Perfluoroheptane Loaded Hollow Gold Nanoshells Reduce Nanobubble Threshold Flux

Abstract

The threshold flux for nanobubble formation and liposome rupture is reduced by 50 – 60% by adding a liquid mixture of tetradecanol and perfluoroheptane to the interior cavity of 40 nm diameter hollow gold nanoshells (HGN), and allowing the tetradecanol to solidify to hold the perfluoroheptane in place. On absorption of picosecond pulses of near infrared (NIR) light, the perfluoroheptane vaporizes to initiate cavitation-like nanobubbles as the HGN temperature increases. The lower spinodal temperature and heat capacity of perfluoroheptane relative to water causes the threshold flux for nanobubble formation to decrease. The perfluoroheptane containing HGN can be linked via thiol-PEG-lipid tethers to carboxyfluorescein containing liposomes and shows a similar decreased flux necessary for liposome contents release.

Graphical Abstract

Carboxyfluorescein release from liposomes triggered by NIR irradiation of perfluoroheptane core gold nanoshells.

Introduction

Thin-walled (1 −5 nm) hollow gold nanoshells (HGN), 10 – 50 nm in diameter, red-shift the local surface plasmon resonance (LSPR) from the 530 nm resonance of solid gold spheres to 650 – 950 nm ^[1] where biological materials adsorb weakly ^[2]. Transient vapor nanobubbles nucleated around HGN by absorption of nanosecond or shorter NIR light pulses ^[1,3–9] provide a unique pathway for endosome escape of proteins ^[10–12], genetic materials ^[13–16] and liposomes attached to the HGN ^[17], increasing the delivery efficiency and cell viability relative to other synthetic vectors ^[10–17]. Nanobubble formation requires a threshold laser flux, which our previous work has shown depends on the HGN size and shape as well as the overlap of the irradiation wavelength with the LSPR ^[1].

We show that by incorporating perfluoroheptane into the HGN interior cavity, the threshold flux for nanobubble formation is reduced by 60%. Minimizing the laser flux to generate nanobubbles decreases the impact on cell viability, and increases the penetration depth over which nanobubbles can be formed, which will enhance HGN applications in drug delivery ^[2]. To retain the perfluoroheptane in the HGN, tetradecanol is mixed with the perfluoroheptane in the liquid state then allowed to diffuse into the hollow core of the HGN, where the tetradecanol-perfluoroheptane mixture solidifies, trapping the mixture inside the HGN cavity (Fig. 1) ^[18]. The threshold flux reduction is likely not due to the large difference in latent heat of vaporization of perfluoroheptane (0.18 kJ/ml) relative to water (2.2 kJ/ml). Instead vaporization occurs at the spinodal temperature at which the latent heat is zero, due to the superheating necessary to nucleate nanometer dimension perfluoroheptane bubbles ^[1]. Perfluoroheptane has a lower spinodal temperature (400 – 430 K vs 550 K) and heat capacity (1.9 J/cm³-K vs 4.2 J/cm³-K) than water, which lowers the laser flux needed to initiate the nanobubbles. Our calculations show that the experimental 60% reduction in threshold fluxes consistent with spinodal decomposition. We also observe a 60% decreased flux necessary for nanobubble induced model drug release from liposomes.

Figure 1.

A) TEM image of silver nanocrystal templates. B) Hollow gold nanoshells (HGN) formed by galvanic replacement reaction from silver templates in (A). Note the hollow interior and multiple bright pores that provide access to the hollow core. C) PFC-HGN after filling with tetradecanol-perfluoroheptane mixture. D) UV/vis spectra of the silver templates (black), the unfilled HGN (red) and the tetradecanol-perfluorocarbon filled HGN (blue). Replacing water in the hollow core of the HGN with the higher permittivity tetradecanol leads to a small red-shift of the plasmon resonance peak. E) Full field electromagnetic simulations of the normalized extinction as a function of wavelength for the silver templates and the bare and filled gold nanoshells ^[38–40]. 40 inside diameter HGN with a 2.5 nm thick shell (45 nm outside diameter) gave the best agreement with the experimental spectra. Exchanging the higher permittivity tetradecanol ($\mathfrak{R}\left(\epsilon \right)=$ 2.088 compared to water permittivity of $\mathfrak{R}\left(\epsilon \right)= 1.77$) for water in the HGN cavity led to a slight red shift for the tetradecanol-perfluoroheptane loaded nanoparticles, which is consistent with the experimental spectra (D). F) Size distribution of the silver templates, unfilled HGN and HGN filled with tetradecanol and perfluoroheptane (PFC-HGN) obtained by particle counting. There are minimal differences in the mean sizes and the size distributions.

Several groups have added gold nanoparticles, dyes or other light absorbers to droplets of liquid perfluorocarbons in aqueous suspensions ^[19–21]. The liquid PFC droplets range from 200 nm to several microns in diameter and must be stabilized against coalescence with a protein, surfactant or lipid shell ^[22]. Wilson et al. made 200 – 400 nm perfluoropentane nanodroplets with a bovine serum albumin (BSA) shell in which light absorbing gold nanorods had been suspended ^[19]. Like our PFC-HGN, the minimal laser flux to initiate bubble formation from these PFC liquid droplets was ~ 5 ${\mathrm{m}\mathrm{J}/\mathrm{c}\mathrm{m}}^2$, confirming that the mechanism of nanobubble formation is similar when the PFC is located within or without the plasmon resonant nanoparticles ^[19]. Hannah et al. made stabilized PFC nanocapsules (BLInCs) in which NIR light absorbing dye was suspended in 200 nm PFC droplets. In response to NIR irradiation, the BLInCs undergo rapid vaporization and re-condensation at body temperature (37 °C) ^[20,23]. We expect that our PFC-HGN would be beneficial in these applications as they are significantly smaller (10 – 40 nm), can be tethered to proteins ^[10–12], genetic materials ^[5,13–16], or liposomes ^{[1,3,4,24,25]} and can be modified for intra- or extra-cellular delivery and do not coalesce.

Experimental

Hollow Gold Nanoshell Synthesis

The galvanic replacement of silver templates (Fig. 1A) by gold (III) chloride hydrate (HAuCl₄) to form HGN (Fig. 1B) was described previously ^[1,24,26]. Silver and gold concentrations and reaction times were chosen to center the HGN LSPR at 750 nm (Fig. 1D, details in Supplemental Materials).

Perfluoroheptane-Tetradecanol HGN (PFC-HGN)

Perfluoroheptane was chosen as its boiling point, 84 °C, is well above the temperatures used in this process. To retain perfluoroheptane in the hollow core of the HGN, we modified the method described in Moon et al. ^[18]. 200 mg of 1-tetradecanol was mixed with 20 $\mathrm{\mu }$ l of perfluoroheptane for 30 min at 50 °C. 1 ml of HGN suspension (Supplemental Materials) in water was centrifuged at 13,000g for 5 minutes and the pellet was re-dispersed in 1 ml of methanol (MeOH). The tetradecanol/perfluoroheptane mixture was added to the HGN in MeOH with magnetic stirring at 200 rpm for 3 hours at 70 °C. The methanol rapidly evaporates, leaving behind the mixture of tetradecanol and perfluoroheptane, which diffused into the HGN central cavities (See Fig. 1C). On adding water at 70 °C, the solution separated into two phases, one containing the unencapsulated tetradecanol and perfluoroheptane, which are insoluble in water, the second containing the hydrophilic PFC-HGN suspended in water. The two-phase mixture was centrifuged at 13,000g for 5 min, leaving a blue pellet that contained the PFC-HGN. The pellet was washed in ice-cold water followed by a room temperature water wash. The PFC-HGN were further washed by dispersing the pellet in Millipore water at room temperature followed by centrifugation at 13,000g and repeated a minimum of 5 times (Fig. 1C).

Fig. 1D shows the large red shift between the silver templates and the HGN. There is a smaller red-shift in the spectra on replacing the lower real permittivity water of $\mathfrak{R}\left(\epsilon \right)= 1.77$ in the hollow core with the higher real permittivity tetradecanol of $\mathfrak{R}\left(\epsilon \right)= 2.088$ refractive. Fig. 1E shows the optical properties of the silver templates and empty and filled HGNs using full field electromagnetic simulations. The hollow structure of the HGN is responsible for the dramatic red shift in plasmon resonance peak between the silver templates and the gold nanoshells ^[1,27]. A Drude function based on a 50% Ag – 50% Au alloy was used to model the optical properties of the gold-silver alloy shells ^[28] (Details in Supplementary Materials). Water with a constant real permittivity of $\mathfrak{R}\left(\epsilon \right)= 1.77$ was assumed to be both inside and outside the HGN. For the PFC-HGN, tetradecanol with $\mathfrak{ }\mathfrak{R}\left(\epsilon \right)= 2.088$ was used inside the PFC- HGN with water on the outside. Previous work has shown that the LSPR of the hollow nanoshells is primarily governed by the ratio of the inner radius to the outer radius of the HGN ^[1,27]. The bare HGN LSPR peak at 750 nm was used to determine an average shell thickness of ~ 2.5 nm for a 40 nm inside diameter HGN (outside diameter 45 nm). Replacing the water on the inside of HGN with the same diameter and wall thickness with tetradecanol induces a small red-shift consistent with the experimental spectra as has been previously observed ^[18,29].

The silver template, HGN, and PFC-HGN concentrations and size distributions were measured using single particle tracking (Fig. 1F). Only minimal differences in the mean size of ~ 45 nm and size distributions were seen. To minimize aggregation, 750 Da methoxy-PEG-thiol was added to the HGN or PFC-HGN at a ratio of 1:10 mol PEG:Au and allowed to react overnight at room temperature. The excess PEG was removed by repeated washing and centrifugation.

Attaching HGN to Liposomes

100 nm diameter dipalmitoylphosphatidylcholine (DPPC) (Avanti Polar Lipids; Alabaster, AI) and DSPE-PEG-2000-SH (Nanocs; New York, NY) at a 95:5 molar ratio liposomes in TES buffer (51 mM NaCl, 50mM TES, NaN₃ 0.02 wt%, and 10 mM CaCl₂; pH 7.4) at a total concentration of 25 mg/ml were prepared as described previously ^[1] The liposomes contained 25 mM carboxyfluorescein (at this concentration, CF fluorescence is quenched ^[30]). HGN or PFC-HGN were tethered via the thiol-PEG linker to the liposomes by reacting overnight at room temperature ^{[1,3,4,24,25]}. Untethered HGN or PFC-HGN were removed by size-exclusion chromatography using a PD MidiTrap G-25 column (GE Healthcare).

Nanobubble Generation and Detection in Flow

Details of the irradiating and probe lasers and the nanobubble detection scheme are described in the Supplemental Materials. Briefly, 28 picosecond pulses of NIR light at 750 nm (Ekspla, Vilnius, Lithuania) irradiated a 0.2 mm ID square, hollow glass capillary (#8320 Vitro Tubes, VitroCom, Mountain Lakes, NJ). The flow rate in the capillary (7.5 mm/sec) was matched to the beam size (300 μm) and repetition rate (50 Hz) to provide a single NIR laser pulse to any given volume of solution. A continuous wave Helium-Neon (632.8 nm,2 mW, polarized, HNL020L-JP, Thorlabs, Inc.) laser focused at the same spot was scattered by the refractive index difference between nanobubbles and liquid water, which caused a reduction in laser intensity which was recorded by a photodetector (Luka, Andor Technology, Northern Ireland) coupled to an oscilloscope (Teledyne LeCroy, Wavesurfer MXs-8) ^[7]. The recording time was limited to 200 nsec in order to resolve the nanobubble threshold intensities. The beam flux was measured with a pulse energy meter (Ophir Optronics, Ltd., Israel).

Quantifying Dye Release from Liposomes

Following irradiation, the liposome sample was collected and analyzed for carboxyfluorescein (CF) release. The emission spectra were integrated over 510 to 530 nm, and dye release was calculated as $\mathrm{\%}\mathrm{ }\mathrm{R}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{e}=\mathrm{ }\frac{\mathrm{I}\left(\mathrm{t}\right)-{\mathrm{I}}_{\mathrm{o}}}{{\mathrm{I}}_{\mathrm{L}\mathrm{y}\mathrm{s}\mathrm{i}\mathrm{s}}-{\mathrm{I}}_{\mathrm{o}}}$, where I(t) was the measured intensity following irradiation, I_o was the intensity prior to irradiation, and I_Lysis was the intensity accompanying complete release following lysis with Triton X-100. Control liposomes lacking either HGN or PFC-HGN showed no CF release on irradiation.

Results

The threshold flux is detected by the large change in local refractive index on bubble formation, which scatters a probe laser beam, decreasing the transmitted intensity. Fig. 2A shows the fractional decrease in transmitted probe beam intensity for 45 nm PFC-HGN (PFC interior) irradiated under flow at increasing flux with single 28 ps NIR light pulses at 750 nm (the LSPR peak was ~ 750 nm, Fig. 1D). Nanobubbles form for fluxes ≥ 5 mJ/cm² as indicated by the decrease in the signal detected by the oscilloscope (arrow, Fig. 2A). Increasing the flux to 12 mJ/cm² decreased the scattered intensity further, consistent with larger nanobubbles. At fluxes ≥ 40 mJ/cm² the intensity went to zero, and did not recover over the subsequent 100 nsec. Fig. 2B (arrow) shows that for the HGN with water interiors we saw no nanobubbles at 5 mJ/cm^{2 [1]}. Increasing the flux to 12 mJ/cm² was needed to initiate nanobubbles (arrows, Fig. 2B). For the HGN with water interiors, the nanobubbles were consistently smaller and disappeared more rapidly than for the PFC-HGN. Incorporating perfluoroheptane-tetradecanol into the HGN interior resulted in a ~ 60% decrease (from 12 to 5 mJ/cm²) in the threshold flux for 40 nm HGN. Fig. 3 shows optical images of the bubbles generated from the PFC-HGN for fluxes of 40 mJ/cm²; the perfluoroheptane bubbles persisted long enough for them to be imaged. We could not image nanobubbles generated from water containing HGN even at high flux; the nanobubbles apparently collapsed too quickly to be recorded by our camera.

Figure 2.

A) Fractional decrease in the optical signal from the continuous HeNe probe laser due to scattering from nanobubbles. For the perfluorocarbon loaded HGN (PFC-HGN), the first deflection of the signal signifying detection of nanobubbles occurs at 5 mJ/cm². Increasing the laser flux increases the scattering, decreasing the signal. The PFC nanobubbles persist longer and are larger for a given flux compared to comparable water nanobubbles. B) For water containing HGN, no bubbles are detected at 5 mJ/cm², and it requires 12 mJ/cm² to see the same nanobubble signal generated by the PFC-HGN at 5 mJ/cm².

Figure 3.

Optical microscope image of PFC-HGN following irradiation within a 200 μm wide capillary. The bright spots are scattered light from the nanobubbles forming. This scattered light is prevented from reaching the photodetector and results in the traces shown in Fig. 2. We do not see similar nanobubbles in water containing HGN; the bubbles do not last long enough for imaging at normal optical frame rates.

Fig. 4 shows that it is the perfluoroheptane that leads to the increase in nanobubble formation. For HGN loaded with tetradecanol only, there was almost no change in the bubble signal following irradiation with NIR pulsed light compared to the HGN containing water (the small differences are likely do to small changes in the LSPR as in Fig. 1D). However, when perfluoroheptane is co-loaded with tetradecanol in HGN, the lifetime of bubbles increases from ~ 10 nanoseconds to ~ 20 nanoseconds and the amplitude of the scattering signal increases at 30.6 mJ/cm² of laser flux.

Figure 4.

Unscattered probe beam intensity for bare HGN (water interior, black), HGN containing only tetradecanol (PCM, red) and tetradecanol mixed with perfluoro-heptane (PCM+PCF, blue). The scattering from the HGN and HGN plus PCM are similar, but the addition of PFC greatly increases the bubble lifetime and size. This confirms that the PFC is the cause of the threshold decrease and nanobubble growth.

Previous work has shown that the threshold flux decreases with decreasing HGN size; for 10 nm diameter HGN, the threshold flux is 5 mJ/cm² compared to 12 mJ/cm² for 45 nm HGN as shown in Fig. 2 ^[1]. However, as Fig. 5 shows, for 10 nm PFC-HGN (perfluoroheptane-tetradecanol interior) irradiated under flow at increasing flux at 800 nm (the LSPR peak for these PFC-HGN was also ~ 800 nm), nanobubbles formed for fluxes ≥ 2.5 mJ/cm² as indicated by the decrease in the signal detected by the oscilloscope. For fluxes ≤ 1.23 mJ/cm², there was no detectable decrease in the probe laser signal ^[1]. Increasing the flux further increased both the change in scattering amplitude and bubble lifetime. The net decrease in threshold flux of the 10 nm PFC-HGN to the 5 nm HGN was about 50%, similar to the larger HGNs.

Figure 5.

Unscattered probe beam intensity for 10 nm PFC-HGN containing tetradecanol mixed with perfluoro-heptane as a function of increasing 800 nm laser flux. Previous work has shown that the threshold flux for HGN decreases with decreasing size^[1]; for 10 nm HGN, the threshold flux in 5 mJ/cm². However, for PFC-HGN, the threshold flux decreases by 50% to 2.47 mJ/cm².

Liposome Release Threshold

To determine if the PFC-HGN generated nanobubbles were capable of lysing liposomes to release small molecules, we encapsulated self-quenching carboxyfluorescein (CF) at 25 mM concentration into 100 nm diameter dipalmitoylphosphatidylcholine (DPPC) liposomes tethered to either HGN or PFC-HGN (Fig. 1) by thiol-PEG-lipid tethers ^[1,3,4,24]. Previous work has shown that nanobubbles of sufficient size can rupture liposome membranes to release liposome contents ^[1,3,4]. The liposome-HGN or liposome-PFC-HGN suspensions were flowed through the capillary at 7.5 mm/sec and were irradiated with, on average, a single laser pulse of controlled flux. After irradiation, the liposome suspension was collected and the CF fluorescence intensity was determined to evaluate the fractional CF release relative to lysing the liposomes with Triton-X ^[1,3,4]. We found that PFC-HGN-liposomes began to release CF at a flux of 12 mJ/cm² compared to 25 mJ/cm² for the HGN-liposomes, a decrease of 52%, which was consistent with the decrease in threshold flux found for the nanoparticles themselves. We find that the threshold to rupture a liposome is greater than that required to initiate nanobubbles, which suggests that liposome lysis requires larger more than twice the fraction of CF than the HGN-liposomes. The inset image shows a fluorescence micrograph taken immediately following irradiation of the CF-containing liposomes. At the 25 mM CF concentration in the liposomes, fluorescence is quenched. However, following liposome rupture, the CF diffuses and convects away from the liposome, decreasing the CF concentration to the point that fluorescence increases and becomes visible as spherical waves of fluorescence intensity. Detecting CF release from liposome rupture provides confirmation that nanobubbles are forming and that the addition of perfluoroheptane nanobubbles ^[1,3,4]. Fig. 6 shows that at all fluxes tested, the PFC-HGN-liposomes released helps nanobubbles grow to larger sizes for smaller laser fluxes as well as lowering the threshold flux for nanobubble formation.

Figure 6.

Fractional release of carboxyfluorescein (CF) (relative to total lysis with Triton X) at different laser fluences with bare HGN or PCM+PCF loaded HGN. At all laser fluences, more CF was released by the PCM+PCF loaded HGN showing that the nanobubbles formed from the perfluoroheptane were capable of lysing liposomes and that the perfluoroheptane initiated nanobubbles at lower laser fluences as shown in Fig. 2, 3. Higher fluences, and hence larger nanobubbles are needed to rupture liposomes than at the nanobubble threshold shown in Figs. 2,3. The inset shows an optical micrograph of the liposomes rupturing and releasing CF following laser irradiation. CF is quenched at the 25 mM concentration in the liposomes, but following release and diffusion and dilution in the surrounding water, the fluorescence intensity increases.

Discussion

Adding perfluoroheptane in tetradecanol to the cavity of surface plasmon resonant hollow gold nanoparticles (PFC-HGN) decreases the flux necessary for nanobubble formation by 50 – 60%. For a given flux, the PFC-HGN generate larger nanobubbles than the HGN, and more efficiently rupture liposomes to release more of their contents. We hypothesize that on absorption of the pulsed NIR, the tetradecanol melts and the perfluoroheptane vaporizes to initiate the nanobubble as the HGN temperature increases due to the conversion of optical to thermal energy.

This conversion of picosecond light pulses to thermal energy occurs faster than dissipation of this energy to the surrounding liquid, initially confining the optical energy to heating the HGN or PFC-HGN (See discussion in Supplementary Materials) ^{[1,26,33–36]}. This initial temperature increase is proportional to the light energy absorbed ^[1,33], ${Q= \sigma }_{abs}F,$ in which ${\sigma }_{abs}$ is the absorption cross section of the HGN and F is the laser flux and Q the net energy absorbed. Nanobubbles then form as the hot HGN begins to transfer heat to the material in the hollow core and the surroundings in the following nanoseconds. However, nanoscale bubbles require significant superheating to nucleate due to the large contribution of vapor-liquid surface tension, γ :

$$\Delta H_{vap}\left(\frac{T-T_B}{T{T}_B}\right)=bln\left(1+\frac{2\gamma }{R{P}_B}\right)$$ (1)

Eqn. 1 shows that to initiate a nanobubble of radius R requires superheating to a temperature, T, above the normal liquid boiling point, T_B, at atmospheric pressure, P_B. $\Delta H_{vap}$ is the latent heat of vaporization and b is the gas constant. For perfluoroheptane, γ = 13 mN/m, T_B = 357 K at P_B = 1 bar, and $\Delta H_{vap}$ = 31 kJ/mol. We estimate that for a bubble of R = 10 nm, the liquid perfluoroheptane would need to be heated to ~520 K, which is above the critical temperature of perfluoroheptane, T_c = 475 K. This suggests that instead of conventional boiling, the actual stability limit of liquid perfluoroheptane is the spinodal temperature, T_S. A reasonable estimate of the spinodal temperature is 85–90% of the critical temperature, $T_s\approx .85- .95 T_c$;^[22] which gives an estimate of T_S ~ 400 – 430 K for perfluoroheptane. At the spinodal temperature, liquids become mechanically unstable and spontaneously convert to the vapor phase with no heat of vaporization ^[36]. Hence, the perfluoroheptane likely vaporizes between 400 – 430 K when confined in small droplets inside the HGN. In comparison, the spinodal temperature of water is T_S = 277 °C or 550 K, which requires substantially greater HGN temperature to initiate a nanobubble.

Following irradiation and conversion of light to thermal energy, the hot metal shell transfers heat to the material in the nanoshell core, which is initially at T₀. This causes the shell temperature to fall until the core and shell have equilibrated (this ignores losses to the surroundings). Hence the maximum temperature, T_max, that can be reached by the material in the core is for an initial gold shell temperature of T_G is:

$$T_{max}=\mathrm{ }\frac{\left[\left(4\pi R^{2}t\right){\rho C}_p{T}_G\right]+\left[\left(\frac{4}{3}\pi R^3\right){{\rho }_{M}C}_{pM}{T}_0\right]}{\left(4\pi R^{2}t\right){\rho C}_p+\left(\frac{4}{3}\pi R^3\right){{\rho }_{M}C}_{pM}}$$ (2)

ρC_p is the volumetric heat capacity of gold (2.5 × 10⁶ J-m⁻³-K⁻¹) and ${{\rho }_{M}C}_{pM}$ is the volumetric heat capacity of the material of radius R in the HGN core, with t being the gold shell thickness. For perfluoroheptane, ${{\rho }_{M}C}_{pM}$ = 1.9 × 10⁶ J-m⁻³-K⁻¹, ^[37] and for water, ${{\rho }_{M}C}_{pM}$ = 4.2 × 10⁶ J-m⁻³-K^-1. For R = 20 nm and t = 4 nm, for T_max to reach the spinodal temperature of water of 550 K, the gold-silver alloy shell must be heated to ~ 1000 °C. However, to reach the lower spinodal temperature of perfluoroheptane of 400 – 430 K, the nanoshell only needs to be heat to about 300 °C (due to the large difference in volumetric heat capacity). This suggests that the threshold flux for perfluoroheptane nanobubble formation should be about 30 – 40% that of water nanobubbles, which is consistent with the reduction in nanobubble thresholds we observe of 50 – 60% (Fig. 2, 4, 6). As the nanoshell is also losing heat to the surrounding water for both water and perfluorocarbon in the core, in practice, the metal shell must reach a higher temperature to initiate nanobubbles, and the reduction in nanobubble threshold is less. For both HGN and PFC-HGN, the gold nanoparticles must be heated sufficiently to anneal from their original hollow structure to a condensed, solid-core structure, as evidenced by the change in the UV/vis spectra following laser irradiation (Fig. 7).

Figure 7.

UV/vis spectra before (red) and after (green) irradiation above the threshold flux for the PFC-HGN nanoparticles. The blue shift of the plasmon resonance is consistent with a change from a hollow gold-silver alloy nanoshell to an annealed solid gold-silver alloy nanosphere ^[1] (Compare to Fig 1D).

Conclusions

The threshold flux for nanobubble formation and liposome bilayer rupture is reduced by 50 – 60% by incorporating perfluoroheptane-tetradecanol mixtures into the hollow cavity of plasmon resonant hollow gold nanoshells. Minimizing the laser flux required to generate nanobubbles could help to decrease the light impact on cell viability and increase the tissue thickness over which nanobubbles could be formed ^[2]. On absorption of picosecond pulsed NIR light, the HGN is rapidly heated, the tetradecanol melts and the perfluoroheptane vaporizes to initiate the nanobubbles. Perfluoroheptane has a lower spinodal temperature (400 – 430 K vs 550 K) and heat capacity (1.9 J/cm³-K vs 4.2 J/cm³-K) than water, which lowers the laser flux needed to initiate the nanobubbles. The perfluoroheptane HGN (PFC-HGN) can be tethered via thiol-PEG-lipid tethers to liposomes ^[1,3,4]; the PFC-HGN also show a 60% decreased flux necessary for liposome contents release.

The perfluoroheptane bubbles persist significantly longer than water vapor nanobubbles, which may be of interest in applications including photoacoustic imaging. Since vaporization provides a stronger photoacoustic signal than thermal expansion, it may be possible to use smaller numbers of nanoparticles in future biological and clinical applications. The longer lifetime of the PFC containing nanobubbles is likely due to the inhibited nucleation and growth of perfluorocarbon liquid droplets as the nanobubbles shrink and cool. As water and perfluoroheptane are immiscible, nucleating a purified perfluoroheptane liquid droplet from the vapor mixture of perfluoroheptane and water will be limited. The small size and likely necessity of homogeneous liquid nucleation will likely require significant supercooling and limit the condensation rate of the nanobubbles.

It has been established, as water vapor nanobubbles collapse, asymmetric liquid water jets are generated that lead to the mechanical forces that rupture lipid membranes ^[1,3–5,17], similar to the cavitation bubbles produced by sonication. While PFC-water nanobubbles last significantly longer, their ability to rupture liposomes leads us to speculate that similar liquid jets are available and rupture liposomes in a similar fashion. Because each PFC-HGN induces nanobubble formation, there is no need for any collective interaction between nanoparticles and even a single PFC-HGN attached to a liposome or contained in an endosome can induce bilayer rupture. Moreover, while the laser power needed to initiate nanobubble formation is dramatically large, $Power= \frac{\mu J/pulse}{28 psec/pulse}=~ Megawatt,$ the overall energy input to the sample (typically of the order of μJ/pulse) is sufficiently small that the overall temperature remains constant.