TY - JOUR
AU - Nagayama, Kuniaki
AB - Abstract It has been six decades since the concept of phase-plate electron microscopy was first reported by Boersch, but an experimental report on a phase plate with a theoretically rational performance has only recently been released by a group including the present author. Currently, many laboratories around the world are attempting to develop a wide range of phase plates to enhance the capabilities of transmission electron microscopy. They are reporting not only advantages of their own developments but also a fundamental problem inherent to electron beam devices, namely charging, i.e. the accumulation of electrostatic charge. In this report, we review the 60-year history of phase-plate development, with a particular focus on the fundamental issue of phase-plate charging. Next, we review biological applications of qualified phase plates, which have been successful in avoiding charging to some extent. Finally, we compare and discuss electron microscopic images, taken with or without phase plates, of biological targets such as proteins (GroEL and TRPV4), protein complexes (flagellar motor), viruses (T4 phage, ɛ-15 phage and herpes simplex virus), bacterial (cyanobacteria) and mammalian (PtK2) cells. phase contrast, phase plate, Zernike, single particle, tomography 1. Introduction Objects thin enough for observation with an electron microscope (EM) are phase objects, which change the phase of electron incident wave rather than its amplitude. No matter what type of microscope is used, whether it is a light microscope (LM) or an EM, the process of visualization of phase objects can be described using a minimum microscopic model that assumes only coherent illumination with the incident beam, elastic scattering of incident waves and phase disturbance arising from the phase objects. In a scalar theory of image formation, the spatially dependent phase disturbance by an object, θ(r), exerted on the incident wave appears in a two-dimensional transparency function as eiθ(r). In electron microscopy, θ(r) may be related to the constituent atomic potential of the object. On the basis of wave optics, the formation of the corresponding image with electron waves can be formulated as shown in the top trace of Fig. 1, which schematized a mathematical interpretation of Abbe’s notion for the image formation in microscopy, namely two successive Fourier transform (FT) for two successive diffractions [1]. In this scheme, phase plates are symbolized as a spatial filter that modulates diffracted electron waves in a spatial-frequency-dependent manner at the back focal plane of an objective lens. Fig. 1. View largeDownload slide The image formation process in phase contrast microscopy. (a) A minimum model for microscopic image formation. 2DFT, two-dimensional (2D) Fourier transform (FT); FT[f], FT operation for a function f; H(k), spatial filter; 2DFT−1, inverse 2DFT; FT−1[f], inverse FT operation for a function f; and *, convolution. (b) A spatial filter schematic for defocus phase contrast (DPC). (c) A spatial filter schematic for Zernike phase contrast (ZPC). (d) A spatial filter schematic for Schlieren optics (SO) or single-sideband imaging (SSI). (e) A spatial filter schematic for Hilbert differential contrast (HDC). The bottom trace represents a typical contrast transfer function (CTF), reflecting each of the spatial filter functions. |k|, magnitude of wave number vector (two-dimensional spatial frequency); kx, xth component of wave number vector. Fig. 1. View largeDownload slide The image formation process in phase contrast microscopy. (a) A minimum model for microscopic image formation. 2DFT, two-dimensional (2D) Fourier transform (FT); FT[f], FT operation for a function f; H(k), spatial filter; 2DFT−1, inverse 2DFT; FT−1[f], inverse FT operation for a function f; and *, convolution. (b) A spatial filter schematic for defocus phase contrast (DPC). (c) A spatial filter schematic for Zernike phase contrast (ZPC). (d) A spatial filter schematic for Schlieren optics (SO) or single-sideband imaging (SSI). (e) A spatial filter schematic for Hilbert differential contrast (HDC). The bottom trace represents a typical contrast transfer function (CTF), reflecting each of the spatial filter functions. |k|, magnitude of wave number vector (two-dimensional spatial frequency); kx, xth component of wave number vector. With this minimum microscopic model, the principle of phase contrast for pure phase objects can be simply described as follows: 1. The exit-wave function at the exit plane immediately after the phase object is described by the transparency function itself, eiθ(r),when a plane wave is employed as the incident. The phase object can scatter the incident wave with no absorbance and hence appear to be transparent because there is no change in the intensity of the transmitted incident. A mathematical abstraction is shown in Eq. (1), where images obtained with an intensity detection (|·|2) of such an exit-wave (eiθ(r)) become featureless, mathematically expressed as follows:   (1) 2. Phase-contrast microscopy can generate a detailed image with contrast by modulating transmitted waves with an optical device, spatial filter H(k), which can be inserted at the back focal plane of an objective lens (Fig. 1a). The argument k is a wave number vector, i.e. a two-dimensional wave number defined in a spatial frequency space. The image with phase contrast is expressed as   (2) where FT−1 represents the inverse Fourier transform and * a convolution. Equation (2) represents a general definition for phase contrast. According to the choice of spatial filters, four types of phase-plate schemes can be devised (Fig. 1b–e): defocus phase contrast (DPC), Zernike phase contrast (ZPC), Schlieren optics (SO) and Hilbert differential contrast (HDC). For the case of weak phase objects (|θ(r)| << 1 or sin θ(r) ∼ θ(r)), the image contrast generated on a screen after detection (Eq. 2, left-hand side) can be simplified under the assumption that terms higher than second order of θ(r) can be ignored [1].   (3)  (4) where γ(k) represents a frequency-dependent phase retardance reflecting the lens conditions, such as the defocus setting and lens aberrations. A simple version of γ(k) can be expressed as   (5) where Δz is the defocus, λ is the wave length and Cs is the spherical aberration coefficient of the objective lens. The first term, 1, in expression (3) or (4) corresponds to a bright-field background, and the second term denotes the phase contrast. For weak phase objects, the ideal phase contrast must be θ(r) itself. But as clearly evident from expression (3) or (4), the phase contrast is heavily modulated by a function of sin γ(k) or cos γ(k). Traditionally, the former is called the ‘phase-contrast transfer function (CTF)’ and the latter is called the ‘amplitude CTF’. A typical form of the CTF corresponding to each of the four phase plates is shown in the bottom trace of Fig. 1. CTFs are defined in a spatial frequency space (k), and can determine the fate of waves propagated within a microscope in a spatial-frequency-dependent manner. Lower frequency components of transmitted waves determine the overall features of formed images, whereas higher frequency components determine the detailed internal morphology. Due to the nature of the sine function, the phase-CTF sin γ(k) fails to transmit lower frequency components, and DPC images lose features, resulting in lowered contrast. In the case of the cosine CTF, cos γ(k), images appear the other way around, and ZPC images can maintain a higher contrast. Like phase plates widely utilized in light microscopy, phase plates for EM appear to be simple and easy to make. In reality, however, construction of a functional phase plate is no simple matter. Manufacturing the phase plate itself is not severely demanding, particularly nowadays when researchers have easy access to sophisticated nanotechnology, but creating a functional one inside an EM has been suffering from a single hazard, phase-plate charging, for more than half a century. Technological evolution, like biological evolution, often follows a road of twists and turns; we often see the extinction of one technology and the sudden emergence of another. The development of phase plates over last 60 years seems to be one example of such a tortuous evolutionary path. Section 2 summarizes the history of the past 60 years, with a particular focus on the development of the thin-film type of phase plates. Section 3 describes the emergence of a functional thin-film phase plate and a technological pitfall that prevents the stable supply of such a plate. Section 4 investigates to what extent image quality is improved in biological EM when qualified phase plates are used. Section 5 describes promising applications of phase plates to organic systems other than biological ones. Section 6 surveys various phase plates recently studied as the next-generation thin-film phase plates. Section 7 describes the future of phase plates and phase-plate EM. 2. 60-Year history of phase-plate development In 1947, ∼10 years after Zernike’s invention of phase-contrast LM [2], Boersch published his ideas on phase-contrast EM, considering two types of phase plates [3] (refer to Fig. 2a). His proposal led to an early prototype for the thin-film or electrostatic type of phase plate studied worldwide at present. At that moment, his idea could not be fully tested, partly because of the difficulty in engineering of phase plates, but mainly because of the advent of other types of contrast schemes such as scattering contrast and DPC that seemed to be technically less demanding (refer to Fig. 1b). Scattering contrast was easily realized by setting an aperture stop with a tiny hole at the back-focal plane; this method became a standard, specifically in biology, when combined with heavy metal staining techniques. DPC became a standard especially in the materials science [4]. The first experimental result with a Zernike phase plate was reported in 1958, ∼10 years after the Boersch proposal, by a Japanese group. In that study, a carbon film with a hole diameter of 20 µm was used as a thin-film-type phase plate [5]. As shown by the photographs in Fig. 2b, the resulting image recorded, of a specimen of a plastic film, did not remarkably improve contrast compared with conventional EM. However, their efforts must be admired as a first experimental effort, but were doomed to fail due to the poor design of the phase plate they employed. The hole size in the center of the Zernike phase plate (refer to Fig. 1c), 20 µm, is too large to recover the lower frequency components responsible for high contrast. This size corresponds to a cut-on frequency (the lowest recoverable frequency due to the Zernike phase plate) of ≥5 nm−1, i.e. almost no change in contrast, particularly for an object with geometry >1 µm (refer to Fig. 2b). Fig. 2. View largeDownload slide Phase plates and experimental results reported in the early days of phase plate development. (a) Boersch’s proposal for two types of phase plate [Figures 6a, 6b reproduced from the article published in: Z.Naturforsch. 2 a, 615 (1947) see ref. 3], (b) The first experimental results, with (1) and without (2) a Zernike phase plate made from a carbon thin film [Reprinted with permission from Kanaya K, Kawakatsu H, Ito K, and Yotsumoto H. Experiment on the electron phase microscope. J. Appl. Phys. 29: 1046–1049. Copyright 1958, American Institute of Physics. from Fig. 2 of ref. 5], (c) Unwin’s proposal for a wire-type electrostatic phase plate, and experimental results with (2) and without (3) the electrostatic phase plate [figure reproduced from Figs. 1 and 7 of ref. 8]. Fig. 2. View largeDownload slide Phase plates and experimental results reported in the early days of phase plate development. (a) Boersch’s proposal for two types of phase plate [Figures 6a, 6b reproduced from the article published in: Z.Naturforsch. 2 a, 615 (1947) see ref. 3], (b) The first experimental results, with (1) and without (2) a Zernike phase plate made from a carbon thin film [Reprinted with permission from Kanaya K, Kawakatsu H, Ito K, and Yotsumoto H. Experiment on the electron phase microscope. J. Appl. Phys. 29: 1046–1049. Copyright 1958, American Institute of Physics. from Fig. 2 of ref. 5], (c) Unwin’s proposal for a wire-type electrostatic phase plate, and experimental results with (2) and without (3) the electrostatic phase plate [figure reproduced from Figs. 1 and 7 of ref. 8]. This research direction within electron microscopy had several problems: not only the engineering challenge of making a fine phase plate with a smaller central hole, but also the issue of phase-plate charging. Taking into account the problem of severe charging due to contamination by organic material, Faget et al. employed a heating holder for their phase plate [6]; however, they could not confirm an improvement in contrast. In the 1970s, after the dawn years, one could observe the EM community entering a boom of phase-plate development. A slightly improved contrast was reported with a Zernike phase plate made of a carbon film with a hole of 10 µm diameter [7]. Unwin reported an experimental result using a wire-type electrostatic phase plate with a 0.3-µm width (Fig. 2c); however, once again, contrast improvement could not be confirmed, as illustrated in Fig. 2c(1) and c(2) [8]. A similar type of electrostatic phase plate was tested later by Krakow and Siegel, again with a little success in contrast improvement [9]. By using a Zernike phase plate with a smaller hole (3 µm in diameter), Johnson and Parson reported a moderately improved contrast for unstained biological materials [10]. As an extension from the simple design of the Zernike phase plate, a profiled phase plate with many concentric circular gaps was tested in an attempt to correct the spherical aberration and simultaneously confer phase contrast, but the expected effect was not confirmed [11]. After the last study published during the boom, in which Ballosier and Bonnet tested a wire-type electrostatic phase plate [12], a long silence of ∼20 years followed. We could find only two theoretical reports on the Zernike phase plate published during this silent period. Both were published by a single group at Hitachi Co., directed by Tonomura. In these papers, the authors investigate the idea of a magnetic type of Zernike phase plate that would exploit the Aharonov–Bohm (AB) effect. The concept was filed as a patent without experimental verification; in a theoretical paper, the authors calculated the potential profile for a genuine Boersch phase plate made from an Einzel lens (refer to Fig. 13b) [13]. To survey the history of phase-plate development, including its revival after a 20-year silence, a chronological table of the history is illustrated in Fig. 3. This table may not be complete, and may be much biased to emphasize the works reported from the author’s group. Regardless, however, there is clear evidence that there have been two booms in the phase-plate development; one in the 1970s and one in the 2000s. Fig. 3. View largeDownload slide Sixty-year history of phase-plate development in electron microscopy. Fig. 3. View largeDownload slide Sixty-year history of phase-plate development in electron microscopy. 3. Functional phase plates and phase-plate charging As shown in Fig. 3, the first boom in phase-plate development in the 1970s suddenly ceased, followed by a long silence. Such technological species as the thin-film type of phase plate simply died out. One major cause of this extinction among various conceivable causes may be the poor image quality attainable via the use of phase plates, which could in no sense justify the huge efforts required to develop and produce them. In the materials science, DPC has successfully been used to observe individual atomic elements; in the biological sciences, the scattering contrast has been good enough to observe highly contrasting cellular structures, once they have been stained with heavy metals. The extinction of the phase plate might be quite natural, an evolutionary process that started at the beginning of the 1980s. New species often emerge in an isolated area, far away from the center of the evolutionary battlefield. The sudden emergence of an experimental paper published in 2001 [14] on a Zernike thin-film phase plate might be an example of such a new species. This study sought to awaken the EM community from its long silence. As shown by the photograph in Fig. 4a, a phase-plate EM image of negatively stained ferritin molecules, recorded under cryogenic conditions, looks far better than those seen in the past reports described previously. DPC cryo-EM images taken for the same sample without the phase plate, with various defocuses, are shown for comparison. Particularly, for the case of a deep defocus (2250 nm), the DCP image looks fine, with good contrast. But there must be a significant amount of image deformation arising from the deep defocus setting, which is inevitably associated with severe frequency modulation in the Thon ring feature, as shown in the inset. The whiteout in the center of the ferritin molecules (Fig. 4b) can be interpreted as such an effect of deep defocus. Fig. 4. View largeDownload slide The 300 kV cryo-TEM images of a sample from negatively stained horse spleen ferritin [from Fig. 6 of ref. 14]. (a) ZPC-TEM image acquired using a Zernike phase plate. (b–d) Conventional TEM images at a defocus of 2550, 540 and 130 nm, respectively (underfocus). The insets show the diffractogram for each image. The scale bars in the insets correspond to 1 nm−1. Fig. 4. View largeDownload slide The 300 kV cryo-TEM images of a sample from negatively stained horse spleen ferritin [from Fig. 6 of ref. 14]. (a) ZPC-TEM image acquired using a Zernike phase plate. (b–d) Conventional TEM images at a defocus of 2550, 540 and 130 nm, respectively (underfocus). The insets show the diffractogram for each image. The scale bars in the insets correspond to 1 nm−1. Today, we can view the ‘wake-up’ paper from four evolutionary viewpoints: (i) 20 years had passed since the last experimental report on phase plates; (ii) the first report of a functional phase plate was published by a group that began the EM work with no legacy in the field; (iii) because the group was isolated, the negative view of phase plates then prevailing in the EM community was very little disturbing; and (iv) the Zernike phase plate was collaterally developed to materialize a novel kind of EM experiment, so-called a complex observation. To express the wave function of electron waves as a complex-valued signal, two experiments consisted of the DPC and ZPC transmission electron microscopy (TEM) had to be combined as an organic unity [15, 16]. When we recall the tremendous effort required to solve the problem of phase-plate charging, the successful result of the ‘wake-up’ paper may be viewed as the type of luck that occasionally falls to beginners who are utterly ignorant of what had actually happened in the past. Soon after its publication [14], two of its authors, Danev and Nagayama, had realized the genuine difficulty of the charging problem. It had been so hard to reproduce Zernike phase plates of which performance was comparable with that utilized to obtain image data published in the 2001 paper. To understand how severely phase-plate performance was affected by charging, and how image deterioration is related to the level of charge, three authors (Danov, Danev and Nagayama) tried to estimate the amount of charge-induced phase retardance definable in the frequency space [17, 18]. The result of the estimation was depressing, though it did explain why past efforts to create phase plates had failed and research efforts had to cease. The result proved that a countable number of elementary charges on a phase plate would be enough to destroy the phase-plate performance. After the acknowledgement of the issue’s severity, the hunt for the origin of these charges started immediately. As many earlier workers had pointed out, the contamination of the surface with insulating materials was immediately found to be a major source of charge. The phase plate itself may not be charged when it is made of conducting material, such as carbon. The identification of the type of contamination materials was rather a difficult task. From the steady but blind efforts, such as varying the phase-plate materials, heating the phase plate, cleaning the surface with acid or organic solvents, rinsing, ion sputtering and oxygen etching, we have concluded that there are three sources of charge contamination: organic materials, metal oxides and inorganic materials. After this identification, subsequent studies aimed at identifying where and when those contaminating materials are deposited onto the phase-plate surface. The results of the hunt for the sources of contamination are summarized in Table 1. Table 1. Source specification of contamination-inducing charges in phase plates Origins  Possible sources  Organic materials  Backflow of oil mist from vacuum pumps  Silicone grease inside the TEM column  Metal oxides  Surface oxidation of phase plates made of metals  Gallium oxide deposited onto the phase plate during focused ion beam fabrication  Sputtering from metal walls of device components inside the vacuum evaporator  Inorganic materials  Mica flakes and minerals from the mica surface that adhere during the exfoliation of a carbon film from mica  Salt deposition from water during exfoliation of a carbon-film on the water surface  Origins  Possible sources  Organic materials  Backflow of oil mist from vacuum pumps  Silicone grease inside the TEM column  Metal oxides  Surface oxidation of phase plates made of metals  Gallium oxide deposited onto the phase plate during focused ion beam fabrication  Sputtering from metal walls of device components inside the vacuum evaporator  Inorganic materials  Mica flakes and minerals from the mica surface that adhere during the exfoliation of a carbon film from mica  Salt deposition from water during exfoliation of a carbon-film on the water surface  Metal oxides are very severe sources of charging; this is the major reason why carbon is the best source of material for the phase plate. Inorganic contamination, particularly from the mica surface, is almost impossible to avoid. We have had to carefully think about this, because the exfoliation of carbon films from the mica surface is an unavoidable consequence of the nature of multilayered 2D crystalline mica. To solve the phase-plate issue, we finally came to the conclusion that we need to wrap contaminants with conductive materials (here again, carbon) to electrostatically shield their charge-generating potential [1]. The essence of the remedy is the recognition that the existence of charges and the charges’ effects are separable. In what ultimately became the final step of phase-plate production, both sides of the phase plate – with the attendant contaminants of organic materials, metal oxides and inorganic materials – are wrapped with carbon in a vacuum evaporator. The wrapped carbon stays clean inside the EM column; once grounded, the sources of charging are eliminated, and the charging problem itself is solved. A schematic of a three-layered phase plate made of carbon is shown in Fig. 5b, together with a conventional Zernike phase plate for comparison (Fig. 5a). Carbon films thicker than 5–10 nm are fairly effective at shielding the electrostatic potential. Fig. 5. View largeDownload slide Shielding of electrostatic charges by a carbon wrapping finish, in order to fix the charging problem [adopted from Fig. 3 of ref. 1]. (a) Typical design of a Zernike phase plate made of a single-layer carbon film for an acceleration voltage of 200 kV. (b) Three-layer carbon film designed for a Zernike phase plate in order to shield the charging effect for an acceleration voltage of 200 kV. Fig. 5. View largeDownload slide Shielding of electrostatic charges by a carbon wrapping finish, in order to fix the charging problem [adopted from Fig. 3 of ref. 1]. (a) Typical design of a Zernike phase plate made of a single-layer carbon film for an acceleration voltage of 200 kV. (b) Three-layer carbon film designed for a Zernike phase plate in order to shield the charging effect for an acceleration voltage of 200 kV. 4. Biological applications of functional thin-film phase plates Careful fabrication procedures that avoid contaminants (Section 3) improve the success rate of functional phase-plate generation. In biological applications, however, the samples themselves are a significant source of charge inside the EM column. As is well known, organic compounds evaporate during the process of electron irradiation. Due to the location of the phase plate immediately below the specimen, some deposition of contaminants evaporated from specimens is unavoidable. To address this issue, a combination of a heating holder for phase plates and a set of cryo-shields was devised, as shown in the schematic in Fig. 6a. In Fig. 6, closed-up views of the complicated development of a Zernike phase plate are also illustrated in a stepwise fashion. Fig. 6. View largeDownload slide An overall schematic inside the objective lens pole piece of a phase plate EM, and stepwise close-ups zooming in on the Zernike phase plate. (a) A schematic side-view showing relative positions of various holders in the pole piece gap (1, specimen holder; 2, aperture holder; 3, phase plate holder; 4, cryo-shields). (b) A close-up view of the phase plate holder [from Fig. 2 of ref. 24]. (c) A close-up view of the multi-hole phase plate grid (2 mm in diameter) covered with a carbon thin film. The diameter of each hole is 100 µm. A diagonal can be seen in the film on the grid. (d) A close-up view of the Zernike phase plate [from Fig. 4 of ref. 14]. Fig. 6. View largeDownload slide An overall schematic inside the objective lens pole piece of a phase plate EM, and stepwise close-ups zooming in on the Zernike phase plate. (a) A schematic side-view showing relative positions of various holders in the pole piece gap (1, specimen holder; 2, aperture holder; 3, phase plate holder; 4, cryo-shields). (b) A close-up view of the phase plate holder [from Fig. 2 of ref. 24]. (c) A close-up view of the multi-hole phase plate grid (2 mm in diameter) covered with a carbon thin film. The diameter of each hole is 100 µm. A diagonal can be seen in the film on the grid. (d) A close-up view of the Zernike phase plate [from Fig. 4 of ref. 14]. The phase plate has been surrounded by three layers of cryo shield from gaseous contaminants (Fig. 6a), which can exclude organic compounds emitted from the biological specimen. Moreover, heating the phase-plate holder at 200–300°C can expel molecules deposited from the gas phase onto the phase-plate surface. A phase plate prepared with carbon wrapping finish to shield contaminant charging (Fig. 5b) can be inserted into an EM column with a composite design (Fig 6a). With such a setup, it is possible to image a range of ice-embedded biological specimens. Next, we will discuss examples of applications of functional phase plates to biological systems. When imaging biological samples, it is difficult to recover higher frequency components, which must arise from minute parasitic charges still remaining in qualified phase plates. Therefore, the biological applications are initially conducted in an ice-embedded state, and the first to be recognized by the biological EM community involved imaging of larger systems, such as samples of cultured bacterial and mammalian cells. These subjects were demanding of high resolution, and were therefore relatively safe from the phase-plate charging issue. Instead of high resolution, such large-scale specimens pose challenges regarding higher contrast in lower frequency components. In that respect, HDC, which is advantageous for recovering lower frequency components, was preferred for such investigations. HDC displays nanostructures of thin specimens in a topographical manner [19]. The specific electron-optical device for manipulating electron waves in HDC experiments is a half-plane π-phase plate (Fig. 1e). This technique was once named difference contrast TEM [19], but has been renamed HDC TEM (HDC-TEM) [20]. Two examples of HDC cryo-EM images shown in this report are experimental results obtained with a 300-kV EM system under liquid helium temperatures. Applications of Hilbert phase plates to cryo-EM The first example of HDC-TEM shown involves cyanobacteria, well-studied bacteria related to chloroplasts, which have a cylindrical geometry of 3 µm (length) × 1 µm (diameter) [21, 22]. Figure 7 shows a comparison of images obtained for an unstained, ice-embedded whole cell and a stained sectioned cell [21]. A characteristic feature of an HDC cryo-EM image, together with the high contrast, is demonstrated for an ice-embedded whole cell in Fig. 7a. Counterexamples obtained with DPC or scattering contrast with staining are shown in Figs. 7b and c. An obscure structureless image is observed in the DPC cryo-EM image, which was taken under the same experimental conditions as for the HDC image in Fig. 7a, except for the phase plate and the defocus. Because almost no contrast was recorded with DPC under near-focus conditions, the DPC image was taken at a deep defocus of 15 µm. Comparing another pair of images, the ice-embedded whole cell (Fig. 7a) and the resin-embedded sectioned cell (Fig. 7c), we observe a large difference in the image appearance, which may be attributable to the enormous difference in specimen treatment. In the sectioned cell, a ragged cell wall, which indicates that some shrinkage of the cell has occurred during TEM preparation, can be seen. Many aggregates and associated voids can also be recognized; these are inevitably induced by chemical treatment, such as dehydration and selective staining of cellular organelles. On the other hand, the images of the ice-embedded cell are smoothly round, and all spaces are recognizably filled. Notice here that the rapid freezing is expected to preserve the overall structure, such as the cell shape and subcellular structures. The preserved roundness of cyanobacterial cells allows us to estimate the specimen thickness to be ∼1 µm. Fig. 7. View largeDownload slide Comparison of TEM images of vitrified cyanobacterial cells [from Fig. 1 of ref. 21]. (a) A 300 kV HDC-TEM image of an ice-embedded unstained whole cell (near-focus). (b) A 300 kV DPC-TEM image of the same ice-embedded unstained whole as shown in (a) (∼15 µm defocus). (c) A 100 kV DPC-TEM image of a resin-embedded, sectioned and stained cell. Fig. 7. View largeDownload slide Comparison of TEM images of vitrified cyanobacterial cells [from Fig. 1 of ref. 21]. (a) A 300 kV HDC-TEM image of an ice-embedded unstained whole cell (near-focus). (b) A 300 kV DPC-TEM image of the same ice-embedded unstained whole as shown in (a) (∼15 µm defocus). (c) A 100 kV DPC-TEM image of a resin-embedded, sectioned and stained cell. Another example of HDC-TEM is shown for the case of a cultured mammalian cell, PtK2, taken from the kidney of kangaroo rat [23]. PtK2 cells have profusely developed microtubules, nano-scale molecular rails for the molecular motor-driven nanomachinery. First, treatment with nocodazole was confirmed with a fluorescent light microscope (insets in Fig. 8). Nocodazole is a drug that depolymerizes microtubules. Under these conditions, actin stress fibers are not affected, but microtubules are depolymerized. Comparison of the two fluorescent LM images (Figs. 8a and b) clearly demonstrates the depolymerization effect of nocodazole, which results in a change in the apparent morphology of fluorescent microtubules from a fibrous to a dispersed state. Fig. 8. View largeDownload slide ZPC-TEM images for PtK2 cells [from Figs. 2 and 3 of ref. 23]. (a) An untreated Ptk2 cell revealing a long spreading of microtubules, and different types of attached membranous organelles (mt, microtubulines; af, actin fibers; m, mitochondria; v, vesicles). (b) A nocodazole-treated vitrified Ptk2 revealing the elimination of the long filamentous structure, suggesting that the long filaments were microtubules (af, actin fibers; m, mitochondria; v, vesicles). Insets in (a) and (b) correspond to fluorescent microscopic images of PtK2 cells. Microtubules were stained with antitubulin monoclonal antibody DM1A (green), actin with phalloidin (red) and mitochondria by MitoTracker (cyan). Fig. 8. View largeDownload slide ZPC-TEM images for PtK2 cells [from Figs. 2 and 3 of ref. 23]. (a) An untreated Ptk2 cell revealing a long spreading of microtubules, and different types of attached membranous organelles (mt, microtubulines; af, actin fibers; m, mitochondria; v, vesicles). (b) A nocodazole-treated vitrified Ptk2 revealing the elimination of the long filamentous structure, suggesting that the long filaments were microtubules (af, actin fibers; m, mitochondria; v, vesicles). Insets in (a) and (b) correspond to fluorescent microscopic images of PtK2 cells. Microtubules were stained with antitubulin monoclonal antibody DM1A (green), actin with phalloidin (red) and mitochondria by MitoTracker (cyan). In Fig. 8a, an HDC cryo-EM image of a mammalian whole cell, long filaments running down to the peripheral region of the cell can be observed, and various membranous organelles can be recognized. From the geometrical characteristics of the filament (width of 25 nm, length of several µm), these structures were regarded as microtubules. Several membranous structures were also observed, and we identified some of them as mitochondria based on their striking inner membrane structure. The mitochondria were located near the microtubules, and sometimes overlapped as EM images of a thick whole cell go. The filaments were abolished by nocodazole treatment (photograph, Fig. 8b), confirming that these structures were microtubules. Applications of Zernike phase plates to single particle analyses Publication-worthy results of the application of phase plates to smaller biological systems, such as protein molecules, protein molecular machines and viruses, demanded a higher resolution in order to delineate their subjects’ internal fine structures. Such results could only be obtained when phase plates able to recover higher frequency components became available; this occurred after the use of carbon-wrapping-finish phase plates became routine [1, 24]. ZPC cryo-EM seems to be more suitable than HDC cryo-EM for smaller biological systems. Examples of ZPC images of various biological specimens, which were studied via single particle analyses (SPA) or cryo-electron tomography (cryo-ET), are collectively shown in Figs. 9 and 10. Fig. 9. View largeDownload slide Four examples of single-particle analysis based on cryo-electron microscopy. (a) A protein, GroEL (1: DPC image (300 kV), 2: ZPC image (300 kV), 3: 3D model) [from Fig. 2 of ref. 25]. (b) A membrane protein, TRPV4 (1: DPC image (300 kV), 2: ZPC image (300 kV), 3: 3D model) [from Figs. 3 and 6 of ref. 26]. (c) A bacteriophage, epsilon 15 (1: DPC image (200 kV), 2: ZPC image (200 kV), 3: 3D model) [from Figs. 2 and 3 of ref. 27]. (d) A capsid of herpes simplex virus type I (1: DPC image (200 kV), 2: ZPC image (200 kV), 3: 3D model) [from Figs. 1 and 2 of ref. 29]. Fig. 9. View largeDownload slide Four examples of single-particle analysis based on cryo-electron microscopy. (a) A protein, GroEL (1: DPC image (300 kV), 2: ZPC image (300 kV), 3: 3D model) [from Fig. 2 of ref. 25]. (b) A membrane protein, TRPV4 (1: DPC image (300 kV), 2: ZPC image (300 kV), 3: 3D model) [from Figs. 3 and 6 of ref. 26]. (c) A bacteriophage, epsilon 15 (1: DPC image (200 kV), 2: ZPC image (200 kV), 3: 3D model) [from Figs. 2 and 3 of ref. 27]. (d) A capsid of herpes simplex virus type I (1: DPC image (200 kV), 2: ZPC image (200 kV), 3: 3D model) [from Figs. 1 and 2 of ref. 29]. Fig. 10. View largeDownload slide Two examples of cryo-electron tomography based on DPC-TEM and ZPC-TEM (200 kV). (a) Projection images of part of a flagellar motor (HBB), taken at near zero tilt (1: DPC projection image, 2: ZPC projection image) [from Fig. 2 of ref. 34]. (b) Tomographic images sliced from a 3D reconstruction of HBB (1: DPC slice image, 2: ZPC slice image) [from Fig. 2 of ref. 34]. (c) Closed-up views of tomograms for the wild-type (1) and mutant (2) HBB. (d) Projection images of T4 phage, taken at near zero tilt. (1: DPC projection image, 2: ZPC projection image) [from Fig. 2 of ref. 33]. (e) Tomographic images sliced from a 3D reconstructed T4 phage (1: DPC slice image, 2: ZPC slice image) [from Fig. 2 of ref. 33]. (f) Close-up tomograms sampled from the 3D reconstruction shown in (e). Fig. 10. View largeDownload slide Two examples of cryo-electron tomography based on DPC-TEM and ZPC-TEM (200 kV). (a) Projection images of part of a flagellar motor (HBB), taken at near zero tilt (1: DPC projection image, 2: ZPC projection image) [from Fig. 2 of ref. 34]. (b) Tomographic images sliced from a 3D reconstruction of HBB (1: DPC slice image, 2: ZPC slice image) [from Fig. 2 of ref. 34]. (c) Closed-up views of tomograms for the wild-type (1) and mutant (2) HBB. (d) Projection images of T4 phage, taken at near zero tilt. (1: DPC projection image, 2: ZPC projection image) [from Fig. 2 of ref. 33]. (e) Tomographic images sliced from a 3D reconstructed T4 phage (1: DPC slice image, 2: ZPC slice image) [from Fig. 2 of ref. 33]. (f) Close-up tomograms sampled from the 3D reconstruction shown in (e). All the EM data except for the data corresponding to GroEL (for which a 300 kV EM system with a helium stage was used) were acquired with a 200 kV EM system under liquid nitrogen temperatures. Our first example of SPA is the case of protein that has widely been studied by EM, GroEL, which forms a complex that has a high molecular weight (800 kDa) [25]. The advantage of ZPC cryo-EM is illustrated in Fig. 9a, which compares the DPC and ZPC cryo-EM images of GroEL. In the conventional approach, the most challenging aspect of SPA is the first step: selecting the images that correspond to protein molecules. The difficulty is due to the fact that individual protein molecules must be identified using noisy raw images, which usually have very poor contrast, as shown in Fig. 9a(1). The handling of ZPC images with higher contrast, as shown in Fig. 9b(2), is less prone to human bias. Once a sufficient number of particles have been sampled and particle selection is completed, analysis becomes more computer-based and straightforward. A 13Å-resolution 3D map was reconstructed without a CTF correction from ∼1500 raw particles picked directly from raw ZPC images shown in the photograph in Fig. 9a(3). Danev et al. [24] compared the results of SPA with and without a Zernike phase plate, and attempted to quantify the efficiency of analysis by assessing the attainable resolution at different levels of sampling. Judging by the correlation between spatial resolution and the number of particles, analysis with ZPC was found to be 30% more efficient than with DPC (i.e. conventional analysis). The second example of SPA is the case of a membrane protein, transient receptor potential varioloid 4 (TRPV4), which has a molecular weight of ∼400 kDa [26]. A ZPC cryo-EM image of purified His-rTRPV4 is shown in Fig. 9b(2). Enlarged views of the particles picked from the area shown in Fig. 9b(2) are placed at the right of the figure. Cryo-EM imaging of membrane proteins in the detergent-solubilized state is difficult due to their structural heterogeneity and low molecular weight. Furthermore, cryo-specimens sometimes have an excessive thickness of vitreous ice, resulting in lowered contrast. Phase effects made it extraordinarily difficult to visualize TRPV4 particles with a conventional microscope (refer to Fig. 9b(1)). As mentioned, Zernike phase-contrast imaging has a much higher contrast at lower spatial frequencies, making it easier to identify and align the particles (refer to enlarged views), and obviating the need for CTF correction and defocus values. Employing C4 symmetry, a 35-Å resolution 3D map was reconstituted from ∼3900 raw particles (Fig. 9b(3)). The third example of SPA is the case of a phage, epsilon-15 [27]. ZPC cryo-EM images of ice-embedded epsilon-15 bacteriophages exhibited quite high contrast, as shown in the photograph in Fig. 9c(2). In this image, the tail hub and surrounding tail spikes can be clearly seen. The white halo surrounding each particle is particularly distinctive in ZPC images of viruses. The poor transfer of information below the cut-on frequency, corresponding to the area inside the phase-plate hole, is responsible for the fringe pattern (‘fringing’). The higher the particle image contrast, the stronger the fringe and hence the white halo appears. This is the reason why the white halo surrounding each of the virus particles so distinctive compared with the case of smaller particles of proteins such as GroEL (Fig. 9a(2)) and TRPV4 (the right side of Fig. 9b(2)). A 10-Å resolution icosahedral 3D map was reconstructed without a CTF correction from 4500 good particle images, which were selected from ∼6200 raw particle images. More interestingly, a 13-Å resolution 3D map was reconstructed without imposing symmetry and without any CTF correction (Fig. 9c(3)) [27]. In order to eliminate model bias, the asymmetric reconstruction was performed without using a known epsilon 15 tail model. The portal vertex complex (the gate structure for genomic DNA), located at 1 of the 12 five-fold vertices, contains several proteins: tail hub, tail spikes, portal and core proteins (Fig. 9c(3)). This map reveals several new structural features that were not resolved in the previous 20-Å resolution asymmetric 3D reconstruction [28]. Figure 9c(3) clearly shows that the tail hub has six subunits surrounded by the six tail spikes. The fourth example of SPA is the case of the capsid of herpes simplex virus type 1 B (HSV-1B) [29]. The ZPC cryo-EM image was targeted in a near defocus condition in order to obtain maximum contrast enhancement. The contrast of the resulting images is substantially higher than that of the DPC image (Fig. 9d(1)). As noted previously for images of virus particles, the white halo around each particle is distinct. From 353 CCD frames, ∼6000 single-particle images of HSV-1 B-capsids were collected. An icosahedral map without the CTF correction can be reconstructed with the use of ZPC images. A total of 2308 particles whose asymmetric orientations satisfied statistical conditions were used for reconstructing the final asymmetric map (Fig. 9d(3)). What is most remarkable in the asymmetric 3D model is that the exact position of the portal (the gate for genomic DNA), an important subject of recent research [30–32], is now finally fixed inside the capsid, as shown in Fig. 9d(3). Applications of Zernike phase plates to cryo–ET The application of phase plates application to Zernike cryo-tomography has been more technically challenging than application to SPA, for several reasons: (i) the tedious procedure of realignment of the phase plate, which is required in association with sample tilting; (ii) the need to write software to cope semiautomatically with the realignment and (iii) the need to occasionally replace the phase plate with a fresh one in order to compensate for phase-plate deterioration, which occurs during the lengthy data acquisition required for tomography. After overcoming these difficulties, the first cryo-tomographic data collected with ZPC-TEM was published in 2010 for the case of a virus, phage T4 [33]. The example of cryo-ET shown in Fig. 10 is the case of the hook-basal body (HBB), a protein supramolecule extracted from a bacterial strain, Vibrio alginolyticus, which uses flagella to swim [34]. A flagellum consists of a filament, a hook and a basal body. The HBB is made up of a rod and several ring structures. Figure 10a shows examples of DPC and ZPC cryo-EM images (projection images), taken close to a zero tilt, from a tilt series of purified HBBs recorded under low-dose conditions. From the conventional (DPC) projection, even with a deep defocus (∼5 µm), it was difficult to determine the location of the HBBs (Fig. 10a(1)). In contrast, the ZPC projection had much higher contrast and hence a clearer visibility (Fig. 10a(2)). Figure 10b shows tomograms of purified HBB samples, specifically z-slices of 3D reconstructed images of purified HBBs. In both of DPC and ZPC tomograms, the major features of individual HBBs were identifiable; however, the contrast of the ZPC slice was greatly improved compared with the conventional tomogram slice. By virtue of the improved contrast, structural details of HBBs could be compared between a wild-type strain and a deletion mutant strain lacking two constituent proteins, Motx and MotY, which are considered to form a ring structure specific to V. alginolyticus (T ring). In Fig. 10c, closed-up views of HBB tomograms of the wild type and mutant are shown. By inspection, one can easily see what has been lost from the structure of the mutant strain HBB; thus, the location of the T ring could be unambiguously determined by this comparison. The second example of cryo-ET is the case of T4 phage [33]. Figure 10c shows examples of sample images, taken close to zero-tilt, from DPC and ZPC tilt series of ice-embedded T4 phage. By comparing the DPC image (Fig. 10d(1)) with the ZPC image (Fig. 10d(2)) it is easy to see the overall contrast improvement produced by the phase plate. In the DPC image specimen, features are difficult to identify. In the ZPC image, as pointed out before, fringes appear around strong-contrast edges. Note, for example, the heavy dark fringe just inside the viral capsid, the thin light fringe outside the capsid and the light fringe close to the edge of the carbon film in the upper right corner. In addition, there is little difference in contrast between empty and filled capsids, indicating an imbalance of the frequency components in the image. All of these features arise from the finite size of the central hole, which introduces a finite cut-on frequency, as explained above. Comparisons of tomograms of bacteriophage T4, acquired under experimental conditions that were identical except for defocus and the presence or the absence of a phase plate, are shown in Fig. 10e. It is important to note the significant advantages in contrast and overall details in images, obtained by ZPC, relative to DPC images. Two tomograms sliced at different z positions from the reconstituted 3D model are shown in Fig. 10f(1) and f(2) as a closed-up view, in which internal structural details of the T4 phage are visible. 5. Applications of functional thin-film phase plates to artificial organic materials In parallel to the biological applications of phase-plate EM, the present author’s group also attempted to accumulate extensive experience in imaging of organic materials that were made artificially and supplied from outside in the context of collaborations. In comparison to biological imaging, such applications appear rather prosaic; nevertheless, these efforts are important to researchers who study structural aspects of organic materials, such as polymer and colloid systems. The latter are particularly attractive as mimics of biological systems, since they are usually prepared in the form of a vesicle, liposome or a detergent solution system, as with biological samples. Because artificial systems are less complicated than biological ones, they can be observed via EM using simpler methods, such as negative staining. Phase-plate EM combined with cryo-technology, however, can similarly provide the kinds of advantages obtained during imaging of bio-samples, such as high-contrast visualization of the undisturbed internal structures. In Fig. 11, we summarize ZPC cryo-EM data recently obtained for aqueous colloidal systems. Importantly, these data were collected using 300 kV ZPC-TEM under cryogenic conditions with no staining. Fig. 11. View largeDownload slide Three examples of cryo-EM (300 kV) image analysis of colloidal solution systems. (a) A model for the aggregated states in a detergent aqueous solution observed with ZPC-TEM (1–1: transformation from tubular structures to dispersed micelle structures) [from Fig. 8 of ref. 35]. (b) The internal structure of a lipid nanotube observed with ZPC-TEM (1: an image showing the growth of both inner and outer bilayer sheets around the core nanotube. The bilayer sheet winds at an angle of 45° relative to the tube axis (shown by white arrows). 2: A close-up of seven stacked lamellar sheets) [from Fig. 4 of ref. 36]. (c) A ternary complex (liposome, DNA and magnesium ion) structure observed with ZPC-TEM [from Figs. 2 and 3 of ref. 37] (1: spectrophotometric monitoring of liposomal aggregation induced by DNA and magnesium ion, 2 and 3: the morphological details observed with ZPC-TEM for samples extracted from a liposomal aggregation process, whose result is shown as the curve A in Fig. 1 as 2 and 3). Fig. 11. View largeDownload slide Three examples of cryo-EM (300 kV) image analysis of colloidal solution systems. (a) A model for the aggregated states in a detergent aqueous solution observed with ZPC-TEM (1–1: transformation from tubular structures to dispersed micelle structures) [from Fig. 8 of ref. 35]. (b) The internal structure of a lipid nanotube observed with ZPC-TEM (1: an image showing the growth of both inner and outer bilayer sheets around the core nanotube. The bilayer sheet winds at an angle of 45° relative to the tube axis (shown by white arrows). 2: A close-up of seven stacked lamellar sheets) [from Fig. 4 of ref. 36]. (c) A ternary complex (liposome, DNA and magnesium ion) structure observed with ZPC-TEM [from Figs. 2 and 3 of ref. 37] (1: spectrophotometric monitoring of liposomal aggregation induced by DNA and magnesium ion, 2 and 3: the morphological details observed with ZPC-TEM for samples extracted from a liposomal aggregation process, whose result is shown as the curve A in Fig. 1 as 2 and 3). One illustrative study explored the differential aggregation behavior of a detergent, potassium N-acyl phenylalaninate, as a function of its solution concentration. Contrary to normal colloidal systems, this compound exhibited large aggregates at lower concentrations, which then transformed to micelles at a higher concentration. To inspect this surprising phenomenon, Ohta et al. applied ZPC-TEM [35]; they found that the large aggregates observed in the initial stage corresponded to tubes with bilayer structures, which then transformed into spherical micelles via thread-like micelle intermediates, as shown in Fig. 11a. The morphological transformation detailed in the ZPC cryo-images took place in a stepwise fashion, with progressive aggregation of a helical ribbon shape first, a tubular shape second, a deformed tube made of thread-like micelles third and finally dispersed micelles (refer to Fig. 11a(1) to a(5)). This kind of new finding about colloidal morphology could only be pursued via the combination of cryo-EM and phase-plate EM used in this study. A lipid nanotube made of N-(11-cis-octadecenoyl)-β-d-glucopyranosylamine was investigated via ZPC-TEM [36]. In this study, the quick-freezing method helped the authors understand the initial growth process of a self-assembled lipid nanotube in water. Due to the high contrast, the thin lamellar edge structure could be identified, and the very fast growth of the newborn edge into a thin tube was elucidated. The thin tube acted as a core structure for further growth into a thick complete lipid nanotube (Fig. 11b(1)). The initially formed nanotube structure, denoted as a core tube, had a uniform wall structure that consists of several lamellar layers; the inner and outer diameters of the core tube are 130 and 180 nm, respectively. The evaluated lamellar spacing of 4.6 nm is compatible with the spacing measured by X-ray diffraction. The molecular packing of the nanotube, determined from the ZPC images together with X-ray wide-angle diffraction and IR spectroscopy, was classified as orthorhombic. The formation of ternary complexes (TCs) from a mixture of DNA, phosphatidylcholine (PC) liposome and divalent metal cations was spectrophotometrically studied, as illustrated in Fig. 11c(1). In this study, the authors observed the aggregation process of PC liposomes as a result of the addition of double-stranded DNA molecules and/or divalent cations [37]. As seen from Fig. 11c(1)B, a divalent cation, Mg2+, is more effective at inducing liposomal aggregation than another cation, Ba2+. An obvious question while studying these complexes is whether DNA condensation is observed on the PC liposome surface or net, as has been previously shown for cationic lipoplexes [38]. Figure 11c(2) shows a ZPC image of TC formed in the presence of Mg2+ and in the absence of Ba2+; Fig. 11c(3) shows a TC formed upon addition of Ba2+, leading to a clear visualization of DNA compacting on the liposome surface. The DNA threads were found both on the PC liposome surface and in solution (Fig. 11c(3)). Further analyses revealed bridges between the large liposomes, formed by fibrils that could be interpreted as DNA molecules. The advantages of EM observation using phase plates, but without heavy metal staining, could be enjoyed in imaging of solid polymer materials, since the detailed intact morphology inside solid polymers could be delineated in the same manner as for biopolymers. Two examples of polymer systems studied with a 300-kV ZPC-TEM are shown in Fig. 12. Fig. 12. View largeDownload slide Two examples of EM image analysis for solid polymer systems. (a) TEM images of a rubber and carbon black mixture, without staining (1: DPC image (300 kV), 2: ZPC image (300 kV), 3: An enlarged ZPC image visualizing stearic acid crystals [from Figs. 1 and 2 of ref. 39]). (b) TEM image of polyethylene (PE) film casted on PTFE [1: conventional (DPC) image (200 kV) of a metal shadowed sample of PE (Mw = 36 500 Da), 2: conventional (DPC) image (200 kV) of a metal shadowed sample of PE (Mw = 11 550 Da), 3: ZPC image (300 kV) of an ice-embedded sample of PE (Mw = 36 500 Da)] [from Figs. 1, 4 and 5 of ref. 40]. Fig. 12. View largeDownload slide Two examples of EM image analysis for solid polymer systems. (a) TEM images of a rubber and carbon black mixture, without staining (1: DPC image (300 kV), 2: ZPC image (300 kV), 3: An enlarged ZPC image visualizing stearic acid crystals [from Figs. 1 and 2 of ref. 39]). (b) TEM image of polyethylene (PE) film casted on PTFE [1: conventional (DPC) image (200 kV) of a metal shadowed sample of PE (Mw = 36 500 Da), 2: conventional (DPC) image (200 kV) of a metal shadowed sample of PE (Mw = 11 550 Da), 3: ZPC image (300 kV) of an ice-embedded sample of PE (Mw = 36 500 Da)] [from Figs. 1, 4 and 5 of ref. 40]. In Fig. 12a, cryo-EM images of a natural rubber sample filled with carbon black (NR–CB) are shown. The DPC image required strong defocusing to recover reasonable contrast (Fig. 12a(1)), whereas the ZPC image exhibited high contrast (Fig. 12a(2)) at the same location under near-focus conditions [39]. The ZPC-TEM image shown in Fig. 12a(3) was taken from a different part of the NR–CB sample. In this figure, the (001) lattice fringes of stearic acid crystals, which were added to the mixture of NR and CB as a dispersant, can be observed together with the carbon black particles. It is noteworthy that samples with a rubber matrix are usually unsuitable for electron staining, because the heavy metal atoms are deposited in the matrix phase and the entire sample becomes dark. ZPC-TEM made it possible for the first time to obtain clear images of the NR–CB system, revealing such structural details as those depicted in the photographs in Fig. 12a(2) and a(3). Figure 12b(1) shows an example of a nano-sized array prepared from a dilute xylene solution of polyethylene (PE) casted onto a polytetrafluoroethylene (PTFE) substrate. The image was taken using a conventional TEM with metal shadowing [40]. The remarkable structures look like shish kebabs. The arrays (kebabs) in question are found along the ridge of the PTFE layer; they are composed of rod-like entities, which are roughly parallel with one another. The periodicity of the arrays is ∼100 nm. Close inspection of Fig. 12b(1) revealed the stacked lamellae in each array (kebab) as shown in Fig 12b(2), which is a metal-shadowed conventional image. The top faces of the ‘kebabs’ are flat, with a contrast feature that can be interpreted as dents at the center. On the other hand, ZPC-TEM allowed observation of different features for the arrays in question. From the ZPC image of the arrays of PE shown in Fig. 12b(3), the rods are known to be composed of grains smaller than ∼30 nm. Moreover, the dent at the center of each rod-like structure can be visualized as a region of sparse grains. This detailed feature is only accessible via the high contrast attained when using ZPC-TEM on untreated polymer samples. 6. Various phase-plate devices recently proposed This section introduces different types of phase plates other than the thin-film one, organized according to their design and the underlying physical principles [41]. The gallery of designs of the phase plates is shown in Fig. 13. Fig. 13. View largeDownload slide Gallery of various phase plate designs currently under development [from Figs. 5–9 of ref. 41]. (a and b) A Boersch electrostatic phase plate [42–45]. (c and d) An AB effect phase plate with a magnetic thin wire [49]. (e and f) A photonic phase plate using laser irradiation [50]. (g) An electrostatic mirror phase plate [51]. (h and i) An anamorphotic phase plate [52]. Fig. 13. View largeDownload slide Gallery of various phase plate designs currently under development [from Figs. 5–9 of ref. 41]. (a and b) A Boersch electrostatic phase plate [42–45]. (c and d) An AB effect phase plate with a magnetic thin wire [49]. (e and f) A photonic phase plate using laser irradiation [50]. (g) An electrostatic mirror phase plate [51]. (h and i) An anamorphotic phase plate [52]. Electrostatic phase plates The use of an electrostatic field in the construction of a phase plate was suggested by Boersch [3] (Fig. 2a). Electrons passing through the Einzel lens experience a phase shift with a magnitude determined by the excitation voltage of the central electrode. Although this design was proposed >60 years ago, the fabrication challenges were overcome only recently, allowing practical research and experiments to begin. Figure 13b illustrates a typical implementation of a Boersch phase plate. At present, there are several groups working on Boersch-type devices [42–45]. Most devices are based on the design proposed 15 years ago by Matsumoto and Tonomura [13], consisting of inner and outer electrodes separated by an insulating layer. An exception is the work of Cambie et al., who propose a ‘drift tube’ device in which the electrodes are separated by vacuum [46]. Unlike thin-film phase plates, there is no loss of electrons at higher spatial frequencies due to scattering by a material film. The main disadvantage of electrostatic phase plates is obstruction of parts of the diffraction plane by the mechanical components of the device, especially the area around the zero-order beam where the Einzel lens is located. The recovery of lower frequency components, which are responsible for higher contrast, will be a major challenge for the development of electrostatic phase plates. One possible remedy for this problem would be to insert a transfer lens after the objective, which would ease the geometrical requirements for the Einzel lens. Such a transfer lens has already been tested with a TEM machine dedicated to phase plate EM, and has been proven to be quite effective [47]. Magnetic phase plates The ability of magnetic potential to shift the phase of electron waves was first demonstrated by Tonomura et al. [48]. These authors used trapped magnetic flux in a ring covered by a superconducting material to prevent the leakage of the field. Their experiment was aimed at decisively proving the AB effect: a quantum-mechanical phenomenon in which charged particles interact with the electromagnetic potential (vector potential) in a field-free space. The group proposed to use the AB effect in phase plates, but there had been no prior experimental trials in that direction. In order to avoid the challenges of manufacturing and supporting a small magnetic ring, Nagayama proposed and is currently investigating a device utilizing a linear magnet [49]. The magnet in question is a thin wire made by evaporation of magnetic metals onto the blade edge of a platinum wire that had been trimmed to a width of ∼500 nm from its original thickness of 10 µm by a focused ion beam apparatus. Figure 13c shows a top view, and Fig. 13d shows an angled view, of the thinned part of the platinum wire. The wire is supported on an aperture disk (not shown). The magnetic flux confined through the magnetic thin wire induces a phase-shift difference between electron waves passing on both sides of the wire. Positioning the zero-order beam very close to one side of the strip, without touching it, will create a phase shift difference between the two halves of the diffraction plane. This asymmetric phase modulation makes this device an HDC-type phase plate. Photonic phase plates A very recently proposed idea is to use high-intensity focused laser light to introduce a phase shift in the electron wave [50]. Figure 13e shows the optical arrangement for such a device. A simple semi-classical explanation can be given for the principle underlying the device’s operation. The intense oscillating electrostatic field at the laser focus causes the electrons to ‘wiggle’ along their path. This elongates their optical paths and simultaneously causes a small increase in their kinetic energy, which in turn shortens their wavelength. The combination of the two effects leads to phase retardance of the electron wave. Figure 13f shows two simplified scenarios: the left side illustrates an electron in a linearly polarized electrostatic field; and the right side illustrates an electron in a circularly polarized field. The biggest practical challenge of the photonic phase plate is the required laser power, which is in the order of several kiloWatts. The power requirement increases rapidly as a function of acceleration voltage. One possible solution would be to increase the intensity at the laser focal spot by an optical resonance cavity, as has been proposed [50]. Electrostatic mirror pixel-wise phase shifter Another novel idea for a TEM phase plate is to use an electrostatic mirror at an appropriate place along the beam path [51]. Figure 13g illustrates one possible configuration for such a device. The electron beam coming from the TEM column is deflected by the magnetic beam separator and directed toward the electrostatic mirror. There may be focusing lenses in front of the mirror. Inside the mirror, the electrons are gradually decelerated until they stop at the zero-energy plane. After momentarily stopping, the electrons are accelerated backwards by the mirror and deflected toward the remainder of the TEM column by the beam separator. Close behind the zero-energy plane is a pixel-wise phase shifting device. The voltage of each electrode/pixel is controlled independently. The 2D array of pixels can be excited to a predefined voltage pattern, producing a desired electrostatic potential distribution at the zero-energy plane, located just in front of the array. Consequently, the electron wavefront can be modulated by a freely controllable 2D phase-shift pattern that can be varied in real time. Anamorphotic phase plates The anamorphotic phase plate for TEM was recently proposed by Schröder et al. [52]. Figure 13h illustrates the construction of such a device. Embedded electrodes are exposed on the inner wall of a narrow, slit-shaped aperture. The number and arrangement of the electrodes can vary depending on the desired behavior of the device and the maker’s manufacturing capabilities. The anamorphotic phase plate requires an optical plane exhibiting anisotropic magnification in which the diffraction pattern is highly compressed in one direction. Figure 13i shows the optical arrangement of the anamorphotic diffraction pattern inside the phase-plate aperture. The narrow slit helps to localize the electrostatic field generated by the electrodes. The aspect ratio of the slit determines its ability to confine the field. The phase shift profile is controlled in only one dimension, across the long side of the slit, resulting in strip-wise modulation in the Fourier space; this may introduce some trouble with the fringing effect, as discussed in previous sections. 7. Discussion and conclusions Phase-plate EM is still in its infancy with respect to its technological and methodological development. For example, the image modulation due to fringing that arises from the finite size of the central hole in a Zernike phase plate has not yet been fully solved, although some attempts have already been made to ease it, either by introducing a spatial filter [34] or by reducing the size of the central hole [53]. In addition to the fringe issue, the technology of thin-film phase plate has other problems, such as inconsistency of manufacturing, resulting in difficulties in avoiding parasitic contaminant charging; aging, specifically charge evolution in the phase plate after its installation in the EM; and information loss due to the interception of electrons by plate materials. Mass production techniques and phase-plate airlock systems [44] could, in the near future, alleviate the former two problems, but the information loss is an inherent feature of the thin-film type of phase plate, as previously pointed out. Several groups are advancing in the development of electrostatic phase plates, as seen in Section 6. This type of phase plate does not suffer from information loss at high spatial frequencies; furthermore, it has the advantage of real-time phase shift control, which can open the doors to new methods such as complex observation [15, 16, 54], which could recover the electron wave function itself in a complex form. Complex observation, even though technically demanding, will represent a complete solution to the inherent problem in the current modalities of phase-plate EM, specifically the nonlinear appearance of phase contrast in strong objects [24] or more precisely, the inability to extract the phase information, θ(r) itself, for strong objects [15]. At present, phase-plate experiments require manual operation, greatly reducing their attractiveness. Hardware and software are much-needed for automated phase-plate alignment, evaluation, focusing and data collection; such developments would make phase-plate EM more accessible to groups studying in biology or medicine. Phase-plate EM is also in its infancy in its biological applications. At present, only the thin-film phase plate is being used for cryo-observations of biological targets; this approach has demonstrated practical benefits, as shown in Sections 3–5. We expect that some electrostatic phase plates will soon be tested in biological applications, solving the problems of information loss and charging effects that appear at higher spatial frequencies. At current levels of performance, the highest attainable resolution is limited to within 0.8 nm [27]; nevertheless, the examples of biological applications shown in this report clearly illustrate the benefits of higher contrast, which has resulted in novel findings that may not have been possible without phase plates. We will continue to see an intimate conversation between two sectors – technological development of phase plates and biological imaging using phase plates – and we foresee in the near future the emergence of communication between phase-plate EM and the materials science. Acknowledgements I owe the development and biological applications of phase-plate EM to the following collaborators: Development: Radostin Danev, Shozo Sugitani, Rasmus Schroeder, Kenneth Holmes, Hiroshi Okawara, Toshiyuki Itoh, Masahiro Ohara, Toshikazu Honda, Toshiaki Suzuki, Yoshiyasu Harada, Yoshihiro Arai, Fumio Hosokawa, Sohei Motoki, Kazuo Ishizuka, Robert Glaeser, and Michael Marko. Applications: Yasuko Kaneko, Koji Nitta, Radostin Danev, Nobuteru Usuda, Kimie Atsuzawa, Naoki Hosogi, Hiroyuki Terashima, Michio Honma, Hideki Shigematsu, Makoto Tominaga, Takaaki Sokabe, Shuji Kanemaru, Yoshie Maitani, Hideki Akita, Kentaro Kogure, Kentaro Sasaki, Kunitada Shimotohno, Tomoya Masuda, Hideyoshi Harashima, Masashi Yamaguchi, Mitsutoshi Setou, Akio Ohta, Hiroharu Yui, Vasily Kuvichkin, Masatoshi Tosaka, Atsushi Shimada, Shigeyuki Yokoyama, Yoshiyuki Fukuda, Ayami Nakazawa, Kiyokazu Kametani, Masashi Tanaka, Hideo Hirokawa, Fumio Arisaka, Koki Taniguchi, Holland Cheng, Xing Li, Ryan Rochet, Xiangan Liu, Frazor Rixon, Joanita Jakana, Michael Shmid, Jonathan King, Kazuyoshi Murata and Wah Chiu. References 1 Nagayama K.  Phase contrast enhancement with phase plates in electron microscopy,  Adv. Imaging Electron Phys. ,  2005, vol.  135 (pg.  69- 146)  doi:10.1016/S1076-5670(05)38002-5 2 Zernike F.  Phase contrast, a new method for the microscopic observation of transparent objects,  Physica ,  1942, vol.  9 (pg.  686- 698)  doi:10.1016/S0031-8914(42)80035-X Google Scholar CrossRef Search ADS   3 Boersch H.  Über die Kontraste von Atomen in Elektronmikroskop,  Z. Naturforschg. ,  1947, vol.  2a (pg.  615- 633) 4 Scherzer O.  The theoretical resolution limit of the electron microscope,  J. Appl. Phys. ,  1949, vol.  20 (pg.  20- 29)  doi:10.1063/1.1698233 Google Scholar CrossRef Search ADS   5 Kanaya K,  Kawakatsu H,  Ito K,  Yotsumoto H.  Experiment on the electron phase microscope,  J. Appl. Phys. ,  1958, vol.  29 (pg.  1046- 1049)  doi:10.1063/1.1723360 Google Scholar CrossRef Search ADS   6 Faget J,  Fagot M,  Ferre J,  Fert C.  Microscopie Electronique a Contraste de Phase,  1962 Proceedings of the 5th International Congress Electron Microscopy A-7 New York Academic Press 7 Badde H G,  Reimer L.  Der Einfluß einer streuenden Phasenplatte auf das elektronen mikroskopische Bild,  Z. Naturforschg. ,  1970, vol.  25a (pg.  760- 765) 8 Unwin P N T.  An electrostatic phase plate for the electron microscope,  Bunsen-Gesellschaft ,  1970, vol.  74 (pg.  1137- 1141) 9 Krakow W,  Siegel B M.  Phase contrast in electron microscope images with an electrostatic phase plate,  Optik ,  1975, vol.  44 (pg.  245- 268) 10 Johnson H,  Parsons D.  Enhanced contrast in electron microscopy of unstained biological material,  J. Microsc. ,  1973, vol.  98 (pg.  1- 17) Google Scholar CrossRef Search ADS PubMed  11 Willasch D.  High resolution electron microscopy with profiled phase plates,  Optik ,  1975, vol.  44 (pg.  17- 36) 12 Ballosier G,  Bonnet N.  Use of electrostatic phase plate in TEM. Transmission electron microscopy: improvement of phase and topographical contrast,  Optik ,  1981, vol.  58 (pg.  361- 376) 13 Matsumoto T,  Tonomura A.  The phase constancy of electron waves travelling through Boersch’s electrostatic phase plate,  Ultramicroscopy ,  1996, vol.  63 (pg.  5- 10)  doi:10.1016/0304-3991(96)00033-2 Google Scholar CrossRef Search ADS   14 Danev R,  Nagayama K.  Transmission electron microscopy with Zernike phase plate,  Ultramicroscopy ,  2001, vol.  88 (pg.  243- 252)  doi:10.1016/S0304-3991(01)00088-2 Google Scholar CrossRef Search ADS PubMed  15 Nagayama K.  Complex observation in electron microscopy. Phase retrieval for strong objects with Foucault knife-edge scanning,  J. Phys. Soc. Jpn. ,  1999, vol.  73 (pg.  2725- 2731)  doi:10.1143/JPSJ.73.2725 Google Scholar CrossRef Search ADS   16 Danev R,  Nagayama K.  Complex observation in electron microscopy. II. Direct visualization of phases and amplitudes of exit wave function,  J. Phys. Soc. Jpn. ,  2004, vol.  70 (pg.  696- 702)  doi:10.1143/JPSJ.70.696 Google Scholar CrossRef Search ADS   17 Danov K,  Danev R,  Nagayama K.  Electric charging of thin films measured using the contrast transfer function,  Ultramicroscopy ,  2001, vol.  87 (pg.  45- 54)  doi:10.1016/S0304-3991(00)00061-9 Google Scholar CrossRef Search ADS PubMed  18 Danov K,  Danev R,  Nagayama K.  Reconstruction of the electric charge density in thin films from the contrast transfer function measurements,  Ultramicroscopy ,  2002, vol.  90 (pg.  85- 95)  doi:10.1016/S0304-3991(01)00143-7 Google Scholar CrossRef Search ADS   19 Danev R,  Okawara H,  Usuda N,  Kametani K,  Nagayama K.  A novel phase-contrast transmission electron microscopy producing high-contrast topographic images of weak objects,  J. Biol. Phys. ,  2002, vol.  28 (pg.  627- 635)  doi:10.1023/A:1021234621466 Google Scholar CrossRef Search ADS PubMed  20 Danev R,  Nagayama K.  Complex observation in electron microscopy. IV. Reconstruction of complex object wave from conventional and half plane phase plate image pair,  J. Phys. Soc. Jpn. ,  2004, vol.  73 (pg.  2718- 2724)  doi:10.1143/JPSJ.73.2718 Google Scholar CrossRef Search ADS   21 Kaneko Y,  Danev R,  Nitta K,  Nagayama K.  In vivo subcellular ultrastructures recognized with Hilbert-differential-contrast transmission electron microscopy,  J. Electron Microsc. ,  2005, vol.  54 (pg.  79- 84)  doi:10.1093/jmicro/dfh105 Google Scholar CrossRef Search ADS   22 Kaneko Y,  Danev R,  Nagayama K,  Nakamoto H.  Intact carboxysomes in a cyanobacterial cell visualized by Hilbert differential contrast transmission electron microscopy,  J. Bacterol. ,  2006, vol.  188 (pg.  805- 808)  doi:10.1128/JB.188.2.805-808.2006 Google Scholar CrossRef Search ADS   23 Setou M,  Danev R,  Atsuzawa K,  Yao I,  Fukuda Y,  Usuda N,  Nagayama K.  Mammalian cell nano structures visualized by cryo Hilbert differential contrast transmission electron microscopy,  Med. Mol. Morphol. ,  2006, vol.  39 (pg.  176- 180)  doi:10.1007/s00795-006-0341-8 Google Scholar CrossRef Search ADS PubMed  24 Danev R,  Glaeser R M,  Nagayama K.  Practical factors affecting the performance of a thin-film phase plate for transmission electron microscopy,  Ultramicroscopy ,  2009, vol.  109 (pg.  312- 325)  doi:10.1016/j.ultramic.2008.12.006 Google Scholar CrossRef Search ADS PubMed  25 Danev R,  Nagayama K.  Single particle analysis based on Zernike phase contrast microscopy,  J. Struct. Biol. ,  2008, vol.  161 (pg.  211- 218)  doi:10.1016/j.jsb.2007.10.015 Google Scholar CrossRef Search ADS PubMed  26 Shigematsu H,  Sokabe T,  Danev R,  Tominaga M,  Nagayama K.  A 3.5-nm structure of rat TRPV4 cation channel revealed by Zernike phase-contrast cryo-EM,  J. Biol. Chem. ,  2010, vol.  285 (pg.  11210- 11218)  doi:10.1074/jbc.M109.090712 Google Scholar CrossRef Search ADS PubMed  27 Murata K,  Liu X,  Danev R,  Jakana J,  Schmid M F,  King J,  Nagayama K,  Chiu W.  Zernike phase contrast cryo-electron microscopy and tomography for structure determination at nanometer and subnanometer resolutions,  Structure ,  2010, vol.  18 (pg.  903- 912)  doi:10.1016/j.str.2010.06.006 Google Scholar CrossRef Search ADS PubMed  28 Jiang W,  Chang J,  Jakana J,  Weigele P R,  King J,  Chiu W.  Structure of epsilon 15 bacteriophage reveals genome organization and DNA packaging/injection apparatus,  Nature ,  2006, vol.  439 (pg.  612- 616)  doi:10.1038/nature04487 Google Scholar CrossRef Search ADS PubMed  29 Rochat R H,  Liu X,  Murata K,  Nagayama K,  Rixon F,  Chiu W.  Seeing the genome packaging apparatus in herpes simplex virus type I (HSV-1) B-capsids,  J. Virol. ,  2011, vol.  85 (pg.  1871- 1874) Google Scholar CrossRef Search ADS PubMed  30 Trus B,  Cheng N,  Newcomb W,  Homa F,  Brown J,  Steven A.  Structure and polymorphism of the UL6 portal protein of herpes simplex virus type 1,  J. Virol. ,  2004, vol.  78 (pg.  12668- 12671)  doi:10.1128/JVI.78.22.12668-12671.2004 Google Scholar CrossRef Search ADS PubMed  31 Chang J,  Schmid M,  Rixon F,  Chiu W.  Electron cryotomography reveals the portal in the herpesvirus capsid,  J. Virol. ,  2007, vol.  81 (pg.  2065- 2068)  doi:10.1128/JVI.02053-06 Google Scholar CrossRef Search ADS PubMed  32 Cardone G,  Winkler D,  Trus B,  Cheng N,  Heuser J,  Newcomb W,  Brown J,  Steven A.  Visualization of the herpes simplex virus portal in situ by cryo-electron tomography,  Virology ,  2007, vol.  361 (pg.  426- 434)  doi:10.1016/j.virol.2006.10.047 Google Scholar CrossRef Search ADS PubMed  33 Danev R,  Kanamaru S,  Marko M,  Nagayama K.  Zernike phase contrast cryo-electron tomography,  J. Struct. Biol. ,  2010, vol.  171 (pg.  174- 181)  doi:10.1016/j.jsb.2010.03.013 Google Scholar CrossRef Search ADS PubMed  34 Hosogi N,  Shigematsu H,  Terashima H,  Homma M,  Nagayama K.  Zernike phase contrast cryo-electron tomography of sodium-driven flagellar hook-basal bodies from Vibrio Alginolyticus,  J. Struct. Biol. ,  2011, vol.  173 (pg.  67- 76)  doi:10.1016/j.jsb.2010.08.004 Google Scholar CrossRef Search ADS PubMed  35 Ohta A,  Danev R,  Nagayama K,  Mita T,  Asakawa T,  Miyagishi S.  Transition from nanotubes to micelles with increasing concentration in dilute aqueous solution of potassium N-acyl phenylalaninate,  Langmuir ,  2006, vol.  22 (pg.  8472- 8477)  doi:10.1021/la0611110 Google Scholar CrossRef Search ADS PubMed  36 Yui H,  Minamikawa H,  Danev R,  Nagayama K,  Kamiya S,  Shimizu T.  Growth process and molecular packing of a self-assembled lipid nanotube: phase-contrast transmission electron microscopy and XRD analysis,  Langmuir ,  2008, vol.  24 (pg.  709- 713)  doi:10.1021/la702488u Google Scholar CrossRef Search ADS PubMed  37 Kuvichkin V,  Danev R,  Shigematsu H,  Nagayama K.  DNA-induced aggregation and fusion of phosphatidylcholine liposomes in the presence of multivalent cations observed by the Cryo-TEM technique,  J. Membr. Biol. ,  2009, vol.  227 (pg.  95- 103)  doi:10.1007/s00232-008-9146-0 Google Scholar CrossRef Search ADS PubMed  38 Sennato S,  Bordi F,  Cametti C,  Diociaiuti M,  Malaspina P.  Charge patch attraction and reentrant condensation in DNA–liposome complexes,  Biochim. Biophys. Acta ,  2005, vol.  1714 (pg.  11- 24)  doi:10.1016/j.bbamem.2005.06.004 Google Scholar CrossRef Search ADS PubMed  39 Tosaka M,  Danev R,  Nagayama K.  Application of phase contrast transmission microscopic methods to polymer materials,  Macromolecules ,  2005, vol.  38 (pg.  7884- 7886)  doi:10.1021/ma0512197 Google Scholar CrossRef Search ADS   40 Tosaka M,  Tsuji M,  Ogawa T,  Kitano H,  Nakano K,  Kohjiya S,  Danev R,  Nagayama K.  Self-assembly of nano-sized arrays on highly oriented thin films of poly (tetrafluoroethylene),  Polymer ,  2006, vol.  47 (pg.  951- 955)  doi:10.1016/j.polymer.2005.12.044 Google Scholar CrossRef Search ADS   41 Danev R,  Nagayama K.  Phase plates for transmission electron microscopy,  Meth. Enzymol. ,  2010, vol.  481 (pg.  343- 369)  doi:10.1016/S0076-6879(10)81014-6 Google Scholar PubMed  42 Huang S H,  Wang W J,  Chang C S,  Hwu Y K,  Tseng F G,  Kai J J,  Chen F R.  The fabrication and application of Zernike electrostatic phase plate,  J. Electron Microsc. ,  2006, vol.  54 (pg.  317- 324) 43 Majorovits E,  Barton B,  Schultheiß K,  Pérez-Willard F,  Gerthsen D,  Schröder R R.  Optimizing phase contrast in transmission electron microscopy with an electrostatic (Boersch) phase plate,  Ultramicroscopy ,  2007, vol.  107 (pg.  213- 226)  doi:10.1016/j.ultramic.2006.07.006 Google Scholar CrossRef Search ADS PubMed  44 Shiue J,  Chang C S,  Huang S H,  Hsu C H,  Tsai J S,  Chang W H,  Wu Y M,  Lin Y C,  Kuo P C,  Huang Y S,  Hwu Y,  Kai J J,  Tseng F G,  Chen F R.  Phase TEM for biological imaging utilizing a Boersch electrostatic phase plate: theory and practice,  J. Electron Microsc. ,  2009, vol.  58 (pg.  137- 145)  doi:10.1093/jmicro/dfp006 Google Scholar CrossRef Search ADS   45 Alloyeau D,  Hsieh W K,  Anderson E H,  Hilken L,  Benner G,  Meng X,  Chen F R,  Kisielowski C.  Imaging of soft and hard materials using a Boersch phase plate in a transmission electron microscope,  Ultramicroscopy ,  2010, vol.  110 (pg.  563- 570)  doi:10.1016/j.ultramic.2009.11.016 Google Scholar CrossRef Search ADS   46 Cambie R,  Downing K H,  Typke D,  Glaeser R M,  Jin J.  Design of a microfabricated, two-electrode phase-contrast element suitable for electron microscopy,  Ultramicroscopy ,  2007, vol.  107 (pg.  329- 339)  doi:10.1016/j.ultramic.2006.09.001 Google Scholar CrossRef Search ADS PubMed  47 Hosokawa F,  Danev R,  Arai Y,  Nagayama K.  Transfer doublet and an elaborated phase plate holder for 120kv electron-phase microscope,  J. Electron Microsc. ,  2005, vol.  54 (pg.  317- 324)  doi:10.1093/jmicro/dfi049 Google Scholar CrossRef Search ADS   48 Tonomura A,  Osakabe N,  Matsuda T,  Kawasaki T,  Endo J.  Evidence for Aharonov–Bohm Effect with magnetic field completely shielded from electron wave,  Phys. Rev. Lett. ,  1986, vol.  56 (pg.  792- 795)  doi:10.1103/PhysRevLett.56.792 Google Scholar CrossRef Search ADS PubMed  49 Nagayama K.  Development of phase plates for electron microscopes and their biological application,  Eur. Biophys. J. ,  2008, vol.  37 (pg.  345- 358)  doi:10.1007/s00249-008-0264-5 Google Scholar CrossRef Search ADS PubMed  50 Müller H,  Jin J,  Danev R,  Spence J,  Padmore H,  Glaeser R M.  Design of an electron microscope phase plate using a focused continuous-wave laser,  New J. Phys. ,  2010, vol.  12 pg.  073011   doi:10.1088/1367-2630/12/7/073011 Google Scholar CrossRef Search ADS PubMed  51 Okamoto H.  Adaptive quantum measurement for low-dose electron microscopy,  Phys. Rev. A ,  2010, vol.  81 pg.  id. 043807  Google Scholar CrossRef Search ADS   52 Schröder R R,  Barton B,  Rose H,  Benner G.  Contrast enhancement by anamorphotic phase plates in an aberration corrected TEM,  Microsc. Microanal. ,  2007, vol.  13 (pg.  136- 137) 53 Fukuda Y,  Fukazawa Y,  Danev R,  Shigemoto R,  Nagayama K.  Tuning of the Zernike phase plate for visualization of detailed ultrastructure in complex biological specimens,  J. Struct. Biol. ,  2009, vol.  168 (pg.  476- 484)  doi:10.1016/j.jsb.2009.08.011 Google Scholar CrossRef Search ADS PubMed  54 Van Dyck D.  Wave reconstruction in TEM using a variable phase plate,  Ultramicroscopy ,  2010, vol.  110 (pg.  571- 572)  doi:10.1016/j.ultramic.2009.12.016 Google Scholar CrossRef Search ADS   © The Author 2011. Published by Oxford University Press [on behalf of Japanese Society of Microscopy]. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
TI - Another 60 years in electron microscopy: development of phase-plate electron microscopy and biological applications
JF - Journal of Electron Microscopy
DO - 10.1093/jmicro/dfr037
DA - 2011-08-01
UR - https://www.deepdyve.com/lp/oxford-university-press/another-60-years-in-electron-microscopy-development-of-phase-plate-KzFTL19Mo9
SP - S43
EP - S62
VL - 60
IS - suppl_1
DP - DeepDyve
ER -