Optimum periodicity of repeated contractile actions applied in mass transport

Dynamically repeated periodic patterns are abundant in natural and artificial systems, such as tides, heart beats, stock prices, and the like. The characteristic repeatability and periodicity are expected to be optimized in effective system-specific functions. In this study, such optimum periodicity is experimentally evaluated in terms of effective mass transport using one-valve and multi-valve systems working in contractile fluid flows. A set of nanoscale gating functions is utilized, operating in nanocomposite networks through which permeates selectively pass under characteristic contractile actions. Optimized contractile periodicity exists for effective energy impartment to flow in a one-valve system. In the sequential contractile actions for a multi-valve system, synchronization with the fluid flow is critical for effective mass transport. This study provides fundamental understanding on the various repeated periodic patterns and dynamic repeatability occurring in nature and mechanical systems, which are useful for broad applications.

Dynamically repeated periodic patterns are abundant in natural and artificial systems, such as tides, heart beats, stock prices, and the like. The characteristic repeatability and periodicity are expected to be optimized in effective system-specific functions. In this study, such optimum periodicity is experimentally evaluated in terms of effective mass transport using one-valve and multi-valve systems working in contractile fluid flows. A set of nanoscale gating functions is utilized, operating in nanocomposite networks through which permeates selectively pass under characteristic contractile actions. Optimized contractile periodicity exists for effective energy impartment to flow in a one-valve system. In the sequential contractile actions for a multi-valve system, synchronization with the fluid flow is critical for effective mass transport. This study provides fundamental understanding on the various repeated periodic patterns and dynamic repeatability occurring in nature and mechanical systems, which are useful for broad applications. P eriodically repeated patterns are pervasive in natural and artificial systems where characteristic spatial or temporal information can be determined. Gravitational forces cause tides in the ocean and Earth with diurnal and semidiurnal periods 1 . Earthquakes in a specific region are anticipated by repeated oscillation patterns 2 . Among the biological periodic processes, the rhythmic process is a central function of life, such as limb motions in walking, heart beating, metabolism, growth, hormone regulation, and circadian clocks 3,4 . In this context, pulsatile pressure and dynamic flow in arteries have attracted increasing interest to explain diseaserelated hemodynamic phenomena 5 .
To investigate the physiological dynamic phenomena occurring in nature, in vitro model systems have been usefully employed. In many model systems, the water-based fluid flows through rigid pipes have been handled. However, these studies have technical limitation in their plausible adoptability for actual dynamics. Selective mass transport with controllable flexibility 6 thus has a great advantage in mimicking the real dynamics actually occurring in many biological systems. For example, human skin, a typical natural responsive membrane, acts as a permeable multifunctional membrane which interacts and responds to the surrounding environments such as light, heat, cold, humidity, chemicals, and mechanical stress 7 . Responsive materials have a great advantages in broad application areas such as controlled-release agents [8][9] , responsive coatings 10 and artificial organs 11 . In addition to the molecular separation by size and charge for which traditional membranes are employed, responsive membranes are designed to dynamically respond to the varying environments. Although this dynamic aspect is essential in the behaviors of responsive materials, it has not been thoroughly investigated because suitable model systems are lacking.
Unlike energy transfer mechanisms (e.g., tides and earthquakes) that employ Fourier's law, mass transport is mainly governed by conventional physical laws including acceleration and gravitational rules. In this study, dynamic mass transport through stimuli responsive soft condensed matters (nanocomposites) is investigated as a new model system to study the dynamically repeated patterns occurring in contractile fluid flows. By changing the dynamic factors, the output patterns of permeate molecules are systematically modulated. Based on the obtained results, we propose a certain optimized pattern for periodic control in dynamic mass transport. This study is the first systematic investigation in which periodic repeatability and effective mass transport are conceptually combined. The obtained results would be useful for understanding various periodic patterns and dynamic mechanics which are abundant in various natural and artificial systems.

Results
Temperature-responsive nanocomposite. Citrate-stabilized colloidal gold nanoparticles (AuNPs) are prepared in aqueous solution. The concentration is adjusted to be around 2.4 3 10 12 AuNPs/mL (Supporting Information) [12][13] . The transmission electron microscopy (TEM) mage in Figure 1a[I] confirms the average diameter of the prepared single AuNP is 20 nm diameter. Binary thiol end-capped functional polyethylene oxide (PEO) are incorporated for AuNP interconnection. The molecular weight between the junctions is adjusted to be M p 5 10,000 ((EO) n , n 5 227). The TEM image in Figure 1a[II] shows that the average size of the clusters in nanometerscale [14][15] . Considering the multi-reactive sites on a single AuNP, the number of incorporated PEO molecules is controlled to 310, 350, and 3100 times that of AuNPs (Supporting Information). A network is hardly formed if binary PEO monomers are only activated. However, highly interconnected composite clusters are formed as a result of multiple reactions of AuNP surface to thiol groups at the ends of the PEO. Multiple PEO linkages are attached to the surface of an AuNP to form networked assemblies. They also possess additional levels of complexity and anisotropy that can be exploited in self-assembly. The crosslink density (r) of the fully-linked network is inversely proportional to the molecular weight between the junction points (M p ).
Nano-scale images are obtained at the 7C X-ray nanoimaging (XNI) beamline at Pohang Accelerator Laboratory (PAL, Pohang, Korea) using high X-ray absorption coefficient of Au 13 (Supporting Information) (Figure 1a[III]). Beam size is adjusted to 100 mm 3 100 mm at 6.7 keV energy. Spatial resolution is approximately 100 nm. A texture of a composite cluster is prominent as black dots because of embedded AuNPs of high X-ray absorption. In Figure  1a[IV], the X-ray micro imaging results obtained at the 6D X-ray Micro Imaging (XMI) beamline at PAL provide a larger field-of-view than XNI (Supporting Information). Its spatial resolution is approximately 4 mm at the sample-to-detector distance of 30 cm and FOV is adjusted to 1200 mm 3 900 mm. Discrete clusters are detected by XMI and the size of the cluster is evaluated. Assuming that the cluster is in an ellipsoidal shape, the longer length R 1 is determined on a cluster and then the shorter R 2 is evaluated at the center point of the R 1 at a perpendicular direction. Arithmetically averaged cluster sizes from 10 clusters are summarized using standard deviation (Fig. S1). Increased PEO to AuNP values from 310, 350 to 3100 (PEG/ AuNP) can enhance the size of the clusters. Figure 1b presents the small-angle X-ray scattering (SAXS) results conducted at the 4C beamline at the PAL with the designed nanocomposite networks in aqueous solution. Depending on the molecular weight and architecture, aqueous PEO solutions exhibit unique physical property: the solubility decrease and/or phase separation occurs above the critical temperature (lower critical solution temperature, LCST) [16][17] . At a fixed PEG concentration (100 mmol/L), tem- perature-responsiveness of the designed nanocomposites is compared at three different temperature conditions (20, 40 and 60uC). A characteristic q value (q* 5 0.4 nm 21 ) indicates that the solubility of the nanocomposite network decreases according to the increase in temperature from 20 to 60uC, exhibiting the characteristic LCST. As temperature increases further, it moves to a lower q region, resulting in the formation of large structures. However, the q* point moves to a higher q region as the incorporated PEO amount increases ( Figure  S2) The organic2inorganic hybrid composites exhibit deviation from the typical elasticity of rubber, due to characteristic contribution of individual components. The designed nanocomposite networks are composed of dual domains, different from typical responsive materials of isotropic scale changes. The X-ray scattering intensity I(q) is experimentally determined as a function of the scattering vector q whose modulus is given by q 5 (4p/l)sin(h/2), where l is the X-ray wavelength and h is the scattering angle. Since our composites are microscopically isotropic the intensities will depend only on the modulus of q that for SAXS is given by q < (2p/l) h. From the obtained peaks at the SAXS profile, the size-definable structures marked by q 1 5 2.3 nm 21 , q 2 5 2.6 nm 21 , and q 3 5 3.0 nm 21 are generated at 60, 40 and 20uC, exhibiting larger size formation at increased temperature conditions.
Based on the obtained SAXS results, temperature-responsive nano-scale networks are suggested as illustrated in Figure 1c (left).
The correlation length (f), the distance between the junction points of the network, changes according to temperature. Compared with the lower temperature condition of highly swollen state [I], the f values increase due to the shrinkage at an increased temperature [II]. The selected permeates pass the nanocomposites more effectively through larger f [II] with increasing temperature than through smaller f at a lower temperature [I]. The designed nanocomposites are embedded in poly(vinyl alcohol) (PVA) gel matrix (3wt%, M w ,72,000 Merck), keeping the temperature-responsiveness of nanocomposites without leakage and shape deformation 18 . PVA matrix is ideal as a nanocomposite matrix, because no significant swelling or shrinking occurs under the designed experimental conditions. As the temperature increases several times above and below the transition point, the release of nanocomposites releasing from the PVA matrix is not detected. The designed nanocomposites in PVA matrix exhibits differentiated paths for the selected permeates according to the temperature condition 5 .
Controlled mass transport through responsive nanocomposites. Mass transport is effectively controlled using the designed apparatus ( Fig. 2a). Selected permeate molecules [19][20] pass through the nanocomposites loaded in a designed diffusion cell. The outlet of permeates is a function of permeate input frequency (f in ), inlet flow rate (r), and geometry of nanocomposite controlled by temperature (T). The concentration-controlled permeate solutions are loaded at a designed time interval to generate periodic input. Approximately 0.01 mL of 0.1 g/mL rhodamine 6G aqueous solution is loaded for one shot, and the number of shots in a minute is varied by changing the f in from 1 min 21 to 8 min 21 . The signal from the eluted permeates is recorded in a continuous mode (see Methods). The nanocomposite is embedded in the PVA matrix for molding into a cylinder (diameter: 0.1 cm; length: 10 cm). This cylinder-shaped nanocomposite pellet is loaded into a humiditycontrolled 0.1 cm-thick glass tube. The tube has heating coils, and it is connected to the inlet pressure and signal detecting devices. The durability and repeatability of the nanocomposite are verified up to 500 min of successive mechanical repetitions. Thus, all the experiments in the previous study are performed under this condition (Supporting Information).
The time for the permeate outlet is normalized to eliminate the condition-specific R t . Therefore, the time is expressed as Dt 5 t2t 0 in the graph, where the zero point (t 0 ) is the time at which injected permeates are first detected as an outlet. At the increased temperature condition of 60uC, the permeate is released in the detector (on position), and the permeate release is stopped at 20uC (off position). The temperature-controlled part in the tube is compartmented using individual heating coils to generate spatio-temporal modulation. Thus the structures can be regionally controlled by artificial gating functions. The input and output signals are highly synchronized as shown in Figure 2b. The left graph is the temperature signal, while the right graph exhibits detection signal of the outlet. The interval of the input signal is controlled to be longer than 15 s to provide sufficient time for the responsiveness of the nanocomposite. With carefully controlled experimental conditions, the results exhibit sharp synchronization responding to the applied temperature stimuli.
At the inlet flow rate (r) of 10 mL/min (10 22 cm 3 /min), proper repetition signals of rhodamine 6G are obtained for a periodic pattern beyond the characteristic retention time (R t ) during which permeates are retained inside the diffusion cell (Figure 2c). By carefully controlling the volumetric inlet flow rate (r) for stable base line, modulated signals of the injected permeate molecules are obtained at a designed time interval. A sharp and discrete outlet signal is verified for a detectable concentration of permeate at a given experimental condition. As long as the flow rate (r) is fixed, normalized frequency of outlet to inlet (f out /f in ) maintains unity (Figure 2c and 2d). However, the R t decreases linearly as r increases (Figure 2d). R t exhibits a typical linear relation according to the van Deemter equation as follows: L , D 1 1 D 2 /r 1 D 3 r. In this equation, L is the permeate path length, r is the flow rate, D 1 is the multiple diffusion parameter, D 2 is the longitudinal diffusion and D 3 is the mass transport parameters 43 . When the mass transfer effect is dominant, which corresponds to the system designed in this study, R t is inversely proportional to the r in a linear relation. When the flow rate is controlled from 0.01 to 0.1 cm 3 /min, R t changes from 16 min to 4 min (Figure 2d). When the total volume of the column as of 0.08 cm 3 , the efficiency factor (H) passing through the designed nanocomposite is close to 2 based on the following relation: Flow-controlled mass transport. Representative synchronized frequency patterns are displayed where the two inlet (f in ) and outlet (f out ) frequencies are highly synchronized ( Figure 2B and 2c). At different r values (0.01 to 0.1 cm 3 /min), the normalized frequency (f out /f in ) pattern is constant (f out /f in 5 1) as long as the acceleration rate is zero (a 5 dr in /dt 5 0). However, when r in is continuously changed at a fixed acceleration (a 5 dr in /dt), f out /f in varies systematically (Figure 3a). Under the given condition, the variation of f out /f in exhibits linear relation according to a (Figure  3b). At a given pass length condition (L), this linear relation with a directly indicates the linear proportionality to the kinetic energy. Therefore, the designed mass (rhodamine 6G) flow behaves as a conventional mass for which the general physical laws are applied: E(energy) , F(force) 3 L(length) , m(mass) 3 a(acceleration) 3 L. As r in continuously increases or decreases at a constant permeate injection frequency (f in 5 4 min 21 ), f out increases or decreases accordingly. The results for two representative cases are displayed in Figure 3a. As r changes with different a 5 60.005 cm 3 /min 2 and a 5 60.01 cm 3 /min 2 , the variations in f out are displayed at the middle and the right graphs in Figure 3a. Under no acceleration (a 5 0), the inlet frequency (f in ) and the first outlet frequency (f out ) has no difference (f out /f in 5 1). The normalized frequency (f out /f in ) exhibit linear proportionality for continuous change of a. This finding is attributed to to the mass-energy relation where the energy (E) is proportional to volumetric acceleration (a): E , a. Therefore, the volumetric acceleration/deceleration, rather than the absolute flow rate, has a significant influence on the permeate output. Even for densely packed viscous soft condensed matter media designed in this www.nature.com/scientificreports study, the energy (E) relation is satisfied as the permeates freely pass through The passing molecules are accelerated following the conventional mass-energy relation applied in typical particle-like physical behaviors.
Mass transport by contractile actions of one-valve system. With a fixed f in 5 4 (min 21 ) and r 5 0.01 cm 3 /min, temperature at the middle of the designed nanocomposite pellet is controlled ( Figure  4a). The off position in the valve function is provided by switching the temperature from 60uC to 20uC. The cooling process is only applied to the middle section on 2 cm length region (out of 10 cm total tube length) to induce temperature gradient. The frequency maintains a regular pattern (4 min 21 ) before changing to the offstate. However, during the responding time (t respond ), permeate movement is suddenly delayed by the contractile actions in the offstate. This phenomenon is caused by the swelling procedure of the nanocomposite, which increases the back pressure caused by sudden water absorption. After the abrupt stop of permeate movement, the decreased water flow induces an incremental decrease in f out , forming a tapering zone (t taper ). Once most of the permeates exit the off-zone, the permeate elution is completely stopped until the designed t off ends. When the on-position state and flow are both regained after t off , the permeates are detected in a high frequency (f recover ) because of accumulative congestion during the period of t off .
The frequency stabilizes during the recovering time (t recover ) to regain original frequency (f in 5 fi), where fi is the ith frequency after f recover .
With varying duration of t off , the resulting f out of the designed systems is observed accordingly (Figure 4b). t off is controlled from 15 s to 3 minutes in the middle of the stable flow at a fixed r. t respond [I] and t taper [II] are independent of t off , which reflects the characteristic property of the nanocomposite itself. Based on the increase in the t off , the first maximum recovering frequency (f recover ) right after the t off is normalized by the inlet frequency (f recover /f in ) [III]. This normalized value is saturated from a specific point (d(f recover /f in )/dt 5 0) at which the optimum off-time (t opt ) is determined. This finding implies a limit to increase the mass transport rate continuously by increasing the t off . Thus an optimized contractile action exist. To find the critical point from which the increase of the mass transport is limited (or saturated), variations in f recover /f in is investigated with varying parameters (Figure 5c). With decreasing the r, the time to reach the critical point takes longer. Before reaching the critical point, the f recover / f in relation follows an approximate liner relation with a definable slope a as a function of r: (f recover /f in ) 5 a(r) t off 1 w, where w is an intercept (Table S1). However, high f in only increases the magnitude of the f recover /f in whereas the critical point remains unchanged. Therefore, to determine the optimized t off , r in is exclusively decisive, and f in contributes to the magnitude of the f recover /f in . Therefore, the mass transport is controlled by the energy impartment conveyed by the flow acceleration. Since f recover /f in is linearly proportional to r, the following relation is reasonably satisfied: Compared with Eq. (1), which is applicable to stable flows linearly proportional to E without contractile action, the normalized maximum outlet frequency (f recover /f in ) is proportional to the half power of E when a contractile action is applied. Figure 4d shows the variations of the time duration elapsed to recover the original frequency (t recover ) plotted by t off . The time duration is classified into two regions, namely t off and t opt : In Eq. (3), t recover is linearly proportional to t off . In Eq. (4), the difference between the actually performed off-time and the optimum off-time (t off 2 t opt ) is additionally inlcuded. t optinum is the suggested time for which the mass transport rate is maximized, and an extended time period does not increase the mass transport rate. Therefore, t recover is necessary to deliver the overloaded masses. The obtained experimental results are fitted by the linear relation of Eq. (3). In additin, the proprotional factor K(r) is a fucntion of r satisfying a liner relation with r. An intercept value of 1/f in explains that the t recover is inversely proportional to the f in in all the cases in this study. The concept of the optimized periodicity thus summarized in Figure 4e. One period is composed of t optinum and t recover which are the functions of the t off , f recover and f in .
Mass transport by contractile actions in a series of multi-valve system. Mass transport is controlled by consecutive repeated contractile actions (Figure 5a). With a series of the 10 compartments, each section is sequentially heated to generate a partial blockage during the interval time (t interval ) from t 1 to t 10 under the conditions of t 1 5 t 2 5 t 3 5 …5 t 10 . When the contractile actions are performed in a series, a pushing effect on a specific point occurs, illustrated as arrows on the left-side. Heating is applied on each compartmental section of the temperaturecontrolled tube on the right in a series, so consecutively moving contractile actions can be generated. The stability of the system is verified by checking whether the normalized frequency (f out /f in ) of the system maintains unity without temperature modulation ( Figure  5b). When the total elapsed time t interval 5 St i (i 5 1, 2, 3, …10) is determined, then the temperature frequency is defined as f temp 5 1/ t interval . When a partial contractile action is applied in a series (Figure  5c), the mass transport can be either promoted or prohibited. With the increase in inlet flow rate (r), f out /f in value increases until a specific frequency (i.e., f temp 5 0.3) and then significantly drops down. In addition, f out /f in is proportional to the square of the r as curve-fitted in the graph. The experimental conditions and the fitting results are available in Supporting Information (Table S2). This result indicates that mass transport is accelerated linearly proportional to the energy as follows: Therefore, high synchronization of the contractile actions performed in a series is necessary to match with the fluid flow rate.  www.nature.com/scientificreports Discussion Mass transport has been investigated typically based on diffusion process for which Fick's law is applied 21 . Transport through smallscale pores, such as reverse osmosis membranes, has been described by the solution-diffusion model 22 and other modelling techniques 23 . In addition, small ions and large-scale colloidal particles in size ranges of 1 to 100 nm have been analyzed in a single theoretical framework 24 . Although the discrete nature of individual nanoscale molecules is evident, the main transport phenomena in nanofluidic systems are explained based on the continuum and mean-field approaches. Unlike these diffusion-dominant passive concepts, a new molecular transport is recapitulated in this study based on the acceleration by repeated contractile action in fluid flow. Depending on the contractile styles (one-valve and continuous multi-valve systems), the amount of energy imparted by the contractile actions is different. Under continuously increased flow rate (with a fixed acceleration) or sequential contractile multi-valve motion, the mass transport efficiency expressed by normalized frequency (f out /f in ) is linearly proportional to the energy: (f out /f in ) , E. Regarding the contractile action using a onevalve system, the relation (f out /f in ) , E 1/2 is satisfied.
Contractile energy is employed in many natural and artificial systems. Blood flow activated by heart beats and undulatory actions of intestines are representative contractile-promoted actions occurring in biological systems. This repeated contractile phenomena actually occur in the human body. The natural defensive mechanism of the human body such as response to the hypotension hemorrhage and blood vessel occlusion, might be explained based on the physical phenomena occurring in the optimal model. In practical artificial systems, intermittent flow controlled systems are generally relevant to this concept. We can successfully reenact the important dynamic behaviors occurring natural systems by effectively employing stimuli-responsive soft condensed materials 25 . This study would contribute to the basic and general understanding on the optimal mass transport in the dynamics-promoted flow systems.  controlled. Water flow rate (r) and mass transport frequency (f) are controlled through the designed nanocomposites. Rhodamine 6G is an ionic molecule sensitive to laser-based detection. It exhibits a dark orange color in water, so it is used as an effective tracer in water flow. With high water solubility (400 g/L), rhodamine 6G is directly dissolved in water and used as a saturated solution. Concentration-controlled permeate solutions are injected into the inlet line at the designed time interval (Micro-Liter OEM Syringe Pump Modules, Harvard Apparatus, MA, USA). A series of eluted permeates are measured and recorded using a spectrophotometer (2489 UV/visible detector, Waters, MA, USA). Empower 3 Chromatography Data Software is employed for data analysis. Spectra are obtained at a scan rate of 100 nm/s, and intensities are collected at an interval of 0.5 nm. For long-term and quick scanning, the wavelength range is narrowly controlled for rhodamine 6G spectra at an excitation wavelength of 485 nm. Scanning is performed every 5 s so that the possible maximum frequency designed in this study is evaluated as 12 min 21 . All the spectra are obtained at 20uC. To precisely control the local temperature of the designed nanocomposites, a multi-point temperature controller (MPC, Briskheat Corp., OH, USA) with resistance heating coils made of coiled chromel wire (MOR Electric Heating Assoc. Inc., MI, USA) is employed.

Methods
Small-angle X-ray scattering (SAXS). SAXS. Synchrotron SAXS measurements are obtained at the 4C beamline at PAL equipped with a position-sensitive twodimensional (2D) detector. Two energy settings are employed for wavelength modulation: 10 (0.0675 nm 21 ) and 18 keV (0.1217 nm 21 ). Samples of 1 mm-thick are used by stacking five 200 mm-thick Si wafer with SiN3 sample window (window thickness is 1 mm so that it hardly contributes to the sample thickness). The sampleto-detector distance (SDD) of 4 m covers the q range of 0.0679 nm 21 , q , 1.64094 nm 21 , where q 5 (4p/l)sin(h/2) is the magnitude of the scattering vector and h is the scattering angle. The q range is calibrated using polystyrene-blockpoly(ethylene-ran-butylene)-block-polystyrene (SEBS) (q 5 0.19165 nm 21 ). On the other hand, the 1 m SDD covers the q range of 0.346 nm 21 , q , 7.68039 nm 21 . The q range is calibrated using silver behenate (q 5 1.052 nm 21 ). During the measurements, the temperature is precisely controlled at an isothermal condition of 20uC if no heating plan is involved. To increase the sample temperature to 40uC and 60uC, programmed heating coil is employed to connect to the copper-based sample holder body. A W/B4C double multilayer monochromator is installed to deliver monochromatic X-rays with a wavelength of 6.75 nm (18360 keV) and spread of Dl/ l 5 0.01. The 2D scattered X-rays are recorded by a CCD camera (Mar CCD, Mar USA, Inc., CCD165). The collected SAXS data are corrected by subtracting the background and empty cell scattering.
Synchrotron X-ray nanoscopy (XN) 26 . Experiments are carried out at the 7C beamline of PAL. The X-ray source of 10 11 photons/mm 2 /sec consists of undulator with 20 mm period and 70 poles. The beam size is about 100 mm 3 100 mm at 7 keV. The X-ray source is radiated from a 3 GeV bending magnet and then monochromatized using a Ge(111) DCM. To achieve focused images, monochromatic X-ray beam of nominally selected 7 keV is focused on the sample using a condenser zone-plate (CZP, 1 mm dia. Beryllium refractive compound lenses) with innermost and outermost diameters of 4 nm and 100 nm, respectively. The primary X-ray image is magnified 50 times with an objective zone plate lens (140 mm innermost and 50 nm outermost diameter, W). It is then converted into a visible image on a thin scintillator crystal (Tb:LSO, 20 mm thickness). The visible image is further magnified 320, using an optical microscope. This provides a total 31000 magnification of image on a cooled CCD camera (Princeton Instrument VersArray 1300B cooled CCD) of 1340 3 1300 pixels, which corresponds to an equivalent FOV of 21 3 21 mm 2 .
Synchrotron X-ray microscopy (XM). Synchrotron X-ray images are captured at 6D beamline of PAL. The X-ray source is a bending magnet with critical energy of 8.7 keV at 3GeV electron energy operation. The white beam is attenuated by polished beryllium (Be) of 0.5 mm thickness and polished Si wafer of 1 mm thickness. The primary X-ray image is converted into a visible image on the thin scintillator crystal CdWO 4 of 100 mm thickness. X-ray images are captured using a CCD camera (Vieworks, VH-2MC). With a 103 objective lens attached in front of the camera generates the field-of-view of 1.2 mm 3 0.9 mm in physical dimension. The size of each pixel is about 0.74 mm.