Going Bananas: Sponge Cake Variant

 Baking Trials 025, 18 May 2013

After many fruitless forays into gluten-free Golden Sponge Cakes, I have decided to get on course with variations of wheaten formulas that are tried and true.

Thus I thought I would take on a Banana Bread variant to the Sponge Cake. Notice I do not say “Golden”, for after the addition of the banana and other ingredients the golden has given way to a lovely speckled brown.

Also here a gluten-free trial was bolted on, for I was curious about the latest GF mix to hit the market, the Cup4Cup brand: can it really deliver on its claimed/implied promise of successful 1:1 substitution for wheaten flour?

After some tweakage to the usual overly-moist, leaden-textured recipes that characterize classic banana breads, I tried the following:

Banana Pulp      250g (well-freckled, ripe)
 
 either:
A: Cake Flour    200g (wheaten variant)
 or:
B: All-Purp. GF  200g (gluten-free variant: "Cup4Cup" brand)

Raisin Slurry    150g (75g each water & raisins, pureéd)
Oil, Sunflower   109g
Egg_Whites        73g (from 2 eggs, separated)
Honey             60g
Sugar, superfine 120g (60g batter + 60g French meringue)
Egg_Yolks         37g (from 2 eggs, separated)
Baking Powder    9.2g (aluminum-free)
WPC-80           8.5g (whey protein concentrate, lecithinated)
Baking Soda      4.6g
Salt, Sea        3.6g
Vanilla Extract  4.2g (~1 tsp.)
Spice, Cinnamon  2.6g
Spice, Nutmeg    1.2g
Spice, Cloves    0.5g
Ascorbyl Palmitate 400mg (fat-soluble Vitamin C; antioxidant)
Lemon extract      3 drops
Vitamin-E Oil      3 drops (antioxidant)
 
Preparation was similar to prior chiffon-based batter methods (see other entries this weblog): beat egg yolks, prepare an emulsion by whipping in a slow drizzle of oil, mix in wets, stir in sifted drys (GF only: beat the batter well), and finally fold in French chiffon (egg whites to soft peak, then 1/2 sugar).

It was noted that the Cup4Cup GF variant was nasty to mix: the batter tended to “wind up” around the mixer blades, a behavior attributable to an excess of xanthan gum, and possibly pregelatinization of GF mix starch fractions.

Standard baking: 18min. at 320°F in canoe pans over boiling-water pan. Canoe cups prepped with shortening with nonstick overspray. Pan release is clean.

Analysis of ingredients and outcome provides these figures for this trial:

Figure     per 38.8g wheaten cake
Protein    1.9g
Fats       5.5g
Starches  10.7g
Fiber      0.7g
Sugars    11.2g
Sodium     290mg
Energy    136kCal


Yield for the wheaten formula is 22 cakes. Analysis for the Cup4Cup GF substitution is about 9% lower due to the batter having a lower density, yielding 24 cakes.
 
The wheaten variant has good texture, flavor, and aroma. The added spices, not typical in banana bread recipes, seem to compete with the banana aroma, but using dead-ripe bananas may help to offset.

Strangely the GF variant seemed to have less apparent banana aroma. Typically, but no different from all other GF mixes offered today and tried so far, cake rise was scary-dramatic, and postbake shrink remarkable and dreadful.
 
Cup4Cup-formulated sponge cakes were about 25% less in final volume, with a typically rubbery-gummy texture—not the worst results compared to that of other GF mixes, but still not good or satisfactory.

Cup-for-cup—whether as a baking approach or as a silly trademarked name—has proven once again to be a fantasy in naively converting traditional wheaten formulas to gluten-free.

—

Sugar composition, sugar-starch balance, and raisin inclusion have been adjusted within the limits of a workable formula to maximize shelf life vis-à-vis retardation of staling, retention of moisture, and resistance to microbial spoilage.

Sensory evaluation & comments from the Browne Crowe Bakes Sensory Evaluation Team:

• Not heavy, dense, or overly sweet as is usual with banana breads
• Good banana aroma; could be stronger?
• Wheaten variant (nice sponge) better than GF (a bit gummy)
• Tender & moist crumb

Overall the outcome of Trial 025 (wheat) is satisfactory.

The next trials contemplated will leave off xanthan- or guar-gum based GF flour substitutes, and go with a from-scratch rice-potato-tapioca base using milled flaxseed as a binding agent. Previous trials by the author (not yet reported here) of modified Yellow Cake GF formulations using flax have produced encouraging results.

Beekeeping for Poets

 “—” ‘—’’

Gluten-Free & Sorghum Notes: Two-week evaluation

In a pair of trials conducted two weeks ago, competing gluten-free flour replacers were evaluated. As noted here, and as many other GF experimenters and home bakers have discovered, GF baking-mix adaptations generally have problems approaching form, function, and palatability of wheat flour-based traditional recipes.

Trial 010B bake of 2013-02-25:
Namaste GF flour blend. Ingredients list from namastefoods.com:

• Sweet brown rice flour
• Tapioca flour
• Arrowroot flour
• Sorghum flour
• Xanthan gum

11 March 2013 14-day sampling: Cake is still moist, but crumb has become crumbly in the mouth, and there is definitely a slight bitterness or biting undertone that makes it a displeasure to eat. The remaining 010B shelf-life cakes have been tossed into the compost bin!

Trial 010A bake of 2013-02-25:
Beth’s All-Purpose GF Baking Flour. Ingredients list from glutino.com:

• White rice flour
• Potato starch
• Tapioca starch
• Guar gum
• Salt

11 March 2013 14-day sampling: Cake is still moist, harder than initially but still okay. Aroma and flavor good.

So what is all this Sorghum stuff, anyhow?

A common formula for sorghum-based GF cake flour replacer goes as follows, by percent weight, not volume. Note that this is for reference only!

I do not recommend baking with or eating this stuff:

• 62.4% Sorghum flour
• 29.4% Potato starch
•  6.8% Tapioca flour
•  1.4% Xanthan gum powder

Sorghum is a seed crop that was introduced to the New World in the 18th Century—generally presumed to have shadowed the traffic in African slaves. It is a drought-tolerant crop used for animal feed and by poverty-stricken folk who, due to hostile climate, poor soils, and long tradition, have nothing better to feed their beasts or themselves.

Sorghum has been increasingly-heavily promoted in the West as the latest wonder-grain for the gluten-intolerant, despite its relative indigestibility and significant issues with toxins (see this & this); ironically, cooking or fermenting sorghum only makes it less digestible and more toxic. In a 1984 study, male rats fed fermented sorghum meal developed anorexia, alopecia (hair loss), blood disorders, and testicular hypoplasia leading to sterility.

Maybe breeding and processing of sorghum have improved on this of late, but it is not hard to imagine that the aforementioned problems are merely being swept under the rug in the headlong ru$h to promote a politically-correct foodstuff that is bound to unwittingly create more health problems for world populations.

As for gluten-free NOT-winkies, there shall be no sorghum content, ever. IMO this crop cr@p lies firmly in the non-food category.

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 014, 6 March 2013

Continuing with the gluten-free trials, a couple more tweaks to the Spice Cake variant were tried. Here is the ingredients list for the least-unsuccessful:

AP Flour        130.5g (Glutino® Beth’s GF Pantry)
Sugar           120g
Milk, whole     108.8g
Egg whites (3)  103.4g
Egg yolks (3)    52.9g
Oil, sunflower   63g
Raisins          60g
Honey            45g
Mucilage         32.8g from 2.7g dry flaxseed; ~420mg dry matter
Flour, Rice RS   15g
Butter           12g
Baking powder     7.5g
Vanilla extract   4.7g
Spice, cinnamon   3.9g
Spice, nutmeg     2.3g
Spice, allspice   2.0g
Spice, cloves     1.7g
Spice, ginger     0.9g
Salt, sea         2.6g
Rice bran ext.    1.5g
Lecithin          2.4g (dry-powdered from Spanish sunflower)
Baking soda       2g
Ascorbyl Palmitate 400mg (fat-soluble Vitamin C; antioxidant)
Lemon extract    12 drops
Vitamin-E Oil     7 drops (antioxidant)


Analysis of ingredients and outcome provides these figures for this trial:

Figure     per 41.3g (average) cake
Protein    2.0g
Fats       6.1g
Starches   9.2g
Fiber      0.8g
Sugars    12.9g
Sodium      132mg
Energy    149kCal


Baking Loss was 7.1%; not bad. The 4 fluid-oz. portions rose remarkably high above the surface of the “canoe” pan, then subsided in the last two minutes of baking.

Excess batter went into the 4½" springforms at 12 fl. oz. portions each, and on a lark I tried upping the baking temperature to 350°F, which drove the baking time down to a brisk 15 min. from leisurely 23 min. at 325°F. Results were not the best: the finished 4½" cakes peaked and cracked, atypical in these trials. So unless the batter inclusions such as rice bran extract, flax mucilage, or milk were to blame, I suspect low & slow is the way to go.

There were several issues driving changes in the formula and process of Trial 008B. The most significant arose earlier today, this 10th day post-bake: the '008 cakes still had good appearance and flavor, but the crumb had become tough and lacking in cohesion—kinda day-old-muffin-y.
 
Another change was in moving from whey powder to whole milk. I desired the casein proteins in the whole milk for extra emulsification, and a little milk fat to add richness to the crumb; indeed, I traded around the fats by backing off the oil in the initial mayonnaise emulsion, and adding a pat of melted butter before mixing in the drys and meringue.

Still another change was adding mucilage from flaxseed: a marvelously slippery-slimy goop that evolves when flaxseed is immersed in water. This flax hydrocolloid has long been used by the allergy and vegan crowd as an egg-white replacer in some applications (though it won't whip or set like egg white). Instead of the slimy CMC (cellulose gum from wood pulp) and modified starches of the Twinkie for adding body and stabilization to batter, why not use natural slime? Time will tell what the effect will be.

Finally I decided to throw in a touch of commercial rice bran extract (see how it is made here), which is claimed to aid emulsification, make for a softer crumb, improve mouthfeel and moistness, and extend shelf life. Again, time will tell...

—

One gripe I have about GF mixes is that cakes seem to turn a bit rubbery and display a good bit of shrinkage as baking finished and as the cake cools. Clearly the guar gum, xanthan gum, and other wheat gluten replacers are doing the trick in holding up the sponge as it sets, but cannot really maintain the structure as water is lost and water vapor and leavening gases cool.

To be sure GF “flour” makers endlessly tweak and refine their concoctions of various starches (both native and pregelatinized) and faux-gluten gums, colloids, and other slimy muck; just okay so far, but man-o-man do they have a ways to go!

—

It takes about an hour to coax the mucilage slime from a cup of flaxseed: okay for R&D work, but not something I'd willingly pay a union worker too do, if a dime’s worth of powdered extract in a premix would serve instead.

So I e-chatted recently with a very nice and persuasive rep from a major food-ingredients supplier that offers a certain well-characterized flaxseed extract might do the trick. It stands to replace the bacterial barf xanthan gum and the fracking-fluid ingredient guar gum (also BTW. subject to wild price fluctuations) in baking formulations. What the heck, certainly worth a try!


Sensory evaluation & comments from the Browne Crowe Bakes Sensory Evaluation Team:

• Very rich taste, more of a creamed-butter cake impression
• Sweetness and spiciness similar to 008B
• All evaluation samples disappeared quickly
• Tender, moist, slightly chewy crumb

Overall the outcome of Trial 014 is satisfactory.

—

Creamed filling, you ask impatiently? Working on it. There are few prototypes to start with, and almost all of the usual cooked-flour & milk, marshmallow creme, and powdered-sugar based filling formulas available do not even come close.

I do have one candidate in the works, but it was mixed up just yesterday, and needs to “mature” in the fridge for a couple days before being whipped. It might serve as the matrix substrate for a variety of fillings, such as one suitable for a certain Spice Cake.

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 008B 22 February 2013

Going Gluten-Free

Since any Tom, Dick, or Mary can use the Same Old Formulas to make a wheaten cake, I decided to take a break from Luscious Creme Filling R&D and venture into the market segment of what till recently has been quaintly termed “dietetic” foods: you know, the tiny, oddball section in the grocer’s where bland specialty products sit, mostly ignored by all but those desperate for something semi-normal to eat that won’t make them stupid, swell up, or send them to the ER.

So: gluten free is actually on the verge of mainstream, normal, really, ever since gluten intolerance and wheat sensitivities have become better diagnosed, and more food options made available. And while we’re at it, why not address those who are lactose-intolerant by removing that troublesome, indigestible disaccharide from the formula?

AP Flour        87g (Glutino® Beth’s GF Pantry All-Purpose flour:
                     white rice flour
                     potato starch
                     tapioca starch
                     guar gum
                     salt) 
Egg, separated 102g (as 3 Large eggs)
Flour, Rice RS  10g
Honey           30g
Oil, canola     45g
Raisins         40g
Sugar           80g (superfine)
Water           59g
WPC-80           9g (lecithinated/instantized)
Baking powder    5g
Lemon Juice      6g
Baking soda      1g
Lecithin      0.67g (sunflower, hydrolyzed)
Allspice      0.50g
Cinnamon      0.67g
Cloves        0.25g
Nutmeg        0.50g
Vanilla ext.     1 tsp.
Ascorbyl Palmitate 200mg. (fat-soluble Vitamin C; antioxidant)
Vitamin E oil    6 drops (antioxidant)

The tricky thing in deriving this dual-requirement formula was maintaining functional and baking properties while jettisoning the lactose and gluten—and of course making something fit to eat. But since I have a Master Spreadsheet that allows me to adjust and maintain correct macronutrient proportions, by simply increasing the whey protein and tweaking the starch & sugary fractions I was able to keep the mix In The Zone.

Zeroing out the lactose also meant dumping the fresh acid whey drained from yogurt, substituting Whey Protein Concentrate (WPC 80% protein) that has had the lactose substantially reduced. Fortuitously I was able to locate locally-sourced, organic grassfed WPC for our formula trials.

For this initial trial, a retail GF flour premix was used, rather than try to tackle a formula from scratch. The Glutino® brand is well-known in the GF “community” for both finished products and for their line of baking mixes.

Long-term, we would have to locate a functional equivalent in 55-lb. sacks for production. One issue in particular is the global shortage and price increases of guar gum, a kinda-weird commodity derived from ground-up guar beans. It finds thickener-stabilizer applications in food processing, in pharmaceuticals, and, of all things, in fracking. Other galactomannans that might serve just as well in NOT-winkie production might be got from fenugreek seed, long a traditional food plant with a host of dietary-medicinal benefits.

—

Anyway,  mixing and baking the formula was straightforward. I borrowed an old trick from French genoise baking in pre-heating the whey (milk) proteins up to the critical temperature of 165°F—but no hotter—before incorporation into the mix. The resultant denaturation of proteins improves the emulsification of batters, and maintains cake-loft rise and stability in baking.

Prior experience suggested that GF formulas often make a heavier, denser product with more entrained moisture, so for this trial the base proportions were increased 50%, and baking times extended. The results were pretty good.

Spiced Fingers of Delight

Baking time for the canoe cakes was increased to 20min at 320°F. Instead of cupcakes for the excess batter, 4½" (11.4cm) springform pans were employed; baking for these mini-cakes required 23min. at 325°F. for a clean toothpick.

With a custom release coating of palm shortening, 7% sunflower lecithin, and 10% RS rice flour, release from the canoe pans was very clean and unremarkable, requiring little wiping or scrubbing for cleanup.

Test Batch 008B

The 4½" mini-cakes finished with a mostly non-domed top. During baking it was observed that the centermost zone of the batter stayed depressed and wet until the last 3 or 4 minutes of baking, filling out and puffing up level with the rest of the cake just at the end of baking.

The slight surface irregularities, which I think add “character” but would be cause for loathing and derision by other professional bakers, might be fixable with a bit more extensive mixing of meringue and flour phases before dispensing the batter. Maybe a smidgen of Sodium Stearoyl Lactylate would help...

Small-Cake Rise and Structure

For shelf-life testing, samples from the bake were placed in Pyrex® dishes and sealed with consumer-grade commercial shrink-wrap film, providing exposure to ambient diffuse light and shirtsleeve temperatures without unrealistic packaging means.

Shelf-Life Long-Term Testing

Each canoe cake weighs an average 38.5g—curiously, exactly the same as with the 005B formula; weird—and each 4½" mini-cake about 140g. Nutritional breakdown of each 008B canoe cake is:

Protein    2.7g
Fats         6g
Carbs       10g
Fiber      0.5g
Sugars      14g
Sodium     177mg
Energy     159kcal

Multiply these figures by 3 for each 4½" mini-cake. It should be noted that the Fiber fraction includes soluble and insoluble fiber from the raisins, as well as resistant starch from the rice flour.

With the particular WPC-80 used in the 008B formula, lactose content is only about 85mg per canoe cake. Since the USFDA has no fixed definition for “lactose-free” or even “lactose-reduced”, maybe comparing lactose content to regular and 70% lactose-reduced milk in groceries would be in order.

Whole milk contains about 4.8% lactose, so our 85mg per cake would be found in 1 2/3g milk or 1/3 tsp. fluid milk: not very much. And in the case of 70%-reduced milk, raise that to 1 tsp. more or less: still not a lot.

If 85mg./cake were deemed still too much, it would be little trouble to use Whey Protein Isolate (WPI), which contains about as much lactose as you’d inhale walking downwind of a cow. But WPI is much more costly than WPC...


The 008B cake is 100% wheat-free/gluten-free, and 99.7% lactose-free. There are only trace amounts of corn/maize (anticaking agent in the baking powder) and non-GMO soy (Solubilizer in organic WPC); with a little effort, even these could be purged.

—

Sensory evaluation & comments from the Browne Crowe Bakes Sensory Evaluation Team:

• Firm texture, not crumbly, gummy, or sticky
• Clean taste, not overly-sweet or cloying
• Satisfying and wholesome
• Attractive spice-cake appearance
• Distinct, aromatic spicy aroma & taste
• “Goes well with tea”
• Moist, slightly chewy crumb

Overall the outcome of Trial 008B is satisfactory.

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 005A & 005B 11 February 2013

 

I decided to try some formulations that omit the sunflower extract aka “goop”, for several reasons: 1) It is time- and labor-intensive to make; 2) Any unique functional advantages may be purely hypothetical; and 3) There are people out there who do have sensitivities, if not allergies, to seeds and nuts.
 
Base Formula:
Flour, cake     100 g
Flour, RS Rice   10 g
Baking Pdr.       5 g
Baking Soda       1 g
Sugar, gran.     85 g
Oil, canola      50 g
Lecithin, hydr.   2 g
Salt, kosher    1.7 g
Eggs, whites    75-80 g (from 2 large eggs)
Eggs, yolks     42-38 g (from 2 large eggs)
Whey, sour       51 g (from skim milk yogurt)
Vanilla ext.     5 ml
Lemon extract    8 drops
Ascorbyl Palmitate 200mg (fat-soluble vitamin C)
plus
005A variant:
Honey            52 g
or
005B variant:
Honey            26 g
Baking Soda     0.7 g additional
Raisins          36 g
Cinnamon, gr.   1.7 g
Nutmeg, gr.     1.1 g
Allspice, gr.   1.0 g
Cloves, gr.     0.3 g
Whey, sour      7.4 g additional to hydrate spices

Setup and make-up similar to that described in previous posts this weblog (see).

The Canoe Cakes were baked at 320°F for 18min. Cupcakes were baked in a separate oven, 13min. at 325°F, both with water pans.

Batter expansion, as before, showed good habit, rising above pan surface while retaining flat-topped shape integrity; no mounding or “muffin-top” overflow.

Vital to this desirable baking behavior as well as good pan release is a well-greased pan using palm shortening plus a generous application of lecithinated nonstick spray.

Batter and post-bake stats are:


Measurement             005A   005B
Batter per canoe-cake   34.8   38.5 g
Baked, per canoe-cake   32.0   35.6 g
Baking loss              8.2%   7.5%

These loss figures are rather lower than and therefore an improvement over prior trial losses: more moisture is being retained, and so less energy is misspent evaporating water from the batter.

This would make the NOT-winkie a kinda-sorta “green” product, just so long as it is understood that this refers to low-energy processing, not the color of the cake!

Trials 005A and 005B resulted in moist, supple cakes with good color and even open-cell surface structure

—

Sensory evaluation & comments from the Browne Crowe Bakes Sensory Evaluation Team:

• Attractive light-golden and darker spice-cake colors
• Sweet and fluffy
• Good cake structure & integrity; breaks easily in-hand
• Good aroma and flavor
• A fine box cake-like texture (but in the best meaning)
• Good spice balance — reminded one tester of Mom’s Spice Cake
• Moist and slightly chewy crumb

Overall the outcomes of Trials 005A & 005B are satisfactory.

—

We now move on to serious Luscious Creamed Filling R&D.

Beekeeping for Poets

 “—” ‘—’’

Filling the Void:

Musings on the Art and Science
of Creamy Filling

The now-defunct Hostess Suzy-Q snack cake was a Devil’s Food sandwich-slab with a filling that was probably substantially similar to that of the Twinkie. There are rumors of a pale, banana-flavored variant also being sold regionally, but I never encountered one in my many years.

There is one ingredient list for an ersatz Suzy-Q Creme Filling to be found on various baking sites, given below:

• 2 tablespoons all-purpose flour
• ½ cup milk
• ½ teaspoon vanilla extract
• 4 tablespoons salted butter, cold
• 4 tablespoons shortening, cold
• ½ cup granulated sugar

 
Whether this formula actually makes a good filling will be set aside for now. What is interesting is looking at the macronutrient breakdown of this mixture:

• Water       35%
• kCal         4 kCal/gram
• Protein      2%
• Fats        29%
• Misc. Carb   4%
• Sugars      31%

The breakdown of the butter-shortening fat mix is as follows:

• Saturated        38%
• Monounsaturated  31%
• Polyunsaturated  16%
• Trans             8%
• Trans-monoenoic   7%

The reason for all this up-close-and-personal on this filling is that I recently came across a 2008 patent that describes a stable whipped frosting/filling/topping concoction that has a composition suspiciously similar to that of the Suzy-Q filling. Its macronutrient balance goes:

• Water       39%
• kCal       3.9 kCal/gram
• Protein    0.3%
• Fats        30%
• Misc. Carb 0.4%
• Sugars      31%

It is hard to see how the owner of this patent could collect royalties since the cooking world has evidently long known of these proportions.

Note that due to various plant oils and shortenings being used instead of the Suzy-Q's butter-shortening 50:50 mix, the fat analysis of the patent formula is quite different:

• Saturated        86-88%
• Monounsaturated  12-10%
• Polyunsaturated   2%
• Trans             0%
• Trans-monoenoic   0%

The upshot is that in either case the formula specifies about equal thirds of water, sugar, and fats. Of course what makes these come together to whip up into a Luscious Creamed Filling (and stay that way) are the various extras—a pinch of this, a dash of that—that amount to typically 2-3% of the total.

Mixing It Up: Emulsifiers

One class of emulsifiers breaks up (destabilizes) the surface tension between the different fractions. About the only natural destabilizing emulsifier for the NOT-winkie would be lecithin.

Another class of emulsifiers stops recombination of (stabilizes) those same dispersed fractions. For us, the saturated glycerides naturally present in butter, coconut oil, and palm oil would be the most suitable and accessible.

Further natural, non-exotic magic ingredients can and should be included to make the filling a reality:

• Gelatinized Starch (as in the cooked flour of the Suzy-Q filling prototype), which acts as both destabilizer and stabilizer.
• Denatured Proteins, as in the cooked milk and flour, also help destabilize and stabilize.
• Salt reduces the influence of other soluble minerals, improving the properties and stability of the mix.
• Flavors, notably vanilla for a Golden Sponge Cake.

And for shelf-life, natural preservatives such as vitamin E (tocopherols) and vitamin C could and should make their way into the mix.

—

The Top Secret R&D Lab of Browne Crowe Bakes will shortly begin formula and process trials for Luscious Creme Filling.

On the baking Trial 004A and 004B (not previously reported), one week after baking the unfilled cakes are still quite good in flavor and texture. They have been stored unrefrigerated on a plate covered with a glass bowl—a rather severe test of moisture-retention!—and are just beginning to dry out and harden. No sign of spoilage, mold, or rancidity.

The next baking trial will adjust the balance of fats and proteins a bit. Stay tuned...

Beekeeping for Poets

 “—” ‘—’’

The Filling

The Luscious Creamed Filling inside the Golden Sponge Cake, far from being an incidental, nice-to-have bonus, is actually the Main Event of the Twinkie, and to a great extent of the NOT-winkie.

The Creamed Filling, too, is worthy of the Gods.

In Twinkie, Deconstructed, author Ettlinger claims the filling likely comprises more or less sugar, shortening, corn syrup, water, polysorbate 60, and salt, plus the key ingredient cellulose gum.

As described earlier, cellulose gum, aka sodium carboxymethyl cellulose (CMC), is indigestible synthetic stuff derived from wood pulp. Say what you will about the utility of CMC in baked-goods fillings, but the bottom line is that it has no drop-in natural equivalents or replacements, and therefore no relevance to the NOT-winkie.

Sleuthing Again For Fillings

Nosing around for information of and ingredients in commercial snack-cake fillings, the following ingredient list for White Creamed Filling popped up:

• Sugar
• Vegetable shortening:
•   partially hydrogenated soybean and palm oils
•   mono and diglycerides
•   polysorbate 60
• Water
• Corn syrup
• Corn starch
• Modified food starch
• Emulsifier:
•   mono and diglycerides
•   TBHQ (antioxidant)
•   citric acid (antioxidant)
• Salt
• Natural and artifical flavor (including milk derivatives)
• Sorbic acid
• Even more Polysorbate 60

Anybody who has already perused the original Twinkies Ingredients List will recognize most of these ingredients as Old Friends. What is not so obvious is the traditional-recipe basis for these synthetics and additives.

What I’ll call formula Type 4 of commonly-cited home-made or small-bakery Twinkie-like fillings includes the following recognizable foods (additives and their functions):

• Wheat Flour (starch, modified starch: thickening)
• Milk, fluid (water, flavor, emulsifiers: moisture & mixing)
• Vanilla extract (flavor)
• Butter or Margarine (fat, flavor, emulsifiers: fats)
• Shortening (fat, emulsifiers: fats)
• Sugar (sugar, corn syrup: sweet!)
• Salt (salt: flavor)

In the Type 4 process, the flour and milk are cooked until a starchy paste is formed, whereupon the vanilla is added. Separately the sugar, butter/margarine, shortening, and salt are whipped, the paste added, and the final mixture whipped some more.

It should be noted that there are numerous patents pertaining to the use of starch gels as fat-replacers in baked goods. It seems ironic that much of this so-called intellectual property could be invalidated in patent courts, due to the prior traditional use of wet-cooked (gelatinized) flours and starches in fillings etc.

Filling Functional Criteria

A successful Golden Sponge Cake filling formula will lie within the all-natural limitations of the NOT-winkie Project, while striving for that je ne sais quoi of the authentic Twinkies’ Luscious Creamed Filling.

• Excellent taste, texture, and aroma
• Right balance of sweet and salt
• Attractive color
• Not custard-like, heavy, or frothy
• Lack of gummy or greasy mouthfeel
• Stable over reasonable temperatures, freezing & thawing
• Long life in-cake without going rancid or off-flavor
• Not especially supportive of microbial growth
• Will not migrate into cake
• Will not gain or lose moisture
• Will not break down, separate, curdle, or harden

So it is just a matter now of admitting je ne sais pas exactement comment, getting over it, and mucking about until it comes out right. No worries, then!

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 004A 04 February 2013

 

It has been said that Luck favors the prepared and persistent.

This latest formulation is pretty close to good, even though it does not fully address shelf-life issues, and still wants Luscious Creamed Filling:

Flour, cake     100 g
Flour, RS Rice   10 g
Baking Pdr.       5 g
Baking Soda       1 g
Sugar, gran.     85 g
Oil, canola      36 g
Lecithin, hydr.   ¼ tsp (sunflower, Spanish non-GMO origin)
Salt, kosher    1.7 g
Eggs, whites     72 g (from 2 large eggs)
Eggs, yolks      31 g (from 2 large eggs)
Whey, sour       52 g (from skim milk yogurt)
Honey            52 g
Vanilla ext.     5 ml
Lemon extract    8 drops
Extract, Sunflower kernel +20 g (from 1.0M extraction)

Mise-en-place setup as typical chiffon-style:

1. Combine whey, honey, soda, vanilla, lemon, sunflower extract, lecithin, blend until uniform.
2. Sift together cake flour, baking powder, ½ the sugar, and salt.Measure RS rice flour, oil, and set by batter bowl.
3. Separate eggs, whites into meringue bowl with remaining ½ the sugar on the side; yolks into primary mixing bowl.
4. Preheat oven to 320°F; place pan of boiling water in rack under main rack.
5. Baking pan(s): grease plus nonstick oil-lecithin spray.

Make-up as follows:

1. Whisk egg yolks until ribbon stage;
2. Gradually drizzle oil into yolks while whisking, making emulsion.
3. Drizzle liquids mixture into yolk emulsion, whisk until uniform.
4. Disengage beaters and manually sift in drys: first RS rice flour, then wheat flour mixture, mixing gently until fully moistened and uniform.
5. With cleaned utensils, whip the egg whites until soft-peak is reached, then sift in sugar and whip until a medium-peak French meringue is formed.
6. Lighten flour-yolk batter with ¼ the meringue, then pour this over the remaining meringue and fold up & over until uniform.

Batter portioned into prepared canoe pan using a ¼-cup measure to fill the cups halfway. Excess batter divided into six cups of a standard cupcake pan (prepped as with canoe pan).

 

Standard Pan Test Load: 8x Canoes + 6x Cupcakes

The canoe portion was baked at 320°F for 19+min. Cupcakes were baked in a separate oven, 14+min. at 325°F, also with water pan.

Batter expansion showed good habit, rising above pan surface while retaining flat-topped shape integrity; no “muffin-top” overflow. This suggests good moisture balance and a rise that does not cling unduly to the cup sides.

Perfect Rising Shape

Cooled cakes showed no signs of collapse.

Post-bake settling

Depanning was by inversion onto a parchment-covered surface. Release was quite clean this time.

Clean cake release

Trial 004A resulted in a moist, tender cake with good color and even open-cell surface structure. Canoe cake average weight about 28.6g, remarkably similar to earlier trials.

Perfect Shape & Color

Sensory evaluation & comments:

• Attractive golden color
• Sweet and fluffy
• Good aroma and flavor
• Good manual breakage and crumb integrity
• Open-cell interior sponge ready for filing injection!

Overall the outcome of Trial 004A was satisfactory.

—

On the seventh day after baking, the last of the 002 cakes remained fresh-like and good, though they never were really great.

On the fourth day post-bake, the 003 cakes, never very good to start, are no fun at all; I have tossed the remaining ones. Besides being off-flavor, they had become quite gooey on the top (flat) surface, suggesting that they were underbaked in the first place, and possibly made worse by an excess of hygroscopic fructose from the honey.

—

Further cake-baking trials can wait while I ponder the issue of Luscious Creamed Filling. There are quite a few recipes around, generally falling into four classes:

1. Whipped-fats-and-sugars: some variants hereof are rated highly for their resemblance to Twinkies filling; sometimes have milk added.
2. Marshmallow-fluff: containing gelatin and sugar or corn syrups.
3. Cheese or cheese-like: featuring mascarpone or cream cheese, or (*gag*) tofu or imitation soy-based faux-fromage.
4. Boiled-starch-and-milk: these rely on forming an emulsion or suspension of fats within a gelled starch matrix.

I tend to favor the 4th category, thinking that the pre-gelled starch-and-milk cookup would be sterile (leg up on extended shelf life), and that such partial fat replacement would have “organoleptic” (yumminess & “mouthfeel”) and nutritional advantages. Progress on fillings R&D will be reported here...

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 003 03 February 2013

Pushing the boundaries of the Sweet Spot in cake formulations, I went to the brink and promptly fell over. The result was Baking Trial Formulation 003, a failure by all accounts. Shows what comes from making too many changes at once.

It is recommended that nobody else try this because it is rather awful:

Flour, cake      95 g
Flour, RS Rice   15 g
Baking Pdr.       5 g
Baking Soda       - g
Sugar, gran.     82 g
Oil, canola      35 g, melted with
Short., palm    3.6 g (1 tsp.)
Salt, kosher    1.7 g
Eggs, whites     70 g (from 2 large eggs)
Eggs, yolks      38 g (from 2 large eggs)
Whey, sour       52 g (drained from skim milk yogurt)
Honey            22 g
Raisin purée     36 g
Vanilla ext.     5 ml
Lemon extract    8 drops
Extract, Sunflower kernel +25 g (from 1.0M extraction)

Mise-en-place setup was modified chiffon-style:

1. Grind raisins, whey, and honey in blender. Filter, and combine with vanilla, lemon, sunflower extract.
2. Sift together cake flour, RS rice flour, baking powder, 40g. of the sugar, and salt.
3. Separate eggs, whites into meringue bowl with remaining 42g. sugar on the side; yolks into primary mixing bowl.
4. Melt oil with shortening, let cool on the side of egg white.
5. Preheat oven to 320°F; place pan of boiling water in rack under main rack.
6. Baking pans ungreased.

Make-up as follows:

1. Whisk egg yolks until ribbon stage;
2. Gradually drizzle oil-shortening into yolks while whisking, making emulsion.
3. Drizzle liquids mixture into yolk-sunflower mix, whisk until uniform.
4. Disengage beaters and manually sift in flour mixture, mixing gently until fully moistened and uniform.
5. With cleaned utensils, whip the egg whites until soft-peak is reached, then sift in sugar and whip until a medium-peak French meringue is formed.
6. Lighten flour-egg batter with ¼ the meringue, then fold this into the remaining meringue until uniform.

The batter was portioned into the pan using a ¼-cup measure to fill the cups halfway. Each pan portion weighed about 34.7g. Excess batter was divided into six cups of a standard cupcake pan, two paper-lined.

The Canoe shape was baked at 320°F for 18min. Cupcake shapes were baked in a separate oven, 13 min. at 325°F, also with water pan.

Depanning was with an inverted slap against a sturdy wooden surface.

Canoe cake average weight was about 28.6g, for a baking loss of 18%: still a little high. Noticeable decrease in rise and final cake volume.

Trial 003 resulted in a moist, somewhat greyish cake with open-cell surface structure.

Sensory evaluation:

• Fragile crumb, falls apart in mouth
• Sweet, but not too too sweet
• Noticeably heavy, chewy, and gummy
• Weird smell; unexpected
• Uneven sponge cell structure

Overall the consensus was Not Good. It is likely the un-neutralized acid of the raisins adversely affected the flavor. Excess RS rice flour and/or sunflower extract, and probably underbaking conspired to make a rather-less-than-fun experience.

Adjustment for next trial:

• Revert to 002 formula and technique as new baseline.
• Increase canoe baking time to 19-20min.

—

Happily, five days after baking, the 002 cakes remain fresh-like and quite yummy, longing though they are for a Luscious Creamed Filling to complete their existence.

More trial results coming up...

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 002 30 January 2013

Analysis of the sensory defects of Trial 001 proved most insightful. Feedback informed some alterations to the initial formula, and the results are much improved!

The Browne Crowe Bakes Sensory Evaluation Team took another bite of the now-24 hour old Trial 001 cakes, which were sealed in a locking food-storage container over the past day. Regrettably, and unlike fine wines, the NOT-winkie sans remplissage rev 001 has not improved with age, so the latest sampling compels the following to be added to the sensory-defect list:

• gummy
• heavy
• squishy 

The ingredient proportions have been modified as listed, and the batter make-up has been tweaked somewhat. While better, it still falls short and must be adjusted further to address some issues listed later:

Flour, cake     100 g
Flour, RS Rice   10 g
Baking Pdr.       5 g
Baking Soda       1 g
Sugar, gran.     85 g
Oil, canola      36 g
Salt, kosher    1.7 g
Eggs, whites     72 g (from 2 large eggs)
Eggs, yolks      36 g (from 2 large eggs)
Whey, sour       52 g (drained from skim milk yogurt)
Honey            52 g
Vanilla ext.     5 ml

Lemon extract    8 drops
Extract, Sunflower kernel +20 g (from 0.5M extraction)

Mise-en-place setup was modified chiffon-style:

1. Stir together in plastic cup whey, honey, soda, RS rice flour, vanilla, lemon. Stir and let sit for 1hr., stirring occasionally.
2. Sift together cake flour, baking powder, 40g. of the sugar, and salt.
3. Separate eggs, whites into meringue bowl with remaining 45g. sugar on the side; yolks into primary mixing bowl with sunflower extract; oil on the side.
4. Preheat oven to 320°F; place pan of boiling water in rack under main rack.
5. Baking pans ungreased.

Make-up as follows:

1. Whisk egg yolks & sunflower extract; do not expect classic ribbon consistency.
2. Gradually drizzle oil into yolks while whisking, forming a loose emulsion.
3. Drizzle liquid mixture into yolk-sunflower mix, whisk until uniform.
4. Reduce mixer speed and sift in flour mixture, beating until fully moistened and uniform.
5. With cleaned utensils, whip the egg whites until soft-peak is reached, then sift in sugar and whip until a medium-peak French meringue is formed.
6. Lighten flour-egg batter with ¼ the meringue, then fold this into the remaining meringue until uniform.

Egg yolks plus sunflower extract

Lightening the batter with meringue portion

The batter was portioned into the pan using a ¼-cup measure to fill the cups halfway. Each pan portion weighed about 34.4g. Excess batter was divided into five cups of a standard cupcake pan, unlined.

Filled pan cups

The Canoe shape was baked at 320°F, and first tested at 15min. The center four cakes were still undone, so baking continued to 18min., whereupon the toothpick test passed. Cupcake shapes were baked in a separate oven, 12½ min. at 325°F, without water pan.

Cakes at half-done point

Finished baking

The cakes were allowed to cool somewhat, contracting slightly and pulling away from the cup walls a bit.

Depanned cakes, looking very familiar...

Depanning was with an inverted, vigorous slap against a sturdy wooden surface. Due to no release coating having been applied to the pan cups, it took several slaps to depan all cakes.

Cake pan-skin stuck to the cup - quite a cleanup!

Cake average weight was about 27.3g, for a baking loss of  1-(27.3/34.4)=21%: still a little high, but not overbaked this time.

Trial 002 resulted in a moist, somewhat supple cake with the highly desired Twinkie-like external open-cell surface structure, in contrast to the greasy-floury puckered elephant-skin of Trial 001.

Sensory evaluation:

• Sweet, but not too too sweet or cloying; more dessert-like
• Slightly sticky (but so are Twinkies!)
• Squishy, but in a good way
• Puffy texture
• Taste is “way better”
• The note of lemon is a pleasant surprise
• Better sponge fineness and uniformity
• Cupcakes were overbaked with a slight egginess

It would seem that presoaking the RS rice flour eliminated the gritty after-sensation, and that neutralizing the acidic honey and whey with the soda is best done during the presoak, rather than adding the dry soda to the flour mixture.

Even though the egg yolks being whipped with the sunflower extract “ruined” the expected ribbon stage of yolk whipping, that last canonical step’s disappearance did not seem to make much difference to the end result, much less spoil the batter.

Adjustment for next trial:

• Research in-batter additions or ratio tweaks for aiding release from pan; chiffons must cling to pan for gas-sponge support and for resisting shrinkage while cooling, but release is a insufficient for production.
• Ensure that eggs are the freshest possible. One member of The Browne Crowe Bakes Sensory Evaluation Team is especially sensitive to egginess.
• Reduce cupcake baking temperature a bit, and add a hot-water pan to the second oven.

More trial results to follow...

Beekeeping for Poets

 “—” ‘—’’

Baking Trial 001 29 January 2013

First trial! An initial formula was scaled to provide cakes sufficient for a week’s worth of shelf-life testing—assuming the result would be edible in the first place!

The basic shape of the NOT-winkie until further notice will result from the Norpro pan as shown below. For the home and blogger-experimenter this will have to do, even though it has only eight cups and a dark, nonstick coating that interferes with the baking and release processes.

Norpro 3964 “Cream Canoe” pan

Now I’ll list the ingredients below with the understanding that the result falls far short of Golden Sponge Cake perfection. I do not recommend using it:

Flour, cake     100 g
Flour, RS Rice   10 g
Baking Pdr.       5 g
Baking Soda       1 g
Sugar, gran.     85 g
Oil, canola      60 g
Salt, kosher    1.7 g
Eggs, whites     75 g (from 2 large eggs)
Eggs, yolks      34 g (from 2 large eggs) 
Whey, sour       52 g (fresh-drained from skim milk yogurt)
Honey            52 g
Vanilla ext.     5 ml
Extract, Sunflower kernel +10 g

Mise-en-place setup for the batter make-up was basic chiffon-style:

1. Sift together flours, leavenings, ½ the granular sugar, and salt.
2. Separate eggs, whites into meringue bowl with other ½ sugar on the side; yolks into primary mixing bowl.
3. Mix liquids: whey, vanilla, honey, and sunflower extract.
4. Grease (palm shortening) and flour Canoe pan cups (mistake; see below)
5. Preheat oven to 320°F; place pan of boiling water in rack under main rack.

Make-up as follows:

1. Whisk egg yolks until ribbon stage.
2. Gradually drizzle oil into yolks while whisking, forming a mayonnaise-like emulsion.
3. Drizzle liquid mixture into yolk emulsion, whisk until uniform.
4. Reduce mixer speed and sift in flour mixture, beating until fully moistened and uniform.
5. With cleaned utensils, whip the egg whites until soft-peak is reached, then sift in sugar and whip until a medium-peak French meringue is formed.
6. Lighten flour-egg batter with ¼ the meringue, then fold this into the remaining meringue until uniform.

The batter was then portioned into the pan using a ¼-cup measure to fill the cups halfway. Each pan portion weighed about 37.7g. Excess batter was divided into six cups of a standard cupcake pan lined with papers.

Prepped pan and batter

Baking for the Canoe shape was 20min. at 320°F. Cupcake shapes were baked in a separate oven, 15min. at 320°F, without water pan.

First batch looking pretty plump

Remove respective pans after passing toothpick tests, let cool with all-around air circulation.

Pan of canoes post-bake

Depan canoes while still warm by inverting with smart drop onto cutting board.

Torpedoes, Trial 001

Cake average weight was 30.3g, for a baking loss of  1-(30.3/37.7)=20%: a little high, probably overbaked.


Sensory evaluation:

• Too sweet
• Slight tongue-coating greasiness
• Chalky-gritty residue left on teeth
• Good texture
• Sponge cells a little uneven
• Good vanilla flavor
• Notable lack of egg taste or aroma
• Cake skin wrinkly-puckered

Resolution and adjustment for next trial:

• Reduce liquid oil content
• Pre-hydrate RS rice flour
• Blend meringue into egg-flour batter a little longer
• Reduce sugar in meringue; try cream of tartar as foam stabilizer
• Mix soda with whey-honey (neutralize) rather than add to flours etc.
• Do not grease/flour pan; try alternate release means.
• Try hydrolyzed sunflower lecithin
• Add a drop or two of lemon oil
• Test for doneness at 17min.

The Browne Crowe Bakes Sensory Evaluation Team will continue evaluating Trial 001 cakes over the next week, to check for shelf-life related issues. 

So what is this Sunflower Kernel Extract Stuff, Anyhow?

Well,  sunflower kernels contain a lot of lecithin, some good oils, proteins, and a good bit of vitamin E, all of which should have some utility in NOT-winkies.

Grains & seeds being tried for additives

While remaining a little vague about exactly the processes used, I will say that after studying and discarding various patent processes for isolating this or extracting that, I have settled on a simple kitchen-counter method using ordinary ingredients and household appliances.

These suffice to make a smooth “cream” that is yummy and inoffensive for inclusion into NOT-winkies. It is probably also deliriously nutritive, but again lacking a food analysis lab, I’ll have to rely on the USDA NAL Database to guess what goodness resides in the NOT-winkie.

Mysterious extracts of sunflower kernels

Trial 002 due to start soon...

Beekeeping for Poets

 “—” ‘—’’

Canonical Cakes

Things are heating up in the Top Secret R&D Lab of Browne Crowe Bakes. A special resistant-starch preparation of germ-retained rice has been turned into flour, ready for baking trials. As I type I have test quantities of sunflower kernels undergoing alchemical transformation into various emulsifier, protein, and oil preparations, with synergistic combinations possibly workable into the pre-batter phase of a chiffon-like formula.

To this end, one might wonder what the starting point must be in making any cake, especially the Golden Sponge Cake component of the NOT-winkie:

The Fundamentals of Formulæ

Professional bakers have several rules of thumb for “balancing” the proportions of cake ingredients to get the mix In The Zone of desired cake properties. The three general guidelines of cake formulation go as follows—quantities refer to weight, not volume measure:

• Sugar should equal or exceed flour (sugar>flour is called “high ratio”)
• Eggs should equal or exceed fats
• Eggs plus other liquids should equal or exceed sugar

For most published cake formulas, these rules seem to hold up pretty well. As demonstrated in previous weblog posts here (see), no useful information or conclusions can be got about the actual amounts of eggs, flour etc. in Twinkies.

What is more interesting and relevant to anybody with a college education is to look at the breakdown of specific food components—starches, fats, protein etc.—such that batter formulas can be better “tuned” to meet specific functional goals, e.g. the shelf-life of our NOT-winkies.

Analyzing the macronutrient breakdown of eight cakes in the same class as Golden Sponge Cake, and contrasting again with the bare-bones nutritional labeling of the archetypal Twinkie, we find these figures emerging (the ± is attached to the standard deviation, for all you engineers, boffins, and statistics wonks):

Component    Classic Cake    Twinkie
Sugars        26.3% ±3.2%      42%
Starches      19.1% ±2.4%      21%
Protein        4.9% ±0.7%       2%
Water         33.6% ±5.8%      24% (guess)
Fats          15.5% ±4.4%      11%

Digging deeper we can examine the ratios between macronutrient components:

Component Ratio    Classic Cake    Twinkie
Starches:Protein     4.0 ±0.6        9.0
Fats:Protein         3.2 ±0.8        4.5
Fats:Starches        0.8 ±0.3        0.5
Sugars:Starches      1.4 ±0.2        2.0
Sugars:Protein       5.5 ±0.8       18.0
Water:Sugars         1.3 ±0.4        0.6
Water:Starches       1.8 ±0.4        1.1

The notable deviations of the Twinkie well outside “natural” macronutrient measures, in both percentage content and content ratios, are no doubt made possible by all those lovely additives. I seriously doubt anyone could make a scratch-cake using the Twinkie proportions without lots of help from the Better Living Through Chemistry crowd.

Sweet Secrets to Shelf Life

As it turns out, one ratio especially relevant for shelf life is the sugar:starch figure. This should be in the range 1.5 to 3.1, so you can see right away that whereas most classic cakes sit on the low end at 1.4, the Twinkie is right there at a solid 2.0.

In the case of the Twinkie, but irrelevant to the NOT-winkie, is that underlying the rich sugar:starch ratio are the aggressive nonfood shelf-life extenders: caseinates, dextrin, HFCS, MGL & DGL, Polysorbate 60, SSL, and soy flour.


Another little card the NOT-winkie has up its sleeve comes from the old-European tradition of Honey Cakes, well-known for their staying moist and good to eat for many days if not weeks after baking. These were devised in a era when cane or beet sugar was scarce, preservatives as we know them were non-existent, and shelf-life was not a dirty word.

The trick is to replace 100% sugar with a sugar mix having a disaccharide : monosaccharide ratio of about 2:1. More likely than not, this is what Hostess did with Twinkies—although the exact proportions are unknown—where the disaccharides comprise sugar and whey lactose, and the monosaccharides glucose and fructose (lævulose) are supplied by added dextrose and HFCS. Recall that fructose is humectant, that is, it draws and holds water molecules aka moisture.

For the NOT-winkie, our natural equivalent would be to substitute, say, 1 cup (225g) of sugar with 2/3 cup sugar (150g) plus 75g÷82% sugars÷1.425g/ml = 64.2ml (1/4 cup plus 1 tsp.) honey, adjusting the formula to account for the 3 1/3 tsp. water present in the honey. Courtesy of floral sources and honey bees’ industry, honey sugars are about half fructose, so the moistness we expect will be present in the NOT-winkie as well.

I am not opposed to the use of glucose/dextrose per se, but if a non-maize dextrose source such as rice or tapioca could be found, it might find its way into the sponge and/or filling of our next-generation Golden Sponge Cake as a means of optimizing sweetness (honey can be too sweet), while maintaining a reasonable shelf life.

 Beekeeping for Poets

 “—” ‘—’’