MICROSCOPE ALIGNMENT

The gantry system translates the microscope along 3 axes (X,Y,Z) and rotation (θ).  Ideally, the beam should not deviate during movement of any of the four translations.  Any movement results in a deviation of the beam onto the back aperture objective and consequently affects the scanned image.  Therefore, a multi-mirror periscope which consists of 6 mirrors (See picture below) was engineered in order to allow sufficient control of the beam alignment onto the vertical breadboard.  Completion of this alignment defines the beam propagation on the vertical breadboard to which all vertical breadboard optics should be aligned to.

Coarse alignment of gantry periscope:  The first step is a coarse alignment, this just ensures that the beam is roughly going through the center of each of the six periscope mirrors.  This safeguards that the beam won’t clip on the housing when fine adjustment commences. 

-Adjust the last two mirror from the prism compressor such that the beam is parallel to the plane of the table and hits periscope mirror #1 at 45° directing the beam up to periscope mirror #2. 

-Adjust the knobs on periscope mirror #1 such that the beam is roughly centered on periscope mirror #2

-Adjust the knobs on periscope mirror #2 such that the beam is roughly centered on periscope mirror #3

-Adjust the knobs on periscope mirror #3 such that the beam is roughly centered on periscope mirror #4

-Adjust the knobs on periscope mirror #4 such that the beam is roughly centered on periscope mirror #5

-Adjust the knobs on periscope mirror #5 such that the beam is roughly centered on periscope mirror #6

It is essential that the beam does not deviate when any of the four translations of the gantry are moved.  In order to confirm this property each axis must be tested and aligned individually.  While one could monitor the deviation coarsely by eye for fine alignment a digital readout is essential.  Therefore, it is best to use a camera, beam profiler or position sensitive detector to monitor beam deviations.

Place the sensor far away from the periscope.  Ideally, this sensor is placed after the remote focus and objective jig. **At this point the telescope, remote focus objective and voice coil should be removed.  Use the vertical breadboard periscope to roughly center the beam through both irises from the jig onto the center of the detector.  This large separation ensures any small beam deviation will result in a detectable change in the position on the sensor.  First, start the gantry to one extreme of its travel and note the location of the beam on the sensor.  Next, move the gantry to the other extreme and record its position on the sensor.  If the input beam is perfectly aligned to the axis of travel the two positions should not change.  In practice some difference will always be observed albeit small since the alignment is limited by the linearity of travel along the axis.

A particular axis alignment is controlled by a specific periscope mirror for linear axes X, Y and Z and 2 periscope mirrors for rotation.  These relationships were defined by how the mirror mounts are attached to the gantry.  Consequently, there is a particular order and a specific beam alignment to how the periscope is aligned to the mounted breadboard.  The order is as followed;

-Periscope mirror #2 for Y axis.

-Periscope mirror #3 for X axis.

-Periscope mirror #4 and #5 for rotation.

-Periscope mirror #6 for Z axis.

1. At this point, if the the first quarter waveplate (QWP #1), outside of the microscope cage, is not rotated ideally, then there will be some change in transmitted intensity through the PBS as the microscope rotation is changed.  To fine tune the angle of QWP #1, use a power meter to check the beam power after the PBS (remote focus objective is removed) while rotating the microscope back and forth.  Observe the power difference, and make small adjustments to the rotation of QWP #1 (rotating QWP #2 each time afterwards to maximize transmission) until the power difference upon microscope rotation is minimized.

1. Remove the MainAlignmentJigInterfaceAssby and mount the imaging objective.
2. Prepare slide with large, deep well of fluorescein dye and proper coverglass thickness.  This slide needs to be mounted relative to the imaging objective at the right distance, and orthogonal to the optical axis.  There are two ways to do this:
   1. Using AccessoryHolder_Simple (part J005700A). Attach this to the objective, and attach dye pool slide to it, as shown in AccessoryHolder_Simple_Assby.dwfx.
   2. Taking advantage of the known beam collimation and alignment entering the imaging objective.  See video attached below.  The slide is mounted to the optical table and approached by the imaging objective, with field curvature correction turned off in ScanImage.  Use a full FOV, striped MROI setup in ScanImage such as roiTile10umPixels.roi. With the beam collimated at the imaging objective, the microscope should be lowered until the focus is 355um below the top of the dye pool.  This is the natural focus position, as designed, and all imaging of slides or animals should occur at this coverglass-objective separation (+/- 100um, see below) for ideal imaging.  Then, the remote focus (fast Z in SI) should be adjusted until the focus is just below the top of the dye pool (depth/V scaling in SI MDF should be adjusted so that the proper values, in um of depth, can be used in SI).   A spot smaller than the FOV should be seen in the image, which has an outline that is the intersection of the curved imaging plane with the top of the dye pool.  This spot will move in the imaged FOV as the objective-slide relative angle is changed, and should be centered when the optic axis and slide are orthogonal, as long as the beam is aligned to the optic axis at the center of the scan field (previous alignment step).  Adjust the objective-slide angle (one axis is accomplished through microscope rotation, and the other can be adjusted by putting the slide on a goniometer or other tilt stage) until the imaged spot is centered.Alignmenttomicroscopeusingdyepoolimage-1.m4v
      1. At Janelia, it was found that the two methods above did not agree very well.  That is, when the beam was adjusted to be collimated and angularly aligned at the objective, and when using the AccessoryHolder_Simple, the focus was not found to be 355um below the top of the dye pool (there was a ~200um difference from what was expected here), and the spot showing the intersection of the imaging sphere and dye pool top was not centered in the imaged FOV.  It's not known what caused these discrepancies, but some amount of compensator adjustment was carried out using both methods of initial alignment, and it seemed that similar final performance should be achievable in either case. It may be a good idea to check both alignment methods, but as long as the discrepancies between them are less than ~250um in distance, and 1 deg in angle, similar final performance can probably be achieved, and it is probably best to use the alignment method that is most like what will be used when doing daily experiments.  I.e. if the animal is attached to the optical table, the latter method can be used to align the objective and animal, and should be used to do the compensator adjustment; if the animal is attached to the objective somehow with a position and accuracy similar to that designed into AccessoryHolder_Simple, then the former method should be used.
      2. A THIRD method for aligning the slide angle is using the Coverglass Alignment module in the accessory optical system.  See attached video below.  This system can be calibrated using the AccessoryHolder_Simple part, most easily by mounting a piece of glass such as an OD 4 ND filter on it (this will show a reflection that is imaged onto the coverglass alignment camera only from the top surface).  With this alignment step, using this system should give an alignment equivalent to mounting the slide onto the AccessoryHolder_Simple part, and using this system should allow a greater alignment sensitivity than the method of observing the intersection of the image sphere with the coverglass bottom.  That being said, this extra sensitivity may not lead to much increase in signal level, as the objective-coverglass alignment was not found to very sensitively affect the signal level.Coverglassalignmentsystem-1.m4v
3. Turn on field curvature correction in SI (see adjustment video below), adjust Fast Z (remote focusing) until focus is a little below the top of the dye pool, such that it's easy to see how well the field curvature correction is working.  Adjust the field curvature correction radius of curvature (in MDF) and Fast Z Lag Time (in SI Fast Z dialog box) until field curvature correction works as well as possible.Fieldcurvaturecorrectionadjustment-1.m4v
4. With slide and objective properly aligned, begin MROI imaging with field curvature correction on.  Set up ~15 small (500 x 500 um) ROIs around the entire volume FOV.  At Janelia we use 5 ROIs–top, bottom, left, right, center–in three planes.  As to the precise location of the ROIs, they should be close to the edges of the used FOV, so that performance at these locations is optimized, but not TOO far out.  Edges at 2.4mm from the center, and planes at 100um below the top (make sure no ROIs image above the top of the dye pool), 300 um, and 600 um should work fine.  See example roi group roigroupOptimize_NotFullResWidth_ToEdges_100-500-1000.roi (these perhaps go too deep–we were checking limits of remote focusing performance).  Make sure there are no bubbles, artifacts, or dye pool walls near any of the ROIs, as these may wander into the ROIs during optimization.  Image all ROIs in MROI mode, and with a few frames rolling average to reduce noise, then adjust compensator knobs (putting arm through small door in side of cage while room is darkened) to maximize the average signal over all ROIs.  The SI user function userFunctionBar.m can be used to display the integrated signal across each ROI in real-time (first n bars, where n is the number of ROIs), plus the average signal over all ROIs (n+1 bar).  At first, use the bars function with norm=0 (parameter in userFuntionBar.m file) to simultaneously roughly maximize the signal across all ROIs and make sure that the signal is not highly non-uniform across the ROIs.  It will probably be impossible to make the signal highly uniform across the ROIs, and there may be some trade-off between uniformity and maximum total average signal.  To fine tune, switch to norm=1 to see the signal relative to the last acquisition made.  See video below.  Here is the basic compensator adjustment strategy:
   1. There are two sets of adjusters which works more-or-less orthogonally to each other–the translation and tilt along one axis, and the translation and tilt along the other axis.  Look at the fold mirror mount to figure out which tilt goes along with which translation axis.  ALL 4 knobs can misalign the remote focus objective and imaging objective pupils, so the signal intensity of the ROIs will depend on each of them sensitively.  However, it will be found that for each pair of adjusters there is a global maximum that can be found by tweaking one knob, remaximizing with the other, and then looking at how the signal level changes as this is repeated.  In effect, this is adjusting the center of the scan range for this axis while keeping the pupils aligned, and this has a much smaller signal sensitivity than the pupil alignment itself.  Iterate back and forth a few times between optimizing with two adjusters, and then the other two.Compensatoradjustmentsetup-2.m4v
5. At any point in the MROI signal optimization imaging, the position of the retroreflector in the prism pair compressor can be optimized.  This should be done.  Due to the properties of the retroreflector, it can be freely moved without adjusting the beam alignment downstream of the PPC.  Unless this was already set very close to the optimum location (optimum GDD), a large gain in signal level should be seen.
6. After optimizing the signal using the 4 compensator knobs, try adjusting the galvo Y1 gain adjustment, which should in effect adjust the location off the Y galvo scan pupil location.  If the signal can be improved with this adjustment, then iterate back to the 4 alignment compensator knobs again.
7. Finally, see if the signal can be improved through adjusting the beam alignment entering the microscope using the 4 knobs on the small beam periscope.  Ideally, the sensitivity should be low here, and the knobs should be at or near their optimum positions, since moving these should misalign the beam going into the remote focus objective.  If a large improvement in signal can be made, then perhaps the beam got misaligned before starting the entire procedure.
8. NOTE! After adjusting the compensator knobs, the center of the scan range will probably no longer be aligned as well to the optical axis of the imaging objective.  With the field curvature correction off and the focus slightly below the surface of the dye pool, look at where the image sphere/dye pool top intersection spot is located in the FOV.  If the location of this spot is used in the future to align the objective-coverglass angle for imaging, then the current position (NOT center position) should be used as the goal.