Dynamic Manipulation of Droplets on Liquid-Infused Surfaces Using Photoresponsive Surfactant

Fast and programmable transport of droplets on a substrate is desirable in microfluidic, thermal, biomedical, and energy devices. Photoresponsive surfactants are promising candidates to manipulate droplet motion due to their ability to modify interfacial tension and generate “photo-Marangoni” flow under light stimuli. Previous works have demonstrated photo-Marangoni droplet migration in liquid media; however, migration on other substrates, including solid and liquid-infused surfaces (LIS), remains an outstanding challenge. Moreover, models of photo-Marangoni migration are still needed to identify optimal photoswitches and assess the feasibility of new applications. In this work, we demonstrate 2D droplet motion on liquid surfaces and on LIS, as well as rectilinear motion in solid capillary tubes. We synthesize photoswitches based on spiropyran and merocyanine, capable of tension changes of up to 5.5 mN/m across time scales as short as 1.7 s. A millimeter-sized droplet migrates at up to 5.5 mm/s on a liquid, and 0.25 mm/s on LIS. We observe an optimal droplet size for fast migration, which we explain by developing a scaling model. The model also predicts that faster migration is enabled by surfactants that maximize the ratio between the tension change and the photoswitching time. To better understand migration on LIS, we visualize the droplet flow using tracer particles, and we develop corresponding numerical simulations, finding reasonable agreement. The methods and insights demonstrated in this study enable advances for manipulation of droplets for microfluidic, thermal and water harvesting devices.

This PDF file includes:  S1 Legends for Movies S1 to S7

SI References
Other supporting materials for this manuscript include the following:

Materials and Instrumentation
All reagents were obtained from Sigma Aldrich, Oakwood Chemical, or Fisher Scientific and were used as received without further purification unless specified.Acetonitrile and dichloromethane were dispensed from a solvent purification system immediately before use or stored over 3 Å molecular sieves.Ethanol and chloroform were stored over 3 Å molecular sieves before use.Analytical thinlayer chromatography (TLC) was performed with Merck silica gel 60 F254 pre-coated glass plates and visualized by UV light or stained with p-anisaldehyde. 1 H and 13 C NMR spectra were recorded on a Bruker 500 MHz NMR spectrometer.Chemical shifts are reported relative to residual solvent peaks (δ 7.26 ppm for CDCl 3 , 2.50 ppm for DMSO-d 6 :) in 1 H NMR, and δ 77.20 ppm for CDCl 3 , 39.52 ppm for DMSO-d 6 in 13 C NMR).

UV-Vis spectroscopy and UV-Vis kinetic measurements
UV-Vis absorption spectra were recorded on an Agilent 8453 UV-Vis spectrometer from 200 to 1200 nm wavelengths.The photoinduced optical absorption kinetics of SP-DA-PEG and MCH-para were measured on a custom pump-probe setup.The pump beam was generated by a high-power LED (light-emitting diode) source (Thorlabs) coupled into a multimode optical fiber terminated with an output collimator.The UV pump source was generated by the fiber-coupled UV LED (Thorlabs M365FP1, 365 nm, 9.8 mW Min Fiber-Coupled LED); the green pump source was produced by green LED (Thorlabs M530F2, 530 nm 6.8 mW Min Fiber-Coupled LED) and the blue pump source by blue LED (M470F3, 17.2 mW Min Fiber-Coupled LED), which was positioned to illuminate the sample directly.The probe beam was generated by a High-Power MINI Deuterium Tungsten Halogen Source (Ocean Optics DH-MINI) w/shutter 200-2000 nm coupled into a multimode optical fiber terminated with an output collimator.The probe light was altered by a shutter (Uniblitz CS25) and controlled manually or through a digital output port (National Instruments USB-6009) using the LabVIEW   in a 2 mL cuvette by adding 0.6 mL of DI water followed by 0.4 mL of corresponding buffer solution (pH 1-4) and equilibrated for 10 minutes.After which, 1 mL of MCH-para stock solution was added and equilibrated for 15 minutes in the dark.

UV-Vis samples were prepared as follows:
The stock solution of the MCH-para (C = 0.25 mM) was prepared by adding 14.6 mg (Molecular Weight = 584.38g/mol, 0.025 mmol) in 100 mL of DI water into a volumetric flask wrapped in aluminum foil and stored at 4 • C. Solutions for UV-Vis analysis were prepared in a 1 cm cuvette by adding 0.57 mL of the corresponding phosphate buffer (pH 1-10) followed by 2.052 mL of DI water, placed in the spectrophotometer, and allowed to equilibrate thermally for 10 minutes.Then 0.228 mL of MCH-para stock solution was added to the final concentration of C = 0.020 mM, with continuous stirring of the solution.Scans were collected after 15 minutes of equilibration in the dark.Spectral data matches the reported literature (1 ).(1 ).To an oven-dry round bottom flask equipped with a magnetic stir bar, a condenser was added (3-bromopropyl)trimethyl-ammonium dibromide (1.5 g, 5.75 mmol, 1.10 eq) followed by a 5-methoxy-2,3,3-trimethyl-3H-indole (0.99 g, 5.22 mmol, 1.00 eq.) in deoxygenated absolute acetonitrile (25 mL).The reaction mixture was heated to 100 • C and refluxed under a nitrogen atmosphere for three days.After cooling down to room temperature, acetonitrile was removed under reduced pressure.The resulting purple solid was dissolved in 70 ml of water and washed with diethyl ether (3 × 40 mL) to remove impurities.The aqueous phase was evaporated in a vacuum, and the product was dissolved in a minimum amount of methanol and precipitated in cold diethyl ether, filtered, and dried in a vacuum oven overnight.The final product contains a small amount of (3-bromopropyl)trimethyl-ammonium dibromide, which due to similar solubility, was impossible to remove.Therefore, it was used without further purification.1.69 g (72% yield).

Merocyanine form:
Surfactants are amphiphilic molecules that are typically classified by their hydrophobic and hydrophilic parts.Our hypothesis is that surface tension correlates with hydrophilicity, which is influenced by the polarity of the molecules.For SP-DA-PEG, the transition to the merocyanine form involves a shift from the less polar spiropyran to the more polar trans-merocyanine, which makes the molecules have a stronger affinity for water due to electrostatic interactions with water molecules.
Consequently, this increase in hydrophilicity causes the molecules to be less likely to accumulate at the interface but more dispersed in the bulk water, helping to maintain the cohesion between water molecules at the interface and sustain a high surface tension value.Similarly, the surface tension decreases when MCH-para switches from the more hydrophilic trans-form to the less hydrophilic cis-form.
Regarding the orientation of the molecules accumulated at the interface, we hypothesize that charged polar alkylammonium and hydroxy groups will be oriented toward the aqueous phase, while non-polar aromatic groups will be aligned toward the Krytox phase.Further investigation may be acquired by Cryo-EM in future studies.
In general, spiropyran is a photoswitchable molecule that changes under UV-light irradiation to Merocyanine form (MCH/MC, depending on conditions the molecule will be either protonated/deprotonated).
When the molecule has an electron-withdrawing -NO 2 group, the equilibrium is shifted to the nonpolar Spiropyran (SP) form in the dark (4 , 5 ); however, with the introduction of electron-donating methoxy groups to the light-responsive molecule, the molecular equilibrium is shifted to the more polar MCH/MC) form (1 ).So, in the case of photosurfactant on the left (Figure 1D), we switch between less polar Spiropyran and more polar trans-Merocyanine (trans-MC).In the second case, the molecule switches between more polar trans-Merocyanine (trans-MCH-para) and less polar cis-Merocyanine (cis-MCH-para).Previously, the cis-Merocyanine molecule ( 6) was attributed to protonated Spiropyran (SPH).However, it has been recently demonstrated (7 ) that it is more correct to refer to it as cis-Merocyanine (cis-MCH), which behaves similarly to SPH due to the apparent barrierless transition from SP to cis-MCH at low pH.

Temperature rise of droplet due to heating from light illumination
It is possible for the thermal Marangoni effect induced by irradiating the substances with light to partially contributes to the droplet motion.Surface tension of an aqueous solution containing the photo-responsive surfactants as a function of temperature was measured using the standard pendant drop method on a commercial tensiometer.The increase in temperature caused by illumination of light with specific intensity was estimated using an infrared (IR) camera (Telops M3K).To decouple the photo-Marangoni effect from heating effect under illumination, we first heated the solution to varying temperatures using a custom syringe heating system and then measured the surface tension.Several film heater plates were attached to the exterior of the syringe to heat the solution, and a thermocouple was attached to the needle to measure the temperature of the injected droplet (Figure S9A).As a result, the relationship between surface tension and temperature is linear, ranging from 20 • C to 60 • C (Figure S9B).As temperature increases, surface tension decreases: To further assess the thermal effect, a series of temperature measurement experiments have been conducted.The emissivity of an aqueous solution containing 1 mM MCH-para was calibrated with reference to the emissivity of deionized water.Images of the two droplets show comparable temperature profiles and heat distributions when two identically sized droplets of deionized water and MCH-para solution are subjected to the same room temperature and light conditions (Figure S9C).As a result, the emissivity of water, ϵ = 1, can be used to estimate the emissivity of MCH-para.The temperature rise in MCH-para droplet due to blue LED illumination with an intensity of 31.8 mW/cm 2 is small.
The collimated 470 nm LED used to illuminate the droplet is the same one utilized for measuring surface tension (Figure S10A).As comparison, the LED was first used to heat a water droplet of equal size, resulting in a temperature change of less than 1 • C (Figure S10B).Then, the MCH-para droplet was stabilized at room temperature.After 150 s of heating, the maximum temperature within the In the case of SP-DA-PEG, UV illumination has been observed to cause an increase in both surface tension and interfacial tension.Even this suggests that the direction of the photo-Marangoni stress must be reversed to overcome thermal influences and lead to movement away from the light source, it is essential to study how much the temperature increase when exposed to UV light.This study helps to figure out how to make bigger changes in surface tension by reducing thermal Marangoni effect that could interfere with this process.We utilized a T-type thermocouple probe to measure the temperature increase in the aqueous droplet.The experimental set-ups for both UV LED and laser illumination experiments are depicted in Figure S11.In these experiments, we placed a equal-sized droplet onto the surface.Uniform illumination was ensured on the droplet, preventing any movement across the surface during the procedure.For data precision, temperature readings were taken at 0.5-second intervals using a data acquisition system (Keysight DAQ973A).
In Figure S12A, we observe that the temperature rises and stabilizes after an increase of approximately 1.8 • C over 100 seconds.This stabilization is attributable to the relatively lower power of the UV LED in comparison to the UV laser.Thus, it is important to note that the migration of droplets under the UV laser is still influenced by thermal effects.This was further investigated using the same type of thermocouple probe to record the temperature changes, as depicted in Figure S12B.
Upon 50 seconds of direct exposure to UV heating, there was a measured increase in temperature by 3.1 • C.This modest rise was insufficient to induce a change greater than 1 mN/m in surface tension, referencing the response of water under varying temperatures (8 ).The minimal fluctuation can be ascribed to the effective thermal conductivity of the silicon surface, which was positioned on an optical table with a metallic surface, facilitating heat dissipation.
Figure S12C-D in the study present data concerning the temperature rise in a droplet floating on liquid lubricant under different conditions.In Figure S12C, the droplet experiences a temperature increase of 5.9 • C.However, this increase is also minimal and insufficient to bring about a reduction in surface tension of 1 mN/m.This observation underscores the relatively subdued impact of temperature From these observations, it becomes evident that despite the presence of the heating from light, its influence on the aqueous droplets with SP-DA-PEG is trivial.The more significant factor appears to be the photo-Marangoni effect, attributable to the compositional gradient within the droplet.We use Eqs.( 6), ( 7) and ( 8) in Eq. ( 3) and eliminate γ 1 , γ hi , and γ lo , obtaining: If Re ≪ 1, we can assume V I ∼ V .Introducing scaling constants c 1 and c 2 , we obtain an implicit equation for V : Eq. ( 10) above is solved by iteration and is used to plot the dotted line in Fig. 4E.We use γ B − γ A = 0.73 mN/m, as this is the value observed after repeated illumination cycles (as exemplified in Fig. S7B), and τ fwd = 0.437 s, τ rev = 2.19 s, µ = 1.24 • 10 −2 / kg/(m s), d = 1.3 mm and R = ( 4 3π V) 1/3 , where V is the drop volume.The constants we use are c 1 = 0.2 and c 2 = 0.15; we note these are of order-one, as should be expected for a physically plausible scaling theory.
It is instructive to seek an approximation of Eq. ( 10) for small drops.We assume that, as R approaches zero, V decreases more slowly than R, such that the ratio R/(V I τ rev ) also approaches zero.
Since in practice the beam diameter d is essentially constant, we assume that the ratio d/(V I τ fwd ) ∼ t i /τ rev ≫ 1, such that, in Eq. ( 10), we neglect the second exponential in the numerator, as well as the exponential in the denominator: Introducing a Taylor expansion in t d /τ rev , neglecting second-order terms, and solving for V , we obtain the explicit expression for the low-Re velocity: suggesting that for small drops the velocity is highly sensitive to volume, as At large drop volumes, we observe empirically that the velocity decreases slightly.There are several potential explanations for this observation.We tentatively consider here the effect of finite Reynolds number, as the larger drops have Re = ρV 2R/µ ∼ 4. For a self-propelled object, the propulsive thrust T scales as T ∼ ṁ∆V , where ṁ is the mass flow rate that is accelerated by a velocity increment ∆V .
Here ṁ ∼ ρ(V + ∆V )δ 2 , where δ is the half-width of the faster-moving fluid region behind the drop.
Equating the thrust with the Marangoni force ∼ (γ hi − γ lo )R we find For large Re ≫ 1, we use the canonical laminar flow scaling δ 2 ∼ R 2 Re −1 ∆V ∼ Rν∆V −1 .Assuming ∆V ≫ V , we approximate (V + ∆V ) ∼ ∆V in (13), such that the expression for γ hi − γ lo simplifies to To relate V and ∆V , we assume that at Re ≫ 1 the fluid resistance follows an inertial scaling, and is therefore proportional to ρV 2 R 2 .Balancing this fluid resistance with the Marangoni thrust T , Using again the assumption ∆V ≫ V , Eq. ( 15) reduces to Note that Eq. ( 16) can be rewritten to give which supports our assumption ∆V ≫ V .
Figure S1 to S21TableS1 program.A 10x10 mm 2 rectangular spectrophotometer cell was placed in a sample holder.The solution was continuously stirred by a miniature stirring plate inserted into the sample holder (Starna Cells SCS 1.11).A system of lenses directed the probe beam into the detector (Ocean Optics Flame-S1-XR spectrometer), which acquired the spectra of the probe light.The detector was connected to a PC via a USB port.The experiment was controlled by a National Instrument LabVIEW program which collected the probe light spectra, determined sample optical absorption spectra, and controlled pump and probe light sources.Kinetics experiments with SP-DA-PEG were performed in 0.05 mM solutions (SI Appendix, Figure S1A-B).

Figure S1 :
Figure S1: Pump-probe kinetics measurements of SP-DA-PEG in (A) water with alternative 365 nm and 530 nm lights and (B) toluene irradiated with 365 nm light.

Figure S2 :
Figure S2: Switching mechanism of SP-DA-PEG in toluene solution.

droplet rises from 19 . 6 •
C to 20.6 • C (Figure S10C-D), resulting in a 0.26 mN/m decrease in surface tension.This decrease is negligible compared to the total change of 4.4 mN/m in surface tension (pH = 3) depicted in Figure 1H.The thermal effect caused by irradiation only accounts for 5.9% of the total change in surface tension.For blue light with similar intensities, the photo Marangoni effect is the primary factor that induces shear flow and causes droplet motions.

Figure S9 :
Figure S9: (A) Experimental setup for measuring surface tension at various temperatures.(B) The response of MCH-para surface tension to temperature.(C) IR images of the deionized water (right) and the MCH-para aqueous solution droplet (left; pH = 3) at room temperatrue.Ranges on the legend bar represent temperatures in • C.

Figure S10 :
Figure S10: (A) Schematic of single droplet IR image capture under collimated LED illumination (31.8 mW/cm 2 ).(B) IR images of blue light-heated deionized water droplets.(C) IR images of 1 mM MCH-para water droplet heated by blue light illuminations.(D) Maximum temperature occurs within the MCH-para droplet over six minutes.

Figure S11 :
Figure S11: Measuring the temperature rise of 0.2 mM aqueous solution due to UV LED and laser heating.(A) Experimental set-up of UV LED heated aqueous droplet.(B) Experimental set-up of UV laser-heated aqueous droplet.

Figure S12 :
Figure S12: Temperature rise of 0.2 mM aqueous droplet on silicon wafer (A) upon UV LED illumination.(B) upon UV laser illumination.Temperature rise of 0.2 mM aqueous droplet on lubricant (C) upon UV laser illumination.(D) upon pulsed UV laser.

Figure S13 :
Figure S13: The changes in the color of SP-DA-PEG in toluene under UV illumination.(A) 0.1 mM SP toluene solution transforms from transparent (SP) to dark blue (MC) under 365 nm UV illumination.(B) After achieving equilibrium, the MC-rich solution under white light illumination changes color from dark blue to transparent.

Figure S14 :
Figure S14: (A) Experimental setup for measuring surface tension under repeated exposure to corresponding light.(B) Surface tension response of 0.2 mM SP-DA-PEG in water under UV (365 nm) illumination with optical intensities in the range of 7.7 to 37.1 mW/cm 2 .

Figure S15 :Figure S16 :
Figure S15: (A) Schematic of a liquid droplet on LIS under UV illumination.(B) Time-lapse optical images (top-down view) of water droplets containing SP-DAPEG on LIS, changing moving direction.)

Figure S17 :
Figure S17: Contact angle measurements of 1 mM MCH-para in water at pH 3. The contact angle of MCH-para water droplet on LIS reduced from 117°to 113°upon blue light (470 nm, 31.8 mW/cm 2 ) irradiation.

Figure S18 :
Figure S18: Optical images (side view) of a stationary pure DI water droplet illuminated by UV light with an intensity gradient after passing through an ND filter.The intensity of the light is greater on the left and weaker on the right.

Figure S19 :
Figure S19: Time-lapse optical images (side view) of linear movement of a water droplet containing MCH-para on LIS directed by blue light with an intensity gradient.

Figure S20 :
Figure S20: Time-lapse optical images (side view) of the internal flow field indicated by tracer particles of a water droplet containing MCH-para.Blue light was illuminated on the right side of the droplet.The directions and lengths of arrows indicate the directions and magnitudes of the velocity of the tracer particles.

Figure S21 :
Figure S21: Summary of literature studies: correlation between light intensity and droplet velocity in droplet manipulation via thermal and photo effects.