The goals of this pre-lab are to:
White blood cells have a further unique property that you will come to appreciate. They are non-adherent. This means that (ideally!) they will not stick to the surface of the chip. Most of the cells in your body adhere to other cells or to molecules; this helps keep you from being a bowl of soup. Your blood cells, however, do not (usually) adhere to other cells. You will find that the cells will occasionally stick to the surfaces, due to interactions between the proteins on the cell's membrane and the chip surface. You may find this slightly annoying, but trust us, it would be much worse if we were using other cell types.
Log in to the computer with the user name "camscope"
and the password "camscope". If there is
not already a command prompt open, click on the "Terminal Program"
icon (which looks like a small black screen) near the lower left corner of the
screen to open one.
You will want to keep your files separate from those of other groups,
so first create a new directory based on your name using the command
"mkdir ~/Your_Name" or something similar. Type "cd ~/Your_Name" to
move into that folder (the ~ refers to your home directory, so these
commands refer to a directory inside your home directory that is
called "Your_Name").
Make sure that the camera is properly mounted on top of the microscope and is
plugged into the computer. Turn on the light by flipping the switch on the back
of the microscope. Start the camscope software by
typing "camscope" in the command prompt. If
it has started properly, you should see a screen that looks like this:
The main window shows the image from the camera. Across the top of the screen are buttons that control camera settings (the + and - buttons change manual settings, the circle sets the camera feature to automatic). Down along the right side of the screen are buttons for data collection and analysis, the functions of which should be fairly straightforward. The upper left part of the screen shows the current image and data set for recorded images, and the lower right part shows the image format. There should be an appendix at the end of this handout that labels the buttons and lists the corresponding keyboard shortcut.
The camscope software was designed specifically for this lab, so you are likely to find its features useful when taking measurements. This pre-lab will introduce some of these features; we refer you to the camscope users' guide for more details.
Our experiment chambers are 2-dimensional microscopic channels formed on the surface of a transparent silicone rubber (polydimethylsiloxane: PDMS). The silicone rubber is then bonded to glass slide. Fabrication methods can be found in the literature (Whitesides-1998), which is a very standard method for making microfluidic devices now. This method is popular because it is inexpensive and quick method of making devices, compared with silicon/glass based micromachining. However, PDMS microfluidic devices have several drawbacks, one of them is that it is a soft material. A small force can distort, deform and even completely collapse the microchannel made in the device.
A picture of the microfluidic device we shall be using in this lab is shown below. The device has five reservoirs divided into two sides: the "3-input" side and the "2-input" side. Fluid channels emerge from reservoirs. All channels on a given side join together at a mixing point to form a main (mixing) channel. Explore your own device and see if you can identify each of the labeled features.
The five reservoirs were intentionally distanced from the microchannels in order to enable free movement of the objective lense, but you should be careful not to hit the reservoirs with microscope objectives. Three solutions will merge at the merging point (beginning of the observation channel). From that point on, the distance is marked by tickers for easy calculation of the distance (time) from the beginning of the mixing point.
The distance markers are positioned every 2mm of the observation channel, and the numbers are the mean distance (in cm) from the beginning of the observation channel. The cell injection channel is 50 microns wide, the two buffer channels are all 100 microns wide, so the observation channel is 250 microns wide. The depth of the microchannel is 75 microns overall.
One of the most important steps in getting reliable data from this system is to
adjust the camera settings properly. These settings are controlled by the
buttons across the top of the screen. Although the default settings should be
reasonable if the condenser is positioned properly, you can optimize the
settings in the following way:
Many of the camera settings can
also be controlled automatically; for example, setting the color balance to
"Auto" tells the camera to guess what it thinks the RGB values
should be. When making quantitative measurements, this kind of automatic
control can often interfere with the measurements. However, you can let the
camera control most of these settings automatically if you wish, by clicking on
the round button next to a particular setting (or by using the corresponding
keyboard shortcuts).
Move the stage around to look at different points along the length of the
channel. How long is the channel? How much of it can you see? How would you
measure your position along the length of the channel?
While looking at the channel, use the fine focus knob (the smaller one on the
microscope base) to change focus. Focus in on the top of the channel (you
should be able to see detailed structure of the channel edges, and some dirt or
debris on both the top and bottom of the channel), and read the position of the
focus knob (the numbers are in micrometers). Now bring the bottom of the
channel into focus. How deep (approximately) is the channel?
Move the device until you can see the channel again. Click and drag the left mouse button to measure the width of the channel (in pixels). The observation channel, the widest channel, is 250 micrometers wide; what is the conversion from pixels to micrometers? Evaluate this conversion factor for both 4x and 10x zoom settings (objectives) on the microscope.
Orient the device with the side with the reservoirs closest to you. Add the recommended amount of microbead solution with to the two leftmost reservoirs (the "2-input" side; ask the TA how much to add). Mix the solution thoroughly and fill both reservoirs at least halfway. You'll probably need to move the chamber away from the microscope objective to do this. Use the xy translation knobs to move the microscope stage; you can use the stage position readouts (in mm) to return to the same position later. Be careful when manipulating the chambers!
Get the TA to help you apply suction to start flow in the channels. Once you get the flow started, it continues to be driven by gravity. That is, the difference in height between the input and output reservoirs creates a net force on the fluid; this force pushes the fluid down the channel. You can change this force (and thus the flow rate) by changing the height of fluid in the input reservoirs. However, the serpentine sections of the microfluidic channel provide increased flow resistance, which makes the flow velocity less sensitive to reservoir fluid height. Are there other ways to increase flow velocity?
Focus on the channel again. Can you see the beads moving along the channel? How fast do you think they are moving (take a guess)? Do you think the bead velocity accurately represents the fluid velocity? What factors might complicate the relation between the two?
While it is possible to measure flow rate from live images if the flow rate is slow enough, it is much easier to measure the rate from pre-recorded data. To capture a sequence of ten images, press the record button (the red circle, or hold down ctrl and press R). Later on, if you need to capture more than ten at once, your TA can show you how to increase this number. Now click on the "analyze mode" button to switch the software into Analyze Mode. You should see the first captured image. You can play the captured images in sequence by clicking the "Play" button, or you can manually switch between imgaes by pressing the "Previous" and "Next" buttons (or hold down Ctrl and press the arrow keys).
To measure bead velocity, move the pointer on top of a bead, and click and hold the left mouse button. While holding this button down, hold ctrl and press the right arrow to switch to the next image. Drag the mouse pointer to the new bead location and write down the velocity displayed on the bottom of the screen. Keep the button pressed and switch to the next image, then record the velocity again. Did it change? If so, which answer is more believable?
HINT: By default, the camera captures images in RGB mode, which operates at 15 images/sec. If the beads move too much between one image and the next, you can switch to either YUV411 or MONO modes, which have less (or no) color resolution but capture images at 30 images/sec. Click in the lower right where it says "RGB" (or press Ctrl-/) to switch modes.
Using a suction pipette, remove half of the fluid from each input reservoir, then measure bead velocity again.
Until now you have had two identical fluids in the two channels. In this section we will ask you to add a marker dye to one of those fluids. When you do make this change, what do you predict will happen at the Y junction where the two fluids meet? Try it and see if your prediction was correct.
Before the two fluids have completely mixed, there should be a color gradation from one side to the other. How steep is this gradation?
Move the stage so you are viewing a location further along the channel. Look at the gradation of color at this location and note how it differs from the gradation near the Y junction. What causes this difference? Obtain color profiles across the channel at different distances from the junction (mixing point) directly from camscope by saving them as text files. Plot them using a program such as MATLAB. Do you see a trend? Is the trend consistent with what you expect?
The color of the dye clearly depends on the concentration of dye. Do you think this dependence is linear? How would you test your hypothesis?
Fluid mixing can occur by two processes - diffusion (a microscopic mixing discussed in class) and convection (the bulk motion of fluid).
The equations for diffusion depend on both space and time. The microfluidic channel was designed to convert the time dependence into a dependence on position. Two fluids come into contact at the Y junction, and the time for which the two fluids have been in contact increases as the fluids flow down the channel. How would you determine the conversion factor between position along the channel and contact time of the two fluids?
The microfluidic channel also causes the fluid flow to be parabolic; that is, the velocity is highest in the center of the channel and falls to zero at the walls (and the floor and ceiling). However, the flow profile right at the Y junction is far more complicated, so we would like to ignore this region. If your flow rate is too slow, most of the diffusion will take place in this complicated region. If your flow rate is too fast, the beads will move too quickly for you to measure. How fast can you get the flow rate while still being able to measure bead velocity? At this speed, over what length of the channel can you see a concentration gradient across the channel?
Using a suction pipette, remove all of the fluid in the two leftmost reservoirs. Do not worry about the fluid left inside the microchannel.
*PBS = Phosphate Buffer Solution:
| Components | mg/liter | Mol. Wt. | Mol. (mM) |
| Calcium Chloride [CaCl2 2H2O] Dihydrate | 132.5 | 147 | 0.90 |
| Magnesium Sulfate [MgSO4] | 59.2 | 120.4 | 0.49 |
| Potassium Chloride [KCl] | 200.00 | 74.55 | 2.68 |
| Potassium Phosphate Monobasic [KH2PO4] | 200.00 | 136.09 | 1.47 |
| Sodium Chloride [NaCl] | 8000.00 | 58.44 | 136.89 |
| Sodium Phosphate Dibasic [Na2HPO4] | 1150.00 | 141.96 | 8.10 |
For this section, we will be using the three rightmost reservoirs (the "three-input" side) as inputs assuming the device is oriented with the reservoirs closest to you. Pipette out the contents of the mouse white blood cells vial into the central input reservoir. You'll probably need to move the chamber away from the microscope objective before doing this. Now fill the two side inlets with PBS buffer. Ensure that the levels of liquid in the three reservoirs are equal. What happens to fluid flow directions at the junction/mixing point if they are not? Try adding dye to one of the input reservoirs to find out.
Be careful when manipulating the chambers! Use the xy translation knobs to move the microscope stage; you can use the stage position readouts (in mm) to return to the same position later.
Get the TA to help you apply suction to start flow in the channels. Once you get the flow started, it continues to be driven by gravity. That is, the difference in height between the input and output reservoirs creates a net force on the fluid; this force pushes the fluid down the channel. You can change this force (and thus the flow rate) by changing the height of fluid in the input reservoirs. How else might you change the flow rate?
Focus on the channel again. Can you see the cells moving along the channel, coming out of the central channel of the three-way junction (it will take a few minutes for them to get there)? How fast do you think they are moving (take a guess)? Do you think the cell velocity accurately represents the fluid velocity? What factors might complicate the relation between the two?
You may notice that the cells appear to be rolling across the bottom of the chamber. Indeed, cells are more dense than water, and thus will tend to sink. Once on the bottom of the chamber, they will no longer migrate at the same velocity as the average flowrate would dictate. In fact, the flow profile along the height of the chamber is not uniform, but is parabolic, with a maximum velocity at the midline of the chamber. This is known as the Poiseuille flow profile, and is described in Appendix 4.2 of Weiss Vol 1. Thus, you cannot assume that the cell velocity is the same as the fluid average velocity.
One can increase the density of water by adding solutes. See for example the table on "Concentrative Properties of Aqueous Solutions" in the CRC Online Handbook of Chemistry and Physics.
While it is possible to measure flow rate from live images if the flow rate is slow enough, it is much easier to measure the rate from pre-recorded data. To capture a sequence of ten images, press the record button (the red circle, or hold down ctrl and press R). Later on, if you need to capture more than ten at once, your TA can show you how to increase this number. Now click on the "analyze mode" button to switch the software into Analyze Mode. You should see the first captured image. You can play the captured images in sequence by clicking the "Play" button, or you can manually switch between imgaes by pressing the "Previous" and "Next" buttons (or hold down Ctrl and press the arrow keys).
To measure cell velocity, move the pointer on top of a bead, and click and hold the left mouse button. While holding this button down, hold ctrl and press the right arrow to switch to the next image. Drag the mouse pointer to the new bead location and write down the velocity displayed on the bottom of the screen. Keep the button pressed and switch to the next image, then record the velocity again. Did it change? If so, which answer is more believable?
HINT: By default, the camera captures images in RGB mode, which operates at 15 images/sec. If the beads move too much between one image and the next, you can switch to either YUV411 or MONO modes, which have less (or no) color resolution but capture images at 30 images/sec. Click in the lower right where it says "RGB" (or press Ctrl-/) to switch modes.
Fluid mixing can occur by two processes - diffusion (a microscopic mixing discussed in class) and convection (the bulk motion of fluid). Fluid flow along the channels is driven by gravity; as a result, the fluid velocity depends on the height of fluid in the input reservoirs. Doubling (halving) the height of both reservoirs should double (halve) the fluid velocity in the entire channel. To test whether this also applies to cell velocity, remove half of the fluid from each input reservoir using a suction pipette, then measure cell velocity again.
Until now you have had fluids with the same osmolarity in the channels. In this section we will ask you to add a fluid with different osmolarity (either higher or lower) from the standard buffer solution to the two side inlets. Keep the reservoir to the central channel intact, (containing the same cell / media solution), while exchanging the PBS solution in both side inlet reservoirs with DI water. Initiate the flow, and let the chamber flow to stabilize (for a couple of minutes) and re-establish a flow profile similar to the previous exercise. Can you see any changes in the sizes of the cells, as they move along within the microchannel? What would be a good way to measure the changes in cell size from the (sequences of) images of cells, perhaps at different locations? Keep in mind that cell sizes vary naturally, so measuring average cell size change across an ensemble may be necessary.
The equations for diffusion depend on both space and time. The microfluidic channel was designed to convert the time dependence into a dependence on position. Three fluids come into contact at the junction, and the time for which the fluids have been in contact increases as the fluids flow down the channel. How would you determine the conversion factor between position along the channel and contact time of the fluids?
The microfluidic channel also causes the fluid flow to be parabolic; that is, the velocity is highest in the center of the channel and falls to zero at the walls (and the floor and ceiling). However, the flow profile right at the junction is far more complicated, so we would like to ignore this region. If your flow rate is too slow, most of the diffusion (of salts) will take place in this complicated region. A rate that is too fast may not allow the cells to reach equilibrium by the end of the channel. What is the optimum flow rate to visualize the osmotic swelling behavior?
Try following a single cell as it flows through the input channel, junction and main channel. Take frequent pictures and try to plot the radius of the cell as a function of time by postprocessing the images. Note that the cells, which spin around unpredictably in the channel need not be spherically symmetric.
You can copy your data to another machine via scp.
If you wish to save your data on a CD, quit camscope and run the program k3b. Click on "New Data CD", then browse the folders in the top left window until you find your directory. Drag this directory into the bottom half of the screen, put a blank CD in the drive, and click the "Burn" button in the lower right. This process should burn a CD (if your data files are too large to fit on a single CD, get a TA to help you split them into multiple folders).
As a courtesy to the next group using your lab station, please clean up your work space. The microfluidic device needs to be rinsed thoroughly to remove any dyes or cells that may be have been trapped inside. To rinse the chamber, drain the input and output reservoirs and refill them with distilled water, then repeat twice more. Finally, use the syringe to apply positive pressure to each input reservoir to flush any remaining dye or cells from the channel. When applying pressure, you should push the syringe no further than 1cm into the input (or output) chamber, or else you will damage the device. Move down along the length of the channel using the microscope and camscope software to ensure that the channel is clean. Don't forget to wipe up any spills, put things back where you originally found them, etc.
Using the colored tape, label your microfluidic device with your name. Finally, these devices are best stored wet. Fill the three reservoirs on the 3-side with DI water and store your device for the next use. Make sure to leave the reservoirs on the 2-side empty or there will not be any flow!
Good luck with your experiment!
Buttons on the right of the camscope window
provide the following functions (top to bottom, with keyboard shortcuts in
parentheses):
· EXIT (ctrl-q): terminate camscope and return to Linux prompt.
· PLAY (ctrl-p): display live images from video camera, or loop through previously recorded images if the program is in Analyze mode.
· PAUSE (ctrl-z): stop displaying new images and hold last one displayed.
· RECORD (ctrl-r): start recording images and brightness traces. Records ten images and then stops (your TA can change this behavior if needed, but it is your responsibility to make sure your measurements fit on a CD). Note: pressing the record button takes you out of Analyze mode.
· SNAPSHOT (ctrl-s): record a singe image and associated brightness traces.
· STOP (ctrl-b): stop recording.
· Analyze (ctrl-a): toggles Analyze mode, which displays previously recorded images.
· BACK (ctrl-left_arrow): display previous recorded image (only works in Analyze mode).
· NEXT (ctrl-right_arrow): display next recorded image (only works in Analyze mode).
· Screen Capture (ctrl-d): save image of screen (screendump).
In addition to these buttons, the upper left corner of the screen has + and - buttons (with keyboard shortcuts ctrl-. (period) and ctrl-, (comma) respectively). These buttons control which data set is currently being recorded or analyzed. All of camscope's other features (record, analyze, playback, etc) operate on only the current data set (which can be changed on the fly). Although you can do the entire lab without using multiple data sets, they are extremely useful for organizing your data. You should use a separate data set for each type of measurement, and for each measurement location. Be sure to make a note in your protocol book about which data set corresponds to which measurement!
By default, the camera uses automatic gain control to set the brightness and
exposure time to levels that give a good image. However, in some cases you may
wish to manually set these values in order to quantitatively compare brightnesses under different lighting conditions (for
example, comparing brightnesses with two different
dye concentrations in the chamber). Although there are no buttons to adjust
these values, you can use the following keyboard shortcuts:
· Ctrl-1: turn down the brightness by 3 points (brightness varies from 0 to 511).
· Ctrl-2: turn up the brightness by 3 points.
· Ctrl-3: turn on auto-brightness (Ctrl-1 and Ctrl-2 both turn it off).
· Ctrl-5: turn down the exposure by 3 points (exposure varies from 0 to 511).
· Ctrl-6: turn up the exposure by 3 points.
· Ctrl-7: turn on auto-exposure (Ctrl-1 and Ctrl-2 both turn it off).
· Ctrl-9: turn down the color saturation by 3 points (saturation varies from 0 to 255, and has no auto mode).
· Ctrl-0: turn up the saturation by 3 points. When any of these keys is pressed, the current values of these parameters are printed to the console.
Camscope allows you to make quantitative measurements both in real-time and from previously recorded images, by switching to/from Analyze mode. Several types of measurements can be made:
Images are recorded sequentially as .BMP files. The first image recorded into an empty directory is named image.000.000000.bmp, the second is named image.000.000001.bmp, and so on. Changing the data set changes the three-digit number near the beginning of the file, and resets the six-digit counter to zero. If the cursor is in the image at the time of recording, a text file is also created that contains the red, green, and blue intensities that were displayed below and to the right of the image. The text file associated with image.000.000000.bmp is named info.000.000000.txt. Screen dumps are saved in a similar sequential fashion, with the first one named print.000.000000.bmp.
When you are ready to leave the lab, you should save your data. We cannot guarantee that your data will be preserved on the lab computer, so copy everything before you leave. The easiest way to save your data is to record it on a CD. Put a blank CD into the CD burner and type
k3b
at the linux prompt. This command will bring up a graphical interface that lets you save your data to CD. If your directory is larger than the size of one CD (about 700 MB), you may find it helpful to split your data into multiple subdirectories and copy each onto a separate CD. Your TA can help you do this.
The camscope program is intended to provide sufficient information for you to monitor how your experiment is proceeding while you are in the lab. This is important so that you can discover problems and fix the problems before taking your final data. It is also important so that you have a good idea that your experiment was successful or not before you leave the lab. If you discover problems with your data, it is generally easier to take new data now than it is to come back at a later time and start again.
Results from camscope can also be directly incorporated into your final report. However, quantitative statistics to test specific hypotheses generally require additional analyses, based on images and brightness profiles saved by the camscope program. Every time camscope takes a picture, it saves the image as a .BMP file. When the cursor marks a "line-of-interest" and a "column-of-interest" at the time that an image is recorded, the red, green, and blue intensitives of those lines and columns of interest are also recorded as "info" files (see "File Formats" above).
The numbers in the info files can also be read by text processing / spreadsheet / plotting programs and used to generate plots. Some trends are easy to see just from the traces. However, the traces are noisy, and averaging can give more reliable statistics. Brightnesses can be averaged in the horizontal or vertical direction from the line-of-interest and column-of-interest numbers contained in a single info file. Brightnesses can be averaged across time by combining information contained in many info files. Arbitrary analyses are possible by calculating directly from the images. Many software packages, including Matlab, provide functions that access brightness values in .bmp image files.