Confocal Wiki

From Course Wiki
Jump to: navigation, search

Measuring Morphology of Biological Specimen using a Low-Cost Confocal Microscope: An Undergraduate Teaching Experiment

Stephanie Bachar, Sivakami Sambasivam, Steven Wassermann and Steven Nagle.

1 Motivation

1.1 Background and Motivation

The main advantage of a confocal microscope is its ability to image a very narrow depth of field, advantageously excluding light from above or below the focal plane of the image. This is done by using a pinhole, strategically placed where the light from the focal plane comes to a focal point while light from below and above the sample is converging or diverging. As a result, confocal microscopes have increased contrast compared to wide-field microscopes. Additionally, they can be used to generate 3D images by taking a series of 2-dimensional images along the z-axis of a specimen. Also, confocal microscopes have a marginally higher resolution than wide-field microscopes.

Currently, confocal microscopes cost on the order of $100,000 - quite an expensive purchase for a teaching laboratory or a low budget laboratory. Recently, there has been a new advent of creating low-cost imaging systems to allow for higher quality imaging in resource constrained environments. Xi et. al (2006) created a $8561 addition to a $20,000 inverted microscope allowing confocal imaging at a much lower cost than confocal microscopes available on the market. Further, researchers at the Massachusetts Institute of Technology have created an atomic force microscope that costs less than 10% of research-grade atomic force microscopes and other MIT researchers have created a cost-efficient optical trap for use by undergraduates.

The objective of this project is to create a teaching confocal microscope with associated documentation and suggested experiments for approximately $20,000. Based on our literature review, such a system does not currently exist.

1.2 Previous MIT Prototypes

In Fall 2009 of 20.309, Laskey, Ramanthan, Rurak and Wilder began the construction of a confocal microscope in the 20.309 laboratory. Stephanie Bachar continued work on the microscope during January of 2011. When beginning our project, the confocal microscope was functional but had a number of imaging and processing limitations. There were also a number of additional components which could be added and characterizations which could be performed to augment the device.

There were some stability issues in the optical setup, as well as unnecessarily complex construction. The signal generated by a fluorescent reference slide was about 2V, which is somewhat low. The images generated by the confocal microscope had blurry edges and inconsistent intensity due to fluctuations in the laser output. While there had been successfully imaged patterns on a fluorescent reference slide, the setup was unable to pick up signal from fluorescent beads. This was likely due to the resolution capabilities of the microscope as well as the low density and dimmed fluorescence of the beads used. Matlab code existed to generate a laser scan over the sample and to recompose the associated output from the PMT into an image. However, the scanning and re-composition code was quite inefficient, resulting in very slow image generation. Furthermore, the scan code generated a single 2D image of the sample instead of 3D imaging.

2 Design

2.1 Optical Setup

Beam path confocal.JPG

A confocal microscope generates images of sample fluorescence by scanning a laser across the field of view. The fluorescent signal at each point of the scan is measured by a photomultiplier tube and the intensities of the signal are used to recompose a fluorescent image. Each component of our optical design is described below.

(A) Our setup uses a 532nm laser with beam diameter of 1.1mm. (C) The laser is shone onto a dichroic which reflects the 532nm light of the laser into 2D galvanometer scanning mirrors. The galvos scan the laser back and forth across the desired field of view and at the desired resolution. (D) The laser beam diameter is expanded 7.14-fold by the 35mm, 250mm lens beam expander. The larger the beam diameter going into the objective, the smaller the beam size on the sample plane. The 35mm lens is placed ~35mm away from the center of galvos to ensure that the angle of deflection of the laser as it is coming out of the galvos is the same angle as the beam entering the objective lens. The 35mm lens is placed (35mm+250mm=) 285mm away from the 250mm objective lens to ensure that the light exiting the beam expander is collimated. (E) The beam is focused on the sample plane using a 40x objective. Fluorescence signal traces back through the beam expander (which shrinks the beam 7.1-fold) and the galvos and then passes through the dichroic. (G) The fluorescence signal is passed through a barrier filter to exclude any reflected laser light and through a 6.7-fold beam expander. Finally, the beam is focused such that the signal from the plane-of-focus of interest comes to a focus at a small pinhole (20-50um), thereby excluding out-of-focus light. The photons from the beam reach a photomultiplier tube (PMT) which translates the intensity of the fluorescence into current.

(B) We additionally measure the laser output to correct for variation in laser intensity. A fraction of the laser beam is diverted using a beamsplitter (a microscope slide cover slip) into a photodiode which reads the laser intensity.

(F) Finally, the microscope images bright-field as well. A near-IR LED illuminates the sample. The bright-field image is reflected off of an IR dichroic and focused onto a CCD camera by a 200mm tube lens.

Design Considerations The 35mm lens is a positive lens not a negative lens because the beam expander also serves as a relay lens, thereby keeping more of the laser scan within the confines of the 1” optics.

We chose a 7-fold expansion for the laser beam because this approximately fills the maximum diameter of the 40x objective, thus minimizing the size of the scanning beam spot.

Suggestions for changes It may be beneficial to have two beam expanders, one before the galvos and one after the galvos. This will shorten the beam path needed for the laser beam expansion leading to the objective as well as the beam expansion needed for the fluorescent signal leading to the PMT.

It should also be possible to use a 100x objective lens with minimal changes. This will give better resolution images. However, 100x objective lenses are more difficult to use since the ones we have in lab are oil immersion lenses. Thus, they should only be used when most of the rest of the microscope has been worked out.

2.2 Circuit Setup

Confocal Circuit Diagram.JPG

(B) PMT Circuit: In order to create a confocal image, the temporal array of PMT signal intensities must be reconfigured into a 2D image based on the scan of the sample. The current generated by the PMT is low and must be amplified. The gain of the PMT circuit is: $ \frac{R3*R1}{R2} = \frac{(10kiloOhms * 3 kiloOhms)} {100 Ohms} = 3*10^5 $

The resistors were chosen to amplify the PMT signal so the output to the DAQ is between 1-10V, and consistently less than 10V (since that is the maximum voltage the DAQ can read).

The capacitor was chosen to act as a low pass filter to minimize shot noise.

The oscilloscope measures the voltage coming out of the Op Amp. Since the voltage from the PMT is negative, the voltage measured from the output of the Op Amp will be positive. The oscilloscope measurement is taken from the Op Amp output in order to prevent interfering with the amplification circuit.

(A) Photodiode Circuit: Similarly, the signal from the photodiode is amplified to be in the range of 1-10 V. The gain is:

$ \frac{R3*R1}{R2} = \frac{(2.2kiloOhms * 2.2 kiloOhms)} {100 Ohms} = 4.8*10^4 $


(E) Stepper Motor Driver In order to take 3D scans of samples, the Z-axis of the stage is connected to a stepper motor which allows the sample to be moved incrementally along the Z-axis. The stepper motor is controlled by the DAQ through the stepper motor driver which transmits voltage signals from the MS1, MS2, MS2, DIR and STEP pins to motor coils 1 and 2 through pins 2B, 2A, 1B and 1A. The input the the driver from the DAQ is described in the table below.

MS1 MS1, MS2, and MS3 control the step size:

full step = low, low, low

MS2 ½ step = high, low, low

¼ step = low, high, low

MS3 ⅛ step = high, high, low

1/16 step = high, high, high

DIR DIR controls the direction of the stage up or down
STEP When STEP switches from low to high the motor takes a step.

These five motor inputs are controlled by 5 digital output channels of the DAQ. The four motor outputs are connected to four pins (4, 5, 6, 7) on the cable to the stepper motor. Additionally, pins 1 and 3 may be used to set a voltage limit to the stepper motor. However, this part of the circuit has not yet been connected.

Pin Description
1 Limit Switch Ground
3 CW Limit Switch
4 Motor Phase 2B
5 Motor Phase 2A
6 Motor Phase 1B
7 Motor Phase 1A

Figure 3 Display of pins on the Stepper Motor Cable

The stepper motor driver is powered by a 12V power supply. The current run through the driver and to the stepper motor is controlled by turning Δ. The current is calculated by measuring VREF and converting to current:

$ ITripMAX = \frac{(VREF)} {8 * R_s} $

where RS = 0.05 Ω. The maximum current tolerated by the stepper motor is 1A, so we ran the device at ITripMAX ~0.5A or VREF~0.2V.

GalvosThe galvanometers scan the laser across the sample as shown in the graph below. Figure4.jpgFigure 4 Voltage (Theta) of laser scanning across sample


AO0 controls the x mirror and AO1 controls the y mirror. A one-volt change in the galvo input corresponds to a 1 degree change of the angle of the galvanometers.

It is important that the PMT data collection and the galvanometer scan begin at the same time if an accurate image is to be created. The PMT data collection is triggered by a trigger pulse at the beginning of the scan (as can be seen in Figure 4 Voltage (Theta) of laser scanning across sample). The galvanometer output is fed into Analog Input 3 (AI3) so that the trigger pulse can be detected. When the trigger pulse occurs, the DAQ begins collecting the PMT data.

Though not noted in Figure 2 A diagram of our PMT Circuit., capacitors were placed between GND – (+)15V and GND – (-)15V to decrease noise.

2.3 Code http://web.mit.edu/sbachar/Public/Images/ThreeDconfocalScanwithpd3.m

This code scans the desired field of view.

1. Set 3D scan parameters. First, the number of scan slices wanted is set. The distance between scans can also be specified as 2.5, 1.25, 0.625, 0.31 or 0.16um. Under the assumption that the sample is currently in focus, the stage will move up half the total height of the scan. Throughout the rest of the code, the stage will step down through the center height and then down the same height below the sample focus.

2. Creating and modifying the scanning function. The scan function dimensions are determined the the field of view, samples per line and number of lines which the user specifies in the code. A wave form is generated which contains a series of horizontal lines connected by sine curves. This x-y distance wave form is converted to the angle (in degrees) of the galvanometers needed to make the laser reach that location on the sample field using the formula:

$ thetaX = atan(\frac{\frac{x}{objectiveFocalLength}} {pi} * 180) $

Since 1V=1degree, this is tantamount to converting the wave form into a voltage feed for the galvos. The wave form is multiplied by the maximum theta (determined by the FOV) to scale the scan accordingly.

Then, the wave form is translated such that the center of the square field of view to be scanned is the same as the laser when it is going down the center of the beam path.

Finally, a trigger pulse is added to the beginning of the vertical drive wave form. Half of the first data point is subtracted from the value of the first data point and this value is added as the first in the wave form. The trigger value is set to half the distance between the trigger pulse and the first value. When the voltage outuput the galvanometers rises through this point, the PD and PMT will begin collecting data. This ensures simultaneous data collection.

3. Collect a 3D stack of data. The stage moves down a step and repeats step 2.

4. Process Data and create image stack. After all the data has been collected, the array of intensities for each z slice is recomposed into a 2D image. The PMT data is corrected for laser variation using the laser intensity data gathered by the photodiode. Finally, the data is normalized by the maximum intensity value found in all slices.

http://web.mit.edu/sbachar/Public/Images/z_stage_Con.m This code was adapted from Dragos Guggiana.

Based on inputs into the z_stage function, this code moves the stepper motor either up or down the specified distance using the specified step size

3. List of parts

http://web.mit.edu/sivakami/Public/20.345/20.345-%20Confocal%20Budget/

4. Methodology

The following sections detail the methodology used to construct the confocal microscope, run experiments and gather data. These sections may also be regarded as instructions for future use of the microscope.

4.1 Aligning the beam path (No lenses)

  1. Make sure that all the lenses and mirrors and beam path is level. In the future, this can be more easily achieved by ordering/making pucks of the exact height that is the offset from the galvonometer mirrors. Then all the components can be mounted at the lowest level possible and the pucks will offset the steel bars for the components after the galvonometer mirrors.
  2. After ensuring that the construction is straight, ensure that the laser beam is straight. Remove the dichroic and adjust the first two mirrors near the laser so that the laser beam is aligned straightly through the center of the back of the box of the dichroic straight out the back of the box of the dichroic by using two points (t-shirts) to check for straightness, because one t-shirt will only provide you with information on whether the laser is center at that location, but not whether it is straight.
  3. Replace the dichroic and adjust it so that the laser beam proceeds into the center of the galvonometer mirror box. In the future, lengthen the beam path between the dichroic and the galvonometer for easier laser alignment.
  4. Simultaneously adjust the combination of dichroic mirros and galvonometer mirrors so that the beam exits the galvanometer box in a straight path down the center of the beam path towards the 35mm lens.
  5. Use minimal adjustments of the galvonometer mirrors and the following mirror so that the beam shines straight into the center of the objective.
  6. Adjust the final mirror that is closest to the objective to straighten the beam to roughly in the center of the view if you were not able to make it straight in step 5. In the future, replace this mirror with a static 45 degree mirror which would provide additional stability as the extra adjustment at this point is most likely unnecessary.
  7. Once the above 6 steps have been accomplished, very little adjustment should be necessary for additional use of the microscope. In the future, it should try to be ensured that the optical axis is orthogonal to the stage. Our current setup has the beam centered in the FOV, but the optical axis is not orthogonal to the stage as was displayed with the movement of our image view in our 3D image stacks of the 4um beads.

4.2 Positioning the relay lenses/beam expander lenses

  1. After the laser beam is roughly straight, place the 250mm lens in its theoretical location and place the objective lens. Adjust the 250mm lens until the laser beam is collimated. Use lens paper to ensure a collimated beam as the ceiling is too far away to ensure collimation.
  2. Remove the objective and add the 35 mm lens Be sure to use a different 35 mm lens than the lens that was previously used because it was scratched. Be careful with placing and removing the 35mm lens because it can be very easily scratched because the lens tube is not long enough to cover the lens. in its theoretical location. Adjust the 35 mm lens until the light is collimated once again.
  3. Adjust the length of the beam path to ensure that the 35mm lens the right distance from the galvos.
  4. Readjust the beam path using the galvanometer mirros and the mirror after the galvanometer to ensure that the laser is moving straight through the center of the 35mm lens as any aberrations off of center will cause a drastic change in the angle of the laser beam as this is a 35mm lens.
  5. The values that we used to align our galvanometer mirrors to the center were -.375 for zeroX and 1.25 for zeroY.

4.3 Reading laser intensity

  1. Use a cover slip as your beam splitter to split the beam.
  2. Mount the cover slip on the microscope part that is used to place dichrocie mirrors in the beam path.
  3. Mount a photodiode in the exact center of the side of the box that holds the cover slip and its mounting component.
  4. Connect the photodiode using a cable and the circuit shown below to the DAQ so that the photodiode signal can be read by the computer.

4.4 Aligning the PMT

  1. Before using the PMT, place in the Barrier Filter and the -30mm lens in arbitrary positions.
  2. Put a 200mm lens approximately 170mm from the -30 mm lens and
  3. Adjust the dichroic mirror to get the laser beam going straight down the beam path.
  4. Move the 200mm lens to make the beam collimated.
  5. Replace the 200mm lens with a 100mm lens.
  6. Put the PMT ~200mm away from the 100mm lens.
  7. Move the dichroic back into place and use a fluorescent reference slide to get a signal. With the lights off, you should be able to see the signal brightly on white paper. Center the signal at the center of the pinhole.
  8. Adjust the mirror knobs and the location of the PMT (distance from the mirror) to get the maximum signal.
  • Note: Never turn on PMT when pinhole isn’t on -- too much light can ruin it.
  • Note: PMT is powered by -760V from the high voltage power supply.

4.5 Making sample slides of fluorescent beads

  1. Take your stock solution of 4 micron beads and dilute it 1:1000 by using 1 uL of the solution and 1000uL of distilled water.
  2. Mount the diluted solution on the slide by using 5uL of glycerin and 1uL of the solution.

Note: The glycerin solution that can be used can be pure glycerin.

  1. Place a cover slip on the slide and use valap to seal the cover slip.

If you are attempting to create PSF slides:

  1. Dilute the stock solution of PSF beads to 1:500 by using 1 uL of the PSF solution and 500uL of glycerol.

Note: The glycerin solution that should be used should be glycerin mixed with 1 micron fluorescent beads so that the user is able to focus the slide on the 1 micron beads which should place them within very close focus of the PSF beads.

  1. Mount the diluted solution on the slide by using 5uL of the diluted (1:500) solution.
  2. Place a cover slip on the slide and use valap to seal the cover slip.

Note: The beads can also be diluted in agar or mounted with agar; however, slightly agar autofluoresces which is detrimental to our imaging of the microscope. In addition, the user may feel free to dilute the solution in distilled water or glycerine. The benefit of glycerin is that it prevents the beads from moving which is very beneficial for the PSF beads. However, glycerin is also extremely difficult to deal with due to its viscosity.

4.6 Creating 3D image stacks

  1. Write the desired name of your image file (.tif) in the imwrite section at the bottom of the code.
  2. Find the optical focus of the item you want to image.
  3. Turn on the laser and adjust the x-y stage location to shine the laser directly on the item.

Note: It is helpful to let the laser equilibrate before scanning by leaving it on for a few minutes prior. However, fluorescent samples photobleach, so cover the laser beam with a mechanical cover.

  1. Adjust the height of the stage for maximum PMT output. Next, adjust the mirror leading to the PMT for the maximum output. The adjustments for this mirror should be slight.

Note: If you want to adjust the z-location of the stage manually, you will need to unplug the power to the stepper motor driver.

  1. Select the desired field of view, samples per line, number of lines, number of vertical slices, and distance between slices.
  2. Run the code.
Note: It is ideal to place the item you want to scan (particularly if it is a bead) slightly away from the center of the field of view to prevent unnecessary photobleaching.
  1. In order to uniformly process all images in a 3D stack, the image intensity should normalize the maximum intensity of the brightest scan. Unfortunately, the code does not do this automatically, so it must be done manually. This should be improved upon in future iterations of the code.
  2. Identify the unprocessed image with the maximum intensity by finding the highest value of max(max(stackPMT(image,:))) among all images in the stack. Uncomment the code before the image processing for-loop and change maxImage in datamodPMT2c=stackPMT(maxImage,:)'./datamodPD; to be the value of the highest intensity point.
  3. Run only the image processing section of the code (highlight + F9) to produce the final image stack.