The interaction of ultrashort laser pulses above the ablation threshold of thin-film indium tin oxide (ITO) is examined with pump-probe microscopy. We are able to observe photomechanical spallation at delay times of hundreds of picoseconds, which plays a stronger role near the ablation threshold of 0.17 J/cm2. A phase explosion may also be observed at tens of picoseconds, playing a stronger role for increasing peak fluences. As one exceeds the material removal efficiency maximum near 0.6 J/cm2, a second spallation is observable in the center of the irradiated spot at a delay time of one nanosecond and corresponds to a crater depth of 50 nanometers. No discernable ridge formation has been observed. We recommend an industrial processing window of at least two pulses per position with a peak fluence between 0.6–1.0 J/cm2.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Indium tin oxide (ITO) belongs to a family of materials known as transparent conducting oxides (TCOs) which offer minimal absorption in the visible spectrum in addition to their ability to act as a conducting electrode for various applications. Of the transparent conducting materials present, ITO has seen great commercial success and finds use in solar cells, LEDs, and display technology [1,2]. A growing interest in ITO has also been seen for photonics applications [3–5]. Whereas photolithography and wet chemical etching still exists as a processing method for ITO for device functionalization, this has largely been replaced by laser structuring [1,6]. Laser structuring of ITO thin films enables selective etch process without the need of etchant chemicals or photolithography for the production of functional layer stacks. Furthermore, as this is a maskless process, the user may freely select the etch pattern before the etch process.
Ultrashort laser pulses may be used to perform localized material removal with minimal heat transfer to neighboring material. Laser structuring typically is performed by scribing electrically isolating lines using lasers with a high repetition rate, scanning across the ITO film. The entire film may be reliably removed by overlapping pulses when performing line scribes [2,7–14]. The scanning speed can be optimized with the goal of enhancing the overall throughput in industrial processes. Known problems exist for laser ablation of ITO, for example ridge or crack formation, or the contamination of neighboring material, all of which may be traced to thermal effects that are more dominant for pulse durations in the nanosecond regime . Particularly problematic is the formation of vertical ridges on the edge of an ablated crater which may interfere with any subsequently deposited layers on top of the ITO. Ridge formation has been hypothesized as a result of a surface tension gradient existing near the outer edge of the irradiated spot due to thermal effects .
The bulk of public research regarding the selective structuring of ITO with pulsed lasers began in the 1990s with nanosecond pulse durations, overlaps ranging from about 20-50%, and repetition rates of tens of kHz [1,16–18]. UV, VIS, and NIR wavelengths were examined with irradiated peak fluences typically ranging from about hundreds of mJ/cm2 to tens of J/cm2. Here, it was observed that shoulders with heights of tens of nanometers, and in some case cracking, form as a result of the laser processing. A degree of ITO residue was also observed in some studies, which may form at the bottom of the crater in neighboring ablated spots or in the ablated crater itself. In some cases, this residue was not seen to impact the electrical isolation and was therefore tolerated. The use of UV wavelengths was seen to give an improvement in the processing quality of the ITO but an increase in the amount of glass that was etched or damaged was also observed. Following the turn of the century, investigations into ITO ablation using femtosecond and picosecond pulsed lasers were performed. Harrison et al. performs a short literature review of existing work up until 2010 . A dramatic reduction in the shoulder height was observed for femtosecond and picosecond pulse durations, as well as less contamination and cracking effects [2,7,9–14,19,20]. This reduction may be attributed to the material heating which takes place on a time-scale that is smaller than the time required for heat conduction, which allows thermomechanical effects to play larger role in material removal . The ablation threshold for film-side irradiation of ITO for sub-picosecond and picosecond pulse durations was reported in the mentioned literature to lie between about 0.1 and 1 J/cm2. These studies suggest that optimum line-scribing parameters have an individual irradiated peak fluence of about 1 J/cm2, an overlap of about 50-90%, repetition rates typically ranging from kHz to hundreds of kHz, and result in about 2-10 J/cm2 of an accumulated irradiated fluence. There are mixed reports as to whether or not ITO films, typically in the order of 100 nm thick, could be removed with a single pulse within the fluence range of the experiments. Glass substrates tend to have a much higher ablation threshold, typically 3 to 10 J/cm2 for pulse durations ranging from 500 femtoseconds to 10 picoseconds in the NIR . For ITO on glass substrates, it is seen that ITO has a stronger absorption coefficient for visible and near-infrared wavelengths compared to that of glass [22–24]. This is namely due to the higher band gap of glass as well as the existence of free carriers in ITO at room temperature. It is generally advisable to work above the ITO ablation threshold, yet below the glass damage threshold in order to prevent unwanted damage to the glass substrate. Studies also exist for substrate-side irradiation of the films, which is seen to be more efficient in terms of material removal per irradiated accumulated fluence [13,25]. In this work, however, we focus on the film-side irradiation, as this is most commonplace.
Ultrafast changes to material and subsequent material heating tend to lead to spallation when the pulse duration and the electron-phonon relaxation time are sufficiently below the mechanical relaxation time due to stress confinement [26,27]. When the heating time is longer than the mechanical expansion time, predominantly a phase explosion with a larger contribution to vapor generation may be observed. Spallation, a photomechanical process, preferentially takes place at fluences just above the ablation threshold, whereas a phase explosion, an evaporative process, takes place at fluences considerably above this threshold [14,28]. Spallation may be observed in PPM as film bulging, which typically results in Newton’s rings, whereas the liquid-gas phase explosion tends to scatter and highly absorb light as a result of following multiple reflections on the droplets [28–35].
A purely photothermal process in which material is removed via vaporization may be termed as a direct ablative process. A photomechanical process, where thermodynamic interactions result in mechanical force responsible for the removal of material, may be much more efficient if the energy imparted into the material is only sufficient enough to break the atomic bonds between the outer surface of the removed layer and the surrounding material. This does not need to overcome the latent heat in order to vaporize the entire material. Furthermore, directly ablative processes of ITO tend to cause contamination issues for most applications, which is especially prominent for nanosecond pulse laser ablation of ITO .
Because a confined energy situation leading to a photomechanical ablation process tends to be much more efficient, threshold fluences are occurring a factor of ten lower than the threshold fluence required direct ablation for film thicknesses tens to hundreds of nanometers thick [37,38]. For a photomechanical process, the absorbed energy per volume ablated lies in the order of the melt enthalpy of a few J/mm3 to tens of J/mm3 whereas for a direct process, the energy required to evaporate the same volume ranges from tens to hundreds of J/mm3 .
Pump-probe microscopy (PPM) may be used to visualize ultrafast ablation by utilizing a delayed illumination probe signal with respect to the pump pulse, which imparts the energy into the material required for ablation. This allows for dynamic characterization of the optical properties of materials. The localized material response may also be spatially resolved by employing an imaging system using a camera and a microscope objective. Furthermore, in this time-resolved microscopic setup, one may also observe in situ material removal from the ablation process as well as shock wave propagation .
Time-resolved pump-probe measurements for the irradiation of ITO exists in literature below its ablation threshold, particularly to develop photonic applications due to ITO’s strong nonlinear properties [5,40]. Although ITO ablation processing parameters and their respective results exist in literature, time-resolved measurements of ITO above the ablation threshold are missing to date.
In this study, we seek to understand what physical processes occur during ultrafast laser ablation of ITO. Can a confined energy situation be observed for fluences near the ablation threshold, giving way to an evaporative thermal process at higher fluences, as is seen for metals and semiconductors? A better understanding of the process parameter criteria to enable precise selective structuring which yield isolating process lines that minimizes or eliminates film cracking or peeling, substrate damage, as well as ridge or spike formation is needed. Here, single pulse experiments are performed to understand the basic physical processes that occur for scribing lines, debarring incubation and heat accumulation effects that may be present for line processing by spatially overlapping pulses with high repetition rates.
2. Materials and methods
We used a commercially prepared ITO sample with an ITO thickness of about 100 nm sputtered onto a SiO2 passivation layer of about 25 nm on a 1.1 mm float glass substrate (CEC020S, Präzisions Glas & Optik GmbH, Iserlohn, Germany). The sheet resistance is specified to be below 20 Ω/□ and we have selected a float glass substrate in order to maintain industrial relevance . The film thickness was verified by means of spectrophotometer measurements and an optical model fit, which gave a ITO film thickness of 105 ± 4 nm. Any material composition differences in the ITO film as a function of the film depth are not taken into account in this study. Further information about the film thickness measurement may be found in the Supplement 1 section S2.
A Nd:glass femtosecond laser (femtoREGEN, Spectra-Physics, Inc.) with a central wavelength of (1056 ± 0.5) nm with a spectral FWHW of (5 ± 0.5) nm (BLUE-Wave, StellarNet Inc.) and a pulse duration of (0.70 ± 0.1) ps (pulseCheck, APE GmbH) was used for PPM. Detailed information about the optical setup of the PPM may be found in the Supplement 1 section S1 and in . The pulse-to-pulse energy stability of the laser source was measured to be below 1% standard deviation by means of a photodiode (DET100A/M, Thorlabs, Inc.) and an oscilloscope. A plano-convex lens with a focal length of 100 mm was used to focus the pump beam with p-polarization, where a beam waist of (15 ± 1) µm corresponding to an intensity drop of 1/e2, an M2 of 1.4, and a Rayleigh length of (0.5 ± 0.05) mm was measured at normal incidence (MicroSpotMonitor, PRIMES GmbH). For pump-probe experiments on ITO, the pump beam is focused with an angle of incidence of 38.7°. The change in the peak fluence due to the non-normal angle of incidence was corrected by multiplying the calculated peak fluences at normal incidence by a factor of cos(38.7°) ≈ 0.78 . An optical delay line was used to provide delay times of up to 3500 ps, where the probe beam which was diverted through a second harmonic generator (SHG) to yield a wavelength of 528 nm. Here, the Kerr effect observed in ITO was used to calibrate the delay by defining a delay time zero as the ΔR/R maximum value observed for a pump irradiation intensity of 50 GW/cm2. For imaging of the sample surface, a Mitutoyo 50x M Plan Apo SL objective (NA = 0.42) was used in combination with a pco.pixelfly usb CCD camera which gives an optical resolution of about 630 nm as per the Abbe diffraction limit. We applied a matte adhesive tape to the backside of the ITO samples in order to suppress reflections of the probe from the backside of glass substrate.
A separate Innolas Picolo AOT laser with a central wavelength of 532 nm and a pulse duration of 600 ps was used for measurements with an electronic delay to probe delay times greater than 3500 ps with an external jitter of less than 1 ns. Here the pump parameters were the same as above, with the exception of a pump pulse duration of (0.5 ± 0.1) ps, a slightly larger beam waist radius at 1/e2 of (15.5 ± 1) µm, and an angle of incidence of 35.5° as these experiments were performed on a different optical setup of the same design. Considering these factors, one may expect a slightly larger and rounder spot for the electronically triggered measurements.
Two images were captured with the PPM camera for a given delay time: one image microseconds before the pump irradiates the sample surface, Rbef, and one image at the specified delay time with regards to the pump pulse, Rdur. The Rdur image is then normalized by multiplying Rdur by a normalization factor determined by comparing the average pixel value at for pixels at least 10 µm away from the edge of the irradiated spot to compensate for probe-pulse fluctuations. ΔR/R was then calculated by the relation (Rdur−Rbef)/Rbef. In order to suppress noise, an adaptive 2-D Wiener filter (MATLAB, The MathWorks, Inc.) was then applied to the resulting ΔR/R image with a neighborhood pixel size of 3 × 3, where one pixel corresponds to an area of 129 nm by 129 nm.
In addition to creating the ΔR/R images, additional processing was performed in order average the ΔR/R response for all pixels on the same given local fluence of the pump pulse. Here we assume a Gaussian intensity distribution projected to an ellipse given by the angle of incidence. This was done by averaging over pixels that correspond to points on the sample exposed to the same fluence or fluence isophote. One point on the minor axis length of an ellipse is defined as a reference for the local fluence. For a given eccentricity of the irradiated spot, the pixels along the perimeter of the corresponding ellipse were averaged in order to determine the localized ΔR/R response for a given fluence isophote. The resolution of the measured ΔR/R values were determined by calculating the standard deviation between an averaged group of 700 pixels in the center of the camera for separate images, absent of pump irradiation. This results in a standard deviation of 0.005 for the relative reflectivity. Additional details regarding the PPM setup and data processing may be found in [28,42].
In order to determine the ablation threshold fluence, crater profiles, depths, and removed volumes, measurements of the craters’ morphology were performed using a Sensofar Plµ2300 with an interferometric 50x objective (NA = 0.55), which gives sub-nm vertical resolution. These measurements were then subsequently evaluated with the program Gwyddion (Ver. 2.58, released February 9th, 2021) . Crater profiles were evaluated by averaging the depths of five craters per peak fluence along the minor axes of the elliptical spots. Fluence isophotes were not examined in this case. The shape of crater profiles may give an initial indication of the characteristics of the ablation process that took place . Ten craters per peak fluence were examined for the determination of D2eff, the crater depths, and the volumes.
Similar to the D2 method, the Beer-Lambert law may be applied to describe the crater depth with threshold-like ablation behavior. A short-coming of this approach is that spallation typically gives a minimum crater depth near the ablation threshold, which is not accurately described by exponential behavior . We therefore extend the Beer-Lambert approach with an additional spallation depth term dspl in Eq. (2) to give:
Once the effective penetration depth has been established, one can apply stress confinement criteria given by
As a result of stress confinement, a shockwave would propagate away from the irradiated spot following rapid material expansion. This shockwave dissipates energy and develops into a pressure wave that propagates at the speed of sound in the material. With PPM, one can calculate the velocity of a pressure wave propagating through a material. This is calculated by oscillations in the ΔR/R signal for given delay times as a result of optical interference between the wave front and reflections from the media interfaces. This is defined as49,50]. With this information, one is able to confirm the presence of a pressure wave propagating in the material.
3.1 Crater morphology
The crater morphology was examined in order to give indications of the ablation processes at hand visible from the final state. Figure 1 shows the crater profiles for selected peak fluences, D2eff and its respective fit, crater depths and the respective Beer-Lambert fit, as well as the removed volume per pulse energy. The fit results are given in Table 1. In Fig. 1(a), a flat crater profile with a depth of about 25 nm for a peak fluence of 0.23 J/cm2 may be observed, which gives an indication of a spallation . Around 25 nm for peak fluences above 0.23 J/cm2, a notable bend in the crater profile may be observed where there is a transition from the steep crater profile near the edge. Between depths of 25 and 50 nm, we observed that the crater profile takes on a relatively parabolic shape, which may be indicative of a direct ablation process that reflects the Gaussian laser intensity profile . For a peak fluence of 1.0 J/cm2, we observed that the crater depth once again saturates and becomes flat in the center of the crater at a depth of about 50 nm. This may be indicative of a second spallation. This once again gives way to a parabolic crater shape below 50 nm for centermost portion of the crater irradiated with 1.2 J/cm2. Minimal ridge formation of a few nanometers was observed outside of the crater for the higher peak fluences shown in Fig. 1(a).
The results of the D2eff fit show that the ablation threshold value was determined to be (0.17 ± 0.01) J/cm2, in agreement with literature values from section 1. The effective beam waist radius weff was fit to a value of (14.9 ± 0.1) µm at an angle of incidence of 38.7° whereas the measured beam waist radius at normal incidence was (15 ± 1) µm. Here, we see a small discrepancy as we would expect a slightly larger weff due to the spot elongation given by the angle of incidence. The results of the Beer-Lambert fit may be seen in Tab. 1 and Fig. 1(c); it is seen that the effective penetration depth was about as large as dspl ≈ 20 nm at the threshold fluence of 0.17 J/cm2. The oscillatory pattern at about 0.3 and 1 J/cm2 observed in Fig. 1(c) may reflect the flattening of the crater profiles as observed in Fig. 1(b) speculated from the first and second spallation regime. Furthermore, it may be noted that the bottom of the 105 nm thick ITO layer was not reached even for the highest fluence given by the laser source at 8.4 J/cm2.
The energy specific ablation volume (ESAV) may be used to characterize the efficiency of the ablation process defined as the removed volume Vabl per irradiated pulse energy Ep [12,51]. The ESAV maximum found in Fig. 1(d) at (6.1 ± 0.4) µm3/µJ for a peak fluence of 0.6 J/cm2 indicates the most efficient processing peak fluence in terms of material removal as seen in Fig. 1(d). This efficiency peak is preceded by a sharp increase in removed material per pulse energy for lower peak fluences and is followed by a steady drop for further increases in the peak fluence. The ESAV maximum is located at irradiated fluences where we found a transitory stage between the first and second spallation regime, which will be highlighted in section 3.2. Here we study the dynamics of the ablation process with PPM to examine the mechanisms of the material removal.
3.2 Pump-probe microscopy
To study the mechanisms of material removal in detail, a time-resolved study of the relative reflectivity with pump-probe microscopy was performed. Two peak fluences were chosen as they clearly demonstrated the effects for the two spallative regimes as seen in section 3.1. Seven distinct numerated phenomena for the ΔR/R images may be observed in Fig. 2. Here it is seen that warm colors represent an increase in reflectivity whereas cold colors represent a decrease in the reflectivity.
Near the zero time-delay, a sharp increase in relative reflectivity of the irradiated spot to a maximum of 1.58 at a delay time of 500 fs for phenomenon 1a and a maximum of 2.63 at 400 fs for phenomenon 1b was observed. This is a result of free electron generation in the conduction band, which results in ITO taking on metallic-like behavior at the probe wavelength.
One was also able to directly observe the Kerr effect marked by phenomena 2a and 2b for low spatial fluences near delay time zero at the outer edge of the irradiated spot. Local ΔR/R maxima were found at 2a with a value of 0.06 ± 0.01 for a peak fluence 0.39 J/cm2 at a local intensity of 50 GW/cm2 and at 2b with a value of 0.06 ± 0.01 for a peak fluence of 1.0 J/cm2 at a local intensity of 90 GW/cm2. This positive change of ΔR/R was observed for lithium niobate in a study by Garcia-Lechuga et al. . The ΔR/R maximum as a result of the Kerr effect is somewhat negligible in comparison to the ΔR/R maxima observed at the zero delay time for the center of the spots observed in Fig. 2. Therefore, due to the lack of contrast, the change in ΔR/R as a result of the Kerr effect is somewhat difficult to observe in Fig. 2, effects 2a and 2b. Note that the non-linearity saturates at a ΔR/R of 6% and directly competes with free electron generation for higher local fluences, as was observed in  in a similar pump-probe setup. Saturation of the Kerr effect in ITO was additionally observed for z-scan pump-probe experiments in .
Following the sharp rise in relative reflectivity at a delay time of zero, a drop to negative relative reflectivity values occurred for delay times of tens of ps as indicated by effects 3a and 3b. The relative reflectivity of the irradiated spot dropped to 0 after a 19 ps delay for phenomenon 3a and after 10 ps for phenomenon 3b. A minimum value relative reflectivity was observed at −0.62 for a delay time of 175 ps for phenomenon 3a and at −0.89 for a delay time of 60 ps for phenomenon 3b. Here it was observed for 3b that the onset of negative relative reflectivity occurs much sooner than for 3a and that the center of the irradiated spot remained at negative relative reflectivity values for hundreds of ps for a peak fluence of 1.0 J/cm2.
For phenomena 4a and 4b, Newton’s rings may be observed at the irradiated spot from about 100 ps until about 1 ns. For phenomenon 4b, these Newton’s rings may only be observed on the outer edge of the irradiated spot whereas the center of the spot remained at negative ΔR/R values, −0.83 for the shown delay time of 600 ps. Here it is notable that the Newton’s rings maximum ΔR/R values were higher for the outermost portion of the irradiated spot, most remarkably for 4a.
For the same delay times of about 100−1000 ps, phenomena 5a and 5b consist of diffraction rings that may be observed outside of the irradiated spot where the Newton’s rings are present resulting from the diffraction-limited system characterized by the numerical aperture of the objective. Rings seen outside of the irradiated spot for these hundreds of picoseconds may be understood as diffraction rings as a result of a steep surface at the edge of the irradiated spot during film bulging and the diffraction limited optical system in Fig. 2 and 3 for phenomena 5a and 5b. Here, the optical system was unable to resolve the surface discontinuity and its corresponding reflected intensity that would have otherwise been encompassed by diffracted light at angles higher than the objective was able to accept. Further analysis of the film bulging speed and the diffraction ring spacing is given in the Supplement 1 sections S3 and S4.
For effects 6a and 6b, the Newton’s rings pattern became distorted and one may observe brightness fluctuations near the edge of the irradiated spot after about 1 ns. This is a result of the spallation layer disintegration and particle generation where the bright spots observed have a diameter of about 5 pixels, which corresponds to 645 nm. This is comparable to the Abbe diffraction limit of the system at about 630 nm. The particles observed are therefore estimated to have a grain size in this order of magnitude, although these spots may consist of multiple particles that were unable to be optically resolved.
A second bulging may be observed as Newton’s rings in the center of the crater for peak fluences greater than 0.6 J/cm2 become observable for delay times greater than 1 ns in Fig. 2 and 3, phenomenon 7. Videos found in Visualization 1 and Visualization 2 clearly demonstrate the mentioned observations. Here, one is also able to observe the propagation of a shockwave in air seen for tens of nanoseconds followed by scattering of the probe pulse as the disintegrated film expands outwards. The observable ablation process ends around 500 ns and no more changes to the sample surface were detected after that delay time.
Figure 3 demonstrates the ΔR/R response for fluence isophotes denoted by the distance from the center of the spot vs. delay time, as defined in section 2. Figure 3 uses the same ΔR/R color scheme and effect numeration as Fig. 2 with the addition of phenomena 8a and 8b. Outside of the irradiated spot, temporal oscillations for ΔR/R values on the order of magnitude of 0.1 may be observed just before the onset of the Newton’s rings. These are observable in Fig. 3 as horizontal lines, most easily observed outside of the center of the irradiated spots. Similar to the case of the Kerr effect, this effect is somewhat difficult to observe in Fig. 2 and 3 due to the lack of contrast. For phenomena 8a and 8b in Fig. 3, it is seen that the oscillations outside of the irradiated spot had a frequency of about 33.4 GHz, as determined by Fourier analysis. Note that this phenomenon does not use chronological number ordering, as it is only visible in Fig. 3 after the other phenomena have been introduced in Fig. 2. We use Eq. (4) where the refractive index of glass at the probe wavelength was determined to be 1.52 from the spectrophotometer measurements. The resulting shockwave velocity at 5800 m/s is comparable to measured literature values for float glass .
With Fig. 3, we were able to observe a different perspective of the ablation dynamics. Here it is easy to see that the maximum ΔR/R value was obtained sooner for a higher peak fluence when one compares effect 1b to effect 1a. The Kerr effect was also faintly visible here, displayed as effects 2a and 2b. One can also see the sooner onset and longer duration of negative ΔR/R values for effect 3b in comparison to 3a. We were also able to observe how effects 4a, 4b and 7, the Newton’s rings, evolved throughout time and how in the center of the spot that this signal was suppressed. We observed that the first minimum for 4a was observable at about 50 ps in the center of the spot. The curvature of the Newton’s rings may be understood in that as time progresses, the film continues to bulge, which effectively moves the rings towards the outer edge of the spot as its height increases. Effectively, these act as contour lines. Here we, once again, observe that the ΔR/R values are greater for the Newton’s rings maxima near the outer edge of the irradiated spot. Vertical lines outside of irradiated spot for effects 5a and 5b represent the diffraction rings outside of the spot, which remained for the most part constant in time for the duration of their existence. For effects 6a and 6b, we see that the Newton’s rings became scattered and that the diffraction rings faded; in Fig. 3, we were not able to see the individual grains.
From these results, we were clearly able to see evidence of both film bulging as well as the absorption and scattering of our probe pulse. For a peak fluence of 1.0 J/cm2, we were able to observe a second film bulging, described as phenomenon 7, observable from about 1 ns onwards. We now seek to interpret these results and draw correlations between the PPM data and the final crater state.
Time-resolved relative reflectivity measurements gives us great insight into the physical processes that occur during ablation. Here, we summarize the observed effects and their origins in chronological order, beginning with free electron generation. As the laser pulse impacts the ITO film, a sharp rise in the relative reflectivity is observed as a result of free electron generation. The increase in relative reflectivity persists for about 10 ps, as demonstrated by effect 1 in Fig. 3 and 4. During this time, thermalized electrons impart energy onto the ITO lattice structure, on a time-scale governed by the electron-phonon coupling time, which typically occurs on a time scale of approximately 16 picoseconds for similar laser intensities . A significant increase in the free electron density results in the increase of reflectivity for visible light, which in turn gives us metallic-like material properties. This is understood by means of the Drude model. Similarities may therefore be drawn between the ultrafast ablation of ITO and metals following infrared pulse irradiation.
After a sufficient degree of electron-phonon coupling has taken place, the irradiated film then begins to expand and any superheated material creates a scattering and absorbing gas-liquid mixture, a characteristic feature of a phase explosion. This effect is observable for tens of picoseconds and is demonstrated by a drop of the relative reflectivity due to multiple scattering and absorption processes as illustrated by effect 3 in Fig. 3 and 4. A phase explosion is only expected for fluences considerably above the ablation threshold, as discussed in section 1. Notably, from section 3.2, the rate at which the relative reflectivity decreases was observed to be greater for the higher peak fluence of 1.0 J/cm2 compared to that of the 0.39 J/cm2 irradiation. This may be understood as a result of the sooner onset of liquid-gas formation which scatters and absorbs the probe signal.
The mechanical relaxation time may be calculated with Eq. (3) using the effective penetration depth fit, and the longitudinal velocity of sound for ITO from , which gives us a mechanical relaxation time of approximately 3.3 ps. We observe that the material heating time is limited primarily by the electron-phonon coupling time as the pulse duration is more than an order of magnitude smaller. Although the electron-phonon time is approximately 5 times larger than the calculated mechanical relaxation time, stress confinement conditions are at least partially fulfilled .
For delay times of hundreds of picoseconds, evidence of a travelling shockwave in the media is observable just before film bulging may be observed in the form of Newton’s rings for phenomena 8a and 8b in Fig. 3. Here, it is understood that a rarefaction wave that trails a pressure wave in ITO may initiate void nucleation which results in film bulging, effect 4, and the eventual spallation of the film [26,27,29,55,56]. The second spallation, effect 7, that is observed is likely initiated by a second pass of a rarefaction wave after the shockwave has reflected off an interface, such as the boundary between ITO and the glass substrate. Due to the scattering present in the center of the spot, it is difficult to estimate when exactly the second bulging begins.
The cessation of the Newton’s rings marks the beginning of the film disintegration at phenomenon 6 in Fig. 4 for a delay time of approximately 1000 ps. The fluctuations in brightness seen near the edge of the irradiated spot for delay times over 1000 ps may be understood as a result of light being scattered as the edge of the spall layer begins to disintegrate into nanoparticles. The nanoparticles are seen to disperse away from the crater until about 500 ns, as shown in Fig. 2. Scattering of the subsequent pulses in multi-pulse processing by the nanoparticles would therefore be minimized if a repetition rate below 2 MHz were to be used.
The fact that the ΔR/R intensity of the Newton’s rings in the center of the spot is lower than that of the rings near the edge of the irradiated spot where the local fluence is closer to that of the ablation threshold has an interesting implication. At the edge, the absence of the scattering and absorbing gas-liquid mixture allows for a larger portion of the light reflected below the bulged film to interfere with the light reflected from the bulged film. This suggests that a purely spallative process takes place near the edge of the irradiated spot whereas in the case of 0.39 J/cm2 both a spallation and a phase explosion take place in the center of the spot. This is due to the fact that Newton’s rings could be observed throughout the entire spot, yet in the center of the spot a degree of scattering and absorption took place due to the presence of a liquid-gas mixture. Here it is, however, unclear whether the liquid-gas mixture is found above or below the film bulging, if not both.
By examining both PPM for ultrafast ITO ablation and its final crater state with an optical profiler, we are able to see similar trends regarding the ablation mechanisms involved. This serves to strengthen the interpretations of the effects observed in PPM. We observed the thickness given by the Beer-Lambert fit result of dspl in Tab. 1, with a value of (19.6 ± 1.9) nm is comparable to the depth of the crater profile seen in Fig. 1(a) for a peak fluence of 0.23 J/cm2 of about 25 nm. Here, a direct ablation process is responsible for the removal of the additional 5 nm of material, as observed in PPM. The additional increase in peak fluence is required to surpass the spallation threshold fluence in order to remove a larger surface area with a Gaussian beam. The parabolic shape below 25 nm and sharp crater edges we see in Fig. 1(a) for a peak fluence of 0.39 J/cm2 is understood as a direct result of the combination of a spallation over the entire spot and a phase explosion in the center.
We understand the lack of ridge or spike formation as a result of stress confinement. As mentioned in section 1, ridge formation has seen to be driven by liquid film motion driven by a surface tension gradient, taking place on a microsecond timescale for an evaporative removal process . In the case of a spallation, as is observed to be the sole removal mechanism at the outer edge of the irradiated spot, the majority of the liquid film is ejected from the surface long before liquid motion can take place and the heating of neighboring film material is minimized. This effectively removes the surface tension gradient and ridge formation does not take place. If we calculate the heat diffusion length using literature values of ITO heat diffusivity at room temperature before the spallation takes place around 1 ns, we obtain a length of only about 50-100 nm .
In Fig. 1(a), the distance across the minor axis that corresponds to the bend around 55 nm in depth for is about 10 µm for a peak fluence of 1.0 J/cm2. This is comparable to that of the length of the minor axis of the inner ring seen in the PPM images for 1.0 J/cm2 in Fig. 2 and 3, phenomenon 7, for delay times above 1000 ps. Here it is thought that a second spallation takes place in the center of the irradiated spot after liquid-gas mixture has dispersed, as illustrated in Fig. 4, effect 7. One can assume that for peak fluences between 0.2 J/cm2 and 0.6 J/cm2 that the ratio of evaporated to spalled materials grows with the applied peak fluence until the next spallation depth of approximately 50 nm is reached with a peak fluence of about 0.6 J/cm2. This peak fluence corresponds to the material removal efficiency peak, as given by the ESAV data in Fig. 1(d). For further increases in the fluence following this plateau at 50 nm in Fig. 1(a) and 1(c), one may expect an additional transition from the second spallation to a phase explosion. Overall, it is seen however that for peak fluences above 1 J/cm2 that a large drop in the efficiency of the material removal may be observed in Fig. 1(d). Here it may be assumed that most of the ITO removal is thermal in nature. An evaporative process is assumed to remove most material the final state of the craters are greater than 50 nm in depth.
In order to gain a better understanding of the physical processes that occur during ultrafast laser ablation of ITO, we utilize both crater profile analysis and time-resolved measurements of the dynamic changes in ITO reflectance. It is seen that the ITO takes on metallic-like behavior following infrared pump irradiation. We observe film bulging followed by a spallation process as the predominate ablation mechanism for fluences near and at the ablation threshold of 0.17 J/cm2, where a spall layer of 20 nm is removed. As one increases the applied peak fluence, a transition to a phase explosion is observed, where photothermal ablation takes place. At what depth the liquid-gas mixture is formed, remains to be examined.
Remarkably, we observed a second spallation, visible approximately 1 ns after the pulse impact for peak fluences above 0.6 J/cm2. This second spallation may be observed in the center of the irradiated spot in PPM, which corresponds to a final crater depth of about 50 nm. The exact origins of the second spallation remain unclear, but is purported to be initiated by a rarefaction wave reflected from the interface between the ITO and the glass substrate. We also observe a phase explosion dominant regime from PPM and a consequent drop in the material removal efficiency for further increases to the peak fluence. Beyond 0.6 J/cm2, however, a larger surface area of the second spall layer may be attained.
As a result of this study, we have understood the reasoning behind the laser parameter suggestions that exist for previous studies of the ultrafast laser processing of ITO. One can therefore agreeably suggest a processing window of about 0.5-1.0 J/cm2, where 0.6 J/cm2 is seen as the efficiency peak with little to no ridge formation. In order to remove the entire film at the irradiated spot in an industrial process, it is necessary to work with pulse overlap. A repetition rate below 2 MHz is suggested in order to avoid the dispersing disintegrated film nanoparticles from scattering subsequent pulses. As the single pulse ablation depth at maximum efficiency fluence is only about 50 nm, at least 2 pulses per position are required to remove a 100 nm film of ITO from float glass.
Deutsche Forschungsgemeinschaft (423531130).
Michael Kaiser is to be thanked for facilitating the optical profiler measurements. The interesting discussions with Maximillian Spellauge, Jan Winter, Ludwig Pongratz, and Sönke Vogel are also greatly appreciated.
The authors declare no conflicts of interest.
Due to its large size, the raw data underlying the results presented in this paper are not publicly available at this time. Specific datasets of interest may be obtained from the authors upon reasonable request.
See Supplement 1 for supporting content.
1. P. M. Harrison, N. Hay, and D. P. Hand, “A study of stitch line formation during high speed laser patterning of thin film indium tin oxide transparent electrodes,” Appl. Surf. Sci. 256(23), 7276–7284 (2010). [CrossRef]
2. M. Park, B. H. Chon, H. S. Kim, S. C. Jeoung, D. Kim, J. I. Lee, H. Y. Chu, and H. R. Kim, “Ultrafast laser ablation of indium tin oxide thin films for organic light-emitting diode application,” Opt. Lasers Eng. 44(2), 138–146 (2006). [CrossRef]
3. N. Kinsey, C. DeVault, A. Boltasseva, and V. M. Shalaev, “Near-zero-index materials for photonics,” Nat. Rev. Mater. 4(12), 742–760 (2019). [CrossRef]
4. G. V. Naik, V. M. Shalaev, and A. Boltasseva, “Alternative plasmonic materials: Beyond gold and silver,” Adv. Mater. 25(24), 3264–3294 (2013). [CrossRef]
5. O. Reshef, I. De Leon, M. Z. Alam, and R. W. Boyd, “Nonlinear optical effects in epsilon-near-zero media,” Nat. Rev. Mater. 4(8), 535–551 (2019). [CrossRef]
6. M. Aleksandrova, G. Kolev, I. Cholakova, G. Dobrikov, and G. Bodurov, “Photolithography versus lift off process for patterning of sputtered indium tin oxide for flexible displays,” Int. J. Thin Film. Sci. Technol. 2(2), 67–75 (2013). [CrossRef]
7. D. Ashkenasi, G. Müller, A. Rosenfeld, R. Stoian, I. V. Hertel, N. M. Bulgakova, and E. E. B. Campbell, “Fundamentals and advantages of ultrafast micro-structuring of transparent materials,” Appl. Phys. A: Mater. Sci. Process. 77(2), 223–228 (2003). [CrossRef]
8. Q. Bian, X. Yu, B. Zhao, Z. Chang, and S. Lei, “Femtosecond laser ablation of indium tin-oxide narrow grooves for thin film solar cells,” Opt. Laser Technol. 45(1), 395–401 (2013). [CrossRef]
9. H. W. Choi, D. F. Farson, J. Bovatsek, A. Arai, and D. Ashkenasi, “Direct-write patterning of indium-tin-oxide film by high pulse repetition frequency femtosecond laser ablation,” Appl. Opt. 46(23), 5792–5799 (2007). [CrossRef]
10. H. Y. Kim, W. S. Choi, S. Y. Ji, Y. G. Shin, J. W. Jeon, S. Ahn, and S. H. Cho, “Morphologies of femtosecond laser ablation of ITO thin films using gaussian or quasi-flat top beams for OLED repair,” Appl. Phys. A: Mater. Sci. Process. 124(2), 123 (2018). [CrossRef]
11. C. Y. Leong, S. S. Yap, G. L. Ong, T. S. Ong, S. L. Yap, Y. T. Chin, S. F. Lee, T. Y. Tou, and C. H. Nee, “Single pulse laser removal of indium tin oxide film on glass and polyethylene terephthalate by nanosecond and femtosecond laser,” Nanotechnol. Rev. 9(1), 1539–1549 (2020). [CrossRef]
12. G. Račiukaitis, M. Brikas, P. Gečys, B. Voisiat, and M. Gedvilas, “Use of high repetition rate and high power lasers in microfabrication: How to keep the efficiency high?” J. Laser Micro/Nanoeng. 4(3), 186–191 (2009). [CrossRef]
13. A. Risch and R. Hellmann, “Picosecond laser patterning of ITO thin films,” Phys. Procedia 12(PART 2), 133–140 (2011). [CrossRef]
14. N. Farid, H. Chan, D. Milne, A. Brunton, and G. M. O’Connor, “Stress assisted selective ablation of ITO thin film by picosecond laser,” Appl. Surf. Sci. 427, 499–504 (2018). [CrossRef]
15. O. Yavas and M. Takai, “Effect of substrate absorption on the efficiency of laser patterning of indium tin oxide thin films,” J. Appl. Phys. 85(8), 4207–4212 (1999). [CrossRef]
16. O. Yavas, C. Ochiai, and M. Takai, “Substrate-assisted laser patterning of indium tin oxide thin films,” Appl. Phys. A: Mater. Sci. Process. 69(7), S875–S878 (1999). [CrossRef]
17. M. Takai, D. Bollmann, and K. Haberger, “Maskless patterning of indium tin oxide layer for flat panel displays by diode-pumped Nd:YLF laser irradiation,” Appl. Phys. Lett. 64(19), 2560–2562 (1994). [CrossRef]
18. T. Szörényi, Z. Kántor, and L. D. Laude, “Atypical characteristics of KrF excimer laser ablation of indium-tin oxide films,” Appl. Surf. Sci. 86(1-4), 219–222 (1995). [CrossRef]
19. S. Xiao, E. L. Gurevich, and A. Ostendorf, “Incubation effect and its influence on laser patterning of ITO thin film,” Appl. Phys. A: Mater. Sci. Process. 107(2), 333–338 (2012). [CrossRef]
20. C. Mcdonnell, D. Milne, H. Chan, D. Rostohar, and G. M. O’Connor, “Part 2: Ultra-short pulse laser patterning of very thin indium tin oxide on glass substrates,” Opt. Lasers Eng. 81, 70–78 (2016). [CrossRef]
21. D. Nieto, J. Arines, G. M. O’Connor, and M. T. Flores-Arias, “Single-pulse laser ablation threshold of borosilicate, fused silica, sapphire, and soda-lime glass for pulse widths of 500 fs, 10 ps, 20 ns,” Appl. Opt. 54(29), 8596 (2015). [CrossRef]
22. R. A. Synowicki, B. D. Johs, and A. C. Martin, “Optical properties of soda-lime float glass from spectroscopic ellipsometry,” Thin Solid Films 519(9), 2907–2913 (2011). [CrossRef]
23. H. Fujiwara and R. W. Collins, Spectroscopic Ellipsometry for Photovoltaics Volume 1: Fundamental Principles and Solar Cell Characterization (Springer Series in Optical Sciences, 2018, Chap. 18.2.3).
24. L. Jing, W. Hong, and Y. You-Ran, “Characterization of spectral selectivity of tin-doped In2O3 thin films using spectroscopic ellipsometry,” Surf. Coat. Technol. 267, 45–52 (2015). [CrossRef]
25. P. Liu, W. Wang, X. Mei, B. Liu, and W. Zhao, “Rear-side picosecond laser ablation of indium tin oxide micro-grooves,” Opt. Lasers Eng. 69, 35–39 (2015). [CrossRef]
26. L. V. Zhigilei and B. J. Garrison, “Microscopic mechanisms of laser ablation of organic solids in the thermal and stress confinement irradiation regimes,” J. Appl. Phys. 88(3), 1281–1298 (2000). [CrossRef]
27. L. V. Zhigilei, Z. Lin, and D. S. Ivanov, “Atomistic modeling of short pulse laser ablation of metals: Connections between melting, spallation, and phase explosion,” J. Phys. Chem. C 113(27), 11892–11906 (2009). [CrossRef]
28. M. Domke, S. Rapp, M. Schmidt, and H. P. Huber, “Ultra-fast movies of thin-film laser ablation,” Appl. Phys. A: Mater. Sci. Process. 109(2), 409–420 (2012). [CrossRef]
29. C. Wu and L. V. Zhigilei, “Microscopic mechanisms of laser spallation and ablation of metal targets from large-scale molecular dynamics simulations,” Appl. Phys. A: Mater. Sci. Process. 114(1), 11–32 (2014). [CrossRef]
30. S. Rapp, M. Schmidt, and H. P. Huber, “Selective femtosecond laser structuring of dielectric thin films with different band gaps: a time-resolved study of ablation mechanisms,” Appl. Phys. A: Mater. Sci. Process. 122(12), 1035 (2016). [CrossRef]
31. S. Rapp, G. Heinrich, M. Wollgarten, H. P. Huber, and M. Schmidt, “Physical mechanisms of SiNx layer structuring with ultrafast lasers by direct and confined laser ablation,” J. Appl. Phys. 117(10), 105304 (2015). [CrossRef]
32. S. Rapp, J. Rosenberger, M. Domke, G. Heise, H. P. Huber, and M. Schmidt, “Ultrafast pump-probe microscopy reveals the mechanism of selective fs laser structuring of transparent thin films for maskless micropatterning,” Appl. Surf. Sci. 290, 368–372 (2014). [CrossRef]
33. S. Rapp, G. Heinrich, M. Domke, and H. P. Huber, “The combination of direct and confined laser ablation mechanisms for the selective structuring of thin silicon nitride layers,” Phys. Procedia 56(C), 998–1006 (2014). [CrossRef]
34. J. Winter, M. Spellauge, J. Hermann, C. Eulenkamp, H. P. Huber, and M. Schmidt, “Ultrashort single-pulse laser ablation of Stainless Steel, Aluminium, Copper and its dependence on the pulse duration,” Opt. Express 29(10), 14561–14581 (2021).
35. M. Spellauge, J. Winter, S. Rapp, C. McDonnell, F. Sotier, M. Schmidt, and H. P. Huber, “Influence of stress confinement, particle shielding and re-deposition on the ultrashort pulse laser ablation of metals revealed by ultrafast time-resolved experiments,” Appl. Surf. Sci. 545, 148930 (2021). [CrossRef]
36. C. McDonnell, D. Milne, H. Chan, D. Rostohar, and G. M. O’Connor, “Part 1: Wavelength dependent nanosecond laser patterning of very thin indium tin oxide films on glass substrates,” Opt. Lasers Eng. 80, 73–82 (2016). [CrossRef]
37. S. Krause, P. T. Miclea, F. Steudel, S. Schweizer, and G. Seifert, “Precise microstructuring of indium-tin oxide thin films on glass by selective femtosecond laser ablation,” EPJ Photovolt. 4, 40601 (2013). [CrossRef]
38. J. Sotrop, A. Kersch, M. Domke, G. Heise, and H. P. Huber, “Numerical simulation of ultrafast expansion as the driving mechanism for confined laser ablation with ultra-short laser pulses,” Appl. Phys. A: Mater. Sci. Process. 113(2), 397–411 (2013). [CrossRef]
39. G. Heise, M. Englmaier, C. Hellwig, T. Kuznicki, S. Sarrach, and H. P. Huber, “Laser ablation of thin molybdenum films on transparent substrates at low fluences,” Appl. Phys. A: Mater. Sci. Process. 102(1), 173–178 (2011). [CrossRef]
40. M. Z. Alam, I. De Leon, and R. W. Boyd, “Large optical nonlinearity of indium tin oxide in its epsilon-near-zero region,” Science 352(6287), 795–797 (2016). [CrossRef]
41. K. Burrows and V. Fthenakis, “Glass needs for a growing photovoltaics industry,” Sol. Energy Mater. Sol. Cells 132, 455–459 (2015). [CrossRef]
42. M. Domke, S. Rapp, M. Schmidt, and H. P. Huber, “Ultrafast pump-probe microscopy with high temporal dynamic range,” Opt. Express 20(9), 10330 (2012). [CrossRef]
43. P. Boerner, M. Hajri, T. Wahl, J. Weixler, and K. Wegener, “Picosecond pulsed laser ablation of dielectric rods: Angle-dependent ablation process model for laser micromachining,” J. Appl. Phys. 125(23), 234902 (2019). [CrossRef]
44. D. Nečas and P. Klapetek, “Gwyddion: An open-source software for SPM data analysis,” Cent. Eur. J. Phys. 10(1), 181–188 (2012). [CrossRef]
45. D. Redka, C. Gadelmeier, J. Winter, M. Spellauge, C. Eulenkamp, P. Calta, U. Glatzel, J. Minár, and H. P. Huber, “Sub-picosecond single-pulse laser ablation of the CrMnFeCoNi high entropy alloy and comparison to stainless steel AISI 304,” Appl. Surf. Sci. 544, 148839 (2021). [CrossRef]
46. J. M. Liu, “Simple technique for measurements of pulsed Gaussian-beam spot sizes,” Opt. Lett. 7(5), 196–198 (1982). [CrossRef]
47. S. Xiao, B. Schöps, and A. Ostendorf, “Selective Ablation of Thin Films by Ultrashort Laser Pulses,” Phys. Procedia 39, 594–602 (2012). [CrossRef]
48. B. H. Christensen, K. Vestentoft, and P. Balling, “Short-pulse ablation rates and the two-temperature model,” Appl. Surf. Sci. 253(15), 6347–6352 (2007). [CrossRef]
49. M. Domke, J. Sotrop, S. Rapp, M. Börger, D. Felsl, and H. P. Huber, “Transient temperature modeling and shock wave observation in confined laser ablation of thin molybdenum films,” Proc. SPIE 8611(March 2013), 86111B (2013). [CrossRef]
50. M. Domke, D. Felsl, S. Rapp, J. Sotrop, H. P. Huber, and M. Schmidt, “Evidence of pressure waves in confined laser ablation,” J. Laser Micro/Nanoeng. 10(2), 119–123 (2015). [CrossRef]
51. J. Furmanski, A. M. Rubenchik, M. D. Shirk, and B. C. Stuart, “Deterministic processing of alumina with ultrashort laser pulses,” J. Appl. Phys. 102(7), 073112 (2007). [CrossRef]
52. M. Garcia-Lechuga, J. Siegel, J. Hernandez-Rueda, and J. Solis, “Imaging the ultrafast Kerr effect, free carrier generation, relaxation and ablation dynamics of Lithium Niobate irradiated with femtosecond laser pulses,” J. Appl. Phys. 116(11), 113502 (2014). [CrossRef]
53. J. E. Gorfain, C. T. Key, and C. S. Alexander, “Application of a Computational Glass Model to the Shock Response of Soda-Lime Glass,” J. Dyn. Behav. Mater. 2(3), 283–305 (2016). [CrossRef]
54. N. Preissler, O. Bierwagen, A. T. Ramu, and J. S. Speck, “Electrical transport, electrothermal transport, and effective electron mass in single-crystalline In2O3 films,” Phys. Rev. B: Condens. Matter Mater. Phys. 88(8), 085305 (2013). [CrossRef]
55. G. Paltauf and P. E. Dyer, “Photomechanical processes and effects in ablation,” Chem. Rev. 103(2), 487–518 (2003). [CrossRef]
56. E. Leveugle, D. S. Ivanov, and L. V. Zhigilei, “Photomechanical spallation of molecular and metal targets: Molecular dynamics study,” Appl. Phys. A: Mater. Sci. Process. 79(7), 1643–1655 (2004). [CrossRef]
57. T. Ashida, A. Miyamura, N. Oka, Y. Sato, T. Yagi, N. Taketoshi, T. Baba, and Y. Shigesato, “Thermal transport properties of polycrystalline tin-doped indium oxide films,” J. Appl. Phys. 105(7), 073709 (2009). [CrossRef]