment 1: Varying Concentrations of Sourdough Starter

Firstly, the metabolic activity of the sourdough starter at different concentrations was investigated; factors like temperature, amount of water and the amount and type of flour used were the control variables. This allowed the most appropriate concentration to be found, which would then be used in future experiments which investigated the effect of changing other factors like flour type.

The vials were filled with a 1:1 ratio of water and flour, which would feed the bacteria present in the sourdough starter. A 5-step 10-fold dilution series was then completed where 440 µl of water was mixed with 0.2 g of sourdough starter. 40 µl of this mixture was then taken and place into a sample tube of water (400 µl) to create the next dilution. This process was repeated for all the concentration we were testing, which can be seen in Figure 1

Figure 1: Shows the real-time heat profiles for each concentration of sourdough starter.

100 µl of each dilution was placed in a vial. These vials were then sealed to create individual incubation systems, placed into the instrument, and left to ferment over 3 days, where the heat flow profiles were simultaneously being produced by the connected software.

The heat profiles of these concentrations can be seen in Figure 1. This data can be used to observe detailed metabolic activity within the samples, such as:

  • Peak of Line: Shows the maximum heat produced by the bacteria, which relates to when they were most active. We can also see at what time during the experiment this took place.
  • Decline Slope: For each concentration the slope after the maximum is reached shows how quickly the bacteria died off, meaning their feed supplies ran out.
  • Incline Slope: In Figure 1 for some concentrations, we see a first smaller peak, a small decline and then the start of our maximum activity peak.

This allowed us to pick a suitable dilution for future experiments in this series.



Experiment 2: Effect of Flour Type on the Metabolic Activity within a Sourdough Starter

Next, the effect of flour type was investigated; the sourdough starter has been continually fed with plain flour so there was interest to see if different flour types would affect the metabolic activity of the bacteria present. The chosen dilution (10-2) from the previous experiment was prepared and added to the vials in a 1:1 ratio with the flour types, which were plain white, strong white and rye. It was thought that plain flour and strong plain would give similar results as the starter would continue to metabolise as normal. However, with the rye flour we would expect to see the lowest metabolic activity due to being a different composition to what the sourdough starter is adapted to. So, this would mean the bacteria would metabolise less or take longer to reach that maximum point.

Figure 2: Shows the average results from the three repeats of each flour type. Red = Plain, Purple = Strong, Green = Rye

Our hypothesis wasn’t seen in these results, with all the flour types reaching a similar maximum at round the same time. Rye both reached its maximum and dropped off slightly before the other flours, possibly showing that the bacteria used up the food resources quicker here or potentially run out of what it could metabolise, whereas in the other cases there was more they were capable of metabolising leading to longer activity times. The difference was slight though, and so full statistical analysis will be needed to ascertain any significance.


These results show that the calScreener is an ideal instrument to conduct our experiments on sourdough, with easy sample preparation, simple procedure, and detailed real-time results. It will also allow us to continue experimenting with the parameters that affect sourdough to deepen our research with experiments like varying the salt concentration or using different source of water to be used as feed, which could potentially introduce new bacteria.