Nanoscale contact line visualization based on total internal reflection fluorescence microscopy

We describe a novel measurement method to study the contact line of a droplet at nanoscale level. The method is based on Total Internal Reflection Fluorescence Microscopy (TIRFM), which uses an evanescent excitation field produced by total internal reflection of light. The evanescent field depends on the angle of the incident light and has an exponential intensity decay, characterized by the penetration depth. The penetration depth is determined by imaging a fluorescent particle probe that is traversed using an Atomic Force Microscopy (AFM) setup. The result confirms the exponential behavior of the evanescent field intensity, and the value of the penetration depth also corresponds with the value predicted based on the optical configuration. By using the intensity distribution of a fluorescent dye and the value for the penetration depth of the evanescent wave, it is possible to reconstruct the interface of a partial wetting droplet. The reconstructed interface based on TIRFM is in good agreement with the interface obtained from two reference measurements: non-disturbing AFM-imaging and conventional contact angle measurement. The latter lacks spatial resolution, while the former is limited to particular droplets. This new non-contact measurement does not suffer from these drawbacks, making it a very useful tool to study the fundamental wetting behavior of both stationary and dynamic interfaces. © 2013 Optical Society of America OCIS codes: (180.2520) Fluorescence microscopy; (300.2530) Laser induced fluorescence; (260.6970) Total internal reflection; (180.5810) Scanning microscopy. References and links 1. C. Huh and L. E. Scriven, “Hydrodynamic model of steady movement of a solid/liquid/fluid contact line,” J. Colloid Interface Sci. 35(1), 85-101(1971). 2. P. G. de Gennes, “Wetting: statics and dynamics,” Rev. Mod. Phys. 57(3), 827-863 (1985). 3. L. M. Hocking. “A moving fluid interface. Part 2. The removal of the force singularity by a slip flow,” J. Fluid Mech. 79(2), 209-229 (1977). 4. J. D. Chen and N. Wada, “Wetting dynamics of the edge of a spreading drop,” Phys. Rev. Lett. 62(26), 30503053 (1989). 5. H. P. Kavehpour, B. Ovryn, and G. H. McKinley, “Microscopic and macroscopic structure of the precursor layer in spreading viscous drops,” Phys. Rev. Lett. 91(19), 196104 (2003). 6. A. Hoang and H. P. Kavehpour, “Dynamics of nanoscale precursor film near a moving contact line of spreading drops,” Phys. Rev. Lett. 106(25), 254501 (2011). 7. A. Hoang, G. Berteloot, P. Sharif-Kashani, and H. P. Kavehpour, “Dynamic measurement of microfilms and nanofilms of fluids using fluorescence microscopy,” Exp. Fluids 52(6), 1657-1662 (2012). #194600 $15.00 USD Received 25 Jul 2013; revised 6 Sep 2013; accepted 7 Sep 2013; published 24 Oct 2013 (C) 2013 OSA 4 November 2013 | Vol. 21, No. 22 | DOI:10.1364/OE.21.026093 | OPTICS EXPRESS 26093 8. M. N. Popescu, G. Oshanin, S. Dietrich, and A. M. Cazabat, “Precursor films in wetting phenomena,” J. Phys.: Condens. Matter 24, 243102 (2012). 9. D. Axelrod, T. P. Burghardt, and N. L. Thompson,“Total internal reflection fluorescence,” Annu. Rev. Biophys. Bioeng. 13(1), 247-268 (1984). 10. F. de Fornel, Evanescent Waves from Newtonian Optics to Atomic Optics (Springer, 2001). 11. A. Sarkar, R. B. Robertson, and J. M. Fernandez, “Simultaneous atomic force microscope and fluorescence measurements of protein unfolding using a calibrated evanescent wave,” Proc. Natl. Acad. Sci. U.S.A. 101(35), 12882-12886 (2004). 12. S. Harlepp, J. Robert, N. C. Darnton, and D. Chatenay, “Subnanometric measurements of evanescent wave penetration depth using total internal reflection microscopy combined with fluorescent correlation spectroscopy,” Appl. Phys. Lett. 85(17), 3917-3919 (2004). 13. A. Fery, T. Pompe, and S. Herminghaus, “Nanometer resolution of liquid surface topography by scanning force microscopy,” J. Adhes. Sci. Technol. 13(10), 1071-1083 (1999). 14. S. Herminghaus, T. Pompe, and A. Fery, “Scanning force microscopy investigation of liquid structures and its application to fundamental wetting research,” J. Adhes. Sci. Technol. 14(14), 1767-1782 (2000). 15. F. Mugele, T. Becker, R. Nikopoulos, M. Kohonen, and S. Herminghaus, “Capillarity at the nanoscale: an AFM view,” J. Adhes. Sci. Technol. 16(7), 951-964 (2002). 16. A. Checco, P. Guenoun, and J. Daillant, “Nonlinear dependence of the contact angle of nanodroplets on contact line curvature,” Phys. Rev. Lett. 91(18), 186101 (2003). 17. P. Fontaine, P. Guenoun, and J. Daillant, “A critical look at surface force measurement using a commercial atomic force microscope in the noncontact mode,” Rev. Sci. Instrum. 68(11), 4145-4151 (1997). 18. M. Franken, C. Poelma, and J. Westerweel, “Dynamics of precursor films,” presented at the 65th Annual Meeting of the APS Division of Fluid Dynamics, San Diego, USA, 18-20 Nov. 2012.


Introduction
The shape of a spreading droplet that partially wets a surface is described by the macroscopic contact angle.For moving contact lines, the assumption of a wedge profile for the droplet height leads to a singularity at the contact line [1].In order to overcome this predicted singularity, long range van der Waals forces [2] and slip velocities [3] have been introduced in models.Light interference patterns have been used to determine the shape of a spreading droplet [4].From these interference patterns, the meniscus shape of the droplet edge was reconstructed for the first time, and the advancing dynamic contact angle was measured.However, the effect of van der Waals forces on the shape of the interface has not been observed in the measurement range attainable by interference due to the inherent limitation of the wavelength of light.Nevertheless, the precursor film length could be obtained using interferometry [5], which is in the order of 10 −4 m.This precursor film is a thin layer that precedes the macroscopic contact line, and is thought to be a result of van der Waals forces.Epifluorescence inverted microscopy has been used to study the evolution of a precursor film [6,7].However, in these studies only a completely wetting fluid could be measured, with measurements restricted to observations during very long time scales (i.e.very slow variation of the contact line position).Up to now, this dynamic contact angle is undetermined below a film thickness of about 1 µm for a partial wetting fluid.Although a good understanding of the behavior of precursor films in the case of complete wetting is present, there is no experimental data regarding (the film profile of) adiabatic precursor films [8].This is related to the inherent difficulty of probing the region of a precursor film due to the great disparity of length and time scales involved.However, understanding the behavior of moving precursor films is the key to understanding moving contact lines.In order to obtain information of the interface of a partial wetting droplet below 1 µm, a new technique is required for doing these near-wall measurements.A technique that is capable of doing near-wall measurements is Total Internal Reflection Fluorescence Microscopy (TIRFM), which is not limited to the wavelength of the light in the direction perpendicular to the substrate, and therefore has the required sub-wavelength resolution to accurately measure film thickness.This paper describes a measurement technique with sufficient spatial as well as temporal resolution that enables, for the first time, precise characterization of the profile of an adiabatic precursor film and study its dynamics.

Total internal reflection fluorescence microscopy
The principle of TIRFM is based on the generation of an evanescent excitation field at the interface of two media with different refractive indices [9,10].When the incident light is beyond the critical angle, the light undergoes total internal reflection and generates a thin electromagnetic field perpendicular to the surface, as illustrated in Fig. 1(a).This electromagnetic field is called an evanescent wave and has a frequency identical to the incident light.Furthermore, the intensity of the evanescent wave decays exponentially with the distance from the substrate along the optical axis, as illustrated in the inset of Fig. 1(a).By limiting the illumination to a very thin layer at the substrate, this method offers the advantage of an unprecedented signal-to-noise ratio since only the particles or fluorophores within the penetration depth of the evanescent wave are producing signal towards the detector.profile of the evanescent field in TIR excitation as measured by a combined TIRFM-AFM setup (green).This curve was obtained by moving a fluorescent particle in the evanescent field while recording its intensity in water (similar to Sarkar et al. [11]) in order to validate the penetration depth for the incident angle setting (see section Total Internal Reflection Fluorescence Microscopy for details).For the first 200 nm the agreement between experiment and theory (black) is good, however beyond 200 nm deviations arise due to background noise of the camera.
The critical angle for total internal reflection (TIR) is given by Snell's law where n 1 and n 2 (n 2 < n 1 ) are the refractive indices of the substrate and medium respectively.The evanescent field intensity has an exponential decay with perpendicular distance z from the substrate where I 0 is the intensity of the evanescent field at the surface of the substrate (z = 0) and depends on the intensity of the incident beam, the incident angle and the polarization [9,10].The penetration depth d is defined as the distance from the substrate where the intensity of the evanescent wave decays to 1/e of its original value, as illustrated in the inset of Fig. 1(a).It can be calculated using [9,10]: where λ and θ are the wavelength and incident angle of the light respectively.The penetration depth is in the order of λ or smaller, as shown in Fig. 1(b).
TIRFM is a near-field microscopy technique that operates less than one λ away from the image plane, which typically lies on the surface of the substrate.Its resolution in z is higher than a far-field technique can attain in a near-wall region.This is because the evanescent excitation field contains information beyond the diffraction limit.Therefore, a result is obtained with a higher out-of-plane resolution.Similar to a far-field technique, the in-plane resolution is represented by the Point Spread Function (PSF).The PSF is experimentally obtained from the diffraction limited spot of a 40 nm fluorescent bead, and is approximately 300 nm, which is slightly larger than the theoretical diffraction limit: the diffraction limit is estimated to be λ /2NA, which is approximately 200 nm for the objective and fluorescent dye used.

Penetration depth validation
The novel approach that is proposed here for obtaining the contact angle of a partial wetting droplet at a nanoscale is based on TIRFM.In particular, the technique relies on the fact that the illumination, confined to the region close to the surface, has a well-defined intensity decay.By adding a fluorescent dye to a droplet, the local fluorescent signal can be used to determine the local droplet height (see Section 3.2 for details).This technique requires exact knowledge of the penetration depth.For a given system, the penetration depth is a function of incident angle only (Eq. ( 3)).Since the incident angle can be difficult to determine accurately [12], the penetration depth is calibrated by combined TIRFM-AFM, as illustrated in Fig. 2(a).Detailed information about the equipment is given in Section 4.2.The characterization was done in order to validate the maximum incident angle setting in demineralized water (refractive index equals 1.332 ± 0.0005 at 22 • C).Water is chosen as a working fluid for this calibration, because it is compatible with the fluorescent particle (and the adhesive attaching the particle to the cantilever), as well as the optics of the liquid scanner of the AFM-head.The system is operated at its maximum incident angle, which is given by θ max = sin −1 (NA/n 1 ), where NA and n 1 are the numerical aperture of the objective and refractive index of the substrate respectively.Figure 2(b) shows an experimentally determined particle intensity (I p ) versus distance (z) curve.This curve is then normalized with the particle intensity at z = 0 (I 0 ) to compare it with the theoretical result.The experimentally determined penetration depth equals 58 ± 4 nm, which is much smaller than the value where the curve starts to deviate, as shown in Fig. 1(b).Furthermore, the experimentally determined value for the penetration depth is close to the theoretical value of 63 nm in water (estimated using the effective NA of the system).During the experiments, we keep the incident angle at the maximum angle (θ max ), yet replace the fluid.As we know the value of the refractive index of the fluid (hexaethylene glycol (HEG), n = 1.462 ± 0.0005), we can use Eq. ( 3) to determine the value of the penetration depth of this medium (d = 147 nm).

Intensity profile at droplet interface
By using a high-magnification objective, the field of view (FOV) is small compared to the radius of the droplet.This allows us to average the intensity distribution over a number of pixels (typically 10) along ∆y, without the need for introducing a polar coordinate system, as illustrated in Fig. 3(a).The recorded images (typically 250) of the edge of a HEG droplet were processed in order to obtain a curve of the intensity profile along the edge of the droplet, as shown in Fig. 3(b).At the edge of the droplet, the imaged intensity profile is a function of spatial coordinates (x, y).The imaged or recorded intensity in gray values (per pixel) I im (x, y) is used together with the nominal value of the penetration depth of the evanescent field to convert this intensity profile to an interface h(x, y).The recorded intensity follows from the integral of the evanescent intensity (that excites fluorescence) along the optical (z-) axis at each (x, y) point.For clarity, we omit the dependence on (x, y), it follows where I is the intensity integrated over the pixel area, and is equivalent to the image gray value.By substituting Eq. ( 2) and splitting the integral at the local surface height h(x, y) at the edge of the droplet, it follows I 0 e −z/d dz . ( Any light collected on the CCD of the camera can only result from fluorophores, which are only present in the droplet (see the inset of Fig. 3(a)).The second term on the right-hand side is therefore zero.Evaluating the integral in Eq. ( 5) gives an expression for the imaged intensity at the droplet edge with d the penetration depth of the evanescent field and I im,background the equivalent intensity due to dark background current of the image sensor of the camera.The value for I im,background can be obtained from parts of the image that are known to be sufficiently far away from the surface (e.g. the far right in the image in the inset of Fig. 3(a)) or by recording a separate set of images.By recording the intensity in the bulk of the droplet I im,bulk , which is the region where the interface h is much larger than the penetration depth d (i.e.h → ∞ effectively), I 0 can be determined using Eq. ( 6) Note that I 0 is a function of (x, y) and Eq. ( 7) thus takes care of non-uniform illumination effects; in a correctly aligned TIRF-microscope, the evanescent illumination has an elliptical Gaussian intensity profile [9].The interface of the droplet h(x, y) is determined by substituting Eq. ( 7) in Eq. ( 6) and rewriting this to By measuring the intensity profile of fluorophores in a partial wetting droplet, as illustrated in Fig. 3(b), the nanoscale interface can be reconstructed using Eq. ( 8).

AFM-imaging of droplet interface
AFM-imaging is well known for its unprecedented resolution and accuracy in the scanning of surfaces.The scanning range attainable in z is in the order of µm with nm accuracy, and therefore suitable as a reference for the TIRFM measurements.However, AFM-imaging of liquid droplets deposited on a substrate is only possible under strict conditions, such as a pinned contact line of a viscous droplet with high surface tension.By operating the AFM-head in non-contact mode, the interface of a partially wetting droplet can be measured without perturbing it.In non-contact mode, the cantilever is driven to oscillate at its resonance frequency.The oscillating amplitude is kept constant in a feedback loop of the system, by changing the distance between tip and sample as it scans the interface of the droplet.A detailed procedure of the imaging process of very small droplets by AFM is described in several other studies [13][14][15][16].

AFM probes
For the calibration of the evanescent wave (see Section 3.1), custom-made single particleattached probes were used.Standard silicon tipless contact-mode AFM probes (Novascan Technologies, Ames, USA) with a spring constant of 0.95 N/m formed the basis of these particle-attached probes.A 545 nm polystyrene Rhodamine B labeled bead (Microparticles GmbH, Berlin, Germany) was attached to the cantilever.
The shape of the droplet interface was determined using commercial ATEC-NCAu (160 × 46 × 4.6 µm 3 ) non-contact (or tapping) mode cantilevers with a tetrahedral tip (with a height of ≈ 15 µm) that protrudes from the very end of the cantilever.This particular kind of tip-geometry was required in order to avoid cantilever-sample interaction.The cantilevers had a typical resonance frequency of 335 kHz and a tip radius of about 63 nm according to the supplier (NanoAndMore GmbH, Wetzlar, Germany).Furthermore, the geometry of the tip has the advantage that acoustic effects due to viscous damping are minimized.These effects can cause additional damping and thus affect the amplitude of oscillation, and therefore the accuracy of surface determination [17].

Optical imaging by combined TIRFM-AFM
The experiments were conducted with an inverted Eclipse Ti TIRF-microscope (Nikon Inc, Tokyo, Japan) combined with a CombiScope 1000 AFM-head (AIST-NT BV, Apeldoorn, the Netherlands) positioned on an optical table (Newport RS 4000, Irvine, CA, USA).The inverted microscope was equipped with a Nikon APO60×/1.49oil-immersion objective and created an evanescent excitation field by a through-the-objective method [9,11].In this method the incident light is directed though the objective, and is reflected back into the objective, while creating an evanescent field.Consequently, a refractive index-matched immersion oil (type NF, refractive index = 1.515;Nikon Inc, Tokyo, Japan) is required for the standard glass substrates (Menzel-Gläser, Braunschweig, Germany).After cleaning, the coverslips had a typical RMS surface roughness of smaller than 1 nm as measured by AFM.The evanescent excitation field was generated by a continuous wave 150 mW Nd:YAG laser (Coherent, Compass 315M; Santa Clara, CA, USA), which was coupled with the microscope using a monomode fiber.The emitted light from the fluorophores passed through a 532 nm dichroic mirror (Chroma Technology, Bellows Falls, USA) and a 532 nm RazorEdge long-pass edge filter (Semrock, Rochester, NY, USA).These filters separate the longer-wavelength fluorescence from the evanescent wave illumination so that only fluorescence was imaged.A cooled 12-bit dynamic range high resolution charged-coupled device (CCD) camera (FlowMaster, 1376 × 1040 pixels, individual pixel size 6.45 × 6.45 µm 2 , LaVision GmbH, Goettingen, Germany) was used to capture the emitted light.This camera has a very linear response to incident light (non-linearity 1% according to the supplier), making it suitable for quantitative photometric analysis.For the 1300 nm laserdiode of the AFM-head, no additional filters were necessary since the CCD of the camera was not sensitive to light with a wavelength larger than 1100 nm.Physically decoupling the camera from the microscope, in order to minimize vibrations from the camera fan, reduced the amplitude of the vibrations to smaller than 1 nm.In addition, an identical camera equipped with a long-distance microscope with magnifications between 0.8× and 4× (Macro Vario Lens, LINOS Photonics GmbH, Munich, Germany) was used for the macroscopic contact angle measurements.The images were recorded using DaVis 7.2 (LaVision GmbH, Goettingen, Germany).

Working fluid
In order to avoid significant evaporation of the droplet in the timescale of the set of experiments (typically less than a hour), a liquid is required that has a sufficiently low vapor pressure at room temperature (i.e.non-volatile under ambient conditions).Furthermore, high surface energy is desirable in order to have stable AFM-imaging conditions for the validation study [15].The TIRFM measurements do not require these strict conditions, because the time required for the measurement is much shorter than AFM-imaging and it uses the intensity distributions instead of tip-sample interaction.Any fluid with a fluorescent dye will work for TIRFM as long as the refractive index is smaller than the one of the glass substrate.The intensity distributions as well as the interface scans are obtained from a hexaethylene glycol (HEG) droplet (Sigma-Aldrich, Saint Louis, MO, USA) with a Rhodamine B fluorescent dye (Sigma-Aldrich, Saint Louis, MO, USA).It should be noted that for the interface determination with both methods, the same droplet is used.The fluorescent dye has a concentration in the order of 10 −4 mol/liter.The concentration is adjusted in such a way that for large incident angles sufficient intensity remains, but self-absorption remains negligible.Therefore, measurements with sufficient signal-to-noise ratio were possible.Furthermore, an Abbe refractometer is used to determine the refractive index of HEG with and without Rhodamine B and equals 1.462 ± 0.0005 in both cases.Note that all the measurements were performed in ambient conditions at room temperature (around 22 • C) and a relative humidity of about 40%.

Results
The experiments were done for different partial wetting conditions, which were obtained by cleaning the glass substrates differently, such that two different macroscopic contact angles were measured.The reconstructed interface of a partial wetting droplet by TIRFM is shown in Fig. 4, and clearly shows the different length scales involved (a ratio of 1000 between the scales of the axes).The results were obtained by illuminating a partial wetting droplet by an evanescent excitation field with a nominal penetration depth of 147 nm.This value was obtained using Eq.(3) for the maximum incident angle (θ max , which was validated using a calibrated intensity-versus-distance curve in water).A clear saturation of the fluorescent signal is represented as the horizontal part (i.e.x = 0 to 2.1 µm) of the interface that approaches the penetration depth, as shown in Fig. 4.
Furthermore, the results from non-disturbing AFM-imaging and conventional contact angle measurement using the side view of the droplet (CA macro ) are also plotted in Fig. 4. Multiple measurements on several sides of the droplet were conducted in order to ensure that the experimental data is representative.A direct comparison between the reconstructed interface with the different measurement techniques shows that these are in good agreement.Possible influence of, for example, reflections at the interface are therefore negligible.The scatter in the reconstructed interface by AFM-imaging is in the order of several nm, and is likely caused by the tip sticking to the droplet interface.The fluctuations on the interface reconstructed by TIRFM are in the order of nm as well.
Figure 4(a) demonstrates that contact line visualization by TIRFM is in good agreement with AFM-imaging.Both techniques capture the shape of the nanoscale interface well at a relatively small macroscopic contact angle.However, at large contact angles AFM-imaging has difficulties following the interface of the droplet due to the transition from a very soft, inclined surface to a flat, hard surface, as can be seen in Fig. 4(b).Consequently, for larger contact angles the near-wall results deviate from TIRFM (i.e.x = 2.4 to 3.5 µm), because TIRFM does not suffer from this transition since it relies on the fluorescent signal.As can be seen in Fig. 4, at approximately h = 30 nm the droplet interface changes from straight to curved due to perturbation by van der Waals interaction.This occurs for both contact angles, and is consistent between both TIRFM measurements.However, no direct comparison with other experimental data for a partial wetting fluid is possible.This is because there is no data available of the measured interface below 30 nm for a partial wetting fluid [8].It should be noted that the purpose of this paper is the description of the developed measurement technique and its validation by AFM-imaging.Further discussion of wetting is beyond the scope of this paper.In both cases the length of the precursor film is much larger than the PSF of the system (which is approximately 300 nm, see text).An error of about 7% is expected for h due to the uncertainty in determination of the penetration depth d.
In order to see if any effect of photobleaching is present, a very long recording of 1000 images at a recording rate of 1 Hz is done (i.e. about 20 minutes).Both the nanoscale interface obtained from the first and last image agree very well with the interface averaged over the total amount of images.From this result it is evident that photobleaching does not occur or does not affect the measurements.This is expected, because in evanescent wave based illumination only a fraction of the 150 mW laser power is used for the excitation of the fluorescent dye.

Conclusion
The results demonstrate that TIRFM is able to reconstruct the interface of a partial wetting droplet at nanoscale level.Moreover, the results are in good agreement with the results from non-contact AFM-imaging.Furthermore, a sensitivity analysis indicate that the technique is not sensitive to photobleaching.By using short exposure times O(ms) TIRFM is not limited to reconstructing static interfaces at nanoscale level in the case of a pinned droplet edge.Hence, with this technique it now becomes possible to study the fundamental behavior of a dynamic interface of a partial wetting droplet at nanoscale level.Such measurements are part of ongoing studies [18].

Fig. 1 .
Fig.1.(a) TIR illumination scheme and coordinate system used, (b) Normalized intensity profile of the evanescent field in TIR excitation as measured by a combined TIRFM-AFM setup (green).This curve was obtained by moving a fluorescent particle in the evanescent field while recording its intensity in water (similar to Sarkar et al.[11]) in order to validate the penetration depth for the incident angle setting (see section Total Internal Reflection Fluorescence Microscopy for details).For the first 200 nm the agreement between experiment and theory (black) is good, however beyond 200 nm deviations arise due to background noise of the camera.

Fig. 2 .
Fig. 2. (a) Schematics of TIRFM-AFM setup for the evanescent wave calibration, (b) fluorescent particle intensity I p as a function of distance z of the evanescent field in TIR excitation as measured by a combined TIRFM-AFM setup.The inset show a 0.545 µm fluorescent particle within the field-of-view providing signal towards the CCD at various z-locations (NB: the particle images are inverted for clarity).

Fig. 3 .
Fig. 3. (a) Schematics of the measurement of the intensity profile at the interface of a partial wetting droplet.The inset shows the contour of the illuminated millimeter-sized HEG droplet within the field of view (FOV), (b) interface detection of a partial wetting droplet using an evanescent excitation field and the resulting intensity profile.

Fig. 4 .
Fig.4.Nanoscale interface h as a function of x (distance with respect to an arbitrary reference point) of a partial wetting HEG droplet on different substrates with a macroscopic contact angle of (a) 8 • , (b) 33 • , as measured with a side view camera.In both cases the length of the precursor film is much larger than the PSF of the system (which is approximately 300 nm, see text).An error of about 7% is expected for h due to the uncertainty in determination of the penetration depth d.