Crystal Clear Ice

[MaxH]

... I am not sure that freezing water can selectively concentrate impurities. Also, it seems to me that the actions of "zone refining" in chip making is different than water freezing. Zone refining uses progressive thawing and refreezing to move impurities in a chrystaline structure.

dcarch, you may wish to look into freeze distillation, an application of fractional freezing (I mentioned it upthread). It explains how freezing of water is well established to concentrate dissolved impurities to one side of the liquid-solid boundary, a consequence of the impurities altering water's melting temperature. (I mentioned a very practical application, concentrating sugars in Riesling grape juice.)

Zone refining of semiconductor ingots indeed entails both a melting and a re-freezing boundary, which move in unison through the material. Techniques here are simpler, with only one phase boundary, not two, moving through the water in what eje dubbed "directional freezing," a common feature of successful experiments reported here. Fractional freezing explains how, as a mass of water freezes along one direction, the moving boundary could "push" solute impurities along, leaving purified ice.

Kohai's goal is basically identical to that of semiconductor ingot makers (albeit with different working material). Both are growing large pure single crystals. Strictly speaking, water is even a "semiconductor" (in purest form its electrical conductivity is very low; conductivity comes from ions contributed by impurities).

[dcarch]

I have been to a chip making factory and have seen “zone refining” and doping of semi-conductors. I didn’t think that would apply to ice making in water. But you are correct, “fractional freezing” is a way to isolate impurities in water.

It appears to me that impurities in water, if concentrated by fraction freezing, would not appear as white bubbles in the ice.

Yes, pure water is a very poor electrical conductor. Interestingly, molten glass is an excellent electrical conductor.

Maxh, you have a very good engineering/science mind!

dcarch

[jmfangio]

I decided to try a little experiment based on what I've been reading in this thread, wondering if I could get clear (or, at least, significantly clearer) ice balls from my ice ball mold. I pulled out a bottle shipping box from my closet that's a perfect fit for the mold, filled the mold with water just taken off the boil, carefully slipped it into the styrofoam, then into its cardboard box, and into the freezer overnight.

The result:

There's still a cluster of bubbles, but the rest of the ball is perfectly clear.

For comparison, here's another ball with the same water, but stuck into the freezer without insulation:

[peterb]

I think this method is too complicated. I've always found that using hot water results in clearer ice. I think starting with just-boiled water (for "normal" sized ice cubes) is all you need to do.

[vice]

For crystal clear (not just clearer) ice in large blocks, simply starting with hot water has been shown to be insufficient.

[Kohai]

Right. I've never noticed much difference between boiled and unboiled (tap) water; call it zero difference.

Keep in mind, I am specifically addressing block ice, not cubes. With block ice, you get expansion fractures, which the method described above virtually eliminates in addition to making markedly clearer ice.

From what we're seeing with jmfangio's and Chris's photographs, as well as what the scientifically-gifted among us have explained upthread, slower freezing (by insulating the molds, for example) seems to be another key. Unfortunately, that is not ideal for a commercial bar application. Wonder if there's a workaround.

[jmfangio]

From what we're seeing with jmfangio's and Chris's photographs, as well as what the scientifically-gifted among us have explained upthread, slower freezing (by insulating the molds, for example) seems to be another key. Unfortunately, that is not ideal for a commercial bar application. Wonder if there's a workaround.

I've been wondering about the slower freezing for a while, but hadn't thought about insulating the ice mold until I saw Eje's link to Camper English's experiments, and Chris' post.

My curiosity about slow freezing came from watching the Japanese cooking show Cooking Showdown, when one week the 'secret ingredient' was ice from a several hundred year's old producer. As winter approached, they would divert water from an underground stream into a shallow pool, and let it slowly freeze in the open air. When cut, the chunks were as clear as glass. Not exactly practical for us, I know, but it was beautiful to see, and was the first thing that made me think that slow freezing was one key.

This isn't exactly the most practical idea either, but I suppose that if you have a spare freezer, you could try to replicate this effect by starting with it off, then bring it up (or down, as the case may be) to temperature as slowly as possible? Maybe use a voltage regulator, like some folks use on rice cookers or crock pots for sous vide?

[Chris Amirault]

Well, I guess we finally figured out what Muji means by "freeze slowly."

[jk1002]

Freeze slowly I heard over and over. I think one can argue what it actually means, starting with hot water or starting with very cold water ....

I think taking a fridge and gradually lowering the temperature of the water over time, I think this is in the end what sort of happens with melting the hollow ice cubes approach, the water comes very very close to freezing this way and putting it into the freezer is just the last tiny coffin before freezing ......

I don't have a spare fridge but I believe some can actually get to freezing temperature ..... this would also explain how it can be done automatically by machines .........

[dcarch]

Several points:

1. A normal refrigerator compressor is either on or off. There is no adjustment for temperature. An automobile compressor can, by varying the speed of rotation. A solid state (peltier junction thermoelectric heat pump) device can by varying electric current. Thermostatic control in a freezer is not accurate. It can vary by more than 10 to 20 degrees.

2. As long as there is ice in water, it does not matter how big the ice chunk is, the temperature of the water will always be the same.

3. Cooling pure water slowly can result in super cooled water. The super cooled water can freeze instantly, resulting in a very cloudy block of ice.

4. I am not sure about this: Water never chrystalizes when going from water to ice. It only chrystalizes by scintilation, as in snow flakes, as in ice forming on the freezer walls, i.e. from vapor to solid without the liquid phase.

dcarch

[slkinsey]

Ice is the crystalline phase of water.

Edited August 22, 2010 by slkinsey (log)

[MaxH]

I have been to a chip making factory and have seen “zone refining” and doping of semi-conductors. ... It appears to me that impurities in water, if concentrated by fraction freezing, would not appear as white bubbles in the ice. ...

Maxh, you have a very good engineering/science mind!

Thanks. It's my background and occupation. Unfortunately I could tell you much more about semiconductors than about ice, the issue here! (I was intrigued by SLK's original mention of hexagonal crystals, but he seems to've changed his mind.)

One thing about dissolved impurities is that when they precipitate out (as they do at low temps. - solubilities tend to decrease radically as water temperature falls), they become solid particles within the ice. There may be further consequences then as the ice forms, like nucleating crystal flaws or even gas bubbles (just as a dust particle or a scratch catalyzes gas bubbles in a glass of Champagne). Whatever the details, when I used specially purified water to make ice cubes, they appeared consistently clearer.

An easy "field improvised" test for certain impurities in water, by the way, is to brew tea with it. This only detects certain impurities, like heavy metals, which aren't the only ones relevant here. But if they're present, a glass of tea brewed with the water will be darker than a "control" glass made the same way but with purified or distilled water. The tannic acid in tea forms dark inert pigments with heavy metals. (That property has been variously exploited to make classic iron-based ink and to counteract heavy-metal poisonings.)

[Chris Amirault]

Is it my imagination, or are these clear ice chunks (a) harder, even at the same temperature, as my usual cubes, and (b) likely to melt more slowly? I will absolutely defer to any argument that suggests both are fetish effects and not science.

[Kohai]

The guy I talked to at M&H about their double-freezing method said that the second freeze compacts the ice further. I don't quite get that myself (partly because I'm not 100% clear on how they do it there - is the ice completely thawed, or is it only half-thawed?) but it seems like that might apply to this, too. So you could indeed be right.

(Though now that I'm wondering, I can't say that I've noticed much difference in the hardness of the ice, meself.)

[vice]

The guy I talked to at M&H about their double-freezing method said that the second freeze compacts the ice further.

I doubt that there's any difference in the volume of the ice between first and second freeze. I'd guess that the fault lines in cloudy ice, presumably caused by trapped air, should weaken it relative to clear, air-free ice. Someone else'll have to take a stab at melting rate. I don't have a clue.

[Kohai]

Just to nitpick, fault lines aren't created by air trapped in water. They're caused by the pressure of expanding water in a rigid container. Cloudy ice and aeration are two different things; you can have cloudy ice without fault lines; you can also have fault lines in otherwise clear ice.

As far as whether M&H's double-freezing method works, I am certainly not the person to be vouching for it since I've only been there once. But I have heard it verified by multiple sources that I trust, and I have difficult imagining anyone going to the trouble unless it did work.

Edited August 24, 2010 by Kohai (log)

[vice]

As far as whether M&H's double-freezing method works, I am certainly not the person to be vouching for it since I've only been there once. But I have heard it verified by multiple sources that I trust, and I have difficult imagining anyone going to the trouble unless it did work.

Oh, I agree. I'm not doubtful that the method works, just of the explanation that refreezing somehow compacts the ice more than when it was first frozen.

[slkinsey]

Is it my imagination, or are these clear ice chunks (a) harder, even at the same temperature, as my usual cubes, and (b) likely to melt more slowly? I will absolutely defer to any argument that suggests both are fetish effects and not science.

It's extremely doubtful that there would be any meaningfully noticeable difference based only on this variable.

The guy I talked to at M&H about their double-freezing method said that the second freeze compacts the ice further. I don't quite get that myself (partly because I'm not 100% clear on how they do it there - is the ice completely thawed, or is it only half-thawed?) but it seems like that might apply to this, too. So you could indeed be right.

What, exactly, is this "double-freezing method"? I have heard tell of double freezing described two ways. One way is to freeze the ice, allow that ice to melt, and then re-freeze the melt-water. I can't see how this would possibly have a positive benefit versus other methods that seem easier to me, except perhaps that precipitated solids might sink to the bottom of the freezing containers. The second way I have heard this described is to make the ice (e.g., in a machine) and then temper that ice to a colder temperature in a freezer.

The latter technique could confer some benefits, namely that the ice will be colder and therefore harder. The idea that re-freezing melt water "compacts the ice" sounds like nonsense to me. The best you can hope for is to have less air trapped in the ice, although it is not clear to me that freezing, melting and re-freezing works any better at getting rid of trapped air than double-boiling water and freezing it directly from hot. In either case, what you need to figure out how to prevent is the reabsorption of gas into the cold liquid (solubility of gas in water increasing with lower temperatures right up to the point where the water freezes and the gas precipitates out). So, for example, filling a container all the way up to the top with double-boiled boiling water, then sealing the container and freezing the contents without allowing any air to enter would be one way of doing that (you'd want a flexible container to prevent stress fractures from clouding the ice). A different way of doing something similar would be to precipitate out all the gas by freezing the water, then allowing the ice to melt while taking measures to minimize reabsorption of gas, then re-freezing the water. Compared to freezing water directly out of the tap, if you're careful to avoid agitating the freezing container, double-freezing ice should produce some improvements. But it's by no means a magical solution to this problem. As for density... frozen water, in the forms which can be found outside of a laboratory, is 0.92 g/mL. Period. Re-freezing melt water isn't somehow going to make it come out to 0.93 g/mL.

Edited August 24, 2010 by slkinsey (log)

[slkinsey]

I've thought that "bottom up freezing" ought to be a pretty good way to get clear ice. My thinking is that you could construct a freezing container with a thick aluminum base and thick-but-flexible latex creating the four vertical walls around it. Then, you put the empty container into the freezer until it gets good and cold. And when you fill the container with water, the aluminum starts freezing the water from the bottom up due to conduction. If you were able to jiggle the walls every so often, that would prevent ice from forming over the surface.

[dcarch]

I've thought that "bottom up freezing" ought to be a pretty good way to get clear ice. My thinking is that you could construct a freezing container with a thick aluminum base and thick-but-flexible latex creating the four vertical walls around it. Then, you put the empty container into the freezer until it gets good and cold. And when you fill the container with water, the aluminum starts freezing the water from the bottom up due to conduction. If you were able to jiggle the walls every so often, that would prevent ice from forming over the surface.

IMHO, that will not work.

Just before water starts to freeze, it expands and becomes lighter. It floats to the top and freezes from the top down just the same.

dcarch

[slkinsey]

Interesting. I thought it expanded when it froze. But now I see that the expansion actually starts to happen around 4C.

[MaxH]

I see that the expansion actually starts to happen around 4C.

Yes, water hits its maximum density before freezing. (Molecular cause of that is enlightening, by the way; despite intuition, for materials to expand when heated is not fundamental -- it just works out that way most of the time.) Many (38?) years ago I heard technical information that water's maximum-density temp. was related to chilled food preservation, consequently to the nominal setting of refrigerators (not freezers) to 3-4 C (37-39 °F).

A problem with "bottom-up" freezing is that dissolved mineral impurities tend to be heavier than water, therefore to move downwards when they precipitate. The freezing process may still push them up, but gravity fights this, vs. the top-down freezing demonstrated in practice (upthread) where gravity inherently assists freeze purification.

Also, though it's just the kind of ritual many folk-alchemists would latch onto and then eloquently rationalize, I'm skeptical of real benefit from putting still-warm water into a freezer, as follows. Any cooling system works better (even beyond Carnot-type efficiency limits) when the desired temperature change is smaller, i.e., the water goes into the freezer as cold as possible. Hot water also evaporates an unusual volume of moisture into the freezer, which doesn't help your objective (but does lean on the auto-defrost feature). Why not let it cool first to room, then refrigerator, temperature? This just bypasses the period during which the freezer itself would be cooling the water down to those temperatures, before freezing begins. (Although I resist the hubristic conceit, including among scientists, that not understanding something implies it can't be so, still I recall no evidence in this thread that putting hot liquid directly into a freezer gives any benefit, vs. those side effects.)

[Luke]

Has anyone thought if polished water has an impact on ice clarity?

Fill a fish tank with water from your cold water tap. Note that the water does not look crystal clear. This is unpolished water. Leave the water in the tank for 1 week. Note that the water now looks crystal clear. It's called polished in aquarium terms.

So I just wondered if the problem is the use of water strait from a tap. Water frozen from a lake, or slow moving water, would be polished.

Luke

[bostonapothecary]

this thread seems to be reinventing the wheel a bit. the machines used to making clear blocks for ice carving are very simple. they basically fill a container with no special water, agitate from the top with pumps, apply cold "energy" from the bottom, and get all the results we are looking for. there are youtube videos if anyone is really curious.

there probably is no better way to do this. the question really becomes how can we scale this down in size with off the shelf parts.

do we need a temperature differential from the top to the bottom for it to really work, or is agitation at the top alone enough? they never freeze all the water but rather leave a few inches at the top. but what prevents this from happening? if they are dedicated to watching the machine they might just stop the process before the entire things freezes and the pumps become endangered.

if all we need is agitation, maybe we could find some sort of pump or stirrer that could be powered by the light bulb socket of the freezer.

i do have a small GE reach in freezer that i bought on sale for $120 to help with random catering jobs. i use it to freeze big cloudy & cracked blocks which i crush for the simple catered drinks.

i was wondering if i could find an off the shelf pump used to long run times like a hot water circulating pump and convert this freezer into a mini block maker replicating all the principles that makes the carving block machines work.

any ideas? i'm willing to drill some holes and spend a tiny bit of money.

[MaxH]

they basically ... apply cold "energy" from the bottom, and get all the results we are looking for.

Speaking as engineer rather than bartender, given that clear ice freezes most readily from the top down (for two separate reasons recently mentioned here -- also how it's done in past demonstrations and writings on the subject cited earlier); and further that normal heat flow direction is out of the top of things rather than the bottom, I continue to wonder why people suggest cooling from the bottom, other than convenience from some further factor.

In convection cooling (i.e., freezers with fans, which is most freezers), even if the cold source is below, I believe that if you measure or simulate, you'll find that much of the heat comes out the top of the water.