A while ago, I came across this project on the Make website. The author had done a really nice job of milling two blocks of aluminum to create a spherical chamber. A couple guide holes and pins help with alignment. The purpose is to take advantage of the neat phenomenon when you bring a big block of warm aluminum in contact with ice.
You put ice in contact with a prewarmed block of aluminum and it starts to melt. So far, so what? With most materials, the ice starts off melting quickly, but before long, the surface of the block has cooled and the melting just about stops. Since aluminum is so thermally conductive, more heat can travel from the interior, and the melting barely slows down. That's why this trick won't work with wood or plastic, even though they may start off at the exact same temperature.
The other day I came across roughly the exact same design (minus the drainage hole) over on Williams Sonoma. It's pretty steep at $200. (edit: Just in case that's not sufficiently first class for you, it's been brought to my attention that Williams Sonoma also sells this version. $700 for a 2" sphere and $1100 for a 2 3/4" sphere) That's a lot for a pretty basic party trick, but then again it is a big block of aluminum that's been precision milled and anodized. Here's the video of it in action.
Ok. Time for my rant. There are a bunch of ways of making spherical ice cubes, as I'll get into in a bit, but why would you want to? Both the above sites make the same argument. The starting fact is that a sphere is the shape with the lowest surface area to mass ratio. The conclusion they then jump to is that spherical ice cube will
keep drinks cold longer and …
dilute the drinks less
To me, it's a prefect example of starting with some absolutely correct facts and then getting it completely wrong. Unfortunately, this is not uncommon when engineers move outside their normal areas of expertise. I could go all mathy on you and pull out the equations, but let's just stick to the variables. The value that gives us ice's coldification is it's heat of fusion. One mass unit of ice will absorb so many units of heat from the drink when it melts. The drink will have a specific heat capacity, which is a measure of how many heat units you need to remove in order to lower one mass unit of drink by a degree.
You'll notice I'm not using numbers because it doesn't matter. What matters is what is missing, namely any quantity to describe the shape of ice. X milliliters of drink will require Y units of ice to melt to get it's temperature from Tyucky_warm to Tyummy_cold. The equations are easy to find and easy to do if you're just talking about water. The details get more complicated if you start adding things like sugar and alcohol that change the heat capacity of the drink and the effective melting point of the ice, but you'll notice I still haven't mentioned shape. It just doesn't matter.
So using the aluminum ice mold thingy makes no difference.
Well… not quite. The single review (one star) on the Williams Sonoma site pointed out the down side. The ice you take out of the freezer isn't at the freezing point, it's colder. When you drop it in your drink, it can pull a certain amount of heat out of the drink before it even starts to melt. This is "free" cooling without any dilution. When you "melt carve" the ice to turn a cube into a sphere, you're starting with ice whose surface has already been warmed to its melting point. You'll actually get more dilution than if you just dropped the cube in uncarved.
So why would you want spherical ice? Well it's cool, that's why. Fortunately there are much cheaper and easier ways to do it by using spherical molds. I'd recommend a set made of silicone. It's flexibility and durability can't be beat when you're releasing the ice, and a set of 4 comes in at 1/20th the price of a single melt carver.
If you read the reviews of the molds, you'll see some complaints about what the ice looks like when it comes out. That's not the fault of the mold, but the water. Getting clear ice is tricky. You might not care so much with a dinky little cube, but when you're making big 2" diameter balls, you'd really like it to be nice and clear. Starting with low mineral content water helps, but dissolved gasses is the biggest culprit. For best results, put the water under vacuum for at least 2 minutes before putting it in the freezer. Better yet, freeze it in a chilled vacuum chamber.
What, you don't have rough pump and cold trap in your kitchen? Ok. Get an air stone from the pet store and bubble helium or hydrogen through the water for a 5 minutes before you freeze it. It's a neat and counter-intuitive trick. Helium has a very low solubility in water, and all the tiny bubbles provide lots of surface area for the other gasses to escape.
No He tank? It's like you guys live in a mud hut or something.
Ok. Looks like boiling is your best bet. Boiling before freezing will remove most of the dissolved gases and give you nice unblemished balls. It's not exactly carbon neutral, but after a few drinks some of your friends will be carpooling home.