I am trying to make a veloute sauce which involves adding roux to boiling hot water mixed with chicken base (such as Tones or McCormick). This sauce is then mixed with heavy cream at a later stage to create a sauce for potatoes au gratin.

After you have made the roux and cooked it to a blonde stage (3-5 minutes), the chef that taught me this instructed me to cool the roux to room temperature and then add the room temperature roux to the boiling liquid stock. I let the stock boil for a few minutes after the addition of the roux and then reduce it to a softer boil (simmer).

The problem that I have encountered is that when simmering the sauce, the roux eventually breaks down releasing the fat (butter) into the sauce and thickening the "center" portion of the sauce. The fat essentially separates from the flour due to extreme heat (simmering). What causes this, and is there any way around it? Most veloute/supreme sauces call for a 30-50 minute simmering of the sauce to reduce any starchy taste and grainy texture.. but if the roux breaks down in the sauce before this how is this even possible?

If I don't leave the sauce at a boil initially for at least 3 minutes and then don't simmer for at least another 2-3 minutes, whatever remains from the lump of roux that was added settles to the bottom of the pan or forms lumps which eventually settle out. The sauce also has a "grainy" texture and seems to have too much of a starchy or floury taste.

Any advice on this would be appreciated. Thanks.

  • 1
    I have never simmered a roux based sauce for 30 minutes, much less 50, and never had trouble with starchy taste - this gets cooked out during the roux making, even for blondes. And a veloute should not have grains unless it broke during making, after which I doubt that simmering would help in any way. Have you tried just turning off the heat after the sauce has blubberred once?
    – rumtscho
    Commented Jun 8, 2013 at 17:15
  • Yes, if the sauce is taken off the heat too soon the roux will basically settle to the bottom of the pan and not be fully incorporated into the sauce. The sauce needs to be boiling for at least a few minutes. A lot of the veloute sauce recipes that I have seen all call for a simmer of at least 20 minutes. What causes the fat to separate from the roux? Shouldn't a properly made roux be stable during the simmering process?
    – Graupel
    Commented Jun 8, 2013 at 17:40
  • 1
    The way I make it, I mix the liquid with the roux and stir vigorously. It first breaks into grains, but by the time the starch gelates, the sauce is perfectly smooth. I leave it on high heat until I see bubbles, which takes 2-3 minutes (I use the hot/hot method) and then remove immediately. Except for a few early tries where I got clumping, I have never had separation issues.
    – rumtscho
    Commented Jun 8, 2013 at 17:48
  • I will note that when I did have issues with it breaking in the sauce, I was preparing the roux with the wrong proportions, tending to add more flour and making it into an extremely thick, dry paste. I have made it a little thinner now but haven't tested it yet to see how stable it is during a simmering sauce.
    – Graupel
    Commented Jun 8, 2013 at 17:57
  • McGee, p. 618, on why the long simmer: "The mixture is allowed to simmer for quite a while--two hours for a veloute... During this time [...], the starch granules dissolve and disperse among the gelatin molecules, with a very smooth texture the result." He comments that the long period is to ensure that no vestige of the granular structure remains (and so coagulated proteins can be skimmed). This is for the very finest, smoothest sauces. I imagine most home cooks would be fine bringing it to the simmer for a minute or two, as commented above.
    – SAJ14SAJ
    Commented Jun 16, 2013 at 16:42

4 Answers 4


Per SAJ14SAJ's comment, the suspension in a traditional veloute relies on the gelatin to keep everything in place. Stocks are naturally rich in gelatin, due to the bones and connective tissue used to make them. Using a base to make the sauce is probably not providing the gelatin necessary. Other rich liquids frequently do separate when simmered for a long time (think curries and gumbo), so it's to be expected that your emulsion will break down eventually. I think your solution is to either use a real stock instead of the base, or perhaps fortify your liquids with a bit of gelatin before adding the roux (or barring those, cook only until your starch is gelatinized).


The idea of letting the roux come to room temperature is based on the idea that the roux and the liquid should be at the same temperature. Instead, heat the liquid (chicken stock, broth, etc) to about the same temp as the cooked roux, and instead of adding the roux to the liquid, slowly add the hot liquid to the roux, whisking it until smooth.


Your sauce is breaking. Your basic veloute recipe calls for adding the stock to the roux, not the roux to the stock. If, however, you are dead-set on adding roux to stock, I would suggest slowly incorporating a small amount of the stock into the roux and whisking until it's smooth before adding that mixture to the rest of the stock (as if you were making gravy, for instance).

Also, I would humbly suggest that your stock is probably too hot. When flour-based sauces get too hot (in other words, boiling), they can break. Think like if you've ever tried to make gravy from drippings that were too hot, and your roux ends up all grainy. A liquid that has been boiled then brought down to a simmer will be closer to boiling that a liquid that has been brought from cold to simmer. I'd try the cold-simmer route and see if you have better results.

One last thing: are you using clarified butter for your roux? Sometimes the water in unclarified butter can cause the roux to break at higher temperatures.

  • It is very, very common to add roux (made ahead of time) to stock. Why would the order matter, exactly? Is there some science here? Same question on at least briefly boiling sauces which are flour based; normally this is required to eliminate the raw starchy taste for slurry/beurre manié thickened sauces--why would roux be different? Lastly, part of making roux is letting the water boil out of the butter. In any case, it is being added to a huge volume of water, so how on earth would using clarified butter make any difference at all?
    – SAJ14SAJ
    Commented Jun 16, 2013 at 5:17
  • Working backwards: the classic recipe for roux does call for clarified butter, because it doesn't bind together as well when the water is incorporated at the start. It makes a difference in this case, because it is the roux acting as the thickener, not just the flour and butter separately. Otherwise, you'd just add them to the stock and skip the extra step. On the boiling flour-based sauces bit, I am by no means an expert in every sauce out there, but as far as I know, those recipes call for constant stirring while boiling, thereby reducing the amount of heat concentrated in the sauce Commented Jun 16, 2013 at 16:25
  • 1
    (via heat distribution). You don't generally just let a flour-based sauce boil merrily away, both because they'll scorch and because you run the risk of breaking them. Even gravies that call for a high boil are usually corn starch-based. And lastly, the order matters because by adding the roux, a small amount of a thickened substance, to stock, a large amount of what is essentially water, you run the risk of making dumplings, basically. It takes a lot of patience to add a roux that hasn't been thinned to stock, whereas you can add a thinned roux much quicker, or the stock to the roux. Commented Jun 16, 2013 at 16:30
  • And how exactly does it not "bind together" with unclarified butter. I am afraid your answer and clarification simply don't make sense. There are three main points to roux: 1) it coats the flour particles so that they are less likely to lump in the sauce; 2) it cooks out the starchy flavor; 3) the heat breaks down some of the starch chains making it easier improving the thickening qualities (more than I can cover in a comment). McGee, On Food and Cooking, p 617. I am sorry, most of your answer and comments simply don't seem to have a basis. Downvoting pending citations.
    – SAJ14SAJ
    Commented Jun 16, 2013 at 16:32
  • That's fine if that's your opinion. Frankly, my responses generally come from years of experience combined with my interest in food science. I'm not a chef, I don't have a degree in the subject. What I do have is experience with very similar problems, and I offer the solutions that worked for me as well as what reasoning I could find at the time as to why they worked. For that reason, I couldn't begin to offer citations (which is why I'm not getting up in arms about you downvoting my answer). That said, I generally suggest at least trying an alternative before slamming it into the ground. Commented Jun 16, 2013 at 17:04

If your sauce breaks, too much water has been cooked off. Add water in small amounts while stirring until the sauce tightens back up.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.