Downloaded from http://cshprotocols.cshlp.

org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by

Cold Spring Harbor Laboratory Press

Polarized Light Microscopy: Principles and Practice

Rudolf Oldenbourg

Cold Spring Harb Protoc; doi: 10.1101/pdb.top078600

Email Alerting Receive free email alerts when new articles cite this article - click here.
Service

Subject Browse articles on similar topics from Cold Spring Harbor Protocols.
Categories
Cell Imaging (454 articles)
Imaging/Microscopy, general (531 articles)
Light Microscopy (54 articles)

To subscribe to Cold Spring Harbor Protocols go to:

http://cshprotocols.cshlp.org/subscriptions

© 2013 Cold Spring Harbor Laboratory Press

 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Topic Introduction

Polarized Light Microscopy: Principles and Practice

Rudolf Oldenbourg

Polarized light microscopy provides unique opportunities for analyzing the molecular order in het-
erogeneous systems, such as living cells and tissues, without using exogenous dyes or labels. This article
briefly discusses the theory of polarized light microscopy and elaborates on its practice using a tradi-
tional polarized light microscope and more specialized polarization microscopes such as the LC-
PolScope, Oosight, or Abrio. The microscope components specific to analyzing the polarization of
light, such as polarizer and compensator, are introduced, and quantitative techniques for measuring
the birefringence of the specimen point by point using a traditional polarizing microscope are dis-
cussed. The new LC-PolScope greatly improves the analytic power of the technique, providing quan-
titative birefringence data simultaneously for every image point, thereby revealing molecular order
with unprecedented sensitivity and at the highest resolution of the light microscope. Practical aspects
discussed include the choice of optics, sample preparation, and combining polarized light with dif-
ferential interference contrast and fluorescence microscopy. A glossary of polarization optical terms is
also included to facilitate the discussion of observations made with a polarized light microscope.

INTRODUCTION

Polarized light microscopy probes the local anisotropy of a specimen’s optical properties such as
refraction and absorption. Anisotropy of the refractive index is called birefringence, whereas anisot-
ropy of the absorption coefficient is called dichroism. Specimens with either property show a char-
acteristic variation of intensity as they are rotated between crossed linear polarizing filters. In this short
introduction to the principles and practice of polarized light microscopy, we will describe the basic
components of a polarizing microscope and its use for identifying specimen anisotropy and for
measuring the birefringence of anisotropic structures that typically occur in biological tissues and
living cells.
Optical anisotropy is a consequence of molecular order, such as is found in crystals. In materials
with molecular order, the absorption, refraction, and scattering of light typically become dependent on
the orientation of the material with respect to the polarization of the light. Polarized light microscopy
exploits this dependency and provides a sensitive tool to analyze the alignment of molecular bonds or
fine structural form in a specimen. By its very nature, polarizing microscopy provides structural
information at a submicroscopic level, including the alignment of polymers in plastics and filaments
in biological tissues and cells. Neuronal processes like axons and dendrites are birefringent because
microtubules and other filaments align parallel to the extended processes. Hence, the polarizing
microscope gives us information not only about where and when certain structures form and
change inside organisms, tissues, and cells, but also about some of the submicroscopic features of
these structures. Most importantly, the polarizing microscope can perform live cell imaging dynam-
ically, repeatedly in short time intervals and over long time periods, at the highest resolution of the

Adapted from Imaging: A Laboratory Manual (ed. Yuste). CSHL Press, Cold Spring Harbor, NY, USA, 2011.
© 2013 Cold Spring Harbor Laboratory Press
Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600

1023
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

light microscope, and with only minimal interference to the physiological conditions required for
live specimens.
This article gives a brief discussion of the theory of polarized light microscopy and its practice
using a traditional polarized light microscope and more specialized polarization microscopes such as
the LC-PolScope (Cambridge Research and Instrumentation, Inc. [now part of PerkinElmer Life
Sciences]; http://cri-inc.com), Oosight, or Abrio. For lucid discussions of polarized light microscopy
and its application to biophysical inquiries inside living cells, the reader is referred to articles and
books by Shinya Inoué, including the recent publication of his collected works (Inoué 1986, 2002,
2007, 2008; Inoué and Oldenbourg 1998).

TRADITIONAL POLARIZED LIGHT MICROSCOPY

Basic Setup
The polarized light microscope (also called polarizing microscope or polarization microscope; Fig. 1)
generally differs from a standard transilluminating microscope by the addition of a polarizer before
the condenser; a compensator slot and analyzer behind the objective lens; strain-free optics; a grad-
uated, revolving stage; centerable lens mounts; crosshairs in the ocular aligned parallel or at 45˚ to the
polarizer axes; and a focusable Bertrand lens that can be inserted for conoscopic observation of
interference patterns in the back aperture of the objective lens.

Polarizers
Most light sources (halogen bulb, arc burner, light-emitting diode) generate unpolarized light; hence,
the first polarizer located before the condenser optics polarizes the light that illuminates the specimen.
The second polarizer serves to analyze the polarization of the light after it passed through the speci-

A B
Eye P

Ocular

z y
x
Analyzer
(Crossed to polarizer) A

Objective lens
Sample C C
on rotatable stage
P
Condenser lens
ts
Fa

Compensator Slow
in rotatable mount
Polarizer

Light source A

FIGURE 1. Traditional polarized light microscope schematic and image cartoons. (A) Schematic of optical arrange-
ment of a conventional polarizing microscope. (B) Cartoon depicting at its center the image of an aster as it appears
when located between a crossed polarizer P and analyzer A. The arrows on the polarizer and analyzer sheet indicate
their transmission directions. An aster is made of birefringent microtubule (MT) arrays radiating from a centrosome
(aster diameter = 15 µm). MTs that run diagonal to the polarizer and analyzer appear bright, whereas MTs that run
parallel to the polarizer or analyzer appear dark. (C ) Cartoon of an aster as it appears when located between polarizer,
analyzer, and a compensator C, which is made of a uniformly birefringent plate. The arrow in the compensator plate
indicates its slow axis direction. Microtubules that are nearly parallel to the slow axis of the compensator appear bright,
whereas those that are more perpendicular to the slow axis are dark. Therefore, the birefringence of microtubules has a
slow axis that is parallel to the polymer axis, as is the case for many biopolymers.

1024 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Polarized Light Microscopy

men; therefore, it is called the analyzer. Usually the polarizer and analyzer are in crossed orientation,
so that the analyzer blocks (absorbs) most if not all of the light that has passed through the specimen.
Therefore, the image of the specimen looks mostly dark, except for structures that are birefringent or
otherwise optically anisotropic, and appears to be bright against the dark background. When the
specimen is rotated on a revolving stage (around the axis of the microscope), the birefringent parts
change brightness, changing from dark to bright and back to dark four times during a full 360˚
rotation. A uniformly birefringent specimen part appears darkest when its optical axes are parallel to
polarizer and analyzer. This is called the extinction orientation (or extinction position). Rotating the
specimen by 45˚ away from the extinction orientation makes the birefringent part appear to be
brightest. Not all birefringent parts in the field of view will turn dark at the same time, because in
general each part has different axis orientations. In summary, by rotating the specimen between
crossed polarizers, one can recognize birefringent components and determine their axis orientations.

Compensator
Although not absolutely necessary for some basic observations, especially of highly birefringent
objects, the compensator (1) can significantly improve the detection and visibility of weakly birefrin-
gent objects, (2) is required to determine the slow and fast axis of specimen birefringence, and (3) is an
indispensable tool for the quantitative measurement of object birefringence. There are several types of
compensators; most of them are named for their original inventors. For the observation of weak
birefringence, typically encountered in biological specimens, the Brace–Köhler compensator is most
widely used. It consists of a thin birefringent plate, often made from mica, with a retardance of 1/10 to
1/30 of a wavelength (λ/10 to λ/30; for definition of retardance, see below). The birefringent plate is
placed in a graduated rotatable mount and inserted either between the polarizer and condenser, as in
Figure 1, or between the objective lens and the analyzer. The location varies between microscope
manufacturers and specific microscope types. In either location, the effect of the Brace–Köhler
compensator on the observed image is the same and its standard usage is independent of these
locations. In general, the birefringence of the compensator, when inserted into the optical path,
causes the image background to become brighter, whereas birefringent specimen parts can turn
either brighter or darker than the background, depending on their orientation with respect to the
compensator. If the birefringent structure becomes brighter, its slow axis aligns more parallel to the
compensator slow axis, whereas if the structure turns darker, then its slow axis aligns more perpen-
dicular to the compensator slow axis. Hence, the compensator with known slow axis orientation can
be used to determine the slow axis of birefringent parts in the specimen. The compensator can also
enhance specimen contrast, and it is used to quantify specimen birefringence by measuring its
retardance, as discussed next.

Birefringence, Retardance, and Slow Axis

Birefringence provides the unique opportunity to measure and analyze the structural order of spec-
imens without having to treat them with exogenous dyes, fluorescence labels, or stains, or to resort to
destructive and expensive electron microscopy. Birefringence occurs when there is a molecular order,
that is, when the average molecular orientation is nonrandom, as in crystals or in aligned polymeric
materials. Molecular order usually gives the material two orthogonal optical axes, with the index of
refraction along one axis being different from that along the other axis. The difference between the two
indices of refraction is called birefringence and is an intrinsic property of the material:

birefringence = Dn = n|| − n⊥ ,

where n|| and n1 are the refractive indices for light polarized parallel or perpendicular to the two
optical axes.
Light polarized parallel to one axis travels at a different speed through the sample than does light
polarized parallel to the orthogonal axis. As a result, these two light components, which were in phase

Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600 1025
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

when they entered the sample, are retarded at a different rate and exit the sample out of phase. Using
the polarizing microscope, one can measure this differential retardation, also called retardance, and
thereby quantify the magnitude and orientation of molecular order in the specimen. Retardance is an
extrinsic property of the material and a product of the birefringence and the path length l through the
material:

R = Dn · l.

In addition to retardance, birefringence has an orientation associated with it. The orientation
refers to the specimen’s optical axes: One is called the fast axis and the other is called the slow axis.
Light polarized parallel to the slow axis experiences a higher refractive index and travels more slowly
than light polarized parallel to the fast axis. In materials that are built from aligned filamentous
molecules, the slow axis is typically parallel to the average orientation of the filaments. Birefringence
orientation always correlates with molecular orientation, which usually changes from point to point in
the specimen (see the aster in Fig. 1B,C).

Quantitative Analysis of Specimen Retardance

Birefringence and retardance are optical properties that can be directly related to molecular order,
such as the alignment of polymeric material. Using the traditional polarizing microscope, Sato, Ellis,
and Inoué, for example, have measured the retardance of mitotic spindles in living cells and
definitively concluded that the birefringence is caused by the array of aligned spindle microtubules
(Sato et al. 1975). Their measurements were made possible by the careful analysis of specimen
birefringence using a traditional compensator in addition to a polarizer and analyzer in the mi-
croscope optical train. Using a Brace–Köhler compensator, the retardance of a resolved image point
or uniformly birefringent area in the field of view can be measured by carrying out the following
steps.
1. With the specimen in place and viewed through the eyepiece (or on the monitor of a video camera),
make sure that polarizer and analyzer are crossed and the compensator is in the extinction position.
Given those settings, background areas that have no birefringence appear dark, and birefringent
structures change from dark to bright four times when rotating the specimen on the rotatable stage
by 360˚.
2. Again, using the rotatable stage, orient the birefringent structure of interest so that it appears
darkest (extinction position). Note the orientation and then rotate the specimen 45˚ away from the
extinction position. At that orientation the structure appears brightest. (By eye, it is usually easier to
determine the orientation that leads to the lowest intensity than the highest one. In any case, the
two orientations are 45˚ apart.)
3. Now rotate the compensator either clockwise or counterclockwise, until the birefringent structure
of interest appears dark. When rotating the compensator, any background area becomes brighter,
whereas birefringent structures become either brighter or darker. For structures that turn darker
than the background, the birefringence of the compensator is said to “compensate” the birefrin-
gence of the sample structure. Take as an example the aster image in Figure 1C. The radial array of
microtubules shows two bright quadrants and two dark quadrants. In the bright quadrants, the
microtubules’ slow axis runs more parallel to the compensator slow axis, whereas in the dark
quadrants the slow axis of the microtubule birefringence is oriented nearly perpendicular to the
slow axis of the compensator.
As the compensator is rotated from its extinction position, the specimen turns darkest at a
characteristic angle θmin. Given the known compensator retardance Rcmp, the angle θmin is a direct
measure of the specimen retardance Rspc:

Rspc = Rcmp sin(2umin ).

1026 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Polarized Light Microscopy

Intensity Ispc – Ibg

A B Contrast =
0.15 Ispc + Ibg
Ibg(Rspc = 0) Imin = 0 1.0
Ispc(Rspc = λ/50) Imin = 0.002
Imin = 0.01
0.1

0.5 –30° –20° –10° 10° 20° 30°

Compensator angle θ

Imin = 0.005
–90° –45° 0 45° 90°
θmin
–1.0
Compensator angle θ

FIGURE 2. (A) The graph shows intensities detected in a single image point or uniform area of the specimen versus the
rotation angle of a Brace–Köhler compensator. The dashed curve shows the background intensity Ibg of a specimen
area that shows no birefringence. The solid curve shows the intensity Ispc of a specimen point that shows a small
retardance (l/50). The transmitted intensity is given as a fraction of the amount of light that has passed through the first
polarizer. Imin represents the spurious intensity that is detected when the compensator is in the extinction position.
(B) The graph shows the contrast of a birefringent specimen point (Rspc = l/50) as a function of compensator angle, for
three different values of Imin. Imin is affected by the quality of the polarizers used and the polarization distortions
introduced by the intervening optics (e.g., condenser and objective lens). Both graphs were generated computationally
using the Jones calculus and assuming a retardance of l/10 for the birefringent crystal plate of the Brace–Köhler
compensator and a specimen retardance of l/50 with slow axis oriented at 45˚ to polarizer and analyzer.

Figure 2A shows a graph of the intensity calculated for a uniformly birefringent specimen area (or
single specimen point) as a function of the rotation angle θ of a Brace–Köhler compensator. At the
rotation angle θmin the intensity is a minimum. For other rotation angles the intensity varies according
to a complex expression of trigonometric functions that can be derived using the Jones calculus. This
expression can also be used to calculate the expected contrast of a birefringent object against its
nonbirefringent background. Figure 2B shows the computed contrast of the birefringent specimen
area versus θ. The contrast is defined as the intensity difference between the specimen area and the
background, divided by their sum:

Ispc − Ibg
contrast = .
Ispc + Ibg

As is apparent from Figure 2B, the highest contrast and therefore the best visibility of the bire-
fringent specimen are achieved when adjusting for a compensator angle of ±θmin.
Figure 2B further illustrates the importance of using so-called high extinction optics when ob-
serving a specimen with low birefringence. Figure 2B introduces the parameter Imin, which represents
the spurious intensity observed when polarizer and analyzer are crossed. The better the quality of the
polarizers used and of the intervening optical components (including condenser and objective lens),
the lower Imin will be and the higher the contrast that can be achieved for a specimen structure of given
retardance. High extinction optics will produce high contrast images that also result in images with
high signal-to-noise ratio (Oldenbourg 1999).

THE LC-POLSCOPE

The LC-PolScope is a birefringence imaging technique that was first developed in the author’s
laboratory at the MBL and is commercially available from Cambridge Research and Instrumentation,
Inc. (CRi, Inc.; now part of PerkinElmer Life Sciences; http://cri-inc.com) under the trade names
Abrio and Oosight. The optical design of the LC-PolScope builds on the traditional polarized light

Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600 1027
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

microscope, introducing two essential modifications: The specimen is illuminated with nearly circu-
larly polarized light and the traditional compensator is replaced by a liquid crystal-based universal
compensator. The LC-PolScope also requires the use of narrow bandwidth (≤40 nm) or monochro-
matic light. In Figure 3, the schematic of the optical train shows the universal compensator located
between the monochromatic light source (arc lamp with interference filter) and the condenser lens.
The analyzer for circularly polarized light is placed after the objective lens. The universal compensator
is built from two variable retarder plates and a linear polarizer. The variable retarder plates are
implemented as electrooptical devices made from two liquid crystal plates. Each liquid crystal plate
has a uniform retardance that depends on the voltage applied to the device. A computer-controlled
electronic circuitry supplies the voltage for each plate. The computer is also connected to the elec-
tronic camera, typically a charge-coupled device (CCD) camera for recording the specimen images
projected onto the camera by the microscope optics. Specialized software synchronizes the image
acquisition process with the liquid crystal settings and implements image-processing algorithms.
These algorithms compute images that represent the retardance and slow-axis orientation at each
resolved image point.
The commercial system, developed and distributed by CRi, Inc., is available as an accessory to
microscope stands of all major microscope manufacturers (Leica, Nikon, Olympus, Zeiss). It usually
includes the universal compensator, circular polarizer, a camera with control electronics, and a
computer with software for image acquisition and processing. Three slightly differing versions are
available, each optimized for research in the life sciences (Abrio LS), for research in industrial

Raw LC-PolScope images (intensity images)

CCD camera
Desktop computer
Circular analyzer

z y
x

Linear polarizer
Quarter wave plate

Objective

Retardance
Specimen

Condenser
Universal compensator

t
s

LC-B Variable retarder

Fa

Slow
slow axis 0° Retarder
LC-A Variable retarder controller
Slow axis orientation
t
Fas

Slow
slow axis 45°
Linear polarizer
transmission axis 0°
Interference filter
Arc lamp

Computed PolScope images

FIGURE 3. Schematic of the LC-PolScope. The optical design (left) builds on the traditional polarized light microscope
with the conventional compensator replaced by two variable retarders LC-A and LC-B. The polarization analyzer
passes circularly polarized light and is typically built from a linear polarizer and a quarter wave plate. Images of the
specimen (top row, aster isolated from surf clam egg) are captured at five predetermined retarder settings, which cause
the specimen to be illuminated with circularly polarized light (top row, leftmost image) and with elliptically polarized
light of different axis orientations (top row, second to fifth images). Based on the raw PolScope images, the computer
calculates the retardance image and the slow-axis orientation or azimuth image shown on the lower right.

1028 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Polarized Light Microscopy

FIGURE 4. Retardance images recorded with the LC-PolScope. (A) Living growth cone of an Aplysia bag cell neuron.
The peripheral lamellar domain contains radially aligned fibers composed of 15–40 actin filaments. A detailed
description of the birefringent fine structure including a time-lapse movie of the growth cone dynamics is published
in Katoh et al. (1999). Image brightness represents measured retardance between 0 nm (black) and 0.5 nm or larger
(white). Scale bar, 10 µm. (Reprinted, with permission, from Katoh et al. 1999, ©American Society of Cell Biology.) (B)
Meiotic spindle in living spermatocyte of crane fly Nephrotoma suturalis. At metaphase, chromosomes have con-
gressed to the spindle equator and kinetochore fibers, each composed of 60 microtubules, connect each bivalent to
opposite poles. A cage of birefringent mitochondria surrounds the spindle. (Reprinted from LaFountain and Old-
enbourg 2004, ©American Society of Cell Biology.) (Time-lapse movies of meiosis I and II recorded with the LC-
PolScope are available at http://cellimages.ascb.org/.) Image brightness represents measured retardance between 0 nm
(black) and 2.0 nm or larger (white). Scale bar, 10 µm.

metrology (Abrio IM), and for in vitro fertilization and related laboratory techniques (Oosight).
Figure 4 shows retardance images recorded with the LC-PolScope, illustrating the clarity, resolution,
and analytic potential of analyzing the birefringent fine structure in living cells.

Practical Considerations
Choice of Optics
As indicated earlier, the polarization distortions introduced by the objective and condenser lenses
limit the extinction that can be achieved in a polarizing microscope. Most microscope manufacturers
offer lenses that are designated “Pol” to indicate low polarization distortions, which can arise from a
number of factors, including stress or crystalline inclusions in the lens glass and the type of antire-
flection coatings used on lens surfaces. Some lens types are available only with “DIC” (differential
interference contrast) designation. “DIC” lenses do not meet the more stringent Pol requirements but
pass for use in DIC and can also be used with the LC-PolScope (without the Wollaston or Nomarski
prisms specific to DIC, of course).
The polarization performance of the most highly corrected lenses, so-called Plan-Apochromat
objectives, is often compromised by a large number of lens elements, special antireflection coatings
and, in some cases, special types of glass used to construct the lenses (Oldenbourg and Shribak
2010). For some applications, these high-quality lenses, which provide a large, highly corrected
viewing field over a wide spectral range, might not be required. If the objective lens is to be used only
with the PolScope that requires monochromatic light and typically acquires images from a region near
the center of the viewing field, a less stringent correction might well suffice. To find out what works best
for a particular imaging situation, several lenses should be tested. It is also helpful to be able to select the
best performing combination of condenser and objective lens from a batch of the same lens types.
Whenever possible or practical, oil-immersion lenses should be used. This is because the transition
of a light ray between two media of different refractive index (nair = 1.00, nglass = 1.52) introduces
polarization distortions, especially for high-numerical-aperture (high-NA) lenses. The peripheral rays
leaving the condenser front lens are highly tilted to the slide surface and their polarization is typically
rotated when traversing the air–glass interface. Oil and, to a lesser degree, water and other immersion
liquids greatly reduce polarization aberrations caused by air–glass interfaces between the specimen

Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600 1029
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

and the lenses. For a discussion of the origin of polarization distortions in lenses and optical systems,
and of ways to reduce them, see Shribak et al. 2002. This publication also discusses the various
conoscopic images that can be observed in a polarizing microscope equipped with crossed, or
nearly crossed, polarizers and a Bertrand lens.
Lenses and other optical elements that are part of the optical train but not placed between the
crossed polarizers do not affect the polarization performance of the microscope setup. Filter cubes for
epi-illumination, for example, should be placed outside the polarization optical train. (The polariza-
tion optical train is defined as the stretch between the crossed polarizers.) In a dissecting microscope,
the polarizing elements and compensator should be placed in front of the objective lens instead of
behind it. The objective of a dissecting scope has very low NA and the image-forming rays have a small
angle of divergence (in contrast to high-NA lenses.) Rays with a small tilt angle with respect to the
normal of the polarizer and compensator plates do not appreciably affect the performance of
these devices.

Specimen Preparation
The following are recommendations for preparing living cells for observation with a polarizing
microscope. Other resources for polarized light microscopy of living cells include Inoué (2002)
and Oldenbourg (1999), which have sections on suitable cell types and their preparation.
• If cells have to be kept in plastic Petri dishes during observations, only use dishes with glass
coverslip bottoms. The strong birefringence of a plastic bottom ruins the polarization of the light.
• Prepare specimens such that cells or structures of interest are as close to the coverslip as possible.
When using high-NA (>0.5) oil-immersion or dry optics, a layer of >10 µm of aqueous medium
introduces enough spherical aberration to noticeably reduce the resolution of cell images. The use
of water-immersion optics alleviates this problem.
• For observations of structures near the cell surface, it might be necessary to increase the medium’s
refractive index to match the refractive index of the cytoplasm and reduce the effect of edge
birefringence (Oldenbourg 1991). Adding polyethyleneglycol, polyvinylpyrrolidone, or some
other harmless polymer or protein substitute to the medium will increase its refractive index
to match that of the cytoplasm inside the cell. The concentration of substitute required might
vary from a few percent to >10%, depending on the average refractive index of the cytoplasm
of the cells of interest. It is best to test various media containing different polymer concentrations
by suspending cells in them and examining the refractive index mismatch in a slide and coverslip
preparation with DIC or polarized light. At the optimal polymer concentration, no distinct cell
boundary is visible, giving the impression that the organelles and cytoskeleton are freely suspend-
ed. After determining the optimal concentration for imaging, the medium should be tested
for compatibility with growing cells. Many cells grow and develop normally in media with inert
additives. The molecular mass of the polymer needs to be 40 kDa or more to prevent its uptake
into cells.
• For observations with the LC-PolScope, prepare the specimen so that, when mounted in the
microscope, at least one clear area without cells or birefringent material can be identified and
moved into the viewing field when required. A clear area is needed for calibrating the PolScope and
for recording background images. The background images are used to remove spurious back-
ground retardance from specimen images, allowing precise measurement and imaging of the cells
and structures of interest. Many cell cultures and cell-free systems can be prepared without special
attention to this requirement. For example, a free area can often be found around sparsely plated
cells. Cultures of free-swimming cells might have to be diluted before mounting a small drop of the
suspension between slide and coverslip to observe free areas. Sometimes it is helpful to add a tiny
drop of oil or other nontoxic, immiscible liquid to the preparation. This clear drop can provide an
area for calibrating the instrument and taking background images in a preparation that is other-
wise dense with birefringent structures.

1030 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Polarized Light Microscopy

Combining Polarized Light with DIC and Fluorescence Imaging

For DIC, two matching prisms are added to the polarization optical train (for the discussion of
a standard DIC arrangement, see Salmon and Tran [1998] and Oldenbourg and Shribak [2010]).
The two prisms, called Wollaston or Nomarski prisms, are specially designed to fit in specific
positions, one before the condenser and the other after the objective lens, and to match specific
lens combinations. The prisms in their regular positions can also be combined with the LC-PolScope
setup, in which case the universal compensator and circular analyzer take the places of the linear
polarizer and analyzer in a standard DIC arrangement. In fact, the same DIC prisms function equally
well using either linearly or circularly polarized light. Note, however, that Carl Zeiss has recently
combined linear polarizers with their condenser DIC prisms, thus preventing their use with circularly
polarized light.
For best results, the LC-PolScope is first calibrated without prisms to find the extinction setting for
the universal compensator. With the compensator set to extinction, the prisms are entered into the
optical path. Typically, one of the two prisms has a mechanism for fine-adjusting its position, which in
turn affects the brightness and contrast of the DIC image. In a PolScope setup, the mechanical fine
adjustment or the universal compensator can be used to add a bias retardance to change the brightness
and contrast of the DIC image. When using the universal compensator, retardance must be added or
subtracted from the extinction setting of either the LC-A or LC-B retarder (Fig. 3), depending on the
orientation of the shear direction of the prisms.
For live cell imaging, it is very useful to be able to switch easily between polarized light and DIC
imaging. DIC provides good contrast of many morphological features in cells using direct viewing
through the eyepiece or by video imaging. When viewing specimens that have low polarization
contrast, it is more effective to align the optics, including the visualization of the specimen, using
DIC. After the optics have been aligned and the specimen is in focus, the DIC prisms are removed
from the optical path and the PolScope-specific adjustments completed.
Fluorescence imaging can be combined with the PolScope in several ways. Fluorescence is com-
monly observed using epi-illumination, which requires a filter cube in the imaging path. The filter
cube includes a dichromatic mirror and interference filters for separating the excitation and emission
wavelengths. For best results the filter cube should be removed for PolScope observations to avoid the
polarization distortions caused by the dichromatic mirror. For observing fluorescence, on the other
hand, the PolScope analyzer should be removed, because it attenuates the fluorescence emission by at
least 50%. To meet both requirements, the PolScope analyzer can be mounted in an otherwise empty
filter cube holder, which is moved into the optical path as the fluorescence cube is moved out, and
vice versa.
If, however, the option of removing the fluorescence cube is unavailable, the cube can remain in
the optical path for observations with polarized light and the LC-PolScope. In this case, the following
points should be considered.
• For the light source of the PolScope, choose a wavelength that is compatible with the emission
wavelength filter in the fluorescence cube. For example, fluorescein isothiocyanate (FITC) requires
excitation with blue light (485 nm) and fluoresces in green. The FITC dichromatic beam splitter
and a barrier filter passing green fluorescence light with wavelength >510 nm would be compatible
with 546-nm light for polarized light observations using a mercury arc burner for the transmission
light path.
• The light source for one imaging mode must be blocked while observing with the other imaging
mode.
• Removing the polarization analyzer while observing fluorescence more than doubles the fluores-
cence intensity.
• The LC-PolScope must be calibrated with the fluorescence filter cube in place. The polarization
distortions caused by the dichromatic mirror are partially counteracted by the calibrated settings
of the universal compensator.

Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600 1031
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

CONCLUSION AND FUTURE PROSPECTS

Polarized light microscopy provides complementary information to the many imaging techniques
discussed in this series of manuals. In contrast to fluorescence microscopy, imaging with polarized
light reveals information about the organization of the endogenous molecules that built the complex
and highly dynamic architecture of cells and tissues. Although fluorescence identifies the chemical
nature of tagged structures, such as f-actin and microtubules, it does not provide information about
their submicroscopic architecture. Polarized light microscopy, on the other hand, is not specific to the
chemical nature but to the structural nature of macromolecular assemblies such as the submicro-
scopic alignment of molecular bonds and the architectural fine structure of the specimen.
Recently, the chemical specificity of fluorescence tagging and the structural information avail-
able through polarization analysis were combined in a study to reveal septin organization in yeast
bud necks (Vrabioiu and Mitchison 2006, 2007). The study took advantage of the polarized flores-
cence emitted by every single GFP fluorophore. GFP was fused to the protein of interest, in this
case septin, with a rigid linker that limits the GFP’s ability to rotate relative to the septin molecule.
When septin assembled into an ordered structure, a consistent orientation of the GFP dipoles was
detected throughout the structure. In the future we can expect more studies of this kind that
highlight specific structures through fluorescence tagging and analyze their organization by polar-
ized fluorescence.
From an optical point of view, polarized light microscopy is related to DIC and phase-contrast
microscopy, because all of them were designed to highlight changes of the refractive index of the
specimen, affecting the phase of the transmitted light. Phase-contrast and DIC microscopy highlight
changes of refractive index that occur from point to point in the specimen, whereas polarized light
microscopy highlights changes in refractive index that occur when changing the polarization of light.
The change of refractive index due to polarization is called birefringence and it turns that it can be
measured very sensitively using relatively simple optical means such as two polarizers and a compen-
sator. The measured quantity is retardance, which can be used to highlight and analyze the molecular
architecture, including the estimation of the number of filaments in a bundle or the alignment of
molecular bonds in a complex structure. It is this submicroscopic information, gleaned from live,
actively functioning cells, tissues, and whole organisms that makes polarized light microscopy an
indispensable tool in the array of analytic and quantitative experimental techniques that provide us a
glimpse into the secrets of life itself.

GLOSSARY OF POLARIZATION OPTICAL TERMS

The following is a brief introduction to terms that are relevant for observations with a polarized light
microscope. A more detailed explanation of these and other terms that describe physical phenomena
or optical devices can be found in the following references (Shurcliff 1962; Born and Wolf 1980;
Chipman 1995; Hecht 1998).

Analyzer
An analyzer is a polarizer that is used to analyze the polarization state of light (see Polarizer).

Azimuth
The azimuth is an angle that refers to the orientation of the slow axis of a uniformly birefringent
region. The azimuth image refers to the array of azimuth values of a birefringent specimen imaged
with the LC-PolScope. The azimuth is typically measured from the horizontal orientation with values
increasing for counterclockwise rotation. Angles range between 0˚ and 180˚, where both end points
indicate horizontal orientation.

1032 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Polarized Light Microscopy

Birefringence
Birefringence is a material property that can occur when there is molecular order, that is, when the
average molecular orientation is nonrandom, as in crystals or in aligned polymeric materials such as
those comprising the mitotic spindle in a dividing cell. Molecular order usually renders the material
optically anisotropic, leading to a refractive index that, in general, changes with the polarization of the
light. For example, assume that a beam of light passes through a mitotic spindle in a direction
perpendicular to the spindle axis. The light that is polarized parallel to the spindle axis experiences
a higher refractive index than the light that is polarized perpendicular to the spindle axis (i.e., the
mitotic spindle is birefringent). The birefringence originates from the parallel alignment of micro-
tubules (Sato et al. 1975). The difference between the two indices of refraction is called birefringence:

birefringence = Dn = n|| − n⊥ ,

where n|| and n1 are the refractive indices for light polarized parallel and perpendicular to one of the
principal axes.

Compensator
A compensator is an optical device that includes one or more retarder plates and is commonly used to
analyze the birefringence of a specimen. For a traditional polarizing microscope, several types of
compensators exist that typically use a single fixed retarder plate mounted in a mechanical rotation
stage. With the help of a compensator, it is possible to distinguish between the slow- and fast-axis
direction and to measure the retardance of a birefringent object after orienting it at 45˚ with respect to
the polarizers of the microscope.
The LC-PolScope uses a universal compensator that includes two electrooptically controlled,
variable retarder plates. Using the universal compensator it is possible to measure the retardance
and slow-axis orientation of birefringent objects that have any orientation in the plane of focus.

Dichroism
Dichroism is a material property that can occur in absorbing materials in which the light-absorbing
molecules are arranged in a nonrandom orientation. Dichroism refers to the difference in the ab-
sorption coefficients for light polarized parallel and perpendicular to the principal axis of alignment.
The measurement of optical anisotropy by the LC-PolScope is affected by the dichroism of
absorbing materials. In nonabsorbing, clear specimens, however, dichroism vanishes and birefrin-
gence is the dominant optical anisotropy measured by the LC-PolScope. Like absorption, dichroism is
strongly wavelength dependent, whereas birefringence only weakly depends on wavelength.

Extinction
Extinction is defined as the ratio of maximum-to-minimum transmission of a beam of light that
passes through a polarization optical train. Given a pair of linear polarizers, for example, the extinc-
tion is the ratio of intensities measured for parallel versus perpendicular orientation of the transmis-
sion axes of the polarizers (extinction = I||/I1). In addition to the polarizers, the polarization optical
train can also include other optical components, which usually affect the extinction of the complete
train. In a polarizing microscope, the objective and condenser lens are located between the polarizers
and significantly reduce the extinction of the whole setup.

Fast Axis
The fast axis describes an orientation in a birefringent material. For a given propagation direction,
light that is polarized parallel to the fast axis experiences the lowest refractive index, and hence travels
the fastest in the material (see also Slow Axis).

Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600 1033
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

Optic Axis
The optic axis refers to a direction in a birefringent material. Light propagating along the optic axis
does not change its polarization; hence, for light propagating along the optic axis, the birefringent
material behaves as if it were optically isotropic.

Polarized Light
A beam of light is said to be polarized when its electric field is distributed nonrandomly in the plane
perpendicular to the beam axis. In unpolarized light, the orientation of the electric field is random and
unpredictable. In partially polarized light, some fraction of the light is polarized, whereas the remain-
ing fraction is unpolarized. Most natural light is unpolarized (sun, incandescent light), but can
become partially or fully polarized by scattering, reflection, or interaction with optically anisotropic
materials. These phenomena are used to build devices to produce polarized light (see Polarizer).

Linearly Polarized Light

In a linearly polarized light beam, the electric field is oriented along a single axis in the plane
perpendicular to the propagation direction.

Circularly Polarized Light

In circularly polarized light, the electric field direction rotates either clockwise (right-circularly) or
counterclockwise (left-circularly) when looking toward the source. Although the field direction
rotates, the field strength remains constant. Hence, the end point of the field vector describes a circle.

Elliptically Polarized Light

In elliptically polarized light, as in circularly polarized light, the electric field direction rotates either
clockwise or counterclockwise when looking toward the source. However, although the field direction
rotates, the field strength varies in such a way that the end point of the field vector describes an ellipse.
The ellipse has a long and short principal axes that are orthogonal to each other and have fixed
orientation. Any type of polarization (linear, circular, or elliptical) can be transformed into any other
type of polarization by means of polarizers and retarders.

Polarizer
A polarizer, sometimes called a polar, is a device that produces polarized light of a certain kind. The
most common polar is a linear polarizer made from dichroic material (e.g., a plastic film with small,
embedded iodine crystals that have been aligned by stretching the plastic), which transmits light of one
electric field direction while absorbing the orthogonal field direction. Crystal polarizers are made of
birefringent crystals that split the light beam into orthogonal linear polarization components. A
polarizer that produces circularly polarized light, a circular polarizer, is typically built from a linear
polarizer followed by a quarter-wave plate.
The LC-PolScope uses a universal compensator that also serves as a universal polarizer in that it
converts linear polarization into any other type of polarization by means of two variable retarders.

Retardance
Retardance is a measure of the relative optical path difference, or phase change, suffered by two
orthogonal polarization components of light that has interacted with an optically anisotropic material.
Retardance is the primary quantity measured by the LC-PolScope. Assume a nearly collimated beam
of light traversing a birefringent material. The light component that is polarized parallel to the high
refractive index axis travels at a slower speed through the birefringent material than the component
polarized perpendicular to that axis. As a result, the two components, which were in phase when they

1034 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

Polarized Light Microscopy

entered the material, exit the material out of phase. The relative phase difference, expressed as the
distance between the respective wave fronts, is called the retardance:

retardance = R = (n|| − n⊥ ) · l = Dn · l,

where l is the physical path length or thickness of the birefringent material. Hence, retardance has the
dimension of a distance and is often expressed in nanometers. Sometimes it is convenient to express
that distance as a fraction of the wavelength λ, such as λ/4 or λ/2. Retardance can also be defined as a
differential phase angle, in which case λ/4 corresponds to 90˚ and λ/2 to 180˚ phase difference.
As a practical example consider a mitotic spindle observed in a microscope that is equipped with
low-NA lenses (NA ≤ 0.5). When the spindle axis is contained in the focal plane, the illuminating and
imaging beams run nearly perpendicular to the spindle axis. Under those conditions, the retardance
measured in the center of the spindle is proportional to the average birefringence induced by the dense
array of aligned spindle microtubules. To determine Δn, it is possible to estimate the thickness, l, either
by focusing on spindle fibers located on top and bottom of the spindle and noting the distance
between the two focus positions or by measuring the lateral extent of the spindle when focusing
through its center. The latter approach assumes a rotationally symmetric shape of the spindle. Typical
values for the spindle retardance of crane fly spermatocytes (Fig. 4B) and of other cells is 3–5 nm and
the spindle diameter is 30–40 µm, leading to an average birefringence of around 10−4. It has been
found that the retardance value of the spindle is largely independent of the NA for imaging systems
using NA ≤ 0.5 (Sato et al. 1975).
On the other hand, when using an imaging setup that uses high-NA optics (NA > 0.5) for illu-
minating and imaging the sample, the measured retardance takes on a somewhat different context.
For example, the retardance measured in the center of a microtubule image recorded with a LC-
PolScope equipped with a high-NA objective and condenser lens is 0.07 nm. A detailed study showed
that the peak retardance decreased inversely with the NA of the lenses. However, the retardance
integrated over the cross section of the microtubule image was independent of the NA (Oldenbourg
et al. 1998). Although a conceptual understanding of the measured retardance of submicroscopic
filaments has been worked out in the aforementioned publication, a detailed theory of these and other
findings about the retardance measured with high-NA optics has yet to materialize.

Retarder
A retarder, or wave plate, is an optical device that is typically made of a birefringent plate. The
retardance of the plate is the product of the birefringence of the material and the thickness of the
plate. Fixed retarder plates are either cut from crystalline materials such as quartz, calcite, or mica or
they are made of aligned polymeric material. If the retardance of the plate is λ/4, for example, the
retarder is called a quarter-wave plate.
A variable retarder can be made from a liquid crystal device. A thin layer of highly birefringent
liquid crystal material is sandwiched between two glass windows, each bearing a transparent electrode.
A voltage applied between the electrodes produces an electric field across the liquid crystal layer that
reorients the liquid crystal molecules. This reorientation changes the birefringence of the layer without
affecting its thickness or the direction of its slow axis.

Slow Axis
The slow axis describes an orientation in a birefringent material. For a given propagation direction,
light polarized parallel to the slow axis experiences the highest refractive index and hence travels the
slowest in the material (see also Fast Axis).

Wave Plate
See Retarder.

Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600 1035
 Downloaded from http://cshprotocols.cshlp.org/ at SERIALS/BIOMED LIB0175B on December 9, 2013 - Published by
Cold Spring Harbor Laboratory Press

R. Oldenbourg

ACKNOWLEDGMENTS

I gratefully acknowledge many years of illuminating discussions on polarized light microscopy with
Shinya Inoué and Michael Shribak of the MBL. This work was supported by funds from the National
Institute of Biomedical Imaging and Bioengineering (grant EB002045).

REFERENCES

Born M, Wolf E. 1980. Principles of optics: Electromagnetic theory of propa- Oldenbourg R. 1999. Polarized light microscopy of spindles. Methods Cell
gation, interference and diffraction of light. Pergamon, Elmsford, NY. Biol 61: 175–208.
Chipman RA. 1995. Polarimetry. In Handbook of optics (ed. Bass M), Oldenbourg R, Shribak M. 2010. Microscopes. In Handbook of optics (ed.
pp. 22.21–22.37. McGraw-Hill, New York. Bass M), pp. 28.1–28.62. McGraw-Hill, New York.
Hecht E. 1998. Optics. Addison-Wesley, Reading, MA. Oldenbourg R, Salmon ED, Tran PT. 1998. Birefringence of single and
Inoué S. 1986. Video microscopy. Plenum, New York. bundled microtubules. Biophys J 74: 645–654.
Inoué S. 2002. Polarization microscopy. In Current protocols in cell biology Salmon ED, Tran PT. 1998. High-resolution video-enhanced differential
(ed. Bonifacino JS, et al.), pp. 4.9.1–4.9.27. Wiley, New York. interference contrast (VE-DIC) light microscopy. Methods Cell Biol
Inoué S. 2007. Exploring living cells and molecular dynamics with polarized 56: 153–184.
light microscopy. In Optical imaging and microscopy (ed. Török P, Kao Sato H, Ellis GW, Inoué S. 1975. Microtubular origin of mitotic spindle form
FJ), pp. 3–20. Springer, Berlin. birefringence. Demonstration of the applicability of Wiener’s equation.
Inoué S. 2008. Collected works of Shinya Inoué. World Scientific, Singapore. J Cell Biol 67: 501–517.
Inoué S, Oldenbourg R. 1998. Microtubule dynamics in mitotic spindle Shribak M, Inoué S, Oldenbourg R. 2002. Polarization aberrations caused by
displayed by polarized light microscopy. Mol Biol Cell 9: 1603–1607. differential transmission and phase shift in high NA lenses: Theory,
Katoh K, Hammar K, Smith PJ, Oldenbourg R. 1999. Birefringence imaging measurement and rectification. Opt Eng 41: 943–954.
directly reveals architectural dynamics of filamentous actin in living Shurcliff WA. 1962. Polarized light, production and use. Harvard University
growth cones. Mol Biol Cell 10: 197–210. Press, Cambridge, MA.
LaFountain JR Jr, Oldenbourg R. 2004. Maloriented bivalents have meta- Vrabioiu AM, Mitchison TJ. 2006. Structural insights into yeast septin orga-
phase positions at the spindle equator with more kinetochore micro- nization from polarized fluorescence microscopy. Nature 443: 466–469.
tubules to one pole than to the other. Mol Biol Cell 15: 5346–5355. Vrabioiu AM, Mitchison TJ. 2007. Symmetry of septin hourglass and ring
Oldenbourg R. 1991. Analysis of edge birefringence. Biophys J 60: 629–641. structures. J Mol Biol 372: 37–49.

1036 Cite this introduction as Cold Spring Harb Protoc; 2013; doi:10.1101/pdb.top078600