We present and demonstrate the use of an extreme ultraviolet (EUV) microscope that was developed in-house. Images are acquired using Bragg reflection multilayer optics and a laser-produced plasma light source. The upper-limit spatial resolution of the EUV microscope is 130 nm with a 10 ns exposure time and 250 × 250 μm2 field of view. Resolution is superior to that of visible microscopes with the same size of field of view, and the exposure time is short enough to observe fine structures in-vivo. Observation of the cerebral cortex of a mouse is demonstrated.
© 2010 OSA
Living organisms have an internal structural hierarchy. The structures within each hierarchy have characteristic shapes and functions related to its shape. Some small structures within a hierarchy combine interactively, and make new structures by which a new characteristic function is achieved in the upper hierarchy. It is therefore important to understand the functions of a structure to simultaneously observe component structures and clarify their specific role .
We usually use visible microscopes (VMs) to observe structures of living organisms. Spatial resolution of VMs is represented by Rayleigh’s criterion , the value is 290 nm using Plan-Apochromat × 60 objective with 0.95 numerical apertures (NA) in 550 nm wavelength. Field of view in VMs is determined by field-of-view number that is characteristic value in a VM system, and the value of field-of-view is 330 μm in diameter with the use of × 60 objective in a 20 field-of-view number system. These values suggest that large organelle-level (more than 1 μm) and cell-level structures can be observed simultaneously by VMs. Similarly, protein-level structures can be observed with smaller organelle-level ones (less than 1 μm) by using transmission electron microscopes (TEMs). Spatial resolution of TEM is also represented by the Rayleigh’s criterion and is proportional to the kinetic energy of incident electrons . The value of a standard TEM for biological observation is about 0.2 nm at 100 keV electron energy. The field of view depends on both magnification and area size of an imaging device installed in a TEM, the value is about 100 nm in diameter. Therefore, there is no tool to observe simultaneously both protein- or small organelle-level structures and tissue-or organ-level ones especially for living organisms .
Adding to the simultaneous observation, observation of active cell in water will be desirable. Because water is transparent in visible wavelength region, the relationship between component structures of organelle-level hierarchy and composed structures of the cell-level hierarchy has been studied extensively with VMs. On the other hand, biological samples are usually fixed and stained for observation with TEMs. In addition, a short exposure time is necessary to observe active cells. When an active cell in water is displaced by Brownian motion, its displacement during sufficiently punctual observation is small enough to observe its fine structures, which are below the spatial resolution of a microscope. For example, with an active 1 μm diameter cell the time required to observe fine structures less than 20 nm in size is less than 1 ms when using a microscope with a spatial resolution of 20 nm or less . A microscope that can observe active and living organisms across several hierarchies can be called “super-hierarchical”, whose features include a wide field of view where multi-hierarchical observation can be achieved and high spatial resolution by which the smallest organelle can be observed.
Because of the short wavelength of extreme ultraviolet (EUV) light, the spatial resolution of EUV microscopes (EUVMs) is 1 to 2 orders of magnitude higher than that of VMs. Microscopes using EUV light have been developed mainly with Fresnel zone plates (FZPs) as objective lenses . Because of the small NA of FZPs, microscopes using them require a bright, pencil-shaped light beam. Therefore, microscopes using FZPs are usually developed using a synchrotron radiation (SR) insertion device as a light source. At present, the highest spatial resolution attained with a soft X-ray (1.75 nm) microscope using both a FZP and a SR light source is 12 nm, and the field of view is limited to 2.5 μm in diameter .
In spite of this development, practical use of microscopes utilizing laboratory-type light sources is desirable. It is possible to observe living cells using a laser-produced plasma (LPP) as a light source for an EUVM because the source supplies EUV pulses shorter than the 1 ms required because of Brownian motion. Interaction of an intense laser pulse with a solid target generates high-temperature plasma, which emits EUV radiation in order that the pulse duration of the EUV radiation from an LPP light source is equal to that of the excitation laser pulse . An EUVM with an LPP light source is thus an extremely valuable analytical tool for observing active living samples.
An LPP produces EUV light from the focal point of an excitation laser. The small NA of a FZP makes it unsuitable as an objective lens for an EUVM with an LPP light source . Although a FZP objective having large NA was recently developed , the NA of a Schwarzschild objective (SO) is twice larger than that of the FZP objective. Therefore, a SO proves to be well suited to image the EUV light from the laser focal point because of the large NA . It should also be added that an SO objective is favorable in that the field-of-view provided by SOs is large compared with that of FZPs in EUV wavelength region. Because the field-of-view of SOs is proportional to the square root of incident wavelength λ and inversely proportional to the NA [10,11], when spatical resolution of objectives is given by Rayleigh's criterion. The field-of-view of FZPs is inversely proportional to the square of wavelength λ . Therefore 13.4 nm wavelength, which we have been used in the present study, will give a wide field-of-view which is needed in “super-hierarchical” microscopes. In addition, LPP light sources and multilayer optics are well studied for next generation projection lithography in the wavelength . We therefore combine an LPP light source with an SO in our newly developed EUVM. For the optics to reflect EUV light from their surfaces, reflective multilayers must be deposited on the surface of the optics. To accomplish this, we developed an ion-beam deposition system that enables us to accurately control the d-spacing of the reflection multilayers . In the current publication, we present details of our EUVM, which include spatial resolution, field of view, exposure time and a variety of sample images.
2. Development of transmission x-ray multilayer mirror microscope
The EUVM has been named the Transmission X-ray Multilayer Mirror Microscope (TXM3 have been used as the acronym of this name in our previous publication. In this publication, we use TXM-CUBE to avoid confusion of the acronym TXM for transmission X-ray microscope.). The schematic layout of the TXM-CUBE is shown in Fig. 1 . It consists of an LPP light source, illumination optics to collect EUV light from this source, an optical filter to block visible light emanating from the light source, a SO to form an EUV image of the sample under observation and an EUV-sensitive CCD camera to record the EUV image. A visible light source, a VI/EUV switching mirror and a CCD camera sensitive to visible light are used for positioning each sample. The SO is shared by the EUV, and the visible optics.
A pulsed Nd:YAG laser (LAB-170, Spectra-Physics, Japan) is used to create the LPP light source. The fundamental wavelength is 1064 nm, the maximum pulse energy is 850 mJ, the pulse repetition frequency (repeat rate) is 10 Hz and the pulse width is 8 to 12 ns. The laser is focused at the surface of a W target where the plasma is generated. EUV light with conversion efficiency from the fundamental wavelength of 0.4% was obtained experimentally in 2% bandwidth.
We designed and fabricated illumination optics to collect the EUV light emanating from the LPP light source . The optics consists of a primary and a secondary reflecting spherical mirror of equal radius of curvature, each facing the other (see Fig. 1). The primary optic generates a parallel EUV beam from the LPP light source by reflection, while the secondary optic focuses the parallel EUV beam by reflection to the focus position.
EUV light passing through the sample is magnified and focused by the SO. The numerical aperture and magnification of the SO is 0.24, and 50, respectively . We estimate through optical calculation that the allowable eccentric error between the two SO mirrors is ± 100 nm. To correct the eccentric error, we developed a precision mirror holder with a two-axis piezoelectric stage .
To reflect 13.4 nm EUV light, Mo/Si multilayers were deposited on the spherical substrates of the illumination optics and the SO. To satisfy the Bragg condition, the period of the Mo/Si multilayer was graded according to the incident angle of the EUV light . The absolute reflectance of each mirror was determined from multilayers that had been deposited simultaneously on Si substrates due to the spherical shape of each mirror. The Si substrates were placed on jigs that emulate spherical mirror substrates, with the coincidence of reflectance between the two confirmed in advance . The measured reflectance from the reflective coatings at 5° from normal incidence was measured between 0.56 and 0.61. The throughput at 13.4 nm through the entire optical system is 0.13.
Unfortunately, the deposited Mo/Si multilayer reflects wavelengths above 100 nm by specular reflection, which then pollutes the Bragg reflected EUV radiation. To block this light, a 200 nm thick Zr membrane filter is placed in front of the sample between the illumination optics and the SO. Transmittance through the Zr membrane filter is 0.001 or less in the visible region, and is 0.1 at 13.4 nm.
The EUV image formed after passing through the sample is expanded and imaged by the SO on a back-illuminated, EUV-sensitive CCD (C4742-98-24KADV, Hamamatsu Photonics, Japan). The CCD pixel size is 13.5 × 13.5 μm2, with the entire imaging area at 1024 × 1024 pixels. Using a Peltier cooler, the CCD is cooled to −60 °C.
3. Evaluation of TXM-CUBE
EUV images of a micro-grid sample holder for TEMs were taken to evaluate the spatial resolution of the TXM-CUBE. To obtain an EUV image, the sample was first positioned at the EUV beam focus using a visible observation system, then switched to EUV observation. The EUV image of the net structure of the micro-grid, shown in the inset of Fig. 2 , was taken using 100 mJ/pulse and with a 10 ns exposure time. The polymer net structure is not clearly imaged by the TEM because of weak absorption of high-energy electrons by the high polymer. However, the high polymer strongly absorbs in the EUV region so the net structure is clear in the EUV image shown in the inset of Fig. 2.
Rectangular area in the inset to Fig. 2 is a nano-wire element made from a mesh structure. The main panel of Fig. 2 shows the EUV intensity change across the nano-wire. The abscissa is calibrated according to the number of pixels in the CCD image, and the magnification of the SO, with the ordinate having an average response of eight neighboring pixels. Figure 2 indicates that the intensity change caused by the nano-wire is clearly detected by only two pixels. The spatial resolution is evaluated by fitting the intensity change to a Gaussian, with its HWHM value at 130 nm. The value obtained from the nano-wire response corresponds to the spatial resolution estimated by using a knife edge response method. In addition, the Airy disk radius is estimated from both the measured wave front profile with the use of VI light (Fig. 3 ) and the Maréchal condition, and corresponds to the 130 nm spatial resolution obtained experimentally . The estimated value is 107 nm, which is nearly equal to the value obtained using the nano-wire response method. Therefore, we estimate that the spatial resolution of TXM-CUBE is better than 130 nm.
An EUV image of the cerebral cortex of a mouse was taken to evaluate the image quality of a biological object. The mouse brain was fixed and stained to observe the neural circuit through X-ray imaging . The fixed and stained cerebrum cortex was embedded in a resin and then polymerized. The polymerized tissue was sliced by the use of a ultramicrotome to produce films approximately 500 nm thick. The EUV image of one such slice is presented in Fig. 4 . The scale of the image was presented by a white bar in the figure. The field of view of the image was limited by the area of the CCD, which was 250 × 250 μm2. The number of laser shot to take the image was 10. The obtained image was a grayscale and the color of the image was changed from gray to purple for increased observational clarity.
The cerebral cortex structures examined are noted in Fig. 4. The image contains axons that have expanded to the inner side of the cerebrum (white matter), and extended neurites through the cell bodies that are intertwined complexly. Moreover, nuclei and other minute structures in the cell bodies are also evident in the image. The EUV image was compared in Fig. 5 with a visible image with a tissue of the same cerebral cortex region dyed for VMs that was obtained with almost the same magnification ( × 40). The comparison indicates that the intertwined structure of neurites is clearly imaged, and the EUV image was contrasted from the visible image. Since a portion of the sample was out of the focal depth, a portion of the obtained image was out of focus (for example, the left part at the center column, upper row.) When the spatial resolution is given by Rayleigh’s criterion, the focal depth is given by 0.61λ/NA2 and the value of the present SO objective is 142 nm. The ruggedness of the sample, which was prepared by the usual procedures with the use of a standard VM , will be more than the focal depth.
It is important to mention the sample thickness of the obtained image. The transmittance value of the obtained image depends on the region of the sample. The values of nuclei and axons lightly stained were 0.08 and 0.13, respectively. The extinction coefficients obtained from the transmittance and the sample thickness were 0.005 and 0.004 at nuclei and axons, respectively. These values are one order of magnitude smaller than graphite crystal at 0.071, and smaller than deoxyribose at 0.007, which are estimated from the atomic scattering factors . Small extinction coefficients will be caused by decreasing density that was promoted by sample preparation. Small extinction coefficients of cell structures in 13.5 nm wavelength enable the observation of a sample, which has a thickness greater than that of a typical TEM sample. The difference of both the extinction coefficients and the thick thickness may give different information from observations using VMs or TEMs.
A transmission type EUVM was developed using multilayer optics, and an LPP light source. The upper limit of spatial resolution obtained was 130 nm when using a single pulse (10 ns) of an excitation laser. EUV images of a cerebral cortex sample from a mouse were taken under a 30 ns exposure time using a 250 × 250 μm2 field of view. The spatial resolution and exposure time data indicate that the highest spatial resolution is obtained under the shortest exposure time with the largest field of view. Images taken through the use of the TXM-CUBE show that EUVMs using diffraction-limit normal incident optics will be a promising advancement as a “super-hierarchical” observational tool for examining living organisms.
This work was supported by The Ministry of Education, Culture, Sports, Science and, Technology (MEXT), Grant-in-Aid for Specially promoted Research 15002001.
References and links
1. For example, B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts, and P. Walter, Molecular biology of the cell, 5th ed., (Garland Science, Taylor & Francis Group, New York, 2008). [PubMed]
2. M. Born and E. Wolf, Principles of Optics, 7th ed., (Cambridge University Press, Cambridge, 1999) Sec. 8.6.2.
3. L. Reimer, Transmission Electron Microscope, 3rd ed., (Springer-Verlag, Berlin Heidelberg, 1984, 1989, and 1993) Chap. 1.
4. A. Ito and K. Shinohara, “Image blurring by thermal diffusion in the observation of hydrated biomolecules with soft X-ray microscopy,” Cell Struct. Funct. 17(4), 209–212 (1992). [CrossRef] [PubMed]
5. For example, D. Attwood, Soft X-rays and extreme ultraviolet radiation, (Cambridge University Press, Cambridge, 2000) Chap. 9.
6. W. Chao, J. Kim, S. Rekawa, P. Fischer, and E. H. Anderson, “Demonstration of 12 nm resolution Fresnel zone plate lens based soft x-ray microscopy,” Opt. Express 17(20), 17669–17677 (2009). [CrossRef] [PubMed]
8. C. A. Brewer, F. Brizuela, P. Wachulak, D. H. Martz, W. Chao, E. H. Anderson, D. T. Attwood, A. V. Vinogradov, I. A. Artyukov, A. G. Ponomareko, V. V. Kondratenko, M. C. Marconi, J. J. Rocca, and C. S. Menoni, “Single-shot extreme ultraviolet laser imaging of nanostructures with wavelength resolution,” Opt. Lett. 33(5), 518–520 (2008). [CrossRef] [PubMed]
9. I. A. Artioukov, A. V. Vinogradov, V. E. Asadchikov, Yu. S. Kas’yanov, R. V. Serov, A. I. Fedorenko, V. V. Kondratenko, and S. A. Yulin, “Schwarzschild soft-x-ray microscope for imaging of nonradiating objects,” Opt. Lett. 20(24), 2451–2453 (1995). [CrossRef] [PubMed]
10. M. Toyoda and M. Yamamoto, “Analytical designing of two-aspherical-mirror anastigmats permitting practical misalighnments in a soft-X-ray optics,” Opt. Rev. 13(3), 149–157 (2006). [CrossRef]
11. When spatical resolution is given by Rayleigh’s criterion , the effective field of view provided by SOs is represented as (1.22λ/(P × NA))1/2 with the use of wavelength λ, coefficient of field curvature P, and numerical aperture NA . Applying the sine condition to a high magnification Fresnel zone plate (FZP) imaging system that the spherical aberration is corrected, the outer zone width Δ of the FZP is represented as ymaxNA’2/2, where ymax is the maximum image height at image plane and NA’ is the numerical aperture of the FZP. The outer zone width Δ of FZP systems is also represented as λ/2NA’ at wavelength λ . When the former two equations equal to each other, effective field of view of a high magnification FZP system is represented by a maximum image height and the notation is given by 8Δ3/λ2.
12. For example, D. Attwood, Soft X-rays and extreme ultraviolet radiation, (Cambridge University Press, Cambridge, 2000) Chaps. 4, 6, and 10.
13. T. Harada, Dr. of Eng. thesis, Tohoku University, 2007 (in Japanese).
14. M. Toyoda, JPN Patent pending.
15. M. Toyoda, Y. Shitani, M. Yanagihara, T. Ejima, M. Yamamoto, and M. Watanabe, “A soft-X-ray imaging microscope with a multilayer-coated Schwarzschild objective: Imaging tests,” Jpn. J. Appl. Phys. 39(Part 1, No. 4A), 1926–1929 (2000). [CrossRef]
16. A custom product of Nikon Engineering Co., Ltd., email@example.com or http://www.ave.nikon.co.jp/n-eng/
17. M. Born and E. Wolf, Principles of Optics, 7th ed. (Cambridge University Press, Cambridge, UK, 1999) p.528.
18. H. Mizutani, Y. Takeda, A. Momose, A. Takeuchi, and T. Takagi, “X-ray microscopy for neural circuit reconstruction,” J. Phys. Conf. Series 186, 012092 (2009). [CrossRef]
19. The extinction coefficient values are computed from the values of atomic scattering factor and density by the use of the computer program IMD. The details of the computer program IMD is described in the paper:D. L. Windt, “IMD-Software for modeling the optical properties of multilayer films,” Comput. Phys. 12(4), 360–370 (1998). [CrossRef]