A Sauce Separation Problem

[project]

  I have a sauce that's not working. Wonder if anyone

  else has encountered this problem and what some good

  ideas are for fixing it.

 

  I'm using a white flour-butter roux to thicken a

  reduced braising liquid. For years, the proportions

  of flour, butter, and liquid have yielded a smooth

  flowing sauce for me, but here the sauce wants to

  separate.

 

  For some detail, I make a French style chicken

  stock, 2 'roasting' chickens, 6.5 pounds each, for 6

  quarts of stock, a stock that will start to gel when

  cold. I brown a 7 pound chicken and then braise it

  in 4 C of the chicken stock plus 4 C of dry white

  wine.

 

  Next I use the braising liquid to poach 3 pounds of

  vegetables -- onions, carrots, peeled celery (to be

  combined with the chicken, diced, and the sauce to

  make a casserole).

 

  I strain the braising liquid and reduce it to 1 1/2

  C at which time it will gel even when warm.

 

  I make a white roux of 10 T of flour and 8 T of

  butter, bubble for 30 seconds, and immediately add

  the simmering 1 1/2 C of braising liquid all at

  once.

 

  Ordinarily at this point, I get a sauce that is very

  thick but smooth and homogeneous; now I get a lot of

  inhomogeneity and some separation.

 

  I whip in 2 1/2 C of hot light cream. Ordinarily I

  get a nice smooth flowing sauce; now the sauce still

  wants to separate.

 

  The last trial, for a fix, I just whipped in 2 C

  more chicken stock and 1 1/2 C more light cream.

  Got a nice sauce but with less intense flavor.

 

  My guess is that the reduced braising liquid has too

  much gelatine and that is interfering with the

  action of the roux.

 

  So, could gelatine be the problem? Is this

  standard, that a reduced braising liquid with a lot

  of gelatine, enough to begin to gel while still

  warm, won't combine well with a roux?

 

  I'm guessing that maybe a fix is just to combine the

  cream with the roux, get essentially a 'Bechamel'

  (but I don't intend long simmering and reduction of

  it), and then add the reduced braising liquid to the

  Bechamel.

 

  Sounds like it might work, but it may be that, no

  matter how I change the order of events, in the end

  the cream will still result in too much butter fat

  present for the gelatine.

 

  I could be tempted to convert from cream to just

  milk, and that may get the butter fat low enough to

  solve the problems, especially if I make the

  'Bechamel' first and then add the reduced braising

  liquid to it. Using milk instead of cream seems to

  be giving up.

 

  Besides, some recent news reports from some recent

  research tell us a way to handle more butter and

  cream: Drink wine with the dish! Ah, goooood

  research!

 

  Maybe the fix is to use just 5 T of flour and 4 T of

  butter.

 

  It would seem that there should be a way for the

  gelatine to help the texture of the sauce, not hurt

  it.

 

  This sauce separation problem sounds like something

  moderately general. This problem was new to me and

  a surprise, but maybe readers on eGullet know this

  problem well and a solution.

[Suzanne F]

IMHO, it's not the gelatine -- if anything, that might act as a binder.

I don't see any mention of having removed the excess fat from the braising liquid. Did you forget to mention doing it, or did you not skim? That's the first thing that comes to mind: too much fat, which separates out. Then when you add the cream, you just exacerbate the problem. And any fat that leaches out of the chicken when it's heated in the sauce will just float on top.

Also, when you add the reduced braising liquid to the roux "all at once" do you just pour it all in and then stir, or do you whisk it as you're pouring?

Your final proportions of flour and butter to liquid should yield a somewhat thick velouté -- medium weight would be 2 T each flour and fat to 1 cup liquid; you've got 2 extra T flour for your 4 cups total liquid. I don't see a problem there.

Oh, please, PLEASE don't switch to milk. Yuck. Goodbye, flavor.

Finally: did you try burr-mixing the sauce with a stick blender after adding the cream? If so, did it stay together, or break later?

[Louisa Chu]

I don't know why if this worked for you before why it's not working now. But having said that usually the rule is cold roux/hot liquid or hot roux/cold liquid. Perhaps try that.

And I don't know if this will help in your case but a quick fix for a sauce separated from overheating is to re-bind it a little at a time with cold water. Take your sauce off the heat, tip the pot at about a 45 degree angle if possible, drizzle in a little cold water/liquid, whisk gently as it starts to bind, start pulling in the separated sauce into the fixed sauce, repeat as needed until completely reincorporated.

I think Suzanne F's pinpointed one of the common problems - too much fat - the other is overheating. Try addressing those two factors.

Good luck.

[schaem]

I rarely use roux, but I would try adding the roux to the sauce, not sauce to the roux. And do it little by little, as if you were mounting with butter. You can mount a lot of butter in a little sauce (gelatin not a factor) so you should be able to with roux (which must be even more stable than butter).

[Malawry]

Roux-thickened sauces need a little time to work, in my experience. By adding the liquid all at once to the roux, it's less likely to emulsify. When we learned roux-thickened sauces at school, we were taught to do as Loufood said and have either the roux or the liquid hot, and the other substance room temp. Then work the liquid into the roux slowly (not the other way around, which I think will create a lumpy sauce...though I admit I have worked very little with beurre manie, which is essentially raw roux and which is added to sauces later in their cooking to enrich and thicken).

When you add liquid to roux, start with just a little drizzle and stir until smooth over low heat. Keep adding more, stopping and stirring until smooth after each addition of liquid. Don't add more liquid until what you have in there is already emulsified. As the quantity of thickened sauce grows, you can add the liquid more quickly and also increase the heat level. Then whip in your cream and such as usual to finish the sauce. Sounds delicious.

I agree with Suzanne that gelatin is an unlikely culprit, and that defatting your sauce will help if you're not already doing so. But I think the biggest problem is adding all your liquid at once to the roux. Even if you're whisking as you do it, it seems like the most likely issue to me.

[ngatti]

FWIW, Sounds like you haven't stretched the gluten in the roux enough. In other words cook it and stir it a bit longer. If your chicken stock is reduced to as gelatinous a stage as you say then the additional cooking of the roux won't sully the color.

Instead of merely bubbling for 30 seconds. Vigorously cook and stir until you detect a faint odor of almonds. That's a blonde roux. Sounds like what you're currently using is essentially warm Beurre Manie. As Suzanne said, good for a thickening at the finish, but very likely to break if cooked for any length of time, and your description sounds like a broken sauce.

Cheers

Nick

[project]

  Suzanne F:

 

  Thanks for your:

 

  IMHO, it's not the gelatine -- if

  anything, that might act as a binder.

 

  I, too, was hoping that the gelatine would help

  the texture of the sauce, not hurt it.

 

  Thanks also for your:

 

  I don't see any mention of having removed

  the excess fat from the braising liquid.

  Did you forget to mention doing it, or did

  you not skim?

 

  Well, I did several drafts of that question,

  and some of them included lots of details but,

  alas, were too long.

 

  Yes, in this dish, I was concerned with both

  skimming for scum early in the simmering of the

  braising and skimming for fat in the braising

  liquid after the braising.

 

  I did try skimming, both times.

 

  The skimming didn't yield much either time.

  The early skimming was with a genuine skimming

  tool, 'spoon' with a very flat round bowl with

  many little holes, and the late skimming was

  with just a cooking spoon.

 

  A Fat Skimming Procedure: For the late

  skimming, I borrow a little from standard

  industrial separation techniques. With my

  crude pragmatic kitchen approximation, the

  procedure becomes: (1) From the braising

  liquid, take off a lot, to be sure to get

  essentially all of the fat, and don't mind

  getting maybe more of the water based liquid

  than the fat. So, now the original pot has no

  fat and have a small bowl with some water-based

  liquid but all the fat, and with the fat in a

  much higher concentration than in the original

  pot. (2) With this small bowl, skim getting

  essentially only fat. (3) For what is left in

  that small bowl, skim getting all the fat but

  likely including some of the water based

  liquid. Etc. Bowls with all fat, combine.

  Bowls with no fat, combine. The rest put in a

  third bowl. Continue until the volume in the

  third bowl is too small in total volume to

  worry with. Done.

 

  I did this procedure, and there was too little

  fat to be concerned with, in total.

 

  For the scum, the browning may have done enough

  to cook the proteins to reduce the scum and the

  simmering may have been gentle enough not to

  generate a lot of scum.

 

  The big chunk of fat in the tail of the chicken

  I did toss into the frying pan (actually used a

  steel wok) at the beginning of the browning.

  The browning was quite hot, the wok over

  170,000 BTU/hour burner, so that chunk of fat

  likely went to smoke.

 

  I did put the bones, skin, scraps in a pot and

  simmered for a day or two. That pot did have a

  fairly thick layer of fat, likely from the

  skin.

 

  For the gelatine, one guess is that it grabs

  the water and leaves too little for the roux to

  bind with. Just a guess.

 

  For your:

 

  And any fat that leaches out of the

  chicken when it's heated in the sauce will

  just float on top.

 

  Well, the chicken doesn't get into the sauce

  until 'final assembly'. By then, the full

  collection of chicken from the braising has been

  separated into bones, skin, scraps in one pot

  and nice chunks of just chicken, very low in

  fat, in another pot. Only the chicken in this

  second pot gets combined with the sauce. This

  chicken is so low in fat that it can't

  contribute much to a separation problem after

  final assembly. It may be that the sauce has

  enough salt that this chicken contributes some

  water-based liquid after final assembly, but

  with a little stirring that should join with

  the sauce well enough -- and that's my

  experience so far.

 

  Net, the separation problem is not any worse

  after final assembly.

 

  For your

 

  Also, when you add the reduced braising

  liquid to the roux "all at once" do you

  just pour it all in and then stir, or do

  you whisk it as you're pouring?

 

  Wow! I'm getting a serious response from a

  serious cook that's been there, done that,

  gotten the T-shirt!

 

  There is some huge collection of recommended

  procedures for adding a liquid to a roux. In

  my question I outlined the procedure that has

  worked for me for decades. I worked this out

  back in the 1960s from direct experiments just

  with the basic ingredients and not connected

  with anything to eat. Made a mess in the

  kitchen for days, found something that worked,

  made notes, and have used the results since.

  The procedure has always worked fine until this

  case with this braising liquid.

 

  Back in the 1960s, there was a fairly serious

  kitchen supply store on the east side of CT

  Avenue in DC -- and that likely uniquely

  identifies it. They sold some long wooden

  cooking 'spatulas' with nearly all convex

  surfaces. They are a lot like wooden cooking

  spoons except have no 'bowls', just an oval

  'blade' instead. Also, the quality of the wood

  is excellent, birch, maple, or some such.

  Actually, there were two woods available, and I

  have a supply of each. One of the woods is

  nearly furniture quality and I keep for folding

  egg whites, etc. The other has a little more

  grain and is better for sauce making. Naw, not

  for sale!

 

  So, the spatulas with the wood with more grain

  are excellent for stirring a roux because the

  wooden end of the oval blade does well wiping

  the roux off the bottom of the pot, keeping the

  flour and butter mixed, and not letting

  anything get too hot.

 

  So, I use one of these spatulas during the

  heating of the roux. I start by just putting

  the flour (room temperature) and the butter

  (refrigerator temperature) in a 2 quart (old)

  Farberware pot. I heat gently so that the

  butter does not separate and stir with the

  spatula. I get the liquid to be added --

  reduced braising liquid, milk, cream, whatever

  -- simmering. Then I turn up the heat on the

  roux, get it bubbling gently, look at the

  seconds counter on my watch, often lift the pot

  off the burner to regulate the heat, and bubble

  gently with constant rapid stirring for 30

  seconds when I'm in a hurry and 60 seconds

  otherwise. Right away (no delay -- have found

  that any delay here can hurt the action of the

  roux), I put the pot down off heat and dump in

  the liquid all at once. I just dump. All at

  once. Quickly I pick up the pot and the

  spatula, stir quickly, put the pot back on the

  burner, keep stirring quickly, wipe the spatula

  blade clean on the edge of the pot, put down

  the spatula, pick up a wire whip (of stainless

  steel, also bought in the 1960s), and whip

  violently.

 

  The usual result is a thick, homogeneous,

  glossy, slowly bubbling sauce -- VICTORY.

  Until now.

 

  For your

 

  Your final proportions of flour and butter

  to liquid should yield a somewhat thick

  velout‚ -- medium weight would be 2 T each

  flour and fat to 1 cup liquid; you've got

  2 extra T flour for your 4 cups total

  liquid. I don't see a problem there.

 

  Yes, you get the picture. It has worked well,

  for decades, back to 'Coquille St. Jacques

  Parisienne', crab meat moistened with Bechamel,

  wrapped in crepes, topped with such a sauce,

  various efforts at poaching frozen skinless

  boneless chicken breasts, 'gravy' for

  Thanksgiving turkey, sauce from braising goose,

  etc.

 

  Yes, you are correct about the 2 T of extra

  flour -- it is a little cheating, but it has

  worked well.

 

  For

 

  Finally: did you try burr-mixing the

  sauce with a stick blender after adding

  the cream? If so, did it stay together,

  or break later?

 

  Don't have a stick blender. I just use a wire

  whip, but it's a good one, and I have a strong

  arm and whip hard. Besides, I didn't really

  try later: I added stock and cream and got a

  thinner but reasonably stable sauce right away.

 

  Also, with just the 4 C of liquid, it really

  looks like it has no real hope of being stable.

 

  So, the curious point, and the base of the

  question, is that the difference appears to be

  the gelatine.

 

  For

 

  Oh, please, PLEASE don't switch to milk.

  Yuck. Goodbye, flavor.

 

  Exactly!

 

  loufood:

 

  Thanks for

 

  And I don't know if this will help in your case

  but a quick fix for a sauce separated from

  overheating is to re-bind it a little at a time

  with cold water. Take your sauce off the heat,

  tip the pot at about a 45 degree angle if

  possible, drizzle in a little cold

  water/liquid, whisk gently as it starts to

  bind, start pulling in the separated sauce into

  the fixed sauce, repeat as needed until

  completely reincorporated.

 

  I've done that with hollandaise. But for this

  sauce for chicken, the separation was not

  total.

 

  Supposedly the universe is sheets and strings

  of galaxies surrounding large voids. Well, the

  sauce was sheets and strings of liquid butter

  fat surrounding large globs of thick sauce.

 

  So, it was enough just to pour in some hot

  chicken stock (from the pot of 6 quarts) and

  pour in some more cream (in this case, still

  cool) and whip. Then the sauce came together

  and was reasonably stable. Overnight in the

  refrigerator, it gets nearly solid. Heated in

  a microwave oven, it starts to separate again,

  but a little stirring makes it homogeneous

  again.

 

  But, without the extra liquid (e.g., stock and

  cream, more than the 4 C I was planning), so

  far I don't see the sauce becoming homogeneous

  or stable. These proportions always worked

  well before. The only difference I can think

  of is the gelatine.

 

  Thanks.

 

  schaem:

 

  Thanks for

 

  I rarely use roux, but I would try adding

  the roux to the sauce, not sauce to the

  roux. And do it little by little, as if

  you were mounting with butter. You can

  mount a lot of butter in a little sauce

  (gelatin not a factor) so you should be

  able to with roux (which must be even more

  stable than butter).

 

  WOW! Actually, in part I am borrowing from

  page 384 of

 

  Jacques P‚pin, 'Jacques P‚pin's Complete

  Techniques', ISBN 1-57912-165-9, Black Dog &

  Leventhal Publishers, New York, 2001.

 

  where he does "Chicken Pie", and much like you

  are suggesting he adds a 'buerre mani‚'.

 

  For more detail: He tops with puff pastry, and

  I'm not. He starts with a 3 1/2 pound chicken,

  and mine is 7 pounds. He reduces the braising

  liquid to 1 C, and I reduced to 1 1/2 C,

  although for the first trial I braised only the

  breast, legs and wings and did not include the

  back, neck, heart, and gizzard which I have

  included since (and, thus gotten more flavor

  and gelatine). To the reduced braising liquid

  he adds a 'buerre mani‚' of 1 t of flour with 1

  t of butter. I didn't see the point of so

  little butter and flour and in my first trial

  used a roux of 2 T of flour with 2 T of butter.

  He adds heavy cream, and in my first trial so

  did I. In my first trial, I concluded that

  when the sauce was hot, it was too thin. In

  the refrigerator, cold, the sauce looks great;

  as it warms, it gets too thin for my tastes.

 

  Malawry:

 

  Thanks for your suggestions. You wrote:

 

  I agree with Suzanne that gelatin is an

  unlikely culprit, and that defatting your

  sauce will help if you're not already

  doing so. But I think the biggest problem

  is adding all your liquid at once to the

  roux. Even if you're whisking as you do

  it, it seems like the most likely issue to

  me.

 

  My roux-liquid combination technique has

  apparently been working fine for decades. But,

  even when the technique has appeared to work,

  maybe I've been too close to the edge and

  actually gotten sauces less stable than I

  should have. Now I'm over the edge.

 

  It is true that when I continued and added egg

  yolks and softened butter I got a less stable

  sauce, but I've always thought that this was to

  be accepted.

 

  I've been guessing that here something

  different has been going on.

 

  Maybe setting aside my personal roux-liquid

  technique and using a more traditional

  technique would make a stable combination.

 

  ngatti:

 

  You wrote:

 

  FWIW, Sounds like you haven't stretched

  the gluten in the roux enough. In other

  words cook it and stir it a bit longer.

  If your chicken stock is reduced to as

  gelatinous a stage as you say then the

  additional cooking of the roux won't sully

  the color.

 

  Instead of merely bubbling for 30 seconds.

  Vigorously cook and stir until you detect

  a faint odor of almonds. That's a blonde

  roux. Sounds like what you're currently

  using is essentially warm Beurre Manie.

  As Suzanne said, good for a thickening at

  the finish, but very likely to break if

  cooked for any length of time, and your

  description sounds like a broken sauce.

 

  WOW! Yes, you are correct; my 'roux' is just a

  'buerre mani‚' that has received what used to

  be commonly described as some 'pre-cooking' of

  the flour to reduce the taste of "raw flour".

  The idea of developing the gluten in a roux is

  new to me but sounds promising.

 

  You're right about the color: By the time the

  chicken is browned darkly in a steel wok over

  170,000 BTU/hour and braised and the braising

  liquid reduced, the liquid is quite dark, and

  nothing could hurt the color. A few drops of

  the reduced braising liquid that splatter

  quickly gel into at least a medium brown.

 

  Also, in the last trial, having run out of

  fresh parsley, I included 1/4 C of dry parsley

  flakes in the braising, and they mostly didn't

  want to strain out, made the reduced braising

  liquid still darker, and basically stayed in

  the final dish. Also, in one of the trials, I

  concluded that having ground black pepper

  present when the chicken is browned generates

  some interesting aromas, so have been adding

  about 2 T of such pepper to the chicken before

  browning, and that pepper, browned, makes the

  braising liquid darker, still. I am including

  the back of the chicken in the browning and the

  braising, and browning all that back skin

  should add a lot of browned flavor. The back

  should add some 'chicken broth' to the braising

  liquid. Similarly for the neck, heart, and

  gizzard. The back does contribute a little

  meat to the final dish but not much. And, I

  cut the wing at all its joints so that the wing

  skin can contact the wok surface better and get

  more brown.

 

  Also, somewhat related, after I boil the

  mushroom slices until they give up their water

  and boil the water away, I continue cooking the

  slices in a dry steel wok over high heat (the

  170,000 BTU/hour propane burner) until the

  surfaces of the mushrooms toast. Here are some

  more browned flavors.

 

  So, I'm going for some 'rustic, robust' flavors

  and not some delicate chicken version of a

  'Blanquette de Veau'. So, sure, cooking the

  roux until it starts to smell a little of

  toasted flour wouldn't hurt the color of this

  dish.

 

  These color and flavor points, then, partly

  explain why I'd rather stay with the cream, or

  even use P‚pin's heavy cream, and not go to

  milk: I'm guessing that the robust flavors of

  the braising liquid and then the delicate

  flavors of the cream create an interesting

  balanced combination, sort of the cream's white

  purity providing culinary 'retribution' for the

  dark sins of the browning!

 

  Actually, the robust flavors are not as

  overwhelming as might be expected: Including

  either parsnips or too many peas can dominate

  the dish. The cream can compete.

 

  Thanks for all the suggestions!

 

  Time to get another chicken, some more mushrooms,

  peas, parsley, thyme, butter, and cream, thaw out 4

  C of chicken stock, shovel the snow from the propane

  cooker, and try again!