High-average-power water window soft X-rays from an Ar laser plasma

Soft X-rays (SXRs) in the water window region between the K absorption edges of oxygen at 2.3 nm and of carbon at 4.4 nm are used in high contrast SXR microscope imaging of live biological samples, because of the difference in absorption lengths of water and proteins in biological specimens. At present, however, the only X-ray sources able to continuously generate such water window SXRs are synchrotron radiation (SR) sources. Many studies have therefore been performed with the intent of producing water window SXRs using laser plasma X-ray (LPX) sources, which are both compact and inexpensive compared to conventional SR sources. An LPX output from a high-density, high-temperature plasma, produced by illuminating a target with high-peak-power laser radiation, constitutes an attractive high-brightness point source. To generate water window SXRs, various materials such as copper,¹⁾ ytterbium,²⁾ carbon,³⁾ nitrogen,^4–7) and argon^4,5,8) have been studied as candidate targets for LPX sources. We previously studied an LPX source that can generate continuous repetitive pulses on a long-term basis for industrial use in extreme ultraviolet lithography applications.^9,10) Following that study, we considered development of a continuous LPX source in the water window using a solid Ar plasma target. Among the aforementioned candidate target materials, N and Ar are considered ideal deposition-free targets because they are formed from inert gases instead of metals, and are thus chemically inactive and not deposited on mirrors near the plasma. This reduction in plasma debris is a major advantage for continuous operation. Compared with the line spectrum of a N target, an Ar target has a broader emission spectrum;⁴⁾ Ar is also expected to generate higher power SXRs in the water window than a N target. To provide a continuous Ar supply at the laser's focal point, several promising approaches such as Ar gas puff targets,⁴⁾ Ar liquid jets,⁵⁾ and Ar solid filaments⁸⁾ have been tested previously. While these schemes demonstrated suitable abilities to supply Ar targets continuously, average powers produced by continuously operated LPX sources have not been studied. In this study, we use a cryogenic Ar solid target to provide high conversion efficiency and high brightness because of the target's high solid density. In addition, a smaller gas load for evacuation by an exhaust pump system was expected from solid-state targets compared with gas targets and liquid jets. Peth et al.⁸⁾ reported a supply system using a nozzle to generate solid Ar filaments, but this system has a disadvantage of nozzle erosion in long-term operation, which occurs at short distances between the nozzle and laser focal point because of close proximity to the plasma. We therefore initially developed a translating cryogenic target system to continuously supply the solid Ar targets without the use of a nozzle. Using this system, we successfully produced a continuous supply of Ar solid targets and demonstrated continuous generation of LPX emissions at repetition rates of up to 10 Hz irradiated by a commercial Nd:YAG Q-switched rod laser.¹¹⁾ In this paper, we report plasma SXR emission characteristics in the water window of a solid Ar plasma target supplied by the system previously developed. To maximize the conversion efficiency (CE) from laser energy to SXR energy, laser intensity and energy conditions were optimized. High average SXR power was achieved continuously under these optimized conditions. Moreover, the generated LPX power was compared with that produced by traditional SR-based sources.

Figure 1 shows the side view of the developed LPX source, including the system used for a continuous supply of cryogenic targets. The target system is described in detail in Ref. 11. A Cu substrate was attached to the tip of a cryostat head with a He gas closed loop, enabling the substrate surface to be cooled to a temperature of 15 K. The Ar target gas was blown onto the substrate surface and thus condensed to form a solid Ar layer. The substrate coated with the solid Ar layer was translated up and down with movement only in one dimension, meaning that a fresh target surface was supplied for each laser shot. In this experiment, the Ar gas flow rate and target speed were set to 250 mL/min and 3 or 4 mm/s, respectively. Wipers were used to adjust the Ar solid layer thickness to 500 µm, sufficient to prevent a 1 J laser shot from damaging the substrate surface, and these wipers also assisted in re-covering laser craters on the target. The entire vacuum chamber was evacuated using a 190 l/s turbomolecular pump, and the vacuum was maintained at less than 1 Pa. A conventional Q-switched Nd:YAG rod laser (Spectra-Physics PRO-230) was used as a pump source and could deliver pulses at a wavelength of 1064 nm with a pulse width of 10 ns. The maximum pulse energy and repetition rate were 1.2 J and 10 Hz, respectively. The pulses were expanded using a beam expander before being passed through a window, and were then focused perpendicularly on the target using a lens with a focal length of 500 mm. Finally, plasmas were produced and SXR radiation was emitted.

Zoom In Zoom Out Reset image size

The SXR spectra were measured using a spectrometer with a resolution of 0.2 nm. The spectrometer consisted of a transmission grating (TG) and back-illuminated two-dimensional charge-coupled device. The TG was composed of a 1000 line/mm pattern over a circular aperture with a diameter of 46 µm. The spectrometer was located at an angle of 45° to the laser incident axis. The spectrum measured at a laser intensity of 5 × 10¹² W/cm² is shown in Fig. 2. It can be seen that the Ar spectrum has strong emissions across the water window region (2.3–4.4 nm) and some recognizable characteristic emission peaks that can be attributed to transitions of Ar XIII–Ar IX ions. These transitions were identified using data in Ref. 12 as (1) 2.937 nm (Ar XIII 2s²2p²–2s²2p3d), (2) 3.166 nm (Ar XII 2s²2p³–2s²2p²(¹D)3d), (3) 3.507 nm (Ar XI 2s²2p⁴–2s²2p³(⁴S⁰)3d), (4) 3.84 nm (Ar X 2s²2p⁵–2s²2p⁴(³P)3d), (5) 4.148 nm (Ar IX 2s²2p⁶–2s²2p⁵(²P⁰_1/2)3d), and (6) 4.918 nm (Ar IX 2s²2p⁶–2s²2p⁵(²P⁰_3/2)3s) emissions.

Zoom In Zoom Out Reset image size

To measure the generated SXR energy, an absolutely calibrated X-ray diode (XRD) with 200 nm titanium and 50 nm carbon filters (Opto Diode AXUV100Ti/C2) was used. The XRD has a detection range of 1–12 nm and thus is suited for SXR energy measurement of Ar emission. By measuring the SXRs using the XRD located at 45° to the laser incident axis, we were able to optimize operating conditions in terms of laser intensity and energy to maximize the CE. By changing the position of the focusing lens, laser intensity on the target could be adjusted to determine the optimum intensity. Figure 3(a) shows the CE per unit solid angle as a function of lens position (i.e., laser intensity). The laser pulse energy used was 0.5 J at 1 Hz. The lens position (LP) at the best focus position was denoted by zero; before the best focus (in focus), by a negative number; and beyond the best focus (out of focus), by a positive number. Figure 3(a) shows the maximum CE at the best focus position LP = 0 mm, at which the laser spot diameter was 50 µm. Next, the lens position was fixed at LP = 0 mm and dependence of the CE on laser pulse energy was investigated. Figure 3(b) shows the CE per unit solid angle as a function of laser energy. The result indicates that the CE was maximized at a laser energy of 1 J. Thus, we concluded that the maximum CE per unit solid angle was achieved at LP = 0 mm and at a laser energy of 1 J, and the laser intensity was estimated to be 5 × 10¹² W/cm². This laser intensity is in agreement with other experimental results^2,8) and produces a plasma with an optimized density and temperature for SXR emission.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Spatial distribution of the SXR emission was measured by scanning the XRD on a rotating stage around the plasma. The measurements showed that the distribution can be fitted by (cos θ)^0.41, as shown in Fig. 4. Taking this distribution into account, we obtained a maximum spatially integrated CE of 19% over the full XRD detection range (1–12 nm). We consider this value to be reasonable compared with a maximum CE of 25% in the 5–17 nm range obtained from a solid Xe target in a previous experiment.¹³⁾ With regard to the CE in the water window region, it can be obtained from the ratio of the area in the water window to that in the full detection range of the Ar spectrum shape, as shown in Fig. 2. This ratio was calculated to be 0.76, and we concluded that the maximum spatially integrated CE was 14% in the water window (2.3–4.4 nm) at a laser energy of 1 J. The CE value of 14% is significantly higher than the value of 3.8% achieved using an Ar gas puff target,⁴⁾ and our LPX source thus demonstrates a very high efficiency.

Zoom In Zoom Out Reset image size

We then tested for repetitive generation of plasma SXR pulses by focusing laser pulses with an energy of 1 J at 1 Hz on the Ar target surface. The combination of Ar gas flow rate and target speed enables a re-covered target surface to be supplied for each laser pulse. The continuously generated repetitive SXR pulses were monitored using the XRD, which was positioned at 45° to the laser incident axis. The results are shown in Fig. 5. The output power was stable without any effect arising from translating the target up and down, even for as many as eight round trips. This indicates that a freshly re-covered target surface was supplied for each laser shot as designed. Therefore, our source, which had no nozzle, can be operated continuously on a long-term basis in the same way as our previously developed LPX source.¹⁴⁾ In this experiment, we achieved continuous SXR generation with a high average power of 140 mW (= 1 J × 1 Hz × 14%) in the water window (2.3–4.4 nm). The energy stability was estimated to be 4% (1σ).

Zoom In Zoom Out Reset image size

We compared the obtained LPX characteristics with typical SR characteristics using an in-house synchrotron facility with an electron storage ring called NewSUBARU (storage energy 1–1.5 GeV, stored current 250 mA).¹⁵⁾ Characteristics of the SR produced in the water window using a bending magnet in NewSUBARU are listed in Table I and compared with those of the LPX. It can be seen that the brilliance of the SR is considerably higher than that of the LPX, because the SR intrinsically has very low beam divergence. However, the power (i.e., photon number) of the SR was calculated to be 8 mW; it is evident that the power of the LPX was more than one order of magnitude higher than that of the SR.

We reported the performance of an Ar LPX source, based on a system developed for continuous supply of solid Ar targets and a commercial Nd:YAG Q-switched rod laser. The source demonstrated a high average power of 140 mW and high efficiency of 14% in the water window wavelength region from 2.3 to 4.4 nm. The LPX power was shown to be higher than that of an SR source. The proposed LPX source is expected to offer a powerful alternative to an SR source in numerous applications such as SXR microscopy.

Acknowledgments