Time-resolved measurement of the three-dimensional motion of gold nanocrystals in water using diffracted electron tracking

Ultramicroscopy 140 (2014) 1–8

Contents lists available at ScienceDirect

Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic

Time-resolved measurement of the three-dimensional motion of gold nanocrystals in water using diffracted electron tracking Naoki Ogawa a,b,n, Yasuhisa Hirohata a, Yuji C. Sasaki c, Akira Ishikawa d,nn a Department of Integrated Science in Physics and Biology, College of Humanities and Sciences, Nihon University, 3-25-40 Sakurajosui, Setagaya-ku, Tokyo 156-8550, Japan b Graduate School for Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan c Graduate School of Frontier Sciences, University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8561, Japan d Department of Physics, College of Humanities and Sciences, Nihon University, 3-25-40 Sakurajosui, Setagaya-ku, Tokyo 156-8550, Japan

art ic l e i nf o

a b s t r a c t

Article history: Received 13 May 2013 Received in revised form 27 December 2013 Accepted 27 January 2014 Available online 5 February 2014

We introduce diffracted electron tracking (DET), which combines two electron microscopy techniques, electron backscatter diffraction and the use of an environmental cell in a scanning electron microscope to measure changes in nanocrystal-orientation. The accuracy of DET was verified by measuring the motion of a flat gold crystal caused by the rotation or tilting of the specimen stage. DET was applied to measure the motion of semifixed gold nanocrystals in various environments. In addition to large motions induced in water environment, DET could detect small differences in the three-dimensional (3D) motion amplitude between vacuum environment and an Ar gas environment. DET promises to be a useful method for measuring the motion of single nanocrystals in various environments. This measuring technique may be used in a wide range of scientific fields; for example, DET may be a prospective method to track the single molecule dynamics of molecules labeled with gold nanocrystals. & 2014 Elsevier B.V. All rights reserved.

Keywords: DET Single molecular dynamics Three-dimensional motion Environmental cell SEM EBSD

1. Introduction Recently, biophysical research has benefited from numerous analytical tools designed to measure single molecule dynamics [1]. Among the most commonly used techniques is laser optical microscopy, which measures the motion of fluorescently labeled single molecules [1]. Motion can also be quantified using a high-speed atomic force microscope (AFM) [2] or using diffracted X-ray tracking (DXT) methods [3,4]. Previously, our research group had applied DXT to the dynamics of single-protein molecules [4,5]. In this method, the rotation of a single nanocrystal, which is linked to a specific site in the molecule, is tracked using a time-resolved Laue diffraction

Abbreviations: DET, Diffracted electron tracking; 3D, Three-dimensional; AFM, Atomic force microscope; DXT, Diffracted X-ray tracking; EC, Environmental cell; SEM, Scanning electron microscope; EBSD, Electron backscatter diffraction; EBSP, Electron backscatter diffraction pattern; SE, Secondary electron; TAC, Triacetyl cellulose; MOPS, 3-Morpholinopropanesulfonic acid; CHAPS, 3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate; CI, Confidence index; MSD, Meansquare displacement n Corresponding author at: Graduate School for Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan. nn Corresponding author. E-mail addresses: [email protected] (N. Ogawa), [email protected] (A. Ishikawa). http://dx.doi.org/10.1016/j.ultramic.2014.01.012 0304-3991 & 2014 Elsevier B.V. All rights reserved.

technique. However, because DXT requires a very strong white X-ray source, in a large institution, competition for the use of the machine by many researchers is fierce. To alleviate this problem, we have developed a compact instrument that uses electron beams in place of X-rays. The nanocrystal-orientation is measured using electron backscatter diffraction (EBSD) instead of X-ray diffraction spots. We refer to this measuring technique as diffracted electron tracking (DET). DET can be performed with a commercially available scanning electron microscope (SEM) in a small laboratory setting and provides instantaneous experimental feedback on the measurement results. In DXT, the Laue diffraction from a gold nanocrystal appears as bright spots on a fluorescent screen. Conversely, in DET, the electron diffraction produces a widespread band pattern, termed an electron backscatter diffraction pattern (EBSP), which contains 3D information regarding the specimen movement. A further advantage of DET is that it allows in situ observation of a specimen. Because nanocrystals can be observed and identified directly using the secondary electron (SE) imaging mode of an SEM, an EBSP can be obtained from targeted gold nanocrystals of identical sizes (approximately 40–60 nm), excluding larger or aggregated gold nanocrystals. In this study, the accuracy of DET was verified by tracking the motion of a flat gold crystal caused by the rotation and tilting of the SEM specimen stage. The motion of gold nanocrystals in various circumstances—in vacuum, in Ar gas and in water—was

2

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

measured using DET. DET is a promising method for single molecule observations.

2. Materials and methods 2.1. Flat gold crystal and verifying the performance of DET A flat gold crystal was grown on a silica glass surface using a vacuum evaporation method [6]; only the evaporation time differed from the procedure used to grow the gold nanocrystals. The flat gold crystal was attached to a general specimen stub with conductive double-sided tape (Nissin EM Ltd.) and silver conductive paste (Silvest P-255; Tokuriki Chemical Research Ltd.). The SEM (JSM-7001F; JEOL Ltd.) specimen stage was rotated from 01 to 31 in the clockwise direction or tilted from 701 to 671 relative to the home angle in 0.21 steps, and the EBSP signal was collected at each step at the same crystallographic face. To detect the EBSP signal, we used a 30 kV accelerating voltage, an 87 pA beam current, and a 17 mm working distance. The lattice orientations were calculated in the same manner as for the gold nanocrystals, which is described in Section 2.4. Each measurement was performed for three different crystal faces. 2.2. Carbon sealing-film for the environmental cell (EC) The sealing-film for an EC must be sufficiently thin to allow electron transmission and sufficiently tough to withstand the pressure differences in excess of atmospheric pressure. The development of the 20 nm thick carbon sealing-film used in this study has been previously described [7]. Briefly, a 1% (w/v) triacetyl cellulose (TAC) (Okenshoji Ltd.) solution was prepared by dissolving TAC in a mixed solvent that contained 90% (v/v) 1,2-dichroloethane and 10% (v/v) methanol. To fabricate the TAC thin film that was used as a support for the carbon film, a glass slide was soaked in the TAC solution and extracted at 2.5 mm/s. A 20 nm thick carbon film was deposited onto the TAC film on the glass slide using the vacuum evaporation method. The thickness of the carbon film was chosen as a compromise between electron permeability and the ability to withstand pressure. The carbon/ TAC combination film was cut into 2
2 mm square pieces on the glass slide. The combination-film pieces were drawn from the glass slide onto the surface of a water bath by surface tension. Each piece of combination film was scooped up with a copper ring (diameter ¼5 mm), dipped in water and placed on a special threeslit grid made of phosphor bronze (Fig. 3c), set on a piece of wet filter paper, which is described in detail in Section 3.3. After drying, the TAC film was dissolved at the slits of the grids by soaking the grids in acetone. The ability of the carbon sealing-film covering the slits to withstand pressure was evaluated using a standard pressure test. Once it was confirmed that the carbon film on each grid could withstand up to 1.3 times atmospheric pressure, an EC was sealed with the grid. 2.3. Specimen preparation for DET Gold nanocrystals were grown on a cleaved NaCl (100) surface using the vacuum evaporation method [6]. The sealing-film on the three-slit grid (Fig. 3c) was processed with a mercapto-silane coupling reagent (KBM-803; Shin-Etsu Chemical Ltd.) using the vapor deposition method [8]. Briefly, 1 ml of mercapto-silane was evaporated to vapor pressure in a well-sealed container. The grids were left standing for 1 h in the mercapto-silane vapor. After incubation, the grids were heated in air at 95 1C for 5 min, after which they were stored under vacuum until they were used. To ensure that thiol group of the mercapto-silane would react with the gold nanocrystals, the grids

were made hydrophilic by exposure to glow discharge (HDT-400; JEOL Ltd.). To prevent the gold nanocrystals from aggregating on the sealing-film, the side of the gold nanocrystals that had been in contact with the NaCl crystals was exposed to a thin buffer solution (50 mM MOPS–NaOH (pH 8.0), 50 mM CHAPS) layered onto the sealing-film for 10 min. After incubation, the remaining NaCl crystal was removed by soaking in distilled water. The EC that contained a water layer was assembled as follows. The three-slit grid with carbon sealing-film and gold nanocrystals was enclosed with 0.2 μl of degassed distilled water as shown in Fig. 3a and b. The EC that contained Ar gas was assembled in an Ar-gas-filled glove box with no water present. The vacuum EC was prepared by removing the O-ring that was used to seal the EC. Except for the vacuum EC, the sealing of the ECs was checked by observing the SEM images of the ECs (Fig. 4a). 2.4. Tracking the 3D motion of gold nanocrystals using DET To detect the distribution of the gold nanocrystals, an SEM image was first observed with an SE detector by scanning the minimum necessary electron beam over the specimen. Each individual gold nanocrystal of suitable size (approximately 40–60 nm in diameter) was chosen from the SEM image, which was obtained at magnification of 50,000 (Fig. 4c). To detect the EBSP signal from a targeted gold nanocrystal, the electron beam was fixed at an edge point on the crystal for 2 s. The irradiation conditions were as follows: a 30 kV accelerating voltage, an 87 pA beam current and a 17 mm working distance at an angle of 701 relative to the normal direction of the specimen stage. The EBSD was projected on the phosphor-screen detector, and the produced EBSP was recorded with a CCD camera after being intensified with an image intensifier (V8070U-74; Hamamatsu Photonics Ltd.), which was operating at a shutter speed of 60 ms. Thus, 26 frames of time-resolved sequential EBSP images were obtained from each gold nanocrystal. The orientations of the gold nanocrystals were calculated from the EBSPs using OIM™ Data Collection version 5.31 and OIM™ Analysis version 5 software (EDAX; AMETEK, Inc.). Both software packages were used as directed by their instruction manuals and nearly automatically determined the orientations of the lattices of the gold nanocrystals. To eliminate the pseudoorientations that appeared in the results because of low signal quality, we set the criterion that the confidence index (CI) value [9] must be greater than 0.2 (average of sequential 26 data). Moreover, the results that contained incorrect orientations were estimated by averaging the values of adjacent steps. To analyze the motion of the gold nanocrystals, the time-resolved lattice orientations in terms of Euler angles [9] were calculated using OIM™ Analysis. The Euler angles indicate the rotational relationship between the specimen coordinate system Cs (Xs, Ys, Zs) (in which the specimen is spatially fixed) and the crystal coordinate system Cc (Xc, Yc, Zc) (in which the gold nanocrystal is fixed). The Euler angles (ϕ1 ; Φ; ϕ2 ) are defined as follows. The Cs system can be made coincident with the Cc system by a three-phase rotational operation. First, Cs is rotated by ϕ1 around the Zs axis, which corresponds to the normal direction of the specimen stage, so that the Xs axis lies on the (Xc, Yc) plane. Then, the resultant Cs is rotated by Φ around the new Xs axis to make the Zs axis coincident with the Zc axis, and finally, the resultant Cs is rotated by ϕ2 around the new Zs axis, namely, the Zc axis, to make the Cs axis coincident with the Cc axis. Thus, in terms of the Euler angles, the orientation matrix g of the crystal can be given as follows: EBSD is commonly used for the analysis of the crystal-orientation mapping of a polycrystalline material by electron beam scanning on the material surface. However, we used the EBSD approach to investigate the motion of the nanocrystals, not via spatial scanning, but via time-sequential beam irradiation on a fixed point of each nanocrystal. In this way, we were able to obtain the changes in the crystal-orientation, namely, the Euler angles, from the EBSP and the

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

rotation angles of the nanocrystal. From the Euler angles, the singleaxis rotation angle ω and the angles α, β, and γ of the principal ! ! ! lattice vectors a , b , and c , respectively, of the gold nanocrystal were deduced for adjacent EBSP frames that were recorded by the !! ! CCD camera. The vectors a , b , and c correspond to the Xc, Yc, and Zc axes of Cc, respectively. To obtain the amplitude of the rotation angle during a time interval, the rotation matrix Δgij between the ith and jth time step can be calculated from the orientation matrices gj and gi 1 using the following formula:

Δgij ¼ g j g i 1

ð2Þ

! ! ! for vectors a , b , and c as follows: 0! 1 0 1 0 ai lai mai nai 1 0 B! C B C B b C ¼ @ lbi mbi nbi A ¼ g B @0 1 @ iA ! l m n 0 0 ci ci ci ci

3

1 0 C 0A

ð4Þ

1

Angles α, β, and γ can be calculated from the cosine law as follows: 0 1 0 1 0 1 ðlai  laj Þ2 þ ðmai  maj Þ2 þ ðnai  naj Þ2 cos αi 1 C B cos β C B C 1B ðlbi  lbj Þ2 þ ðmbi  mbj Þ2 þ ðnbi  nbj Þ2 C @ i A ¼ @ 1 A B A 2@ cos γ i 1 ðlci  lcj Þ2 þ ðmci  mcj Þ2 þ ðnci  ncj Þ2 ð5Þ

The single-axis rotation angle ωij between time steps i and j, which is referred to as the misorientation in orientation mapping, is given by

where (li, mi , ni ) and (lj, mj, nj) are the directional cosine matrices of the principal lattice vectors at time steps i and j, respectively.

2 cos ωij ¼ TraceðΔg ij Þ  1

2.5. Mean-square displacement curves

ð3Þ

To obtain 3D information regarding the nanocrystal motion, the angles α, β, and γ were calculated from the directional cosine matrices of the ! ! ! principal lattice vectors a , b , and c , respectively, of the gold nanocrystal, which were calculated from the Euler angles. The directional cosine matrices (li, mi, ni) of the lattice vectors at the ith step, after rotation by rotation matrix g i , can be calculated

For analysis of the motion of the gold nanocrystals, we plotted the amplitude mean-square displacements (MSD) curves [10] of the values of ω and of α, β, and γ that were calculated for various time steps. For the time interval between the ith step and the jth step, these values were determined as described in Section 2.4. The MSDs of these parameters were calculated and data from approximately 300

Electron beam (30 keV)

Band shift by tilting of lattice

EBSD

Lattice structure

Fluorescent screen

ω

Actual EBSD pattern

Fig. 1. Principle of DET and the parameters to be measured. (a) When the electron beam irradiates a nanocrystal, inelastically scattered primary electrons arrive at the Bragg angle for a set of specific planes and form a band pattern on the fluorescent screen. This band indicates the orientation of the source plane and shifts in relation to the nanocrystal motion. Essentially, bands are formed by every set of lattice planes and collectively constitute the EBSP. On the basis of this principle, the 3D motion of nanocrystals can be tracked using the sequential changes in the EBSP. (b) A measured EBSD pattern of a gold nanocrystal. Each band resulted from a pair of parallel lines (Kikuchi lines) corresponding to the crystal lattice planes. From this pattern, the orientation of the gold lattice could be determined as illustrated. (c) The rotation angle ! ! ! ω around a single axis and the rotation angles α, β, and γ of the principal lattice vectors a , b , and c of the nanocrystal, respectively, between time steps i and i þ1. The angle ω provides a suitable comparison of the magnitude of the motion regardless of the nanocrystal-orientation, while the angles α, β, and γ provide directional information regarding the motion.

4

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

nanocrystals were used to plot each MSD for the time intervals. The MSD curve of ω represents the amplitude of the nanocrystal motion, while the MSD curves of α, β, and γ represent the 3D information of the nanocrystal motion. When the MSD curves of α, β, and γ are nearly coincident with each other, the motion of the nanocrystal should be random. The best-fit curve to the measured MSD curve was approximated using the method of least squares, as described previously [10].

3. Results and discussion 3.1. Principle of DET Fig. 1 shows the principle of DET and the parameters to be measured. When a specimen is irradiated with an electron beam,

some of the incident primary electrons are deflected backward from the specimen surface. These electrons, termed backscattered electrons, are used for SEM imaging and crystallographic analysis. After striking crystalline specimens, primary electrons are inelastically scattered and diffused in all directions. Some of these primary electrons intercept crystals at the Bragg angle θB for every set of lattice planes and are elastically scattered to provide a strong reinforced beam. Because electron diffraction through the Bragg angle occurs in all directions, two cones of radiation result from each family of planes, as shown in Fig. 1a. When the diffraction cones strike the fluorescent screen, multiple pairs of parallel conic sections are formed, which collectively constitute the EBSP (Fig. 1b). The EBSP moves across the fluorescent screen as the crystal rotates. From the EBSP dynamics, the orientation of the crystallographic axis of the specimen can be determined.

Flat gold crystal Crystal glass basal plate

Pole piece of Objective lens

Carbon conductive double-sidedtape

Electron beam

Silver conductive paste

Specimen stab

Brass spesimen stub

Specimen stage (Tilted 70°)

2 μm

Fluorescent screen

Image intensifier

Electron beam(30 keV)

Tilt

Fluorescent screen and CCD camera

EBSD

Tilt Flat gold crystal Crystal glassbasal plate

Rotation

Brass specimen stab

CCD camera

3

Integrated value of α ,β,γ angle (deg.)

Integrated value of ω angle (deg.)

Fluorescent screen

2

1

0

3

2

1

0 0

5

10

Rotating steps (0.2 degree/step)

15

0

5

10

Tilting steps (0.2 degree/step)

15

0

5

10

15

Rotating steps (0.2 degree/step)

Fig. 2. Verification of DET using a flat gold crystal. The lattice orientation of the flat gold crystal measured using DET was investigated as a function of the travel distance of the specimen stage. (a) A photograph of the flat gold crystal. The flat gold crystal was attached to a specimen stub with conductive materials. (b) An SE image of the flat gold crystal. Each flat surface had the same lattice orientation. (c) The flat gold crystal on the specimen stage (tilted at an angle of 701 to the phosphor screen) and the EBSD detection system. (d) The custom-designed EBSD detection system is equipped with an image intensifier and a CCD camera for recording the EBSP. The sensitivity is ten times higher than that of standard systems. (e) A schematic diagram that depicts the measurement of the lattice orientation of the flat gold crystal using DET. The EBSD signals were obtained while rotating or tilting the specimen stage in 0.21 steps. (f) The rotation angle ω of the flat gold crystal varies as the specimen stage on which the crystal is held is tilted and rotated. The results of this experiment exhibited a close agreement with the theoretical values for both tilt and rotation. (g) An illustration of the variation of the rotation angles α, β, and γ of the flat gold crystal with the tilting and rotating of the specimen stage. From the results, the validity of using α, β, and γ as the parameters that describe the rotational direction of the specimen was confirmed.

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

For adjacent frames, we measured the rotation angle ω relative to a single axis and the rotation angles α, β, and γ of the principal ! ! ! lattice vectors a , b , and c , respectively, of each gold nanocrystal. The angles ω and α, β, and γ are visualized in Fig. 1c. From the time sequence of the measured rotation angles, the motion of the unit cell of the crystal can be derived. 3.2. Verification of DET The most important goal of DET is to track the time-resolved changes in the orientation of a gold crystal lattice in a precise manner. To estimate the accuracy of DET, a control measurement of a flat gold crystal (on the order of a micrometer) was performed in vacuum as follows. The flat gold crystal was attached to a normal specimen stub in such a way as to ensure electrical conductivity using a conductive carbon double-sided tape and silver conductive paste. A photograph of the flat gold crystal from a diagonal upward view is shown in Fig. 2a. Fig. 2b shows an SE image of the flat gold crystal. This crystal had the same crystallographic orientation at each flat surface region (approximately 4 μm squares). Fig. 2c shows the internal configuration of the SEM (JSM-7001; JEOL Ltd.). The flat gold crystal on the specimen stub was tilted at an angle of 701 for EBSD detection. The working distance of the specimen stage was 17 mm, and the distance from the crystal to the fluorescent screen of the EBSD detector was approximately 40 mm. Fig. 2d illustrates the EBSD detection system (DVC1412-FW-T1-EX) attached to the SEM. This system was custom designed by TSL Solutions. The image intensifier (V8070U-74; Hamamatsu Photonics K.K.) renders this system more sensitive than standard systems by a factor of 10. Fig. 2e shows a schematic view of the verification experiments for DET. First, an SE image of the flat gold crystal was obtained via electron beam scanning with an SE image detector, as shown in the photograph in Fig. 2b. Then, an irradiation spot was chosen from this image. The electron beam spot was directed at the center of a preferably broad flat surface using a beam of 30 keV and 87 pA. With the electron beam focused on the flat gold crystal, the EBSP was projected onto the fluorescent screen, amplified by the image intensifier and recorded with the CCD camera. According to the SEM specification sheet, the spot size at 30 keV is approximately 2 nm. This size was sufficient for the spotted irradiation of the electron beam at the same crystallographic face for every measured tilt or rotation of the specimen stage. After recording the EBSP and determining the orientation of the gold lattice, the SEM specimen stage was tilted or rotated in 0.21 steps. At each step, the ω angle was determined as discussed in Section 2.4. Each integrated value of the ω angle was plotted, as shown in Fig. 2f. The theoretical

washer

Grid

value for this experiment using 15 steps is 31 (0.21
15). The results of this experiment revealed a close agreement with the theoretical value for both tilt and rotation. The slight difference between tilt and rotation may be thought of as a mechanical error of the travel distance of the specimen stage. Therefore, this method could be also used for very precise calibration in the field of machine engineering. As shown in Fig. 2g, the values of α, β, and γ were obtained from the flat gold crystal by tilting and rotating the specimen stage in 0.21 steps, and the integrated values were plotted. The rotating and tilting mechanisms correspond to rotation around the Zs and Ys axes of Cs, respectively. The results strongly agree with our expectations, thereby confirming the validity of using α, β, and γ as the parameters that represent the rotational direction of the specimen. From the results, we conclude that DET can measure the 3D motion of a crystal regardless of the true crystal-orientation. 3.3. ECs for DET Fig. 3 shows the configurations of the ECs in water, Ar gas and under vacuum conditions. Fig. 3a shows a photograph of an assembled EC from a diagonal upward view. The three ellipsoidal slits at the top center are windows to allow the passage of electrons. The rear sides of these windows are sealed with a carbon film (20 nm in thickness) to prevent liquid or gas leakage from the EC. The film acts as a barrier against atmospheric air pressure from the high-vacuum conditions of the SEM column. Moreover, in this study, we used this carbon sealingfilm as a support for the gold nanocrystals, as shown in Fig. 4d. Fig. 3b shows the schematic cross section of the EC. A specially designed small O-ring (inner diameter¼2.0 mm and thickness¼0.45 mm) was placed on the upper surface of the brass holder. A phosphor–bronze grid designed for EBSD measurements was tilted at 701 to the incident beam. The grid (diameter¼3.5 mm and thickness¼ 0.2 mm) has three slits (each of length¼ 0.3 mm, width¼ 0.1 mm, and depth¼ 0.02 mm) to allow the electron beam to pass through the center of the grid, as shown in Fig. 3c. Previously, many types of ECs have been reported by many groups [11]. Recently, silicon-nitride has been widely used for EC windows. Silicon-nitride is obtainable from commercial companies, in contrast with our hand-made carbon sealing-film. We tested a silicon-nitride window for use in an EC for DET in a preliminary stage of development; the silicon-nitride window was too fragile when used in water EC. Therefore, we chose to use the carbon sealing-film, which we have developed over the past two decades, instead of siliconnitride. In future, if the manufacturing technology of silicon-nitride develops to the point that it can produce sufficiently robust films for use in an EC for DET, it may be a suitable candidate for this purpose. Furthermore, the use of graphene as a liquid-enclosure material for use in a transmission electron microscope has been reported [12].

screw O ring

Carbon sealing-film (thickness ~ 20 nm) 0.1 mm

8.0 mm

Carbon sealingfilm

5

Water, Ar or Vacuum condition

0.3 mm

0.2 mm 12.3 mm φ

3.5 mm φ

Fig. 3. Construction of environmental cells that enable the SEM to obtain the EBSPs of nanocrystals in various environments. (a) Photographs of an assembled EC. The three ellipsoidal slits at the top center are sealed with a carbon sealing-film (20 nm in thickness) on the rear side. This thin film is able to endure the pressure difference between vacuum and atmospheric pressure, and the electron beam and the signal from the specimen can pass through it easily. (b) A schematic cross section of the environmental cell. Gold nanocrystals were fixed beneath the sealing-film, which was processed with the thiol groups of mercapto-silane. The cell, in which the specimen is housed in the desired environment, is indicated by arrow. The cell is sealed with a special grid containing three slits covered with carbon sealing-film. (c) Photographs of both sides of the grid showing the sealing-film, the grid, and the slit dimensions.

6

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

10 µm

100 µm

500 nm

Electron beam (30 keV)

EBSD Gold nanocrystal (approx. 40 to 60 nm of diameter)

Fluorescent screen and CCD camera Carbon sealing-film (20 nm thickness) Mercapto-silane polymer layer

Fig. 4. The motions of gold nanocrystals were measured by DET. Shown here are SEM images and EBSPs of gold nanocrystals in water sealed in an EC tilted at 701. (a,b) Swelling of the sealing-film caused by the difference between the internal and external pressure, and the thick water sections beneath the film appear brighter in the slit regions. (c) Gold nanocrystals in water. DET was performed only on isolated individual gold nanocrystals. (d) Cross-sectional model of DET. The gold nanocrystals were semifixed on the carbon film using the thiol groups of mercapto-silane polymer. (e) Example of EBSP obtained from gold nanocrystals recorded at a shutter speed of 60 ms under an electron beam of 30 kV and 87 pA. Although the EBSP was more blurred than the most commonly observed EBSPs because of the low electron dose and high shutter speed, the orientations of the gold nanocrystals could be determined from signals of this quality using our processing method.

At this time, the intact area of a graphene sheet is too small to cover the three-slit EC windows, but if large-area intact graphene becomes available for use, it may be used in place of our sealing-film. The crystallinity of a sealing-film material may be an important factor in determining its suitability for use in an EC for DET because diffraction from the sealing-film may obscure the diffraction signal from the gold nanocrystals, causing the signal-to-noise ratio to decrease drastically. In this respect, our amorphous carbon sealingfilm has good properties for DET. 3.4. DET measurement of the motion of gold nanocrystals in various environments The motions of the gold nanocrystals were tracked via DET in three environments: water, Ar gas, and vacuum. The gold nanocrystals, which were grown on the cleaved surface of an NaCl crystal [6], were transferred to the carbon sealing-film of the EC, where they were processed with a mercapto-silane coupling reagent. The NaCl surface was dissolved in a buffer solution (50 mM MOPS  NaOH (pH 8.0), 50 mM CHAPS). The gold crystals were immersed in a thin layer of degassed distilled water set onto the sealing-film, and the EC was sealed. Fig. 4a and b shows SEM images of the slit portion of the water-filled EC, tilted at 701 for the EBSD measurements. The pressure difference between the inside and outside of the EC caused the carbon sealing-film to expand. The thicker sections of the water appear brighter in the image. Fig. 4c and d shows the magnified immersed gold crystals and the cross section of the EC when the motion of the gold nanocrystals in water is tracked using DET, respectively. The gold nanocrystals were fabricated by vacuum evaporation, as

mentioned above, so the shape of each individual particle was a nearly oval sphere [6]. Moreover, our SE image revealed that many isolated individual particles were bound to much of the surface of the carbon sealing-film. These results indicate that a comparatively large number of thiol groups of the mercaptosilane were involved in binding to the gold nanocrystals. DET was performed using an electron beam operating at 30 keV and 87 pA. The shutter speed of the CCD camera was 60 ms/frame. Each measurement was collected for a duration of 2 s because contamination occurred at the beam-irradiated site on the vacuum side, affecting the detection of the EBSP signal and causing the EBSP signal to deteriorate with time. Electron irradiation damage, especially for biological specimens such as proteins, is usually a major limiting factor of electron beam measurements. In DET, however, the electron beam can be focused to a nanospot and controlled to irradiate only a nanocrystal that is attached to the protein molecule. Therefore, EBSP can, in principle, be applied without exposing the protein to excessive irradiation. We wished to further protect the specimen by reducing the electron beam current to the minimum required level for EBSP. To this end, the EBSP detector attached to our SEM is equipped with a specially designed image intensifier through which the image passes before it is recorded by the CCD camera, as illustrated in Fig. 2d. From these considerations, we chose the electron beam conditions of 30 keV and 87 pA, so the irradiated electron dose for single DET measurement is calculated as 174 pC in 2 s. A typical example of an EBSP frame obtained from a gold nanocrystal is shown in Fig. 4e. An example of the time-sequential EBSP shift caused by gold-nanocrystal movement and the corresponding motion of the unit cell of the crystal is shown in Fig. 5. This motion

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

EBSP

Unit cell of the gold nanocrystal

Time course 0.00 s

0.18 s

0.36 s

0.54 s

0.72 s

0.90 s

1.08 s

7

value that was obtained for the immobile flat gold crystal, as discussed in Section 3.2, is also shown in Fig. 6a. The motions of the gold nanocrystals in vacuum and in Ar gas were very small. The MSD curve collected for the gold nanocrystals in water, however, appears to be parabolic, indicating that the motion was of the type known as directed Brownian motion [13]. Several types of forces can affect this motion. One of these is the pressure force of the electron beam radiation; in fact, the pressure force of the X-ray beam has been demonstrated in DXT [14]. The observed difference between “Not Moving” and “Vac” should be attributable to the electron-pressure force. However, “Ar gas” and “Water” should be affected by not only the electron-pressure force but also collision with environmental molecules. In fact, these molecules exhibit thermodynamically based Brownian motion [15]. Moreover, it is possible that this directed Brownian motion can be reinforced by the electron beam or the disturbance of the water molecules [16,17]. In conclusion, using MSD analysis, we can sufficiently estimate the differences in the motional magnitude among the “Not Moving” “Vac” and “Ar gas” cases. Thus, DET can detect 3D motion differences in various environments at an angular detection limit of less than 10 mrad. DET therefore promises to be a useful method for tracking the motion of single molecules labeled with gold nanocrystals. The damage of the specimen by the electron irradiation is one of the most serious problems (for a fragile specimen). To solve this, we will have to develop a more sensitive EBSP detection system to reduce the electron dose. Moreover, some recent reports have demonstrated methods of acquiring high-angular-resolution EBSPs [18,19]. The use of these methods will facilitate the improvement of DET detection limits. For gold nanocrystals, whose size and shape are uniform, the accuracy of DET is also expected to be improved.

4. Conclusions

1.26 s

Fig. 5. An example of the time-sequential EBSP shift (left column) caused by goldnanocrystal movement and the corresponding motion of the unit cell of the crystal (right column). The time from the beginning of the measurement is indicated on the right-hand side. The actual data were obtained at 60 ms per frame, so two frames are skipped between each pair of shown frames. This motion of this ! nanocrystal was primarily composed of rotation around Ys ( b axis).

of the nanocrystal was primarily composed of rotation around Yc ! ( b ) axis. 3.5. Statistical analysis of DET measurements The Euler angles were obtained from the EBSPs using the OIMTM Data Collection and OIMTM Analysis software packages (EDAX; AMETEK, Inc.). The Euler angles indicate the rotational relationships between the specimen coordinate system Cs and the crystal lattice coordinate system Cc. The specimen movement was analyzed using the Euler angles by incrementally calculating the above-mentioned rotation angles ω and α, β, and γ of the nanocrystals (time step¼ 60 ms). From approximately 300 individual gold nanocrystals, the MSDs of the rotation angles were obtained via statistical analysis at each step. Fig. 6 shows the variation of the DET-derived MSDs of the nanocrystals across the time steps. The MSD measurements of the ω values of the gold nanocrystals in vacuum, Ar gas, and water are shown in Fig. 6a. Each MSD curve, except “Not Moving”, was obtained from approximately 300 independent gold nanocrystals. To illustrate the accuracy of the DET method, the MSD of the ω

We present a newly developed method of measuring single molecule dynamics under various conditions, which is named DET. DET can be performed using a commercially available scanning electron microscope equipped with only a highly sensitive EBSD detector. To verify the performance of DET, we measured the motion of a flat gold crystal caused by the rotation or tilting of the specimen stage. The DET results corresponded to the motion of the specimen stage. It was confirmed that the change in the crystal-orientation measured by DET was in good agreement with the actual rotation angle of the specimen. An EC designed for an SEM was applied for DET to detect the motion of gold nanocrystals in a gaseous or liquid environment. Using DET with an EC, we applied DET to measure the Brownian motion of gold nanocrystals that were semi-fixed on a carbon thin film in vacuum, in Ar gas, and in water. In water, the motions of the gold nanocrystals were approximately 10 times larger than the motions in Ar gas or vacuum. In addition to the large motions induced by the water molecules, DET was also able to detect the small differences in motion amplitude between vacuum and Ar gas environments. In conclusion, DET is able to detect 3D motion differences in various environments with an angle detection limit of less than 10 mrad. DET is a promising method for the measurement of the motion of single molecules labeled with gold nanocrystals.

Acknowledgment This research was supported by the Japan Science and Technology Agency under the Core Research for Evolutional Science and Technology (CREST) program.

8

N. Ogawa et al. / Ultramicroscopy 140 (2014) 1–8

(X10-3) 2.0

(X10-3)

MSD of α , β , γ (rad2)

2

MSD of ω (rad )

0.8 1.5

1.0

0.5

0.6

0.4

0.2

0

0 0

5 10 Time steps (60 ms/step)

15

0

5 10 15 Time steps (60 ms/step)

Fig. 6. Motion of gold nanocrystals bound to sealing-film with mercapto-silane measured in various environments. (a) MSD of the rotation angle ω measured using DET. Over 300 individual gold nanocrystals, except for the “Not Moving” case, were measured in each environment and used for the MSD calculation. The “Not Moving” case was derived from the flat gold crystal, which was measured three times. Large motion was observed only in the water environment. DET can detect small differences between the gold-nanocrystal movement in Ar gas and in vacuum. In addition, the observed motion in vacuum was slightly larger than in the “Not Moving” case. The motion of the gold nanocrystals in water and in gaseous environments is caused by collisions between the nanocrystals and the water or gas molecules, while in vacuum, the motion is induced only by the pressure force of the electron beam. The MSD curve appears to be parabolic, indicating that the motion is of the type known as directed Brownian motion [13]. (b) MSDs of rotation angles α, β, and γ of gold nanocrystals in water. No definite anisotropic motion was observed in this specimen. Therefore, these results indicate that each gold nanocrystal has some directed motion, but the average motion of approximately 300 nanocrystals exhibits no directional specificity.

References [1] T. Komatsuzaki, M. Kawakami, S. Takahashi, H. Yang, J.R. Silbey (Eds.), Advances in Chemical Physics, Single Molecule Biophysics: Experiments and Theory, Wiley, New Jersey, 2012. [2] N. Kodera, D. Yamamoto, R. Ishikawa, T. Ando, Video imaging of walking myosin V by high-speed atomic force microscopy, Nature 468 (2010) 72–76. [3] Y.C. Sasaki, Y. Okumura, S. Adachi, H. Suda, Y. Taniguchi, N. Yagi, Picometerscale dynamical X-ray imaging of single DNA molecules, Phys. Rev. Lett. 87 (2001) 248102–248106. [4] H. Shimizu, M. Iwamoto, T. Konno, A. Nihei, Y.C. Sasaki, S. Oiki, Global twisting motion of single molecular KcsA potassium channel upon gating, Cell 132 (2008) 67–78. [5] H. Sekiguchi, A. Nakagawa, K. Moriya, K. Makabe, K. Ichiyanagi, S. Nozawa, T. Sato, S. Adachi, K. Kuwajima, M. Yohda, Y.C. Sasaki, ATP dependent rotational motion of group II chaperonin observed by X-ray single molecule tracking, PLOS One 8 (2013) e64176. [6] Y. Okumura, T. Miyazaki, Y. Taniguchi, Y.C. Sasaki, Fabrications of dispersive gold one-dimensional nanocrystals using vacuum evaporation, Thin Solid Films 471 (2005) 91–95. [7] A. Ishikawa, H. Miyata, Closed type environmental cell with carbon sealing film withstanding atmosphere pressure, Proceeding of China-Japan SALC, 2003, pp. 20–25. [8] K.C. Popata, R.W. Johnsonb, T.A. Desaia, Characterization of vapor deposited thin silane films on silicon substrates for biomedical microdevices, Surf. Coat. Technol. 154 (2002) 253–261. [9] A.J. Schwartz, M. Kumar, B.L. Adams, D. Field (Eds.), Electron Backscatter Diffraction in Materials Science, Springer, New York, 2000.

[10] A. Kusumi, Y. Sako, M. Yamamoto, Confined lateral diffusion of membrane receptors as studied by single particle tracking (nanovid microscopy). Effects of calcium-induced differentiation in cultured epithelial cells, Biophys. J. 65 (1993) 2021–2040. [11] N. de Jonge, F.M. Ross, Electron microscopy of specimens in liquid, Nat. Nanotechnol. 6 (2011) 695–704. [12] M. Krueger, S. Berg, D. Stone, E. Strelcov, D.A. Dikin, J. Kim, L.J. Cote, J. Huang, A. Kolmakov, Drop-casted self-assembling graphene oxide membranes for scanning electron microscopy on wet and dense gaseous samples, ACS Nano 12 (2011) 10047–10054. [13] M.J. Saxton, K. Jacobson, Single-particle tracking: applications to membrane dynamics, Annu. Rev. Biophys. Biomol. Struct. 26 (1997) 373–399. [14] Y.C. Sasaki, Y. Okumura, T. Miyazaki, T. Higurashi, N. Oishi, Observations of xray radiation pressure force on individual gold nanocrystals, Appl. Phys. Lett. 89 (2006) 053121–053124. [15] S.P. Jang, S.U.S. Choi, Role of Brownian motion in the enhanced thermal conductivity of nanofluids, Appl. Phys. Lett. 84 (2004) 4316. [16] H. Zheng, U.M. Mirsaidov, L. Wang, P. Matsudaira, Electron beam manipulation of nanoparticles, Nano Lett. 12 (2012) 5644–5648. [17] U.M. Mirsaidov, H. Zheng, D. Bhattacharya, Y. Casana, P. Matsudaira, Direct observation of stick-slip movements of water nanodroplets induced by an electron beam, Proc. Natl. Acad. Sci. 109 (2012) 7187–7190. [18] Jon Alkorta, Limits of simulation based high resolution EBSD, Ultramicroscopy 131 (2013) 33–38. [19] B. Britton, I. Holton, G. Meaden, D. Dingley, High angular resolution electron backscatter diffraction: measurement of strain in functional and structural materials, Microsc. Anal. 98 (2013) 8–13.