Transcript lecture 6 confocal top down

object
image
detector
scanning microscope, magnification defined by
display size / scan size
full field microscope, magnification defined by image
distance/object distance
image
Focal plane of lens
In epi-illuminated light
microscopy, , all arrows are
double headed.
object
Focal plane of lens
object
Probe formation and signal collection in light microscopy
(case study using Marshall U. MRC 1024 CSLM)
laser in, signal
from sample out
In conventional microscopy
the image point intensities of
tiny objects are defined
primarily by the collection
pathway, in CSLM the
illumination and detection
pathways are equally
important to defining this spot
that is imaged. Both
pathways travel through the
objective lens and the
confocal scanhead.
X
Y
Z not shown (height is intensity)
sample
Where do the
excitation and
emission
filters go?
Where is the
confocal iris?
Illumination and detection path in MRC1024 confocal
gas, KrAr laser
PMT detector
optical filters (Neutral Density and excitation λ)
confocal iris
optical fiber to carry laser to scanhead
emission λ filter
dichroic mirror (beam splitter) split excitation
from emission )
dichroic mirror
scan mirrors (galvo driven x & y)
scan mirrors
objective lens
objective lens
sample
sample
EX EM
Different light sources have different spectra
488nm
568nm
647nm
This is just one
of the Hg lines
at much higher
spectral
resolution.
Low pressure gasses emit discreet lines of color as in this low pressure Hg
spectrum (grey) overlaid with 3 of the major lines from a KrAr mixed gas laser,
also low pressure (color).
gas
gas
gas
black body
black body
Metal halide are very similar to gas arc but
include metal iodides and bromides in the
gas mix. Longer life and tunable spectrum
(based on the metal used).
All adapted from Murphy 2001
High pressure gasses in arc lamps emit relatively broad spectrum light (many more vibrational
energy states than low pressure). Black body radiation as emitted from the tungsten filament in
halogen/tungsten or standard tungsten filament sources varies with filament temperature and are
UV and blue poor as well as being inefficient (heat).
Neutral Density Filtering (all λ are affected equally)
Beer / Lambert relationship
Transmittance: T = P / P0
% Transmittance: %T = 100 T
Absorbance (OD):
A=-log T
A = log10 P0 / P
A = log10 1 / %T
P0
P
A = ε (pathlength) concentration
ε depends on the filter material, pathlength is thickness
of filter, concentration is the conc. of material with ε in the filter
Lets say this filter blocks 70% of the incident light so its T=.30
So, this filter has an A=.125
What happens if we double the width of the filter? T= ? A=?
What is the A if the filter blocks out 99.7% of the incident light?
Short and long pass filtering with colored (colloidal) glass filters; these
do not affect all wavelengths equally
Absorbance
<515nm
A<515=-log T<515
515 LP or
OG 515
Filtering with interference filters (most but not all band pass filters are
interference filters)
All of our confocal band pass filters and dichroic mirrors are interference filters, see website
below or Slayter & Slayter 4.3. Also see the optical thin film example on course website.
These filters are given
absorbance and OD values
but they DO NOT follow the
Beer Lambert relationship
with thickness and
concentration.
http://hyperphysics.phy-astr.gsu.edu/hbase/hph.html
Detector
Signal in
Signal out
DM
sample on
microscope
Emission light (pink color) follows the same path as the
excitation light (red) until we reach the dichroic mirror
(DM). This DM prevents reflected laser signal from
causing decreased contrast of our fluorescence signal.
Laser
Purdue Univ.
Many lens elements with complex
shapes are needed to bend and
fold the light so that rays of
different colors and those closer or
further from the optical axis are
focused to the same place. When
the objective lens is perfectly
designed and built, only then do
we achieve ‘diffraction limited
resolution’ as defined by the next
slide.
Murphy 2001
These objective lenses are corrected for
chromatic and spherical aberrations and
curvature of field. Apochromats are corrected
at more wavelengths (colors) than achromats.
Gaussian image points vs. Abbe’s theory of image formation; Gaussian ray tracing
brings rays to an infinitely small and unacheivable point in space (what is dmin).
Abbe’s theory of image formation states that a small interference pattern is formed
by the perfect lens, this is the diffraction limited resolution. If you can develop
technology that elegantly overcomes this barrier, you could win the Nobel prize.
dmin = 1.22(λ) / NAcond + NAobj
dmin = 1.22(λ) / 2 sin α η
dmin = .61(λ) / NAobj (for epi- microscopy)
NA (numericalaperture) = sin α η
Gaussian theory works well
for image formation of
features larger than the dmin.
Abbe’s theory works for describing
features that are approximately the
size of dmin or ~ the wavelength of
light (or electrons) used.
What does the confocal iris or pinhole do?
Airy pattern with
central disk.
Iris and diaphragm
This is a 2D view of a 3D phenomenon.
Murphy 2001
An interference pattern like this is projected at
the PMT for each object/ image point as the
beam scans the sample. The confocal iris or
pinhole selects only part of this Airy pattern to
reach the PMT, this can improve resolution.
Its effect on the z-axis diffraction pattern (not
shown) is even more dramatic. This is why
with confocal we can take optical z-sections.
Y
X
Z
From Pawley, 1995
Yes, these are secondary e- like in SEM.
Total voltage drop in a PMT may
equal thousands of volts.
Photocathode, usually coated,
must be sensitive to Einstein’s
photoelectric effect.
(+++++)
(++++)
(+++)
(++)
(+)
(-)
This PMT detector gives no
spatial information on its own,
only gives counts. The eye
does give spatial information
directly. It has an array of
detectors (neurons). A ccd or
film camera also directly
record spatial information.
(- -)
(- - -)
emission filter
Adapted from university of British Columbia Physics
Eye is most sensitive to green
light, PMTs can have
photocathodes made of materials
sensitive to specific colors
(Murphy spectra p.25 )
http://www.olympusmicro.com/primer/flash/photomultiplier/index.html
Now lets include system noise into the detector. Lets assume that
we get the equivalent of 1 photon/second of noise due to a light
leak (this could also be electronic noise). Lets also assume that
we get 16 photons/sec in real signal. Our PMT converts each
photon into 10 electronic counts (gain of 10).
Given the above information, what is the signal to noise (S/N)
ratio (in electronic counts) per pixel in two cases;
1-scan rate= 1 pixel/second
2- scan rate = 1 pixel/10seconds?
This can also be called signal/background ratio.
(+++++)
One step further, all detector systems have some inherent
ability to detect differences in intensity. It is not the signal /noise
ratio that is important here, it is the absolute difference in signal
intensity between 2 sample points.
(++++)
(+++)
Try this with this 4x4 pixel, 8 bit image. Which pairs of pixels
can be differentiated from each other?
Each pixel pair has the same signal / signal ratio but the absolute
differences vary.
1
2
(++)
(+)
(-)
4
8
10
20
40
80
(- -)
(- - -)
emission filter
Adapted from university of British Columbia Physics
fast (? sec/scan)
slow (4 sec/scan)
Slowest (? sec/scan)
When sampling with more pixels, adjust scan rate to increase photons/pixel; this is a
good idea unless you are worried about beam damage! The top row of 3 scans (not in
box) was done at constant scan rate (what is the scan rate for the top row?) The
photon flux from the sample in all cases is 16 photons/second.