Microscopy, 2015, 181–187 doi: 10.1093/jmicro/dfv011 Advance Access Publication Date: 6 March 2015
Article
Phase-contrast scanning transmission electron microscopy Hiroki Minoda1,*, Takayuki Tamai1, Hirofumi Iijima2, Fumio Hosokawa2, and Yukihito Kondo2 Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
1
Department of Applied Physics, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan, and 2JEOL Ltd, 3-1-2 Musashino, Akishima, Tokyo 196-8558, Japan
*To whom correspondence should be addressed. E-mail: [email protected] Received 17 October 2014; Accepted 11 February 2015
Abstract This report introduces the first results obtained using phase-contrast scanning transmission electron microscopy (P-STEM). A carbon-film phase plate (PP) with a small center hole is placed in the condenser aperture plane so that a phase shift is introduced in the incident electron waves except those passing through the center hole. A cosine-type phase-contrast transfer function emerges when the phase-shifted scattered waves interfere with the non-phase-shifted unscattered waves, which passed through the center hole before incidence onto the specimen. The phase contrast resulting in P-STEM is optically identical to that in phase-contrast transmission electron microscopy that is used to provide high contrast for weak phase objects. Therefore, the use of PPs can enhance the phase contrast of the STEM images of specimens in principle. The phase shift resulting from the PP, whose thickness corresponds to a phase shift of π, has been confirmed using interference fringes displayed in the Ronchigram of a silicon single crystal specimen. The interference fringes were found to abruptly shift at the edge of the PP hole by π. Key words: phase plate, STEM, phase contrast, contrast transfer function, the principles of reciprocity, Ronchigram
Introduction Transmission electron microscopy (TEM) is a powerful technique for nanoscale imaging. However, it is difficult to obtain high-contrast images for biological materials because of the weak interaction between the electron waves and the specimens. Because the specimens are transparent for the high-energy electrons used in TEM, an electron wave passing through the specimen changes its phase, but its amplitude change is minimal. Therefore, in-focus images of such specimens show weak contrast. Phase-contrast TEM (P-TEM) could provide contrast enhancement for weakphase objects (such as biological materials) by converting
the amount of the phase shift induced by the specimens into an amount of intensity in the image [1]. Consequently, P-TEM gives more information about specimen structures than conventional TEM [1]. In P-TEM, several types of phase plates (PPs) have been proposed [2–8]. A thin-film PP is the most successful type among them, and most results in biology were obtained with thin-film PPs [1,7,8]. A Zernike PP (ZPP), a type of thin-film PP, has a hole in the center and is situated in the back focal plane (BFP) of the objective lens (OL) of the microscope. The ZPP provides a phase shift of π/2 to scattered waves, and the phase shift may be adjusted with the
© The Author 2015. Published by Oxford University Press on behalf of The Japanese Society of Microscopy. All rights reserved. For permissions, please e-mail: [email protected]
181
182
Fig. 1. Comparison of the optical conditions of (a) P-TEM and (b) P-STEM.
Figure 1 compares the optical ray diagrams for P-TEM and P-STEM. In P-STEM, a PP can be set in the condenser aperture (CA) plane, which is adjusted to be optically equivalent to the front focal plane (FFP) of an OL. In this report, we describe the setup of a P-STEM and the experimental results obtained from it.
Materials and methods The experiments in this study were performed with a JEOL JEM-2100F electron microscope, which is equipped with a 200-kV field emission electron gun. For P-STEM imaging, a ZPP made of carbon film was used. The carbon film PP was prepared by a similar procedure to that used for a P-TEM [9]. The ZPP has a small hole with a diameter of 5 μm. The thickness of the ZPP is ∼25 nm, which corresponds to a phase shift of π/2 for 200 kV electrons. We also prepare a PP with a π-phase-shift thickness. The PP has a circular center hole of 10 μm diameter. The holes were fabricated using a scanning focused ion beam microscope (JIB-4500) with a Ga ion source. The grids with the PPs were mounted on a standard CA holder situated at the CA plane and that is adjusted to be optically equivalent to the FFP of the OL, as shown in Fig. 2. In the normal STEM experiment, the FFP of the OL and the CA plane is not optically equivalent as the ray path shown in the left panel of Fig. 2. In our experiment, we
Fig. 2. Comparison of the optical conditions of C-STEM (left) and modified P-STEM. The CA plane is not optically conjugate to the detector plane (FFP of the OL) in STEM, but using a CM lens, the CA plane may be modified to be the conjugate to the detector plane in P-STEM.
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
film thickness and the mean inner potential of the material of the film. Another type of PP, an electrostatic one, consists of three layered electrodes. The central electrode is connected to a voltage source, and the other two are connected to the ground to prevent the leakage of the field applied with the center electrode. Each electrode has a hole for the unscattered electrons to pass through. This configuration forms an electrostatic potential that modifies the phase of the unscattered electrons passing through the holes of the stacked electrodes. In both the thin-film ZPPs and the electrostatic PPs, a cosine-type phase-contrast transfer function (PCTF) emerges when the electron waves, phase-shifted by the PP, interfere with the non-phase-shifted electron waves. Accordingly, these PPs enhance the image contrast especially at low spatial frequency, because the PCTF of cosine type is high at low frequency range. The OL pole piece design is restricted to a wide gap, because the PP is placed in the BFP of the OL in a P-TEM. Installation of a PP is difficult for an ultrahigh resolution OL that usually employs a narrow gap pole piece. Electrostatic charging of the PP is also an important problem to consider for a P-TEM. The size of the non-scattered beam, located at the center in the BFP of the OL, is very small. As a result the current density is very high at the center beam. The high-current-density electron beam incident onto the PP may cause electrostatic charging when we insert the PP into the optical axis or when errors occur in the placement of the PP. The charging causes image degradation because it gives an unintended phase shift to the electron waves. Therefore, this charging issue must be overcome for reproducible data acquisition and stable observation. Applying the principle of reciprocity to scanning transmission electron microscopy (STEM), STEM imaging optics is equivalent to the imaging optics in conventional TEM. Therefore in principle, phase-contrast STEM (P-STEM) can be used to enhance the phase contrast of phase objects.
Microscopy, 2015, Vol. 64, No. 3
Microscopy, 2015, Vol. 64, No. 3
Fig. 3. A schematic representation of the optical condition of the Ronchigram showing interference fringes. (a) and (b) show the optical conditions without and with the PP, respectively. The interference fringes shift owing to the effect of the PP around the point P corresponding to the center of the PP.
is a crossover of the incident electrons, and the virtual source is located at an extrapolated position of the diffracted electrons. Interference fringes can be observed in the overlapped areas. The lengths from the sources O and O0 to the screen are the same (L) as shown in Fig. 3. When the reflection condition was determined, the fringe spacing may also be determined from the phase difference arising from the difference of paths (|OP–O0 P|) from the real (O) and the virtual (O0 ) sources of electrons to a point (P) in the Ronchigram plane. In the case of P-STEM shown in Fig. 3b, where the ZPP was placed in the CA plane, the phase of the incident electrons is shifted by the ZPP except for those electrons passing through the center hole. Thus, the fringes in the circles shift depending on the amount of phase shift generated by the ZPP.
Results and discussion Figure 4 compares the experimental (d) and simulated (a–c) Ronchigrams from a silicon single crystal in which a phase shift is introduced by the PP placed in the CA plane. In the simulation, the values 200 kV, 16 mrad and 1 mm were used for the acceleration voltage of the electrons, the convergence angle of the incident beam and the spherical aberration coefficient (Cs) of the OL, respectively. These parameters were selected to match the experimental conditions. We ignored chromatic aberration in this calculation because the degradation of temporal coherence affects the resulting Ronchigrams minimally. Figure 4a is a Ronchigram with no PP and Fig. 4b and c are Ronchigrams with PPs giving phase shifts of π/2 and π, respectively. The detailed fringe patterns outside the hole in (a)–(c) is not the
Fig. 4. (a–c) Simulated and (d) experimental Ronchigrams showing the displacement of the silicon single crystal’s interference fringes, which results from the phase shift provided by the PP.
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
employed the ray path using the condenser mini (CM) lens shown in the right panel of Fig. 2, where the FFP of OL and the CA plane are optically equivalent. If the PP is located out of the FFP, the electron waves under the C-STEM and P-STEM conditions are mixed and the contrast enhancement could not be expected. With this configuration, a phase shift is given to the electrons except those passing through the center hole, before a specimen scatters the incident electrons. An OL with a focal length of 2.3 mm was used. This focal length allows the beam to make a focused spot on a specimen to perform STEM imaging. We performed an experiment to confirm a phase shift resulting from a PP in a P-STEM with Ronchigram observations of a Si [110] crystalline specimen. Figure 3a and b illustrates the schematics of the Ronchigram observations in conventional STEM (C-STEM) and P-STEM. In Ronchigrams of C-STEM using a crystalline specimen, we usually observe interference fringes in a leaf-shaped overlapping region of diffracted and transmitted disks. Because an appropriate camera length is chosen to obtain a Ronchigram, the diameter of the disks is determined with the incident convergence angle for a STEM probe and the defocus length for the Ronchigram. A crossover of the convergent incident wave (at a point O) is formed before/after the samples (before the sample in the illustration) and it diverges from/to the crossover to illuminate a specimen. The transmitted wave can be illustrated as a cone from a real source O. Electron waves diffracted by the sample with a reflection index can be illustrated as an inclined cone from a virtual source O0 . The real source
183
Microscopy, 2015, Vol. 64, No. 3
184
The pitch of the fringes on the screen is determined by a reflection (hkl) index that determines the reflection angle, a defocus length that determines the diameter of an illuminated area on the specimen and a camera length (sources-to-screen distance). The white circles shown in the simulated Ronchigrams of Fig. 4b and c correspond to shadows of a hole in a PP illuminated from the real (direct) and virtual (reflected) crossovers. In the experimental Ronchigram shown in Fig. 4d, the interference fringes are clearly observable as the simulated Ronchigrams. The thickness of the PP was selected to provide a phase shift of π for 200 kV, because the phase shift of π gives the largest fringe displacement and it is suitable for demonstration. The fringe displaces abruptly at the edge of the PP. The phase shift approximately corresponds to half of a period of the fringes (∼0.85π). Similar abrupt fringe displacements at the edge of the PP can also be observed in the discs of the diffracted waves. There is no local distortion in the interference pattern, which implies that there are no charged areas in the PP. This is the first result of the measurement of phase shift in a P-STEM.
Fig. 5. Comparison of the images of the amorphous carbon and their power spectra at the initial stage of the PP (a) and after 100 h (b). Plot (c) shows a comparison of the PCTFs of (a) and (b).
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
same as that in (d) because the simulation used simple model. However, the patterns inside the hole is the two beam interference fringes and the discussion below can be applicable. The bright and dark lines correspond to fringes formed by the interference between the transmitted wave and the [111] reflected wave. The transmitted (unscattered) waves and the [111] reflected waves constructively interfere when the difference of their optical paths (Δd) is (n + 1/2)λ (where n is an integer and λ is the wavelength of electrons) and destructively interfere when Δd = nλ. However, with a PP, the transmitted (unscattered) waves and the [111] reflected waves with a phase shift of π (π/2 phase shift) introduced by the PP constructively interfere when Δd = nλ (Δd = (n + 1/4)λ) and destructively interfere when Δd = (n + 1/2)λ (Δd = (n + 3/4)λ). Therefore, the phase shift introduced by the PP can be determined by measuring the shift of the fringes at the shadows of a hole of the PP. This is not affected either by the spherical aberration of the OL or coherency of the electron source because spherical aberration would change the size of the visible area of the interference fringe and coherency affects the contrast of the fringes.
Microscopy, 2015, Vol. 64, No. 3
There is no direct means of observation of the phase shift in a P-TEM and the only means of evaluating the state of the PP is to calculate a fast Fourier transform (FFT) of the image. Utilization of the Ronchigram is one of the advantages for the P-STEM, because we can check or measure the amount of phase shift before or after the phase-contrast imaging.
185
If the PP is partially charged, the interference fringes around the charged area are distorted in the Ronchigram in P-STEM. In contrast, in P-TEM, only the statistical electrostatic potential due to the charging from the distortion in the FFT pattern of the images may be estimated. An analysis of the Thon ring is also required to evaluate
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
Fig. 6. The C-STEM and P-STEM images are exhibited in (a) and (b), respectively, and their power spectra plotted in (a0 ) and (b0 ). Graph (c) shows a comparison of the PCTFs for the C-STEM (red (dark) lines) and P-STEM (blue (bright) lines). The calculated sine-type and cosine-type PCTFs are also shown as broken red (dark) and blue (bright) lines, respectively.
Microscopy, 2015, Vol. 64, No. 3
186
It also suggests that the image contrast of the phase object whose size is larger than several nanometers is enhanced in a P-STEM image. Solid lines in Fig. 6c show the power spectrum for the C-STEM [red (dark)] and P-STEM [blue (bright)] images. The calculated PCTF for the C-STEM (sine-type PCTF) and the P-STEM (cosine-type PCTF) are also represented as red and blue broken lines in the same image. The positions of the peaks in the power spectrum are consistent with those in the PCTFs. This clearly proves that P-STEM enhances contrast of the phase object and changes the PCTF from sine to cosine type.
Conclusions In the this article, we report on the first demonstration of P-STEM. P-STEM requires two major components: one is a probe-forming optical ray that makes the CA plane optically equivalent to the FFP of the OL, and the other is an imageforming ray that magnifies the diameter of the ZPP hole to a size larger than the STEM BF detector. Under the experimental conditions discussed in this article, the contrast enhancement of the phase object was evident. If ZPP can be applied to Cs-corrected STEM, the PCTF can be written as unity for almost all spatial frequency ranges in principle. This condition would be the best combination for highcontrast and high-resolution imaging in the STEM mode. Electrical charging was not found to be a serious issue in P-STEM, because no charging was observed for 100 h. The reason for this is that the current density irradiating onto the PP in P-STEM is approximately uniform over the entire PP, and there is no location with a particularly high electron density. In P-STEM, the Ronchigram can be used to evaluate the state of the PP, because it becomes possible to directly measure the phase shift of the PP from the Ronchigram using a crystal sample from the displacement in the interference fringes of direct and diffracted waves. The Ronchigram also shows the charging state from an observation of the bend interference fringe. Electron damage is considered to be a more serious issue in STEM than in TEM. So far, STEM has not been found suitable for observing biological molecules. However, STEM observations have been useful for wet specimens [12]. Moreover, a report indicated that electron damage is actually a less serious issue in STEM than in TEM [13]. These studies suggest the effectiveness of STEM for biological specimens. The use of P-STEM may thus open new frontiers in biological imaging.
Acknowledgements We greatly thank Mr Ohkura and Mr Ishikawa in JEOL for their support.
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
the phase shift. In P-STEM, the phase shift and the change in its state (resulting from electrostatic charging) can directly be obtained from an image of the interference fringes using the Ronchigram in addition to the FFT. The use of Ronchigrams to evaluate the condition of the PP is an important advantage of P-STEM. The power spectrum of the image for an amorphous carbon film showed almost no change during a 100 h electron irradiation as shown in Fig. 5 that demonstrates almost no change by the long period of electron irradiation. Thus, electronic charging is not as serious in P-STEM as it is in P-TEM. This is because the current density illuminated onto the PP is uniform in P-STEM and such an illumination does not result in the local charging of the PP, in spite of the fact that the total current is higher than that of the P-TEM. This is one of the big advantages in P-STEM [10]. Contamination of the PP is one of the origins of charging in P-TEM. The main sources of contamination are expected to be the samples. However, the distance from the samples to the CA, where the PP for P-STEM is placed, is much longer than that to the BFP of the OL, where the PP for P-TEM is located. This implies that the transfer lens system for anticontamination of the PP is unnecessary in P-STEM [11]. Figure 6a and b shows the bright field (BF) images of C-STEM and P-STEM for a sample consisting of amorphous carbon. The defocus values for both the images were the same (∼40 nm). The smallest angle (the cut-on angle) where the phase shift was added by the ZPP was determined by the diameter of the center hole in the ZPP and the focal length of the OL. It was 0.4 mrad in this study. Between the cut-on angle and the convergence angle limited by the CA, the PP imparts a certain phase shift to the incident electron wave. The thickness of the ZPP shown in Fig. 6b was set to provide a phase shift of π/2 for 200 kV electrons. The phase shift changes the sine-type phase-contrast transfer to a cosine-type one. The optical ray, in the image-forming lens system, was adjusted so that the collection angle by STEM BF detector is approximately equal to the cut-on angle. With this image forming lens settings, the P-STEM images could be obtained. The amorphous carbon film is thin enough to produce a phase contrast. Therefore, the image contrasts of these images are phase-originated and determined with the PCTF. Comparison of the observed images in Fig. 6a and b shows that contrast of a relatively large pattern of the amorphous film is enhanced in Fig. 6b. In contrast, the fine pattern of the film is clearer in Fig. 5a. This tendency is consistent with the changes in the PCTFs of the C-STEM and the P-STEM. Power spectra of both the images are shown in Fig. 6a0 and b0 . The intensity around the center peak in Fig. 6a0 is lower than that in Fig. 6b0 and the diameter of the bright ring in Fig. 6a0 is larger than that in 6b0 . This clearly shows the change in the PCTF by the PP: cosine-type to sine-type.
Microscopy, 2015, Vol. 64, No. 3
187
Funding This work was supported by SENTAN, JST.
7.
References 1. Danev R, Okawara H, Usuda N, Kametani K, Nagayama K (2002) A novel phase-contrast transmission electron microscopy producing high-contrast topographic images of weak objects. J. Biol. Phys. 28: 627–635. 2. Huang S-H, Wang W-J, Chang C-S, Hwu Y-K, Tseng F-G, Kai J-J, Chen F-R (2007) The fabrication and application of Zernike electrostatic phase plate. J. Electron Microsc. 55: 273–280.
9.
10.
11.
4. Cuambie R, Downing K H, Typke D, Glaeser R M, Jin J (2007) Design of a microfabricated, two-electrode phase-contrast element suitable for electron microscopy. Ultramicroscopy 107: 329–339.
12.
5. Matsumoto T, Tonomura A (1996) The phase constancy of electron waves traveling through Boersch’s electrostatic phase plate. Ultramicroscopy 63: 5–10.
13.
6. Buijsse B, van Laarhoven F M H M, Schmid A K, Cambie R, Cabrini S, Jin J, Glaeser R M (2011) Design of a hybrid
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
3. Majorovits E, Barton B, Schultheiß K, Pérez-Willard F, Gerthsen D, Schröder R R (2007) Optimizing phase contrast in transmission electron microscopy with an electrostatic (Boersch) phase plate. Ultramicroscopy 107: 213–226.
8.
double-sideband/single-sideband (schlieren) objective aperture suitable for electron microscopy. Ultramicroscopy 111: 1688–1695. Minoda H, Okabe T, Iijima H (2011) Contrast enhancement in the phase plate transmission electron microscopy using an objective lens with a long focal length. J. Electron Microsc. 60: 337–343. Inayoshi Y, Minoda H, Arai Y, Nagayama K (2012) Direct observation of biological molecules in liquid by environmental phaseplate transmission electron microscopy. Micron 43: 1091–1098. Danev R, Nagayama K (2001) Complex observation in electron microscopy. II. Direct visualization of phase and amplitudes of exit wave functions. J. Phys. Soc. Jpn. 70: 696–702. Danev R, Gleaser R M, Nagayama K (2009) Practical factors affecting the performance of a thin-film phase plate for transmission electron microscopy. Ultramicroscopy 109: 312–325. Hosokawa F, Danev R, Arai Y, Nagayama K (2005) Transfer doublet and an elaborated phase plate holder for 120 kV electronphase microscope. J. Electron Microsc. 54: 317–324. Aoyama K, Takagi T, Hirase A, Miyazawa A (2008) STEM tomography for thick biological specimens. Ultramicroscopy 109: 70–80. Evans J E, Jungjohann K L, Wong Peony C K, Chiu P-L, Dutrow G H, Arslan I, Browning N D (2012) Visualizing macromolecular complexes with in situ liquid scanning transmission electron microscopy. Micron 43: 1085–1090.
Article
Phase-contrast scanning transmission electron microscopy Hiroki Minoda1,*, Takayuki Tamai1, Hirofumi Iijima2, Fumio Hosokawa2, and Yukihito Kondo2 Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
1
Department of Applied Physics, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan, and 2JEOL Ltd, 3-1-2 Musashino, Akishima, Tokyo 196-8558, Japan
*To whom correspondence should be addressed. E-mail: [email protected] Received 17 October 2014; Accepted 11 February 2015
Abstract This report introduces the first results obtained using phase-contrast scanning transmission electron microscopy (P-STEM). A carbon-film phase plate (PP) with a small center hole is placed in the condenser aperture plane so that a phase shift is introduced in the incident electron waves except those passing through the center hole. A cosine-type phase-contrast transfer function emerges when the phase-shifted scattered waves interfere with the non-phase-shifted unscattered waves, which passed through the center hole before incidence onto the specimen. The phase contrast resulting in P-STEM is optically identical to that in phase-contrast transmission electron microscopy that is used to provide high contrast for weak phase objects. Therefore, the use of PPs can enhance the phase contrast of the STEM images of specimens in principle. The phase shift resulting from the PP, whose thickness corresponds to a phase shift of π, has been confirmed using interference fringes displayed in the Ronchigram of a silicon single crystal specimen. The interference fringes were found to abruptly shift at the edge of the PP hole by π. Key words: phase plate, STEM, phase contrast, contrast transfer function, the principles of reciprocity, Ronchigram
Introduction Transmission electron microscopy (TEM) is a powerful technique for nanoscale imaging. However, it is difficult to obtain high-contrast images for biological materials because of the weak interaction between the electron waves and the specimens. Because the specimens are transparent for the high-energy electrons used in TEM, an electron wave passing through the specimen changes its phase, but its amplitude change is minimal. Therefore, in-focus images of such specimens show weak contrast. Phase-contrast TEM (P-TEM) could provide contrast enhancement for weakphase objects (such as biological materials) by converting
the amount of the phase shift induced by the specimens into an amount of intensity in the image [1]. Consequently, P-TEM gives more information about specimen structures than conventional TEM [1]. In P-TEM, several types of phase plates (PPs) have been proposed [2–8]. A thin-film PP is the most successful type among them, and most results in biology were obtained with thin-film PPs [1,7,8]. A Zernike PP (ZPP), a type of thin-film PP, has a hole in the center and is situated in the back focal plane (BFP) of the objective lens (OL) of the microscope. The ZPP provides a phase shift of π/2 to scattered waves, and the phase shift may be adjusted with the
© The Author 2015. Published by Oxford University Press on behalf of The Japanese Society of Microscopy. All rights reserved. For permissions, please e-mail: [email protected]
181
182
Fig. 1. Comparison of the optical conditions of (a) P-TEM and (b) P-STEM.
Figure 1 compares the optical ray diagrams for P-TEM and P-STEM. In P-STEM, a PP can be set in the condenser aperture (CA) plane, which is adjusted to be optically equivalent to the front focal plane (FFP) of an OL. In this report, we describe the setup of a P-STEM and the experimental results obtained from it.
Materials and methods The experiments in this study were performed with a JEOL JEM-2100F electron microscope, which is equipped with a 200-kV field emission electron gun. For P-STEM imaging, a ZPP made of carbon film was used. The carbon film PP was prepared by a similar procedure to that used for a P-TEM [9]. The ZPP has a small hole with a diameter of 5 μm. The thickness of the ZPP is ∼25 nm, which corresponds to a phase shift of π/2 for 200 kV electrons. We also prepare a PP with a π-phase-shift thickness. The PP has a circular center hole of 10 μm diameter. The holes were fabricated using a scanning focused ion beam microscope (JIB-4500) with a Ga ion source. The grids with the PPs were mounted on a standard CA holder situated at the CA plane and that is adjusted to be optically equivalent to the FFP of the OL, as shown in Fig. 2. In the normal STEM experiment, the FFP of the OL and the CA plane is not optically equivalent as the ray path shown in the left panel of Fig. 2. In our experiment, we
Fig. 2. Comparison of the optical conditions of C-STEM (left) and modified P-STEM. The CA plane is not optically conjugate to the detector plane (FFP of the OL) in STEM, but using a CM lens, the CA plane may be modified to be the conjugate to the detector plane in P-STEM.
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
film thickness and the mean inner potential of the material of the film. Another type of PP, an electrostatic one, consists of three layered electrodes. The central electrode is connected to a voltage source, and the other two are connected to the ground to prevent the leakage of the field applied with the center electrode. Each electrode has a hole for the unscattered electrons to pass through. This configuration forms an electrostatic potential that modifies the phase of the unscattered electrons passing through the holes of the stacked electrodes. In both the thin-film ZPPs and the electrostatic PPs, a cosine-type phase-contrast transfer function (PCTF) emerges when the electron waves, phase-shifted by the PP, interfere with the non-phase-shifted electron waves. Accordingly, these PPs enhance the image contrast especially at low spatial frequency, because the PCTF of cosine type is high at low frequency range. The OL pole piece design is restricted to a wide gap, because the PP is placed in the BFP of the OL in a P-TEM. Installation of a PP is difficult for an ultrahigh resolution OL that usually employs a narrow gap pole piece. Electrostatic charging of the PP is also an important problem to consider for a P-TEM. The size of the non-scattered beam, located at the center in the BFP of the OL, is very small. As a result the current density is very high at the center beam. The high-current-density electron beam incident onto the PP may cause electrostatic charging when we insert the PP into the optical axis or when errors occur in the placement of the PP. The charging causes image degradation because it gives an unintended phase shift to the electron waves. Therefore, this charging issue must be overcome for reproducible data acquisition and stable observation. Applying the principle of reciprocity to scanning transmission electron microscopy (STEM), STEM imaging optics is equivalent to the imaging optics in conventional TEM. Therefore in principle, phase-contrast STEM (P-STEM) can be used to enhance the phase contrast of phase objects.
Microscopy, 2015, Vol. 64, No. 3
Microscopy, 2015, Vol. 64, No. 3
Fig. 3. A schematic representation of the optical condition of the Ronchigram showing interference fringes. (a) and (b) show the optical conditions without and with the PP, respectively. The interference fringes shift owing to the effect of the PP around the point P corresponding to the center of the PP.
is a crossover of the incident electrons, and the virtual source is located at an extrapolated position of the diffracted electrons. Interference fringes can be observed in the overlapped areas. The lengths from the sources O and O0 to the screen are the same (L) as shown in Fig. 3. When the reflection condition was determined, the fringe spacing may also be determined from the phase difference arising from the difference of paths (|OP–O0 P|) from the real (O) and the virtual (O0 ) sources of electrons to a point (P) in the Ronchigram plane. In the case of P-STEM shown in Fig. 3b, where the ZPP was placed in the CA plane, the phase of the incident electrons is shifted by the ZPP except for those electrons passing through the center hole. Thus, the fringes in the circles shift depending on the amount of phase shift generated by the ZPP.
Results and discussion Figure 4 compares the experimental (d) and simulated (a–c) Ronchigrams from a silicon single crystal in which a phase shift is introduced by the PP placed in the CA plane. In the simulation, the values 200 kV, 16 mrad and 1 mm were used for the acceleration voltage of the electrons, the convergence angle of the incident beam and the spherical aberration coefficient (Cs) of the OL, respectively. These parameters were selected to match the experimental conditions. We ignored chromatic aberration in this calculation because the degradation of temporal coherence affects the resulting Ronchigrams minimally. Figure 4a is a Ronchigram with no PP and Fig. 4b and c are Ronchigrams with PPs giving phase shifts of π/2 and π, respectively. The detailed fringe patterns outside the hole in (a)–(c) is not the
Fig. 4. (a–c) Simulated and (d) experimental Ronchigrams showing the displacement of the silicon single crystal’s interference fringes, which results from the phase shift provided by the PP.
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
employed the ray path using the condenser mini (CM) lens shown in the right panel of Fig. 2, where the FFP of OL and the CA plane are optically equivalent. If the PP is located out of the FFP, the electron waves under the C-STEM and P-STEM conditions are mixed and the contrast enhancement could not be expected. With this configuration, a phase shift is given to the electrons except those passing through the center hole, before a specimen scatters the incident electrons. An OL with a focal length of 2.3 mm was used. This focal length allows the beam to make a focused spot on a specimen to perform STEM imaging. We performed an experiment to confirm a phase shift resulting from a PP in a P-STEM with Ronchigram observations of a Si [110] crystalline specimen. Figure 3a and b illustrates the schematics of the Ronchigram observations in conventional STEM (C-STEM) and P-STEM. In Ronchigrams of C-STEM using a crystalline specimen, we usually observe interference fringes in a leaf-shaped overlapping region of diffracted and transmitted disks. Because an appropriate camera length is chosen to obtain a Ronchigram, the diameter of the disks is determined with the incident convergence angle for a STEM probe and the defocus length for the Ronchigram. A crossover of the convergent incident wave (at a point O) is formed before/after the samples (before the sample in the illustration) and it diverges from/to the crossover to illuminate a specimen. The transmitted wave can be illustrated as a cone from a real source O. Electron waves diffracted by the sample with a reflection index can be illustrated as an inclined cone from a virtual source O0 . The real source
183
Microscopy, 2015, Vol. 64, No. 3
184
The pitch of the fringes on the screen is determined by a reflection (hkl) index that determines the reflection angle, a defocus length that determines the diameter of an illuminated area on the specimen and a camera length (sources-to-screen distance). The white circles shown in the simulated Ronchigrams of Fig. 4b and c correspond to shadows of a hole in a PP illuminated from the real (direct) and virtual (reflected) crossovers. In the experimental Ronchigram shown in Fig. 4d, the interference fringes are clearly observable as the simulated Ronchigrams. The thickness of the PP was selected to provide a phase shift of π for 200 kV, because the phase shift of π gives the largest fringe displacement and it is suitable for demonstration. The fringe displaces abruptly at the edge of the PP. The phase shift approximately corresponds to half of a period of the fringes (∼0.85π). Similar abrupt fringe displacements at the edge of the PP can also be observed in the discs of the diffracted waves. There is no local distortion in the interference pattern, which implies that there are no charged areas in the PP. This is the first result of the measurement of phase shift in a P-STEM.
Fig. 5. Comparison of the images of the amorphous carbon and their power spectra at the initial stage of the PP (a) and after 100 h (b). Plot (c) shows a comparison of the PCTFs of (a) and (b).
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
same as that in (d) because the simulation used simple model. However, the patterns inside the hole is the two beam interference fringes and the discussion below can be applicable. The bright and dark lines correspond to fringes formed by the interference between the transmitted wave and the [111] reflected wave. The transmitted (unscattered) waves and the [111] reflected waves constructively interfere when the difference of their optical paths (Δd) is (n + 1/2)λ (where n is an integer and λ is the wavelength of electrons) and destructively interfere when Δd = nλ. However, with a PP, the transmitted (unscattered) waves and the [111] reflected waves with a phase shift of π (π/2 phase shift) introduced by the PP constructively interfere when Δd = nλ (Δd = (n + 1/4)λ) and destructively interfere when Δd = (n + 1/2)λ (Δd = (n + 3/4)λ). Therefore, the phase shift introduced by the PP can be determined by measuring the shift of the fringes at the shadows of a hole of the PP. This is not affected either by the spherical aberration of the OL or coherency of the electron source because spherical aberration would change the size of the visible area of the interference fringe and coherency affects the contrast of the fringes.
Microscopy, 2015, Vol. 64, No. 3
There is no direct means of observation of the phase shift in a P-TEM and the only means of evaluating the state of the PP is to calculate a fast Fourier transform (FFT) of the image. Utilization of the Ronchigram is one of the advantages for the P-STEM, because we can check or measure the amount of phase shift before or after the phase-contrast imaging.
185
If the PP is partially charged, the interference fringes around the charged area are distorted in the Ronchigram in P-STEM. In contrast, in P-TEM, only the statistical electrostatic potential due to the charging from the distortion in the FFT pattern of the images may be estimated. An analysis of the Thon ring is also required to evaluate
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
Fig. 6. The C-STEM and P-STEM images are exhibited in (a) and (b), respectively, and their power spectra plotted in (a0 ) and (b0 ). Graph (c) shows a comparison of the PCTFs for the C-STEM (red (dark) lines) and P-STEM (blue (bright) lines). The calculated sine-type and cosine-type PCTFs are also shown as broken red (dark) and blue (bright) lines, respectively.
Microscopy, 2015, Vol. 64, No. 3
186
It also suggests that the image contrast of the phase object whose size is larger than several nanometers is enhanced in a P-STEM image. Solid lines in Fig. 6c show the power spectrum for the C-STEM [red (dark)] and P-STEM [blue (bright)] images. The calculated PCTF for the C-STEM (sine-type PCTF) and the P-STEM (cosine-type PCTF) are also represented as red and blue broken lines in the same image. The positions of the peaks in the power spectrum are consistent with those in the PCTFs. This clearly proves that P-STEM enhances contrast of the phase object and changes the PCTF from sine to cosine type.
Conclusions In the this article, we report on the first demonstration of P-STEM. P-STEM requires two major components: one is a probe-forming optical ray that makes the CA plane optically equivalent to the FFP of the OL, and the other is an imageforming ray that magnifies the diameter of the ZPP hole to a size larger than the STEM BF detector. Under the experimental conditions discussed in this article, the contrast enhancement of the phase object was evident. If ZPP can be applied to Cs-corrected STEM, the PCTF can be written as unity for almost all spatial frequency ranges in principle. This condition would be the best combination for highcontrast and high-resolution imaging in the STEM mode. Electrical charging was not found to be a serious issue in P-STEM, because no charging was observed for 100 h. The reason for this is that the current density irradiating onto the PP in P-STEM is approximately uniform over the entire PP, and there is no location with a particularly high electron density. In P-STEM, the Ronchigram can be used to evaluate the state of the PP, because it becomes possible to directly measure the phase shift of the PP from the Ronchigram using a crystal sample from the displacement in the interference fringes of direct and diffracted waves. The Ronchigram also shows the charging state from an observation of the bend interference fringe. Electron damage is considered to be a more serious issue in STEM than in TEM. So far, STEM has not been found suitable for observing biological molecules. However, STEM observations have been useful for wet specimens [12]. Moreover, a report indicated that electron damage is actually a less serious issue in STEM than in TEM [13]. These studies suggest the effectiveness of STEM for biological specimens. The use of P-STEM may thus open new frontiers in biological imaging.
Acknowledgements We greatly thank Mr Ohkura and Mr Ishikawa in JEOL for their support.
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
the phase shift. In P-STEM, the phase shift and the change in its state (resulting from electrostatic charging) can directly be obtained from an image of the interference fringes using the Ronchigram in addition to the FFT. The use of Ronchigrams to evaluate the condition of the PP is an important advantage of P-STEM. The power spectrum of the image for an amorphous carbon film showed almost no change during a 100 h electron irradiation as shown in Fig. 5 that demonstrates almost no change by the long period of electron irradiation. Thus, electronic charging is not as serious in P-STEM as it is in P-TEM. This is because the current density illuminated onto the PP is uniform in P-STEM and such an illumination does not result in the local charging of the PP, in spite of the fact that the total current is higher than that of the P-TEM. This is one of the big advantages in P-STEM [10]. Contamination of the PP is one of the origins of charging in P-TEM. The main sources of contamination are expected to be the samples. However, the distance from the samples to the CA, where the PP for P-STEM is placed, is much longer than that to the BFP of the OL, where the PP for P-TEM is located. This implies that the transfer lens system for anticontamination of the PP is unnecessary in P-STEM [11]. Figure 6a and b shows the bright field (BF) images of C-STEM and P-STEM for a sample consisting of amorphous carbon. The defocus values for both the images were the same (∼40 nm). The smallest angle (the cut-on angle) where the phase shift was added by the ZPP was determined by the diameter of the center hole in the ZPP and the focal length of the OL. It was 0.4 mrad in this study. Between the cut-on angle and the convergence angle limited by the CA, the PP imparts a certain phase shift to the incident electron wave. The thickness of the ZPP shown in Fig. 6b was set to provide a phase shift of π/2 for 200 kV electrons. The phase shift changes the sine-type phase-contrast transfer to a cosine-type one. The optical ray, in the image-forming lens system, was adjusted so that the collection angle by STEM BF detector is approximately equal to the cut-on angle. With this image forming lens settings, the P-STEM images could be obtained. The amorphous carbon film is thin enough to produce a phase contrast. Therefore, the image contrasts of these images are phase-originated and determined with the PCTF. Comparison of the observed images in Fig. 6a and b shows that contrast of a relatively large pattern of the amorphous film is enhanced in Fig. 6b. In contrast, the fine pattern of the film is clearer in Fig. 5a. This tendency is consistent with the changes in the PCTFs of the C-STEM and the P-STEM. Power spectra of both the images are shown in Fig. 6a0 and b0 . The intensity around the center peak in Fig. 6a0 is lower than that in Fig. 6b0 and the diameter of the bright ring in Fig. 6a0 is larger than that in 6b0 . This clearly shows the change in the PCTF by the PP: cosine-type to sine-type.
Microscopy, 2015, Vol. 64, No. 3
187
Funding This work was supported by SENTAN, JST.
7.
References 1. Danev R, Okawara H, Usuda N, Kametani K, Nagayama K (2002) A novel phase-contrast transmission electron microscopy producing high-contrast topographic images of weak objects. J. Biol. Phys. 28: 627–635. 2. Huang S-H, Wang W-J, Chang C-S, Hwu Y-K, Tseng F-G, Kai J-J, Chen F-R (2007) The fabrication and application of Zernike electrostatic phase plate. J. Electron Microsc. 55: 273–280.
9.
10.
11.
4. Cuambie R, Downing K H, Typke D, Glaeser R M, Jin J (2007) Design of a microfabricated, two-electrode phase-contrast element suitable for electron microscopy. Ultramicroscopy 107: 329–339.
12.
5. Matsumoto T, Tonomura A (1996) The phase constancy of electron waves traveling through Boersch’s electrostatic phase plate. Ultramicroscopy 63: 5–10.
13.
6. Buijsse B, van Laarhoven F M H M, Schmid A K, Cambie R, Cabrini S, Jin J, Glaeser R M (2011) Design of a hybrid
Downloaded from http://jmicro.oxfordjournals.org/ at Dalhousie University on June 19, 2015
3. Majorovits E, Barton B, Schultheiß K, Pérez-Willard F, Gerthsen D, Schröder R R (2007) Optimizing phase contrast in transmission electron microscopy with an electrostatic (Boersch) phase plate. Ultramicroscopy 107: 213–226.
8.
double-sideband/single-sideband (schlieren) objective aperture suitable for electron microscopy. Ultramicroscopy 111: 1688–1695. Minoda H, Okabe T, Iijima H (2011) Contrast enhancement in the phase plate transmission electron microscopy using an objective lens with a long focal length. J. Electron Microsc. 60: 337–343. Inayoshi Y, Minoda H, Arai Y, Nagayama K (2012) Direct observation of biological molecules in liquid by environmental phaseplate transmission electron microscopy. Micron 43: 1091–1098. Danev R, Nagayama K (2001) Complex observation in electron microscopy. II. Direct visualization of phase and amplitudes of exit wave functions. J. Phys. Soc. Jpn. 70: 696–702. Danev R, Gleaser R M, Nagayama K (2009) Practical factors affecting the performance of a thin-film phase plate for transmission electron microscopy. Ultramicroscopy 109: 312–325. Hosokawa F, Danev R, Arai Y, Nagayama K (2005) Transfer doublet and an elaborated phase plate holder for 120 kV electronphase microscope. J. Electron Microsc. 54: 317–324. Aoyama K, Takagi T, Hirase A, Miyazawa A (2008) STEM tomography for thick biological specimens. Ultramicroscopy 109: 70–80. Evans J E, Jungjohann K L, Wong Peony C K, Chiu P-L, Dutrow G H, Arslan I, Browning N D (2012) Visualizing macromolecular complexes with in situ liquid scanning transmission electron microscopy. Micron 43: 1085–1090.
