To add to your survey:
Yes I have tried it. Yes it does work "better" for glassifying cryobuffers.
I saw Rob Thorne give a talk on this some time ago and immediately went
home to give it a try. I used a cryobuffer (25% (v/v) glycerol; 0.6N
NaOH; 25 mM SeMet) I had been having hit-or-miss ice/amorphous results
with what I thought was systematic plunge cooling into liquid N2. The
tricky bit was that the cooled drop was always optically transparent,
but the diffraction pattern reveals if it is amorphous (diffuse rings
at: 3.57 and 2.08A) or nanocrystalline ice Ic (sharper rings at: 3.67,
2.247 and 1.917A)
The long and short of it is that by filling my flash-cooling dewar all
the way up to the brim and (gently) blowing the gas layer away (with my
breath) pretty much ALWAYS gave me a amorphous droplet (~150um thick).
You can't blow too hard or you will blow the drop out of the loop
before it hits the lN2. For the same solution, cooling in the cryo
stream ALWAYS gave me a highly nanocrystalline droplet. A singular
result, yes, but still an example.
The droplets looked the same by eye, but the diffraction patterns were
different. This differnece in solvent structure turned out to be
important for my experiment at the time (J. Syncrotron Rad. (2007) 14
51-72). In that work I had to "sort" droplets based on how they
diffracted. It is now nice to have the Warkentin et. al. technique (and
citation) to get 100% amorphous droplets when I want them.
It is perhaps interesting to note that at higher %glycerol (concentrated
by letting the droplet evaporate a bit), I can occasionally get a
stream-cooled droplet that is amorphous on the leading edge, but
nanocrystalline on the leeward side (opposite the nozzle). You can
optically see the boundary between the two transparent substances in
these drops. The nanocrystalline ice has a bumpy surface, whereas the
amorphous solid water has a smooth surface. Don't really know why that
is. I suppose the outer skin of the droplet can be expected to be
amorphous so there is a final-density mismatch between the amorphous
"skin" and the nanocrystalline core, causing the "skin" to shrivel.
Personally, I suspect that a similar mismatch in the final density of a
cooled protein crystal lattice and the cryobuffer in its solvent
channels is probably a major reason for "bad freezes".
Indeed, cooling to an amorphous solid is not always the best way to
cryopreserve a protein crystal. I do have one example case of a crystal
that consistently diffracted better (1.75A) when the cryo was cooled to
be (partly) nanocrystalline ice Ic, but not as well when cooled to an
amorphous solid (2.2A). Again, a singular result, but an example
nontheless.
-James Holton
MAD Scientist
Florian Brückner wrote:
Dear everyone,
has anyone tried removing the cold gas layer above liquid nitrogen (or
propane) by blowing or sucking to improve cooling rates in the
cryocooling of crystals. I would like to try and I would be happy if
someone could share some experience.
Florian.
P.S.: there is an interesting paper: J. Appl. Cryst. (2006). 39, 805-811.
