Demonstration of Near Edge X-ray Absorption Fine Structure Spectroscopy of Transition Metals Using Xe/He Double Stream Gas Puff Target Soft X-ray Source

A near 1-keV photons from the Xe/He plasma produced by the interaction of laser beam with a double stream gas puff target were employed for studies of L absorption edges of period 4 transitional metals with atomic number Z from 26 to 30. The dual-channel, compact NEXAFS system was employed for the acquisition of the absorption spectra. L1–3 absorption edges of the samples were identified in transmission mode using broadband emission from the Xe/He plasma to show the applicability of such source and measurement system to the NEXAFS studies of the transition metals, including magnetic materials.


Introduction
Material investigations are typically performed employing beamlines at synchrotrons. Such studies are often based on an investigation of subtle changes in the absorption coefficient spectrum in the X-ray and soft X-ray (SXR) spectral ranges near the absorption edge, i.e., using near edge X-ray absorption fine structure spectroscopy (NEXAFS) [1,2]. Measurement of changes in the absorption spectrum near the absorption edge of a given material allows for obtaining information not only about its elemental composition, but also electronic structure, oxidation states, as well as a local bonding environment [3]. So far, synchrotrons make it possible to test most of the materials near the K and L edges [4]. However, one of the biggest disadvantages of this type of measurement is the availability of the synchrotron measurement time, i.e., the need to apply for beamtime without a guarantee of obtaining such a possibility, and the measurement time itself is strictly defined, while the cost of such measurements for i.e., commercial applications, is very high.
Due to the above arguments, several research groups have started work and put significant effort into the development of the laboratory measurement systems for NEXAFS analysis. Such solutions are based on X-ray lamps [5,6] or laser-plasma EUV and SXR radiation sources [7][8][9][10].
Laser plasma sources, depending on the used combination of irradiation parameters (pulse energy, pulse duration, focusing parameters) and target type (solid or gas), generate enough radiation to conduct the NEXAFS measurements. However, such sources are typically operating at the photon energy range, around C-K edge (284 eV), up to O-K Figure 1. Schematic representation (a), modified from [45], and the photograph (b) of the experimental setup for transition metals NEXAFS measurements using near 1 keV from laser-produced Xe plasma, indicating major components. Orange arrows indicate the radiation propagation direction to the SXR spectrometer. The SXR radiation is generated through the interaction of a compact Nd:YAG laser (NL 303 HT, EKSPLA, Vilnius, Lithuania, pulse energy of EL = 0.6 J and 3 ns time duration at a repetition rate of 10-Hz) with a target. The radiation was focused by a 12.7 mm in diameter, 25 mm focal length best-form lens to a spot with ~100 µ m diameter, reaching a power density in the focal point ~2.5 × 10 12 W/cm 2 . The target was produced in a form of Figure 1. Schematic representation (a), modified from [45], and the photograph (b) of the experimental setup for transition metals NEXAFS measurements using near 1 keV from laser-produced Xe plasma, indicating major components. Orange arrows indicate the radiation propagation direction to the SXR spectrometer.  [45], and the photograph (b) of the experimental setup for transition metals NEXAFS measurements using near 1 keV from laser-produced Xe plasma, indicating major components. Orange arrows indicate the radiation propagation direction to the SXR spectrometer. The SXR radiation is generated through the interaction of a compact Nd:YAG laser (NL 303 HT, EKSPLA, Vilnius, Lithuania, pulse energy of EL = 0.6 J and 3 ns time duration at a repetition rate of 10-Hz) with a target. The radiation was focused by a 12.7 mm in diameter, 25 mm focal length best-form lens to a spot with ~100 µ m diameter, reaching a power density in the focal point ~2.5 × 10 12 W/cm 2 . The target was produced in a form of Figure 2. SXR reference spectrum in the photon energy range of 200-1500 eV, obtained for Xe and He gas backing pressures of 10 bar. The Blue region indicated well known "water window" spectral range that can be used for NEXAFS of organic materials (C and O-K edges are indicated). Gray region indicates 680-1100 eV region used for L-edge NEXAFS of transition metals, reported in this study.
The SXR radiation is generated through the interaction of a compact Nd:YAG laser (NL 303 HT, EKSPLA, Vilnius, Lithuania, pulse energy of E L = 0.6 J and 3 ns time duration at a repetition rate of 10-Hz) with a target. The radiation was focused by a 12.7 mm in diameter, 25 mm focal length best-form lens to a spot with~100 µm diameter, reaching a power density in the focal point~2.5 × 10 12 W/cm 2 . The target was produced in a form of double stream Xe/He gas puff (cloud), where Xe and He backing pressures were optimized for maximum photon yield and were equal to 10 bar for each gas. The details about the geometry, timing [46], and synchronization can be found in [45] and are presented schematically in Figure 1a. The laser-target interaction conditions were chosen in such a way as to achieve the maximum photon yield in the SXR spectral region from 680-1100 eV, as can be seen in Figure 2. The yield was measured using the SXR spectrometer (MUT, Warsaw, Poland). The spectrum was obtained by integrating 40 X-ray pulses by the camera chip cooled down to −20 • C. A small circular aperture with~5 mm in diameter, located 50 mm from the plasma was used to limit the reabsorption of the produced SXR radiation in the residual gas. It was achieved by separating the region of relatively high pressure inside the source chamber (P = 10 −2 mbar during the source operation and P = 10 −4 mbar otherwise) from the pressure inside the SXR spectrometer (typically 10 −5 -10 −6 mbar). The gas puff target source provides the SXR energy used for NEXFAS studies with spectral stability of~5% for the optimized source. Previous measurements resulted in estimation of the photon yield in the photon energy range from 440 eV to 1550 eV (selected by using a 3 µm thick Be filter) to be equal to 2.3 × 10 9 photons/pulse/µm 2 of the source surface emitted into a 4π solid angle. The SXR source size was 43.8 × 254.9 µm 2 and an SXR pulse duration~1.3 ns. For this experiment no filter has been used to increase the photon number, preserving, however, the spectral character of the SXR emission for parallel acquisition of the absorption spectrum.
The emission spectra were recorded using a homemade SXR spectrometer, equipped with a grazing-incidence flat-field diffraction grating (Hitachi High Technologies America, Inc., Schaumburg, IL, USA). The spectrometer configuration was previously reported in [47]. The spectrometer is also equipped with a 100 µm wide entrance slit, located~715 mm from the plasma source. For detection and storage of the SXR spectrum a GE 20482048 (greateyes, Berlin, Germany) CCD camera (2 k × 2 k chip, each pixel is 13 × 13 µm 2 in size), as depicted in Figures 1a and 3, was used. double stream Xe/He gas puff (cloud), where Xe and He backing pressures were optimized for maximum photon yield and were equal to 10 bar for each gas. The details about the geometry, timing [46], and synchronization can be found in [45] and are presented schematically in Figure 1a. The laser-target interaction conditions were chosen in such a way as to achieve the maximum photon yield in the SXR spectral region from 680-1100 eV, as can be seen in Figure 2. The yield was measured using the SXR spectrometer (MUT, Warsaw, Poland). The spectrum was obtained by integrating 40 X-ray pulses by the camera chip cooled down to −20 °C. A small circular aperture with ~5 mm in diameter, located 50 mm from the plasma was used to limit the reabsorption of the produced SXR radiation in the residual gas. It was achieved by separating the region of relatively high pressure inside the source chamber (P = 10 −2 mbar during the source operation and P = 10 −4 mbar otherwise) from the pressure inside the SXR spectrometer (typically 10 −5 -10 −6 mbar). The gas puff target source provides the SXR energy used for NEXFAS studies with spectral stability of ~5% for the optimized source. Previous measurements resulted in estimation of the photon yield in the photon energy range from 440 eV to 1550 eV (selected by using a 3 µ m thick Be filter) to be equal to 2.3 × 10 9 photons/pulse/µ m 2 of the source surface emitted into a 4π solid angle. The SXR source size was 43.8 × 254.9 µ m 2 and an SXR pulse duration ~1.3 ns. For this experiment no filter has been used to increase the photon number, preserving, however, the spectral character of the SXR emission for parallel acquisition of the absorption spectrum. The emission spectra were recorded using a homemade SXR spectrometer, equipped with a grazing-incidence flat-field diffraction grating (Hitachi High Technologies America, Inc., Schaumburg, IL, USA). The spectrometer configuration was previously reported in [47]. The spectrometer is also equipped with a 100 µ m wide entrance slit, located ~715 mm from the plasma source. For detection and storage of the SXR spectrum a GE 20482048 (greateyes, Berlin , Germany) CCD camera (2 k × 2 k chip, each pixel is 13 × 13 µ m 2 in size), as depicted in Figures 1a and 3, was used. Figure 3. Scheme of the dual-channel SXR NEXAFS approach, indicating sample and reference SXR beams forming two independent spectra on the CCD camera. The spectra are used for the calculation of the OD spectrum.
The SXR radiation from Xe/He plasma illuminates the thin sample located in a small load-lock downstream the beamline, as can be seen in Figure 3, which depicts the optical arrangement used to obtain simultaneously the sample and the reference beams. The sample holder allows illumination of the sample by the SXR light from the source. The SXR light is then transmitted through the sample forming a sample beam. Due to its design, the holder allows also some portion of the SXR wavefront, called here a reference beam, Figure 3. Scheme of the dual-channel SXR NEXAFS approach, indicating sample and reference SXR beams forming two independent spectra on the CCD camera. The spectra are used for the calculation of the OD spectrum.
The SXR radiation from Xe/He plasma illuminates the thin sample located in a small load-lock downstream the beamline, as can be seen in Figure 3, which depicts the optical arrangement used to obtain simultaneously the sample and the reference beams. The sample holder allows illumination of the sample by the SXR light from the source. The SXR light is then transmitted through the sample forming a sample beam. Due to its design, the holder allows also some portion of the SXR wavefront, called here a reference beam, to directly enter, undisturbed, the spectrometer's entrance slit. Thus, using this approach, contrary to other compact systems, i.e., [48], simultaneous acquisition of the two spectra has been achieved. We called this approach a dual-channel SXR spectrometer. Such a solution allows for the spectral acquisition and NEXAFS measurements to remain unaffected by the mechanical instabilities or the energy fluctuations of the source. These in turn can lead to unpredicted spectral shifts as well as serious difficulties in calculating the optical density in the presence of spatially or spectrally uncorrelated sample and reference signals.
The samples were prepared by the vacuum deposition method from metal pellets (ONYXMET, Olsztyn Poland) using e-beam evaporation (Syrus III 1100, Bühler Leybold Optics, Alzenau, Germany). The base pressure of the system was 3.5 × 10 −6 mbar. The zinc sample was also prepared by e-beam evaporation; however, a different system was used (PVD 75, Kurt J. Lesker, St. Leonards-on-Sea, UK). The base pressure of the system was 2.7 × 10 −6 mbar. The thicknesses of all layers were measured with quartz microbalance. Commercially available SiN membranes (Silson Ltd., Southam, UK), with a thickness of 100 nm and window size of 1.5 × 1.5 mm 2 in a 5 × 5 mm 2 frame were used as a target for deposition of various transition metals from period 4 of the periodic table of elements with thicknesses ranging from~100 nm to~220 nm. For this work, we selected iron (Fe), Z = 26, cobalt (Co), Z = 27, nickel (Ni), Z = 28, copper (Cu), Z = 29, and zinc (Zn), Z = 30. These metals, due to their electronic configurations, exhibit L absorption edges in the spectral region of interest from 680 eV to 1100 eV [49]. The characteristic parameters of the samples are listed in Tables 1 and 2 A typical procedure for the SXR spectrometer calibration was performed using the Ar:N 2 :O 2 (1:1:1 by volume) gas mixture and SF 6 gas injected into the laser-target interaction region. The positions of some most prominent emission lines spanning the spectral region of interest were found, namely, the fluorine line at λ = 1.6807 nm from F 8+ ions, oxygen line λ = 2.1602 nm from O 6+ ions, two SXR nitrogen lines: λ = 2.489 nm from N 6+ (hydrogen-like) ions and λ = 2.878 nm from N 5+ (helium-like) ions, and argon lines λ = 4.873 and λ = 4.918 nm from Ar 8+ (neon-like) ions were used to construct the calibration curve. Additionally, the resolving power of the SXR spectrometer, in the spectral region of interest for this work, at the photon energy of 738 eV (isolated F line at λ = 1.6807 nm from F 8+ ions with a measured spectral width of 3.7 eV) was equal to E/∆E~200.

Experimental Results
The reference spectrum emitted from Xe/He plasma for Xe and He gas backing pressures of 10 bar is presented in Figure 2. The blue region indicates the well-known "water window" spectral range that is typically used for SXR microscopy applications [50][51][52], but also can be used for NEXAFS studies of organic materials [10,53] (C-K and O-K edges are indicated at 284 and 540 eV respectively, N-K edge is in the middle of that range at~410 eV). In this case, however, we are more interested in the higher energy range of the Xe emission. The gray region in Figure 2 indicates the main energy range of interest, 680-1100 eV, that is used in this work for L-edge NEXAFS of transition metals. The spectral range of our interest is dominated by one, most intense, high energy band from 850 eV to 950 eV, with the peak at~900 eV, [45]. This emission originates from 3p 5 3d 10 4s, or 3p 6 3d 9 4f down to 3p 6 3d 10 level in XeXXVII ions [54].

Thickness Influence on NEXAFS Measurements
The relation between the optical density spectrum OD(E) and the attenuation coefficient spectrum Att(E) is presented as Equation (1). The spectral attenuation coefficient was obtained from the optical density spectrum, taking into account the thickness of sample d: where S sam (E) is the sample spectrum and S ref (E) is the reference spectrum, as was explained using Figure 3. The attenuation spectrum and the optical density spectrum were achieved by slightly smoothing the raw data, obtained using Equation (1), using the Golay-Savitzky filter (order = 3 and frame length = 11, 3 × 11 filter) in a typical post-processing approach, described in [7]. A typical optical density spectrum (a) and attenuation coefficient spectrum (b) is presented in Figure 4 for various thicknesses of the sample. 40 SXR pulses were used to acquire a one CCD image with an exposure time of 5 s. A total of 16 images were added together to improve the signal-to-noise ratio by a factor of 4, except for a case of Cu L 2,3 edges, where 25 images were integrated. Two regions, 200 pixels in width, in the CCD image, corresponding to the sample spectrum region and reference spectrum region were further integrated. This allowed us to extract the sample spectrum S sam (E) and the reference spectrum S ref (E) and use them to compute OD(E) and Att(E).
where Ssam(E) is the sample spectrum and Sref(E) is the reference spectrum, as was explained using Figure 3. The attenuation spectrum and the optical density spectrum were achieved by slightly smoothing the raw data, obtained using Equation (1), using the Golay-Savitzky filter (order = 3 and frame length = 11, 3 × 11 filter) in a typical post-processing approach, described in [7]. A typical optical density spectrum (a) and attenuation coefficient spectrum (b) is presented in Figure 4 for various thicknesses of the sample. 40 SXR pulses were used to acquire a one CCD image with an exposure time of 5 seconds. A total of 16 images were added together to improve the signal-to-noise ratio by a factor of 4, except for a case of Cu L2,3 edges, where 25 images were integrated. Two regions, 200 pixels in width, in the CCD image, corresponding to the sample spectrum region and reference spectrum region were further integrated. This allowed us to extract the sample spectrum Ssam(E) and the reference spectrum Sref(E) and use them to compute OD(E) and Att (E). To demonstrate the spectral changes in the OD(E) and Att(E), a Ni sample was chosen with Ni layers of thicknesses ranging from 91 nm to 234 nm, as depicted in Figure 4. The sample thickness was measured using an atomic force microscope (AFM) (Veeco Nano-Scope IV MultiMode AFM, Veeco Metrology, Plainview, NY, USA), based on 10 crosssection measurements. The representative sample morphology is depicted in Figure 5 for the 138 nm thick Ni sample. The AFM scan was performed with a 20 × 20 µ m 2 scan size. As can be noted, changing the sample thickness preserves all the main features of the attenuation spectrum, as well as the positions of the L3 and L2 edges. To demonstrate the spectral changes in the OD(E) and Att(E), a Ni sample was chosen with Ni layers of thicknesses ranging from 91 nm to 234 nm, as depicted in Figure 4. The sample thickness was measured using an atomic force microscope (AFM) (Veeco NanoScope IV MultiMode AFM, Veeco Metrology, Plainview, NY, USA), based on 10 crosssection measurements. The representative sample morphology is depicted in Figure 5 for the 138 nm thick Ni sample. The AFM scan was performed with a 20 × 20 µm 2 scan size. As can be noted, changing the sample thickness preserves all the main features of the attenuation spectrum, as well as the positions of the L 3 and L 2 edges.

NEXAFS L-Edge Spectra of Z = 26-30 Transition Metals
As a demonstration in the course of this work, the transition metals from period 4 of the periodic table of elements were studied with atomic number Z between 26 and 30. The list of elements studied as well as the sample thicknesses are listed in Table 1. There are several commonly accepted methods for determining the position of the absorption edge, depending on the quality of the data obtained. One of them is the determination of the edge position relative to the 50% slope, and this is what was used in this study, similarly to, e.g., in [38,55].
In the case of iron (Fe, Z = 26) all three L absorption edges were identified, as depicted in Figure 6a

NEXAFS L-Edge Spectra of Z = 26-30 Transition Metals
As a demonstration in the course of this work, the transition metals from period 4 of the periodic table of elements were studied with atomic number Z between 26 and 30. The list of elements studied as well as the sample thicknesses are listed in Table 1. There are several commonly accepted methods for determining the position of the absorption edge, depending on the quality of the data obtained. One of them is the determination of the edge position relative to the 50% slope, and this is what was used in this study, similarly to, e.g., in [38,55].
In the case of iron (Fe, Z = 26) all three L absorption edges were identified, as depicted in Figure 6a In the case of cobalt (Co, Z = 27), we could not identify the L1 absorption edge, due to too low absorption in the absorbing Co layer. For this reason, only the energies of the experimental L2,3 edges were found, Figure 6c Table 1.
Possible differences in the energies arise due to the limited spectral resolution of our spectrometer, and hence an additional equipment broadening might also influence the energies slightly (of the order of~0.1 eV to~1 eV at a photon energy of~1 keV, corresponding to 0.01% to 0.1% energy shift for most of the samples). Our sample is not ideal bulk material, but a thin layer deposited on the top of the supporting SiN membrane [23] could be another source of deviation. Importantly, we compared our experimental results so far with only one database [49]. In the literature, however, there are other sources of data reporting the L edges at slightly different energies, e.g., Kaye & Labe [56] or NIST database [57]. In the case of copper, for which we have the largest ∆E, the L 1 edge is at 1103.12 eV, L 2 at 959.58 eV, and L 3 at 939.85 eV [58], resulting in better matching to our experimental data regarding the L 3 edge (∆E = 1.7 eV, or 0.15%), however, increasing the discrepancy for L 2 edge up to ∆E = 9.9 eV (0.89%). Additionally, the Laser-Induced Breakdown Spectroscopy (LIBS) was used to detect the elemental composition of the samples. For that purpose witness samples were used because samples on the SiN membrane were too fragile and easily destroyed during the measurement. The witness samples were obtained by depositing the metal layers on top of 10 × 10 mm 2 Si wafers in the same deposition process (same conditions) as the NEXAFS samples (on SiN membrane support) 5 cm away from each other and at an equal distance from the electron beam spot. In LIBS measurements a 20 Hz 1064 nm laser source (Quantel, Bozeman, MT, USA, BigSky Brio model) with 50 mJ/4 ns pulses was used. The 4 mm in diameter laser beam was directed through a 150 mm focusing lens onto the investigated samples. The lens-to-sample distance was fixed at 145 mm to avoid an accidental plasma formation in the air (laser spark). The energy of a single laser pulse was sufficient to acquire a useful spectrum. Plasma radiation was registered by ESA 4000 Echelle spectrometer equipped with an ICCD camera with Kodak KAF 1001 detector (spectral resolution of the system was about λ/∆λ~20,000). This experimental system has been used before in different experiments with solid or gaseous targets [59][60][61][62].
To capture useful line radiation, data collection lasted 500 ns and began 500 ns after the laser pulse, allowing the initial continuum emission from the hot plasma to weaken significantly. LIBS measurements showed only the main sample material, due to the high purity of metals deposited on the SiN membrane, without any detectable impurities, and did not provide quantitative results. For this reason, a more quantitative analysis was performed by Energy Dispersive Spectroscopy (EDS). The EDS analysis using a Quanta 250 FEG SEM (FEI, Hillsboro, OR, USA) was also performed on witness samples in three different locations. The acquisition parameters of the EDS are an accelerating voltage of 30 kV, spot 4.5 corresponding to the spot diameter on the sample equal approximately to 2 µm. Table 2 shows the results of EDS analysis on all samples. All samples investigated using the EDS show a strong signal from the interaction volume originating from the supporting Si membrane, as well as a weaker signal from the investigated target material ranging from 7 to~20% (by weight). Moreover, no additional elements (admixtures or infusions) were detected by the LIBS and EDS. In the case of copper (~163 nm thick, comparable to cobalt sample) all L edges were detected at the amount of~11% of copper in the investigated by the EDS volume using our compact NEXAFS setup. Copper L 1 edge occurs also at the highest recorded in this experiment energy at~1.1 keV, demonstrating the applicability of gas-based laser plasma source to such measurements above 1 keV photon energy. Copper is also second in thickness among the samples, second only to zinc sample, providing good absorption for SXR radiation. Even better results were obtained for the iron sample (~142 nm thick), in which only~7.5% of iron was measured in the EDS interaction volume while providing data for all three L edges. Although there was more cobalt in the cobalt sample (~160 nm thick), on average 14.2% by weight, the L 1 edge was not detected, probably due to the insufficient signal-to-noise ratio of the absorption spectral data obtained using our system, caused mainly by limited spectral resolution of the spectrometer. A similar case is for the nickel at~9.2% for 140 nm thick sample and zinc at almost 19.5% for quite thick 300 nm sample for which only L 2,3 edges were found.
The L 2,3 edges for all samples (single layers and multilayers) were detected; however, some L 1 edges (for cobalt, nickel, and zinc) were not visible. We believe it is probably due to insufficient signal-to-noise ratio of the absorption spectral data, caused by the fact that the L 2,3 edges are typically 10-100 times more pronounced (higher value of the ratio of the OD(E) below and above the absorption edge) than the L 1 edges [3].

NEXAFS Spectrum of a Multilayer Sample
To demonstrate the possibility to study multi-component materials a multilayer sample composed of Fe, Ni, and Co was also studied. To prepare the sample Co, Ni, and Fe were subsequently deposited on top of a 100 nm thick SiN membrane. Theoretically, each layer should have a thickness of 70 nm; however, the total thickness of the three layers amounted to 216 nm, slightly higher than expected, with a standard deviation of 3.5 nm. The small difference in the thickness was caused by the deposition process. The NEXAFS spectrum of such a multilayer sample is presented in Figure 7.
To demonstrate the possibility to study multi-component materials a multilayer sample composed of Fe, Ni, and Co was also studied. To prepare the sample Co, Ni, and Fe were subsequently deposited on top of a 100 nm thick SiN membrane. Theoretically, each layer should have a thickness of 70 nm; however, the total thickness of the three layers amounted to 216 nm, slightly higher than expected, with a standard deviation of 3.5 nm. The small difference in the thickness was caused by the deposition process. The NEXAFS spectrum of such a multilayer sample is presented in Figure 7. The ΔE2 values are generally small (<0.5 eV), except for the L3 absorption edge of iron, where it exceeds 3 eV. Table 3 summarizes the data for the Co-Ni-Fe sample. The ∆E 2 values are generally small (<0.5 eV), except for the L 3 absorption edge of iron, where it exceeds 3 eV. Table 3 summarizes the data for the Co-Ni-Fe sample.

Conclusions
In conclusion, we have shown the possibility to study the L absorption edges of transition metals from the 4th period of the periodic table of elements with Z numbers ranging from 26 to 30, including the most important elements exhibiting magnetic properties.
The proof experiment employed the SXR emission from a laser-plasma source covering a spectral region near 1-keV. The laser plasma source was based on a double stream Xe/He gas puff target. We have shown that using a dual channel NEXAFS measurement system, recording independently the sample and reference spectra, it was possible to acquire good quality spectra from single and multilayered samples in the transmission mode reaching Cu-L 1 edge around 1.1 keV. Compact NEXAFS was demonstrated before reaching photon energies above 1 keV, e.g., [7]; however, it was performed with solid-state targets, e.g., copper, producing debris from the laser ablation process. The debris prohibits measurements of delicate samples and, with long exposures, may cover the measured samples with an additional metal layer. This layer is originating from target debris, affecting in turn, the quality of the measured spectra. Additionally, the spectrometer used for obtaining NEXAFS data employed two separate (and not identical) dispersion elements-complicated, narrow energy range off-axis zone plates, also affecting the data quality. The NEXAFS data we have shown were not demonstrated so far using a compact laser-plasma system based on a gaseous target, and the dual-channel spectrometer setup we have developed was also employed in this energy range for the first time. The presented results demonstrated that all essential features of the collected absorption spectra remain the same for various sample thicknesses (the nickel data) and the system is stable providing proper calibration in the spectral domain (stable positions of the absorption peaks). The demonstrated absorption spectra near the L 1-3 absorption edges corresponded to transition metals, such as iron, cobalt, nickel, copper, and zinc. The L 2,3 edges were detected for all samples (single and multilayers). However, some L 1 edges were not detected in the experiment, in particular the L 1 edge of cobalt, nickel, and zinc, probably due to the insufficient signal-to-noise ratio of the absorption spectral data. The reason for this seems to be the fact that L 2,3 edges are typically significantly stronger than the L 1 edges [3]. This work proves the possibility to apply a compact gas-based laser plasma source to the NEXAFS technique in a much broader than usual spectral range. As a consequence, the results demonstrate the applicability of the presented technique to spectroscopic studies not only organic materials but also metals, including, important from the industrial point of view, magnetic materials (Co, Ni, Fe), typically studied using synchrotron sources.