Do you frequent fancy bars? Well, then, you may have encountered the Colossal Cube.
Exacting bartenders from Los Angeles to Laval have begun chilling drinks with bergs massive enough to sink the Titanic. Typically served in beverages that call for an old-fashioned glass, the cubes often measure about five centimetres on each side, barely fitting into the glass. The Shangri-La Hotel in Vancouver uses them, as do Colborne Lane and Sidecar in Toronto and Wunderbar in Montreal. The theory is simple. Big cubes melt more slowly, resulting in a less-watery drink.
Or do they?
With a couple of university physics courses under my belt, I was skeptical. Warmth flowing into the drink from the room is the same in both cases. That should produce equal melting, notwithstanding the big cube's relatively smaller surface area. Heat in equals water out. Follow me?
It seemed to me that bartenders have been guilty of bogus science. Just because large cubes stay relatively large for a longer period of time, like a glacier versus icicles in the sun, it doesn't mean that they're not shedding a lot of water in an absolute-volume sense. It's a case of optical delusion.
A fan of the seductive clink of multiple small cubes, and no fan of nose-numbing big ones, I had to find out. That's where Doug Bonn came in. He's a professor in the physics and astronomy department at the University of British Columbia and a specialist in low-temperature physics and superconductors. I liked him immediately because he led me to believe that I wasn't entirely stupid.
"The first thing that you would think is, okay, in that situation, it's the amount of heat coming in from the environment that's going to affect how fast it melts," he told me over the phone. "The glasses are in the same environment."
But everybody knows that abstract physics and $15 won't get you a banana daiquiri in a swank bar these days. He kindly agreed to heed the thirsty call of barflies everywhere. Experiment time!
Using beakers, a scale and a fancy temperature sensor called a thermocouple, he set to work with ice and water. (I offered to pay for gin, but he assured me that alcohol's lower freezing temperature wasn't likely to change the results much.)
To exaggerate the potential difference in melting rates, he used a 7-to-1 ice-size ratio - one 35-gram cube, which is slightly larger than those of your freezer ice trays, versus seven five-gram cubes. That's an equal volume of ice in both drinks but a greater-than-normal number of smaller cubes in the second.
Assisted by PhD student Brad Ramshaw, he placed the cubes in 125 grams, or about 4½ ounces, of water each for an arbitrary 17 minutes. "It was limited by our patience, not a statement about how quickly you should drink a drink," he said. They measured the temperature above both drinks. It was 7 to 8 C. Nothing was left to chance.
The verdict? Ouch. The large cube shed 4.7 grams of water while the smaller ones shed a total of 7.7 grams, a 61-per-cent difference. A second trial using two large cubes and 14 smaller ones yielded a similar result. Big cubes make stiffer drinks.
So, I bow my head in shame, but not completely. Dr. Bonn says the physics is more involved than any bartender or designer-ice vendor probably realizes. He speculates that it paradoxically has to do with the fact that small cubes, thanks to more surface area, yield a more consistently colder beverage. (Anybody who has experienced brain freeze from an ice-chip-laden margarita knows that effect; a slushie tastes colder than bourbon on the rocks.)
Dr. Bonn's theory is that the colder liquid chills the glass walls to a greater degree, setting up a convection system outside the glass. Cold air is heavier than warm air, so it tends to fall, making room for hotter air above. That new, warmer air drives more heat through the glass and into the drink. It's a sort of passive version of the fan-driven action in a convection oven, which circulates hot air away from the heating elements and over your Thanksgiving turkey, resulting in a faster roast. (Only in this case we're talking about melting rather than cooking.)
"If the glass is a little bit colder, that convection runs faster and dumps more heat into the drink," Dr. Bonn said.
I'd like to stress, in partial defence of smaller cubes, that his experiment probably exaggerated the melting difference you would experience in most bars. To keep things tidy, he used ice-cold water and warmed his ice to zero degrees. Starting with warmer liquid, such as room-temperature gin, produces more - and an equal volume of - meltwater in both glasses, at least until the difference due to convection takes over. The small-cube glass generates a more turbulent convection system only after the glass gets really cold - colder than its big-cube counterpart. The same thing happens, for similar reasons, if you start with much colder ice, such as the cubes in your minus-20 freezer.
Dr. Bonn would like to stress something too: The convection theory is speculative. He could have sacrificed a few more lunch breaks verifying the hunch with more elaborate equipment. But he and Mr. Ramshaw had their superconducting day jobs to get back to.
That said, it wasn't all fun and games at the university's expense. They're considering building a first-year lab assignment out of the work. Call it a bar exam for science students.
"Real-world physics problems often are surprisingly subtle and complicated, and this is a spectacular example of that," Dr. Bonn said. Yes, all the chalkboard math in the world can't adequately solve certain vital questions without a bit of prodding and poking. You need a lab for that. Or at least a good watering hole.