FIB-SEM Investigation of Laser-Induced Periodic Surface Structures and Conical Surface Microstructures on D16T (AA2024-T4) Alloy

The use of aluminum alloy AA2024-T4 (Russian designation D16T) in applications requiring a high strength-to-weight ratio and fatigue resistance such as aircraft fuselage often demands the control and modification of surface properties. A promising route to surface conditioning of Al alloys is laser treatment. In the present work, the formation of ripples and conical microstructures under scanning with femtosecond (fs) laser pulses was investigated. Laser treatment was performed using 250 fs pulses of a 1033 nm Yb:YAG laser. The fluence of the pulses varied from 5 to 33 J/cm2. The scanning was repeated from 1 to 5 times for different areas of the sample. Treated areas were evaluated using focused ion beam (FIB)- scanning electron microscopy (SEM) imaging and sectioning, energy-dispersive X-ray (EDX) spectroscopy, atomic force microscopy (AFM), and confocal laser profilometry. The period of laser-induced periodic surface structures (LIPSS) and the average spacing of conical microstructures were deduced from SEM images by FFT. Unevenness of the treated areas was observed that is likely to have been caused by ablation debris. The structural and elemental changes of the material inside the conical microstructures was revealed by FIB-SEM and EDX. The underlying formation mechanisms of observed structures are discussed in this paper.


Introduction
Laser treatment with femtosecond pulses (fs) has emerged as a topic of intensive investigation due to the recent advances in femtosecond pulse laser availability. Femtosecond lasers are an effective tool for the precision processing of metals [1], having multiple advantages over the more conventional nanosecond pulse lasers. Unlike the latter, which create an extended heat-affected zone, femtosecond pulse lasers can provide a very rapid transition from the solid state to vapor or plasma omitting Metals 2020, 10, 144 3 of 16 A detailed coverage of conical microstructure growth on aluminum and other metals at 800 nm 130 fs laser pulses is provided by Nayak [24]. More specifically, bump-like microstructures along with maze-like microstructures were observed on AA2024 [25], and the effects of laser fluence and the number of passes on the roughness and reflectivity of ablated areas were examined [26].
One can conclude that the complex character of matter-radiation interaction over a wide range of fluences causes versatile surface topological and microchemical modification for commonly used engineering alloys and offers a promising dedicated procedure for advanced surface enhancement. This highlights the importance of rigorous characterization of surface structures using modern techniques such as focused ion beam (FIB) milling and lamella cutting.
The present work is devoted to the focused investigation of D16T alloy, the Russian equivalent of AA2024-T4 alloy. Using non-destructive atomic force microscopy (AFM) and confocal laser profilometry, high resolution SEM, and FIB milling, we systematically investigated the formed LIPSS and conical microstructures reaching 50 µm in size and covered by nanoparticles, which were created as a result of laser treatment using 250 fs pulses of a 1033 nm Yb:YAG laser. We discuss how laser scan patterns can lead to non-uniform profiles and surface structure variations. Internal non-uniformity of the micrometer-sized cones is first investigated at the nanometer scale using fine FIB-SEM observations.

Materials and Methods
The flat surface of a mechanically polished cylinder (diameter 20 mm, height 5 mm) of D16T alloy (Table 1) was treated with laser pulses with fluences varying over the range of 5 to 33 J/cm 2 ( Figure 1a and Table 2). Roman numerals correspond to the fluence of the pulses, from I corresponding to 5 J/cm 2 to VII corresponding to 33 J/cm 2 . An MLP1-2106 (Lasers & Apparatus Ltd., Zelenograd, Russia) device in a customized commercial setup was used. Scanning of each individual area was achieved by redirecting the beam with a scan head (HurryScan 14, SCANLAB, Puchheim, Germany), whereas switching between different areas was done by moving the stage. The incident laser beam was vertical and linearly polarized along the y-axis direction (Figure 1b). The laser source used was Yb:YAG (TETA-10, Avesta, Troitsk, Russia) and had a wavelength of 1033 ± 3 nm. The initial beam (diameter 12 mm) was focused by an F-theta lens with a 70 mm focus distance. Elongated (1.5 × 0.5 mm 2 ) rectangular areas were scanned with the laser beam longitudinally in the same x-axis direction for all laser beam tracks ( Figure 1). The focused beam radius was calculated to be 35 µm. Beam overlap along the scanning lines was calculated to be 34%, whereas the overlap between adjacent scanning lines was 71%.  The pulses had a repetition rate of 25 kHz, which for a 580 mm/s scanning speed gives a 23.2 µ m distance between separate pulses. The 50 longitudinal scanning lines were separated by a 10 μm distance, which is less than the separating distance in the direction of the raster scan. For each value of energy, the scan repetition (SR) number varied from 1 to 5. A PVE300 Photovoltaic Device Characterization System (Bentham Instruments, Reading, UK) was used to estimate the reflectivity in the visible range (300-900 nm). The confocal laser profilometer CCS Optima+ (STIL SAS, Aix-en-Provence, France) with a CL2MG140 optical pen having a z-sensitivity of 40 nm was used for profilometry. Tescan VEGA 3 LMU (Brno, Czech Republic) and Thermo Scientific Quattro ESEM (ThermoFischer, Waltham, MA, USA) devices were used to perform SEM imaging in high vacuum mode. A Thermo Scientific Helios G4 PFIB-SEM UXe DualBeam (ThermoFischer, Walthem, MA, USA) was used for ion milling and lamella preparation. EDX was performed using the Octave Elite Super module (EDAX Inc., Mahwah, NJ, USA). An atomic force microscope SMENA (NT-MDT, Zelenograd, Russia) was used for AFM profilometry. The LIPSS period and average distance between separate cones in the conical surface microstructures were estimated using ImageJ (1.52a, NIH,   I  60  5  II  110  9  III  160  13  IV  210  17  V  260  21  VI  310  25  VII  400  33 The pulses had a repetition rate of 25 kHz, which for a 580 mm/s scanning speed gives a 23.2 µm distance between separate pulses. The 50 longitudinal scanning lines were separated by a 10 µm distance, which is less than the separating distance in the direction of the raster scan. For each value of energy, the scan repetition (SR) number varied from 1 to 5. A PVE300 Photovoltaic Device Characterization System (Bentham Instruments, Reading, UK) was used to estimate the reflectivity in the visible range (300-900 nm). The confocal laser profilometer CCS Optima+ (STIL SAS, Aix-en-Provence, France) with a CL2MG140 optical pen having a z-sensitivity of 40 nm was used for profilometry. Tescan VEGA 3 LMU (Brno, Czech Republic) and Thermo Scientific Quattro ESEM (ThermoFischer, Waltham, MA, USA) devices were used to perform SEM imaging in high vacuum mode. A Thermo Scientific Helios G4 PFIB-SEM UXe DualBeam (ThermoFischer, Walthem, MA, USA) was used for ion milling and lamella preparation. EDX was performed using the Octave Elite Super module (EDAX Inc., Mahwah, NJ, USA). An atomic force microscope SMENA (NT-MDT, Zelenograd, Russia) was used for AFM profilometry. The LIPSS period and average distance between separate cones Metals 2020, 10, 144 5 of 16 in the conical surface microstructures were estimated using ImageJ (1.52a, NIH, Bethesda, MD, USA) (https://imagej.nih.gov/ij/index.html) to process SEM and AFM images by means of FFT peak analysis.

Profilometry
Transversal and longitudinal profiles of the ablated areas were investigated by confocal profilometry (Figure 2).
The transversal and longitudinal profiles were clearly different, as the longitudinal profile has its mean line mostly parallel to the sample surface. A dip is observable at the side where the scanning lines start, which is due to the beam not being shut when it switches between scanning lines (Figure 1b). The transversal profile, however, is non-uniform, deeper, and rougher on one side than on the other and therefore its mean line has a noticeable slope. The shallower side corresponds to the side where the laser beam starts the first scanning track. Possible reasons for this effect are discussed later. Bethesda, MD, USA) (https://imagej.nih.gov/ij/index.html) to process SEM and AFM images by means of FFT peak analysis.

Profilometry
Transversal and longitudinal profiles of the ablated areas were investigated by confocal profilometry (Figure 2).
The transversal and longitudinal profiles were clearly different, as the longitudinal profile has its mean line mostly parallel to the sample surface. A dip is observable at the side where the scanning lines start, which is due to the beam not being shut when it switches between scanning lines ( Figure  1b). The transversal profile, however, is non-uniform, deeper, and rougher on one side than on the other and therefore its mean line has a noticeable slope. The shallower side corresponds to the side where the laser beam starts the first scanning track. Possible reasons for this effect are discussed later. A color-coded chart characterizing the profiles of the treated areas is provided (Figure 3a). The full profiles are presented as Supplementary Materials ( Figure S1). The number of scan repetitions varies from 1 to 5 from left to right, the fluence varies from 13 J/cm 2 to 33 J/cm 2 from top to bottom. The profiles corresponding to a lower fluence (less than 13 J/cm 2 ) could not be reliably registered because of the low signal-to-noise ratio. AFM and SEM were applied to study these areas.
Three different trends can be deduced from the diagram. First, for 13 J/cm 2 , the profile mean lines are parallel to the surface and they have relatively low roughness until spiky peaks appear at higher numbers of pulse repetitions; that range is marked in green. An example profile is provided (Figure 3b). The depth of the profiles increases monotonously with the number of pulses ( Figure 4a). Second, at a higher energy treatment, the spiky peaks are not yet identifiable on the profiles after smaller numbers of pulse repetitions, but the profile mean lines are noticeably sloped; those profiles are marked in yellow. An example profile is provided (Figure 3c). The mean slope angle of the profile increases with both energy and the number of pulse repetitions (Figure 4b). The slope angle is small but always positive. The 21 J/cm 2 one-repetition treated area is excluded, since it was too close to the edge and was curved due to polishing.
Finally, at high values of energy and numbers of pulse repetitions, the profiles reveal abrupt peaks and dips, and their height increases with energy and number of pulse repetitions. Those profiles are marked in red. An example profile is provided (Figure 3d) In general, such transversal profiles show uneven roughness-the peak-dip pattern is deeper and rougher at the side of the treated area where the laser beam was scanning later. A color-coded chart characterizing the profiles of the treated areas is provided ( Figure 3a). The full profiles are presented as Supplementary Materials ( Figure S1). The number of scan repetitions varies from 1 to 5 from left to right, the fluence varies from 13 J/cm 2 to 33 J/cm 2 from top to bottom. The profiles corresponding to a lower fluence (less than 13 J/cm 2 ) could not be reliably registered because of the low signal-to-noise ratio. AFM and SEM were applied to study these areas.
Three different trends can be deduced from the diagram. First, for 13 J/cm 2 , the profile mean lines are parallel to the surface and they have relatively low roughness until spiky peaks appear at higher numbers of pulse repetitions; that range is marked in green. An example profile is provided ( Figure 3b). The depth of the profiles increases monotonously with the number of pulses ( Figure 4a).
Second, at a higher energy treatment, the spiky peaks are not yet identifiable on the profiles after smaller numbers of pulse repetitions, but the profile mean lines are noticeably sloped; those profiles are marked in yellow. An example profile is provided (Figure 3c). The mean slope angle of the profile increases with both energy and the number of pulse repetitions (Figure 4b). The slope angle is small but always positive. The 21 J/cm 2 one-repetition treated area is excluded, since it was too close to the edge and was curved due to polishing.
Finally, at high values of energy and numbers of pulse repetitions, the profiles reveal abrupt peaks and dips, and their height increases with energy and number of pulse repetitions. Those profiles are marked in red. An example profile is provided ( Figure 3d) In general, such transversal profiles show Metals 2020, 10, 144 6 of 16 uneven roughness-the peak-dip pattern is deeper and rougher at the side of the treated area where the laser beam was scanning later. R a and R z roughness parameters were calculated for the profiles (Tables 3 and 4, respectively).
Metals 2020, 10, x FOR PEER REVIEW 6 of 16 Ra and Rz roughness parameters were calculated for the profiles (Tables 3 and 4, respectively).  Figure 3. Profiles of laser-treated areas: (a) color-coded chart; (b) example profile for "green" profiles which are parallel to the initil surface; (c) example profile for "yellow" profiles with an observable slope; (d) example profile for "red" profiles with high roughness. SR: scan repetition.  Figure 3. Profiles of laser-treated areas: (a) color-coded chart; (b) example profile for "green" profiles which are parallel to the initil surface; (c) example profile for "yellow" profiles with an observable slope; (d) example profile for "red" profiles with high roughness. SR: scan repetition.

(b) (a)
Metals 2020, 10, x FOR PEER REVIEW 6 of 16 Ra and Rz roughness parameters were calculated for the profiles (Tables 3 and 4, respectively).  Figure 3. Profiles of laser-treated areas: (a) color-coded chart; (b) example profile for "green" profiles which are parallel to the initil surface; (c) example profile for "yellow" profiles with an observable slope; (d) example profile for "red" profiles with high roughness. SR: scan repetition.

LIPSS: SEM Imaging and Analysis
The collage of SEM images characterizing the appearance of treated areas is given in Figure 5 as a function of laser beam energy and number of scan repetitions. Treated areas are generally darker than the surrounding alloy surface. SEM reveals two types of surface motifs formed by femtosecond laser ablation-the relatively even surface is populated with periodic ripples (LIPSS) resolvable at higher magnifications and rough and bright grainy clusters constructed by conical microstructures. Both motifs may be simultaneously present in the treated area; however, generally the conical microstructure is characteristic of a higher beam energy and a higher number of scan repetitions, that is, larger doses of treatment energy applied. The portion of conical microstructure in the treated area is always greater at higher values of y, that is, for the last scan tracks, which strongly correlates with the data for treated area profiles. Separate single cones appear relatively randomly over the surface of treated areas at low fluence and the smaller number of scan repetitions, but they are more frequent at larger y values and may even occur outside the treated area. We believe that the main mechanism triggering the formation of conical microstructures is the redeposition of ablated material; however, the intensity of ablation is stimulated by laser treatment at neighboring tracks, and therefore an avalanche (or chain reaction) effect takes place when the redeposition of material at previous tracks enhances the ablation at the current track and the redeposition at the next track. On the other hand, at low fluence the most representative structures are the LIPSS (Figure 6a,d). The SEM images in secondary electrons had a satisfactory contrast of the ripples, allowing the determination of their period using FFT image processing. The area treated with 9 J/cm 2 fluence with the scanning repeated 5 times is depicted as an example (Figure 6a,b). The LIPSS having obvious periodicity exhibit some irregularity (Figure 6a), which is also demonstrated by the broad peak at the FFT spectrum (Figure 6c,d). The LIPSS period for this area is equal to 0.78 ± 0.13 µm. This value within the range of standard deviation was found to be invariant for all studied treated areas showing almost no dependence on fluence and scan repetitions. On the other hand, at low fluence the most representative structures are the LIPSS (Figure 6a,d). The SEM images in secondary electrons had a satisfactory contrast of the ripples, allowing the determination of their period using FFT image processing. The area treated with 9 J/cm 2 fluence with the scanning repeated 5 times is depicted as an example (Figure 6a,b). The LIPSS having obvious periodicity exhibit some irregularity (Figure 6a), which is also demonstrated by the broad peak at the FFT spectrum (Figure 6c,d). The LIPSS period for this area is equal to 0.78 ± 0.13 µm. This value within the range of standard deviation was found to be invariant for all studied treated areas showing almost no dependence on fluence and scan repetitions.
A granular structure of the ripples, however, is visible at high magnification ( Figure 6b). Perhaps, these granules resulted from the redeposition process, and some of them may play a role as cone precursors during an ongoing laser treatment.
Metals 2020, 10, x FOR PEER REVIEW 9 of 16 A granular structure of the ripples, however, is visible at high magnification ( Figure 6b). Perhaps, these granules resulted from the redeposition process, and some of them may play a role as cone precursors during an ongoing laser treatment.

LIPSS: AFM Profilometry and Analysis
Laser confocal profilometry cannot resolve the LIPSS formed at the lowest energy regime. Therefore, the characteristics of periodic ripples formed at a fluence of 5 J/cm 2 were studied by AFM (Figure 7).

LIPSS: AFM Profilometry and Analysis
Laser confocal profilometry cannot resolve the LIPSS formed at the lowest energy regime. Therefore, the characteristics of periodic ripples formed at a fluence of 5 J/cm 2 were studied by AFM (Figure 7). Metals 2020, 10, x FOR PEER REVIEW 9 of 16 A granular structure of the ripples, however, is visible at high magnification ( Figure 6b). Perhaps, these granules resulted from the redeposition process, and some of them may play a role as cone precursors during an ongoing laser treatment.

LIPSS: AFM Profilometry and Analysis
Laser confocal profilometry cannot resolve the LIPSS formed at the lowest energy regime. Therefore, the characteristics of periodic ripples formed at a fluence of 5 J/cm 2 were studied by AFM (Figure 7).   The calculated LIPSS periods and peak heights are given in Table 5.

Conical Microstructure Imaging and Analysis
The increase of energy and number of pulses leads to the emergence and growth of conical microstructure clusters and the simultaneous depletion of the LIPSS area.
For the laser treatment at 13 J/cm 2 fluence, the conical microstructures become noticeable even after a single scan ( Figure 6). With the rise of pulse energies and numbers of repetitions, the conical microstructures cover an increased portion of the treated area. Moreover, SEM imaging of the conical microstructures indicates that the increase of fluence and number of pulses tends to enhance the aspect factor of a single cone. For example, for the twice-repeated 21 J/cm 2 fluence laser treatment, the conical microstructures are represented by 2.5-3 µm high and less than 10 µm wide cones, whereas for the 33 J/cm 2 fluence and five repetitions, cones reach a height of 30 µm and a width of 20-25 µm (Figure 8a). A pattern of lines is observable on the slopes of some conical structures (Figure 8b). These lines are most likely LIPSS.
The focused ion beam (FIB) cutting of a cone grown after the treatment at 33 J/cm 2 fluence with five scan repetitions reveals that the content of the tip is different from that of the bulk of the alloy (Figure 8e).
A comparison between SEM images acquired before and after cleaning in an ultrasonic bath in acetone and isopropanol is shown in Figure 8c,d. After cleaning, the removal of the loose layer of nanoparticles can be observed on the slopes. The calculated LIPSS periods and peak heights are given in Table 5.

Conical Microstructure Imaging and Analysis
The increase of energy and number of pulses leads to the emergence and growth of conical microstructure clusters and the simultaneous depletion of the LIPSS area.
For the laser treatment at 13 J/cm 2 fluence, the conical microstructures become noticeable even after a single scan ( Figure 6). With the rise of pulse energies and numbers of repetitions, the conical microstructures cover an increased portion of the treated area. Moreover, SEM imaging of the conical microstructures indicates that the increase of fluence and number of pulses tends to enhance the aspect factor of a single cone. For example, for the twice-repeated 21 J/cm 2 fluence laser treatment, the conical microstructures are represented by 2.5-3 µm high and less than 10 µm wide cones, whereas for the 33 J/cm 2 fluence and five repetitions, cones reach a height of 30 µm and a width of 20-25 µm (Figure 8a). A pattern of lines is observable on the slopes of some conical structures ( Figure  8b). These lines are most likely LIPSS.
The focused ion beam (FIB) cutting of a cone grown after the treatment at 33 J/cm 2 fluence with five scan repetitions reveals that the content of the tip is different from that of the bulk of the alloy (Figure 8e).
A comparison between SEM images acquired before and after cleaning in an ultrasonic bath in acetone and isopropanol is shown in Figure 8c,d. After cleaning, the removal of the loose layer of nanoparticles can be observed on the slopes.  By analyzing the position of the small and wide peak in the FFT spectrum, it was possible to estimate the average distance between the cone tops in the conical structures (Figure 9b). The cones are surrounded by deep furrows, which gives a recognizable pattern of dark lines (Figure 9a). The peak in the FFT spectrum does not represent true periodicity, but identifies the most frequent distance between two furrows (Table 6).  The internal structure of a cone formed by laser ablation treatment at 33 J/cm 2 with five scan repetitions was investigated by FIB cross-sectional cutting. A protective Pt layer was applied to avoid curtaining defects. The cut (Figure 10a) was analyzed by EDX (Figure 10b-d), then a thin lamella was cut out for higher resolution SEM (Figure 10e) and EDX (Figure 10f-h) imaging. An interesting area with a brighter SEM signal, deep oxidation, and a reduced Al EDX signal is observable in the crosssection and is marked in red (Figure 10a-c). Further features worth consideration can be seen in the By analyzing the position of the small and wide peak in the FFT spectrum, it was possible to estimate the average distance between the cone tops in the conical structures (Figure 9b). The cones are surrounded by deep furrows, which gives a recognizable pattern of dark lines (Figure 9a). The peak in the FFT spectrum does not represent true periodicity, but identifies the most frequent distance between two furrows (Table 6). By analyzing the position of the small and wide peak in the FFT spectrum, it was possible to estimate the average distance between the cone tops in the conical structures (Figure 9b). The cones are surrounded by deep furrows, which gives a recognizable pattern of dark lines (Figure 9a). The peak in the FFT spectrum does not represent true periodicity, but identifies the most frequent distance between two furrows (Table 6).  The internal structure of a cone formed by laser ablation treatment at 33 J/cm 2 with five scan repetitions was investigated by FIB cross-sectional cutting. A protective Pt layer was applied to avoid curtaining defects. The cut (Figure 10a) was analyzed by EDX (Figure 10b-d), then a thin lamella was cut out for higher resolution SEM (Figure 10e) and EDX (Figure 10f-h) imaging. An interesting area with a brighter SEM signal, deep oxidation, and a reduced Al EDX signal is observable in the crosssection and is marked in red (Figure 10a-c). Further features worth consideration can be seen in the  The internal structure of a cone formed by laser ablation treatment at 33 J/cm 2 with five scan repetitions was investigated by FIB cross-sectional cutting. A protective Pt layer was applied to avoid curtaining defects. The cut (Figure 10a) was analyzed by EDX (Figure 10b-d), then a thin lamella was cut out for higher resolution SEM (Figure 10e) and EDX (Figure 10f-h) imaging. An interesting area with a brighter SEM signal, deep oxidation, and a reduced Al EDX signal is observable in the cross-section and is marked in red (Figure 10a-c). Further features worth consideration can be seen in the thin lamella imaged using SEM and EDX. The feature marked in yellow (Figure 10e-h) reveals an area of darker SEM signal (Figure 10e), an enclave with higher Al concentration (Figure 10f), and a pore. The pore gives a bright signal in the Al EDX image due to multiple electron scattering (Figure 10f). Both the aluminum enclave and the pore show a "shadow" of reduced Cu and O concentration (Figure 10g,h). The anomaly outlined in white is most likely a pore (Figure 10e-h). The black-circled anomaly (Figure 10e-h) does not exhibit noticeable SEM contrast, but EDX reveals a higher concentration of Cu ( Figure 10h) and a decrease in Al concentration (Figure 10g).
Metals 2020, 10, x FOR PEER REVIEW 12 of 16 thin lamella imaged using SEM and EDX. The feature marked in yellow (Figure 10e-h) reveals an area of darker SEM signal (Figure 10e), an enclave with higher Al concentration (Figure 10f), and a pore. The pore gives a bright signal in the Al EDX image due to multiple electron scattering ( Figure  10f). Both the aluminum enclave and the pore show a "shadow" of reduced Cu and O concentration (Figure 10g,h). The anomaly outlined in white is most likely a pore (Figure 10e-h). The black-circled anomaly (Figure 10e-h) does not exhibit noticeable SEM contrast, but EDX reveals a higher concentration of Cu ( Figure 10h) and a decrease in Al concentration (Figure 10g).

Discussion
We believe that the slope of the mean bottom line and the uneven roughness of the treated areas are caused by the influence of redeposited ablated material. Ablated material from the areas exposed to the earlier laser pulses partially redeposits to the areas that are exposed later. This leads to decreased light reflectivity of the latter areas and, as a result, their higher absorption and more intensive further ablation. Moreover, this increases the probability of forming conical microstructures, since the redeposited nanoparticles make the ablation less uniform causing a positive feedback for cone growth in a self-stimulating process. Thus, conical structures starting to grow earlier at the edge of the treated area will have a higher aspect ratio, leading to non-uniform roughness. This effect may be undesirable for engineering applications.
The influence of redeposited ablated material on the formation of conical microstructures was described by Zuhlke et al. [18] in detail. Ablated nanoparticles redeposit when the beam is away, and later, when the scanning beam reaches them again, a portion of the nanoparticles melts. This forms a structure with alternating layers of melted and non-melted particles. The similarity between this work and our work is confirmed when we clearly observe that the conical structures appear to be covered with the layer of sharp nanoparticles, while ultrasonic bath cleaning with acetone removes any loose layer of nanoparticles.
AA2024 that is ablated to the point of conical structure growth has been demonstrated to be a broadband absorber [26]. We could easily observe this in our work. The low reflectivity can be observed with the naked eye, as the treated areas look black (Figure 1a). The decrease in reflectivity was approximately evaluated by an integrating sphere measurement. The sphere PVE300 illuminated a spot of 1 mm diameter that was obviously bigger than the width of the treated area; therefore, the signal was collected from both the treated area and the surrounding untreated surface. The untreated alloy's surface showed a reflectivity of 79 ± 5% in the visible light (300-900 nm) wavelength range, whereas the spot that was positioned in the center of the laser-treated area illuminating both the laser-treated area and the untreated surrounding alloy had an integral reflectivity of 42 ± 2% within the same wavelength range. Applying the linear rule of mixture to the contributions of treated and untreated areas, it was possible to estimate the reflectivity of the treated area as 21 ± 2%.
LIPSS with a 703 nm period were observed on AA2024 in a very recent work for a scan with an 800 fs pulse 1030 nm laser [22]. Similarly, LIPSS on aluminum are mentioned in the work of Umm-i-Kalsoom et al. [3]. Those LIPSS were observed at much higher fluences for a treatment conducted using a 30 fs pulse duration and an incident laser wavelength of 800 nm.
Bashir et al. observed LIPSS on Al for low energy fluences, 25 fs pulses, and an 800 nm incident wavelength. The authors concluded that the LIPSS observed were due to the excitation of surface plasmon polaritons and that the period of the structures depended sensitively on the material selected and the laser fluence [21].
In our work, the LIPSS that we observed were irregular and their period Λ seemed to show no dependence on fluence in the range we observed, equaling approximately to 0.73·λ (λ-laser wavelength of 1033 nm). Accordingly, these LIPSS should be attributed to the category of low spatial frequency LIPSS (LSFLs), having a period higher than 0.5·λ. Existing theories explain the formation of LSFLs by the interference of the incident laser pulses and the surface-scattered electromagnetic waves [20].
A detailed coverage of conical microstructure growth on aluminum and other metals at 800 nm 130 fs laser pulses is presented by Nayak and Gupta [24]. A focused ion beam (FIB) cutting of a conical structure grown on titanium is provided, with no internal structural change was observed. Conical structures on aluminum exhibited less regularity than the ones created on stainless steel and titanium.
Bump-like microstructures along with maze-like microstructures were observed on AA2024 [25]. Models were proposed to fine-tune the laser irradiation parameters for the better reproduction of a desired type of microstructure. The differences in the surface structures observed on different metals were explained by differences in the strength of electron-phonon coupling and the thermal conductivities.
Another work providing significant insight into the properties of microstructures grown on AA2024 via femtosecond ablation examines the effect of laser fluence and number of passes on the roughness and reflectivity of ablated areas [26].
Finally, it is worth mentioning the work revealing the contents of an onion-like aggregated nanoparticle sphere on AA2024 via focused ion beam milling [27].
In our case, we supposed that the observed conical structures and deep valleys surrounding them were mostly caused by preferential ablation of lower areas, as described in [18]. However, while this does define the high aspect ratio of the conical structures for the most part, this is not the only mechanism of structure growth, because profilometry reveals the presence of microstructures higher than the initial surface level. This is most likely caused by the redeposition of ablated material on the cone tops, where less ablation occurs compared to other areas.
While detailed research of conical microstructure growth on AA2024 has already been conducted, to the best of our knowledge, no attempts have been made to study the internal contents of the conical microstructures using FIB-SEM and conducting an EDX study of a lamella cut from the tip of a conical structure. A research paper has conducted an EDX study of the surface of ablated AA2024 [4]. However, there are significant differences between that work and ours. Exposures were static (no scanning was conducted). In that work, it was shown that more significant elemental change and oxidation were observed at the rim of laser impact and not in the center [4]. Section 3.4. of this paper provides insight into the structure of the tip of a conical microstructure, grown above the initial surface by redeposition. EDX data reveals the oxidization of treated areas as deep as 3 µm inside the conical microstructures (Figure 10c). Areas of increased concentration of O and Cu are observed with the respective depletion of Al concentration (Figure 10b,f), and vice versa. The inhomogeneous distribution of O, Cu, and Al inside the tip can be caused by a number of overlapping possible mechanisms. Stain-like patterns visible inside the lamella cannot be easily explained by changes in elemental concentration or a change in density.
The formation of conical microstructures and LIPSS can be speculatively thought of as dissipative structures similar to Rayleigh-Bénard cells which are formed in non-equilibrium conditions of energy and mass transport within the system. Particular shapes, aspects and sizes of the Rayleigh-Bénard cells are the stationary solutions of parametrical differential equations specific for the particular processing conditions. It can be anticipated that other than conical types of structures formed at laser ablation treatment can be created using the technique presented herein [28].

Conclusions
Femtosecond pulse laser treatment of metal alloys represents a number of advantages for the gentle and tunable advanced surface enhancement at an industrial scale. The complex multimodal modification of reflectivity, hydrophobicity, nanohardness, and corrosion resistance can be implemented through purposeful surface structuring. LIPSS and surface conical structures having particular topology and size characteristics are formed in certain ranges of parameters, such as laser wavelength, pulse energy, and total fluence. Femtosecond pulse laser treatment of widely applied aged Al-Cu alloys (2024-T4 or Russian equivalent D16T) is of particular practical interest for systematic characterization and research in technological parameter optimization.
We applied both traditional non-destructive AFM, confocal laser profilometry, and high-resolution SEM and modern techniques such as FIB milling and lamella cutting to fill the gaps in the understanding of self-assembled surface structure formation mechanisms. We found that, under laser treatment using 250 fs pulses of a 1033 nm Yb:YAG laser, the unevenness of treated areas occurred at lower fluences (5 J/cm 2 ) in the form of LIPSS developing toward conical structures, which were highly likely formed as a result of redeposited ablation material from earlier stages of ablation affecting later ablation in other areas. In contrast to other reports, LIPSS with periods of 0.73 λ were observed parallel to the incident irradiation direction. No clear dependence of LIPSS characteristics on laser treatment parameters was observed. Overall, the increase of laser fluence or pulse number (SR) gave rise to the formation conical microstructures. The increment of fluence enhanced the area of conical microstructures and overall roughness. The height of cones reached 20-30 µm and the altitude of some cones surpassed the initial surface level, which indicates redeposition mechanisms of growth in addition to preferential ablation. FIB-SEM observations facilitated with EDX revealed that the ablation and redeposition processes were responsible for internal non-uniformity and chemical inhomogeneity at the nanometer scale in the top 3-5 µm of the cone structures. Corrosion tests were considered to assess the effect of copper enrichment on the top of the cones.
Both LIPSS and conical structures caused a significant decrease in reflectivity of the treated areas. We believe that the modification of both chemical and optical surface properties ultimately expands the application diversity for AA2024 after this type of relatively simple, fast, and low-cost surface treatment.