Particle Imaging Velocimetry of Self
Contents
1. timing diagram for camera on double exposure mode In the laser timing diagram the default Q switch delay in the laser is 200 us In the camera timing diagram TPD refers to the transfer pulse delay which is adjustable and TPW refers to the transfer pulse width which is by default 5 0 us The transfer pulse delay can be set to equal the camera exposure which can be adjusted from 1 us to 33 ms To properly capture images the first laser pulse must occur during the first camera exposure and the second laser pulse must occur during the second camera exposure In the experiments conducted the exposure time was maximized to 33 ms to create a bright image The transfer pulse delay was also set to equal the camera exposure time and the transfer pulse width was left at the default of 5 0 us As such the first exposure started roughly 33 005 ms after the trigger pulse and ended 66 005 ms after the trigger pulse and the second exposure started roughly 66 01 ms after the trigger pulse and ended 99 01 ms after the trigger pulse Since the delay between a trigger pulse to the laser is short slightly more than 200 us the trigger pulses for the lasers were set to occur roughly halfway into the camera exposures Three channels were necessary on the pulse generator to create the trigger pulses required for these experiments The first channel corresponded to the first laser pulse the second to the second laser pulse and th2. 11 8 07 0 1 g of the synthesized silica particles were mixed with 10 mL of ethanol This was added to 0 4 mL of a solution comprising 0 8 mL octadecyltriethoxysilane and 40 mL ethanol The mixture was stirred overnight at room temperature with a magnetic stirring bar in a 50 mL beaker During this time ethanol was released as the octadecyltriethoxysilane molecules reacted with the silica affixing many long carbon chains to the surfaces making them hydrophobic The reacted particles were centrifuged and washed with isopropanol for 10 minutes at 15 000 RPM three times 10 mL of 10 mg mL silica in isopropanol was prepared as a base mixture to be further diluted when experimenting with film preparation Note octyltriethoxysilane as well as other long chain organosilanes may also be used to functionalize the particles though the shorter chain of carbon atoms imparts a lesser degree of hydrophobicity to the particles which may significantly Advisor Dr Richard Riman Joseph Corry decrease the 2D _ packing factor The investigation of 2D packing factor vs functionalizing agent is another task yet to completed within the scope of this project Basic Film Making Procedure All subsequent experiments involved spreading the functionalized silica particles in isopropanol over a water air interface with controlled area at a specified flow rate This was achieved using a Langmuir Blodgett trough as shown in Figure 3 It is comprised of
3. Cylindrical Infusion Pum 45 532nm Lens p mirror Convex lens N 5 5 L i r F Laser head CC s Pump Stand 1L D Syringe Figure 5 side view of laser camera PIV setup In the above setup two laser pulses are fired in quick succession while particles are moving on the water surface enclosed by the Langmuir Blodgett trough The particles are fed at a controlled rate onto the water surface by an infusion pump that slowly compresses the plunger of a syringe filled with a particular volume percentage of particles suspended in isopropanol in a similar manner to the film making experiments described above A multi channel pulse generator is used to time these pulses which are also synchronized with two separate exposures of a camera situated directly above the water surface The laser pulses are reflected upwards and then to the right by two consecutive 45 mounted 532 nm Nd YAG Advisor Dr Richard Riman mirrors The laser beams are then spread horizontally by a cylindrical lens and in turn made parallel again by a convex lens of similar focal length The distance between the two lenses should be arranged such that the final widened parallel beam has roughly the same width as the short dimension of the trough The beam which is vertically aligned with the plane of the water surface then passes through the glass of the beaker and illuminates any particles Situated on the surfa
4. a 1 L beaker filled with deionized water and a top mounted apparatus that controls the area of the water surface A glass slide attached to a controlled vertically oriented continuous motor was inserted vertically into the water adjacent to the Teflon separator in the center of the apparatus The motor controller was a Newport Univeral Motion Controller Model ESP300 Functionalized silica particles in isopropanol were loaded into a B D 20 mL syringe which itself was loaded into a Baxter Model AS40A Infusion Pump capable of controlled flow rates between and 120 mL hr 8 22 2010 PIV Project Joseph Corry Figure 3 the Langmuir Blodgett trough used throughout the film making experiments sketch made using Google SketchUp Advisor Dr Richard Riman 8 22 2010 PIV Project The syringe needle was bent downwards so as to just touch the top of the trough allowing the particles in isopropanol to smoothly spread onto the water air interface and minimally disturb the existing particle arrangement To estimate a good initial withdrawal rate a mass balance calculation was carried out to equalize the area withdrawal rate with the coverage rate The final equation is FIR 480 W PF r rate in mm s fy volume fraction of particles in the liquid Ry feed rate in mL hr W slide width in mm PF 2D particle packing factor and r particle radius in cm For all of the experiments carried out in this report PF was a
5. htm 7 DaVis FlowMaster Manual Not available online Packaged with product http www lavision de en products davis php 8 Optical Microscope Limitations http en wikipedia org wiki Optical_microscope Limitations Advisor Dr Richard Riman 8 22 2010
6. pairs into a single vertically stacked image This program s executable is entitled comboimage and can be found on the computer near the laser remote controllers in the D 107 laboratory One of the output images from this program 1s shown in Figure 26 The procedure from obtaining velocity fields using the DaVis software package 1s listed separated from the above procedure because these steps are not necessary to perform PIV experiments However they are extremely useful to obtain a qualitative visual analysis of what 1s happening in any given image pair Advisor Dr Richard Riman Joseph Corry Figure 26 example image combination used by DaVis to calculate vector fields 8 22 2010 PIV Project Joseph Corry Procedure for Obtaining Vector Displacement Fields Open the DaVis software The initial screen should look something like this 1 File Buffer Rectangle Macro PIV PTY a mr AcgSetup Devices Window E 5 Rectangle D 2 Zoom QQ QA Resolution 4 om Load Save Export s 1 a nes BER Camera Take o 12 5 Optsize 64 y E Interactiv E 7 Buffers el PTU Laser Control yO 00 J In 0 de LO M ia M Auto P v y 10 Compute vectors i gt Parameter 13 14 15 ay Vectors 1 5 H 16 M Auto 18 19 Postprocess 20 Parameter 21 22 Y Postpr f 9 24 we Ba Ga Mask gt 4 E Rect ae Load one or more image File s From disk into the actual buffer and the F
7. Joseph Corry EP Settings PIV Parameter Image Preprocessing Image Correction OB Interrogation Window Correlation Peak Search Yector Scaling Mask buffer Vector Postprocessing Hok Adaptive multrpass C Single pass constant final window size C Multi pass with 2x constant final window size f Multi pass with decreasingly smaller sizes 2 passes 5124512 Iterations 1 Final interrogation window size 256 x 256 iterations 1 Initial interrogation window size Overlap final window 75 0 50 75 Grid spacing 64 pixel Figure 30 the DaVis interrogation window screen a The settings chosen here affect the actual vector calculation The algorithm can be chosen as single pass multi pass with constant window size or multi pass with decreasingly smaller sizes Usually the latter option is preferable since it will have the program calculate vectors across different size scales giving a better representation on the whole b The initial and final interrogation window size are the number of pixels that are analyzed by each pass Vectors are calculated within each interrogation window and summed at the end of the calculation These calculations are very processor intensive and take considerable time to compute as such large interrogation window sizes should be chosen when trying to get a vector field for a large image c The overlap of the final window determines how much different interrogation
8. PIV Project Particle Imaging Velocimetry of Self Assembled Monolayer Formation Abstract Small surface functionalized particles can be made to form densely packed two dimensional layers on a water air interface in a process known as fluid forming Using a laser based particle imaging velocimetry PIV system the formation of these monolayers can be observed in real time This report documents the synthesis and preliminary image analysis of particles suited for these investigations establishing a proof of concept for future experiments Introduction Silica particles are an ideal candidate for this study Using a modified St ber synthesis method highly monodisperse silica particles in the range of 300 nm to 1 um can be easily produced These particles may then be coated with organosilanes with long carbon chains such as octadecyltriethoxysilane to give them a hydrophobic surface character These surface functionalized particles will closely pack on a water air interface in order to minimize contact with the underlying water Silica Particle Synthesis The synthesis procedure used to produce monodisperse silica spheres was adopted from a 2007 paper by M Shishido and D Kitagawa which itself was based on the St ber synthesis method All chemicals used were supplied by Fischer scientific 26 5 mL of 29 wt ammonium hydroxide solution was combined with 17 2 mL of deionized water and 189 7 mL of reagent grade 95 wt e
9. age were measured by hand According to this analysis the mean particle size is 457 nm and the coefficient of variation is only 3 4 In addition the level of agglomeration seems acceptably low in the micrograph especially since this particular sample was not sonicated before the image was taken 8 22 2010 PIV Project Joseph Corry LS 230 Advanced Operator Mode File Window Run Control Preferences Security Help BB 5 A za 8 6 5 ao X WZ 12 i3 Open Qverlay Up Fonts Overlay Save Saveas Print Report Info Edit f Graph Listing Stats Color Cursor Legend Stober Sil_11_01 SIs Stober Sil_12_01 Is Stober Sil_10_01 Sls RunFile Edit Yiew Graph Analyze Display Cursor Differential Volume Stober Sil_11_01 ls Stober Sil_12_01 ls Stober Sil_10_01 ls Stober Sil_10_01 ls 0 040 um to 2000 um Volume 100 Mean 0 468 um Median 0 469 um SD 1 166 um dio 0 383 um doo 0 575 um Volume 4 6 10 20 40 100 200 400 1000 2000 Particle Diameter jm C Documents and Settings CoulterLS230 My Documents Joe11 6 10 1 um EHT 5 00 kV Signal A SE2 Date 8 Jan 2010 m WD 17 0 mm Photo No 1157 Mag 23 62 KX Figure 2 SEM micrograph of synthesized silica particles Advisor Dr Richard Riman 8 22 2010 PIV Project Surface Functionalization of Synthesized Silica Particles The functionalization method employed conditions found by Hyun Jun Kim and used on
10. an Joseph Corry Figure 36 PIV_S102_10um_1213 Shows fast uniform downward displacement Xx y vw w 32 32 o 0 906 32 16 8389 6 51544 32 o 0 96 96 17 3767 15 4833 160 96 16 8825 10 1818 224 96 15 2993 5 63021 Table 3 a DaVis vector data These are only the first few rows of the spreadsheet The x and y columns refer to the x and y pixel of the image while the vx and vy columns refer to the x and y 8 22 2010 PIV Project components of the velocity at that particular location measured in pixels per time difference between image exposure If the size scale of the image as well as the time between exposures 1s known these velocity components can be easily calculated in meters per second Discussion The most pressing problem facing these experiments is the inability of large 10 um particles to self assemble neatly under current functionalization conditions There are a few remedies that may be attempted to correct this issue Functionalization studies may be conducted on large particles to identify better functionalizing agents surface concentrations that yield densely packed particle films In addition smaller particles may be used with known working functionalization conditions but the particles have to be large enough to be seen with the camera As was roughly calculated above each pixel in the images investigated corr
11. ate in mL hr the described experiments were performed at 30 mL hr Adjust the height of the infusion pump so that the needle tip just touches the top of the trough on the beaker The liquid needs to flow smoothly onto the water surface 3 Software preparation a b Start the EPIX XCAP software on the computer If the camera is connected properly the following window should appear EPIX PIXCIO D2X Redlake MASD ES 1 0 8 bit 2c PIC DER Redlake WASD ES 1 0 5 bit chan Preset Capture Communication Buf Res Trig B uffe as RS 232 Port Mone Disabled RS 232 Mode Min Upn Dnload E nt Buff urre utter ae 339 Tire 0 500 sec s Figure 8 The XCAP Frame Buffers RS 232 Log Mone Buffer Screen Field Count Clear Butters Live Snap fe Unive XCAP works with an image buffer system Each time a pair of images is taken the current buffer and the buffer immediately following it will store the images Up to 8 pairs of images may be stored at a time Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry c Click the Live radio button to see a continuous image from the camera Use this mode to focus the camera ideally an object such as a ruler can be used to focus in the same plane that will be investigated as this will also give an indication of the level of magnification to create a scale Once focused click the Unlive
12. ce while the camera captures two images of the illuminated particles in quick succession and sends them to a computer running XCAP software or any other compatible PIV image capturing software via serial RS 232 connection 8 22 2010 PIV Project Joseph Corry Triggering and Timing both the lasers and the camera The following The multi channel pulse generator must diagrams illustrate the timing of these be preconfigured with appropriate timings to components n the configurations used properly conduct PIV experiments These throughout the described PIV experiments timings are dependent on the internal delays in Pulse Width gt 100 usec External Lamp Trigger IN 9 25 Us 2 ms pulse Lamp Synch OUT Adjustable Q Switch Delay Pulse Width gt 100 psec Q Switch Synch OUT 230 ns Same as External Q sw Trigger IN pulse Q Switch Synch OUT 130 ns Laser Pulse Figure 6 timing diagram for laser using external triggering Advisor Dr Richard Riman 8 22 2010 PIV Project Trigger gt e 20 us 100 ns 0 ns 100 Nanosecond Minimum Joseph Corry Strobe Out lt TPW Transler BL Internal TPO i First Image FRME ENA 512 Line Pairs gt lt MR Line ENA M i Second Image 512 Line Pairs lt 4 Lines 1 usec lt TPD lt 33 milliseconds User Programmable 2 usec lt TPW lt 5 usec User Programmable Figure 7
13. cle resolution The image shown at the left appears to be an advancing film front but this 1s questionable since the other image in the pair not included s very different Since the 10 um particles produced much crisper images the analyses were conducted on these image pairs Unfortunately a clear image of an object with features of known dimensions such as a ruler could not be obtained with the camera maintained in the experimental position Luckily the size scale in the images can still be deduced from the size in pixels of a single in focus particle in the images The average particle diameter judged from multiple images is about 4 pixels Therefore one pixel measures about 2 5 um of distance 8 22 2010 PIV Project Each image measures 1000 by 1000 pixels making the total image about 2 5 mm on each side By measuring the distance travelled by a single particle and dividing by the interval of time between the two image exposures or more Joseph Corry specifically the time between laser pulses as these are the only times the particles are illuminated the velocity of individual particles may be calculated An example velocity calculation iS shown below Figure 25 movement of a single particle in two consecutive images close ups of Figures 16 and 17 In Figure 25 the top left corner of each image portion can be taken as the pixel location 0 0 with the first number in the pair corresponding to the hori
14. clearly image closely spaced particles smaller than 1 um optically and this possibility should be investigated Finally it should be noted that the composition selections in these experiments were somewhat arbitrary Silica particles were used simply for ease of synthesis and availability Octadecyltriethoxysilane was used because t was previously demonstrated to provide good film density but an in depth study of film quality vs functionalizing agent has yet to be conducted Finally isopropanol was used because of 1ts high spreading tension but there may be other spreading agents that are able to create particle films more quickly and uniformly There are many different chemical combinations that may yield better results than the ones investigated here and though theoretical considerations can suggest combinations that are more likely to be effective in the end the best combinations which vary according to what effect is desired must be determined experimentally Conclusion This report shows a proof of concept of particle imaging velocimetry of 10 um functionalized silica particles Calculation of vector displacement fields using specialized software has also been demonstrated However many aspects of this realm of experimentation are left to be investigated In particular the effect of variations in the size of particles used and the chemistry of the system components have yet to be studied Regardless a working exp
15. ction In many cases the field may exhibit curvature as seen in Figure 33 or even vast changes in magnitude in the small area imaged as seen in Figure 35 8 22 2010 PIV Project Joseph Corry Figure 31 PIV_Si02_10um_0001 Control Figure 32 PIV_5102_10um_02093 Slight images taken with the system essentially at rest displacement towards the top right of image Figure 33 PIV_S102_10um_0405 Vector Figure 34 PIV_Si02_10um_0607 Shows field shows an interesting curved geometry fast uniform movement towards top left Advisor Dr Richard Riman 8 22 2010 PIV Project Figure 35 PIV_Si02_10um_0809 Shows displacement some areas but not others To obtain the data set describing the above vector fields from DaVis simply select File gt Export Buffer Spreadsheet into SUSE as shown in Figure 37 ee Da F Buffer Rectangle Macro PIVETY Compute Acquisition Acqs Load FO ictangle Save Crl F3 lage Y Statistics Profiles Movie Browse Files Import Export Buffe Export Filefs Export Window preadsnieet into Cipboard E Print Print Active Window Global Options save Settings Customer Settings Personal Labbook Exit Figure 37 how to export DaVis vector data into a spreadsheet Once this done the data may be directly pasted into an open Excel spreadsheet Table 3 shows the data set for PIV_S102_10um_0405 minus the header columns Advisor Dr Richard Rim
16. e third to the camera exposure The following table shows the timing for the three channels The pulse widths were chosen to be well over the minimum time Channel Pulse ee ms Pulse Width ms E first laser pulse O B5 S O second laser pulse exposure Table 2 proper channel timings for pulse generator Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry Particle Imaging Velocimetry Procedure 1 532 nm Nd YAG laser initialization SES f g Turn on both laser power supplies by turning keys Set the flashlamp setting on both power supplies to external triggering Set the Q switch setting on both power supplies to internal triggering Set both lasers to standby mode using the remote control boards Use the remote control boards to fire the lasers briefly and adjust both lasers power levels to bring the brightness to a suitable uniform level If the lasers make a loud click with each firing the power level is probably set too high Always operate within safe power levels and use appropriate eye protection Return both lasers to standby mode when finished adjusting the brightness 2 Particle preparation a b C d e Load the particle containing liquid into the syringe 10 mL is usually enough Attach a clean bent needle to the end of the syringe Load the syringe into the infusion pump Select the proper settings on the infusion pump and choose an appropriate feed r
17. erimental framework has been demonstrated that may be adapted to yield better results in the future Hopefully future experiments will determine a generalized fluid forming method 8 22 2010 PIV Project Joseph Corry that can be used with many different chemistries applications If these goals can be accomplished to create densely packed thin particle films these methods may be shown to be a viable effectively transfer those films onto a large alternative to conventional thin film processing range of useful surfaces and successfully techniques incorporate them into a variety of technical References LW St ber A Fink E Bohn Controlled growth of monodisperse silica spheres in the micron size range J Colloid Interf Sci 26 1968 62 69 M Shishido D Kitagawa Preparation of ordered mono particulate film from colloidal solutions on the surface of water and continuous transcription of film to substrate Colloids and Surfaces A Physicochem Eng Aspects 311 2007 32 41 Hyun Jun Kim s Notebook 11 8 07 Tempest and Gemini PIV Operator s Manual New Wave Reseach March 2004 https engineering purdue edu MICROFLUIDICS lab 90 1083_B_TESPEST 20 amp 20GEMINI pdf gt The MegaPlus Model ES 1 0 Series Cameras User s Manual Redlake MASD Inc August 2001 http alacron com clientuploads directory Cameras REDLAKE ES1 0_manual pdf XCAP Image Processing Software http www epixinc com products xcap
18. esired b Choose the cross correlation mode works for two vertically stacked images 4 Click the Image Preprocessing menu item in the Settings window This window should appear EF Settings E m x P Parameter Image Freprocessing w Do Image preprocessing Image Correction Interrogation Window Correlation Peak Search high pass filter Subtract sliding background scale length 5 Pixel Vector Scaling Mask buffer i Subtract offset counts 00 counts Vector Postprocessing Figure 29 DaVis image preprocessing window a Image preprocessing can be performed to increase the particle background contrast Particularly the preprocessing techniques include a high pass filter to subtract a sliding background and an offset subtract that will reduce the intensity of all pixels in the image b If the particle contrast is low it may be useful to subtract the background intensity in the image To do this hover the mouse over background areas of the image to find an approximation of the average background value and subtract this offset c Click the Execute button to perform the selected image preprocessing The preprocessed image will appear in a buffer that 1s selected determined via a text entry in a pop up d Note image preprocessing should not be necessary in most experiments 5 Click on the Interrogation Window menu item The following window should appear Advisor Dr Richard Riman 8 22 2010 PIV Project
19. esponds to roughly 2 5 um so this should be the smallest particle size that can still be seen directly by the camera A study of particle visibility vs particle size should be conducted to determine the smallest feasible particle size that can be used Finally other possibilities to make the particles more visible should also be investigated For example fluorescent dyes may be incorporated into the particles to make them release wavelengths of light that can be picked up by the camera If this is done 1t may be possible to see individual particles smaller than the resolution of the camera Indeed this may be already possible since even very small particles will exhibit Rayleigh scattering to some degree that may be observed However as seen from the image of 457 nm particles in Figure 24 it is most likely that individual small particles will be very difficult to see when the film is densely packed Also 1t should be noted that the current experiments are limited by the hardware available Better lenses or indeed better cameras may be used to clearly observe Advisor Dr Richard Riman Joseph Corry individual particles smaller than 2 5 um In addition the wavelength of laser light used may be changed to a smaller wavelength to better image smaller particles However regardless of hardware t will be basically impossible to decrease the optical resolution limit below about 200 nm In any case it should be possible to
20. ifference between a quality film and a subpar film is quite obvious A quality film will appear smooth and uniform and also exhibit some diffraction of light making the film appear red or blue at different angles Conversely a poor film will appear blotchy exhibit uneven diffraction and may contain striations These striations are the result of the withdrawal rate being too small causing the particles to build up too quickly on the water air interface and disrupting the arrangement at the addition point In any case the experiments above show that Advisor Dr Richard Riman the conditions that give the most reproducibly good films are 0 038 vol particles with a 10 mL hr feed rate Particle Imaging Velocimetry Setup It should be noted that while 468 nm particles are easily synthesized and self assembled they are quite small on the order of the wavelength of light and therefore very difficult to observe using conventional imaging techniques Optical methods are the only option for performing particle imaging velocimetry experiments without requiring an extremely expensive and specialized experimental setup 8 22 2010 PIV Project As such in order to resolve individual particles and produce more meaningful PIV data 10 um plain silica particles were purchased from Corpuscular Inc and functionalized in the same way as the synthesized particles monodisperse silica particles greater than 1 um are quite difficult to sy
21. lect the first pair of images b Continue to collect images as particles build up on the water surface If possible try to capture the advancing film front if one exists c Click File gt Save Current Buffer to save all the images in the current buffer d Use image processing software such as DaVis to analyze the images Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry Images and Results Two sets of images were taken one equipment and procedure described above was with 10 um SiO particles and one with 457 nm used to collect all of the images Noteworthy SiO particles both functionalized The sets of raw image pairs are shown below Figure 11 PIV_Si02_10um01 Figure 12 PIV_Si02_10um02 Figure 13 PIV_Si02_10um03 Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry Figure 14 PIV_Si02_10um04 Figure 15 PIV_Si02_10um05 Figure 16 PIV_5102_10um06 Figure 17 PIV_Si02_10um07 Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry Figure 18 PIV_Si02_10um08 Figure 19 PIV_5102_10um09 Figure 20 PIV_5102_10um12 Figure 21 PIV_SiO2_10um13 Advisor Dr Richard Riman 8 22 2010 PIV Project Figure 22 PIV_5102_10um14 Figure 24 PIV_S102_457nm04 Advisor Dr Richard Riman Joseph Corry Figure 23 PIV_5102_10um15 As can be seen from the image directly to the left the 457 nm particles produced much lower quality images in terms of individual parti
22. nthesize Due to their larger size these particles do not self assemble as reliably as their smaller counterparts but this was a necessary concession to establish a proof of concept for the PIV experiments as velocity fields and individual particle velocities cannot be calculated using sub micron particles The diagrams below show the physical setup created to perform PIV experiments which was located in D 107 in the engineering building on Busch campus a laboratory under the supervision of Professor Tse in the 1L Beaker 532 nm Cylindrical Convex mirror lens i Tempest amp Gemini PIV 532 nm Nd YAG Laser N lens on Joseph Corry Department of Mechanical and Aerospace Engineering The required components include e Tempest and Gemini 532 nm Nd YAG dual head laser Redlake MASD Inc Kodak MegaPlus Model ES 1 0 Series Camera 2 532 nm specialized mounted mirror 4 mm focal length cylindrical lens 4 mm focal length convex lens IL beaker w Langmuir trough Baxter Model AS40A Infusion Pump Adjustable stand for infusion pump B D 20 mL syringe Any precision multi channel at least 3 pulse generator e Any computer with EPIX XCAP software and a serial connection port Infusion Pump Pump Stand surface TOP VIEW Figure 4 Top view of laser camera PIV setup Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry Megaplus Model ES SIDE VIEW Camera
23. ollowing shortc 42 C Figure 27 DaVis opening screen amp interactive window The window on the left is called the interactive window and is where the majority of manipulation of the program takes place On the right is the buffer window in which input images are loaded by the user and generated images are placed by the program Load a combined image into the chosen buffer by clicking the Load button a Make sure the number in the text box next to the Image graphic in the interactive window matches the buffer number in which the combined image is loaded Click the Parameter button in the interactive window The window featured in Figure 28 should appear This is where many of the vector calculation settings can be modified Advisor Dr Richard Riman 8 22 2010 PIV Project Joseph Corry 2 Settings EIER Pl Parameter Image Preprocessing Image Correction Image preprocessing Subtract 1 00 Interrogation Window Correlation Peak Search Correlation mode cross correlation 1 cam i Data range complete image Vector Scaling Mask buffer Interrogation window 51 2 512 gt 256 2 256 shift 0 0 ow 75 eee POS OE Cor peak search confined Ww B 6 Yu 4 75 pizel Save correlation map into buffer Vector postprocessing Range Q hedian 1 4 F ill Figure 28 DaVis Settings window a Select the appropriate data range usually complete image if a vector field for the entire image is d
24. radio button to turn off the continuous mode e Click on the Trig tab to change the trigger settings The window looks like this o EPIX PIXCI D2X Redlake MASD ES 1 0 8 bit 2c E Pl Cl Dux Redlake MASD ES 1 0 5 bit chan Prezet Port Gain Exp m Stat Exposure Options put Res Trig Trigger Camera Trigger Mode Transter Pulse width i Trigger Input Transter Pulse Delay 5 Figure 9 TPO is Exposure w The XCAP Controlled Frame Rat Trigger Screen it Exposure m Single Shot Exposure Continuous 10 msec Live Snap ie Unlive Be f Set the camera trigger mode to Trig Double Exp g Set the trigger input to Snap Button If this button 1s clicked while the pulse generator is running the camera will take an image on the next incoming pulse h Leave the transfer pulse width at 5 0 us 1 Check the T P D is Exposure box j Choose an appropriate exposure time more time will make each image brighter 4 Pulse generator preparation a Adjust the settings on the pulse generator to match those listed in Table 2 b Press the Run button or equivalent button to start pulse generation 5 Particle spreading a Press the Start button on the infusion pump to begin feeding particles onto the water surface This may take some time as air may be trapped in the syringe needle 6 Image Collecting a When ready click Snap in the above window to col
25. ssumed to be 0 9069 and r was 229 nm or 0 00029 cm where R withdrawal Advisor Dr Richard Riman Joseph Corry Finding Ideal Film Making Conditions For the initial experiments the silica particles in isopropanol were diluted to 10 mg mL roughly equivalent to 0 04 by volume In the first few films made imperfections on the glass slide surface were limiting the overall film quality As such a batch of slides were soaked in 1M nitric acid for 2 hours and subsequently rinsed and stored in deionized water only to be taken out immediately before use At this point 1t was also clear that at some times the withdrawal rate had to be adjusted on the fly the calculated withdrawal rate could be used initially adjustments are necessary once the withdrawal has begun due to slight fluctuations with each experiment such as slide positioning in the trough As such the data collected notes both initial and final withdrawal rates The experiments shown in Table 1 below were performed on 1 15 10 8 22 2010 PIV Project mg particles per mL liquid Particles Joseph Corry Feed Rate Initial Final Comment mL hr Withdrawal Withdrawal Rate mm s Rate mm s 0 153 0 140 0 305 0 300 0 458 0 450 0 076 0 076 0 153 0 153 Table 1 qualitative assessment of film quality vs volume particles and feed withdrawal rate The comment section is a rough estimate of the film quality based on visual inspection The d
26. thanol and briefly stirred in a 500 mL beaker 16 6 mL of Advisor Dr Richard Riman Joseph Corry tetraethylorthosilicate TEOS was then added to beaker which was then covered with a layer of Parafilm The contents were stirred for 2 hours at room temperature with a magnetic stirring bar The liquid containing synthesized particles was then centrifuged for 10 minutes at 15 000 RPM After draining the supernatant and dispersing the caked particles in ethanol the contents were again centrifuged This process was repeated a total of three times yielding silica particles dispersed in ethanol Characterization of Synthesized Silica Particles Dynamic light scattering DLS with a Coulter LS 230 was used for rough characterization of the particles The particles were diluted in deionized water before scattering The DLS analysis showed a mean particle size of 468 nm but a high coefficient of variation of roughly 249 Figure 1 shows the results of the dynamic light scattering analysis Note the symmetrical lognormal distribution This suggests that the normal distribution is skewed to the right which in turn suggests that agglomeration may be significant However the particles were additionally characterized by scanning electron microscopy SEM to get more accurate coefficient of variation and level of agglomeration estimates Figure 2 shows an SEM image of the dried particles The diameters of 20 randomly chosen particles in the im
27. windows in the final pass overlap Without overlap a particle that moves from one interrogation window to the next will not be detected as moving 75 overlap is usually sufficient Choose the appropriate buffer for the input data 1 e preprocessed or not Select the appropriate buffer for the vector field before it is calculated 8 Click the Compute Vectors button to begin vector calculation The process may take up to 5 minutes to complete 9 Computed vector fields may be saved in various image formats using File gt Export Aa Vector Displacement Fields of Select Image Pairs Below are shown some of the vector particles are simply too large to self assemble fields calculated using DaVis from the particle images shown above Note that the flow direction 1s not uniform in all the image pairs this 1s because most of the images were taken in the time period before a film front began to materialize and as such the particle motion was essentially random over large time scales In fact in this set of experiments only 10 um particles could be studied a consistent film front failed to develop most likely because the Advisor Dr Richard Riman into neat hexagonal grids However clear patterns in particle velocities can be seen in some of the images indicating that over small time scales the behavior is more uniform Also note that even over very short time scales the vector displacement fields need not be uniform in dire
28. zontal difference from the top left corner and the second number corresponding to the vertical distance from the top left corner In the image portion on the left the particle encircled in red is located at 144 175 pixels In the image portion on the right the particle now encircled in blue is located at 134 144 pixels In microns these locations can be expressed as 360 437 5 um and 335 360 um The x and y velocity components of the Advisor Dr Richard Riman particle can be calculated by dividing the distance travelled by the time interval which is known to be 82 5 ms 49 5 ms 33ms Therefore _ dx _ 335um 360um wem 0 0 758mm s dt 33 10 s TE _ 9 348mmI s dt 33 10 s The total magnitude of the velocity 1s therefore lv alv y 4 0 758mm1 s 2 348mm1 sy 2 467 mm s 8 22 2010 PIV Project Image Analysis Using DaVis 6 0 In addition specialized image analysis software may be used to calculate a vector field for an image pair The software available for this research was DaVis FlowMaster 6 0 from LaVision a utility that allows for vector calculations of image pairs that can be displayed in a visual format DaVis has a cross correlation mode that calculates a vector field from an image file composed of two vertically stacked individual images In order to create an image file the students in Dr Tse s group have created a MATLAB program that automatically combined image
Download Pdf Manuals
Related Search