Experiment

The experiments were performed with the NHELIX laser system at the GSI near Darmstadt Germany³⁹.

The NHELIX laser system consists of a ns pulsed laser head generating 10 seed pulses per second. These pulses have an approximate duration of 10 ns and a center wavelength at 1064 nm. A pockels cell is used to introduce temporal shaping of the pulse and reduce the seed rate to 5 Hz. Then a mechanical shutter picks one single pulse out of this pulse train for full energy amplification.

The amplification is achieved by five Nd:glas amplifiers with ascending diameters of which the third one is set up to perform a double-pass arrangement. The aperture of the last amplifying rod is 64 mm which limits the diameter of the final laser beam. The amplified laser-pulse of approximately 80 J pulse energy then is directed into the experiment room.

In the experiment room (compare Fig. 5a) the laser beam is redirected to enter a vacuum chamber through an optical quality anti-reflective coated (AR) window from the top. The light within the chamber is, via a highly reflecting (HR) mirror, redirected to a horizontal propagation direction through the center of the chamber. Directly behind the mirror an AR coated three inch diameter lens (Thorlabs LA4246 with C-coating) with 500 mm focal length is used to reduce the beam diameter (64 mm) in the target plane to approximately 3 cm. The beam shape was verified by a burn pattern. The remaining laser beam is dumped on black anodized aluminum foil several ten centimeters after the intermediate focus. Beam polarization at the target was horizontal.

Sketch of the experimental setup. (a) Top view of the vacuum chamber. (b+c) View from color high-speed cameras perspective showing (b) the prepared dropping mechanism and (c) Laser irradiation with indicated falling path from green over orange to blue (along the dotted line). The global coordinate system is indicated by arrows entitled x,y and z. Abbreviations: VAC: vacuum chamber, DM: dropping mechanism, P: periscope, L: Lens, LED: LED-lamp, HAL: halogen lamp, HSM: high-speed camera monochrome, HSC: high-speed camera color, BPC: beam profiling camera, LB: Laser beam, M: mirror, TSC: PTFE screen, HM: holding magnet, T: target, BP: burn pattern foil, PS: pressure spring, DA: dropping arm, HS: holding springs.

To keep track of the beam properties, for every single laser shot the time-dependent signal of the laser pulse was captured with a photodiode (Thorlabs DET01CFC). A beam centering analysis module gave information about the laser pulse energy prior to its amplification. The amplified beam was further analyzed by a PTFE scattering screen which again was exposed to leaking laser light at a folding mirror. The screen was monitored by a CMOS camera (Basler acA2000-50 gm) capturing information about the beam profile and relative intensity (Fig. 4b). In order to generate a reference to these relative intensities, four laser shots were documented dumping all pulse energy onto a calorimeter (Gentec QE95LP-H-MB-QED-D0). The average pulse energy was considered to equal the average pulse energy during the remaining experiments. These two-dimensional profiling measurements were accompanied by burn patterns. Typically these can be achieved by exposing thermal printing paper (Kodak Linagraph direct print type 1895) to the laser. A second type of burn pattern foil, black anodized aluminum foil, was used as a beam dump. After a full energy illumination it documented a shadowgraph of the free falling target at the moment of laser illumination (Fig. 4c).

In order to release the debris targets for free fall, a dropping arm was constructed (Fig. 5b,c) consisting mainly of Thorlabs standard opto-mechanical components. An anodized aluminum breadboard 15 cm × 60 cm in size was fixated within the vacuum chamber. On top of the breadboard, two opposing 1.5 inch diameter metal rods with an absolute height of 56 cm and 51 cm were mounted to hold the dropping mechanism. The lower rod held the dropping arm build with the optical rail system allowing for a flexible adjustment to the desired lengths. The arm was attached via two ball bearings (MDC Vacuum Limited - precision bearings) to grant its free movement away from the target. 10 cm below the arm a pair of tension springs was used to catch the arm at its lowest point and prevent it from being repelled towards the free falling target. The other rod held an electrically demagnetizing magnet and a pressure spring. In its start position the arm was held by the magnet, while bringing pressure onto the pressure spring. As soon as the magnet was demagnetized the pressure spring was able to push the arm downwards and accelerate at a much higher rate than the falling target was accelerated by gravity (compare Fig. 5b,c).

Target movement was captured by a pair of high-speed cameras. Both cameras were placed outside the vacuum chamber recording the free fall at 1000 fps and 0.5 ms exposure time through a glass window each. The color high-speed camera (MotionPro X3 - IDT, Inc.) was placed perpendicular to the laser beam path while the black and white high-speed camera (Redlake MotionScope M-3) was placed in between the incoming laser pulse and the the first camera at a 45° angle. The wide field objectives with f = 1.4 and 12 mm focal length (Thorlabs - MVL12M1) allowed to view the entire free fall path. Heat shielding glasses made from Shott KG5 with a diameter of 50 mm (Edmund Optics - 49-095) were attached to the objectives to shield the cameras from direct laser irradiation. Color sensitivity was not essential for the experiment.

Most targets were made of an aluminum alloy (AL6061) and differently shaped to investigate the influence of the targets shape on its impulse vector. Supplementary Table ^S2 summarizes the targets involved in the experiment. Surfaces were taken as received from the workshop for fine mechanics (compare Fig. 4a). Targets were cleaned using ethanol and handled with disposable gloves.

The tested aluminum plates were weighed before and after ablation. The difference in target mass m_target equals the mass lost to the ablation jet m_jet. To generate a realistic estimate of the uncertainty in weighing, a blind test was performed by dropping three aluminum targets without laser irradiation in the evacuated chamber and measuring the introduced mass difference, which ideally should be expected to be zero. Due to time restrictions the blind test was performed with a delay of one day.

Synchronized triggering of all electrical systems as well as the basic clock signal of the NHELIX laser system was achieved with an Arduino Leonardo based self-designed trigger box. The Arduino Leonardo was extended by a custom shield providing driving ICs to drive several 50 Ohm terminated BNC lines and one 12 V electromagnet. The system was placed within a housing supporting 10 BNC connectors for final wiring. One extra output was foreseen to drive the electrically demagnetizing magnet. An external 12 V power supply was used to provide stand alone functionality. Furthermore, all lines except the electric magnet can be driven by the USB supply line, which was also used to document the trigger procedure via a serial interface protocol. The serial protocol allowed to securely switch between an armed and an unarmed status of the trigger system, without being sensitive to stray induced signal spikes. Once armed, the experimental process can be started by pushing a corresponding button on the outside of the trigger box.

After pushing the trigger button, the box sends a trigger signal to the NHELIX laser system, which starts a countdown to emit the laser pulse. At the same time the box starts counting down to the last 200 ms clock cycle before laser emission. After a delay of approximately 50 ms the electromagnet is demagnetized. As soon as the electrical contact between electromagnet and dropping arm seizes, the high-speed cameras are triggered and the system waits a predefined falling time of approximately 150 ms of free fall. Simultaneously the clock signal of the NHELIX system will be delayed by a few ms to fire the system at the specified time. Shortly before the laser event occurs, the photodiode and the beam profiling camera are triggered to capture the beam properties.

Camera calibration was introduced using Microsoft’s “Easycalib” tool based on the Zhang algorithm⁴⁰ in combination with the preceding ImageJ Plugin “Optic Calib” by Peter Stierlen⁴¹. Tracking of markers was performed manually using ImageJ. Intermediate movements were interpolated and translated to time and position tables. These tables were evaluated by stereoscopic triangulation, determining the midpoint of the closest approach from corresponding projection rays. The algorithms were implemented in a custom Python script, which was followed by a script fitting the parameters of freedom for a rigid object: angular velocity ω→i and velocity v→i before and after the laser irradiation. The used coordinate system is right handed as indicated in Fig. 5.

To give an additional estimate of the effects introduced by vacuum quality the aluminum plate experiments were performed at two different pressure levels. The higher level was chosen slightly below 2·10⁻¹ mbar while the lower level was chosen slightly below 10⁻⁴ mbar.