I apologize in advance for the length of this post, but this is such an interesting topic to mull over. I'll stating up front that these are just my speculations. I haven't delved into the literature to verify whether anything here is true or false, so YMMV.
Here is some initial food for thought:http://www.chemguide.co.uk/qandc/ptfe.html
I encourage you to read that before you read what I have to add.
The author of that webpage points to an apparent discrepency between what many authors say about the low magnitude of PTFE intermolecular (van der Waals) forces and the high melting point of PTFE (~327 C, compared to, say regular old low density polyethylene at around 105-115 C). Why is this an apparent discrepency? Well, let's look briefly at the two facts:
(1) Most websites and other popular literature sources point to the very low magnitude of van der Waals (London dispersion) forces in PTFE as being responsible for the nonreactivity and low friction coefficient of PTFE surfaces
. Many like to point out that PTFE is the only substance to which a Gecko's toes won't adhere.* The stickiness of a Gecko's toes is commonly attributed to van der Waals forces between the high-surface-area toes and the surface
that the Gecko is trying to climb. I don't think we have any reason to doubt this latter fact (about Geckos). In turn, the lack of susceptibility of the PTFE surface
to instantaneous induced dipole moment changes - that is, the lack of van der Waals forces at the PTFE surface - is usually attributed to the very strong C-F single bond, and the extremely low polarizeability of fluorine. Fluorine is so electronegative that it holds its electrons very close to the nucleus, which lowers the probability that some other nearby dipole can temporarily pull an electron farther away. Hence the "nonstickiness" of perfluorinated polymer materials.
Sounds good, right? Uh-oh:
(2) PTFE has a very high melting point. The temperature at which a phase change for a substance occurs is well known to all chemists to be highly influenced by the amount of intermolecular forces holding it together, and typically a higher phase change temperature means MORE and STRONGER intermolcular forces - all things being equal of course.
So, if the non-stickiness of PTFE is attributed to very low magnitude intermolecular forces and the high melting point of PTFE is attributed to very high magnitude intermolecular forces, something clearly ain't right.
The author of the article I linked to solves this discrepency by essentially rejecting (1) completely. From the horse's mouth - er, fingers - "However . . . several web sites talk about PTFE having very weak van der Waals forces. That has got to be completely untrue.
If it had very weak van der Waals forces, it would be a gas - not a fairly high melting point solid!" [Emphasis not mine.]
On the surface this seems reasonable. After all, the melting point is an actual piece of indisputable data that is hard to rationalize without assuming that the intermolecular forces in PTFE must substantial in magnitude.
However, I maintain that this discrepency is due to an inappropriate comparison of bulk and surface properties, as well as a need to explore in more detail what gives rise to the high melting point of PTFE. Let me address these one at a time.
(A) Phase changes and intermolecular forces.
The temperature at which a phase change occurs for a substance is due to the strength of the intermolecular forces that hold the substance together. For boiling points, applying this rule is usually pretty simple. You look at the structure of the molecules making up the substance, determine whether there are likely to be dipole-dipole interactions, hydrogen bonds, or London forces, and make some arbitrary but straightforward assessment of the overall strength of these interactions based on the magnitude and number
of the forces identified as important. It's typically understood that a higher molecular weight corresponds to higher boiling point, for example, because higher molecular weight substances tend to have more overall molecular-scale surface area - more places for molecules to stick together due to transient dipoles. (Many students assume higher weight substances are harder to boil because of gravity, but this couldn't be more wrong.)
Melting points, while still generally dependent on intermolecular forces, are not so straightforward to explain. No longer can we just look at the constituent molecule and make an assessment of the relative strength of intermolecular forces expected, because the way a solid is put together matters. The crystal structure, density, particle size, and so forth all make a huge difference on bulk properties, even among solids which are compositionally identical (and moreso if they aren't). I don't want to reinvent the wheel here, so let me link to an excellent description about how molecular shape and symmetry, in addition to the usual electronic factors, play a role in dictating the strength of intermolecular forces of a solid, and the resulting melting point: http://www.masterorganicchemistry.com/2010/07/09/chemical-tetris/
Essentially the point that is being made there is that while two substances may be chemically similar with respect to molecular polarity, the number of relevant intermolecular forces in the solid state can vary substantially simply because of the way the molecules pack together. A branched alkane, for example, will have a higher melting point than a linear alkane of the same molecular weight due to the effeciency of packing in the former: linear alkanes pack much more closely together than branched alkanes, which means the shared surface area is larger (or, the average distance between molecules is smaller). Strength of intermolecular interactions is due not only to relative differences in charges (permanent or transient) but also to the distances between them. If you compare hexane to 2-methylpentane, the reason the latter melts almost sixty degrees colder is because due almost wholly to packing efficiency.
The same kind of logic can be applied to polymers. Low density and high density polyethylene differ primarily in the extent of branching along the polymer backbone. HDPE has less branching and as a result packs together much more efficiency (it is higher density). This translates into better shared surface area between adjacent polymer strands, and in turn a higher melting point (~120 to ~130 C) compared to LDPE (~110 C), which packs less efficiency and has less shared surface area between adjacent strands. Molecular weight, too, plays a role. Ultra high mol. weight polyethylene (UHMWPE) has an even higher melting point (~135 C) because higher weight strands have a higher degree of tangling and such that increases the relative surface area even more. All these differences in a series of polymers that are nearly identical electronically (i.e., as far as polarity and polarizability is concerned).
The relevance of all this to PTFE is as follows: the melting point of PTFE is 327 C. Corresponding melting point of LDPE is about 110 C. What has stronger intermolecular forces? The clear answer is PTFE. That's kind of indisputable. But it's pertinent to ask why. Fluorine is heavier than hydrogen and thus PTFE strands might be assumed to be higher molecular weight than LDPE. But that's neither here nor there. The real reason is mass and electron density. The specific gravity of LDPE is ~0.92 and the specific gravity of HDPE is ~0.95, accounting for the slightly higher melting point of the latter. The specific gravity of PTFE is a whopping 2.2. Granted, the molecular weight of fluorine probably has a lot to do with that, but so does the packing state. PTFE has so few free volume holes because of the way the individual polymer strands join together, deriving predominantly from fluorine sterics and fluorine-fluorine repulsion. Let me quote from the original article linked to above: "The [fluorine-fluorine] repulsions lock the [PTFE] molecules into a rod-like shape with the fluorines arranged into very gentle spirals - a helical arrangement of the fluorines around the carbon backbone. The rods will then tend to pack together a bit like long thin pencils in a box." So, what this means is that even if the C-F bond is very strong and fluorine very resistant to polarization, the fact that PTFE is so dense and the polymer strands organized so tightly and regularly ensures a high surface area contact between strands. I.e., even though fluorine holds its electrons tight, the fact that everything is crammed so close together in PTFE greatly increases the frequency of transient dipole interaction, and thus gives rise to a high melting temperature.
This is all basically a round-about way of saying that yes, the intermolecular forces in PTFE are large, but they are large because of frequency of occurance and packing structure of the solid - that is, the intermolecular forces are strong despite
the low polarizability of fluorine. This gives us a perfect way to rationalize the apparently contradictory behavior that occurs at the PTFE surface
(2) Surface properties vs. bulk molecular properties.
Several times now I've italicized the word "surface", and I've done this for a reason. What happens at the surface of a solid is a function of the respective free energies of the solid and whatever the solid is interfacing with (even if it's just air). The apparent discrepency that's been discussed above is rooted in the fact that two very different processes are under consideration: cohesion vs. adhesion. Because of tight packging, as we've seen, PTFE has a high energy of cohesion. But, this still doesn't change the fact that the fluorines in PTFE are not very sticky, on a per fluorine basis. At an interface, where packing density perhaps is not so important, the inherent stickiness of the substance is predominantly responsible for determining the magnitude of adhesion when another substance comes into close contact.
In fact, and this is just conjecture, it may be argued the packing density that gives rise to high energy of cohesion in PTFE directly contributes to the low energy of adhesion and especially the low coefficient of friction, due to mechanical effects as well as a higher available surface area for adhesive bonding. If there is a high packing density, this would (I would think) give rise to a smoother, more uniform surface. Thus there would be less surface area available for bonding at the interface, certainly in the event that the bonding substance is a liquid or other substance - like denaturing proteins in a cooking chicken breast - that can conform to whatever geometry is available. Rougher surfaces take more energy to move parallel to each other under an applied perpendicular force - there's your friction aspect. (I believe that Teflon pans lose some of their stickiness as they age due to formation of micro-scale pits and other physical surface defects.)
Even beyond that - let's pretend for a moment you had a perfectly flat surface made of PTFE and a perfectly flat surface made of LDPE. We can agree that the PTFE material would make a good pan and the LDPE would not because PTFE has a higher melting point owing to its much greater polymer packing density and corresponding strength of intermolecular forces. LDPE has a correspondingly low force of cohesion between its molecules and falls apart at a pretty low temperature - this despite the fact that the C-H electrons on LDPE are inherently more sticky (polarizeable) than the C-F electrons in PTFE. There is simply a lot more C-F's bumping up against C-F's in PTFE than there are C-H's bumping up against C-H's in LDPE.
On the other hand, at the interface of our ideally flat, structurally identical surfaces, the PTFE material I think would have roughly the same amount of C-F's available for forming London bonds as the LDPE has C-H's available. What we do here is essentially normalize the materials for density of bonding area, and the sticky LDPE wins out now by a fair margin when it comes to adhesion potential. (And in fact, we err on the conservative here, if my conjecture about the roughness of the LDPE vs. PTFE surface is correct - a real LDPE surface would have a higher surface for London bonding than an PTFE surface of the same area). This is somewhat borne out in the numbers: despite its higher melting point, PTFE has a lower surface free energy (19.4 dyne/cm) and higher water contact angle (109.2 deg) than polyethylene (31.6 dyne/cm, 96 deg), polypropylene (~30.5 dyne/cm, 102.1 deg), polystyrene (34 dyne/cm, 87.4 deg), and polybutadiene (29.3 dyne/cm, 96 deg) (source: http://www.accudynetest.com/polytable_03.html
). Notably, the four comparison polymers are all exclusively C-H containing polymers and thus might be expected to have superficially similar degrees of stickiness despite having very different melting points and other bulk molecular properties, and this is what the numbers kind of show, at least when compared to PTFE. Granted, this is probably a simplification of what is going and it would be easy to nitpick against this argument, particularly since things like contact angle also depend on the cohesive properties of the liquid being dropped on the substrate.
Anyway, to sum: in the bulk, PTFE forms weak intermolecular forces due to the high polarizeability of fluorine, but the dense packing in the solid state ensures that it forms a large number of them, which translates into a high melting point. The cohesive forces holding PTFE together are very strong. However, at the interface between PTFE and a binding substrate, I think this packing density may not be as important, and the inherent resistance of fluorine to sticking creates a low energy of adhesion compared to other materials. That is, the intermolecular forces between the PTFE surface and another substance at the interface are especially weak. The tight packing of PTFE polymer strands may also give rise to a more uniform, smooth surface which limits the extent of mechanical adhesion and provides less surface area for dispersive adhesion compared to other solids.
* I'm not sure if that's the case; there are a few substances with a lower coefficient of friction than PTFE.