As shown in Fig. 1(a), when deionized water (25 μm) is directly added onto the HOPG surface, only bare graphite surface and steps could be seen in the height image. Nevertheless, when we add the cold water which is cooled at 4 °C for 72 h to the fresh HOPG surface and waitsfor more than twenty minutes at room temperature, a large number of spherical cap-shaped objects can be observed via PF-QNM imaging (Fig. 1(b)). The lateral dimensions of these spherical caps arearound 100 nm–200 nm and the contact angles are above 150°. By performing PF-QNM imaging, we are able to acquire the stiffness information at the same time. The stiffness of those spherical cap-like objects is around 50 mN/m–150 mN/m (Fig. 2(a)) which is much softer than the substrate surface. Those results are consistent with the former observations about the morphology and stiffness of interfacial nanobubbles produced by water-ethanol exchange.^[62] During the AFM scanning, we also find that most of the bubbles (size > 50 nm) are stable and hardly changed after scanning for several hours. In some cases, some smaller bubbles (size < 50 nm) disappear or newly form, which may be caused by the AFM tip during the scanning or the effects from the neighboring bubbles.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) AFM height images of HOPG surfaces using (a) room temperature water and (b) cold water after cooled at 4 °C for 72 h. The scan size is 5 μm × 5 μm.

Fig. 2.

Figure Option ViewDownloadNew Window

Fig. 2. (color online) (a) Stiffness image of nanobubbles in Fig. 1(b), and (b) force curves of a spherical cap-shaped object in force volume mode. The insert in panel (b) shows the slice (deflection error) image of the chosen nanobubble. Force curve is obtained on the nanobubble marked by blue circle. Panel (b) shows the three regions, i.e., region 1 (ellipse): linear-response region; region 2 (circle): retraction on substrate surface; region 3 (square): adhesion region.

The force curves of a spherical cap-shaped object acquired by force volume mode are shown in Fig. 2(b). In the experiment the calibrated spring constant of the silicon nitride tip is 0.3556 N/m and trigger threshold is set to 18.25 nm. As a result, the “trigger” peak force which leads to the retraction of the tip is about 6.5 nN. With such a large force, there is a very strong interaction between the tip and the substrate surface before retraction.

In the approaching curve (green line), the AFM tip firstly attached to the spherical cap-shaped object at about 10 nm, which is referred to as “jump-in” process. Then the AFM tip interacting with the object for about 8 nm in depth before reaching the HOPG surface. The tip-object interaction as shown in region 1 (ellipse) exhibits a quasi-linear response before touching the HOPG surface. The slope of the linear part is estimated at 1 nN/8 nm which is 125 mN/m. This is comparable to the surface tension of water 72 mN/m, which may indicate the gaseous property of the object. Those results are consistent with the results reported previously, suggesting that the tip-bubble interaction is dominated by the repulsive force which increases linearly when the tip approaches to the bubble.^[63–65] and agrees with simulation results of interaction between AFM tip and pinned nanobubble by using the lattice density functional theory.^[66]

When retracting from the substrate surface as shown in the region 2 (circle), the AFM tip suffers a strong adhesion force (~1.8 nN). The reason for that may be that the tip has already “touched” the HOPG surface when such a large trigger force is reached. Note that the interaction between the tip and the object has a small plateau before the linear-response starts in region 1 (ellipse). The plateau has also been observed in several researches on force curves of nanobubble.^[63,67] This phenomenon is not expected with a hydrophilic tip according to the dynamic interaction model. But it can be explained if the tip is hydrophobic and penetrated into the bubble near the substrate surface so that the contact line does not contribute to the interaction force.^[63] Considering the fact that the AFM tip is plasma-cleaned to achieve hydrophilicity before the imaging process, neither a hydrophobic tip nor a contaminated tip cannot be the most possible reason for that plateau. A possible explanation may state as follows: the AFM tip which is not so hydrophilic is eventually penetrated into the object due to the large loading force (6.5 nN) while approaching to the substrate surface. Therefore, when the tip is withdrawn from the substrate surface, the contact line is not pinned on the AFM tip and slips several nanometers on the tip before being fully pinned on the tip. In such views, a small plateau then appears before the linear response of tip-object interaction. In the next step as shown in region 3 (square), the contact line still attaches to the “less hydrophilic” tip which leads to an attractive force until the tip is jumped off from the object surface. It is noted that most of the features are exhibited in Fig. 2(b) such as quasi-linear response, jump-in and adhesion force before jump-off, which are all consistent with the previous research results of the tip-nanobubble interaction^[64,65] and also quite different from the forces curves of PDMS nanodroplets as reported recently.^[67,68]

To further verify that those cap-shaped objects are indeed gaseous nanobubbles, we place the cold water into a vacuum chamber and degassed at a pressure lower than 10 mbar (1 bar = 10⁵ Pa) for more than 2 hours. When the degassed cold water is injected into the HOPG surface and kept for more than 30 min before AFM imaging, only smooth graphite surface is observed as shown in Fig. 3(a). In the following experiment, we put the degassed water again into the refrigerator and cooled at 4 °C for another 72 h. When the re-cooled degassed water is added onto the HOPG surface and kept for about 30 min, the spherical cap-like objects appear on the substrate as shown in Fig. 3(b). In the degassing process, most of the dissolved gases including N₂ and O₂ are removed from the water. Meanwhile the degassed water contacts only the ambient environment in the whole cooling process as described above, thus those spherical cap-shaped objects should be gas filled nanobubbles.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) The AFM height images of HOPG surfaces in (a) degassed water and (b) degassed water cooled for at 4 °C for 72 h. The scan size is 5 μm × 5 μm.

In order to find the appropriate experimental conditions for effective and controllable growth of nanobubbles, we further investigate the effects of different cooling time on the formation of interfacial nanobubbles. We cool the room temperature milipore water in the 4 °C refrigerator for 0 h, 2 h, 2 h, 4 h, 6 h, 12 h, 24 h, 48 h, and 72 h, then deposited these cold water onto the HOPG surface. In Fig. 4, when adding the room temperature water onto the HOPG surface, nanobubbles are rarely formed. After adding the cold water with a certain cooling time (2 h, 4 h), some nanobubbles began to form on the substrate surface. More bubbles form on the graphite surface when the cold water is cooled for a longer time. Specifically, nanobubbles formed with longer cooling time (72 h, 48 h, 24 h) are slightly larger in size than those formed with shorter cooling times (12 h, 6 h, 4 h, 2 h). The number of nanobubbles per μm² increases significantly with increasing cooling time. In addition, we find that the nanobubbles formed by our method are quite uniform in their lateral sizes (about 100 nm–200 nm) and distributed evenly on the HOPG surface as compared with those produced by water-ethanol exchange, where the lateral size and distribution of the nanobubbles vary a lot. This phenomenon may be attributed to the observation that the growth and stabilization of nanobubbles in our methods are much slower and moderate (about 30 min) while in other methods the environment near the HOPG surface is severely disturbed by the exchange process^[69] or high substrate temperature.^[60]

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) The AFM images of interfacial nanobubbles formed by cold water after being cooled at 4 °C for 0 h, 2 h, 4 h, 6 h, 8 h, 12 h, 24 h, 48 h, and 72 h. The AFM image of 0 h is obtained by using room temperature water stored in ambient condition.

The densities (number) and volumes of formed nanobubbles in water with different cooling times are analyzed to further explore the physical properties of nanobubbles. It has been generally recognized that the physical and chemical heterogeneities of substrate would affect the formation of the interfacial nanobubbles. The nanobubbles formed on different substrates may be different even on the same substrate the nanobubbles are also different in different regions. For example, many nanobubbles tend to nucleate at the edge of the graphite steps as indicated by our observations. In order to reduce the substrate effects as much as possible and acquire a quantitative description about the nanobubbles formed by cold water with different cooling times, we have the experiments repeated independently at least 5 times for each group, and select the different areas randomly for AFM imaging. After that, those AFM images are analyzed following the particle analysis procedure in the Nanoscope Analysis V1.5 software to count the numbers of nanobubbles and their average lateral size. We only count the nanobubbles with lateral dimension larger than 50 nm and height higher than 2 nm considering the larger error problem of small bubbles. In this process about 150 AFM height images with 13580 nanobubbles counted are analyzed in all the experiments for this work.

Since the nanobubbles formed by cooled water are all spherical cap-shaped and similar to each other in lateral size, we then calculate the density and total volume of the formed gaseous nanobubbles in the area from the following equations:

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. Density and total volume of nanobubbles in the area produced by cold water with different cooling times, room temperature water and degassed water. Error bar is given by analyzing the AFM images scanned at different places on HOPG surface in the experiments. DW: degassed water, RW: room temperature water (0 h).

Table A1.

Table A1.

Table A1. statistics Data of nanobubbles produced by cold water cooled for 48 h at 4 °C. . Count for nanobubbles, area 5 μm × 5 μm, lateral size > 50 nm Numbers Mean size/nm Mean height/nm Volume/nm3 Area/μm2 118 160.916 7.894 9502300.747 2.39977397 251 143.836 1.355 2763500.036 4.07848425 198 133.178 1.098 1514369.118 2.758163719 63 159.007 4.341 2718028.934 1.251016135 46 178.379 4.639 2668830.508 1.149569296 64 199.458 8.302 8320080.723 1.999736474 67 169.628 7.543 5725552.394 1.51411813 45 177.236 9.273 5166289.718 1.110212896 56 177.848 3.528 2455286.98 1.391156125 87 184.477 3.178 3696487.938 2.325378118 85 149.583 5.5 4115173.276 1.493734039 127 149.093 7.29 8107516.463 2.217216501 167 154.304 4.246 6636650.471 3.122918956 169 156.369 5.208 8463709.324 3.245472221 136 144.97 7.315 8238368.815 2.244838283 126 148.637 8.78 9642610.252 2.186322788 114 141.038 9.22 8257252.247 1.781012661 158 150.009 7.069 9899051.2 2.792425484 135 157.44 5.973 7864098.995 2.628172124 103 166.966 4.78 5395794.745 2.255190246 84 173.798 5.064 5051423.059 1.992777031 72 176.912 5.451 4829837.71 1.769852016 127 172.18 4.329 6405934.319 2.957052274 46 161.619 3.447 1627448.098 0.943696917 40 145.277 2.357 781673.8326 0.663045896 36 162.836 3.763 1411583.505 0.749709866 90 139.298 2.814 1930867.569 1.371583163 97 154.993 3.635 3328730.065 1.830146169 48 158.789 1.931 917930.589 0.950543374 55 163.278 1.425 820610.1416 1.151616577 67 148.289 2.959 1712884.281 1.157131056 66 170.384 1.98 1490063.63 1.504843819 55 167.375 1.682 1017860.389 1.210134781 Mean number Mean size/nm Mean volume/nm3 Mean area/μm2 96.90909091 160.5272727 4620539.396 1.88475895 50.05231423 14.49712962 2977539.155 0.797928053 Density/μm2 3.876363636 2.002092569

Table A1. statistics Data of nanobubbles produced by cold water cooled for 48 h at 4 °C. .

In Fig. 5(a), the density of nanobubbles formed by degassed water indicates that long time cooling would greatly improve the efficiency of nanobubble formation. In Fig. 5(b), the average total volume of nanobubbles on each AFM image is calculated from Eq. (2). Our purpose of performing such a calculation is that the original sources of these gaseous interfacial nanobubbles are the gases (mainly N₂, O₂) dissolved in the cold water. Therefore, the total volume of interfacial nanobubbles on AFM image could be directly related to the excess gas dissolved in the cold water when it is warmed up at ambient temperature. In degassed and room temperature water, the total volume of nanobubbles is quiet small, consistent with the previous observations that nanobubbles rarely formed in the room temperature deionized water and degassed water.^[56] Note that the heterogeneity of the substrate also affects the formation of nanobubbles, for example, nanobubbles tend to form near the steps of HOPG surface in many cases. This is why the error bars in Fig. 5 are quiet large. The total volume of nanobubbles increases with cooling time increasing and reaches a maximal value at the 48th hour. This means that the amount of gas dissolved in the cold water is in equilibrium after cooling for more than 48 h. Therefore all the statistical results suggest that room temperature pure water cooled for more than 48 h at 4 °C is sufficient enough to produce interfacial nanobubbles effectively on the HOPG surface.

In our method, the pure water is treated only by 4 °C and the high temperature is avoided in all the processes. Our strategy is based on the fact that at lower temperature such as at 4 °C the gas concentration would be higher than that at room temperature. Thus when the gas-saturated cold water is brought to the room temperature, an oversaturation state will be expected. This could be demonstrated by directly measuring the gas concentration in water.

Although most of the dissolved gases in water are nitrogen and oxygen, it is difficult to accurately measure the dissolved nitrogen in water. Alternatively, the dissolved oxygen is measured to understand how the excess dissolved gasaffects the formation of nanobubbles in our method. As shown in Fig. 6 the concentrations and saturations of dissolved oxygen in cooled water increase when the cold water is cooled for longer hours. The dissolved oxygen increases to more than 10 mg/L (about 120% saturation) after cooling for 72 h. The concentration of dissolved oxygen in cold water is in equilibrium after cooling for more than 96 h. The concentration of equilibrium is about 11 mg/L.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. Concentrations and saturations of the dissolved oxygen in cold water after different cooling times. The saturation rate is acquired by comparing with the 100% saturated water at 25 °C (8 mg/L).

While in the room temperature water (0 h), the dissolved gases are quiet few (4.5 mg/L, 55% saturation) and no nanobubbles can be formed by such water. As shown in Fig. 5(b) with more dissolved gases in cold water, more nanobubbles are produced on the HOPG surface. Thus the total volume of interfacial nanobubbles is directly related to the volume of dissolved gases in cold water.

In the control experiments, room temperature water preserved at the room temperature over 72 h could not produce nanobubbles on HOPG. We further cool this gas-saturated water at room temperature in 4-°C refrigerator for half an hour and find very few nanobubbles on HOPG surface, either. This means that the cold pure water with saturated gas at room temperature is not sufficient to produce a lot of nanobubbles. All the experiments show that only the cold pure water with gas saturated (normally 100%–120% saturation) at 4 °C can achieve the high efficiency of nanobubbles production. The growth mechanism is simple and can be explained as follows. When the cold water on the HOPG surface is warmed up, the gases are in an oversaturated state. This can be seen from the measured gas saturation of the cold water cooled for over 72 h, which is 10 mg/L, about 120% oversaturated, as compared with the saturated concentration 8 mg/L at the room temperature. Then, a large quantity of the dissolved gases would release from the bulk water and aggregate onto the hydrophobic HOPG surface. Therefore nanobubbles could be easily formed with a high efficiency.

The highly oriented pyrolytic graphite (short as HOPG, 12 mm × 12 mm, ZYB grade, Bruker) used as the substrate was freshly cleaved with double faced adhesive tape in each experiment. Deionized pure water (room temperature) was prepared with an ELGA PURELAB Classic water purification system to obtain a conductivity of 18.4 MΩ·cm. Degassed water was obtained by placing the water under 10 mbar for at least 2 h The freeze-pump-thaw process (3 cycles) was also used to acquire fully degassed water. The glass beaker (50 mL, 18 mm in diameter) and glass syringe (5 mL) were pre-cleaned with chromic acid and clean water. The AFM liquid cell contains a translucent O-ring which was used to seal the fluid. Before each experiment these glass containers and AFM liquid cell were ultrasonic cleaned with ethanol (99.8%, GR) and then deionized pure water for 45 min separately and then dried at 70 °C for 1 min before being used. The cold water was injected to the fluid cell through Teflon inlet and outlet tube and silicone connectors. The Teflon tubes and connectors were ultrasonic cleaned with ethanol (99.8%, GR) and then deionized pure water and dried before usage.

In order to produce interfacial nanobubbles, the deionized water was firstly preserved in a clean glass beaker and sealed with Parafilm (Laboratory Film, Parafilm Brand) while leaving several holes to contact the air environment, then the water was cooled in a 4-°C refrigerator for 2 h, 4 h, 6 h, 12 h, 24 h, 48 h, and 72 h. The Parafilm used to seal the glass beaker was a plastic paraffin film which is translucent, waterproof and odorless. Its potential effects on the purity of the water during the long term preservation were assessed in Appendix A (Fig. A1). The fresh cleaved HOPG surface at 25 °C was placed inside the AFM liquid cell before the cooled water was injected. When being removed from refrigerator, the cold water was added to the HOPG surface immediately using a clean glass syringe. After the HOPG surface was immerged in the cold water for about 20 min–30 min, different areas of the surface were imaged. The brief procedure was presented in Fig. 7. Each experiment was repeated at least 5 times independently and for each group at least 20 AFM images were collected for analysis.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. (color online) A brief diagram of the procedure for the nanobubble production.

PF-QNM (PeakForce Quantitative Nanomechanical Mapping) imaging in fluid was performed on a Bruker Multimode 8 SPM with NanoScope version 8.15 software and NanoScope V in order to quantitatively characterize the morphologies and stiffness of interfacial nanobubbles. The NPS-type probe (Bruker’s silicon nitride probes with a nominal spring constant of 0.35 N/m and a tip radius of 10 nm) and SNL-type probe (Bruker’s silicon nitride probes with a nominal spring constant of 0.1 N/m and a tip radius of 2 nm) were treated with a plasma cleaner (HARRICK PLASMA CLEANER PDC-32G) for about 3 min beforehand to achieve hydrophilicity and then immediately used to avoid being contaminated.

Following the standard PF-QNM imaging procedure, the deflection sensitivity and the spring constant of each cantilever were calibrated via PF-QNM ramp, thermal tuning on HOPG in water. The sample was oscillated vertically at a frequency of 2 kHz with an amplitude of 100 nm and a scan rate of 0.977 Hz. The Poisson ratio of the sample was set to be 0.5 as recommended by Multimode user guide to acquire reliable data while the stiffness of the sample was lower than 100 MPa. The peakforce set point was carefully selected and the loading force of imaging was about 200 pN–300 pN. Typical scan sizes were 5 μm and scan time for each image was 4 min–5 min. All the PF-QNM imaging was performed under ambient condition at room temperature and the humidity was about 30%–50%.

Force volume imaging was performed at a resolution of 128 pixels × 128 pixels with a scan rate of 0.0254 Hz. The ramp sizes and trigger thresholds were set to be 100 nm and 18.25 nm. The tip approaching speed was set to be 1.3 μm, and the number of points in a single approach–retraction cycle was 512. The scan size was 5 μm and for each image the scan took about 1 h. The force curves were presented as force versus separation, with the separation larger than 80 nm treated as the base line.

The offline AFM analysis program, NanoScope Analysis 1.5 was used for morphologies and mechanical property measurements and analyses. In particularly, particle analysis was performed to calculate the distribution and mean size of nanobubbles.

The dissolved oxygen (DO) was measured by a dissolved oxygen meter (Orion Versa Star pro, VSTAR94, Thermo Scientific). The room temperature water was cooled at 4 °C for 2 h, 4 h, 6 h, 12 h, 24 h, 48 h, and 72 h before measurement. Because the temperature of the cold water would rise up very soon when removed from the 4-°C refrigerator, to avoid a possible deviation induced by temperature change, during the measurement the DO probe and the bottle of cold water were kept in an ice box to maintain a low temperature. Each measurement took about 2 min before the meter reached its equilibrium. The dissolved oxygen probe was calibrated in air-saturated water at 25 °C. The calibrated dissolved oxygen of air-saturated water at 25 °C was set to be 8 mg/L automatically by the meter. The experiments were repeated several times to acquire reliable data.