In the recipe handed down through my family, the oil used to create the roux was the drippings from the bacon that would eventually be added into the pot of Gumbo. This important distinction added a more enhanced flavor to the dish, as many of the fatty acids in the bacon break down during cooking to introduce flavorful compounds found in the muscle, such as aldehydes, furans, and ketones. But is there any benefit to using butter rather than bacon fat?
To answer this question, we followed a recipe written by the staff at Masterclass, in which equal parts butter and flour were used. We decided to use 1 stick (1/2 cup) of salted butter and the same amount of flour. Around 5 minutes in, we noticed that it had began bubbling significantly more than the bacon fat roux, which could have been due to the fact that the amount bacon fat we had added was eyeballed, and thus was subject to a great amount of human error. It also has to do with the water that is included in the emulsified butter cooking away. After a few minutes of stirring, most of the bubbles dissipated and we were left with a somewhat smooth mixture.
The most noticeable difference, however, was the rate at which the butter roux was browning. After ~12 mins, the butter roux was darker than the bacon fat roux that had been cooking for over an hour! Because the water had almost entirely cooked off of the butter, the milk fats and proteins were left, which continued cooking. Because of this, the Maillard Reaction occurred much faster, as the chemical makeup of the different compounds in butter gave the reaction an opportune environment to occur. After 18 minutes, the roux was a rich, dark chocolate color, and was giving off a nutty, smoky aroma.
Believing we had completed our observations for this trial, Jordan and I began to wash the dish. What happened next was quite shocking and unexpected. As soon as we ran a small amount of water in the pan full of dark roux, it violently bubbled, darkened, and thickened, letting off a strong scent of burnt bread and emitting a sound similar to a blacksmith quenching metal in oil. We deduced that, since the roux was so hot and mainly consisted of hot oil, when we added the water, it quickly evaporated, releasing steam that was now less dense than the oil, and carried both the oil and water vapor into the air. This allowed the flour in the roux to burn.
Overall, our findings for this trial were that, even though the butter roux did not have the same flavor profile as the one made with butter fat, this roux is best for making a gumbo in a time crunch (18 mins vs. >1.5 hrs).