6.2.1 The Challenges behind the Preparation of a Thin Liquid Film

Ideally, in traditional LIFT of liquids, one would like to have thin films with high temporal stability and spatial uniformity over large areas (∼mm²). Despite the preparation of a liquid film being simple and straightforward, this task can be highly challenging. In fact, thin liquid films are often subjected to several instabilities that compromise their spatiotemporal uniformity [40]. Depending on the mechanism that triggers these instabilities, two main groups can be recognized. The first one involves spontaneous instabilities, namely the intrinsic tendency of a thin liquid film to break into smaller droplets (Figure 6.2a). Typical examples are the Rayleigh–Taylor instability, caused by gravity forces that destabilize a suspended (facing down) film, and liquid dewetting, in which the liquid film is ruptured due to its tendency to withdraw from a “hostile” surface [41, 42]. The second group of instabilities encompasses those induced by external agents (Figure 6.2b) [43], as in the case of capillary waves, that is, pressure waves traveling along the phase boundary of the liquid. Eventually, when the external perturbation ceases, these nonuniformities may vanish due to viscous dissipation. For a given magnitude or frequency of the perturbation, though, the instabilities could also result in permanent film breakup.

The occurrence of liquid film instabilities strongly depends on several rheological parameters of the printed material or ink, namely surface tension (γ), mass density (ρ), vapor pressure, and viscosity (υ), as well as on the morphology (e.g., surface roughness) and chemistry (e.g., level of hydrophobicity) of the carrier substrate. Because there exists an interplay between all these parameters, understanding how they affect the donor preparation is not an easy task. To simplify the study, it is useful to distinguish two main stages in which liquid nonuniformities can appear during LIFT. The first one includes the initial preparation of the liquid film, and it is dominated by spontaneous instabilities. The second stage occurs once a uniform film is obtained and involves the formation of capillary waves induced by the laser pulses that act as the external perturbing agent. A more detailed description of each of these stages is presented next.

6.2.1.1 The Role of Spontaneous Instabilities

To understand the challenges behind the initial preparation of a thin liquid film, it is necessary to first clarify the meaning of the term thin. In the context of printing liquids with LIFT, we consider “thin” those films with a thickness in the interval of 5–50 µm. Within this length scale, gravity and consequently the Rayleigh–Taylor instability are negligible. Thus, liquid dewetting is the only potential source of spontaneous liquid instabilities. Importantly, liquid dewetting occurs for films below a critical thickness e_C given by the equation

6.1

where k⁻¹ = (γ/ρg)^1/2 is the capillary length, and γ, ρ, g, and are the ink surface tension, density, gravitational constant, and equilibrium contact angle, respectively [44]. From Eq. (6.1), it is possible to unveil the main parameters that affect film stability, which as stated earlier, are not limited to the ink rheology, but also include the chemical and morphological properties of the carrier substrate.

For the sake of simplicity, we limit the discussion on donor film dewetting to two inks, water ( = 73 mN/m, = 1 × 10³ kg/m³, and = 1 mPa s) and glycerol (γ_glyc = 63 mN/m, ρ_glyc = 1.3 × 10³ kg/m³, and υ_glyc = 1.5 Pa s). Note that these two inks are widely used as model systems in LIFT [45, 46]. Table 6.1 presents a list of the critical thickness e_C for various materials used in LIFT. As expected, only when the wettability between the liquid and the carrier substrate is high, it is possible to obtain stable films with a thickness suitable for LIFT.

Material	(°)	eC
Metal	∼0	∼100 nm
Oxide (e.g., glass)	∼10–30	∼1–10 µm
Polymer	∼60–110	∼50–500 µm

Even if the critical thickness of the two model inks can be, in many cases of practical interest (e.g., metal absorbing layers), well below the typical thickness used in LIFT, the risk of dewetting cannot be completely averted. First of all, any imperfections of the absorbing layer can induce local instabilities that trigger the rupture of the film. To avoid this problem, it is imperative to have clean and uniform carrier substrates. Moreover, several metals oxidize upon exposure to air resulting in a thin oxide film that reduces ink wettability. Thus, inert noble metals such as Au are, a priori, more suitable for LIFT. Finally, because liquid evaporation decreases the film thickness d₀, at least locally, the film stability is limited overtime, until d₀ < e_C. Therefore, the kinetics of film evaporation determines the temporal window during which the donor film remains stable and suitable for printing.

The liquid film evaporation rate () can be estimated for a model case, namely the liquid evaporation at the free vapor–liquid interface. This phenomenon is governed by the general form of Dalton's law that can be written as

6.2

where h_M is the mass transfer coefficient, S the surface area of the film, the relative humidity, with R₀ the gas constant, the vapor molar mass, T the temperature, P_b the vapor pressure, and P_sat the saturated vapor pressure. Importantly, h_M is a function of additional parameters among which the vapor diffusivity D and the velocity U of the surrounding air (convection) are the most relevant. Remarkably, Eq. (6.2) introduces an additional ink parameter, that is, the ink vapor pressure, which further limits the range of the ink candidates for LIFT. Indeed, the ideal ink should have a low evaporation rate. Note that many organic solvents, which typically feature excellent wettability over a wide variety of substrates, can be discarded as ink candidates because of their typically high vapor pressure. For a water film under standard laboratory conditions, = 60%, T = 293 K, and U from 0.01 to 1 mm/s, the evaporation rate ranges from approximately 70 to 700 pg/m² s, which corresponds to a thickness reduction rate ranging from approximately 0.07 to 0.7 µm/s. Thus, a water film with a thickness of about 10 µm prepared on glass becomes unstable () within a timescale ranging from tens of seconds to few minutes. This is impracticable for many applications and could limit future industrial implementations of LIFT. In contrast, a glycerol film exhibits good stability over time (evaporation rate nm/s) thanks to its low vapor pressure (∼10 mPa) and high viscosity (1.5 Pa s). As detailed in the next sections, viscosity can be a drawback during the interaction of the liquid film with the laser irradiation. Highly viscous materials are also difficult to spread, and hence, in practice, it is difficult to prepare them in thin form. This is why glycerol is mainly used as an additive for decreasing the vapor pressure of aqueous solutions.

6.2.1.2 The Role of External Instabilities

Once a stable donor film is obtained, external agents can degrade the spatiotemporal stability of the liquid. In the case of LIFT, the irradiation of the film with laser pulses is the main external factor that can lead to liquid instabilities. Indeed, as a possible consequence of the liquid–matter interaction, capillary waves are usually generated [47]. The dynamics of these waves can degrade the performance of LIFT. For instance, when two laser pulses are fired in close temporal and spatial proximity, the presence of capillary waves induced by the first pulse can affect the printing process initiated by the second pulse [47, 48] (see Chapter 5 for details). To overcome this issue, the second pulse should be fired temporally or spatially separated from the first pulse. A quantification of the minimal separation (in time or space) between consecutive pulses for which interference by capillary waves is avoided can be estimated by using the lubrication approximation. In this case, capillary waves within a thin liquid layer decay exponentially with a relaxation time given by the equation

6.3

where is the absolute or dynamic viscosity of the ink, d₀ the film thickness, and λ the wavelength of the external disturbance with [44]. Thus, two pulses that are temporally separated by less than will interfere. This temporal separation can be easily transformed into a spatial distance by considering the propagation velocity of capillary waves , which is given by

6.4

Thus, the minimal spatial distance between two consecutive pulses (), by combining Eq. (6.3) with Eq. (6.4), is

6.5

Under typical operational conditions, for example, d₀ = 10 µm and λ = 40 µm [47], for a water–glycerol mixture ranges from 35 µs for pure water to 40 ms for pure glycerol, whereas the corresponding would be 180 µm and 180 mm, respectively. A simple estimate of the maximum repetition rate of the printing process () when the spatial separation between events is below , as it occurs in most practical instances, would scale with , which leads to 28 kHz for water and 25 Hz for glycerol. However, in this assumption, the temporal dynamics of the printing process have not been considered. Thus, the time required for liquid ejection, denoted as , further limits . In other words, , where also depends on viscosity. These results further highlight the crucial role of the ink rheology on the performance of LIFT, which does not only affect the preparation of the liquid film but can also later alter the throughput of the technique.

Importantly, in the case of FF-LIFT, the values of and can be greatly reduced. In fact, the damping effects in a bulk liquid help mitigating capillary wave effects [44]. Thus, the use of a film-free approach is less demanding in terms of the rheology of the inks to be printed.

6.2.2 Technologies for Thin-Film Preparation

Even if an ink with the ideal conditions required for LIFT printing was found, one should still look for the optimal existing technology that enables the spreading of the ink into a thin film. Several techniques are now available that enable controlled preparation of thin films at the 5–50 µm range required in LIFT. A brief description of the main features and properties of some of the most used techniques in LIFT is provided next.

1. Spin coating. In this approach, a small amount of ink is deposited on the center of a flat substrate, which is then rotated at high speed to uniformly spread the ink over the substrate surface (Figure 6.3a) [49, 50]. During the rotation step, solvents that are present in the ink totally or partially evaporate, leading to a film whose thickness depends on both the rotation speed (ω) and the ink viscosity (υ). In particular, the film thickness decreases with ω, whereas, it increases with υ. Therefore, the use of spin coating further limits the range of available inks for LIFT. This is particularly crucial for polymer inks, since they usually feature high viscosity that limits the minimum thickness of feasible films. Furthermore, this method is not appropriate for water-based inks since a fine control of the water evaporation rate is difficult.
2. Dip coating. This technique is based on a two-step process (Figure 6.3b). First, the carrier substrate over which the liquid has to be prepared is immersed in the ink. Second, the substrate is withdrawn at constant speed [50]. During this second step, a thin layer of ink deposits on the substrate, and the retracting speed determines the final thickness. In general (for Newtonian fluids, as the model solutions considered earlier), the faster the withdrawal process, the thicker the resulting film. Compared to spin coating, this method is compatible with water-based films and features a simpler implementation. For highly viscous inks, though, the uniformity of the film can be compromised.
3. Blade coating. This method is probably the most used in LIFT thanks to its ease of implementation and limited cost [27, 45]. Initially, a tiny amount of ink is placed in an outer marginal region of a flat substrate. After that, a blade is used to coat the entire substrate with the ink (Figure 6.3c). Because the gap between the blade and the substrate sets the final thickness of the coating, uniform films are feasible regardless of the ink viscosity. Despite its simplicity, any variations of the coating speed compromise the film uniformity.

6.2.3 Wetting of the Receiver Substrate

There is an additional key factor to be considered in LIFT, and that ultimately dictates the outcome of this printing process: the properties of the receiver substrate. Indeed, the amount of ink deposited on the receiver substrate does not exclusively depend on the rheological properties of the ink or the laser parameters, but it can also be affected by the wetting of the receiver substrate. In LIFT, the printing process is mediated by the formation of a jet that impacts upon the receiver substrate (Chapter 4). Thus, the jet–receiver interaction (wettability, adhesion forces) determines aspects as relevant as the laser fluence threshold required for printing [51] or the final amount of deposited material [52]. As a result, an ink that presents rheological properties optimal for the preparation of the donor film may not necessarily result in good printing. For instance, the addition of surfactants in the liquid to be printed can facilitate liquid wettability on the carrier substrate and thus enables the preparation of stable liquid films. However, the decrease of the surface tension of the ink can also promote wetting on the receiver substrate, resulting in printed droplets with a diameter or volume larger than the ones obtained without the use of surfactant, which is detrimental in terms of printing resolution. This limitation is the consequence of the intrinsic dependence of the LIFT performance on the entire carrier/donor/receiver system. In other words, in traditional LIFT, there is a fundamental coupling between the receiver substrate and the carrier/donor system in terms of the ink properties. Thus, the choice of the ink will always be a trade-off between the optimal ink for the receiver substrate and the ideal rheological properties of the ink for its preparation in film form. In this scenario, the reader can catch a glimpse of the great potential of FF-LIFT, where the ink simply stays in a reservoir, and its rheology can be optimized only with respect to the receiver substrate. Therefore, the film-free approach provides LIFT with the ability to tune the ink with respect to a targeted application rather than to optimize the ink for its suitability as a donor film.

6.3.1 Cavitation-Induced Phenomena for Printing

First, we need to consider how to promote significant liquid ejection or large liquid transport using a laser. This essential step for printing is challenging to implement in practice, and it usually imposes strong requirements on the laser system or liquids to be printed. For instance, in highly light-absorbent liquids, laser-induced temperature gradients can drive large-scale fluid flows [53]. This mechanism, though, is incompatible with biological solutions, which are typically transparent to the laser radiation. In contrast, liquid actuation by cavitation bubbles fulfills the requirements desired in a laser printing technique, namely the possibility to operate with a large amount of liquids (provided they are transparent to the laser radiation) and control of the printed material by means of the laser parameters. Indeed, laser-induced cavitation constitutes the fundamental mechanism of FF-LIFT.

The typical sequence of events in the formation of laser cavitation bubbles starts when a laser pulse is focused within a liquid. Provided the laser intensity is high enough (see Section 6.4), optical breakdown in the liquid can be induced [54, 55]. In analogy to the well-known electrical breakdown of a dielectric under strong electric fields, the molecules of a fluid can be ionized in the presence of laser radiation. As a consequence, a confined and hot plasma can be formed that explosively expands at a velocity faster than the speed of sound in the liquid (i.e., shock wave), resulting in the generation of a cavitation bubble [56, 57]. More importantly, the evolution of this bubble, including its initial growth and final collapse, can induce powerful fluid flows. Depending on the particular conditions of cavitation bubble formation, different scenarios can develop. In particular, when the bubble is generated close to a free surface, a jet perpendicular to the surface is formed, which transports a fraction of liquid away from the surface [58]. The formation of these cavitation jets dictates the outcome of the printing process in FF-LIFT. Thus, it is paramount to understand the interplay between cavitation bubble and jet formation in this printing technique.

6.3.2 Jet Formation in Film-Free LIFT

The initial dynamics of jet formation in FF-LIFT is dependent on only one parameter, the so-called standoff distance , defined as [59]:

6.4

where R_max is the maximum bubble radius (depends on the laser pulse energy absorbed), and h is the distance from the free surface to the initial bubble centroid. In general, if > 0.8, the cavitation bubble is far from the free surface for its given R_max, and only a small protrusion in the surface is formed, which impedes proper printing. In addition, if < 0.3, bubble bursting can occur prior to jet formation, resulting in splashing in the receiver substrate. Thus, the suitable range for printing in FF-LIFT is usually 0.3 < < 0.8 [60, 61]. Note that surface tension does not affect the initial formation of a jet, since bubble growth and collapse are inertia-dominated processes. At this stage, the Weber number, which describes the importance of inertial over surface tension forces, is typically large, up to 10⁴ [60]. Moreover, viscosity is also found to have little effects at this initial stage [62]. Indeed, inviscid models have been found to be in good agreement with the experimental observations [36, 58, 59].

Figure 6.4 presents a temporal sequence of the jet dynamics in an FF-LIFT experiment with [60]. Simulations obtained with computational fluid dynamics (CFDs) are also shown for comparison [62]. Despite the different temporal and spatial scales for experiment and simulation, the overall dynamics show a remarkable good agreement. After bubble formation, the cavitation bubble grows caused by hot gases created during plasma formation (Figure 6.4a). Notably, the growing is asymmetrical, due to the presence of the free surface that disrupts the symmetry of bubble expansion and results in a pressure gradient around the bubble. In other words, because it is easier for the bubble to expand against the free surface than against the bulk liquid, a pressure gradient around the bubble is formed (Figure 6.4b). As a consequence, fluid flows are generated that move toward the bubble tip, resulting in the development of a jet. Moreover, bubble expansion causes a decrease in the inner bubble pressure, which finally produces the collapse of the bubble (Figure 6.4c). During this phase, a second jet is formed that moves downward. This second jet, also called counter-jet, passes through the bubble and impacts its bottom surface. At this point, a second gas pocket is formed that detaches from the initial bubble (Figure 6.4d). Note that the initial bubble remains toroidal due to the counter-jet piercing, as clearly appreciated in the simulation. The two bubbles then initiate an expansion phase followed by collapse (Figure 6.4e). Interestingly, during the expansion phase of the toroidal bubble, the same situation as described previously occurs, namely the formation of a pressure gradient around the bubble and the consequent flow of liquid toward the low-pressure region, that is, the free surface. In this case, given the toroidal shape of the bubble, a corona or bulge is formed (Figure 6.4f). Depending on the experimental conditions, this corona could be disrupted due to surface tension forces (as observed in the simulations, where the We number decreases to a value of approximately 50 [62]) or could progress unperturbed (as in the experiment). In the case that inertial forces overcome viscous effects, both jet and corona continue advancing, until impacting the receiver substrate (not shown). At a low laser fluence (just above the threshold for printing), though, the corona may not reach the receiver substrate, even if the jet is energetic enough to do so. This would explain the steep transition observed in the amount of printed volume as the laser fluence increases [38]. Finally, instabilities created by surface tension (Plateau–Rayleigh instability) produce jet breakup [63]. Importantly, during jet propagation, viscosity also plays a crucial role, since it helps jet stability and hence can prevent the formation of satellite droplets during printing. Note that stable jets are highly desirable since they provide flexibility in the positioning of the receiver substrate relative to the liquid surface.

6.3.3 Differences with LIFT of Liquids

As described in Chapter 4, traditional LIFT of liquids is also a printing process based on cavitation jets. Thus, it comes as no surprise that the dynamics between this technique and FF-LIFT regarding cavitation bubble generation and jet development are similar. However, there are distinct peculiarities of FF-LIFT that are worth detailing. First, the corona or bulge observed in FF-LIFT is not typically identified in time-resolved imaging studies of LIFT [35, 48, 64–67]. A possible interpretation of this fact is the use of a thin liquid film in LIFT (5–20 µm), which would prevent a full development of the counter-jet and the consequent creation of the toroidal bubble. Indeed, in LIFT, the counter-jet would hit the bottom of the donor substrate and dissipate its energy during impact, truncating the further development of the corona. For liquid films with a thickness above 50 µm, though, the formation of the corona can be clearly distinguished [68–71]. Thus, in agreement with the aforementioned interpretation, the film thickness plays a crucial role in corona formation. Regarding the typical absence of corona in LIFT, this technique would seem preferable over FF-LIFT for printing small volumes. Nevertheless, the use of low laser fluences in FF-LIFT, probably under conditions in which the corona did not reach the substrate, resulted in the smallest liquid droplets printed to date [39].

A second important difference between FF-LIFT and LIFT is the propagation and dissipation of capillary waves (Section 6.2). Due to the large reservoir used in FF-LIFT, capillary waves can be quickly dissipated, and thus, a second laser pulse fired in close vicinity to the first one could lead to suitable printing. This contrasts with LIFT, in which the presence of capillary waves impedes the proper development of closely spaced jets [47, 48]. In fact, in LIFT printing of highly viscous prepolymers, the minimum lateral distance between two laser shots that resulted in proper printing was about 50 µm [51]. In this case, the time required to completely dissipate the capillary waves was of the order of minutes. Therefore, FF-LIFT appears more suitable for high-throughput printing.

Finally, a more subtle difference between the two techniques concerns the level of understanding of the dynamics of each process. Notably, the formation of bubbles near a free surface is a phenomenon that has been widely studied in fluid mechanics, experimentally as well as computationally. Indeed, literature dating back to more than 30 years ago already studied this problem, given its direct link to cavitation erosion. Thus, the extensive knowledge of the mechanisms of FF-LIFT facilitates the search of optimization pathways for this technique. Instead, the configuration used in the LIFT of liquids is more recent, and parts of the mechanisms that lead to bubble collapse are more challenging to explore. In addition, the particular conditions of LIFT (in particular, the initial formation of the cavitation bubble) also make the computational study of this phenomenon difficult. On the other hand, this can be regarded as an opportunity for scientists in the field of fluid mechanics, for which LIFT represents a very interesting boundary value problem.

6.4.1 Laser Source

What is the suitable laser source for FF-LIFT? To answer this question, one must consider the intensity required to induce optical breakdown in a liquid. This value depends on the optical properties of the liquid as well as on laser pulse duration, wavelength, and focusing optics. For water at a wavelength of approximately 1 µm, the intensity value necessary to reach optical breakdown with nanosecond pulses (6 ns) is approximately 1 × 10¹¹ W/cm², whereas for picosecond (30 ps) and femtosecond pulses (300 fs), the intensity increases to approximately 5 × 10¹¹ W/cm² and 30 × 10¹¹ W/cm², respectively [54, 72]. Despite the increase in intensity as the pulse duration decreases, for a given wavelength and focusing optics, the energy per pulse required to induce optical breakdown with femtosecond pulses (10 s of nJ) is significantly lower than for nanosecond pulses (mJ). Moreover, multiphoton absorption induced by femtosecond pulses facilitates the generation of seed electrons. These electrons are necessary for avalanche ionization and plasma formation (optical breakdown). In the case of nanosecond pulses, the seed electrons are generated by defects in the material, which are stochastically distributed. In contrast, femtosecond pulses enable the deterministic formation of optical breakdown in a liquid and hence a much higher reproducibility of the printing process than in the case of nanosecond pulses. Combined with the low energy per pulse needed with femtosecond pulses, one can generally state that ultrafast lasers are the most appropriate laser source to perform FF-LIFT.

6.4.2 Forward (Inverted) versus Backward (Upright) Systems

FF-LIFT can be implemented in two different layouts, as shown in Figure 6.5. In the first configuration, known as forward [38, 61, 73], the liquid is ejected in the direction of beam propagation. Thus, the laser beam enters the bottom part of the reservoir and passes through the liquid until being focused just beneath the liquid-free surface. Because of the long distance the beam has to travel through the liquid, the liquid optical properties play a crucial role. Indeed, not only liquid absorption has to be very low, but also the chromatic dispersion of the medium has to be small in order to preserve the temporal duration of the laser pulse and the consequent high intensity in the focal volume required to induce optical breakdown. In addition, the reservoir has to be transparent to the laser beam, and ideally, it should have the same refractive index as the liquid in order to prevent unwanted reflections.

In the second configuration of FF-LIFT, known as backward, the liquid is ejected in the direction contrary to laser propagation. In particular, the laser beam first passes through the receiver substrate, it crosses the air gap between the receiver and the liquid film, and then it is focused beneath the liquid-free surface. Because the laser beam only travels through the liquid over a short distance, chromatic dispersion effects are significantly reduced compared to the forward implementation. In addition, the reservoir can be optically opaque without affecting the performance of the technique. In contrast, the receiver substrate must be transparent and present low chromatic dispersion. This strongly limits the potential use of FF-LIFT with many technologically significant substrates, such as silicon or metals, which are opaque but widely used in applications as relevant as lab-on-a-chip [2] or surface patterning [45]. Furthermore, the inevitable mismatch between the refractive index of air and that of the liquid will result in significant loss of intensity within the laser focal volume.

Considering the previous differences between the forward and backward configurations, what would be the suitable system to implement? This strongly depends on the particular demands of a given application. There is, though, an additional aspect that has not been discussed yet and that will probably determine the final system of choice for most users. In both forward and backward FF-LIFT, the reservoir has to be placed upside up. Consequently, the forward implementation requires a beam coming from below the reservoir, that is, an inverted optical system. This is more commonly found in life science laboratories. In contrast, the backward implementation needs an upright optical system, which is the standard system used for laser materials processing.

6.4.3 Spherical Aberration and Chromatic Dispersion

The focal point of a focusing lens (microscope objective) is usually designed to be an aplanatic point. In other words, the focus of most optical systems is designed to present minimal coma and spherical aberrations. However, if a planar dielectric with a different refractive index is placed within the optical path of the focusing length (between the lens and the focus), the aplanatic conditions are no longer fulfilled, and one can observe spherical aberration. The immediate consequence is the axial elongation of the focal volume. In FF-LIFT, the liquid plays the role of the dielectric, and hence, spherical aberration will have to be considered. Indeed, the elongation of the focal volume decreases the overall laser intensity and degrades the control in the positioning of cavitation bubble formation.

By using geometrical optics and Snell's law, it is possible to derive an expression for the lateral spherical aberration (A_s) induced by a planar dielectric. Considering A_s to be defined by the separation between the marginal and paraxial focus (the effective spreading of the focal volume) yields [74]:

6.5

where is the ratio of the refractive index of the medium (typically air) relative to the liquid, is the physical distance travelled by the laser beam within the liquid (in the case of forward configuration, it should also include the thickness of the reservoir, considering it has the same refractive index as the liquid), and , with being the beam waist at the entrance of the focusing lens with focal length . Note that is proportional to the numerical aperture (NA) of the focusing optics. From this equation, it can be observed that A_s scales with the NA of the system and with the distance the beam travels within the liquid (Table 6.2). Thus, the forward configuration will in general produce a more significant elongation of the focal volume than the backward layout. In addition, low NA optics will usually lead to better focus confinement, as experimentally demonstrated [61]. There exists, though, the possibility to greatly minimize spherical aberration in the forward configuration. This can be achieved by using a dipping objective, for example, a water-dipping objective. In this case, if both liquid and reservoir match the operational refractive index of the objective, for example, water, no spherical aberration would be induced.

d (µm)	NA	(µm)
100	0.3	1.6
400	0.3	6.3
1000	0.3	16
100	0.5	4.8
400	0.5	19
1000	0.5	48

Another important optical property of the medium that can affect the printing process is the refractive index dependency with wavelength. Indeed, the relative large wavelength bandwidth of femtosecond pulses makes them prone to chromatic dispersion, and consequent pulse stretching, when passing through different media. In addition, as a consequence of chromatic aberration, the effective intensity in the laser focus may be reduced. The chromatic dispersion of a medium is typically characterized with the Abbe number or constringence (symbol V), which specifies the variations in the refractive index of a medium for a specific spectral range. In particular, V = (n_D−1)/(n_F−n_C), where n_D, n_F, and n_C are the refractive indices of the material at the wavelengths 589.3, 486.1, and 656.3 nm, respectively. The Abbe number can be found in tables. Typically, materials with V < 50 are considered dispersive (V = 34 for polycarbonate plastics), whereas low-dispersive materials can reach V > 70, such as phosphate crown glasses. For water, V = 50, and hence, the amount of chromatic dispersion in FF-LIFT, particularly in the forward configuration, can be significant.

6.5.1 Film-Free LIFT for Printing Biomaterials

A potential niche market for LIFT is the controlled deposition of biological materials such as DNA, [27, 75] proteins, [76–78], and even living cells [26, 28, 30, 37]. Compared to alternative strategies such as inkjet printing or pin spotting, LIFT offers unique features that render this technique particularly suitable for the growing fields of biomolecule microarrays or biosensors. Indeed, LIFT enables the printing of volumes as small as femtoliters while preserving a high throughput. In addition, it is possible to visually inspect the donor film during the printing process. This makes possible to irradiate only a targeted area of the donor film. For instance, if the donor film contains cells, one can select the cells to be ejected toward the receiver substrate. However, the problems related to the preparation of the liquid film are particularly excruciating in the case of biological materials. Notably, biological solutions are often transparent to the common laser radiations – the use of UV lasers is not recommended since it could alter the biological materials (e.g., denaturation of DNA) – and thus, an intermediate absorbing layer or DRL, usually a metal, is utilized to enhance the light absorption and induce liquid ejection. The use of an absorbing layer, though, can lead to sample contamination by micro/nanometallic particles. For instance, DNA damage and cell toxicity could be induced by metallic contaminants [79, 80]. Although the use of a polyimide intermediate layer, as in the case of blister-actuated LIFT, solves this problem [37] (Chapter 5), the risk of liquid dewetting for thin films prepared on polymers (Section 6.2.1.1) is very high, and hence, only specific materials with very low surface tension can be used in practice. Note that it is also possible to use ultrashort laser pulses in LIFT to induce material ejection from weakly absorbing materials without DRL. However, this option is difficult to implement in practice due to the small thickness of the liquid film and the refractive index mismatch between the ink and the carrier substrate. Indeed, in this case, the control of the laser focus beneath the free surface of the liquid film is utterly challenging not only due to spherical aberration and the intrinsic depth of field of the focusing optics (which also exist in FF-LIFT), but especially for the change (jump) in the focus position when traversing two materials each with a different refractive index.

Within this framework, the possibility offered by FF-LIFT to directly deposit the biological material from a reservoir is of clear advantage. Because no absorbing layer is needed, contamination can be greatly reduced. Moreover, by obviating the liquid film, one can save time, prevent potential contamination during the preparation of the liquid film, and more importantly, skip the tight requirements on the rheology of the ink necessary to obtain a spatiotemporal stable thin film. In fact, FF-LIFT could be used with liquids that present a relatively high vapor pressure, provided the laser focus beneath the free surface is adjusted to compensate for liquid evaporation. As stated in Section 6.2.1.1, such inks are incompatible with LIFT, since liquid evaporation can lead to dewetting issues.

The feasibility of FF-LIFT for the fabrication of miniaturized optical biosensors has been recently demonstrated [39]. In this case, the biosensor consisted of a microarray printed on a poly-l-lysine-coated microscope slide with FF-LIFT and containing mouse and rabbit immunoglobulin G (IgG) solutions (Figure 6.6a). Notably, both solutions were directly printed from a 100 µl cylindrical plastic container. After an immunoassay with tagged antibodies, fluorescence microscopy revealed high specific binding for each printed solution. This result demonstrates the ability of accurately placing undamaged proteins onto solid substrates using FF-LIFT. Furthermore, FF-LIFT has been successfully exploited for printing DNA molecules for a hybridization assay [39]. Two DNA solutions were printed by FF-LIFT, directly form a reservoir, each one containing a single human cDNA clone insert: the mitogen-activated protein kinase 3 gene (MAPK3) and the v-ets avian erythroblastosis virus E26 oncogenehomolog 2 (ETS2), respectively. After a standard hybridization protocol, with the complementary strands of the cDNAs tagged with a different fluorochrome (Cy5 for MAPK3 and Cy3 for ETS2), fluorescence images demonstrated the specific binding of each cDNA strand to its perfect complementary. This further confirms the feasibility of FF-LIFT for printing biomolecules.

6.5.2 Film-Free LIFT for Micro-Optical Element Fabrication

The fabrication of micro-optical components plays a key role in the development of functional devices such as detectors, energy-harvesting systems, or CMOS cameras. With the advance of additive manufacturing processes and the trend toward customized production units, methods capable of directly fabricating user-defined microlenses or microlens arrays without masks or molds appear particularly suitable. Given the intrinsic compatibility of LIFT with highly viscous inks such as those traditionally used for microlens fabrication, this technique seems to be an ideal tool of choice for this purpose. In addition, the use of liquid printing techniques has a natural advantage over the methods that operate with solid materials, because surface tension provides liquid droplets with a perfectly spherical shape. This has prompted the use of LIFT for the successful printing of liquid microlenses that provide an enhancement of the focusing capabilities of a laser system and enable deep subwavelength laser nanopatterning [45]. Moreover, LIFT has recently been used for the fabrication of polymeric lenses with tailored geometry and size (Figure 6.6b) [51]. In both cases, the potential advantages of LIFT can be hampered by two serious constraints of this technique. First, the requirement of an absorbing layer can result in debris (micro/nanoparticles) inside the fabricated lens. As a result, the optical quality of the microlenses can be severely deteriorated. Second, the wettability of the microlenses with the receiver substrate dictates the optical properties of the lenses, namely the radius of curvature. Indeed, to obtain lenses with high curvature, one requires inks that present a poor wettability with the receiver substrate. Given the restricted number of available inks with desired optical qualities, an ink with poor wettability on the receiver will typically have poor wettability in most materials, and hence, the preparation of such an ink in film form will be challenging. In contrast, FF-LIFT has the potential to solve all these problems. Indeed, the possibility to generate micro-optical components without an absorbing layer can prevent contamination issues. Furthermore, printing inks with poor wettability is no longer a problem, because the ink can be directly handled from a reservoir or container.