Microscopy, 2014, 333–336 doi: 10.1093/jmicro/dfu017 Advance Access Publication Date: 19 May 2014
Technical Report
A simple way to obtain backscattered electron images in a scanning transmission electron microscope 1
Department of Electronics, 2Ecotopia Science Institute, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan, and 3Department of Electrical and Electronic Engineering, Meijo University, 1-501 Shiogamaguchi, Tempaku-ku, Nagoya 468-8502, Japan *To whom correspondence should be addressed. E-mail: [email protected] Received 4 March 2014; Accepted 23 April 2014
Abstract We have fabricated a simple detector for backscattered electrons (BSEs) and incorporated the detector into a scanning transmission electron microscope (STEM) sample holder. Our detector was made from a 4-mm2 Si chip. The fabrication procedure was easy, and similar to a standard transmission electron microscopy (TEM) sample thinning process based on ion milling. A TEM grid containing particle objects was fixed to the detector with a silver paste. Observations were carried out using samples of Au and latex particles at 75 and 200 kV. Such a detector provides an easy way to obtain BSE images in an STEM. Key words: backscattered electron, BSE, detector, STEM
Introduction Since a backscattered electron (BSE) signal depends on the mean atomic number, phases with different atomic numbers can be recognized by BSE imaging [1]. These imaging techniques are usually equipped with a scanning electron microscope (SEM). The high kinetic energy of BSEs enables them to penetrate thin films. This feature has been used to image small objects submerged in a liquid [2–4]. In this procedure, a thin film of silicon nitride (SiN) film can be used to seal the liquid solution. An imaging electron beam is incident through the film, and BSE signals are then detected for imaging. Because such techniques are typically based on SEM, the accelerating voltage available is up to ∼30 kV. Depending on the thickness of the sealing film, the spatial resolution may be limited owing to scattering of the imaging beam in the sealing film. To use higher accelerating voltage,
we use the scanning transmission electron microscope (STEM). However, BSE detectors are not commonly available for STEMs. There has been little demand for such detectors and they are difficult to incorporate owing to the limited space around the objective lens. In this report, we introduce a simple BSE detector that is easily incorporated into an STEM sample holder, and present some results for BSE imaging using an STEM electron beam with an accelerating voltage of up to 200 kV.
Methods Figure 1 is a schematic representation of our BSE detector. The BSE detector consists of p-type silicon (Si) and a Schottky contact. Transmission electron microscopy (TEM) grids were used to hold particle objects. The grid was placed just below the detector and conveniently held in place with a
© The Author 2014. Published by Oxford University Press on behalf of The Japanese Society of Microscopy. All rights reserved. For permissions, please e-mail: [email protected]
333
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
Hiroki Tsuruta1, Shigeyasu Tanaka2,*, Takayoshi Tanji2, and Chiaki Morita3
Microscopy, 2014, Vol. 63, No. 4
334
sample. BSEs entering the detector from the Schottky electrode surface create minority carriers in the Si chip. The minority carriers that reach the depletion region are driven by the built-in electric field and become a short-circuit current. This is the signal we measured. We used two types of samples. One consists of latex (⌀90 nm) and gold (Au, ⌀60 nm) particles on a carbon filmcoated grid. The other is made of Au (⌀60 nm) particles confined between two SiN membrane window grids. The former sample was prepared by dropping a small amount of latex on a grid, followed by a liquid solution of Au. The sample was then dried. The latter sample was prepared as follows: first, we dropped a small amount of Au in liquid solution on the surface of two membrane grids, which was then dried. The thickness of the membrane window was ∼100 nm, and the window size was ∼0.5 mm. The frame thickness was ∼200 μm. Next, the two grids were glued face-to-face with a single-hole TEM grid in-between the grids. The thickness of the single-hole TEM grid was ∼20 μm. The role of this grid is to ensure that there is sufficient space between the two membrane grids. The gap between the grids was then sealed by glue, trapping air from the atmosphere inside. Additionally, Au particles from the membranes are located in this gap. The schematic illustration of this sample is shown in Fig. 3.
Results and discussion
Fig. 1. A schematic representation of the BSE detector.
Fig. 2. Fabrication procedure of our BSE detector.
Dark-field (DF)-STEM and BSE images of the latex with the Au sample taken at 75 kV are shown in Figure 4a and b, respectively. In this image, the carbon film was faced towards the detector, as shown in Fig. 1, to simulate an environmental experiment. Both latex and Au particles are
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
silver paste. The distance between the bottom surface of the detector and the grid was estimated to be ∼500 μm. Electrical contact was made for signals from the detector to be available. The current signals from the detector were amplified and then acquired by a computer, along with the STEM and synchronization signals, and the BSE and STEM images were obtained. No bias voltage was applied to the detector. Observation experiments were performed using a Hitachi H-8000 STEM (accelerating voltage 75–200 kV). The beam current was ∼0.15 and 1.5 nA at 75 and 200 kV, respectively. The fabrication procedure of the detector is easy, similar to the TEM sample thinning process, as shown in Fig. 2. A square chip of ∼4 mm side length was cut from a Si wafer. First, an ohmic electrode was made by vacuum evaporation of Al and subsequent annealing. Then, a dimple was made in the bottom side of the sample chip. Next, a through-hole with a diameter of ∼300 μm was created by Ar ion milling. Finally, a thin Schottky electrode was made on this surface by vacuum evaporation of Ti. The principle of BSE detection is as follows: the imaging electron beam is incident on a sample through the hole in the detector. This beam creates BSEs via interaction with the
Microscopy, 2014, Vol. 63, No. 4
335
Fig. 3. The structure of the sealed sample. Au (⌀60 nm) particles (along with atmospheric particles) were confined between two SiN membrane window grids. The whole sample was fixed to the detector with a silver paste for BSE observation of the Au particles. Fig. 5. Au particles confined between the two SiN membrane window grids taken at 200 kV. (a) DF-STEM and (b) BSE images, respectively.
visible in the BSE image, and they are distinguishable according to their intensity. Au particles appear brighter than latex particles. The detector current at the bright Au particles was ∼130 nA. However, for the DF-STEM image, the difference in the intensities is not as noticeable, so that it is difficult to distinguish latex and Au particles. The DF-STEM image was formed by electrons scattered at angles larger than about 12 mrad. The reason for the low contrast between particles in the DF-STEM image is not clear. The latex particles seem to agglomerate to form a thick layer, and thus, more scattered electrons may be created from the latex particles. DF-STEM and BSE images of Au particles confined between the two SiN membrane window grids taken at 200 kV are shown in Figure 5a and b, respectively. These particles were located on the upper membrane. Since the kinetic energy of BSEs is as high as 200 kV, despite the presence of the 100 nm thick membrane, the Au particles can be seen clearly. The BSE detector current at the bright Au particles was ∼8 nA. There are several reasons for the low current signal measured in this sample, when compared with the images in Fig. 4. One reason is that the BSEs at 200 kV penetrate deeper into the
Concluding remarks We have fabricated a simple BSE detector and incorporated it into an STEM sample holder. Observations were carried out using samples of Au and latex particles at 75 and 200 kV, respectively. Such a detector provides an easy way to obtain BSE images in an STEM.
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
Fig. 4. Latex and Au samples taken at 75 kV. (a) DF-STEM and (b) BSE images, respectively.
detector. Minority carriers created deep in the detector are located far from the Schottky electrode. They recombine with the majority carriers before being transported to the Schottky junction by diffusion. Thus, these minority carriers cannot produce a current signal. Since the penetration depth in Si for 200 kV is much larger than that for 75 kV, the ratio of such lost minority carriers must be large for 200 kV. Another reason for the low current signal is that the backscattering probability is low for 200 kV, compared with 75 kV. Thirdly, there is a difference in angular distribution of BSEs at 75 and 200 kV. The BSEs at 200 kV are distributed at higher angles than those at 75 kV. Therefore, the collection efficiency of BSEs is larger for 75 kV than that for 200 kV. The presence of the 100-nm thick SiN membrane might affect the BSE image intensity. However, it does not seem to deteriorate the visibility of the Au particles in the STEM image. Therefore, we conclude that the effect of the membrane thickness on the BSE image intensity is small. Our BSE detector was conveniently fixed to the sample grid with a silver paste. However, it was possible to remove the detector from the sample grid with tweezers. Therefore, the detector was reusable until a breakage which happened by mistake. In this report, we did not perform observations of objects suspended in liquid. Nonetheless, such observations should be possible by using the cell structure that is depicted in Fig. 3. Towards this, experiments are ongoing.
336
References 1. Reimer L (1984) Scanning Electron Microscope. Springer Series in Optical Sciences (Springer-Verlag, Berlin Heidelberg New York Tokyo). 2. Bogner A, Thollet G, Basset D, Jouneau P, Cauthier C (2005) Wet STEM: a new development in environmental SEM for imaging nanoobjects included in a liquid phase. Ultramicroscopy 104: 290–301.
Microscopy, 2014, Vol. 63, No. 4 3. Suga M, Nishiyma H, Ebihara T (2009) Atmospheric electron microscope: limits of observable depth. Microsc. Microanal. 15: 294–295. 4. Nishiyama H, Suga M, Ogura T, Maruyama Y, Koizumi M, Mio K, Kitamura S, Sato C (2010) Atmospheric scanning electron microscope observes cells and tissues in open medium through silicon nitride film. J. Struct. Biol. 169: 438–449.
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
Technical Report
A simple way to obtain backscattered electron images in a scanning transmission electron microscope 1
Department of Electronics, 2Ecotopia Science Institute, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan, and 3Department of Electrical and Electronic Engineering, Meijo University, 1-501 Shiogamaguchi, Tempaku-ku, Nagoya 468-8502, Japan *To whom correspondence should be addressed. E-mail: [email protected] Received 4 March 2014; Accepted 23 April 2014
Abstract We have fabricated a simple detector for backscattered electrons (BSEs) and incorporated the detector into a scanning transmission electron microscope (STEM) sample holder. Our detector was made from a 4-mm2 Si chip. The fabrication procedure was easy, and similar to a standard transmission electron microscopy (TEM) sample thinning process based on ion milling. A TEM grid containing particle objects was fixed to the detector with a silver paste. Observations were carried out using samples of Au and latex particles at 75 and 200 kV. Such a detector provides an easy way to obtain BSE images in an STEM. Key words: backscattered electron, BSE, detector, STEM
Introduction Since a backscattered electron (BSE) signal depends on the mean atomic number, phases with different atomic numbers can be recognized by BSE imaging [1]. These imaging techniques are usually equipped with a scanning electron microscope (SEM). The high kinetic energy of BSEs enables them to penetrate thin films. This feature has been used to image small objects submerged in a liquid [2–4]. In this procedure, a thin film of silicon nitride (SiN) film can be used to seal the liquid solution. An imaging electron beam is incident through the film, and BSE signals are then detected for imaging. Because such techniques are typically based on SEM, the accelerating voltage available is up to ∼30 kV. Depending on the thickness of the sealing film, the spatial resolution may be limited owing to scattering of the imaging beam in the sealing film. To use higher accelerating voltage,
we use the scanning transmission electron microscope (STEM). However, BSE detectors are not commonly available for STEMs. There has been little demand for such detectors and they are difficult to incorporate owing to the limited space around the objective lens. In this report, we introduce a simple BSE detector that is easily incorporated into an STEM sample holder, and present some results for BSE imaging using an STEM electron beam with an accelerating voltage of up to 200 kV.
Methods Figure 1 is a schematic representation of our BSE detector. The BSE detector consists of p-type silicon (Si) and a Schottky contact. Transmission electron microscopy (TEM) grids were used to hold particle objects. The grid was placed just below the detector and conveniently held in place with a
© The Author 2014. Published by Oxford University Press on behalf of The Japanese Society of Microscopy. All rights reserved. For permissions, please e-mail: [email protected]
333
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
Hiroki Tsuruta1, Shigeyasu Tanaka2,*, Takayoshi Tanji2, and Chiaki Morita3
Microscopy, 2014, Vol. 63, No. 4
334
sample. BSEs entering the detector from the Schottky electrode surface create minority carriers in the Si chip. The minority carriers that reach the depletion region are driven by the built-in electric field and become a short-circuit current. This is the signal we measured. We used two types of samples. One consists of latex (⌀90 nm) and gold (Au, ⌀60 nm) particles on a carbon filmcoated grid. The other is made of Au (⌀60 nm) particles confined between two SiN membrane window grids. The former sample was prepared by dropping a small amount of latex on a grid, followed by a liquid solution of Au. The sample was then dried. The latter sample was prepared as follows: first, we dropped a small amount of Au in liquid solution on the surface of two membrane grids, which was then dried. The thickness of the membrane window was ∼100 nm, and the window size was ∼0.5 mm. The frame thickness was ∼200 μm. Next, the two grids were glued face-to-face with a single-hole TEM grid in-between the grids. The thickness of the single-hole TEM grid was ∼20 μm. The role of this grid is to ensure that there is sufficient space between the two membrane grids. The gap between the grids was then sealed by glue, trapping air from the atmosphere inside. Additionally, Au particles from the membranes are located in this gap. The schematic illustration of this sample is shown in Fig. 3.
Results and discussion
Fig. 1. A schematic representation of the BSE detector.
Fig. 2. Fabrication procedure of our BSE detector.
Dark-field (DF)-STEM and BSE images of the latex with the Au sample taken at 75 kV are shown in Figure 4a and b, respectively. In this image, the carbon film was faced towards the detector, as shown in Fig. 1, to simulate an environmental experiment. Both latex and Au particles are
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
silver paste. The distance between the bottom surface of the detector and the grid was estimated to be ∼500 μm. Electrical contact was made for signals from the detector to be available. The current signals from the detector were amplified and then acquired by a computer, along with the STEM and synchronization signals, and the BSE and STEM images were obtained. No bias voltage was applied to the detector. Observation experiments were performed using a Hitachi H-8000 STEM (accelerating voltage 75–200 kV). The beam current was ∼0.15 and 1.5 nA at 75 and 200 kV, respectively. The fabrication procedure of the detector is easy, similar to the TEM sample thinning process, as shown in Fig. 2. A square chip of ∼4 mm side length was cut from a Si wafer. First, an ohmic electrode was made by vacuum evaporation of Al and subsequent annealing. Then, a dimple was made in the bottom side of the sample chip. Next, a through-hole with a diameter of ∼300 μm was created by Ar ion milling. Finally, a thin Schottky electrode was made on this surface by vacuum evaporation of Ti. The principle of BSE detection is as follows: the imaging electron beam is incident on a sample through the hole in the detector. This beam creates BSEs via interaction with the
Microscopy, 2014, Vol. 63, No. 4
335
Fig. 3. The structure of the sealed sample. Au (⌀60 nm) particles (along with atmospheric particles) were confined between two SiN membrane window grids. The whole sample was fixed to the detector with a silver paste for BSE observation of the Au particles. Fig. 5. Au particles confined between the two SiN membrane window grids taken at 200 kV. (a) DF-STEM and (b) BSE images, respectively.
visible in the BSE image, and they are distinguishable according to their intensity. Au particles appear brighter than latex particles. The detector current at the bright Au particles was ∼130 nA. However, for the DF-STEM image, the difference in the intensities is not as noticeable, so that it is difficult to distinguish latex and Au particles. The DF-STEM image was formed by electrons scattered at angles larger than about 12 mrad. The reason for the low contrast between particles in the DF-STEM image is not clear. The latex particles seem to agglomerate to form a thick layer, and thus, more scattered electrons may be created from the latex particles. DF-STEM and BSE images of Au particles confined between the two SiN membrane window grids taken at 200 kV are shown in Figure 5a and b, respectively. These particles were located on the upper membrane. Since the kinetic energy of BSEs is as high as 200 kV, despite the presence of the 100 nm thick membrane, the Au particles can be seen clearly. The BSE detector current at the bright Au particles was ∼8 nA. There are several reasons for the low current signal measured in this sample, when compared with the images in Fig. 4. One reason is that the BSEs at 200 kV penetrate deeper into the
Concluding remarks We have fabricated a simple BSE detector and incorporated it into an STEM sample holder. Observations were carried out using samples of Au and latex particles at 75 and 200 kV, respectively. Such a detector provides an easy way to obtain BSE images in an STEM.
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
Fig. 4. Latex and Au samples taken at 75 kV. (a) DF-STEM and (b) BSE images, respectively.
detector. Minority carriers created deep in the detector are located far from the Schottky electrode. They recombine with the majority carriers before being transported to the Schottky junction by diffusion. Thus, these minority carriers cannot produce a current signal. Since the penetration depth in Si for 200 kV is much larger than that for 75 kV, the ratio of such lost minority carriers must be large for 200 kV. Another reason for the low current signal is that the backscattering probability is low for 200 kV, compared with 75 kV. Thirdly, there is a difference in angular distribution of BSEs at 75 and 200 kV. The BSEs at 200 kV are distributed at higher angles than those at 75 kV. Therefore, the collection efficiency of BSEs is larger for 75 kV than that for 200 kV. The presence of the 100-nm thick SiN membrane might affect the BSE image intensity. However, it does not seem to deteriorate the visibility of the Au particles in the STEM image. Therefore, we conclude that the effect of the membrane thickness on the BSE image intensity is small. Our BSE detector was conveniently fixed to the sample grid with a silver paste. However, it was possible to remove the detector from the sample grid with tweezers. Therefore, the detector was reusable until a breakage which happened by mistake. In this report, we did not perform observations of objects suspended in liquid. Nonetheless, such observations should be possible by using the cell structure that is depicted in Fig. 3. Towards this, experiments are ongoing.
336
References 1. Reimer L (1984) Scanning Electron Microscope. Springer Series in Optical Sciences (Springer-Verlag, Berlin Heidelberg New York Tokyo). 2. Bogner A, Thollet G, Basset D, Jouneau P, Cauthier C (2005) Wet STEM: a new development in environmental SEM for imaging nanoobjects included in a liquid phase. Ultramicroscopy 104: 290–301.
Microscopy, 2014, Vol. 63, No. 4 3. Suga M, Nishiyma H, Ebihara T (2009) Atmospheric electron microscope: limits of observable depth. Microsc. Microanal. 15: 294–295. 4. Nishiyama H, Suga M, Ogura T, Maruyama Y, Koizumi M, Mio K, Kitamura S, Sato C (2010) Atmospheric scanning electron microscope observes cells and tissues in open medium through silicon nitride film. J. Struct. Biol. 169: 438–449.
Downloaded from http://jmicro.oxfordjournals.org/ at Serials Department on October 15, 2014
