Important factors affecting the stability of your liposomes are:

1) The thermodynamically stable phase for your lipid mixture and how far your liposomes deviate from it - if it is something else than a lamellar gel or liquid crystalline lamellar L-alpha phase, then it is more likely you will loose your liposomes. Particularly with charged lipids at low salt concentrations you have a myriad of metastable phases through the vesicle solution will develop through days, weeks, and months. Similarly, liposomes with mixture close to e.g. L-alpha -> HII phase transition will be less stable than those firmly in the lamellar phase.

2) Curvature elastic free energy - if your lipid mixture prefers lamellar phase (implying near zero spontaneous curvature), then more curved i.e. smaller vesicles fuse and break on the surfaces more readily (the total curvature elastic free energy for a spherical liposome is the same for every size of sphere, but the free energy per lipid is consequently much higher for smaller vesicles that have fewer lipids (as the number if lipids in the vesicle scales to the R^2, so will the curvature free energy per lipid). This is why supported lipid bilayers are made from (unstable) small unilamellar vesicles rather than large, more stable large unilamellar vesicles. This is probably the most important factor for your question. Area/lipid is more or less the same in all vesicles.

3) Chemical stability of your lipids. Unsaturated and particularly polyunsaturated lipids will oxidize over time. (With DHA-lipids people need to store and handle them under argon gas.) Oxidized acyl chains (typically in the position 2 of the glycerol backbone in the natural lipids because this is where the unsaturated chains tend to be) tend to loop back to the lipid/water interface, so that their shapes favour micellar phase, so when their concentration becomes too high, your liposomes will become unstable. Saturated acyl chains better stand the test of time. 

The transformation from liposomal to true lamellar phase at fridge temperatures for saturated phosphatidylcholines in solution is extremely slow particularly for the multilamellar vesicles that still are in the liposomal phase after four YEARS of incubation in the fridge. (This is longest I have seen.) For LUVs it is rarely a practical matter either, unless you want e.g. long term storage of a drug-releasing liposome for clinical use (then use PEG-lipids in the mixture) and unless you prepare them in sterile conditions. They are anyway good for a week or two, though if it of particular interest to you, you could use The Malvern Zetasizer by Alan's company (excellent instrument with excellent software) to check if the size distribution remains stable over time. 

Good question, Riccardo.  What do you mean by 'stability'? Are you referring (as I think) to propensity to aggregate?  Or to settling?  Or to reactivity?  As particle get smaller attractive forces become large in comparison to inertial forces.  See:

November 3rd, 2015 Adhesion and cohesion  http://tinyurl.com/zwb2wlh

Thank you for your kind answer, Alan; well, rather than aggregation (that, to me, is an interesting aspect, too), I was referring to propensity to swell and/or break down and become weak.

As size decreases then the surface area increases as the square of the size (1/2 the diameter is 4 times the surface).  In general terms this means that reactivity increases to a similar extent.  Depending on the kinetics of hydration/swelling then one would expect the smaller system to absorb water more rapidly - that is a kinetic effect.  If it's rapid then I suspect little difference would be seen. Thermodynamics would tell you if the system was more or less stable.  I would hazard a guess (and a guess it is only) that like-for-like there'd be little difference in stability.  However, in order to (for example) charge stabilize smaller systems then they require a higher concentration of stabilizer (e.g. phosphate for inorganic oxides and other materials) to give a monolayer of the stabilizer.  Smaller systems need higher levels of ionic (more charge) or steric (more coating) stabilizer to confer stability from a negation of van der Waals forces, so one could argue that like-for-like systems that the smaller systems would be less stable.

Important factors affecting the stability of your liposomes are:

1) The thermodynamically stable phase for your lipid mixture and how far your liposomes deviate from it - if it is something else than a lamellar gel or liquid crystalline lamellar L-alpha phase, then it is more likely you will loose your liposomes. Particularly with charged lipids at low salt concentrations you have a myriad of metastable phases through the vesicle solution will develop through days, weeks, and months. Similarly, liposomes with mixture close to e.g. L-alpha -> HII phase transition will be less stable than those firmly in the lamellar phase.

2) Curvature elastic free energy - if your lipid mixture prefers lamellar phase (implying near zero spontaneous curvature), then more curved i.e. smaller vesicles fuse and break on the surfaces more readily (the total curvature elastic free energy for a spherical liposome is the same for every size of sphere, but the free energy per lipid is consequently much higher for smaller vesicles that have fewer lipids (as the number if lipids in the vesicle scales to the R^2, so will the curvature free energy per lipid). This is why supported lipid bilayers are made from (unstable) small unilamellar vesicles rather than large, more stable large unilamellar vesicles. This is probably the most important factor for your question. Area/lipid is more or less the same in all vesicles.

3) Chemical stability of your lipids. Unsaturated and particularly polyunsaturated lipids will oxidize over time. (With DHA-lipids people need to store and handle them under argon gas.) Oxidized acyl chains (typically in the position 2 of the glycerol backbone in the natural lipids because this is where the unsaturated chains tend to be) tend to loop back to the lipid/water interface, so that their shapes favour micellar phase, so when their concentration becomes too high, your liposomes will become unstable. Saturated acyl chains better stand the test of time. 

The transformation from liposomal to true lamellar phase at fridge temperatures for saturated phosphatidylcholines in solution is extremely slow particularly for the multilamellar vesicles that still are in the liposomal phase after four YEARS of incubation in the fridge. (This is longest I have seen.) For LUVs it is rarely a practical matter either, unless you want e.g. long term storage of a drug-releasing liposome for clinical use (then use PEG-lipids in the mixture) and unless you prepare them in sterile conditions. They are anyway good for a week or two, though if it of particular interest to you, you could use The Malvern Zetasizer by Alan's company (excellent instrument with excellent software) to check if the size distribution remains stable over time. 

 P.S. If the aqueous compartments are the same both inside and outside liposomes, they have no tendency to swell (although they will of course swell a bit although for unilamellar vesicles very little, if you have trapped osmolyte inside the liposomes but not outside, although they will break it the concentration gradient is too high), but fusion may occur, i.e. the liposomes may grow in size but only by growing fewer in number. Unless they are oxidized or hydrolyzed by alkali, for instance, or they are inherently kinetically unstable (they all are thermodynamically), they will not become "weak".

Hyvää huomenta, Juha-Matti! Kiitoksia paljon erinomaisesta vastauksestasi!  Kuuluuko hyvää?

Riccardo, Juha-Matti's answer  is excellent and I commend you to read it in detail.  I will now up-vote it (would like to have 5 stars here rather than a single binary '1')

Interesting answers.

Juha-Matti makes an important point on curvature.  A similar effect happens with cavitation in sonication - the radius of curvature is so tight that the generated ultrasound 'bubbles' immediately collapse at around 1011 g (where g is the acceleration due to gravity) generating heat, sound, and light (sonoluminescence).  For more entertainment see (Katso! - Juha-Matille):

June 20th, 2013 Ultrasound, cavitation, and the singing kettle  http://tinyurl.com/olueohz

Nov 11th, 2008. Dispersion and nanotechnology http://tinyurl.com/hpywsge

agreed by Juha-Matti. Very interesting

Dear Riccardo

I will suggest you to read following articles

http://www.ijper.org/sites/default/files/IJPER_45_4_13.pdf

http://www.ncbi.nlm.nih.gov/pubmed/1979618

Thank you everybody. This is all very interesting.

Hi,

Generally speaking, the smaller the size of a sphere, the more the curvature. Therefore, we expect less stability. Have a look at this manuscript:

Size and structure of spontaneously forming liposomes in lipid ... - NCBI

https://www.ncbi.nlm.nih.gov › NCBI › Literature › PubMed Central (PMC)

by M Rovira-Bru - ‎2002

-  -  -

With respect of liposome stability:

Stability of liposomes can be improved by coating the surface of vesicles

with certain polymers, such as polyethylene glycol (PEG) or chitosan, or by using

cationic or anionic ingredients in the structure of liposomes, or employing stabilizing

agents such as cholesterol or glycerol (Mozafari, 2005). Generally, it is desirable to

maximize the repulsive forces between the vesicles in order to keep each vesicle

discrete and prevent them from gathering into larger, faster settling aggregates. This can

be achieved by preparing vesicles of the same charge that repel each other electrostatically.

The application of similarly charged ingredients in the liposome structure reduces

aggregation, fusion, and precipitation of the vesicles and results in increased stability of

the liposomal suspension. This is because lipid vesicles are thermodynamically

unstable. Both liposomes and nanoliposomes will aggregate, fuse, flocculate, and

precipitate during storage. Increasing interparticle repulsion, either electrostatic or

steric, can enhance their stability. In general, particles with zeta potential values greater

than +30 mV or lower than –30 mV possess long-lasting electrostatic stability (De Luca

et al., 2006).