Roux on the fly (split from San Francisco board)
Tip, in case useful:
All roux is based on starch cooked without water, which converts it to dextrin. Professionally, basic roux is sometimes made in bulk by _baking_ a fat-starch combination, but another commercial method, which I've found very useful at home, is simply to roast starch, dry (I've used both corn and potato) to the light tan color it acquires, let it cool, then soon (before it starts absorbing moisture from the air) transfer to a sealed jar. It lasts for years, unrefrigerated, if tightly closed, and you can add your favorite fat (duck, butter, etc.) before or even after combining it with a sauce. Half a pound of cornstarch spread out in a metal cake or pie pan and baked an hour or so in a (preheated!!) 400 deg F oven, occasionally stirring or tossing it, works.
A particular advantage of this is the not-always-publicized fact that some fats (duck and some others) decompose and smoke at temperatures below the 375 F or so needed for efficient dextrination. With the dextrin made in advance, you're free to use absolutely any fat you want. You can then cook it further for a dark (basically burnt) Louisiana roux, or cook the fat separately if you just want to make it smoke, or not.
(Sniffing a jar of finished dextrin reveals a familiar, slightly sweet smell that some people may have trouble placing. It's the smell of gummed envelope flaps, labels, and sheets of traditional postage stamps. All based on dextrin, cheapest and most common of the vegetable gums.)
Note by the way that the method I mentioned is used commercially and also in some home-cooking traditions. (After fiddling around with it myself, I encountered its use in the US Southeast, where probably most home roux in the country is made; starch is cooked in an open skillet on stove-top, and the resulting dextrin powder itself is informally called "roux.") I find the oven easier, the only attention it needs is occasional shaking or stirring, and inspection for color change, from white to light yellow or tan which will be already familiar if you cook much roux. Also, as the starch heats, it first expels its own moisture content (becoming clumpy during that phase) but then the clumps disppear and it becomes a fine powder, once the starch actually cooks into dextrin. The reason I emphasized preheating the oven is that gas ovens specifically, before heating fully, accumulate steam from the combustion, and water is your enemy in dextrin making. An oven preheated to about 400 to 415 degrees F seems to work well. Finally I was moved to experiment with this approach after a frustrating experience trying to make conventional roux with fresh duck fat for a gumbo recipe, without letting the fat break down and smoke too much; the temperature required to keep the fat from breaking down was low enough that the starch took forever to cook and change color (compared to roux made with butter or higher-temperature fats -- my experience is with French rather than Creole cookery). Fat's breakdown also accelerates rancidity, which may be the complaint Melanie Wong observed about a commercial jar roux in the SF Bay forum where this discussion started. When you make things yourself you control the quality, freshness, everything.
To your question, Cary, I gather you mean when using the finished dextrin in a sauce or recipe. You can use dextrin directly (I do) by making a slurry with water. Dextrin, unlike starch, is soluble, but it's a vegetable gum, and they behave kind of like their animal cousins, powdered eggwhite and dry gelatine. Gummy and uneven if quickly mixed with water, but if left in water and occasionally stirred they dissolve over a period of minutes.
When making a basic roux, where the dextrin combines with fat before anything else, this is not a problem because dextrin combines easily with fat, as cornstarch does. Only water leads to lumpiness or incomplete-solution issues. I have yet to try it with the Creole-style roux recipes that first cook flavoring vegetables or herbs in the fat before adding the starch, but as long as it's a basically fatty mixture without loose water, dextrin should mix easily into such a pan and anyway, after you eventually add liquid, any lumpy dextrin will soon dissolve.
While I use corn or other starches dissolved in water as a slurry for thickening where I want a particular appearance in a sauce, I never use it as a starch base combined with a fat for a roux. The definition of roux is specifically a mixture of wheat flour and fat, traditionally clarified butter, cooked slowly and used to thicken mixtures such as soups and sauces. Fats can vary; however, wheat flour in roux is a constant.
Flour can be browned in the same oven baking method described in you post for large quantities of roux, but a white or blond roux would not contain pre browned flour. Oven flour browning is an old restaurant chef's trick for making a large quantity of brown roux without much effort, by starting with browned flour. It should be noted that the darker the roux, the less ability it has to thicken, as carbohydrates break down when heat is applied, and there is a loss of material and particular properties in the original chemical structure.
Baking roux is also a easier, almost scorch proof method of making a brown roux, as the oven heat surrounding the sheet pan is gentler and more controllable than a more direct gas flame, and baked roux requires less stirring than the stove top method. Any type of fat, even one with a lower smoke point, can be used with the baking method. Duck fat, as well as chicken, goose, beef tallow and lard, have a smoke point of 375° or higher, significantly higher than clarified butter at 300°, and are all fine for brown roux and deep frying, given that most deep fried foods are cooked at 365° or lower. I would never consider making a Cajun style brown roux with cornstarch. If I did, I wouldn't tell anyone living in Louisiana about it. I think referring to a Louisiana roux as "basically burnt" might be an issue as well. Basically burnt it is not.
I don't understand what the point is to brown corn or potato starch when they are not a component of roux, and their applications rarely, if ever, call for dry cornstarch, for example, to be browned before being added to water. Beyond that, it's very rare that professional chefs make roux with cornstarch and clarified butter or another fat, although it's doable. Cornstarch is a more efficient thickener than flour; being low in or devoid of protein, starches can thicken with 50% greater efficiency than flour, cook out faster, but there's not many reasons to combine cornstarch with fat or to brown it in this manner. Wheat flour is a better choice for a fat based thickener and while pre-browning the flour is convenient, it is not necessary for achieveing a dark roux, regardless of fat choice.
I'm not sure how important is is to have a starch convert to dextrin when making a roux. Dextrin's use as a starch is important in brewing and in various commercially prepared food products, but gelatinization of the starch molecules by applying heat and the resultant thickening effect that gelling has on a sauce is the point of using a roux.
Actually bushwickgrl (and I'll follow up with original sources when I have more time), French authorities, including Escoffier, long have commended what they call "purer" starches as attractive alternatives to the wheat flour that is (as you mention) the most common and traditional form of starch (simply because it's the most familiar and common source of dry starch in European cooking). The argument for the alternatives is that wheat flour contains mostly the starches that you want (which become dextrin on anhydrous cooking, either in fat or in air), but also includes undesired nitrogenous material ("protein") which becomes some of the scum traditionally skimmed during the making of primary sauces such as Espagnole and demi-glace (#16 and 18 in the standard 5012-recipe _Guide Culinaire,_ Flammarion, 1921). I do know something of the culinary background of this subject, including that standard French repertoire does indeed normally cook the starch in fat (either stove-top or in oven -- I've done both many times these past 40 years), and I also know some of the food-science background. The real essence of the situation, which finds various expressions and rituals in different cooking customs, is dextrin, prepared in or out of the presence of fats. (Clarified butter is traditional in French roux-making, but not in some Creole and other variations. You are making dextrin when you cook any starch in fat, whether you know it or not.) You could, of course, use wheat flour in the same way I described, for convenience or dogma. But I took for granted (along with the background of traditional roux procedures) that not all cooks would wish to be constrained to wheat flour only, as that is far from necessary or universal.
I apologize if I did not explain all this context (and more) in what was originally an offhand alternative suggestion to a query for commercial jar roux. I took traditional roux definitions and rituals for granted, and "cut to the chase" about an interesting variation that , I assure you, works very well in practice. (Especially with poultry fats of low smoke point!! ;-) When I mentioned dextrin being prepared separately and used commercially, I meant in commercial processed foods, not restaurant kitchens. For example, shortcut glace-de-viande approximations or sauce concentrates, such as Aromont from France, often include dextrin as an ingredient, it adds a gelatinous character more economically (shall I say) than increasing the meat and bone content.
By the way I am not sure where you got those numbers for fat smoke points, but I urge you to take a look at Harold McGee's reference book which has a good summary, consistent both with all of my practical experience and with the details I gave above.
Here's some of where I learned points mentioned above. (This is all directly from authoritative books, I haven't looked up anything online.)
Escoffier on roux:
... It is also worth remembering in the study of the making of Roux that the starting point of the thickening of the sauce comes from the starch in the flour and it is on this that the balance of the sauce depends.
A Roux made from a pure starch such as arrowroot would give the same result as one made from flour, the only difference being that it is necessary to take into account the other substances contained in flour which would mean that a smaller amount of pure starch is required.
("Sauces" chapter, _Le Guide Culinaire,_ 4th ed. 1921, in the usual Cracknell-Kaufmann translation.)
Other French sources cite the different pure starches commonly used in different countries, and that in the US, cornstarch is common.
Fat smoke points: Harold McGee evidently isn't the source I remembered with the handy comparison of this data across animal and vegetable fats, but he reports clarified butter "can be heated to 400°F / 200°C before burning." [P. 37]
Enig's fats reference book's table of standard vegetable-oil smoke points shows olive oil smoking at 280°F / 138°C unless processed specifically for high-temperature uses. (Enig also shows how olive oil chemically resembles poultry fats, differing in proportions of the three main fatty acids. These proportions, and consequently also the melting and smoke points, fluctuate further from batch to batch and from bird to bird.)
"Burnt" roux: In late 1980s I exchanged info online (on what was long the sole public Internet cooking forum, still active) with someone claiming categorically that classic French repertoire cooking never allows roux "to brown." In reply I quoted the _Guide Culinaire_ (standard professional handbook of classic French repertoire) recipes 13 and 14 ("brown roux" and "blond roux"). A linguistically inclined friend added that "roux" literally means reddish-brown or russet. The original commentator then explained he had meant "brown" roux in the sense of Louisiana variants whose extra cooking can make them much darker, I believe he used the term burnt. That was my metaphor above.