We optimized the parameters of a laser-produced plasma source based on a solid-state Nd: YAG laser (λ = 1.06 nm, pulse duration 4 ns, energy per pulse up to 500 mJ, repetition rate 10 Hz, lens focus distance 45 mm, maximum power density of laser radiation in focus 9 × 1011 W/cm2) and a double-stream Xe/He gas jet to obtain a maximum of radiation intensity around 11 nm wavelength. It was shown that the key factor determining the ionization composition of the plasma is the jet density. With the decreased density, the ionization composition shifts toward a smaller degree of ionization, which leads to an increase in emission peak intensity around 11 nm. We attribute the dominant spectral feature centred near 11 nm originating from an unidentified 4d-4f transition array in Xe+10…+13 ions. The exact position of the peak and the bandwidth of the emission line were determined. We measured the dependence of the conversion efficiency of laser energy into an EUV in-band energy with a peak at 10.82 nm from the xenon pressure and the distance between the nozzle and the laser focus. The maximum conversion efficiency (CE) into the spectral band of 10–12 nm measured at a distance between the nozzle and the laser beam focus of 0.5 mm was CE = 4.25 ± 0.30%. The conversion efficiencies of the source in-bands of 5 and 12 mirror systems at two wavelengths of 10.8 and 11.2 nm have been evaluated; these efficiencies may be interesting for beyond extreme ultraviolet lithography.

Projection extreme ultraviolet (EUV) lithography at 13.5 nm is becoming increasingly widespread in producing critical layers in chips.1 However, a technology of double patterning is needed since the existing numerical apertures NA = 0.32 for projection objective in scheme are already insufficient to achieve a sub-10 nm resolution. The increase in the numerical aperture is fraught with problems, including those associated with photomask (reticle) shading.2 In particular, when designing an objective with a numerical aperture NA = 0.5, the authors had to resort to a non-standard solution—the development of an objective with a different magnification in perpendicular planes.3 

An obvious solution to the problem is to reduce the wavelength. In articles,4,5 a wavelength of about 7 nm was considered promising. The article6 gives an overview of the actual results in developing effective radiation sources, multilayer mirrors, and EUV resists achieved at 6.7 nm wavelength. This article particularly demonstrated that the productivity of the lithographic process at 6.7 nm is more than an order of magnitude lower than at 13.5 nm. It was also noted that productivity comparable to 13.5-nm lithography can be expected at 11 nm wavelength, with the resolution improving by about 20%. Two different wavelengths were discussed. The first wavelength is 10.8 nm, which corresponds to the peak position of the emission band excited from the Xe+ 10… + 13 ions,7,8 and at which the theoretical reflectivity for Rh/Sr multilayer mirrors is R = 73.9%. The second wavelength is 11.2 nm. Although it is on the decline of the emission band, the theoretical maximum of the reflectivity for Ru/Be multilayer mirrors is R = 78.2%. Considering the large number of reflecting surfaces in the optical system of the lithograph, the lower intensity in emission spectrum at 11.2 nm can be compensated by higher reflectivity of mirrors. In the article it was noted that to determine the prospects of these wavelengths for beyond extreme ultraviolet lithography, research is needed both in the field of developing effective multilayer mirrors and in optimizing the xenon source. Apparently, because of the relatively small change in wavelength, this proposal did not find any response from the manufacturers of lithographic equipment and research on this topic was limited to the work of several research groups in the field of multilayer X-ray mirrors.9,10 However, in connection with recent works on maskless EUV lithography11–15 which provides low cost and availability of equipment for relatively small companies and research organizations that do not require high productivity of lithographic equipment, interest in the spectral region of 11 nm has again revived.

The aim of this paper is to measure the conversion efficiency of the laser energy into radiation within a spectral band with a maximum at 10.8 nm wavelength. Despite the fact that several groups have studied the emission characteristics of xenon plasma in the extreme ultraviolet range in some detail, they were mainly devoted to the problem of optimizing the emission spectra at a wavelength of 13.5 nm.7,8,16–20

There is a research21 in which the measured coefficient of conversion of laser radiation energy into a line with a maximum of 10.8 nm for a cryogenic xenon target is given. It amounted to 0.8%/sr (λ = 10.8 ± 0.27 nm), which, in the isotropic distribution approximation, gives the conversion efficiency into a half-space: CE ≈ 5.0%/(2π sr). However, the value of the conversion coefficient for pulsed Xe gas jet is not known exactly. Nevertheless, based on the spectra and values of the conversion efficiency given in the literature at 13.5 nm, one can expect high values (several percentages) at 10.8 nm. The subject of the analysis is two schemes of a lithographic setup: a 12-mirror classical projection lithography scheme and a 5-mirror scheme for maskless lithography.14 

To achieve this goal the following tasks should be solved. The first is to optimize the parameters for a double-stream Xe/He gas jet to obtain a maximum emission around 11 nm wavelength. The second is to carry out precise measurements of the peak position and the width of the emission line. The third is to determine the spectral transmission bands of both optical schemes at 10.8 and 11.2 nm wavelengths for the most promising multilayer mirror coatings. The last task is to measure the conversion efficiency (within 2π sr) of the laser energy into EUV energy of the emission line, and in the transmission band of optical systems at selected wavelengths.

The investigation of emission properties of a double-stream Xe/He gas jet was carried out on an experimental arrangement that is schematically detailed in Fig. 1. The laser radiation from a solid-state Nd:YAG laser (EKSPLA NL 300, wavelength 1.06 μm, pulse energy up to 500 mJ, pulse duration 4 ns, pulse repetition rate 10 Hz, lens focus distance 45 mm, maximum power density of laser radiation in focus 9 × 1011 W/cm2) was introduced into the vacuum chamber using a prism P and a lens L and focused in the vicinity of the axis of the pulsed gas target. The average power of the laser radiation was measured by the aid of a beam splitter with a known reflection coefficient and a calorimeter.

FIG. 1.

Scheme of the experiment. Nd:YAG laser, BS: beam splitter, C:calorimeter, P: prism, L: short-focus lens, N: coaxial nozzles, V1 and V2: pulse controllers, S1 and S2: input and output slits of the spectrometer, M1 and M2: collimating spherical mirrors, DG: planar diffraction grating, D1: EUV detector, F1 and F2: filters, and D2: calibrated semiconductor detector, TMP 1-TMP 3: turbo molecular pumps, G1-G3: vacuum gauges.

FIG. 1.

Scheme of the experiment. Nd:YAG laser, BS: beam splitter, C:calorimeter, P: prism, L: short-focus lens, N: coaxial nozzles, V1 and V2: pulse controllers, S1 and S2: input and output slits of the spectrometer, M1 and M2: collimating spherical mirrors, DG: planar diffraction grating, D1: EUV detector, F1 and F2: filters, and D2: calibrated semiconductor detector, TMP 1-TMP 3: turbo molecular pumps, G1-G3: vacuum gauges.

Close modal

A gas-jet target was formed by a pulsed nozzle. The nozzle consisted of internal and external tubes with a diameter of din = 0.4 mm and dext = 1.5 mm and a length of 10 mm, that were assembled coaxially. The gas supply to the nozzle was controlled by high-speed pulse valves V1 and V2, which were synchronized with the laser. The valves were opened for 250 μs, which is much longer than the formation time of steady-state flow. The delay between the valve opening time and the laser pulse was 1240 μs. We did not measure the concentration of atoms in the interaction zone with the laser radiation directly, but since we use the same valve and its parameters as in, Refs. 8 and 22 then, up to an order of magnitude, the density was about 1018 atoms/cm3.

The maximum Xe gas pressure was 4 ATA (technical atmosphere). The high concentration of particles in the emission zone allowed us to assume that a significant portion of the incident laser radiation was absorbed by the jet. Nevertheless, this prediction requires additional studies.

Spectral characteristics of the EUV light source were measured using a grazing incidence Czerny-Turner spectrometer with a plane diffraction grating and collimating spherical mirrors M1 and M2.23 A holographic diffraction grating with 900 lines per millimeter was used. The operating spectral range was 4–14 nm with a spectral resolution of 0.095 nm. The EUV radiation intensity was measured with a chevron type microchannel plate detector (MCP) operating in linear detection mode. The spectrometer was pumped out by three turbo pumps (TMP) with a performance of 1000 l/s each. During the experiment, the pressure in the detector chamber was 2 × 10-5 torr (gauge G1), about 4 × 10-4 torr in the spectrometer (gauge G2), and about 1 × 10-3 torr around power meter (gauge G3). The pressure in the source chamber was not measured directly. A rough estimate of the pressure in the source chamber of 2–3 × 10-3 torr was made on the basis of the known conductivity of the vacuum tube between chamber and TMP 3 (1000 l/s).

The conversion efficiency of the laser radiation into EUV in the vicinity of 11 nm wavelength was studied with the help of a calibrated power meter. The main elements of the power meter are a set of thin film filters (F1, F2) and a double mirror monochromator (not shown) with known (±4% accuracy) spectral characteristics and a silicon photodiode (D2) designed for the detection of ultraviolet and EUV radiations. The sensitivity of the photodiode was calibrated throughout the spectral range on the BESSY-II synchrotron.24 The spectral characteristics of the filters were measured using characteristic emission lines from a dismountable X-ray tube with the help of a reflectometer.25 In the experiment, we used multilayer Mo/ZrSi2 (72 nm thick) and monolayer Be (288 nm thick) thin film filters, and a double mirror monochromator with Mo/Si multilayer mirrors with maximum peak reflectivity at 11.2 nm wavelength. The filters were mounted on a special holder and could be exchanged automatically without opening the device. The principle of operation and the features of the power meter are described elsewhere.26 

Typical spectral distributions of the pulsed gas target emission in the dependence of Xe pressure are shown in Fig. 2. The case of pure xenon is displayed in Fig. 2a; the double-stream Xe/He gas jet at a fixed He pressure of 1 ATA in Fig. 2b. The distance between the end of the nozzle and the laser spot was 0.7 mm. As can be seen from the figures, a peak at 10.8 nm wavelength is observed at low xenon pressures, reaches a maximum approximately at 1 ATA and decreases sharply with increasing pressure. In the latter case, a strong growth of the emission in the wavelength band 4–10 nm is observed. When He is added, the peak at 10.8 nm is faint but discernible even at low pressures of Xe.

FIG. 2.

Spectral distribution of EUV emission of Xe plasma (a) and of Xe/He plasma at a He pressure of 1 ATA (b). The distance between the end of the nozzle and the laser spot was 0.7 mm. Laser parameters: wavelength 1.06 μm, pulse energy 470 mJ, pulse duration 4 ns, frequency 10 Hz.

FIG. 2.

Spectral distribution of EUV emission of Xe plasma (a) and of Xe/He plasma at a He pressure of 1 ATA (b). The distance between the end of the nozzle and the laser spot was 0.7 mm. Laser parameters: wavelength 1.06 μm, pulse energy 470 mJ, pulse duration 4 ns, frequency 10 Hz.

Close modal

The influence of the distance between the end of the nozzle and the laser spot on the emission spectra is illustrated in Fig. 3, which shows spectra for pure Xe and with the addition of He (PHe = 1 ATA). As was expected, the signals falls rapidly with increasing distance. However, in the presence of He, the signal at 10.8 nm has slightly grown (Fig. 3b).

FIG. 3.

Spectra for pure xenon-a) and with a He pressure of 1ATA -b) for different distances between the end of the nozzle and the laser spot.

FIG. 3.

Spectra for pure xenon-a) and with a He pressure of 1ATA -b) for different distances between the end of the nozzle and the laser spot.

Close modal

The emission spectra as a function of the energy of the laser pulse are shown at Fig. 4. The distance between the nozzle and the laser spot in both cases is 0.7 mm.

FIG. 4.

Emission spectra as a function of the energy of the laser pulse with the target parameters: PXe = 2 ATA (a) and PXe = 2 ATA, PHe = 1 ATA (b).

FIG. 4.

Emission spectra as a function of the energy of the laser pulse with the target parameters: PXe = 2 ATA (a) and PXe = 2 ATA, PHe = 1 ATA (b).

Close modal

The following conclusions can be drawn from the experimental data so far.

First, in studying the emission spectra of a laser-plasma EUV source with a pulsed gas target, two main features in the spectra can be distinguished: a wide band in the wavelength range of 4–10 nm and a relatively narrow peak with a maximum at 10.8 nm. These spectral features are formed by ions of different degrees of ionization. According to the literature data,20,27–29 a wide wavelength band is formed by the emission of the Xe + 14 ... + 16 ions, while at 10.8 nm by the emission of the Xe + 10 ... + 13 ions.7,8

Secondly, the ionization composition is determined mainly by the density of the gas jet rather than by the power density of the laser radiation in the interaction region. This conclusion has been confirmed by all data. In particular, at a low Xe pressure, a peak at 10.8 nm, which is produced by ions of lower degrees of ionization, is dominant. With an increase in the gas density (with an increase in the Xe pressure at the valve inlet, or when approaching the nozzle, or when adding the He jet thereby preventing the expansion of the Xe jet), an emission is observed at the 4–10 nm wavelength range. Using a He jet is especially effective at increasing the emission at this wavelength range. Conversely, increasing the distance from the nozzle is accompanied by a decline of the density of the gas, a strong drop in the emission at the 4–10 nm wavelength range, and even a slight increase in the absolute value of the emission peak at 10.8 nm (Fig. 3). This statement is confirmed by Fig. 4, from which it can be seen that for both modes the shape (ionization composition) of the spectrum does not change strongly with decreasing energy in the pulse from 0.47 J to 0.18 J.

Thirdly, it was shown in experiments that a double-stream target is more effective for generating radiation in a spectral range of 4–10 nm. For λ = 10.8 nm, the emission efficiency is significantly higher when using a relatively low-dense Xe target.

Separately, the effect of self-absorption of the EUV radiation in its propagation within plasma and the surrounding gas on the shape of the spectra has not been investigated. However, several factors indicate that at the Xe pressures of interest to us, when the emission maximum is observed in the 11 nm region, the self-absorption effect plays no significant role. The estimates of self-absorption made from the measured pressure values in the source chamber and in the spectrometer chamber indicate that self-absorption is not determinative. This is also confirmed by the fact that adding He gas to the jet does not add much to absorption, while a radical change in the shape of the spectrum is observed: the emission practically disappears at 10.8 nm, and the intensity of emission in the wavelength range of 4–10 nm increases sharply. Conversely, a decrease in the intensity within this wavelength range accompanied by a decrease in the gas pressure cannot be related to self-absorption.

The experimental confirmation of this statement is the measured dependence of the signal from the power meter on the displacement of the nozzle perpendicular to the axis of the laser beam and in the direction of the power meter, Fig. 5. The measurements were made both with two Be and MoZrSi2 filters. When moving the nozzle along the direction perpendicular to the axis of the laser beam axis, in the absence of self-absorption, the signal must be symmetric with one maximum corresponding to the position of the focus of the laser beam on the axis of the jet. In the case of self-absorption, the dependence should be distorted: when the laser beam is focused on the edge of the jet from the side opposite to the position of the power meter, the signal should be less than in the case of focusing the laser radiation on the edge of the jet, located closer to the power meter. This effect was not observed.

FIG. 5.

The dependence of the signal from the power meter on the displacement of the nozzle is perpendicular to the axis of the laser beam and in the direction of the power meter. The curves were obtained for two Be and two MoZrSi2 filters.

FIG. 5.

The dependence of the signal from the power meter on the displacement of the nozzle is perpendicular to the axis of the laser beam and in the direction of the power meter. The curves were obtained for two Be and two MoZrSi2 filters.

Close modal

Thus, we believe that ionization processes are responsible for the rearrangement of the spectrum, and not self-absorption of the EUV radiation.

The measurement of the conversion efficiency of laser radiation energy into a spectral line with a maximum of 10.8 nm was performed as follows. Using the radiation power meter (Fig. 1) equipped with a sensitivity-calibrated semiconductor detector (SPD-100UV), filters and a double-mirror Mo/Si monochromator with a resonance wavelength of 11.2 nm, with known spectral transmission bands and a solid angle of the receiving aperture, the total EUV energy incident on the power meter has been measured. Simultaneously, in the perpendicular direction (along the axis of the gas flow), using a Czerny-Turner spectrometer, the spectral dependence of the radiation intensity was measured. To suppress the background radiation, a multilayer Mo/ZrSi2 thin film filter was installed at the output of the spectrometer. Assuming the absence of self-absorption in the plasma and gas, and taking into account the measured spectral efficiency of the diffraction grating, the reflection coefficients of the collimating mirrors M1 and M2 and the transmission of the filters, the sensitivity of the detector based on microchannel plates (MCP), taken from the articles30,31 (all these dependences are shown in Fig. 6), the form of the spectrum at the input of the power meter was restored.

FIG. 6.

Measured spectral dependences of the transmission of the filters and reflectance of the double mirror monochromator (a), the sensitivity of both detectors (b) and reflectance of mirrors and the diffraction grating of the Cherny-Turner spectrometer (c).

FIG. 6.

Measured spectral dependences of the transmission of the filters and reflectance of the double mirror monochromator (a), the sensitivity of both detectors (b) and reflectance of mirrors and the diffraction grating of the Cherny-Turner spectrometer (c).

Close modal

In Fig. 7 the spectra illustrating the methodology for determining conversion efficiency are shown. The upper curve, shown with a break in the ordinate scale, is the reconstructed spectrum at the input of the power meter. The lower curves describe the spectra recorded by the semiconductor detector after passing two identical Mo/ZrSi2 (a thin line) and two identical Be (thick line) filters. Since the sensitivity of a semiconductor detector in a spectral range of 4–12 nm is practically independent of the wavelength, the conversion efficiency (CE) within 2π sr in a spectral range of 10–12 nm can be determined from the following relationships.

FIG. 7.

The spectra illustrating the methodology for determining conversion efficiency. The upper curve, shown with a break in the ordinate scale, is the reconstructed spectrum at the input of the power meter. The lower curves describe the spectra recorded by the semiconductor detector after passing two identical Mo/ZrSi2 (a thin line) and two identical Be (thick line) filters.

FIG. 7.

The spectra illustrating the methodology for determining conversion efficiency. The upper curve, shown with a break in the ordinate scale, is the reconstructed spectrum at the input of the power meter. The lower curves describe the spectra recorded by the semiconductor detector after passing two identical Mo/ZrSi2 (a thin line) and two identical Be (thick line) filters.

Close modal

The power W of the EUV light of the source into a solid angle unit in a spectral range of 10–12 nm was determined by the relation:

WK×VΩ×S×T×I1012Itot[W/(sr)],
(1)

where V is the signal at the detector (in V), Ω is the solid angle from which the radiation was collected (7.5 × 10-4 sr), S is the sensitivity of the detector (0.21 A/W, see Fig. 6), T is the double filter transmittance, I10-12 is the area under a spectral curve at the wavelength range 10–12 nm, Itot is the total area under the curve, and K is the sensitivity of the amplifier (10-6 A/V). Assuming isotropic emissivity, the conversion efficiency CE of the laser radiation power PL into EUV light radiated into the half-space 2π sr, is determined from the relation:

CE=2π×WPL×100[%].
(2)

To achieve the maximum of the conversion efficiency initially at a pressure of Xe PXe = 1.5 ATA and the distance between the end of the nozzle and the laser focus spot of 0.5 mm, optimal parameters of duration of the valve open state τop and the delay time between the opening of the valve and the arriving of the laser pulse τdel were found. They amounted to τop = 250 μs and τdel =1240 μs. As the detector signals were slightly dependent on the change in these parameters, in subsequent experiments, these values remained unchanged.

After optimizing τdel and τop times, we studied the dependence of the conversion efficiency (determined by formulas 1 and 2) on the xenon pressure at the valve inlet for different distances between the nozzle and the focus of the laser beam spot, which was measured with different filters and a two-mirror monochromator. Fig. 8 shows the corresponding dependencies. Since the Mo/ZrSi2 thin film filters have sufficiently higher transmission at the wavelength range of interest to us, they were thus used in most measurements. As follows from the figure, the spectral dependences of the conversion efficiency are fairly smooth with maximal shifting to the region of large xenon pressures with increasing distance from the nozzle to the focus of the laser beam. The maximum conversion efficiency of laser radiation energy into the emission band 10-12nm and 2π solid angle with a peak at 10.8 nm was CE = 4.25 ± 0.30% at a xenon pressure of 1.0 ATA and a distance of 0.5 mm. When we determined the error bars we took into account the signal fluctuation and calibration errors of the optical elements and SPD-100UV semiconductor detector, while the sensitivity of the MCP detector was taken from the literature.

FIG. 8.

Dependences of the conversion efficiency of the laser energy in a 10–12 nm bandwidth on the pressure of Xe at the valve inlet for various distances between the nozzle and the focus of the laser beam, measured with different filters (two pairs of Mo/ZrSi2 or two pairs of Be) and a double-mirror monochromator.

FIG. 8.

Dependences of the conversion efficiency of the laser energy in a 10–12 nm bandwidth on the pressure of Xe at the valve inlet for various distances between the nozzle and the focus of the laser beam, measured with different filters (two pairs of Mo/ZrSi2 or two pairs of Be) and a double-mirror monochromator.

Close modal

The result of measurements obtained with a double-mirror monochromator shown in Fig. 8 (indicated by snowflake) is CE = 5.0 ± 1.5%. This value is 20% greater than the highest CE measured with thin filters for this wavelength range, which may be due to the uncertainty of the sensitivity calibration of the MCP detector and relatively low signal from the double-mirror monochromator.

Since the spectral bandwidth of the optical system depends on the wavelength, composition, and number of multilayer mirrors, the practical use of the obtained data on the conversion of laser radiation energy to a band at 10.8 nm requires precise measurements for both the positions and the widths of this band. For these purposes, the emission spectra of Xe plasma obtained with a Be thin film filter (the sharp absorption K-edge of which lies within the emission line) are analysed. The spectra are given in Fig. 9a. One spectrum was obtained without the Be filter (circles); one was obtained with the filter (rectangles); the dependence (hollow circles), is the multiplication of the upper curve and the calculated transmission of Be filter. The optical constants of Be are taken from Ref. 32. In Fig. 9b, the derivatives of the measured and calculated curves of emission spectra with Be filters are given. In the case of the calculated curve, the width of the bell-shaped curve is determined by the shape of the spectral line (the maximum of the derivative is on the decline of the emission line) and by the dispersion of the imaginary part of the absorption coefficient in the region of the Be absorption K-edge. In the case of the measured dependence, the spectral resolution of the spectrometer also affects the width of the bell-shaped curve. As can be seen from the figure, the half-width of the calculated curve (FWHM) is Δλ = 0.044 nm. The measured curve has a half-width Δλ = 0.105 nm. In the approximation of Gaussian curves, the resolution of the spectrometer can be estimated as Δλ = 0.095 nm.

FIG. 9.

Measured spectral dependences of the intensity of the emission line without Be filter (the curve is labelled with circles) and with Be filter (the curve is labelled with rectangles) and the dependence (hollow circles) which is the multiplication of the upper curve and the calculated dependence of Be filter transmission -a) and the derivatives of the spectrum with Be filters-b).

FIG. 9.

Measured spectral dependences of the intensity of the emission line without Be filter (the curve is labelled with circles) and with Be filter (the curve is labelled with rectangles) and the dependence (hollow circles) which is the multiplication of the upper curve and the calculated dependence of Be filter transmission -a) and the derivatives of the spectrum with Be filters-b).

Close modal

From the known (tabulated) Be absorption K-edge position and bandwidth of the calculated bell-shaped curve, the exact position of the maximum of the emission line have been determined, which was λmax = 10.820 nm, its shape and half-width is Δλ10.8 = 0.398 nm. This line was later used to determine the conversion efficiency of laser radiation into the wavelength band around 11.2 nm (within the bandwidth of Ru/Be and Mo/Be multilayer mirrors) and 10.8 nm (within the bandwidth of Rh/Sr and Rh/Y multilayer mirrors).

In the next section, we will turn to the analysis of 5- and 12-mirror optical systems. The first system is of interest for maskless EUV lithography,14 the latter for classical EUV projection lithography.2 Let’s also consider the two wavelengths: λ = 11.2 nm and λ = 10.8 nm. The first wavelength is interesting in that it lies within the emission band of Xe ions and that at this wavelength it is theoretically possible to obtain high reflectivity (up to 78.2%) for the Ru/Be multilayer mirrors. The wavelength of 10.8 nm is of interest, since it corresponds to the maximum in emission spectrum of the Xe source and that at this wavelength the calculated reflectance for the Rh/Sr multilayers is 73.9% and for the Rh/Y multilayers is 64.2%. The phosphorus-based multilayer structure considered in the article10 may possess high reflectivity in theory but the question of how such a structure can be made remains unanswered. Therefore, this structure does not warrant our consideration.

In Figs. 10 and 11, the emission line and the spectral dependences of the reflectivity of 5 (a) and 12 (b) mirror systems for the Ru/Be and Mo/Be (with the peak at 11.2 nm), and the Rh/Sr and Rh/Y (with the peak at 10.8 nm) multilayer mirrors are shown. The integral under the reflection curves Iopt with respect to the integral under the entire emission line Il

EFos=IoptIl,
(3)

characterizes the effectiveness of a specific optical system EFos and represents the part of EUV energy to be captured by the system in assumption of 100% reflectance of the mirrors. The conversion efficiency corresponding to these fractions are given in Table I.

FIG. 10.

Emission line and spectral dependences of the reflectivity of 5 and 12-mirror systems based on the Ru/Be-a) and Mo/Be-b) multilayer mirrors.

FIG. 10.

Emission line and spectral dependences of the reflectivity of 5 and 12-mirror systems based on the Ru/Be-a) and Mo/Be-b) multilayer mirrors.

Close modal
FIG. 11.

Emission line and spectral dependences of the reflectivity of 5- and 12-mirror systems based on the Rh/Sr-a) and the Rh/Y-b) multilayer mirrors.

FIG. 11.

Emission line and spectral dependences of the reflectivity of 5- and 12-mirror systems based on the Rh/Sr-a) and the Rh/Y-b) multilayer mirrors.

Close modal
TABLE I.

Comparison of the productivity of the most promising optical systems at 10.8 and 11.2 nm wavelengths.

Number of mirrorsN = 5N = 12
Wavelength, nm11.210.811.210.8
StructureRu/BeMo/BeRh/SrRh/YRu/BeMo/BeRh/SrRh/Y
Bandwidth, nm 0.2080 0.1864 0.2473 0.2096 0.1748 0.1508 0.1768 0.1452 
EFos 0.0981 0.0914 0.3878 0.3280 0.0747 0.0727 0.2851 0.2365 
EFos×CEl, % 0.417 0.388 1.648 1.394 0.317 0.309 1.212 1.005 
RN 0.2923 0.2607 0.2202 0.1094 0.0522 0.0399 0.0265 0.0049 
PLPos, % 0.1219 0.1012 0.3629 0.1525 0.0165 0.0123 0.0321 0.0049 
Number of mirrorsN = 5N = 12
Wavelength, nm11.210.811.210.8
StructureRu/BeMo/BeRh/SrRh/YRu/BeMo/BeRh/SrRh/Y
Bandwidth, nm 0.2080 0.1864 0.2473 0.2096 0.1748 0.1508 0.1768 0.1452 
EFos 0.0981 0.0914 0.3878 0.3280 0.0747 0.0727 0.2851 0.2365 
EFos×CEl, % 0.417 0.388 1.648 1.394 0.317 0.309 1.212 1.005 
RN 0.2923 0.2607 0.2202 0.1094 0.0522 0.0399 0.0265 0.0049 
PLPos, % 0.1219 0.1012 0.3629 0.1525 0.0165 0.0123 0.0321 0.0049 

Table I shows the data comparing the productivity (PLPos) of the lithographic process of different lithographic schemes. In this table, as a quantity characterizing the efficiency of the scheme, we used the values

PLPos=RN×EFos×CEl[%],
(4)

where R is the maximum theoretical reflectivity of the multilayer mirrors (expressed in absolute magnitudes), N is the number of mirrors in the optical system, and CEl = 4.25% is the value of the conversion efficiency of the source in the spectral range of 10–12 nm.

In this study, we optimized the parameters of a double-stream Xe/He gas jet in order to achieve a maximum radiation flux ∼11 nm wavelength. It is shown that the presence of the He gas jet substantially increases the emission in a wavelength band of 4–10 nm from Xe ions with degrees of ionization +14 ... + 16. However, with the He gas jet, the He/Xe plasma emission at 11 nm significantly decreases. The moving of the interaction area away from the nozzle leads to a strong drop at the short-wave portion of the EUV spectrum and even to a small increase in emission around 11 nm. Both phenomena are explained by the fact that in the case of dense plasma, the key factor determining the ionization composition of the plasma is the jet density, rather than laser light intensity. Due to the high frequency of electron-ion collisions and lower electron diffusion, the ionization state of denser plasma shifts to the higher charges. With lowered jet density, the rate of the electron-ion collisions decreases and the rate of electron diffusion out of the interaction zone increases. This shifts the ionization state of denser plasma toward smaller ionization states and thus leads to an increase in emission at 11 nm. Generally, our experiments have shown that in the range of studied values of the pressure of Xe (up to 4 ATA) and He (1 ATA) to achieve the maximum conversion efficiency at the range of 11 nm wavelength, it is better to use a single-stream Xe gas jet.

The exact position of the maximum of the emission line, which was equal to λmax = 10.820 nm, its shape and half-width Δλ10.8 = 0.398 nm have been determined. The dependence of the conversion efficiency (CE) of the laser radiation energy into the line on xenon pressure at the valve inlet and on the distance between the nozzle end and the interaction region of the laser radiation with the jet was measured. With the removal from the nozzle, the CE monotonically falls and its maximum is achieved with a greater pressure of xenon at the valve inlet.

The measured CE value differs by about 20% when we measure it with the filters and with the double-mirror monochromator. The observed difference in CE value is associated with the uncertainty of the sensitivity calibration of the MCP detector and relatively low signal from the monochromator. It should be noted that in the case when the double-mirror monochromator is used, virtually all out-of-band radiation is effectively absorbed. Therefore, when measuring the CE, the spectral dependence of the sensitivity of the MCP is not used. However, due to a significant measurement error, we gave a conservative value of CE = 4.25 ± 0.30% when measured with the filters.

The spectral bandpass of 5- and 12-mirror systems for the most promising multilayer mirrors and wavelengths of 10.8 and 11.2 nm were calculated. Taking into account the measured conversion efficiency of the source CE = 4.25% in the spectral range 10–12 nm, the conversion efficiencies of the source in the spectral bandpass of these optical systems were found and the maximum efficiency of each of them was compared. For the five-mirror system at λ = 10.8 and with the Rh/Sr multilayered mirrors, the conversion efficiency of the source was about 1.65%. For the Ru/Be multilayer mirrors at λ =11.2 nm, the conversion efficiency is about 4 times smaller and amounted to CE = 0.42%. The maximal productivity of the lithographic process at λ = 10.8 was almost 3 times higher than the analogous parameter at λ = 11.2 nm.

All this is true only if the maximum reflectivity of multilayer mirrors is achieved. At present, it is known that for Mo/Be multilayer mirrors at 11.2 nm, a reflectance of 70.2% is achieved,33,34 while at 10.8 nm for Rh/Sr and Pd/Y, it is only about 50%.35,36 If we substitute these reflectivity and measured CE values from Table I in relation (4), we obtain the following productivities: at λ = 11.2 nm and with five Mo/Be mirrors, PLP5 = 0.0652 and at λ = 10.8 nm and with five Rh/Sr mirrors PLP5 = 0.0515. Thus, the up-to-date technology for the deposition of multilayer mirrors indicates that despite the relatively low energy conversion efficiency at 11.2 nm due to the higher reflectivity of Be-based multilayer mirrors, the wavelength of 11.2 nm still looks preferable for EUV projection lithography. The final choice can be made after obtaining the limiting values of the reflectance for the Sr- and Y-based multilayer optics.

In conclusion, it should be noted that, in practical use of the results obtained, it is necessary to find the optimal distance between the end of the nozzle and the focus of the laser beam, which will ensure acceptable levels of erosion of the nozzle and contamination of the optics, and an acceptable value for conversion efficiency.

The work was done in frame work of the project 0035-2014-0204 as well as supported by the RSF, grant # 17-12-01227 in part related to development of the LPP source and experimental investigation, and the RSF-DFG, grant # 16-42-01034 in part related to the spectrometer-monochromator development and RFBR, grants ## 18-02-00173, 18-02-00588, 18-07-00633, 17-52-150006, 16-07-00306 and 16-07-00247.

1.
S.-S.
Kim
,
R.
Chalykh
,
H.
Kim
,
S.
Lee
,
C.
Park
,
M.
Hwang
,
J.-O.
Park
,
J.
Park
,
H.
Kim
, and
J.
Jeon
, “
Progress in EUV lithography toward manufacturing
,”
Proceedings of SPIE
10143
,
1014306
(
2017
).
2.
S.
Migura
,
B.
Kneer
,
J. T.
Neumann
,
W.
Kaiser
, and
J.
van Schoot
, “
Anamorphic high- NA EUV lithography optics
,”
Proceedings of SPIE
9661
,
96610T
(
2015
).
3.
A.
Pirati
,
J.
van Schoot
,
K.
Troost
,
R.
van Ballegoij
,
P.
Krabbendam
,
J.
Stoeldraijer
,
E.
Loopstra
,
J.
Benschop
,
J.
Finders
, and
H.
Meiling
, “
The future of EUV lithography: Enabling Moore’s Law in the next decade
,”
Proceedings of SPIE
10583
,
105830R
(
2018
).
4.
N.
Salashchenko
and
N.
Chkhalo
, “
Shortwave projection nanolithography
,”
Herald of the Russian Academy of Sciences
78
,
279
(
2008
).
5.
C.
Wagner
and
N.
Harned
, “
EUV lithography: Lithography gets extreme
,”
Nature Photonics
4
,
24
(
2010
).
6.
N.
Chkhalo
and
N.
Salashchenko
, “
Next generation nanolithography based on Ru/Be and Rh/Sr multilayer optics
,”
AIP Advances
3
,
082130
(
2013
).
7.
E.
Kieft
,
K.
Garloff
,
J.
Van der Mullen
, and
V.
Banine
, “
Comparison of experimental and simulated extreme ultraviolet spectra of xenon and tin discharges
,”
Physical Review E
71
,
036402
(
2005
).
8.
R.
Rakowski
,
A.
Bartnik
,
H.
Fiedorowicz
,
F.
De Gaufridy De Dortan
,
R.
Jarocki
,
J.
Kostecki
,
J.
Mikołajczyk
,
L.
Ryć
,
M.
Szczurek
, and
P.
Wachulak
, “
Characterization and optimization of the laser-produced plasma EUV source at 13.5 nm based on a double-stream Xe/He gas puff target
,”
Applied Physics B
101
,
773
(
2010
).
9.
D. L.
Windt
and
E. M.
Gullikson
, “
Pd/B4C/Y multilayer coatings for extreme ultraviolet applications near 10 nm wavelength
,”
Applied Optics
54
,
5850
(
2015
).
10.
V.
Medvedev
,
A.
Yakshin
,
R. W. E.
van de Kruijs
, and
F.
Bijkerk
, “
Phosphorus-based compounds for EUV multilayer optics materials
,”
Optical Materials Express
5
,
1450
(
2015
).
11.
N.
Choksi
,
D.
Pickard
,
M.
McCord
,
R.
Pease
,
Y.
Shroff
,
Y.
Chen
,
W.
Oldham
, and
D.
Markle
, “
Maskless extreme ultraviolet lithography
,”
Journal of Vacuum Science & Technology B
17
,
3047
(
1999
).
12.
Y.
Chen
, “
Maskless EUV lithography: An already difficult technology made even more complicated
,”
Proceedings of SPIE
8323
,
83231Q
(
2012
).
13.
A. F. G.
Leontovich
,
A. P.
Hitchcock
,
B.
Watts
, and
J.
Raabe
, “
Sub-25 nm direct write (maskless) x-ray nanolithography
,”
Microelectronic Engineering
108
,
5
(
2013
).
14.
N.
Chkhalo
,
V.
Polkovnikov
,
N.
Salashchenko
, and
M.
Toropov
, “
Problems and prospects of maskless (B)EUV lithography
,”
Proceedings of SPIE
10224
,
102241O
(
2016
).
15.
N.
Chkhalo
,
V.
Polkovnikov
,
N.
Salashchenko
, and
M.
Toropov
, “
Deposition of Mo/Si multilayers onto MEMS micromirrors and its utilization for extreme ultraviolet maskless lithography
,”
Journal of Vacuum Science & Technology B
35
,
062002
(
2017
).
16.
B.
Hansson
,
L.
Rymell
,
M.
Berglund
, and
H. M.
Hertz
, “
A liquid-xenon-jet laser plasma x-ray and EUV source
,”
Microelectronic Engineering
53
,
667
(
2000
).
17.
H.
Fiedorowicz
,
A.
Bartnik
,
R.
Jarocki
,
J.
Kostecki
,
J.
Mikolajczyk
,
R.
Rakowski
, and
M.
Szczurek
, “
Compact laser plasma EUV source based on a gas puff target for metrology
,”
Proceedings of SPIE
5037
,
389
(
2003
).
18.
U.
Stamm
,
I.
Ahmad
,
I.
Balogh
,
H.
Birner
,
D.
Bolshukhin
,
J.
Brudermann
,
S.
Enke
,
F.
Flohrer
,
K.
Gäbel
, and
S.
Götze
, “
High-power EUV lithography sources based on gas discharges and laser-produced plasmas
,”
Proceedings of SPIE
5037
,
119
(
2003
).
19.
K. N.
Koshelev
,
V. E.
Banine
, and
N. N.
Salashchenko
, “
Research and development in short-wave radiation sources for new-generation lithography
,”
Uspekhi Fizicheskikh Nauk
50
,
741
(
2007
).
20.
S.
Ter-Avetisyan
,
U.
Vogt
,
H.
Stiel
,
M.
Schnürer
,
I.
Will
, and
P. V.
Nickles
, “
Efficient extreme ultraviolet emission from xenon-cluster jet targets at high repetition rate laser illumination
,”
Journal of Applied Physics
94
,
5489
(
2003
).
21.
A.
Shimoura
,
S.
Amano
,
S.
Miyamoto
, and
T.
Mochizuki
, “
X-ray generation in cryogenic targets irradiated by 1 mμ laser
,”
Applied Physics Letters
72
,
164
(
1998
).
22.
H.
Fiedorowicz
,
A.
Bartnik
,
R.
Jarocki
,
J.
Kostecki
,
J.
Krzywiński
,
J.
Mikołajczyk
,
R.
Rakowski
,
A.
Szczure
, and
M.
Szczurek
, “
Compact laser plasma EUV source based on a gas puff target for metrology applications
,”
Journal of Alloys and Compounds
401
,
99
(
2005
).
23.
S. A.
Garakhin
,
I. G.
Zabrodin
,
S.
Zuev
,
I. A.
Kas’kov
,
A. Y.
Lopatin
,
A.
Nechay
,
V. N.
Polkovnikov
,
N. N.
Salashchenko
,
N. N.
Tsybin
, and
N. I.
Chkhalo
, “
Laboratory reflectometer for the investigation of optical elements in a wavelength range of 5–50 nm: Description and testing results
,”
Quantum Electronics
47
,
385
(
2017
).
24.
P. N.
Aruev
,
M. M.
Barysheva
,
B. Y.
Ber
,
N. V.
Zabrodskaya
,
V. V.
Zabrodskii
,
A. Y.
Lopatin
,
A. E.
Pestov
,
M. V.
Petrenko
,
V. N.
Polkovnikov
, and
N. N.
Salashchenko
, “
Silicon photodiode with selective Zr/Si coating for extreme ultraviolet spectral range
,”
Quantum Electronics
42
,
943
(
2012
).
25.
M.
Bibishkin
,
D.
Chekhonadskih
,
N.
Chkhalo
,
E.
Kluyenkov
,
A.
Pestov
,
N. N.
Salashchenko
,
L.
Shmaenok
,
I.
Zabrodin
, and
S. Y.
Zuev
, “
Laboratory methods for investigations of multilayer mirrors in extreme ultraviolet and soft x-ray region
,”
Proceeding of SPIE
5401
,
8
(
2003
).
26.
N. I.
Chkhalo
,
N. N.
Salashchenko
,
S. V.
Golubev
,
D.
Mansfeld
,
A. V.
Vodopyanov
, and
L.
Sjmaenok
, “
Source for extreme ultraviolet lithography based on plasma sustained by millimeter-wave gyrotron radiation
,”
Journal of Micro/Nanolithography, MEMS, and MOEMS
11
,
021123
(
2012
).
27.
H.
Fiedorowicz
,
A.
Bartnik
,
Z.
Patron
, and
P.
Parys
, “
X-ray emission from laser-irradiated gas puff targets
,”
Applied Physics Letters
62
,
2778
(
1993
).
28.
M.
McGeoch
, “
Radio-frequency-preionized xenon z-pinch source for extreme ultraviolet lithography
,”
Applied Optics
37
,
1651
(
1998
).
29.
N.
Chkhalo
,
S.
Garakhin
,
S.
Golubev
,
A. Y.
Lopatin
,
A.
Nechay
,
A.
Pestov
,
N.
Salashchenko
,
M.
Toropov
,
N.
Tsybin
, and
A.
Vodopyanov
, “
A double-stream Xe: He jet plasma emission in the vicinity of 6.7 nm
,”
Applied Physics Letters
112
,
221101
(
2018
).
30.
M.
Kowalski
,
G.
Fritz
,
R.
Cruddace
,
A.
Unzicker
, and
N.
Swanson
, “
Quantum efficiency of cesium iodide photocathodes at soft x-ray and extreme ultraviolet wavelengths
,”
Applied Optics
25
,
2440
(
1986
).
31.
N.
Yamaguchi
,
S.
Aoki
,
Y.
Kiwamoto
,
M.
Oshiba
,
T.
Saito
, and
S.
Miyoshi
, “
A pinhole camera with microchannel plates for hot electron spatial distribution measurement
,”
Japanese Journal of Applied Physics
24
,
1065
(
1985
).
33.
C.
Montcalm
,
S.
Bajt
,
P. B.
Mirkarimi
,
E. A.
Spiller
,
F. J.
Weber
, and
J. A.
Folta
, “
Multilayer reflective coatings for extreme-ultraviolet lithography
,”
Proceedings of SPIE
3331
,
42
(
1998
).
34.
N. I.
Chkhalo
,
S. A.
Gusev
,
A. N.
Nechay
,
D. E.
Pariev
,
V. N.
Polkovnikov
,
N. N.
Salashchenko
,
F.
Schäfers
,
M. G.
Sertsu
,
A.
Sokolov
, and
M. V.
Svechnikov
, “
High-reflection Mo/Be/Si multilayers for EUV lithography
,”
Optics Letters
42
,
5070
(
2017
).
35.
B.
Sae-Lao
and
C.
Montcalm
, “
Molybdenum–strontium multilayer mirrors for the 8–12-nm extreme-ultraviolet wavelength region
,”
Optics Letters
26
,
468
(
2001
).
36.
D. L.
Windt
, “
EUV multilayer coatings for solar imaging and spectroscopy
,”
Proceedings of SPIE
9604
,
96040P
(
2015
).