Sugar-free ice cream

[Tennessee Cowboy]

There is a NIH-funded study that says glycerine does not trigger and insulin response.  See http://archinte.jamanetwork.com/article.aspx?articleid=574143

 

[JoNorvelleWalker]

Xanthan gum is interesting and useful stuff.  However you have to measure just right or you get rubber ice cream and gastrointestinal excitement.  For real ice cream I have concluded xanthan gum is evil.  Test it first in MC@H gelato.

 

[adrianvm]

I don't use Splenda, but it is a mixture of sucralose and a bulking agent, looks like often maltodextrin.   In particular, if it's maltodextrin, then it's probably very fluffy, so you're not adding enough weight of the product to lower the freezing point significantly.  It's a similar problem to the one I face if I use 1/8 tsp of stevia concentrate. 

 

I think there may be a role for stabilizers, but they're not the key to success. 

 

[paulraphael]

By weight, maltodextrins have between a tenth and a third the freezing point suppression of sucrose. And they're not sweet. So yeah, not helpful.

 

Stabilizers will probably be very useful in these ice creams, but as you've said, they're not the solution to the freezing point issue. Since a lot of the sweeteners in question here work in small quantities, and since people are even talking about reducing the lactose (so no added milk solids) the ice creams will have low solids levels / high water levels. This means they'll need extra help to slow the growth of ice crystals and to have good body.

[Tennessee Cowboy]

Today I tried making a sugar free ice cream that combines the 25 minute technique from our friend Ruben Porto and the gum mixture from an egullet post that I printed out earlier but cant find now.  The recipe is as follows:

80 g egg yolk (approx 4)

70 g xylitol

50 g trim happy mama super sweet erythritol & Stevia blend

25 g poly-D powder

46 g skim milk powder

500 g cream (US)

200 g milk

1 vanilla bean, split and innards raked out.

4t vegetable glycerine

 1.5 g gum mixture (4 parts guar gum, four parts lamda carageenan, six parts locust gum)

 

procedure:  Weigh pan.  Mix egg yolk with dry ingredients.  Add milk and cream a mix for one minute.  Weigh pan again with mixture inside; do the math to measure your liquid (mine was 1020 g).  Bring to 162 degrees farenheit (71 degrees Celsius) for 25 minutes.  Weigh pan with liquid to assure 15% reduction.  

 

Add vegetable glycerine to hot cream mixture.  After starting stick blender, sprinkle gum mixture slowly until completely mixed.  Pour mix into plastic bag through strainer and put bag into ice bath.  I'll run it through the cuisinart tomorrow, freeze it and report results.

 

Some comments:  (1) After rereading some earlier posts, the mix may not have been hot enough to properly activate the gums, so it may not be a fair test.  (2) Forgot the salt.  (3)  After completing the recipe I went to the trim happy mama web site, and it appears their erythritol stevia blend is much sweeter than I thought.  However, my primary goal here is the holy grail of sugar free ice cream--scoopability and mouth feel.  I can adjust for sweetness later.  I am hoping that the combination of Ruben's 25 minute method and the gum mixture will succeed.  Watch this space!

[Tennessee Cowboy]

17 hours ago, Tennessee Cowboy said:

Today I tried making a sugar free ice cream that combines the 25 minute technique from our friend Ruben Porto and the gum mixture from an egullet post that I printed out earlier but cant find now.  The  I am hoping that the combination of Ruben's 25 minute method and the gum mixture will succeed.  Watch this space!

I ran this batch through the cuisinart this morning.  While it is freezing, I put a couple tablespoons of the cream in a small bowl in the freezer for a quick taste.  First impressions:  It has the same incredible smoothness as I got from Ruben's method previously.  The erythritol-stevia blend was too sweet, as feared.  Will have to wait another 24 hours to see if it is scoopable after freezing.

[Tennessee Cowboy]

On 7/28/2016 at 7:42 PM, Tennessee Cowboy said:

Today I tried making a sugar free ice cream that combines the 25 minute technique from our friend Ruben Porto and the gum mixture from an egullet post that I printed out earlier but cant find now.  The  I am hoping that the combination of Ruben's 25 minute method and the gum mixture will succeed.  Watch this space!

I ran this batch through the cuisinart this morning.  While it is freezing, I put a couple tablespoons of the cream in a small bowl in the freezer for a quick taste.  First impressions:  It has the same incredible smoothness as I got from Ruben's method previously.  The erythritol-stevia blend was too sweet, as feared.  Will have to wait another 24 hours to see if it is scoopable after freezing.

[Tennessee Cowboy]

Bad news and good news on my no-sugar-added experiment.  The recipe using sugar alcohols, polydextrose, glycerine and a blend of three gums was not a success.  The result was too scoopable.  It was gummy, almost pudding-like, the stevia taste overwhelmed the tongue, and the family turned thumbs down.  So I eliminated the gums and the glycerine.  Put another way, I used a conventional egg-based custard, a sugar alcohol blend and polydextrose.  I again used the the icecreamscience 25-minute-at-71 degrees celsius method of cooking down the ingredients developed by Ruben Porto.  The result?  Voila!  Sweetness that my stevia-hating spouse said was perfect, smooth mouth feel, AND ice cream that was scoopable as soon as it came out of the freezer!  It was not quite as elastic as I prefer, so I'm going to tinker, but I invite others to try his method and see if you get results siimilar to mine.  If so, we may have stumbled on a significant breakthrough for those of us trying to avoid sucrose in our ice cream. 

[ngotung99]

I don't like to eat much sugar which is unhealthy for me :((

[adrianvm]

That's an interesting result.  Did you omit the Trim Happy Mama sweetener the second time around?  It's not totally clear what your final formula was.   I generally prefer adding stevia separately from my other sweeteners instead of using a blend. 

 

It seems like your results are a bit different from mine.  I didn't find the Poly D was making a huge difference.  I wonder if the xylitol (which I haven't tried) is important.  (Xylitol has a small effect on blood sugar, so I haven't used it.)  Another observation is that you've got 34g of sugar in your recipe in the milk and powdered milk.  I wonder what role that is playing.  I was trying to avoid milk to keep all the sugars down. 

[Tennessee Cowboy]

Sorry about the omission.  The recipe was 500 g cream, 200 g milk, 80 g skimmed milk powder, 80 g egg yolks.  The sugar substitute was 70 g xylitol, 30 g PolyD, 15 g Splenda (my wife's request) and 5 g of Trim Healthy Mama supersweet blend (Erythritol and Stevia, in that order).  I left out the glycerine accidentally and included no gums at all.  The result satisfied my wife (she doesn't like stevia at all, uses xylitol daily in her tea) in both sweetness and mouth feel.  Most important, for me, it was scoop-able right out of the freezer though hard.  After five minutes in the bowl it was acting like the full sugar version I made a few days ago.  I may try the same recipe and add a tiny amount of my blended gum (guard, lambda, locust bean) to see if i can get it to be a tiny bit more elastic when frozen.

 

As of now, i feel that Ruben's method of cooking the mix at 160 fahrenheit for 25 minutes rather than the PolyD may be the key to scoopability.  If so, I've discovered something really important.  BTW, have you ever tried the extended low-temperature cooking method?  

[adrianvm]

How could the extended cooking method affect scoopability?  It seems like it could potentially affect texture and mouthfeel, but I don't see how it could affect scoopability.   The reason for the extended heating appears to be "to promote...protein denaturation, which contributes to texture."    I have never tried this method. 
 

I'm thinking may be I should try a preparation using whey and casein protein to replace the milk and milk powder.   Did you use 36% cream and whole milk?   Under this assumption I get the following formula

 

522 g cream

80 g egg yolks

162 g water

28 g casein protein

7 g whey protein

70 g xylitol

30 g polydextrose

5 g erythritol

 

Then I need to replace 52g lactose and 15g splenda with something else.  I could use 67 g erythritol. 

 

[paulraphael]

On August 3, 2016 at 8:38 AM, adrianvm said:

How could the extended cooking method affect scoopability?  It seems like it could potentially affect texture and mouthfeel, but I don't see how it could affect scoopability.   The reason for the extended heating appears to be "to promote...protein denaturation, which contributes to texture."    I have never tried this method. 

 

 

The idea of extended cooking is somewhat misunderstood around here. It can potentially do just two things:

 

-increase the solids content through evaporation.

-get the milk proteins to work more efficiently as ice cream emulsifiers

 

For the first function, this method is no better than simply adding solids in the form of nonfat milk. I'd argue that it's worse, because it's less controllable and is time-consuming.

 

The second function is irrelevant if you have other emulsifying ingredients in the formula. Like egg yolks. Commercial ice cream makers like Jenni's use heat denaturation because they want to avoid eggs (an esthetic choice), and they want to avoid additional emulsifying ingredients (a marketing choice). If these aren't your concerns, then there's no compelling reason to go through the trouble.

 

There's a slight chance that if you made two versions side-by side with identical final formulas (one made with heat reduction, one made with additional cream and nonfat solids and other solid ingredients) you'd get slight textural and flavor differences. But I've never seen this done—certainly not in a controlled, blind taste test.

 

If you want to experiment with heat denaturization, at the very least do it logically and leave the eggs out.

 

I've written about using denatured proteins as emulsifiers here, and as stabilizers here. The truth is that the published science on these topics, with regards to ice cream, is scarce. Most of the practical knowledge lies behind closed doors at places like Jeni's and Haagen Dazs. Jeni might give you some hints if you ask nicely but the big guys won't. 

[adrianvm]

I went ahead and made this following the extended cooking procedure (even though paulraphael thinks it's mixing up variables).  I used the formula I listed where the milk is replaced by casein protein, whey protein, water and the 67g erythritol.   I froze it this morning and I'm pretty sure it's going to be scoopable, because it was almost soupy after it was frozen (at 16 deg F).  In fact, I predict that it's going to be unnecessarily soft.   I did find that the sweetener combination seems to bother the back of my throat, so I don't know if I'd say this is an acceptable final recipe. 

 

Why is this formula soft?  Is it simply that the quantities of sugar-like substances (xylitol, erythritol and polydextrose) suffice to replace the sugar and lower the freezing point?  Or does it have something to do with protein?.   It seems like the next step would be to eliminate the long cooking step and see what happens.  That's a pretty annoying step.  The recipe would need to be adjusted to account for the water loss in cooking.   For my recipe it's easy enough to remove 15% of the weight in water.  That leaves only 18 g water remaining, which actually also makes it obvious that this recipe is unusually high in fat.  It's got the fat and milk solids without the accompanying water.   (I have assumed you used heavy cream, 36% fat.)   It looks like if you want to do this using without protein powders then you would use 520 g heavy cream, 20 g whole milk and 14.5 g powdered milk to replace the milk and cream.  (You'd add this to your already existing powdered milk quantity.) 

 

[adrianvm]

I sampled the ice cream yesterday evening.  It was quite soft.  I tried it again this morning and it's harder, but still easily scoopable.  It's probably softer than it needs to be. 

 

However, there's an issue.  It has a terrible texture.  It's not smooth.  The texture is kind of grainy.  I'm not sure if it's ice crystals or something else---I suspect it might be something else.  But that's definitely not very encouraging.  

 

[Christen]

I know this is an old thread, but I've used this recipe with good success repeatedly, producing a smooth and easily scoop-able end product:

135g xylitol

1/4t kosher salt

1/4t xanthan gum

13.5oz whole milk (or full fat coconut milk for dairy-free)

2t vanilla extract + seeds from 1 vanilla bean pod OR 4-6t vanilla extract

480g heavy whipping cream (or chilled coconut cream for dairy-free)

1T vodka (omit if using higher amount of vanilla extract)

 

2 options for prep: heat or no heat. End results are comparable, but when heating per original recipe (warm the first 4 ingredients + vanilla bean scrapings until xylitol is dissolved, then chill) you have to wait longer for the mixture to cool before using. So I made a shortcut: 

 

Place xylitol, salt and xanthan gum in blender and blend until powdered. Add the milk and vanilla extract/pod scrapings and blend until well-incorporated. Voila.

 

Whip cold cream to soft peaks. Sometimes I play with this for a more or less dense end product. More whipped = lighter/fluffier ice cream. Less whipped = more dense.

 

Once cream is to desired texture, slow the mixer down and add the milk/sweetener mixture and vodka. The original recipe calls to put this directly into the freezer with no need to churn. I prefer to churn it anyway for 10-15mins and then freeze it.

 

If you use my blender shortcut this comes together very quickly with no cooking. The ice cream is easily scoop-able with no weird textures or flavors. It's lighter than a custard-based ice cream, but I prefer that, personally. You could play with the milk mixture and adding yolks to that though I suspect the fat ratio might get too high and leave you with a heavy mouth feel, and you may need to adjust the amount of cream to reduce fat content accordingly.

 

Something else to play with, anyway, if you're still looking for a solution.

Cheers!

[adrianvm]

I have found that if you use allulose or bochasweet it gives a soft ice cream as well. 

[paulraphael]

On 8/10/2016 at 8:52 AM, adrianvm said:

Why is this formula soft?  Is it simply that the quantities of sugar-like substances (xylitol, erythritol and polydextrose) suffice to replace the sugar and lower the freezing point?  Or does it have something to do with protein?.   It seems like the next step would be to eliminate the long cooking step and see what happens.  That's a pretty annoying step.  The recipe would need to be adjusted to account for the water loss in cooking.   For my recipe it's easy enough to remove 15% of the weight in water.  That leaves only 18 g water remaining, which actually also makes it obvious that this recipe is unusually high in fat.  It's got the fat and milk solids without the accompanying water.   (I have assumed you used heavy cream, 36% fat.)   It looks like if you want to do this using without protein powders then you would use 520 g heavy cream, 20 g whole milk and 14.5 g powdered milk to replace the milk and cream.  (You'd add this to your already existing powdered milk quantity.) 

 

 

Anything dissolved in the water lowers the freezing point. Fat doesn't directly lower the freezing point, but since it doesn't freeze (or at least it doesn't harden to the same degree as water) it softens the ice cream by replacing water that could freeze. We focus on sugars because they have greater powers of freezing point depression than most other ingredients, with the exception of salt and alcohol.

 

This is 100% a property of the size of the molecules. Smaller molecules (quantified as the molecular mass) have greater freezing point depression. This is why glucose and fructose each are more powerful than sucrose, which is a larger molecule made of one of each of the other two. And sugar alcohols, being larger molecules still, are less powerful than sucrose. Proteins are very large, and so are significantly weaker than any sugars. 

 

You're dealing with such an unconventionally balanced recipe that we can also look at just in terms of how little water is in there. There isn't enough of it to freeze and properly harden the ice cream.

[adrianvm]

I have no idea how I started with 162g of water, removed 15% and arrived at 18g.  And it's been a couple years, so I really don't recall what I was doing.  I think the original recipe of 2 parts cream and 1 part milk is fairly conventional. 

 

Note that allulose and bochasweet (pentose) are both small molecules.  The pentose with a molecular mass of 150 is even smaller than fructose at 180. So it's no surprise that they are effective at softening ice cream. 

 

[paulraphael]

I'm not familiar with pentose or any of its uses in food, but I need to correct what wrote above about sugar alcohols: they tend to have smaller molecules than the more common monosaccharides like glucose and fructose. So they have greater freezing point depression.

[adrianvm]

Well, I can tell you from my experience with making ice cream that if you replace sucrose with erythritol (a sugar alcohol with molecular weight 122) it transforms a soft ice cream into one that freezes rock hard, like ice cubes.   So that seems to contradict your theory. 

[paulraphael]

On 8/30/2018 at 4:47 PM, adrianvm said:

Well, I can tell you from my experience with making ice cream that if you replace sucrose with erythritol (a sugar alcohol with molecular weight 122) it transforms a soft ice cream into one that freezes rock hard, like ice cubes.   So that seems to contradict your theory. 

 

I don't know what's going on in your particular formula, but the inverse relationship between freezing point depression and molecular mass isn't just some pet theory of mine. It's a basic principle in chemistry. Chemists actually use it to calculate the molecular mass of unknown chemicals in solution.

[paulraphael]

Ok. Here's a paper on using erythritol in ice cream.

 

It says the hardness comes from crystallization of the the erythritol itself; the freezing point depression factor is actually three times stronger than that of sucrose. The suggested compensations are to combine with other sugar alcohols (sorbitol, sucralose, polydextrose). 

 

The complete sugar reduction is achieved by taking out the crystallisation inhibiting glucose syrup. However, due to erythritol’s strong crystallisation behaviour, a crystallisation inhibitor is indispensable. Several options from the polyol range were tested in the frame of the project. It was found that sorbitol, maltitol, xylitol and their mixtures, all perform quite well alongside erythritol. It was also discovered that all three indeed soften the ice cream structure. This creates a great toolbox where the polyol pair can be varied to modify product hardness and creaminess. The actual second polyol can be chosen according to individual preference as they all will harmonise with erythritol. Texture as well as sensory analysis revealed that the preferred polyol ratio should be in a range of three parts erythritol to two parts of the second polyol.

Edited September 4, 2018 by paulraphael (log)

[adrianvm]

That was an interesting read.  I wonder if there are other things that would inhibit erythritol crystallization.    Note for clarification, the article said that adding a sugar alcohol (they listed xylitol, sorbitol and maltitol) would inhibit crystallization.  The sucralose is added for sweetness and the polydextrose to "[add] bulk and ... creamy sensory attributes". 

 

[paulraphael]

45 minutes ago, adrianvm said:

That was an interesting read.  I wonder if there are other things that would inhibit erythritol crystallization.

 

Certainly dextrose. Not sure about other non-sugar or non-caloric options. I've never done battle with sweetener crystallization.