Phase Contrast Microscopy

Who we are
Paul Selvin (Instructor—Lectures)
Jaya Yodh (Instructor – Labs)
Marco Tjioe– TA (Selvin, Fluorescence, FIONA)
Alex Kreig – Sua Myong’s lab, AFM
Seongin Park (Ha, STORM)
Digvijay Singh (Ha, smFRET)
Jichuan Zhang (Ha, split between STORM, smFRET)
What we’re here for
Give you direct experience in lab manipulations
associated with modern biophysics
We are not here to lecture you! I’m irrelevant!
A big part is you must take responsibility for learning
We’re here to help.
Model is based on summer schools,
Taught for past 5 years, about 40 students/year,
1 week/year, very full-time.
A big emphasis will be on detection of Single Molecules
Physics 598BP
Format: 6 experimental labs will be offered in total. Plus M 4-5pm lecture
Lab 1 (Loomis): Ensemble Fluorescence Basics
Lab 2 (IGB): Bright Field & Fluorescence Microscopy
Lab 3 (Loomis): FIONA
Lab 4 (IGB): AFM
Lab 5 (Loomis): smFRET
Lab 6 (Loomis): STORM
Each experimental lab is taught over 2 weeks to 4 groups (Groups A-D) of 3
students in two consecutive 4-hr lab sessions (Tuesday or Thursday from 1-5pm).
The hands-on experiments and analysis will be mixed over each 2 week period (for
example – 6 hrs total of taking data & 2 hrs of analysis). Students will be expected
to do the remainder of the analysis on their own. In summary, 2 experimental
labs will be taught over a total of 4 weeks to 12 students
You must choose a lab time, Tuesday or Thursday
Good Resources
Our Web site
Course materials --Zeiss web site
(a fair amount of (today’s) lecture taken from this)
Molecular Biology of the Cell
Lab 1: Bright Field and Fluorescence
Optical Microscopy and Sectioning
Lab Concepts
(1) Basic Concepts in Microscopy:
(2) Magnification
Numerical Aperture and Resolution
Point Spread Function and Deconvolution
(2) Bright Field Imaging
Köhler illumination
(3) Enhancing Contrast in Optical Microscopy
Phase Contrast Bright Field Imaging
Differential Interference Contrast (DIC) Reflected Light Microscopy
(4) Fluorescence Imaging
(5) 3D-Imaging of thick specimen
Z-stack wide-field fluorescence Imaging and deconvolution
Apotome Sectioning (Structured Illumination Microscopy)
Introduction to seeing
Lens Maker Equation (for thin lenses)
A lens transfers an object plane to an image plane with some magnification.
Different lenses,
depending on curvature,
o have different magnification
4x  100x
Def’n: Object and image planes are conjugate planes.
An image is formed where one object point goes to one (and only one) image point.
In 3D, you have problems with out of focus light. (Need Deconvolution microscopy)
Numerical Aperture
Objective lens does this with some magnification and collecting
some fraction of the emitted/scattered light
Numerical aperture = NA = nsinq
n = Index of refraction of media
(n= 1.0 air; 1.33 water; 1.5 for immersion oil)
Higher N.A., can detect weaker fluorescence (highest NA= 1.49-1.68)
Also, higher NA gives you better Resolution
Infinity Objective Lenses
(now standard, greater flexibility)
Optical elements
Tube lens
(filters, etc.)
Fixed length (160-220 mm,
depending on company)
Infinity space
Point Spread Function
Even a “point” forms a finite spot on detector.
No matter how small an emitting light, it always forms a finite-sized spot,
PSF ~ l/2NA
You can “never” get better than l/2NA ~ 500 nm/2* 1.4 ~ 175 nm
(Caveat: can do 100x better with single molecules!)
PSF depends on NA
Accuracy & Resolution
You may be able to measure center really well.
Prism-type TIR 0.2 sec integration
Signal ~ N, # of photons
Noise ~ √N
(Photon number inherently varies)
= l/2
Y ax
Z-Data from Columns 1-21
W.E. Moerner, Crater Lake
X Da
How well can you tell where the center
Depends on width and S/N.
Accuracy = width/√N = 250nm/√N
10,000 photons.
Uncertainty = 
Accuracy = 25 nm/ sem
= 250 nm/ 100 = 2.5 nm
You can get a few nanometer accuracy with light that is several hundred
Light from all sources is Poissonian (varies)
N = 100: sometime N=90, 99, 102, 98 etc.
N ± √N
Light varies in two way:
It varies within a point spread function
--it lands anywhere within a PSF with
a certain probability.
-- The number of photons (i.e. the
intensity) varies in time.
Noise—either one varies like √N
Big effort to get a light sources with a fixed # of photons
Resolution: The Abbe or Rayleigh criteria
How well can you resolve two nearby (point) objects?
Light always
spreads out
to ~ l/2NA
The resolution is limited to how well you can separate two overlapping PSFs.
Rayleigh Criteria ~ ~ l/2NA ~ 200-250 nm
But, with single-molecule
imaging, can beat this.
(e.g. have one green and one
Cells discovered with invention of microscope.
Or with CCD
1000x, 0.2 um
Molecular Biology of the Cell. 4th edition.
Alberts B, Johnson A, Lewis J, et al.
New York: Garland Science; 2002.
Bright-field Microscopy (Standard configuration)
Limits to sensitivity
N photon
N- d
detector 1
detector 2
When can you detect that grey object is there?
Tell that detector 2 has more photons than detector 1
Ideal: you have the light source as good as possible—no spatial (or temporal) heterogeneity
You have intensity I: N photons on average:
there is √N variation inherently
Bright-field con’t
N = 100: sometime N=90, 99, 102, 98 etc.
N ± √N
Can tell that grey object is there is if:
N2 – N1 > (N12 + N22)1/2
Ideally, you’d hit the sample with a really bright light source
What if detector isn’t perfect? Noise is larger than √N?
This is very common. Use Koehler illumination
Köhler illumination:
minimize sample variation due to excitation
(used with sources like light bulbs; irrelevant for lasers)
B. Köhler Illumintion:
(old technique)
Conjugate planes are
A. Critical
the illuminating bulb
filament and
Conjugate planes are
the illuminating bulb
diaphragm. Second
filament and sample
conjugate planes are
plane (O). When
the Field diaphragm
adjusted correctly,
and the sample plane.
the image of the
When adjusted
filament is seen
correctly, the image of
coincident with the
the field diaphragm
sample image. A
and the sample are
The trick is to make sure that you are not
diffusing glass filter imaging the light source. The filament is out of coincident. The
(d) is used to blur
the plane of focus, and thus uniformly diffuse. filament is out of the
the filament image.
plane of focus, and
FD: Field diaphragm: CD: Condenser diaphragm thus uniformly
What if you can’t see something by bright-field?
(Contrast too low)
Optical elements
(filters, etc.)
Infinity space
Use dyes to generate more contrast
Enhancing Contract in Optical Microscopy
Investigations dealing with inherently low-contrast specimens, such as unstained
bacteria, thin tissue slices, and adherent live cells, rely on specialized contrastenhancing techniques to assist with imaging these virtually transparent samples.
Use dyes
(w color contrast)
Two Phase techniques
Polarized light requires
birefringence (usually not
present to a significant
degree in animal cells) to
generate contrast. Muscle
cells are birefringent.
Darkfield Microscopy
Great technique if it works
What if you can’t see something by bright-field?
(Contrast too low)
Modulate phase
Infinity space
Rely on phase of light (instead of amplitude)
What if you can’t see something by bright-field?
Light has a phase, (plus an amplitude)
You may be able to see a phase change.
Bright-field Microscopy –Phase Contrast
Limits to sensitivity
N photon
sin wt
A sin wt + f
Asin wt
Have to interfere with each other,
i.e. end up hitting detector at same place.
Two Phase Techniques:
Phase Contrast and Digital Interference
Both rely on phase difference between the sample
and background
Phase contrast yields image intensity values as a function of specimen optical path
length magnitude, with very dense regions (those having large path lengths)
appearing darker than the background. Alternatively, specimen features that have
relatively low thickness, or a refractive index less than the surrounding medium, are
rendered much lighter when superimposed on the medium gray background.
Good for thin samples.
DIC: optical path length gradients are primarily responsible for introducing
contrast into specimen images: Really good for edges. Thick samples; can be
used with high numerical aperture lenses
Phase Contrast Microscopy
Very little absorption, so Brightfield and Darkfield isn’t good.
Phase contrast is an excellent method to increase contrast when viewing or
imaging living cells in culture, but typically results in halos surrounding the
outlines of edge features.
The technique is ideal for thin
unstained specimens such as
culture cells on glass.
(which are approximately 5 to 10
micrometers thick above the nucleus,
but less than a micrometer thick at the
periphery), thick specimens
(such as plant and animal
tissue sections).
The amount of the phase shift
depends on what media
(refractive index) the waves have
passed through on their paths,
and how long the paths were
through these media.
Slight differences in phase are translated
into differences in intensity
Phase contrast microscopy
Notice red line, which contains a different phase due to sample is not phase
shifted. They interfere with light that is unrefracted.
Differential Interference Microscopy
Differential interference contrast microscopy requires plane-polarized light and
additional light-shearing (Nomarski) prisms to exaggerate minute differences in
specimen thickness gradients and refractive index. Good for thick samples.
Can use high numerical apertures (in contrast to Phase contrast.
Lipid bilayers, for example, produce excellent contrast in DIC because of the
difference in refractive index between aqueous and lipid phases of the cell. In
addition, cell boundaries in relatively flat adherent mammalian and plant cells,
including the plasma membrane, nucleus, vacuoles, mitochondria, and stress
fibers, which usually generate significant gradients, are readily imaged with DIC.
DIC—a Polarization-type of microscopy
Condensor splits light
into two orthogonal
polarizations and
slightly shifts them
laterally shifts these partial
beam such a way that a
small lateral displacement
of the wavefronts occurs
where regions of thickness
or refractive index vary. If
the two partial beams now
pass through exactly the
same structures, no further
path difference will occur in
the specimen (Figure 5(a)
and Figure 5(c)). However, if
the two partial beams see
The End

similar documents