Phil Attard contemplating nanobubbles while completing his first marathon, (Adelaide, 1999). Note that due to an unfortunate misunderstanding about the start time, the actual running time was 3:03:09.


Experimental Evidence
Practical Implications
Thermodynamic Stability

See also here for a brief discussion of nanobubbles, colloid science, and surface forces.

Experimental Evidence

Nanobubbles are tiny gas bubbles that have been found on some surfaces in liquids. Typically composed of atmospheric gases, they occur on hydrophobic (water repellent) surfaces in water. The existence of nanobubbles has also been inferred in other solvents, (in this case the surfaces are called lyophobic). The nanobubbles are from 10-100nm in diameter, and, although they appear to be forbidden by thermodynamics, (see stability below), their lifetime is at least of the order of hours. They likely arise by nucleation of gas at the surface from a supersaturated solution.

Nanobubbles appear to be responsible for the long-range attraction that occurs between hydrophobic colloid particles and surfaces. Nanobubbles bridge apposite surfaces and draw the particles into contact, leading to a capillary adhesion that makes aggregation of such particles practically irreversible. Nanobubbles also appear to decrease the drag on water flowing over hydrophobic surfaces and particles and through hydrophobed capillaries.

These images of nanobubbles were obtained by Dr James Tyrrell of the University of South Australia using a scanning probe microscope in tapping mode [J. W. G. Tyrrell and P. Attard, Phys. Rev. Lett. 87, 176104 (2001)]. The first image is the phase representation (10 degrees peak to trough) and the second is the corresponding height profile (30nm peak to trough). The nanobubbles are irregular in shape, are highly interconnected, and form a network that almost completely covers the underlying hydrophobic surface.

Nanobubbles were first proposed by Dr Phil Attard, then at the Australian National University, to explain certain features in the force measurements performed by Dr John Parker at the Institute of Surface Chemistry, Stockholm [J.L. Parker, P.M. Claesson, and P. Attard, J. Phys. Chem. 98, 8468 (1994)]. Prior to Parker's work, a number of laboratories had reported measurements of a long range attractive force between macroscopic hydrophobic surfaces in water, and there had been much debate as to the origin of the attraction. Parker's high resolution data revealed steps or discontinuities in the force at separations as great as 300nm, (see figure at right). Attard suggested that these signified the bridging of bubbles between the surface. The bubbles were thought to exist on one surface and to attach to the other when it was brought within touching distance. Thermodynamic calculations of the force due to such a bridging bubble were in quantitative agreement with the measured data, (the line in the figure is a theoretical fit based on six bubbles located at various positions on one of the surfaces).

Not everyone was convinced by this proposal, with the most common objection being that the internal gas pressure of the nanobubbles is, according to the Laplace-Young equation, so high that they should rapidly dissolve. But consensus began to emerge following the results obtained in the laboratory of Dr Mark Rutland at Sydney University.

Two typical force versus separation curves showing the steep repulsion at long range, (inset; the flat curve is the hydrodynamic drainage repulsion), the jump toward contact from either 60 or 140nm separation, the initial soft compliance regime immediately after the jump, and the final, infinitely steep, hard wall compliance. [ A. Carambassis, L. C. Jonker, P. Attard, and M. W. Rutland, Phys. Rev. Lett. 80, 5357-5360 (1998)].

The data showed prior to the onset of the attraction, (at 60 and 140nm separation in the figure), a sharply increasing repulsion in the force law. This indicated the presence of a pre-existing object that stuck out from the hydrophobic surface and that deformed or disappeared when the surfaces jumped into contact. In addition, the soft compliance regime appeared to be a dynamic effect due to the spreading of the bubble along the surfaces. These features, in conjunction with experiments by others that showed a diminished attraction in de-aerated water, provided strong evidence for the existence of nanobubbles.

Four force versus separation curves at pH 9.4, 9.4, 5.6, and 3, (in order of decreasing repulsion), showing that the position of the jump is approximately equal to the height of the nanobubbles imaged in the first figure, (it is less than that height, presumably due to bubble flattening by the repulsive force). Note the step in the force following the jump out of contact upon reversal of the motion [J. W. G. Tyrrell and P. Attard, Phys. Rev. Lett. 87, 176104 (2001)].

The experimental evidence outlined above provided, in the eyes of many, conclusive proof for the existence of nanobubbles. Others however were still not convinced. The problem was (and is) that nanobubbles appear to be forbidden on thermodynamic grounds because of their high internal pressure. It is for this reason that the images obtained by Tyrrell are especially valuable, since for many scientists seeing is believing. The images were quite difficult to obtain because the nanobubbles are rather delicate and are easily disrupted. There is a narrow range of instrumental parameters that dictate how large an amplitude of oscillation one can apply to the tapping cantilever. The images are also sensitive to the threshold damping: too much causes the tip to penetrate the bubbles and to stick and destroy them, and too little causes it to lose contact with the surface. Because of the ephemeral nature of the phenomenon, the results are at best qualitatively reproducible, and it is difficult to distinguish genuine variability from experimental artefacts.

Tapping mode image of a 3 micron square encompassing a 1 micron square, (lower right corner), imaged in the preceding 20 minutes in contact mode. The latter showed no lumps and less than 0.5nm root mean square roughness. The large features on either side of the small square show no phase lag and appear to be solid debris from the contact mode imaging [J. W. G. Tyrrell and P. Attard, Phys. Rev. Lett. 87, 176104 (2001)].

The main argument that the features exhibited above are indeed nanobubbles is the phase image, since in general a phase lag signifies a soft or deformable feature, which means that the lumps imaged above are materially different to the substrate. In addition, if the decrement in the vibration amplitude used in the feedback loop is increased, the image is severely degraded, (ie. the features are destroyed), but they reappear if the decrement is returned to its original value. A more systematic investigation of this effect involved a preliminary contact mode imaging, which proved to be smooth, (less than 0.5nm root mean square roughness), and featureless. This was interpreted as signifying that the tip had scraped the nanobubbles aside. Within 20 minutes a trapping mode image was acquired, that revealed once again the lumps. It is difficult to understand this behaviour if the lumps were a material property. What it suggests is that the nanobubbles had emerged from the aqueous surrounds, which indicates that they are in some form of equilibrium.

In another series of experiments, after imaging the nanobubbles in water, (contact angle 101 degrees), the cell was filled with ethanol, (for which the surfaces are lyophilic with a contact angle of 29 degrees). Tapping mode images revealed no nanobubbles in the ethanol. The nanobubbles returned when the cell was subsequent refilled with water. In these experiments a jump into contact and a strong adhesion were found in water but neither were present in ethanol.

Practical Implications

Nanobubbles have two distinct practical consequences: they cause a long-range attraction and a strong adhesion between hydrophobic colloid particles, and they decrease the drag on water flowing next to hydrophobic surfaces and in hydrophobed capillaries. In addition their stability raises certain fundamental thermodynamic questions.

When a nanobubble bridges between the surfaces of two approaching hydrophobic particles, it spreads laterally and draws the two particles into contact. This maximises vapour-solid contact and minimises the water-vapour interfacial area. The range of the attraction is essentially the height of the nanobubble above the surface, which can be two orders of magnitude greater than the measurable range of the classical van der Waals attraction. Accordingly, the stability of dispersions of hydrophobic particles is much diminished if they have a covering of nanobubbles. The colloid particles come together to form aggregates that either float or sink out of the solution. Further, the colloids are bound strongly in contact by a vapour neck, and this capillary adhesion is essentially irreversible.

Understanding the origin of the long range hydrophobic attraction offers the possibility of control and exploitation. For example, one might consider destabilising lyophobic suspensions by supersaturating them with gases. This doubtless is already used to some extent in dissolved air flotation. Alternatively, the longevity of dispersions may be enhanced by excluding air or by de-aeration treatments.

The fact that the nanobubbles appear to carpet the hydrophobic surface also has implications. In terms of the surface chemistry of the particles, it is probably more appropriate to consider reactions and adsorption at the air-water interface rather than at the underlying solid surface. Further, the drag of a hydrophobic particle moving through water is likely much reduced by this covering of bubbles. In terms of hydrodynamics, slip boundary conditions would appear to be more appropriate than the stick boundary conditions that are traditionally assumed at solid surfaces. The unexpectedly high flow of water down thin hydrophobic capaillaries is probably a manifestation of this effect. It is also possible that the swimming suits used by elite swimmers are made from hydrophobed fabrics, and that it is a covering of nanobubbles that reduces their drag.

Thermodynamic Stability

The conceptual challenge posed by nanobubbles is to account for their existence and stability. The Laplace-Young equation says that the internal pressure of a bubble is inversely proportional to its radius. A 10nm nanobubble would have an internal pressure of 144 atmospheres, and it could not be in equilibrium with the atmosphere. If such bubbles exist, either the equilibration time is exceedingly long, (ie. the gas diffuses only slowly through the water), or the solution is supersaturated with gas, (the two are not unrelated).

The images obtained by Tyrrell provide insight into the origin of the stability of the nanobubbles. They reveal that although the jump into contact in the force measurements is a good indicator of the height of the nanobubbles above the surface, this is not equal to the radius of curvature of the nanobubbles, as had originally been thought. The radius of curvature is much larger than this, (cf. the large cross-sectional dimensions), and, if they were spherical, only a slice of the sphere would protrude above the surface, (a semisphere rather than a hemisphere). A radius of the order of 100nm, as estimated from the images, gives an internal pressure of 14 atmospheres. This goes a long way to accounting for their long life time.

Another interesting feature revealed by the images is the fact that the nanobubbles are not ideal spheres and their cross-section is irregularly shaped. This shows that there is a pinning or resistance to motion at the three phase line that is larger than the tendency to minimise the liquid-vapour interfacial area by making an ideal spherical shape and circular cross-section. If the solution is supersaturated with air, which appears to be the only logical conclusion that can be drawn from the fact that the bubbles are convex, then one would expect the bubbles to grow with time, leading to the close-packed appearance of the images. It appears that there is an electric double layer repulsion between the surfaces of adjacent bubbles that inhibits their coalescence and further growth.

Why should the water be supersaturated with gas? The answer is not entirely clear but one cannot but observe that heating of the solution usually occurs in force measurements. In the case of the surface forces apparatus, the heating occurs by the incandescent light globe used for the interferometry. In the case of the atomic force microscope, the heating occurs by the laser that is used to detect the cantilever deflection, and by the current passing through the piezo-electric drive on which the substrate is mounted. The significance of heating is of course that the solubility of gases decreases with increasing temperature. In both apparati once the measurement chamber is filled, the solution has little or no access to the atmosphere. Another potential source of supersaturation is the entrainment of gases during passage of the surfaces through the liquid-vapour interface, and during the pumping and pouring of solutions.

To Phil Attard's home page.