Ice Cream: Solids, Water, Ice

Scanning electron micrograph of ice crystals

This post is addendum to the post on How To Build a Recipe, and the post on Sugars. I want to clarify the importance of solids—which is really a reflection of the importance of water. These are key ideas—if you master them, you will be well on your way to texture Ninjahood.

We’ve discussed how ice cream is made up of three physical systems: an emulsion (fats suspended in water); a foam (air dispersed in a solid fat network); and a sol (solid water dispersed in liquid water).
Here we’re going to look at that sol—the interactions of water and ice. And we’re going to look at the effects of all the nonfat solids that are dissolved or suspended in the water.
It’s important—and instructive—that the ice is mostly pure water. This is the case because freezing acts as a purifying process; when ice crystals form, they expel dissolved and suspended solids. The result, besides much lower concentrations of stuff in the ice, is an increased concentration of stuff in the liquid portion of the water. 
This phenomenon is called Fractional Freezing. It creates an interesting system in which there’s no single freezing temperature for ice cream, or for any water-based solution. Instead there’s a temperature range between the extreme points where none of the water is frozen and where all of it is frozen. Between these points, you get a mix. The colder the temperature, the greater the proportion of frozen water, and the higher the concentration of solutes in the liquid water.
Fractional freezing works because of the colligative properties of water regarding freezing point depression.* The stronger a solution, the lower the freezing point. So when a bit of a solution freezes, strengthening the concentration of the remaining liquid portion, that liquid’s freezing point drops. And so on. This process is continuous and self-regulating. And pretty cool. 

So What? 

We need ice to make ice cream, but we don’t want too much. With too much ice, you get a popsicle. In order to have a lower proportion of ice, we need a higher proportion of other stuff. 
Some of that other stuff is fat. But we likewise don’t want too much, or the ice cream will be too rich and the flavors too muted. Some of that other stuff is air, but we really don’t want too much of that, or the ice cream will be too fluffy and insubstantial. 
The remainder is liquid water, and dissolved or dispersed nonfat solids. These two are intimately related, because 
1) the more nonfat solids, the lower the percentage of total water, and
2) the more nonfat solids, the greater the portion of that water that stays liquid
So, in general, solids are good. Solids with low molecular weights** depress the freezing point the most, while all solids displace some of the water. 
We especially like milk solids—specifically the nonfat portion—because they effectively concentrates the milk. We get more of the functional qualities of milk, like emulsification, freezing point depression, and improved body, but without added water. And milk tastes good, the way ice cream should.
Sometimes we get solids from the flavor ingredients: chocolate and cocoa, fruit pulp, nut butters, matcha powder, coffee solids, etc.
Remember from the How to Build a Recipe post, we typically aim for the following levels of solids in a well-balanced recipe:
Nonfat Milk Solids: 10-12%. Or higher for low-fat ice cream.
(everything in milk besides the water and fat)
Total Nonfat Solids: 22–25% 
(everything in the ice cream besides water, alcohol, and fats)
Total Solids: 37–42%
(total nonfat solids plus total fat. everything besides water and alcohol)
If solids levels are too low, ice cream can lack body, freeze too hard, and may have textural problems like iciness. If solids levels are too high, ice cream can be excessively chewy or cakey. If the milk solids specifically are too high, you can get grainy textures. 
Just remember that by managing the solids, you’re managing the water.
Periodic Table of the Elements. The number int the top left of each box is that element’s atomic mass (when a molecule is made of many atoms, we add the individual atomic masses together to find the Molecular Mass). Lower number = smaller molecule = more molecules at a given weight = greater freezing point depression.
*Colligative properties are based on concentration of dissolved solutes in the water, and on the size of their molecules, but not on any special chemical qualities of those solutes. Two equally concentrated solutions of two chemicals that have the same molecular weight will depress the freezing point of water equally. This is handy to know; it gives us just two values to consider when adjusting ice cream’s hardness.
**Molecular Wuh?? The molecular weight (also called molecular mass) of substance is a number that tells us the size and mass of its individual molecules. Molecules are made of atoms, and atoms are made of protons, neutrons, and electrons. The protons and neutrons have equal mass, while the electrons are so tiny and insubstantial that we can ignore them. So the molecular weight is the total number of protons and neutrons. 
A high molecular weight means the individual molecules are relatively massive. This means that a gram of something with a high molecular weight contains fewer molecules than something with a low molecular weight. And this, in turn, is why substances with a low molecular weight are more powerful at depressing the freezing point: for a given mass, there are more individual molecules in dispersion in the water, exerting their colligative influence. 
Next post: infusing flavor (yes, we’re eventually getting to the good stuff)
Notify of

Newest Most Voted
Inline Feedbacks
View all comments
5 years ago

Hi Anon, I'm planning to get more into fruit flavors this summer. You're right, the added water is an issue. So is the added sugar. I'm planning to share a chart that shows typical water and sugar contents of various fruits (including the breakdown of sucrose, glucose, fructose, etc.) so that we can calculate the fruit as a mix of all these ingredients. I think the only problem with your recipe is that you've added sugar rather than subtracting it. Raspberries on average include around 10% sugar by weight, so you're already adding 20g sugar (roughy a third each glucose, sucrose, fructose). The puree is probably also 85% water, so you've got 170g water to contend with. The way to do that is to add enough nonfat dry milk to turn that water into milk … and subtract an equivalent amount of milk from the base recipe (which it looks like you did more or less).

5 years ago

Hi Paul! I'm currently playing around with berry sherbets (right now with raspberries). I use one of your base without egg yolk as a base recipe, because yolk tends to mask the berry flavor. Currently it's looking like this:

300g milk
100g cream
85g non fat milk powder
100g buttermilk
200g raspberry puree
10g freeze dried raspberry

75g sugar
40g dextrose
20g trimoline

2g lecithin
0.8g locust
0.6g guar
0.4g lambda
0.7g salt
1g citric acid

So basically I lowered the amount of cream and milk and add raspberry puree and a little bit of buttermilk/freeze dried raspberry/citric acid to brighten up the flavors. I also add a little bit of sugars because my freezer temperature is around -21c, so the ice cream would not be so hard..

But anyway, do you have any experience about berry/fruit ice creams without egg yolk? Should I add more trimoline, milk powder or stabilizer to compense the added amount of raspberry puree? (raspberry puree = lot of water, not a good thing!) Thanks!

5 years ago

Hello! Does the sugar and protein in cream count as NFMS? Or is it that which is in milk only?

5 years ago
Reply to  Unknown

Proteins can come from the milk solids, or they can come from egg yolk, or possibly from other ingredients. They’re all part of the solids (but only the milk proteins are part of the NFMS.

All the added sugars are likewise part of the total solids. But not the NFMS.

There IS a bit of sugar (lactose) in the milk solids.

5 years ago

It sounds like a lot of sugar also. Have you tried using less? I don't know what's typical for popsicles. You're using double what's conventional for ice cream, and I think most ice cream is too sweet. You might want to look at the post on sugars, and see if you can reduce the total sugars, and find a way to balance sweetness with whatever physical properties of the sugars you feel you need.

5 years ago

Thank you for your prompt reply. I shall surely try to make it with 2grams of stabilizer. But do you think the sticky surface on the popsicles could be cos of excessive sugar or is the excess stabilizer the culprit?

5 years ago

Hi Shruti, thanks for writing. A paleta is like a popsicle? Fully frozen without air? I don't have much experience with this, but suspect you're using way too much stabilizer. That's about double what I use for ice cream, and I'm not sure you even need stabilizer for a popsicle … there's hardly any unfrozen water that needs to be stabilized. Have you tried making them without stabilizers? Or with half as much?

5 years ago

Hiii Paul
Thank you for an insight on ice-cream.
I have been trying to make fruit paletas with 1000ml water, 400 GM sugar n 4 GM's of stabilizer (407,410,412,415,466,471)
My end product has a glue like feel on the outside.
Can you tell me as to where am I going wrong.
Have been going crazy with experimenting. But no success yet.
Looking forward to a solution
Love from india

5 years ago

Hi Ryan, thanks for writing. I don't know anything about dairy-free ice cream, but have had some pretty tasty ones made from coconut milk.

And I have some experience with sorbets, which I'll write about eventually.

5 years ago

I really hope you continue this series, thank you for making it.

I see that you're a photographer, but you don't post pictures of the ice cream you've made! I'd be very interested in seeing your results.

5 years ago
Reply to  Anonymous

Thanks Anon. Food photography isn’t what I do best, but I’ll make an effort …

5 years ago

You mentioned a post about sorbets!
You also mentioned somewhere about a chocolate sorbet you are making.

Any plans on posting the post and/or the recipe?

I really love the work you 've done here.


5 years ago
Reply to  kostelidis

Look for more on sorbets this summer, 2019 …

5 years ago

Hello, I really enjoyed this series of post and hope you find time to continue them.

Do you have any pointers for dairy-free icecream? I'd like a starting point if you have any experience in this area.

5 years ago
Reply to  Ryan

Hi Ryan, thanks for writing. I don’t know anything about dairy-free ice cream, but have had some pretty tasty ones made from coconut milk.

And I have some experience with sorbets, which I’ll write about eventually.

There seems to be some interest in dairy-free bases, so I plan to do some research.

5 years ago

Thank you Paul for responding so quickly and for your helpful suggestions. I use erythritol and a little sugar, a combination of sugars and sugar alternatives – I'm still experimenting. Thank you for answering my question re quality dry milk powder brands and what to look for, and for all your other suggestions. I'll keep working at my recipes and let you know how they turn out. I already like the taste of the base after I've chilled it (except for the almond milk flavor, but even that is not horrible). I'm really looking forward to improved texture. Thank you again. Best regards.

5 years ago
Reply to  Stacy Hackney

You’re welcome, Stacy. Good luck!

5 years ago

Hi Paul. I recently discovered your great blog and already love it and have found it informative and helpful, even though currently I understood almost none of the technical, food science stuff. I'll keep re-reading and asking questions until I figure it out. I'm insulin resistant and poor and love frozen desserts but most are either too pricey, don't use the sugar alternatives I prefer or, like Haagen Dazs, don't sell a low carb dessert, so I've been working to create a lower carb option for myself which currently is mostly unsweetened almond milk because it is almost zero carbs, is what I can most affordably find right now, but which I will be switching out for cashew milk because I just don't like the taste of the commercial almond milks available to me. In addition to the almond milk, soon to be cashew milk, I also add heavy cream and a small amount of milk and buttermilk. Dairy milks are carbier than almond/cashew milk and heavy cream and I only use them because I prefer the flavor dairy milk gives my frozen dessert versus when I use only almond milk/cream. Because I've been insulin resistant for some years now I've experimented with many low carb sweeteners and use those for the sweetening the frozen dessert, with a little dextrose, sugar and reduced carb corn syrup to cover up the aftertaste the sugar alternatives leave behind. I'm now learning to use gums so I'd appreciate any tips you care to provide and I also will be trying lecithin as I saw you recommended it to another commenter. I also use egg yolks and dry milk powder. So far the frozen dessert when it comes out of the ice cream machine tastes good, I add plenty of vanilla (which fortunately I bought some years ago before the price went crazy). The frozen dessert recipe I currently have, though still in flux, is a low carb gelato – something I can eat without it kicking up my blood sugar, and if I get the recipe right, I'll begin a tiny business and sell it at my local farmer's market once a week if I can feel well enough to do that. My current version tastes good when initially out of the ice cream maker, but it sucks when frozen. The texture is horrible, icy and grainy. I eat it because I can't afford to waste food, and also because it actually tastes good, the flavor is good. Can you help point me in the right direction for a combination of gums for my low carb gelato which is going to be mostly nut milk, with almost no sucrose. Because I'm obese I will be creating a frozen dessert that's lower fat, though I will use heavy cream and eggs in it, just not as much cream as in ice cream. I use nonfat milk powder but it's what I can get from Smart & Final. Can you recommend a high quality skim milk powder brand or two? Can you recommend a combination of gums for a gelato that's mostly cashew milk with almost no sugar. Thank you for any help. Best regards.

5 years ago
Reply to  Stacy Hackney

Hi Stacy, thanks for writing. The most important thing for your recipe is going to be the sweeteners. You want to reduce the calories, and reduce the glycemic load, but still have it taste good, and still get enough freezing point depression so that the ice cream doesn’t freeze hard as a rock.

I’ve done just a bit of research into this stuff, and have experimented a couple of times with low-sugar frozen yogurt for my girlfriend. I’ve gotten some decent results but am no expert.

The best low-calorie, low-glycemic sweeteners for this purpose are probably the sugar alcohols. These are sweeteners that end with -ol, like sorbitol, xylitol, and erythritol. The last one is possibly the best for ice cream, but is fairly expensive. They are derived from sugar, but are not digested and converted to calories. The drawbacks are that they don’t taste exactly like sugar, and that in large quantities they can give you gas or a bellyache. I think it might make sense to use a blend of maybe 1/4 sugar, 3/4 sugar alcohol. This is just an educated guess.

I don’t think you need to worry about lecithin. Just make the ice cream with a couple of yolks per quart. This is enough emulsifier to make the texture nice, but not enough to add lots of fat.

For stabilizers, I’d start with the standard blend I recommend in the stabilizer post, or if you want to keep things simple, buy some off-the-shelf stabilizer like Cremodan.

It’s important to get the level of solids high enough to help with the texture. The ideal ingredient for this is nonfat dry milk, as you’ve suggested. I use Now brand organic. You can get it on Amazon. I’ve also used Horizon brand with good luck. Make sure the ingredients say 100% skim milk solids. No other ingredients. Ideally you also want one that’s dried at a low temperature. Now brand does this. But this is less important.

Just be sure to avoid using maltodextrin, which seems innocent enough (it hardly tastes like anything) but is in fact a sugar with a high glycemic index. It’s popular for adding solids to some sorbet recipes.

I hope this helps! Feel free to hit me up with any other questions. And if you work out something that you like, let us know.

5 years ago

Amazing..Really hoping you get a chance to continue this series, especially on flavors and adding fruit!

5 years ago

Fruit flavors are in the works … more to come in 2019 …

Wally Jacob
5 years ago

Hi Paul,
I've been working through your posts and have to say, this is AWESOME! Thanks for putting this together. But, truth be told, I was never great at chemistry, and am having a hard time applying the information you are giving us into my recipes.
I've been working on a recipe for a frozen nonfat Greek yogurt desert that I add citrus flavored fish oil to. Sorry if you just threw up a little, but hear me out. I'm trying to develop a "healthy" desert. I love ice cream, but I don't like what it does to me, so I said, "Hey. Why can't I make it healthier?" So, here's the recipe I've come up with with weights in grams. Oh yeah, I also keep it at a 1:1:1 ratio of calories from carbs, protein, and fats. That's the portion control mechanism I use:
Nonfat Greek Yogurt 684.9285171
MPC 85 684.9285171
Maltodextrin 114.1547529
Honey 296.8023574
Orafti P95 114.1547529
Orange Juice Concentrate 228.3095057
Citrus Flavored Cod Liver Oil 308.2178327
Skim Milk 570.7737643
Guar Gum 10
Locust Bean Gum 10.00
I'll add flavoring to make it more like a sherbet once I get the texture down.

I heat it to 165, cool it in the fridge, and then turn it into sherbet in a condenser style ice cream maker.

It still gets real hard after freezing it in 16oz deli containers to the point that I have to let it sit out for 45 minutes.

The main health benefit from this from the fish oil. Its a tasty way to incorporate it into my diet.

What can I do to enhance the texture while keeping the 1:1:1?


5 years ago
Reply to  Wally Jacob

Thanks for writing, Mr. Jacob.

I applaud your efforts, but have to say you’re way out into unknown territory here, and so my knowledge may not take you very far. I’m not saying that cod liver oil-based Greek yogurt ice cream is some kind of crime against nature (which, of course, it is). It’s that approaching ice cream in terms of macronutrient ratio (carbs, proteins, fats) is unconventional, and not anything I’ve seen done.

Nevertheless, here are a couple of tips. If the biggest problem is that it freezes too hard, you should look at the dissolved solids, and at the molecular mass of those solids. I see you’re using maltodextrin (I assume this is to add bulk). This sugar is a giant molecule that’s quite ineffective at suppressing freezing point. See the chart in the post on sugars. How about replacing with nonfat dry milk? You’ll be adding protein, freezing point suppression, bulk, and it adds very little nutritional sugar. I’m not sure if this is the same as the MPC you’re using, but you can always use more of it. You can also replace some of the sweeter sweeteners with dextrose.

Good luck! With the formula, and with getting anyone to try it ;)Reply

5 years ago

Hey Paul! THIS IS AN AMAZING BLOG! I'm actually starting up my own ice cream truck in Arkansas and am using your techniques on stabilization to make all of my flavors egg-free. What are the major differences between these "home" recipes and places like Oddfellows, Jeni's, etc.? It seems to me that even they could benefit from stabilizers. I'm lucky enough to use VAT pasteurized, grass-fed dairy with a great milk fat content. Yet, I still find using an average of 0.1%-0.15% LBG:Guar:Lambda mixture greatly increases my mouthfeel and scoop for the better! I can't wait for your next blog post!

5 years ago
Reply to  Unknown

Hi Unknown, thanks so much for writing. I don’t know anything about Oddfellows. I know a bit about Jeni’s (Jeni herself was kind enough to answer some questions, and a few years ago her company site shared some details about their process).

Jeni’s does stabilize their ice cream. They use tapioca starch, almost certainly because it’s label-friendly (those words sound more like food than like “chemicals”). But if the quality of your ice cream is the number one priority, you can do better than starches, by using gums like the one’s you’re already using.

Where Jeni’s gets interesting is emulsification. Their ice creams are completely egg-free, and they don’t add any emulsifiers to make up for this. They get all the emulsifier action from the milk proteins.

They do it by using a mix that’s very high in milk solids. I don’t know their process now, because they have their dairy do all the heavy lifting. But back when they made the ice cream in-house, they started with raw milk, and centrifuged it to separate it into cream and skim milk. Then they’d use reverse osmosis to concentrate the milk without adding any heat (you could also use vacuum evaporator). Then they would blend this concentrated, high-solids milk with the cream to get whatever fat level they wanted. They’d maximize the emulsifying power of the milk proteins by cooking at 75°C for an hour, in batches.

Now I believe they use a continuous process, with higher temps and shorter times. Jeni didn’t go into details. She did suggest that it would be very hard to get the results they get if you didn’t start with raw milk—presumably because you’d have no control over the time/temperature of pasteurization, which is essential for getting the proteins to work as emulsifiers. But raw milk is challenging to handle safely—Jeni’s actually had to shut down some years ago because they discovered some contamination.

Anyway. If you don’t have all the industrial-level toys, the best way to add milk solids it to use nonfat milk powder. If you get good stuff (100% skim milk, low-temperature spray-dried) there are no disadvantages.

Then the trick is to replace the emulsifying power of the eggs. Cooking the proteins as I describe here should help. Adding an emulsifier should help. If you’re making ice creams with low fat or high overrun, this can get tricky, but for more deluxe ice creams that have a fair amount of milk fat and not a lot of air, I think lecithin should work well.

Let me know how it goes. If the lecithin doesn’t work for you, I’d suggest calling tech support at TIC Gums or Kelco. They know a lot. If they help out, please share what they say.

Good luck!

5 years ago

Thanks for this Paul, and for all of the other blogposts. This blog, together with the comments people post, has been the single most useful resource in terms of home ice cream making for me.
I recently had the opportunity to buy a well-reviewed compressor-style ice cream maker at a crazily low price (what would be $85!) and it's on it's way to me. I have a freezer bowl style at the moment, but frustrating to an experimenter like me because you have to wait at least a day in between attempts rather than beast it in an afternoon!

Hence I've progressed slowly with my knowledge and attempts – adding some xanthan, LBG and a tablespoon of vodka to an egg base was as far as I'd got (although these were all big improvements). But as I already have some lecithin I'm ready to quit the eggs altogether – never enjoyed the laborious custard-making process, and if I swap out LBG for guar I can totally avoid having to heat (other than a little to help dissolution).

I realise now after reading your blog that what I'd assumed was iceyness was actually chalkiness from too much fat. The recipes I've been using use way too much cream – I'm going to go for a more gelato style and use mostly milk I think. And switch in some invert/dextrose too.

Cannot wait for this machine to arrive – it better work!

Have you ever used tapioca starch or arrowroot as stabilisers/thickeners? I like the idea of the glossiness that the arrowroot might provide (saw it somewhere – could have even been this blog but I think it was somewhere else) and I have a big tub of an "instantised" tapioca starch as a versatile sauce thickener that might be worth a go…

5 years ago
Reply to  Tom

Hi Tom, thanks so much for writing. Nice score on the ice cream machine.

I haven’t used Tapioca starch, but it should work fine. Jeni Britton Bauer of Jeni’s homemade uses it, and her ice creams are great. However, I believe she chooses it because it’s “label friendly” … it sounds like something your grandma might have used. It doesn’t offer any functional advantages over the gums, and will probably not work as well, or offer as much versatility.

I have tried arrowroot, and give it the big thumbs-down. It gave the ice cream an inedible, slimy texture. I’d read a warning about bad textures with arrowroot in dairy, but had to try for myself, since it’s one of my favorite thickeners on the savory side of the kitchen.

My experience was bad enough that I never felt inclined to try again. This isn’t the final word—since then I’ve seen a recipe that included it, from someone I had no reason to believe was crazy. But at this point, I’m not interested in starches at all.

Good luck!

5 years ago

I used Sunflower Lecithin powder for the first time today. Do you have any tips on getting that stuff to dissolve?

5 years ago
Reply to  Barbara

Hi Barbara,

did you try heating it? Most lecithins need to get to around 180°F / 82°C to dissolve. Disperse it first along with the other solid ingredients with a blender.

Noah Baker
3 years ago

In considering the concept of freezing point depression; it brought to mind a health product intended to add trace-elements to mineral deficient water …

[Comment edited to remove product link]

Theoretically; if it didn’t affect the taste; then the addition of these trace elements to the liquid portion of the mix would bring down the freezing point because of the dissolved substances, correct?

3 years ago
Reply to  Noah Baker

Anything that dissolves is going to lower the freezing point. The question is by how much. The saline level of sea water only depresses the freezing point a few degrees. You need more than a trace of something. And those minerals (mostly ionic salts) if there’s more than a trace, you’re going to taste them.