1. INTRODUCTION

The processing of glass with laser radiation compared to conventional processing offers several advantages, such as high geometrical freedom, high lateral and vertical resolution, and fast laser processing without contacting the material. Glass ablation by laser radiation can be realized either by nonlinear effects using ultrashort pulse lasers [1–3] or by heating and evaporating the material using ${\mathrm{CO}}_2$ lasers [4–6]. With ${\mathrm{CO}}_2$ laser radiation, optics of fused silica can be repaired, polished, or form generated by laser ablation [7–9]. In particular, $Q$-switched ${\mathrm{CO}}_2$ lasers are used for ablation and marking processes, but mostly $Q$-switched laser sources are limited to small ablation rates due to their low average power of $P_L<50\mathrm{W}$ [10]. A new high-power $Q$-switched ${\mathrm{CO}}_2$ laser with an average output power of $P_L\ge 200\mathrm{W}$ [11] offers a flexible approach for processing glass materials, especially with a larger ablation rate and thus a reduced processing time. Moreover, due to the pulse duration of $t_{\text{pulse}}\ge 250\mathrm{ns}$, in comparison with continuous wave (cw) ${\mathrm{CO}}_2$ laser radiation, glass materials with several chemical elements and also high thermal expansion coefficients can be processed as well [12]. Hence, this laser can be used for form generating and form correction of optics by ablation of glass materials.

As a result of the processing with high-power $Q$-switched laser radiation, the processing time compared to conventional fabrication methods can be strikingly decreased. Increasing demands on optics with nonspherical shape and high manufacturing times of optics with such surface shape, a high economical potential is identified.

At the Fraunhofer Institute for Laser Technology (ILT), a laser-based process chain for optics manufacturing is being developed. The process chain aims at the economic manufacturing of nonspherical surfaces in small batches, as shown in Fig. 1. By this process chain, advantages in terms of process time, flexibility, and production costs toward the production of a conventional manufactured lens with nonspherical surface are provided [13].

 

Fig. 1. Laser-based process chain for the manufacturing of optics [13].

Download Full Size | PDF

The laser-based process chain consists of three process steps. Starting from a preform, glass material is ablated with pulsed ${\mathrm{CO}}_2$ laser radiation for generating the form of the optics. Subsequently, the resulting roughness of the surface is reduced in a second process step by polishing with defocused cw ${\mathrm{CO}}_2$ laser radiation. Finally, the form of the optics is corrected by selective material ablation (laser beam figuring) with modulated ${\mathrm{CO}}_2$ laser radiation in a third process step. The second and third process steps are iterated with the measuring step until the glass surface has reached the required shape accuracy and roughness. A more detailed description of the process chain is given in [13]. In this paper, results for the first and third steps with pulsed and modulated ${\mathrm{CO}}_2$ laser radiation will be shown.

2. EXPERIMENTAL SETUP

For the experiments, a $Q$-switched ${\mathrm{CO}}_2$ laser source with a maximum output power of $P_{L,\max }\approx 200\mathrm{W}$ built by FEHA LaserTec GmbH is used. The functional principle of the $Q$-switched laser source is described in [11]. An overview of the specifications of the laser source used for the experiments is given in Table 1.

Table 1. Specifications of the High-Power $\mathsf{Q}$-Switched ${\mathsf{CO}}_{\mathsf{2}}$ Laser Source

View Table | View all tables in this article

The main advantage of the $Q$-switched ${\mathrm{CO}}_2$ laser source compared to conventional ${\mathrm{CO}}_2$ laser sources is the short pulse duration of $\ge 250\mathrm{ns}$ in combination with the high average power of $\approx 200\mathrm{W}$ and the resulting high peak power of $\approx 40\mathrm{KW}$.

The experimental setup is shown in Fig. 2. The laser radiation is guided over the workpiece with a 2D galvanometer scanner. The workpiece can be positioned in three dimensions using a manual $x–y–z$ stage. The laser beam is focused using an F-theta lens with a focal length of $f=100\mathrm{mm}$ and $f=200\mathrm{mm}$. An extraction system in combination with a crossjet is used to remove the ablated glass material from the processing area in order to prevent environmental contamination and to ensure a stable process.

 

Fig. 2. Photograph of the experimental setup.

Download Full Size | PDF

With the given laser source, two different pulse forms can be generated. The pulse forms are schematically shown in Fig. 3. Due to an internal acousto-optic modulator (AOM), $Q$-switched pulses can be generated (Fig. 3, Mode 1). The peak pulse laser power can be reduced by an additional external AOM. The external AOM can also be used for pulse picking as well as to modulate rectangular pulses out of cw laser radiation (Fig. 3, Mode 2). The latter pulse form is used for the laser beam figuring.

 

Fig. 3. Schematic drawing of the pulse form generating of $Q$-switched pulses (Mode 1) and modulated pulses (Mode 2).

Download Full Size | PDF

3. PROCESS PARAMETERS

The main process parameters for ablating glass material with pulsed ${\mathrm{CO}}_2$ laser radiation are shown schematically in Fig. 4. During the process, ${\mathrm{CO}}_2$ laser radiation is used to heat up glass material above evaporation temperature and hence locally ablate material from the glass surface. A detailed description of the process parameters and its influence on the ablation result is given in [13].

 

Fig. 4. Scheme of the process parameters and the scanning strategy used for glass ablation.

Download Full Size | PDF

The focused laser beam with an average laser power of $P_{\mathrm{avg}}$, pulse duration $t_{\text{pulse}}$, repetition rate $f_{\text{rep}}$, and focus diameter $d_S$ is moved across the glass surface at the scan speed $v_S$ in a unidirectional scan strategy resulting in a pulse distance $d_x$ [Eq. (1)]. The ablation depth $z_{\mathrm{ab}}$ is determined as the height difference between the initial surface and the ablated surface.

For a homogenously ablated surface, the track pitch $d_y$ as well as the pulse distance $d_x$, calculated with Eq. (1), are set to values smaller than $d_S$:

Table 2. Parameters Used for the Experiments of High-Speed Laser Ablation with a Focus Distance of 100 mm and ${\mathsf{d}}_{\mathsf{s}}\approx \mathsf{300}\mathsf{\mu m}$ ($\mathsf{Q}$-Switched Pulses)

View Table | View all tables in this article

Table 3. Parameters Used for the Experiments of High-Precision Laser Ablation with a Focus Distance of 200 mm and ${\mathsf{d}}_{\mathsf{s}}\approx \mathsf{500}\mathsf{\mu m}$ (Modulated Pulses)

View Table | View all tables in this article

For the experimental implementation, test fields with the dimension of $A=10\mathrm{mm}\times 10\mathrm{mm}$ for high-speed laser ablation and $A=5\mathrm{mm}\times 5\mathrm{mm}$ for laser beam figuring on flat conventional polished fused silica samples are generated. The process time for each layer $t_{\text{Layer}}$ is given by

4. ANALYSIS PROCEDURE

For the measurement of the ablation depth and roughness, a white-light interferometry is used. An exemplary measurement of an ablated $10\mathrm{mm}\times 10\mathrm{mm}$ test area is shown in Fig. 5.

 

Fig. 5. White-light interferometry measurement of an ablated field for measuring ablation depth.

Download Full Size | PDF

For identifying the ablation depth $z_{\mathrm{ab}}$, the height difference of a test field of $2\mathrm{mm}\times 1\mathrm{mm}$ in the middle of the ablated field and a reference field of $1\mathrm{mm}\times 1\mathrm{mm}$ on the initial surface is measured. The small field size reduces the measurement time without influencing the precision of the measurement results.

5. RESULTS

A. High-Speed Laser Ablation

With the presented setup, different modifications of the used parameters are possible. To reduce the time and cost of the test setup, a design of experiment (DoE) approach was applied to investigate the influence of the parameters for the ablation. According to Eq. (3) the pulse energy density can be varied by varying the repetition rates $f_{\text{rep}}$ as well as the average laser powers $P_{\mathrm{avg}}$. Due to the DoE and according to Eqs. (1) and (2), a high repetition rate $f_{\text{rep}}$ at a constant pulse distance $d_x$ allows a high scan speed $v_s$, which influences the ablation rate $\dot{V}$. In Fig. 6, the ablation depth per exposure layer as a function of pulse energy density $H$ for $f_{\text{rep}}=150\mathrm{kHz}$ is shown. The average laser power $P_{\mathrm{avg}}$ is varied from 10 to 140 W.

 

Fig. 6. Ablation depth per exposure layer as a function of pulse energy density for $f_{\text{rep}}=150\mathrm{kHz}$ and $10\mathrm{W}<P_{\mathrm{avg}}<140\mathrm{W}$.

Download Full Size | PDF

Starting from the initial surface, the ablation depth $z_{\mathrm{ab}}$ per exposure layer increases with the pulse energy density $H$. For $H>3\mathrm{kJ}/{\mathrm{m}}^2$, the ablation depth increase is linear. For $H<3\mathrm{kJ}/{\mathrm{m}}^2$, the pulse energy density is not high enough to evaporate the glass material homogenously. By using a linear regression for the values of $H>3\mathrm{kJ}/{\mathrm{m}}^2$, a threshold pulse energy density of $H_{\text{th}}\approx 3.34\mathrm{kJ}/{\mathrm{m}}^2$ for the given parameters can be determined.

In Fig. 7, the influence of the pulse energy density on the ablation rate is shown for different repetition rates of $f_{\text{rep}}=150\mathrm{kHz}$ and $f_{\text{rep}}=20\mathrm{kHz}$. The figure shows that for both repetition rates, the ablation depth $z_{\mathrm{ab}}$ increases linearly with $H$. However, for a repetition rate of $f_{\text{rep}}=150\mathrm{kHz}$, the increase of $z_{\mathrm{ab}}$ is higher compared to $f_{\text{rep}}=20\mathrm{kHz}$.

 

Fig. 7. Ablation depth per exposure layer as a function of pulse energy density for $f_{\text{rep}}=150\mathrm{kHz}$ for $10\mathrm{W}<P_{\mathrm{avg}}<140\mathrm{W}$ and $f_{\text{rep}}=20\mathrm{kHz}$ for $10\mathrm{W}<P_{\mathrm{avg}}<90\mathrm{W}$.

Download Full Size | PDF

To minimize fluctuations at the upper maximum of the laser power, the following investigations for the ablation rate $\dot{V}$ are accomplished with reduced laser power. Therefore, the parameters of $P_{\mathrm{avg}}=115\mathrm{W}$ and $f_{\text{rep}}=150\mathrm{kHz}$ are chosen and kept constant for the following investigations of the influence of track pitch and pulse distance. For this experiment, fields with the dimensions of $10\mathrm{mm}\times 10\mathrm{mm}$ and $n=25$ are ablated. The results of $\dot{V}$ for different track pitches $d_y$ and track distances $d_x$ are shown in Fig. 8. The ablation rate $\dot{V}$ is calculated with Eq. (2).

 

Fig. 8. Contour plot of the ablation rate $\dot{V}$ dependency on the track pitch $d_y$ and track distance $d_x$ for $P_{\mathrm{avg}}=115\mathrm{W}$ and $f_{\text{rep}}=150\mathrm{kHz}$.

Download Full Size | PDF

In Fig. 8, it can be demonstrated that a decreasing of $d_x$ increases $\dot{V}$ while a variation of $d_y$ nearly does not affect the ablation rate. The reason for this behavior is the preheating effect of the previous pulse. With a decreasing of $d_x$, the overlapping of two laser pulses is increased and hence the preheating temperature on the surface at the center of the following laser pulse is increased. Therefore, less energy is necessary for heating up the material from the solid phase to the gas phase. Therefore, more material can be ablated with the same pulse energy.

With the given experimental setup, a maximum ablation rate of ${\dot{V}}_{\max }=2.35{\mathrm{mm}}^3/\mathrm{s}$ can be achieved. Compared to the results given in [14], the track pitch, scan speed, repetition rate, and average laser power has been increased. Due to the high enhancement of the ablation depth per exposure layer, the scan speed and the track pitch the ablation rate [see Eq. (2)] are increased by a factor of 635. The used parameters are given in Table 4.

Table 4. Used Parameters for the Maximum Ablation Rate ${\dot{\mathsf{V}}}_{\mathsf{max}}$

View Table | View all tables in this article

B. Form Generating

Due to a controlled movement of the laser beam and a homogenous ablation depth per layer, complex surface structures can be generated. In this paper, the processing of a hexagonal structure (honeycomb structure) for weight reduction is demonstrated. The computer-aided design (CAD) image is shown in Fig. 9.

 

Fig. 9. CAD image of the honeycomb structure.

Download Full Size | PDF

Based on the CAD image, the laser tool path can be generated with a specially developed program by the Fraunhofer Institute for Laser Technology ILT. Then, the 3D workpiece is sliced into layers with a height of the ablation depth per layer with constant parameters. Due to the selective processing in each layer, a 2.5D form can be generated. The scan and jump vectors of one slice of the workpiece and a magnification is shown in Fig. 10.

 

Fig. 10. Section of the laser tool path with scan vectors (green) and jump vectors (purple) for a honeycomb structure for a layer in the middle of the workpiece and a magnification.

Download Full Size | PDF

A photograph of a honeycomb structure, which is ablated with laser radiation, is shown in Fig. 11. The used parameters for the ablation process are given in Table 4. The maximum measured ablation depth is 1 mm. It is demonstrated that a form generating for weight reduction can be realized.

 

Fig. 11. Photograph of a laser-ablated honeycomb structure on fused silica.

Download Full Size | PDF

The processing time of the workpiece only depends on the ablation rate. To further reduce the processing time, higher ablation rates can be realized by increasing the average power of the $Q$-switched pulses.

C. Laser Beam Figuring

In addition to high-speed laser ablation, a further process is developed to ablate glass material of only a few nanometers but with the same laser source. With laser beam figuring, a laser-based shape correction process of optics by selective ablating glass material can be realized. For a precise ablation process, rectangular-shaped laser pulses are used. By varying the pulse duration, the ablation depth can be controlled. Here, the pulse duration is varied from 20 to 40 μs. The rectangular pulses are generated by an AOM with a ramp time of $<1\mathrm{\mu s}$. An average laser power of 50 W and a laser beam diameter of approximate 500 μm with a Gaussian beam shape is used.

To determine the effect of the pulse duration on the ablation depth, $5\mathrm{mm}\times 5\mathrm{mm}$ fields with constant laser power were processed. The pulse duration varies for each test field. In Fig. 12, four white-light interferometry images are shown for the ablated test fields with a track pitch and a track distance of 20 μm. In the ablated area, the single ablation crater cannot be seen. Hence, a homogenous ablation without increasing the microroughness can be achieved with laser beam figuring.

 

Fig. 12. White-light interferometry images of ablated test fields processed with different pulse durations by laser beam figuring.

Download Full Size | PDF

In Fig. 13, the ablation depth is plotted against the pulse duration. In contrast to the linear dependence of the pulse energy density for the high-speed laser ablation in Fig. 6, here a nonlinear dependence is determined. However, with laser beam figuring, ablation depths down to 3 nm can be achieved.

 

Fig. 13. Dependence of the pulse duration on the depth $z_{\mathrm{ab}}$ with $P_{\mathrm{avg}}=50\mathrm{W}$.

Download Full Size | PDF

For the selective ablation of glass, the pulse duration of each laser pulse is adapted to the necessary ablation depth. Therefore, the characteristic slope shown in Fig. 13 is used to generate a scanning script for the laser processing. Figure 14 shows a demonstration of the laser beam figuring process. Each two adjacent letters are processed with same pulse duration.

 

Fig. 14. False color image taken by a white-light interferometry with varying pulse durations (laser beam figuring).

Download Full Size | PDF

In the laser-based chain for manufacturing optics, the high-precision laser ablation can be used to reduce the waviness and form errors after laser polishing. The results presented above are already sufficient in vertical dimension to reduce the resulting waviness after the second process step. After measuring the surface, the pulse duration of each laser pulse can be set to the depths needed to ablate at the specific position. In further investigations, the high-precision laser ablation has to be applied on laser polished glass surfaces.

6. CONCLUSION AND OUTLOOK

Current results of processing fused silica with pulsed ${\mathrm{CO}}_2$ laser radiation are presented. In comparison to conventional grinding methods, laser ablation offers a small processing time independently from the surface shape to be processed. It is shown that the ablation depth using $Q$-switched ${\mathrm{CO}}_2$ laser pulses increases linearly with an increasing of the pulse energy density for different repetition rates. The results reveals that a higher repetition rate requires less pulse energy density compared to a lower repetition rate to ablate the same amount of material. The reason for this behavior is the preheating effect of the previous pulse. For the form, generating ablation rates up to $\dot{V}\le 2.35{\mathrm{mm}}^3/\mathrm{s}$ can be achieved. An exemplary processing of a honeycomb structure with high-speed laser ablation is also demonstrated in this paper.

For a precise ablation with rectangular ${\mathrm{CO}}_2$ laser pulses, a nonlinear dependence between the ablation depth and the pulse energy density is determined ($z_{\mathrm{ab}}<200\mathrm{nm}$). Homogenous ablation depths down to 3 nm can be achieved. Next steps include the locally selective application of the laser beam figuring process toward laser polished surfaces as well as the combination of all three process steps. A first example for the combination of high-power laser ablation and laser polishing is shown in Fig. 15.

 

Fig. 15. Photograph of locally ablated glass material (left) fused silica with different ablated layers $n$ and different dimensions, with processing time $\approx 3\min$ (right) polished fused silica with laser radiation.

Download Full Size | PDF