Causes of Oven Spring
For a long time I have been puzzled by the variety of explanations offered for oven spring. So finally I decided to search for scientific papers which might shed some light on the subject. Even in the scientific literature there is more than one opinion but I think that it is not hard to figure out the truth. In case anyone else is interested, I thought that it might be useful to record what I found. The full discussion is a bit technical but there is a summary at the end. To make the results more concrete I have calculated the size of the effects for a typical loaf of homemade bread.
In reading books, and articles accessible on the internet I found two main schools of thought. The first was due to Professor R. C. Hoseney of Kansas State University and the second, more recent, one was propounded by A. H. Bloksma of the TNO Cereals, Flour, and Bread Institute of the Netherlands. I have not been able to locate copies of the original papers by these authors but I was able obtain a copy of Chemistry and Physics of Baking, edited by J.M.V. Blanshard, P. J. Frazier, and T. Galliard, 1986, which is the proceedings of a symposium held in 1985 and which includes papers by both Hoseney and Bloksma. A free copy is linked here [1]. Both authors provide some rough calculations of gas production and Bloksma shows a graph, but without any details of the underlying calculations. However, I also found a useful related paper by Fan, Mitchell, & Blanshard (Journal of Food Engineering 41 (1999) p69) which presents the equations for a detailed model of oven spring and also includes the dynamics of bubble growth. The link is here [2].
To provide estimates of the relative importance of the different contributions I will assume the following parameters. The nominal homemade loaf has a hydration of 70%, and is proofed to double its initial volume before baking. If the dough density before proofing is 1.2 g/cc, the initial loaf volume divided by the water volume is (1.7/1.2) / 0.7 = 2.0. After bulk fermentation the dough is saturated with CO2 and therefore contains approximately 1.7 g of CO2 per liter of water. I will assume an initial temperature of 20° C and that the oven spring continues up to a temperature of 70° C.
There are several possible contributors to oven spring:
1. The final burst of yeast fermentation
2. Thermal increase of gas volume as the dough temperature rises from room temperature to about 70° C.
3. CO2 dissolved in the liquid phase of the dough being forced out of solution as the temperature rises.
4. Ethanol evaporating.
5. Water evaporating.
Yeast Fermentation. This is a popular explanation but Hoseney computes that it contributes only 1% of the total rise in the case he was considering. Bloksma does not even provide a value, presumably because it is so small. In my own experiment [3] I saw no indication of a significant burst of yeast activity. Hoseney and Bloksma agree on this one.
Thermal increase of gas volume. To a first approximation the gas volume expands in proportion to the absolute temperature as it rises from room temperature , 20° C or 293 K, up to the end of oven spring at about 70° C or 343 K. The expanded gas volume is thus (343 - 293)/293 = 1.17 times the original gas volume. The expansion is therefore 8% of the original volume of the proofed loaf, which is 50% gas by volume. Hoseney and Bloksma agree on this one, too.
Dissolved CO2.A saturated solution of CO2 in water contains about 1.7 g of CO2 per liter of water at 20° C but only 0.5 g/L at 70° C. Each liter of water therefore expels at least 1.2 g of CO2, which takes up a volume of 0.8 L at 70° C. The expelled gas is thus 80% of the water volume. Since the volume of water is 1/4 of the volume of the proofed loaf, the expelled gas is 20% of the volume of the proofed loaf. Again, Hoseney and Bloksma agree.
By this point, we have accounted for a substantial amount of rise, of about 30% of the volume of the proofed loaf, but this is still short of the observed rise in the examples considered by Hoseney and Bloksma. So there must be another substantial contributor.
Ethanol evaporation. This is the solution suggested by Hoseney, since ethanol is produced in a quantity equal to the CO2. One might guess that it could potentially expand to the same volume as the CO2 above its boiling point of 78° C if it could somehow be separated from the water. However, Hoseney offers no mechanism for this, and Bloksma dismisses ethanol. Doc Dough points out that an ethanol-water mixture boils at a single temperature near 100° C. In other words, the mixture behaves essentially like pure water. In the absence of any plausible mechanism for ethanol, I conclude that ethanol has no significant effect on oven spring, in agreement with Bloksma, but not Hoseney.
Water evaporation. This is the final contribution proposed by Bloksma but completely ignored by Hoseney. At first glance it might seem implausible because unleavened dough does not rise significantly during baking, although there is plenty of water. But the trick is that the evaporation of water into the bubbles of CO2 produces an increase in volume which is proportional to the initial gas volume. So unleavened bread with no CO2 will not show any effect of water vapor while risen dough will expand significantly. In addition, the expansion of the bubbles will reduce the partial pressure of CO2 and thus bring more CO2 out of solution.
To calculate the expansion, we consider a bubble of CO2 in equilibrium with liquid water. At room temperature the saturated water vapor pressure is low. However, at 70° C the saturated water vapor pressure rises to roughly 0.3 atmospheres. As the water vapor enters the bubble, the bubble must expand to maintain the pressure of approximately one atmosphere. For a water vapor pressure of 0.3 atm, the required expansion factor is 1/(1 - 0.3) = 1.4. Taking the previous effects into account, the expanded CO2 volume is 80% of the volume of the proofed loaf volume. The effect of water vapor is to increase this to 1.4 x 80 = 112% so the contribution from water is 32%, about the same as the previous two contributions added together. The total potential oven spring is then 62% of the volume of the proofed loaf.
The relative magnitudes of these three calculated contributions agree fairly well with the graphs shown in Bloksma’s paper, giving some assurance that the calculations are correct. Of course they must be considered only as approximations because they are based on a simple equilibrium model while the temperature inside a real loaf varies dramatically from place to place and over time.
In summary, a typical home-made loaf has the capacity to expand its volume by about 60% in the oven, although the dough may not permit the full amount before it starts losing gas. The three main contributors are 1) thermal expansion of gas (~10%), 2) expulsion of dissolved CO2 (~20%), and 3) evaporation of water (~30%). The contributions from the final burst of fermentation and from ethanol are negligible.