How do snowflakes form so symmetrically?

[Tom Mazanec]

I mean, how does a water molecule "know" that it is "supposed" to join the flake at the right corner of the flange sticking out of the first branch of the flake, to match the flange on the fourth branch of the flake 1 mm away? A millimeter may as well be a megameter to a water molecule.

[Swift]

The same way any molecule "knows" when it crystallizes - it doesn't. Given the intermolecular bonds of a given material, at a certain set of conditions, such as temperature and concentration of the crystallizing substance, there is a position that is the most energetically favorable.

[Perikles]

This doesn't answer your question, but last week I came across these amazing photographs of snowflakes. Presumably I can't reproduce them here for copyright reasons.

[agingjb]

It's still not obvious to me why each of the six branches tends to resemble the others on the same flake more than the branches on other flakes.

[HenrikOlsen]

The very specific conditions governing the growth of each arm more closely resembles those of the other arms in the same flake than those in other flakes because those are farther away?

[Rhaedas]

Rotation may be a key element too. Each flake as it drifts and spins subjects each side to similar conditions.

[Swift]

Good primer from Caltech on why snowflakes adopt specific geometries
There is a nice chart on that page that shows which specific geometry is adopted as a function of temperature and humidity.

Any further variation is from randomness.

[Paul Wally]

I mean, how does a water molecule "know" that it is "supposed" to join the flake at the right corner of the flange sticking out of the first branch of the flake, to match the flange on the fourth branch of the flake 1 mm away? A millimeter may as well be a megameter to a water molecule.

Interesting question. My guess is that the continued evolution of the the crystal is a function of the previous state of the crystal, like a recursive mathematical function or cellular automaton. The initial state or "seed" would be a random symmetrical state on the inter-molecular level. So it's not like the the one branch knows about the other, but rather that the same states must have the same next states.

[Trebuchet]

Slightly OT, but am I the only one who, as a small child, wondered who the heck had examined every single snowflake to determine that no two were alike?

[grapes]

Ha! You're still wondering, aren't you?

[Swift]

Slightly OT, but am I the only one who, as a small child, wondered who the heck had examined every single snowflake to determine that no two were alike?

I've thought, at least since college, that to be precise, the statement that "no two snowflakes are alike" is meaningless.

At some macroscopic level of examination, it is almost certainly incorrect, that at a coarse enough scale, that there are probably lots of "alike" snowflakes.

On the flip side, at a fine enough level of examination, because of random, atomic sized defects found in any material, that yes, no two snowflakes are alike, but then that could be said for any other group of things.

[trinitree88]

Notice the bond angles in the water molecules, which are always 108 degrees. The larger electronegativity of oxygen ~ 3.0 over that of hydrogen ~ 2.1, means that the charge of the electrons in the bond is shared unequally. that makes water a very polar molecule which explains parts of it's physics and it's chemistry. This allows cooling water to begin to form six member "cages" even before it freezes with a large interstitial space inside. This cause the density of water to increase as it approaches 0 degrees C....even before it freezes, reaching a maximum at ~4 degrees, then beginning to decrease as it expands as the "cages" form, resulting in a solid less dense than it's liquid....not the usual case in phase changes.
In Swift's paper the same electrostatics demand hexagons in sublimation with a six member ring structure dictating morphology of the final product....snowflakes.
p.s. It's nice that the universe is built that way, because if ice settled to the bottom of the oceans, the climate models would be a whole lot different and we wouldn't have evolved to be talking about it now. pete


SEE:http://www.google.com/imgres?q=six+w...r:0,s:36,i:195

[HenrikOlsen]

<snip> This cause the density of water to increase as it approaches 0 degrees C....even before it freezes, reaching a maximum at ~4 degrees, then beginning to decrease as it expands as the "cages" form, resulting in a solid less dense than it's liquid....not the usual case in phase changes.<snip>

But remember that those numbers are specific to pure water and will change significantly with the presence of impurities, including salt.

[Delvo]

Only the most symmetrical ones get chosen for use in pictures.

[Grey]

The very specific conditions governing the growth of each arm more closely resembles those of the other arms in the same flake than those in other flakes because those are farther away?

This is exactly correct. Exactly how an ice crystal grows is very strongly dependent on details like the exact temperature and humidity as it is forming. A snowflake forms relatively slowly, and will typically move around pretty chaotically in the storm cloud as it does so. All the parts of that particular snowflake share a common creation history, and so closely resemble each other (and of course the crystal structure itself limits the snowflake to certain shapes and angles), but another snowflake, even from the same cloud, will have taken a different path and experienced a slightly different formation process. Snowflakes from a single cloud will still have relatively similar histories, which is why they'll still resemble each other pretty closely, being big and fluffy, or small and icy, or whatever, while snow from different storms can be dramatically different. (I was curious about this myself some years ago, and spent some time researching the matter.)

[Selfsim]

Yet another example of the formation of a fractal in nature, whose outcome is highly sensitive to initial conditions, and whose formation process is chaotic.

Precise prediction of the outcome, at the scale of the arm detail, is not possible because the precise initial conditions cannot be retraced over time, in order to generalise a 'snowflake formation' formula!

All this unpredictability … in spite of the laws of physics and chemistry being determinable!

One instance of a myriad of similar phenomena in nature .. is life of 'em?

Regards

[HenrikOlsen]

In this case it's not initial conditions, but total history of conditions during growth. Unless you're looking at the cloud as a whole.

[slang]

last week I came across these amazing photographs of snowflakes.

Wow. Just wow. Thanks! (And wow again at some of the comments there... )

[John Jaksich]

As was pointed out by Delvo--not geometries are equal---I did attend a seminar a few years ago in which according to the author---due to increasing particulate (i.e. hydrocarbon soot from gasoline emissions) the crystals of snowflakes were not forming highly symmetrical crystals at all. I had posted this very thing a few years ago---I will attempt to find it among my past posts.

N.B.
But, then again, the finding may have been debunked by now?

[John Jaksich]

As a sidenote--the research of which I cite was conducted at a branch of the Desert Research Institute of University of Nevada Reno. Unfortunately--I am unable to find it, currently.


But I did find as interesting description as to the reason for some of the anomalous weather patterns being found currently

Here is the embedded link to the youtube video----approximately 5 and one-half minute description



youtube video here


tusenfem: Please do not direct link to youtube videos.