We all learned in high school chemistry (and some of us had to re-learn it at university) that water is weird for a lot of reasons, and its weirdness makes life possible. One of the strangest properties about water is that when it cools below its freezing point, instead of contracting and densifying as most materials do, it becomes less dense, and it suddenly expands. Lots of things would be very different if water was like most other materials. That there is probably a cosmic significance to this oddity is a temping digression I shall resist.
Water expands about 9 percent of its unfrozen volume when it changes to ice. If restrained by, say, the cast iron casing of a Goulds jet pump (or the coolant channels of a cast steel engine block), the expansion exerts a pressure on the walls of whatever is trying to contain it. The resulting pressure is enough to split metal clean in two. When freezing water is present in the intergranular (pore) spaces of a soil body, it will expand into air voids, or, if voids are full of water, it will act to make them larger by forcing individual soil particles apart, which can cause the soil to swell. When water in soil freezes, the associated expansion is commonly called frost heave. When a soil body thaws, and water shrinks by 9 percent, the soil body tends to return to its thawed volume. If the voids were mostly full of air as the soil froze, it will likely return to near its thawed void volume and both frost heave and thaw settlement will be small. The combined effects of frost heave and thaw settlement are collectively known as frost action.
Building foundations supported by soil which is subjected to ground freezing can be damaged by frost action. The most common damage results from frost heave. Most damage to conventional steel-reinforced concrete foundations results from heaving stresses acting below the foundations which can lift the building unevenly and cause bending and cracking. Classical soil mechanics (i.e., geotechnical engineering) is the branch of civil engineering that deals with how soil responds loads to the presence of water in liquid form. Cold regions soil mechanics (i.e., frozen ground engineering) is a specialized subset of soil mechanics that deals with the effects freezing and thawing water in soil. With regard to frost action, soils are generally characterized by susceptibility to changes in engineering behavior that arise from ground freezing. All soils containing water are susceptible to changes in their engineering characteristics arising from ground freezing. However, some soils can also expand (heave) large amounts. For soil to be frost susceptible, it need only contain water. For soil to be susceptible to the most heave and by association, the most risk of differential movement and damage to building foundations, three basic components must be present. It helps to think of them in terms of the "three W's."
1. Winter.
2. Water.
3. Wickable soil.
Like the old-school fire triangle, if you kill one of the
three sides of the triangle you kill fire - get rid of any of the three Ws
and you kill frost action.
Winter is easy to kill when it comes to
conventional buildings. Just bury the foundations sufficiently deep and/or heat
the building. Or use a combination of heat and artificial insulation (for
shallower monolithic slabs). For deep foundations in cold regions
supporting warm buildings that are elevated and the space beneath remains
unheated through the freezing season to preserve nasty stuff like permafrost, a
bond breaker (or reducer) is usually applied to each foundation (hint: NOT a
plastic sleeve).
But for utility scale solar, where we have thousands of
acres and many tens or even hundreds of thousands of foundations (I’m looking
at you 1P), heating is definitely out and killing wickable soil is easier said
than done (but sometimes it is the only way in some situations, though mostly
to kill fear, and not frost action). Of the three, water is the most
difficult to kill. Where it is present within or near the zone of ground
freezing and where bitterly cold winters happen more often than not, we’re
extra careful.
To recap and to add detail (deep breath) to frost jacking,
in plain terms adfreeze is the bond between the soil and the foundation
material. Lens ice growth is the strongest driver mobilizing the adfreeze
bond stress and the resultant is a tangential (jacking) force. “Adfreeze”
by itself without lens ice growth (and continued development through the
freezing season, (you don’t just wish all the freezing season’s frozen ground
into place at once)) is a potential stress, at least in the up direction.
True, it can also be a stress normal to the foundation material, but we’ll
simplify things for now, and only talk about the thing we most fear – the up
direction. Without accompanying lens ice growth (in the up direction), it
is stress without strain. Like grabbing hold of a broomstick and gripping
it with (some to) all of your strength. But don’t pick up. True,
lens ice, when it grows under the right conditions and with the right
orientation can mobilize the tangential bond stress and impose a force in the
up (and down) direction. And true, this force, if not considered, can
displace foundations embedded in it. But at this moment, it is only
partly frozen and by virtue of simultaneously being partly unfrozen (it is in
the process of freezing) it is not very strong. So, unless it is growing
under very cold conditions (these conditions diminish rapidly with distance
from the ground surface in temperate climates where the subsurface ground
temperatures are well above freezing (say, 50F, year round)), any ice under load
creeps like a bad dog and smears and fails itself at the interface as the
freezing front grows incrementally. Each incremental freezing layer
suffers this same fate and this must be considered. Simply wishing the
frozen ground into place and applying a late season ice lens at the bottom is a
gross oversimplification of what’s actually happening.
As to the adfreeze bond itself, that’s a long, conversation
(insert beer here) guaranteed to bore your friends. It can’t really be
killed when subfreezing temperatures and water are present in any kind of
soil. But it can be resisted or, when necessary, mortally wounded.
It is strongest in granular soils (e.g., sands which may contain appreciable
fines and water and still expand 9 percent on freezing (that’s 9 percent of the
volumetric water content, not the total volume) but you’d have trouble growing
good lens ice in dirty granular material (GM, SM, SC) not in temperate climes
anyway) even if you wanted to. On the flip side, the adfreeze bond is not
as strong in fine grained soils (especially those containing colloids) at
comparable temperatures. This is partly because at temperatures between
freezing and say, -5C (23F), fine graineds contain decreasing but appreciable
unfrozen water which weakens the bond (think soft ice cream). Lens ice
grows best in layered fine grained soils (ML, CL, CH). Note, that lens
ice need not be horizontal, and in CH soils, it often is vertical. Water
is a snob when it freezes, and it likes to be pure, forcing out impurities that
look different than it. Water molecules present in the intergranular
spaces of clean, coarse-grained materials have little trouble finding each
other in a crowd and congregate into party formation (i.e., freeze) at
comparatively high temperatures (below freezing), and most of them present at
the nucleation site freeze completely at or just below their pure freezing
point. In similar fashion, water molecules present in the intergranular
spaces of fine-grained soils (especially colloids) also prefer to be near each
other, but they have to form a conga line and make their way around tiny soil
particles to congregate into party formation. In a fine-grained
soil, as more water molecules are attracted to the nucleation point, and
effectively expel and leave fine soil particles behind, you get segregated
frozen water. A special case of segregated frozen water is lens
ice. As an aside, the migration of water molecules towards the nucleation
point (and the freezing front) also ends up desiccating the soil behind it,
actually enhancing capillarity. In very cold climates (subarctic to
arctic) where the soil surface temperature may be -70F or colder, some
mistakenly think that water gets an extra push upward against gravity by the
Second Law of Thermodynamics. But that’s probably an old Engineer’s Tale.
More likely, if there is an extra push, it’s water at the phreatic surface more
readily vaporizing across a steep temperature differential (vapor transport)
and traveling through air gaps present in the intergranular spaces
(unsaturated) to help to fill a vacuum left in an unsaturated soil skeleton
created by the water moving towards the freezing front and away from the
nonfrozen soil. SIDENOTE: clay can and usually does have such a low
hydraulic conductivity that the conga line that water molecules need to follow
to get to the party won’t allow them to congregate and lens ice in clay soils
can freeze in all kinds of weird orientations. So as you can see, the
proverbial frost line is more of a frost front and lots of weird phase changes
are happening in that zone all at once. This is what can and does make
frozen ground so bizarre (and ground ice so weirdly neat).
At the end of the day (days in the north are three months
long, so we have time), there’s a lot going on that makes water molecules want
to segregate themselves into ground ice (pore ice, lens ice, or wedge
ice). See ASTM D4083 (attached) for a more of a visual on segregated
ground ice. Vs is the kind of ice we most fear, but mercifully, if it is
a real problem, it is cyclic and because of this, it is simultaneously the
easiest for drillers to spot during the SI program if drilled during or
immediately following a freezing season if they know what they’re looking
for. But segregated ice doesn’t equal lens ice and segregated ice by
itself doesn’t drive the kind of frost action that terrifies us as
engineers. Put another way, defining a soil as frost susceptible or
non-frost susceptible only begins to tell the story.
