Part 3: Experimenting with Bread Dough Process

This is a continuation of Part 2: Experimenting with Bread Dough Process.

I had mentioned I wanted to try adding a longer primary fermentation, a so-called “preferment” or “pre-ferment”. The first one I tried was a simplistic poolish using my typical bread formula and flour amounts. The second batch used too little yeast, and it didn’t rise very well.

The first poolish batch turned into very nice bread, bread that was accidentally ruined after it had been bagged in plastic, ready for slicing the following day and stored in the oven, when I forgot it was there and started warming up the oven for something else. The oven was somewhat over 200F when I remembered the two loaves there, but the plastic had shrunk itself around those loaves. I decided to compost that batch, as I figured that plastic fumes had saturated through the bread.

I wasn’t enthralled with the poolish process, however, because it was more complicated to make, it involved more weighing of ingredients and separate processes (particularly if I had to separately autolyse each), as well it didn’t seem to provide much advantage over the straight dough approach I’d been using judging by the results (but it was and is easy to mix). More work for the same result? That’s when I realized that the poolish methods didn’t autolyse all the flour, at least not a strict autolyse.

So I started wondering if I could autolyse all the flour much the same as the straight-dough process I’d been using, then turn some or all of that dough into a long fermentation that occurred prior to the typical overnight bulk refrigeration. Flipping the question around, it seemed easy enough to simply add a long fermentation step between the straight-dough autolyse and the typical, retarded or refrigerated overnight fermentation. Additionally, I realized that with an additional fermentation step, I could add the salt and oil immediately prior to the secondary or bulk fermentation, and not need to do much kneading the next morning before division and panning for the final proof: there would be nothing that needed to be mixed into the dough at that point in the process.

I had originally hoped the long fermentation would take 8-hours at 72F (winter room temperatures here), but it had tripled in bulk at 4.5 hours due to the 0.15% Instant Dry Yeast added to it at that point-in-time. I’ll have to reduce that amount the next time for a longer fermentation, as the schedule allows an 8-hour primary ferment without needing a third day. For reasons of simplicity of process, I opted to subject all the autolysed flour to this long fermentation at 55% water (only flour, water, vinegar, and yeast added to the dough at this point).

First, a warm autolyse of only flour and water that resulted in 100F degree dough, which was placed in the refrigerator for 1.5 hours, at which point it cooled to about 83F degrees. Then the vinegar was added and kneaded, then a small amount of yeast for the long primary fermentation was added, and this mix was returned to the refrigerator for another 1/2 hour or so, until the dough cooled to 70F, when it was removed to room temperature. This rose for the next 4 hours at room temperature (72F) for a total of 4.5 hours. Because the dough had tripled in volume, I thought that was enough, but my guess is it would still be considered immature. At this point, more yeast was added, a short knead of a minute or so followed by a 10 minute rest, then salt was added and kneaded, then oil was added and separately kneaded with 1/3 of the dough using the prior-mentioned food-processor blade technique, then reincorporated and all of it stand-mixer kneaded for another minute on the slowest speed. The dough was 79F when it entered the refrigerator for the overnight bulk or second fermentation.

This dough process seems to result in the nicest and softest crumb yet, and the crust is a somewhat darker golden color. The crust was slightly tough following baking, but as toast it’s wonderfully crisp:

I’m not certain why the crust was tougher than straight dough versions which have always been crisp shortly after baking. I think I’ll keep playing around with this additional technique for some batches and see if it can’t be better optimized.

This dough was warmed to 79F before the overnight retarded rise, but I find myself doubting whether warming is needed for this step of the process when using a long primary fermentation.

Another question that I’ve had over a number of baking cycles is how fast does the temperature of the dough rise during the final fermentation proof after panning. The dough warms up quite a bit during division and panning, before the proof itself. I recorded some temperature measurements this time.

While those data points with the proof were derived from an 86F thermostat setting, it stands to reason that even if it were set to 96F, and if the initial dough temperature (which I often abbreviate as d.temp) is =< 65F, the dough will still need to warm up, that it will take time in the chamber to do so, therefore it would only warm the cold dough at a faster rate. It further stands to reason that the size of the final proofing dough ball or mass would also affect the rate of warm up, with smaller masses warming up faster.

The baking procedures and temperatures for this formula were the same as they’ve been for the last few batches.

Sundry Reading:
Creating Artisan Breads (General Mills) PDF.
Your Guide to Preferments.

I’m also grateful I was able to obtain a copy of Bread by Jeffrey Hamelman. I haven’t had a chance to read it fully yet, but I’ve picked it up a few times, read a few pages, and each time learned something new!

2009.Nov.26

I note that this last batch using autolyse and “pre-ferment” is a darker baked color, much like the straight dough batch with 2% scalded flour. Because one loaf of the straight-dough batch with scalded flour had “flying crust”, I opted to not further explore that at the time, however it’s possible with additional optimization, particularly with the baking or mixing process, the chance of a “flying crust” could be reduced or increased. Perhaps scalded flour is a fast method of simulating longer-fermented dough.

I also note from the fermentation temperatures that it seems to take instant dry yeast about 1 hour to “wake up”.

With the straight-dough versions, I found it beneficial to warm the dough to 84F before placing in the refrigerator, and while the dough’s cool-off time in the fridge is related to the size of the dough ball (probably a function of surface area to volume), at the current dough ball size (two loaves or about 2700 grams total weight), that allows a cool off period that was long enough so there was significant swelling of the dough before the dough gets below 68F degrees, which seems a temperature threshold where yeasts’ out-gassing activity decreases. Because there’s additional fermentation with the addition of the first ferment (”pre-ferment”) in this latest process, I find myself thinking that it’s less important to get a lot of rise during the first portion of the retarded bulk ferment, a technique that seemed to work well for straight dough. I want to try reducing the initial temperature of bulk fermentation to 75F, particularly to delay and or decrease the rise occurring during the initial portion of bulk fermentation, and to try to synchronize outgassing in such a way that more of it occurs during final proof after the dough has been panned.

Besides lowering the initial temperature of the dough, shortening the bulk fermentation period would also seem a similar outgas-delaying tactic.

I’ve calculated that if I autolyse the dough to 40% hydration for the autolyse, but keep the total formula’s water at 53-55%, and further weigh 25% of the the autolysed dough and add the balance of the water to that 25%, it will end up at 100% hydration, or what is generally known as a poolish. Because 75% of the autolysed dough will be allowed to rest without the addition of yeast until incorporation at the bulk ferment stage, none of the carbohydrates in that 75% portion will be consumed by yeast during the first ferment, so I believe the poolish that does have yeast added can mature more than the last batch’s first fermentation. So, I’ve decided to keep the 0.15% IDY for the 8-hour first fermentation for at least one more try.

Using the flours I have on hand, I’ve found in the past that 40% hydration will probably cause the autolysed dough to remain mealy unless it’s pressed together. It may be too little moisture for a satisfactory autolyse and, if so, the percentage of the dough that can be brought to 100% hydration would decrease from 25% to some lower figure. If the autolyse moisture percentage is the same as the formula’s overall moisture, then any amount of autolysed dough can be used for the first ferment, as there would be no additional water needed. In this case, the first ferment wouldn’t have enough moisture to be considered a poolish, it would be a biga.

So, this will be the next experiment: split the first fermentation from 100% of the dough to 25%, as well as decrease the initial bulk fermentation temperature a few degrees.

Sundry Reading:

The somewhat high pH of hard water is undesirable because the high content of alkaline salts tends to neutralize the normal acidity developed during yeast fermentation. Since the functions of enzymes in dough are significantly affected by the pH of the medium (enzymes act at their optimum at pH levels between 4 and 5) it is evident that excessively alkaline waters, which raise the pH of the dough above the optimum range for enzyme activity, have a detrimental effect on the direction and quality of the fermentation. The alkaline pH can be adjusted by adding acetic acid, lactic acid or monocalcium phosphate.
…
On the other hand, soft waters (10-50 ppm calcium carbonate) are objectionable because they lack the gluten-strengthening minerals and tend to yield soft, sticky dough. Also, the somewhat low pH of soft waters has an accelerating effect on fermentation, requiring some reduction in fermentation time. While soft water may yield bread with fairly good volume and very even grain, its texture and colour are likely to be poor. Corrective steps include an increase in the use of yeast food and dough salt.

Perhaps the sticky dough that I have is due to the soft water I use. It could make an interesting experiment to autolyse first with water that has only had the chlorine removed, and find out if the dough made from the baker’s flour that I’ve been using is still sticky.

2009.Nov.29

I made the poolish percentage as 20% of autolysed dough weight (not a baker’s percent), instead of 25%, which allowed the autolyse hydration of all the flour to be about 43% expressed as baker’s percent.

In spite of having little-to-no oven spring, and falling a little during the bake (slightly overfermented), this bread made the crispest toast so far, and the tough crust, after loaf cooling, has disappeared. There are visible differences in the lower and side crusts in the larger photos when zoomed in.

While it’s my intent to make process and ingredient changes to the second and third ferments due to perceived issues, this particular formula is probably worth saving for the outstanding toast. It seems that crisp toast needs at least three bread characteristics: sufficient oil, open crumb, and some residual sugars remaining during the bake.

With respect to both this batch, and the last, the dough had a different feel during panning, perhaps a characterization is more dissolved feeling, certainly a little more translucent, a slightly different shade of color; both batches at this stage were more extensible.

This last batch was more work to make given the 43% hydration of the autolysed dough. The blade attachment of the food processor seems to work well for this situation, trying to mix something much more liquid into what is essentially hard and dense. The prior batch used the overall formula’s hydration during autolyse, and that makes for much easier mixing.

The next batch will have process changes to bulk and final fermentations, to re-adopt a process I used with the straight dough versions, and see how it compares to these last two batches.

2009.Dec.1

I just noticed the bread dough rise times are partly a Fibonacci sequence: 8, 5, 3.

A yeast-related question that has been simmering in the mind is whether the poolish, were it to be added to the next flour batch without any more yeast added, then how long would it take to rise? Restated, if the 8-hour poolish or sponge comprises 20% of the flour total, it’s brought to maturity or somewhat past it, and the other 80% of the autolysed dough is then added to the poolish without any additional yeast added, then how long will that rise take?