Confocal Microscopy List

Phase contrast microscopy

Classic

List

Threaded

25 messages Options

Andrew York

Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

I have a few questions about phase contrast microscopy:

1. Can a phase-contrast microscope see a single Rayleigh-scattering
particle, for example, a 40 nm glass bead suspended in water?
http://en.wikipedia.org/wiki/Rayleigh_scattering

2. Suppose you've etched a perfectly smooth square trough (or a raised
square ridge) into your coverslip surface, perhaps one micron high and
20x20 microns wide. I'm confident the edges of the trough/ridge would show
up nicely under phase-contrast microscopy; would the middle of the
trough/ridge appear to be a different brightness than the surrounding
coverslip?

I'm primarily interested in first-hand experience rather than theoretical
predictions, so if you know the answers for similar-in-spirit but not
identical samples, please let me know.

I have followup questions, assuming the answers to 1 and 2 are
straightforward.

Shalin Mehta

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

> I have a few questions about phase contrast microscopy:
>
> 1. Can a phase-contrast microscope see a single Rayleigh-scattering
> particle, for example, a 40 nm glass bead suspended in water?
> http://en.wikipedia.org/wiki/Rayleigh_scattering

It is possible to see, but it will be small blip (positive or negative
depending on the sign of phase delay in the phase ring) on a large
background. You will need high dynamic range camera (e.g., 16-bit
sCMOS) so that the intensity modulation because of the bead spans
three bits. The shading in illumination can make things worse. It
should help to move the bead out of the view, acquire background, and
then subtract the background from the image with the bead in the view.
It may be easier to start with dark-field to locate the bead and then
switch to phase-contrast. I am assuming you are interested in optical
assessment rather than practical use.

>
> 2. Suppose you've etched a perfectly smooth square trough (or a raised
> square ridge) into your coverslip surface, perhaps one micron high and
> 20x20 microns wide. I'm confident the edges of the trough/ridge would show
> up nicely under phase-contrast microscopy; would the middle of the
> trough/ridge appear to be a different brightness than the surrounding
> coverslip?
>

Yes, the middle of the square will approach the background. This is
because of shade-off artifact
(http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/).
It is because some low spatial frequencies diffracted by the specimen
get modified by the phase-ring in objective.

> I'm primarily interested in first-hand experience rather than theoretical
> predictions, so if you know the answers for similar-in-spirit but not
> identical samples, please let me know.
>
> I have followup questions, assuming the answers to 1 and 2 are
> straightforward.

Best
Shalin

Guy Cox-2

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Well, I cannot really dispute optics with Shalin! But I can say that I've looked at resin sections under phase contrast and NOT seen this shade-off artefact. If flat and uniform they've appeared uniform in shade, and different from the background. But these would be much larger than 20 x 20 µm.

Guy

Guy Cox, Honorary Associate Professor
School of Medical Sciences

Australian Centre for Microscopy and Microanalysis,
Madsen, F09, University of Sydney, NSW 2006

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Shalin Mehta
Sent: Thursday, 20 March 2014 3:10 AM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

> I have a few questions about phase contrast microscopy:
>
> 1. Can a phase-contrast microscope see a single Rayleigh-scattering
> particle, for example, a 40 nm glass bead suspended in water?
> http://en.wikipedia.org/wiki/Rayleigh_scattering

It is possible to see, but it will be small blip (positive or negative depending on the sign of phase delay in the phase ring) on a large background. You will need high dynamic range camera (e.g., 16-bit
sCMOS) so that the intensity modulation because of the bead spans three bits. The shading in illumination can make things worse. It should help to move the bead out of the view, acquire background, and then subtract the background from the image with the bead in the view.
It may be easier to start with dark-field to locate the bead and then switch to phase-contrast. I am assuming you are interested in optical assessment rather than practical use.

>
> 2. Suppose you've etched a perfectly smooth square trough (or a raised
> square ridge) into your coverslip surface, perhaps one micron high and
> 20x20 microns wide. I'm confident the edges of the trough/ridge would
> show up nicely under phase-contrast microscopy; would the middle of
> the trough/ridge appear to be a different brightness than the
> surrounding coverslip?
>

Yes, the middle of the square will approach the background. This is because of shade-off artifact (http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/).
It is because some low spatial frequencies diffracted by the specimen get modified by the phase-ring in objective.

> I'm primarily interested in first-hand experience rather than
> theoretical predictions, so if you know the answers for
> similar-in-spirit but not identical samples, please let me know.
>
> I have followup questions, assuming the answers to 1 and 2 are
> straightforward.

Best
Shalin

Shalin Mehta

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Well, I cannot dispute experimental observation with Guy!
I should have said 'will approach the background slightly' rather
than 'will approach the background'. The intensity will be
distinguishable from background. I think Figure 1 on this Nikon page
(http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/) is
reasonable. We can observe shadeoff even with 5 um wide flat features
on our MBL/NNF test targets (made by etching patterns in 90nm thick
layer of silica).

If at all interesting, one can simulate such effects with pretty good
accuracy with a MATLAB package I have placed on googlecode
(https://code.google.com/p/microlith/). The diffraction effects (due
to imaging and *illumination*) are accounted for correctly. I am
working on including polychromatic illumination, quantization, and
camera noise.

Shalin

On Thu, Mar 20, 2014 at 2:43 AM, Guy Cox <[hidden email]> wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Well, I cannot really dispute optics with Shalin! But I can say that I've looked at resin sections under phase contrast and NOT seen this shade-off artefact. If flat and uniform they've appeared uniform in shade, and different from the background. But these would be much larger than 20 x 20 µm.
>
> Guy
>
> Guy Cox, Honorary Associate Professor
> School of Medical Sciences
>
> Australian Centre for Microscopy and Microanalysis,
> Madsen, F09, University of Sydney, NSW 2006
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Shalin Mehta
> Sent: Thursday, 20 March 2014 3:10 AM
> To: [hidden email]
> Subject: Re: Phase contrast microscopy
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
>> I have a few questions about phase contrast microscopy:
>>
>> 1. Can a phase-contrast microscope see a single Rayleigh-scattering
>> particle, for example, a 40 nm glass bead suspended in water?
>> http://en.wikipedia.org/wiki/Rayleigh_scattering
>
> It is possible to see, but it will be small blip (positive or negative depending on the sign of phase delay in the phase ring) on a large background. You will need high dynamic range camera (e.g., 16-bit
> sCMOS) so that the intensity modulation because of the bead spans three bits. The shading in illumination can make things worse. It should help to move the bead out of the view, acquire background, and then subtract the background from the image with the bead in the view.
> It may be easier to start with dark-field to locate the bead and then switch to phase-contrast. I am assuming you are interested in optical assessment rather than practical use.
>
>>
>> 2. Suppose you've etched a perfectly smooth square trough (or a raised
>> square ridge) into your coverslip surface, perhaps one micron high and
>> 20x20 microns wide. I'm confident the edges of the trough/ridge would
>> show up nicely under phase-contrast microscopy; would the middle of
>> the trough/ridge appear to be a different brightness than the
>> surrounding coverslip?
>>
>
> Yes, the middle of the square will approach the background. This is because of shade-off artifact (http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/).
> It is because some low spatial frequencies diffracted by the specimen get modified by the phase-ring in objective.
>
>> I'm primarily interested in first-hand experience rather than
>> theoretical predictions, so if you know the answers for
>> similar-in-spirit but not identical samples, please let me know.
>>
>> I have followup questions, assuming the answers to 1 and 2 are
>> straightforward.
>
> Best
> Shalin

Andrew York

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Ok, that lines up with my expectations. Thanks, Shalin! Now, on to my
underlying question:

Is the information in a phase contrast microscope image more like a
dark-field image (sensitive to deflection of light), or more like an
interferometric image (sensitive to optical path length)?

Of course, deflection of light and changes in optical path length are
often related. I asked about a subdiffractive bead because it seemed like a
good sample to produce high-angle deflection of light with minimal change
in optical path length. I asked about a wide, uniform-thickness slab
because it seemed like a good sample to produce a huge change in optical
path length with hardly any deflection (except at the edges).

As far as I can tell, a phase-contrast image shows the portions of a
transparent sample which deflect light (much like a dark-field image). A
phase-contrast image is bright where there is no deflection (instead of
dark), and darker where there *is* deflection (instead of brighter), but in
both types of image, a signal which differs from background implies light
which changes direction because of the sample. (Of course, if the sample is
not perfectly transparent, phase contrast will also dim due to absorption,
but that's not the interesting part. I wonder if absorption is what Guy
observed in resin sections?)

However, many textbooks and websites imply that a phase contrast
microscope shows variations in optical path length, not simply regions of
the sample which deflect light. Shalin's link nicely illustrates this:
http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/

> It would normally be expected that the image of a large phase specimen
> having a constant optical path length across the diameter would appear
> uniformly dark or light in the microscope. Unfortunately, the intensity of
> images produced by a phase contrast microscope does not always bear a
> simple linear relationship to the optical path difference produced by the
> specimen.

It seems to me that phase contrast image intensity is *completely
unrelated* to optical path difference produced by the specimen.
Interferometric imaging (like this: http://goo.gl/UuE5Bz ), depends on
optical path differences directly (assuming the system is aligned with
perfect constructive interference at output 1 and perfect destructive
interference at output 2 before the sample goes in). Phase contrast imaging
(like this: http://goo.gl/NDtckE ), seems to depend only on the sample
changing the direction of the light; the sample-induced 1/4-wave phase
shift is purely an effect of dipole scattering, and has nothing to do with
optical path differences, right?

So, am I wrong? Can a phase contrast microscope see a sample which purely
changes optical path length without deflecting light? If so, how? If not,
why do so many sources imply that it should?

Other examples of phase-contrast explanations that seem (to me) to imply
that phase contrast microscopy is sensitive to optical path length, rather
than simply deflection:

From Hecht's Optics, page 619, 4th edition:
"...phase objects that are thicker or have higher indices appear dark
against a bright background."

From Nikon MicroscopyU:
http://www.microscopyu.com/articles/phasecontrast/phasemicroscopy.html
"In terms of optical path variations between the specimen and its
surrounding medium, the portion of the incident light wavefront that
traverses the specimen (D-wave), but does not pass through the surrounding
medium (S-wave), is slightly retarded...
...For individual cells in tissue culture, the optical path difference is
relatively small ... about a quarter wavelength (of green light)... optical
path differences produce a linear reduction in intensity with increasing
phase shift (the image grows progressively darker) up to a point ... after
which, the specimen image becomes brighter through reversal of contrast."

From NobelPrize.org:
http://www.nobelprize.org/educational/physics/microscopes/phase/
"The phase contrast microscope uses the fact that the light passing trough
a transparent part of the specimen travels slower and, due to this is
shifted compared to the uninfluenced light."

From UC Berkeley:
http://microscopy.berkeley.edu/Resources/instruction/phase_contrast.html
"As light rays pass through areas within the tissue of different optical
path (refractive index and geometric path length) they may be retarded in
phase by up to 1⁄4 wavelength but will remain unchanged in amplitude."

From UCLA:
http://www.gonda.ucla.edu/bri_core/phase.htm
"When a light wave passes through an object, it is deviated, that is, it
becomes phase retarded or phase shifted."

On Thu, Mar 20, 2014 at 11:16 AM, Shalin Mehta <[hidden email]>wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Well, I cannot dispute experimental observation with Guy!
> I should have said 'will approach the background slightly' rather
> than 'will approach the background'. The intensity will be
> distinguishable from background. I think Figure 1 on this Nikon page
> (http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/) is
> reasonable. We can observe shadeoff even with 5 um wide flat features
> on our MBL/NNF test targets (made by etching patterns in 90nm thick
> layer of silica).
>
> If at all interesting, one can simulate such effects with pretty good
> accuracy with a MATLAB package I have placed on googlecode
> (https://code.google.com/p/microlith/). The diffraction effects (due
> to imaging and *illumination*) are accounted for correctly. I am
> working on including polychromatic illumination, quantization, and
> camera noise.
>
> Shalin
>
> On Thu, Mar 20, 2014 at 2:43 AM, Guy Cox <[hidden email]> wrote:
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > *****
> >
> > Well, I cannot really dispute optics with Shalin! But I can say that
> I've looked at resin sections under phase contrast and NOT seen this
> shade-off artefact. If flat and uniform they've appeared uniform in shade,
> and different from the background. But these would be much larger than 20
> x 20 µm.
> >
> > Guy
> >
> > Guy Cox, Honorary Associate Professor
> > School of Medical Sciences
> >
> > Australian Centre for Microscopy and Microanalysis,
> > Madsen, F09, University of Sydney, NSW 2006
> >
> > -----Original Message-----
> > From: Confocal Microscopy List [mailto:[hidden email]]
> On Behalf Of Shalin Mehta
> > Sent: Thursday, 20 March 2014 3:10 AM
> > To: [hidden email]
> > Subject: Re: Phase contrast microscopy
> >
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > *****
> >
> >> I have a few questions about phase contrast microscopy:
> >>
> >> 1. Can a phase-contrast microscope see a single Rayleigh-scattering
> >> particle, for example, a 40 nm glass bead suspended in water?
> >> http://en.wikipedia.org/wiki/Rayleigh_scattering
> >
> > It is possible to see, but it will be small blip (positive or negative
> depending on the sign of phase delay in the phase ring) on a large
> background. You will need high dynamic range camera (e.g., 16-bit
> > sCMOS) so that the intensity modulation because of the bead spans three
> bits. The shading in illumination can make things worse. It should help to
> move the bead out of the view, acquire background, and then subtract the
> background from the image with the bead in the view.
> > It may be easier to start with dark-field to locate the bead and then
> switch to phase-contrast. I am assuming you are interested in optical
> assessment rather than practical use.
> >
> >>
> >> 2. Suppose you've etched a perfectly smooth square trough (or a raised
> >> square ridge) into your coverslip surface, perhaps one micron high and
> >> 20x20 microns wide. I'm confident the edges of the trough/ridge would
> >> show up nicely under phase-contrast microscopy; would the middle of
> >> the trough/ridge appear to be a different brightness than the
> >> surrounding coverslip?
> >>
> >
> > Yes, the middle of the square will approach the background. This is
> because of shade-off artifact (
> http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/).
> > It is because some low spatial frequencies diffracted by the specimen
> get modified by the phase-ring in objective.
> >
> >> I'm primarily interested in first-hand experience rather than
> >> theoretical predictions, so if you know the answers for
> >> similar-in-spirit but not identical samples, please let me know.
> >>
> >> I have followup questions, assuming the answers to 1 and 2 are
> >> straightforward.
> >
> > Best
> > Shalin
>

Craig Brideau

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

I played around with DODT microscopy a while back, and that technique
mainly depends on ray deflection. You create a gradient light source and
cast it onto the sample. As it passes through, it deflects some of the
rays. This disrupts the gradient, enhancing edges and giving you an image
a bit like phase contrast but there are no waveplates or other
phase-controlling elements involved.

On Thu, Mar 20, 2014 at 10:13 AM, Andrew York <
[hidden email]> wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Ok, that lines up with my expectations. Thanks, Shalin! Now, on to my
> underlying question:
>
> Is the information in a phase contrast microscope image more like a
> dark-field image (sensitive to deflection of light), or more like an
> interferometric image (sensitive to optical path length)?
>
> Of course, deflection of light and changes in optical path length are
> often related. I asked about a subdiffractive bead because it seemed like a
> good sample to produce high-angle deflection of light with minimal change
> in optical path length. I asked about a wide, uniform-thickness slab
> because it seemed like a good sample to produce a huge change in optical
> path length with hardly any deflection (except at the edges).
>
> As far as I can tell, a phase-contrast image shows the portions of a
> transparent sample which deflect light (much like a dark-field image). A
> phase-contrast image is bright where there is no deflection (instead of
> dark), and darker where there *is* deflection (instead of brighter), but in
> both types of image, a signal which differs from background implies light
> which changes direction because of the sample. (Of course, if the sample is
> not perfectly transparent, phase contrast will also dim due to absorption,
> but that's not the interesting part. I wonder if absorption is what Guy
> observed in resin sections?)
>
> However, many textbooks and websites imply that a phase contrast
> microscope shows variations in optical path length, not simply regions of
> the sample which deflect light. Shalin's link nicely illustrates this:
> http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/
>
> > It would normally be expected that the image of a large phase specimen
> > having a constant optical path length across the diameter would appear
> > uniformly dark or light in the microscope. Unfortunately, the intensity
> of
> > images produced by a phase contrast microscope does not always bear a
> > simple linear relationship to the optical path difference produced by the
> > specimen.
>
>
> It seems to me that phase contrast image intensity is *completely
> unrelated* to optical path difference produced by the specimen.
> Interferometric imaging (like this: http://goo.gl/UuE5Bz ), depends on
> optical path differences directly (assuming the system is aligned with
> perfect constructive interference at output 1 and perfect destructive
> interference at output 2 before the sample goes in). Phase contrast imaging
> (like this: http://goo.gl/NDtckE ), seems to depend only on the sample
> changing the direction of the light; the sample-induced 1/4-wave phase
> shift is purely an effect of dipole scattering, and has nothing to do with
> optical path differences, right?
>
> So, am I wrong? Can a phase contrast microscope see a sample which purely
> changes optical path length without deflecting light? If so, how? If not,
> why do so many sources imply that it should?
>
> Other examples of phase-contrast explanations that seem (to me) to imply
> that phase contrast microscopy is sensitive to optical path length, rather
> than simply deflection:
>
> From Hecht's Optics, page 619, 4th edition:
> "...phase objects that are thicker or have higher indices appear dark
> against a bright background."
>
> From Nikon MicroscopyU:
> http://www.microscopyu.com/articles/phasecontrast/phasemicroscopy.html
> "In terms of optical path variations between the specimen and its
> surrounding medium, the portion of the incident light wavefront that
> traverses the specimen (D-wave), but does not pass through the surrounding
> medium (S-wave), is slightly retarded...
> ...For individual cells in tissue culture, the optical path difference is
> relatively small ... about a quarter wavelength (of green light)... optical
> path differences produce a linear reduction in intensity with increasing
> phase shift (the image grows progressively darker) up to a point ... after
> which, the specimen image becomes brighter through reversal of contrast."
>
> From NobelPrize.org:
> http://www.nobelprize.org/educational/physics/microscopes/phase/
> "The phase contrast microscope uses the fact that the light passing trough
> a transparent part of the specimen travels slower and, due to this is
> shifted compared to the uninfluenced light."
>
> From UC Berkeley:
> http://microscopy.berkeley.edu/Resources/instruction/phase_contrast.html
> "As light rays pass through areas within the tissue of different optical
> path (refractive index and geometric path length) they may be retarded in
> phase by up to 1⁄4 wavelength but will remain unchanged in amplitude."
>
> From UCLA:
> http://www.gonda.ucla.edu/bri_core/phase.htm
> "When a light wave passes through an object, it is deviated, that is, it
> becomes phase retarded or phase shifted."
>
>
> On Thu, Mar 20, 2014 at 11:16 AM, Shalin Mehta <[hidden email]
> >wrote:
>
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > *****
> >
> > Well, I cannot dispute experimental observation with Guy!
> > I should have said 'will approach the background slightly' rather
> > than 'will approach the background'. The intensity will be
> > distinguishable from background. I think Figure 1 on this Nikon page
> > (http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/) is
> > reasonable. We can observe shadeoff even with 5 um wide flat features
> > on our MBL/NNF test targets (made by etching patterns in 90nm thick
> > layer of silica).
> >
> > If at all interesting, one can simulate such effects with pretty good
> > accuracy with a MATLAB package I have placed on googlecode
> > (https://code.google.com/p/microlith/). The diffraction effects (due
> > to imaging and *illumination*) are accounted for correctly. I am
> > working on including polychromatic illumination, quantization, and
> > camera noise.
> >
> > Shalin
> >
> > On Thu, Mar 20, 2014 at 2:43 AM, Guy Cox <[hidden email]> wrote:
> > > *****
> > > To join, leave or search the confocal microscopy listserv, go to:
> > > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > > *****
> > >
> > > Well, I cannot really dispute optics with Shalin! But I can say that
> > I've looked at resin sections under phase contrast and NOT seen this
> > shade-off artefact. If flat and uniform they've appeared uniform in
> shade,
> > and different from the background. But these would be much larger than
> 20
> > x 20 µm.
> > >
> > > Guy
> > >
> > > Guy Cox, Honorary Associate Professor
> > > School of Medical Sciences
> > >
> > > Australian Centre for Microscopy and Microanalysis,
> > > Madsen, F09, University of Sydney, NSW 2006
> > >
> > > -----Original Message-----
> > > From: Confocal Microscopy List [mailto:
> [hidden email]]
> > On Behalf Of Shalin Mehta
> > > Sent: Thursday, 20 March 2014 3:10 AM
> > > To: [hidden email]
> > > Subject: Re: Phase contrast microscopy
> > >
> > > *****
> > > To join, leave or search the confocal microscopy listserv, go to:
> > > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > > *****
> > >
> > >> I have a few questions about phase contrast microscopy:
> > >>
> > >> 1. Can a phase-contrast microscope see a single Rayleigh-scattering
> > >> particle, for example, a 40 nm glass bead suspended in water?
> > >> http://en.wikipedia.org/wiki/Rayleigh_scattering
> > >
> > > It is possible to see, but it will be small blip (positive or negative
> > depending on the sign of phase delay in the phase ring) on a large
> > background. You will need high dynamic range camera (e.g., 16-bit
> > > sCMOS) so that the intensity modulation because of the bead spans three
> > bits. The shading in illumination can make things worse. It should help
> to
> > move the bead out of the view, acquire background, and then subtract the
> > background from the image with the bead in the view.
> > > It may be easier to start with dark-field to locate the bead and then
> > switch to phase-contrast. I am assuming you are interested in optical
> > assessment rather than practical use.
> > >
> > >>
> > >> 2. Suppose you've etched a perfectly smooth square trough (or a raised
> > >> square ridge) into your coverslip surface, perhaps one micron high and
> > >> 20x20 microns wide. I'm confident the edges of the trough/ridge would
> > >> show up nicely under phase-contrast microscopy; would the middle of
> > >> the trough/ridge appear to be a different brightness than the
> > >> surrounding coverslip?
> > >>
> > >
> > > Yes, the middle of the square will approach the background. This is
> > because of shade-off artifact (
> > http://www.microscopyu.com/tutorials/java/phasecontrast/shadeoff/).
> > > It is because some low spatial frequencies diffracted by the specimen
> > get modified by the phase-ring in objective.
> > >
> > >> I'm primarily interested in first-hand experience rather than
> > >> theoretical predictions, so if you know the answers for
> > >> similar-in-spirit but not identical samples, please let me know.
> > >>
> > >> I have followup questions, assuming the answers to 1 and 2 are
> > >> straightforward.
> > >
> > > Best
> > > Shalin
> >
>

Shalin Mehta

Re: Phase contrast microscopy

In reply to this post by Andrew York

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

On Thu, Mar 20, 2014 at 12:13 PM, Andrew York
<[hidden email]> wrote:

> It seems to me that phase contrast image intensity is *completely
> unrelated* to optical path difference produced by the specimen.
> Interferometric imaging (like this: http://goo.gl/UuE5Bz ), depends on
> optical path differences directly (assuming the system is aligned with
> perfect constructive interference at output 1 and perfect destructive
> interference at output 2 before the sample goes in). Phase contrast imaging
> (like this: http://goo.gl/NDtckE ), seems to depend only on the sample
> changing the direction of the light; the sample-induced 1/4-wave phase
> shift is purely an effect of dipole scattering, and has nothing to do with
> optical path differences, right?

I almost agree with you Andy on your first statement above, but I will
soften it by saying phase contrast image intensity is 'only partially
related' to optical thickness of the specimen along the optical axis
of the microscope.

You are absolutely correct that the 'phase' in phase-contrast is the
1/4-wave phase shift *due to dielectric dipole scattering*. But, the
optical path length does affect the amplitude and phase of the
scattered light on top of the 1/4-wave phase shift due to dielectric
interaction of light and material.

If we read the phasor diagrams (Fig. 1, right, MD') in Zernike's
Nobel lecture, he suggests that there is pi/2 phase shift between
direct and scattered light and THEN SOME. As is correctly documented
at most places, the phase-ring either advances or retards the direct
light by pi/2 so that its superposition with scattered light produces
measurable intensity variation when they come together in the image
plane.

My approach to understanding this is to ask: if I have a pure phase
grating (say a saw tooth or sinusoidal variation in optical
thickness), at what angles and with what amplitudes does the scattered
light (in this case diffraction orders) propagate?

The answer can be obtained by taking 1D Fourier transform of periodic
complex transmission [Exp(i*2pi/lambda*refractive index*
thicknessprofile)].

The spatial frequencies in the spectrum are the sine of the angle at
which diffraction orders leave the specimen, and also the location of
the diffraction orders in the back focal plane of the objective. The
amplitudes of the diffraction orders does change as you increase the
thickness of the grating. If we start with a very small bead or an
extended feature change its size, its diffraction spectrum will change
a little-bit with change in size.

So the optical path length affects the amplitude of the scattered
light (on top of the pi/4 shift due to scattering), which in turn
produces intensity modulation.

I tried to construct a more intuitive (non-Fourier) diffraction model
of phase-contrast using Huygen's wavelets, but didn't get very far.

Shalin

Andrew York

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

It sounds like we agree on my primary point: *Many high-quality,
authoritative sources imply phase-contrast imaging measures path length,
which is incorrect and misleading*.

Statements like this:

It would normally be expected that the image of a large phase specimen
> having a constant optical path length across the diameter would appear
> uniformly dark or light in the [phase contrast] microscope.

...or this:

...optical path differences produce a linear reduction in intensity with
> increasing phase shift (the image grows progressively darker) up to a point
>

...only make sense if you think phase contrast microscopy measures path
length (like interferometric imaging) rather than measuring deflection. If
you believe this, you might use phase contrast imaging to measure which of
two thin films is thicker, and you would get nonsense results.

* Why is this misconception so widespread? What can we do to correct it?*

****
Minor details following up on Shalin's points:
****
I agree with your refinement to my description: if deflected light is
re-deflected, or the different deflection directions encounter
substantially different optical path lengths on the way to the detector,
this will alter a phase contrast image. I think we agree that this doesn't
apply to either of the simple examples I initially asked about:

1. A single, isolated, pointlike bead (very little path length delta, so
almost invisible in interferometric imaging; high-angle scattering, so
visible in phase contrast imaging)

2. A wide, ~micron thick, uniform slab of glass (huge path length delta, so
visible in interferometric imaging; no scattering except at the edges so
almost invisible to phase contrast imaging)

I think we also agree that if light is re-deflected, or different
deflection directions encounter substantially different optical path
lengths, this will alter the image in any form of microscopy that depends
on emission, and we usually neglect these corrections. For example, if the
fluorescent light from a single Alexa 488 encounters optical path lengths
which vary substantially for different emission directions on its way to a
CCD, the image of that Alexa molecule will be distorted, and we can't model
the imaging system as having a spatially invariant PSF.

On Thu, Mar 20, 2014 at 7:53 PM, Shalin Mehta <[hidden email]>wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> On Thu, Mar 20, 2014 at 12:13 PM, Andrew York
> <[hidden email]> wrote:
> > It seems to me that phase contrast image intensity is *completely
> > unrelated* to optical path difference produced by the specimen.
> > Interferometric imaging (like this: http://goo.gl/UuE5Bz ), depends on
> > optical path differences directly (assuming the system is aligned with
> > perfect constructive interference at output 1 and perfect destructive
> > interference at output 2 before the sample goes in). Phase contrast
> imaging
> > (like this: http://goo.gl/NDtckE ), seems to depend only on the sample
> > changing the direction of the light; the sample-induced 1/4-wave phase
> > shift is purely an effect of dipole scattering, and has nothing to do
> with
> > optical path differences, right?
>
> I almost agree with you Andy on your first statement above, but I will
> soften it by saying phase contrast image intensity is 'only partially
> related' to optical thickness of the specimen along the optical axis
> of the microscope.
>
> You are absolutely correct that the 'phase' in phase-contrast is the
> 1/4-wave phase shift *due to dielectric dipole scattering*. But, the
> optical path length does affect the amplitude and phase of the
> scattered light on top of the 1/4-wave phase shift due to dielectric
> interaction of light and material.
>
> If we read the phasor diagrams (Fig. 1, right, MD') in Zernike's
> Nobel lecture, he suggests that there is pi/2 phase shift between
> direct and scattered light and THEN SOME. As is correctly documented
> at most places, the phase-ring either advances or retards the direct
> light by pi/2 so that its superposition with scattered light produces
> measurable intensity variation when they come together in the image
> plane.
>
> My approach to understanding this is to ask: if I have a pure phase
> grating (say a saw tooth or sinusoidal variation in optical
> thickness), at what angles and with what amplitudes does the scattered
> light (in this case diffraction orders) propagate?
>
> The answer can be obtained by taking 1D Fourier transform of periodic
> complex transmission [Exp(i*2pi/lambda*refractive index*
> thicknessprofile)].
>
> The spatial frequencies in the spectrum are the sine of the angle at
> which diffraction orders leave the specimen, and also the location of
> the diffraction orders in the back focal plane of the objective. The
> amplitudes of the diffraction orders does change as you increase the
> thickness of the grating. If we start with a very small bead or an
> extended feature change its size, its diffraction spectrum will change
> a little-bit with change in size.
>
> So the optical path length affects the amplitude of the scattered
> light (on top of the pi/4 shift due to scattering), which in turn
> produces intensity modulation.
>
> I tried to construct a more intuitive (non-Fourier) diffraction model
> of phase-contrast using Huygen's wavelets, but didn't get very far.
>
> Shalin
>

Zdenek Svindrych

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Andrew,

the distinction between phase shift and deflection is not quite clear (at
least to me). You can clearly get phase shift without deflection and
probably the other way round, mut most often you get both (even in your
simple model cases). But most often you get both: an irregular refractive
index distribution leads to phase shifts in near field and deflection in far
field.
And, indeed, phase contrast relies both on phase shift and on scattering
(deflection). I think the description given on Wikipedia is quite
informative.

I agree that one definitely does not measure path length by phase contrast
imaging. You can only measure path length difference (either locally,
interpreting DIC or phase contrast images; or globally, with some sort of
interferometric design). You can measure the thickness difference of two
films (preferably reasonably thin, with no gap between them) by just
observing the transition across the edge where they meet...

The matter is by no means trivial and the models predicting the image of a
given structure are complicated (some links were given earlier in this
thread). After all, Zernike got a Nobel Prize for this... (well... alright,
there was also Nobel Prize for photo-electric effect, but that's a different
story :-).

A final note to your Alaxa 488 molecule: angle-dependent phase shift may
still be spatially invariant if it happens 'far away from the sample plane'
(that means 'near the aperture plane', like all tradidional objective
abberations). Then you can define (crippled) PSF.

Best,
zdenek svindrych

---------- Původní zpráva ----------
Od: Andrew York <[hidden email]>
Komu: [hidden email]
Datum: 24. 3. 2014 22:02:10
Předmět: Re: Phase contrast microscopy

"*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

It sounds like we agree on my primary point: *Many high-quality,
authoritative sources imply phase-contrast imaging measures path length,
which is incorrect and misleading*.

Statements like this:

It would normally be expected that the image of a large phase specimen
> having a constant optical path length across the diameter would appear
> uniformly dark or light in the [phase contrast] microscope.

...or this:

...optical path differences produce a linear reduction in intensity with
> increasing phase shift (the image grows progressively darker) up to a
point
>

...only make sense if you think phase contrast microscopy measures path
length (like interferometric imaging) rather than measuring deflection. If
you believe this, you might use phase contrast imaging to measure which of
two thin films is thicker, and you would get nonsense results.

* Why is this misconception so widespread? What can we do to correct it?*

****
Minor details following up on Shalin's points:
****
I agree with your refinement to my description: if deflected light is
re-deflected, or the different deflection directions encounter
substantially different optical path lengths on the way to the detector,
this will alter a phase contrast image. I think we agree that this doesn't
apply to either of the simple examples I initially asked about:

1. A single, isolated, pointlike bead (very little path length delta, so
almost invisible in interferometric imaging; high-angle scattering, so
visible in phase contrast imaging)

2. A wide, ~micron thick, uniform slab of glass (huge path length delta, so
visible in interferometric imaging; no scattering except at the edges so
almost invisible to phase contrast imaging)

I think we also agree that if light is re-deflected, or different
deflection directions encounter substantially different optical path
lengths, this will alter the image in any form of microscopy that depends
on emission, and we usually neglect these corrections. For example, if the
fluorescent light from a single Alexa 488 encounters optical path lengths
which vary substantially for different emission directions on its way to a
CCD, the image of that Alexa molecule will be distorted, and we can't model
the imaging system as having a spatially invariant PSF.

On Thu, Mar 20, 2014 at 7:53 PM, Shalin Mehta <[hidden email]>wrote:

Tobias Baskin

Re: Phase contrast microscopy

In reply to this post by Andrew York

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Andrew and company,
This is fascinating. I certainly
learned in school that phase contrast generates contrast in the image by
means of the difference in optical path between the sample and the
background, but unlike interference microscopy where the difference is
quantitative, in phase contrast the difference is only qualitative (that
is, not measurable). Because I was taught so, it is a little unsettling
to read that my brilliant profs were wrong. I would *love* to see an
image of a piece of mica (or other opitcal flat thing) in phase contrast
where the background and the mica have same intensity (edges not
included). Not that I don't trust the theory but ...

Thanks,
Tobias

On 3/24/14, 4:42 PM, Andrew York wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> It sounds like we agree on my primary point: *Many high-quality,
> authoritative sources imply phase-contrast imaging measures path length,
> which is incorrect and misleading*.
>
> Statements like this:
>
> It would normally be expected that the image of a large phase specimen
>> having a constant optical path length across the diameter would appear
>> uniformly dark or light in the [phase contrast] microscope.
>
> ...or this:
>
> ...optical path differences produce a linear reduction in intensity with
>> increasing phase shift (the image grows progressively darker) up to a point
>>
>
> ...only make sense if you think phase contrast microscopy measures path
> length (like interferometric imaging) rather than measuring deflection. If
> you believe this, you might use phase contrast imaging to measure which of
> two thin films is thicker, and you would get nonsense results.
>
> * Why is this misconception so widespread? What can we do to correct it?*
>
>
> ****
> Minor details following up on Shalin's points:
> ****
> I agree with your refinement to my description: if deflected light is
> re-deflected, or the different deflection directions encounter
> substantially different optical path lengths on the way to the detector,
> this will alter a phase contrast image. I think we agree that this doesn't
> apply to either of the simple examples I initially asked about:
>
> 1. A single, isolated, pointlike bead (very little path length delta, so
> almost invisible in interferometric imaging; high-angle scattering, so
> visible in phase contrast imaging)
>
> 2. A wide, ~micron thick, uniform slab of glass (huge path length delta, so
> visible in interferometric imaging; no scattering except at the edges so
> almost invisible to phase contrast imaging)
>
> I think we also agree that if light is re-deflected, or different
> deflection directions encounter substantially different optical path
> lengths, this will alter the image in any form of microscopy that depends
> on emission, and we usually neglect these corrections. For example, if the
> fluorescent light from a single Alexa 488 encounters optical path lengths
> which vary substantially for different emission directions on its way to a
> CCD, the image of that Alexa molecule will be distorted, and we can't model
> the imaging system as having a spatially invariant PSF.
>
>
> On Thu, Mar 20, 2014 at 7:53 PM, Shalin Mehta <[hidden email]>wrote:
>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> *****
>>
>> On Thu, Mar 20, 2014 at 12:13 PM, Andrew York
>> <[hidden email]> wrote:
>>> It seems to me that phase contrast image intensity is *completely
>>> unrelated* to optical path difference produced by the specimen.
>>> Interferometric imaging (like this: http://goo.gl/UuE5Bz ), depends on
>>> optical path differences directly (assuming the system is aligned with
>>> perfect constructive interference at output 1 and perfect destructive
>>> interference at output 2 before the sample goes in). Phase contrast
>> imaging
>>> (like this: http://goo.gl/NDtckE ), seems to depend only on the sample
>>> changing the direction of the light; the sample-induced 1/4-wave phase
>>> shift is purely an effect of dipole scattering, and has nothing to do
>> with
>>> optical path differences, right?
>> I almost agree with you Andy on your first statement above, but I will
>> soften it by saying phase contrast image intensity is 'only partially
>> related' to optical thickness of the specimen along the optical axis
>> of the microscope.
>>
>> You are absolutely correct that the 'phase' in phase-contrast is the
>> 1/4-wave phase shift *due to dielectric dipole scattering*. But, the
>> optical path length does affect the amplitude and phase of the
>> scattered light on top of the 1/4-wave phase shift due to dielectric
>> interaction of light and material.
>>
>> If we read the phasor diagrams (Fig. 1, right, MD') in Zernike's
>> Nobel lecture, he suggests that there is pi/2 phase shift between
>> direct and scattered light and THEN SOME. As is correctly documented
>> at most places, the phase-ring either advances or retards the direct
>> light by pi/2 so that its superposition with scattered light produces
>> measurable intensity variation when they come together in the image
>> plane.
>>
>> My approach to understanding this is to ask: if I have a pure phase
>> grating (say a saw tooth or sinusoidal variation in optical
>> thickness), at what angles and with what amplitudes does the scattered
>> light (in this case diffraction orders) propagate?
>>
>> The answer can be obtained by taking 1D Fourier transform of periodic
>> complex transmission [Exp(i*2pi/lambda*refractive index*
>> thicknessprofile)].
>>
>> The spatial frequencies in the spectrum are the sine of the angle at
>> which diffraction orders leave the specimen, and also the location of
>> the diffraction orders in the back focal plane of the objective. The
>> amplitudes of the diffraction orders does change as you increase the
>> thickness of the grating. If we start with a very small bead or an
>> extended feature change its size, its diffraction spectrum will change
>> a little-bit with change in size.
>>
>> So the optical path length affects the amplitude of the scattered
>> light (on top of the pi/4 shift due to scattering), which in turn
>> produces intensity modulation.
>>
>> I tried to construct a more intuitive (non-Fourier) diffraction model
>> of phase-contrast using Huygen's wavelets, but didn't get very far.
>>
>> Shalin
>>

--

__ ___ ___ ___ Tobias I. Baskin
/ \ / /\ / \ Professor
/ / / / \ \ \ Biology Department
/ __/ /__ /___ \ \ \__ 611 N. Pleasant St.
/ / / \ \ \ University of Massachusetts
/ / / \ \ \ Amherst, MA, 01003 USA
/ /___ / \ \___/ \_____ 413 - 545 - 1533
www.bio.umass.edu/biology/baskin

ssss

James Pawley

Re: Phase contrast microscopy

In reply to this post by Zdenek Svindrych

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi all,

Just to muddy the waters a little more, or perhaps, to add a practical point:

We need to remember that most objects delineated
by a change in RI in the XY plane also have
changes in RI in the z plane (i.e., these XY RI
patterns seldom extend without change very far in
the z direction).

Imagine a pattern etched in a 1µm thick layer of
RI 1.4 plastic on the surface of RI 1.5 glass
slide. If we look at this through immersion oil,
light passing through the top and bottom surfaces
of the plastic bits will be reflected back
towards the light source (reflection is
proportional to the square of the difference in
RI). Light passing through only glass and oil
will not be so reflected

At high NA (large phase rings), this effect is
enhanced because the conical illumination then
approaches these horizontal surfaces at a fairly
oblique angle, a second factor that increases
reflection).

Consequently, regardless of what happens with the
phase and intensity shifts induced by the phase
ring, plastic areas will look darker than
"non-plastic" (oil-filled?) areas of the specimen.

In the real world, things are seldom as simple as the math.

JP

>
>Hi Andrew, the distinction between phase shift
>and deflection is not quite clear (at least to
>me). You can clearly get phase shift without
>deflection and probably the other way round, mut
>most often you get both (even in your simple
>model cases). But most often you get both: an
>irregular refractive index distribution leads to
>phase shifts in near field and deflection in far
>field. And, indeed, phase contrast relies both
>on phase shift and on scattering (deflection). I
>think the description given on Wikipedia is
>quite informative. I agree that one definitely
>does not measure path length by phase contrast
>imaging. You can only measure path length
>difference (either locally, interpreting DIC or
>phase contrast images; or globally, with some
>sort of interferometric design). You can measure
>the thickness difference of two films
>(preferably reasonably thin, with no gap between
>them) by just observing the transition across
>the edge where they meet... The matter is by no
>means trivial and the models predicting the
>image of a given structure are complicated (some
>links were given earlier in this thread). After
>all, Zernike got a Nobel Prize for this...
>(well... alright, there was also Nobel Prize for
>photo-electric effect, but that's a different
>story :-). A final note to your Alaxa 488
>molecule: angle-dependent phase shift may still
>be spatially invariant if it happens 'far away
>from the sample plane' (that means 'near the
>aperture plane', like all tradidional objective
>abberations). Then you can define (crippled)
>PSF. Best, zdenek svindrych ---------- PÛvodní
>zpráva ---------- Od: Andrew York
><[hidden email]> Komu:
>[hidden email] Datum: 24. 3.
>2014 22:02:10 PÞedmût: Re: Phase contrast
>microscopy "***** To join, leave or search the
>confocal microscopy listserv, go to:
>http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>***** It sounds like we agree on my primary
>point: *Many high-quality, authoritative sources
>imply phase-contrast imaging measures path
>length, which is incorrect and misleading*.
>Statements like this: It would normally be
>expected that the image of a large phase
>specimen > having a constant optical path length
>across the diameter would appear > uniformly
>dark or light in the [phase contrast]
>microscope. ...or this: ...optical path
>differences produce a linear reduction in
>intensity with > increasing phase shift (the
>image grows progressively darker) up to a
>point > ...only make sense if you think phase
>contrast microscopy measures path length (like
>interferometric imaging) rather than measuring
>deflection. If you believe this, you might use
>phase contrast imaging to measure which of two
>thin films is thicker, and you would get
>nonsense results. * Why is this misconception so
>widespread? What can we do to correct it?* ****
>Minor details following up on Shalin's points:
>**** I agree with your refinement to my
>description: if deflected light is re-deflected,
>or the different deflection directions encounter
>substantially different optical path lengths on
>the way to the detector, this will alter a phase
>contrast image. I think we agree that this
>doesn't apply to either of the simple examples I
>initially asked about: 1. A single, isolated,
>pointlike bead (very little path length delta,
>so almost invisible in interferometric imaging;
>high-angle scattering, so visible in phase
>contrast imaging) 2. A wide, ~micron thick,
>uniform slab of glass (huge path length delta,
>so visible in interferometric imaging; no
>scattering except at the edges so almost
>invisible to phase contrast imaging) I think we
>also agree that if light is re-deflected, or
>different deflection directions encounter
>substantially different optical path lengths,
>this will alter the image in any form of
>microscopy that depends on emission, and we
>usually neglect these corrections. For example,
>if the fluorescent light from a single Alexa 488
>encounters optical path lengths which vary
>substantially for different emission directions
>on its way to a CCD, the image of that Alexa
>molecule will be distorted, and we can't model
>the imaging system as having a spatially
>invariant PSF. On Thu, Mar 20, 2014 at 7:53 PM,
>Shalin Mehta <[hidden email]>wrote: >
>***** > To join, leave or search the confocal
>microscopy listserv, go to: >
>http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy >
>***** > > On Thu, Mar 20, 2014 at 12:13 PM,
>Andrew York >
><[hidden email]>
>wrote: > > It seems to me that phase contrast
>image intensity is *completely > > unrelated* to
>optical path difference produced by the
>specimen. > > Interferometric imaging (like
>this: http://goo.gl/UuE5Bz ), depends on > >
>optical path differences directly (assuming the
>system is aligned with > > perfect constructive
>interference at output 1 and perfect
>destructive > > interference at output 2 before
>the sample goes in). Phase contrast >
>imaging > > (like this: http://goo.gl/NDtckE ),
>seems to depend only on the sample > > changing
>the direction of the light; the sample-induced
>1/4-wave phase > > shift is purely an effect of
>dipole scattering, and has nothing to do >
>with > > optical path differences, right? > > I
>almost agree with you Andy on your first
>statement above, but I will > soften it by
>saying phase contrast image intensity is 'only
>partially > related' to optical thickness of the
>specimen along the optical axis > of the
>microscope. > > You are absolutely correct that
>the 'phase' in phase-contrast is the > 1/4-wave
>phase shift *due to dielectric dipole
>scattering*. But, the > optical path length does
>affect the amplitude and phase of the >
>scattered light on top of the 1/4-wave phase
>shift due to dielectric > interaction of light
>and material. > > If we read the phasor diagrams
>(Fig. 1, right, MD') in Zernike's > Nobel
>lecture, he suggests that there is pi/2 phase
>shift between > direct and scattered light and
>THEN SOME. As is correctly documented > at most
>places, the phase-ring either advances or
>retards the direct > light by pi/2 so that its
>superposition with scattered light produces >
>measurable intensity variation when they come
>together in the image > plane. > > My approach
>to understanding this is to ask: if I have a
>pure phase > grating (say a saw tooth or
>sinusoidal variation in optical > thickness), at
>what angles and with what amplitudes does the
>scattered > light (in this case diffraction
>orders) propagate? > > The answer can be
>obtained by taking 1D Fourier transform of
>periodic > complex transmission
>[Exp(i*2pi/lambda*refractive index* >
>thicknessprofile)]. > > The spatial frequencies
>in the spectrum are the sine of the angle at >
>which diffraction orders leave the specimen, and
>also the location of > the diffraction orders in
>the back focal plane of the objective. The >
>amplitudes of the diffraction orders does change
>as you increase the > thickness of the grating.
>If we start with a very small bead or an >
>extended feature change its size, its
>diffraction spectrum will change > a little-bit
>with change in size. > > So the optical path
>length affects the amplitude of the scattered >
>light (on top of the pi/4 shift due to
>scattering), which in turn > produces intensity
>modulation. > > I tried to construct a more
>intuitive (non-Fourier) diffraction model > of
>phase-contrast using Huygen's wavelets, but
>didn't get very far. > > Shalin >"

--
****************************************
James and Christine Pawley, 5446 Burley Place (PO
Box 2348), Sechelt, BC, Canada, V0N3A0,
Phone 604-885-0840, email <[hidden email]>
NEW! NEW! AND DIFFERENT Cell (when I remember to turn it on!) 1-604-989-6146

mmodel

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

I know there is quantitative DIC and quantitative defocused brightfield and various interferometric methods, digital holography, etc. But I never read about quantitative phase contrast. People must have realized that it's not that simple.

Mike Model

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of James Pawley
Sent: Monday, March 24, 2014 9:25 PM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi all,

Just to muddy the waters a little more, or perhaps, to add a practical point:

We need to remember that most objects delineated by a change in RI in the XY plane also have changes in RI in the z plane (i.e., these XY RI patterns seldom extend without change very far in the z direction).

Imagine a pattern etched in a 1µm thick layer of RI 1.4 plastic on the surface of RI 1.5 glass slide. If we look at this through immersion oil, light passing through the top and bottom surfaces of the plastic bits will be reflected back towards the light source (reflection is proportional to the square of the difference in RI). Light passing through only glass and oil will not be so reflected

At high NA (large phase rings), this effect is enhanced because the conical illumination then approaches these horizontal surfaces at a fairly oblique angle, a second factor that increases reflection).

Consequently, regardless of what happens with the phase and intensity shifts induced by the phase ring, plastic areas will look darker than "non-plastic" (oil-filled?) areas of the specimen.

In the real world, things are seldom as simple as the math.

JP

>
>Hi Andrew, the distinction between phase shift and deflection is not
>quite clear (at least to me). You can clearly get phase shift without
>deflection and probably the other way round, mut most often you get
>both (even in your simple model cases). But most often you get both: an
>irregular refractive index distribution leads to phase shifts in near
>field and deflection in far field. And, indeed, phase contrast relies
>both on phase shift and on scattering (deflection). I think the
>description given on Wikipedia is quite informative. I agree that one
>definitely does not measure path length by phase contrast imaging. You
>can only measure path length difference (either locally, interpreting
>DIC or phase contrast images; or globally, with some sort of
>interferometric design). You can measure the thickness difference of
>two films (preferably reasonably thin, with no gap between
>them) by just observing the transition across the edge where they
>meet... The matter is by no means trivial and the models predicting the
>image of a given structure are complicated (some links were given
>earlier in this thread). After all, Zernike got a Nobel Prize for
>this...
>(well... alright, there was also Nobel Prize for photo-electric effect,
>but that's a different story :-). A final note to your Alaxa 488
>molecule: angle-dependent phase shift may still be spatially invariant
>if it happens 'far away from the sample plane' (that means 'near the
>aperture plane', like all tradidional objective abberations). Then you
>can define (crippled)
>PSF. Best, zdenek svindrych ---------- PÛvodní
>zpráva ---------- Od: Andrew York
><[hidden email]> Komu:
>[hidden email] Datum: 24. 3.
>2014 22:02:10 PÞedmût: Re: Phase contrast microscopy "***** To join,
>leave or search the confocal microscopy listserv, go to:
>http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>***** It sounds like we agree on my primary
>point: *Many high-quality, authoritative sources imply phase-contrast
>imaging measures path length, which is incorrect and misleading*.
>Statements like this: It would normally be expected that the image of a
>large phase specimen > having a constant optical path length across the
>diameter would appear > uniformly dark or light in the [phase contrast]
>microscope. ...or this: ...optical path differences produce a linear
>reduction in intensity with > increasing phase shift (the image grows
>progressively darker) up to a point > ...only make sense if you think
>phase contrast microscopy measures path length (like interferometric
>imaging) rather than measuring deflection. If you believe this, you
>might use phase contrast imaging to measure which of two thin films is
>thicker, and you would get nonsense results. * Why is this
>misconception so widespread? What can we do to correct it?* **** Minor
>details following up on Shalin's points:
>**** I agree with your refinement to my
>description: if deflected light is re-deflected, or the different
>deflection directions encounter substantially different optical path
>lengths on the way to the detector, this will alter a phase contrast
>image. I think we agree that this doesn't apply to either of the simple
>examples I initially asked about: 1. A single, isolated, pointlike bead
>(very little path length delta, so almost invisible in interferometric
>imaging; high-angle scattering, so visible in phase contrast imaging)
>2. A wide, ~micron thick, uniform slab of glass (huge path length
>delta, so visible in interferometric imaging; no scattering except at
>the edges so almost invisible to phase contrast imaging) I think we
>also agree that if light is re-deflected, or different deflection
>directions encounter substantially different optical path lengths, this
>will alter the image in any form of microscopy that depends on
>emission, and we usually neglect these corrections. For example, if the
>fluorescent light from a single Alexa 488 encounters optical path
>lengths which vary substantially for different emission directions on
>its way to a CCD, the image of that Alexa molecule will be distorted,
>and we can't model the imaging system as having a spatially invariant
>PSF. On Thu, Mar 20, 2014 at 7:53 PM, Shalin Mehta
><[hidden email]>wrote: >
>***** > To join, leave or search the confocal microscopy listserv, go
>to: > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy >
>***** > > On Thu, Mar 20, 2014 at 12:13 PM, Andrew York >
><[hidden email]>
>wrote: > > It seems to me that phase contrast image intensity is
>*completely > > unrelated* to optical path difference produced by the
>specimen. > > Interferometric imaging (like
>this: http://goo.gl/UuE5Bz ), depends on > > optical path differences
>directly (assuming the system is aligned with > > perfect constructive
>interference at output 1 and perfect destructive > > interference at
>output 2 before the sample goes in). Phase contrast > imaging > > (like
>this: http://goo.gl/NDtckE ), seems to depend only on the sample > >
>changing the direction of the light; the sample-induced 1/4-wave phase
>> > shift is purely an effect of dipole scattering, and has nothing to
>do > with > > optical path differences, right? > > I almost agree with
>you Andy on your first statement above, but I will > soften it by
>saying phase contrast image intensity is 'only partially > related' to
>optical thickness of the specimen along the optical axis > of the
>microscope. > > You are absolutely correct that the 'phase' in
>phase-contrast is the > 1/4-wave phase shift *due to dielectric dipole
>scattering*. But, the > optical path length does affect the amplitude
>and phase of the > scattered light on top of the 1/4-wave phase shift
>due to dielectric > interaction of light and material. > > If we read
>the phasor diagrams (Fig. 1, right, MD') in Zernike's > Nobel lecture,
>he suggests that there is pi/2 phase shift between > direct and
>scattered light and THEN SOME. As is correctly documented > at most
>places, the phase-ring either advances or retards the direct > light by
>pi/2 so that its superposition with scattered light produces >
>measurable intensity variation when they come together in the image >
>plane. > > My approach to understanding this is to ask: if I have a
>pure phase > grating (say a saw tooth or sinusoidal variation in
>optical > thickness), at what angles and with what amplitudes does the
>scattered > light (in this case diffraction
>orders) propagate? > > The answer can be obtained by taking 1D Fourier
>transform of periodic > complex transmission
>[Exp(i*2pi/lambda*refractive index* > thicknessprofile)]. > > The
>spatial frequencies in the spectrum are the sine of the angle at >
>which diffraction orders leave the specimen, and also the location of >
>the diffraction orders in the back focal plane of the objective. The >
>amplitudes of the diffraction orders does change as you increase the >
>thickness of the grating.
>If we start with a very small bead or an > extended feature change its
>size, its diffraction spectrum will change > a little-bit with change
>in size. > > So the optical path length affects the amplitude of the
>scattered > light (on top of the pi/4 shift due to scattering), which
>in turn > produces intensity modulation. > > I tried to construct a
>more intuitive (non-Fourier) diffraction model > of phase-contrast
>using Huygen's wavelets, but didn't get very far. > > Shalin >"

--
****************************************
James and Christine Pawley, 5446 Burley Place (PO Box 2348), Sechelt, BC, Canada, V0N3A0, Phone 604-885-0840, email <[hidden email]> NEW! NEW! AND DIFFERENT Cell (when I remember to turn it on!) 1-604-989-6146

Philippe clemenceau

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Mike,

I think Prof. Popescu SLIM method is what you are calling quantitative phase
contrast.
It is basically equivalent to a microscope equipped with a modulated phase
contrast objective, meaning that the phase ring can take 4 different phase
values.

4 phase contrast images are then recorded , and then the optical path length
is reconstructed everywhere in the image. The path length sensitivity is
sub-nanometer.

I invite the list to visit the following links for more info :
http://light.ece.illinois.edu/index.html/archives/portfolio-item/spatial-lig
ht-interference-microscopy-slim

A company was even created called PhiOptics which commercializes the SLIM
attachment.
http://www.phioptics.com

Regards,

Philippe Clémenceau, Division Manager, MS in Optical Science

Imagine Optic Inc./Axiom Optics
Ph:+1 (617) 401 2198
Cell: + 1 (310) 597 1347
1 Broadway, 14th floor
Fax: +1(425) 930 9818
Cambridge, MA 02142
www.axiomoptics.com

Metrology, Adaptive Optics, Scientific Imaging, Laser Measurements

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On
Behalf Of MODEL, MICHAEL
Sent: Tuesday, March 25, 2014 1:09 PM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

I know there is quantitative DIC and quantitative defocused brightfield and
various interferometric methods, digital holography, etc. But I never read
about quantitative phase contrast. People must have realized that it's not
that simple.

Mike Model

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On
Behalf Of James Pawley
Sent: Monday, March 24, 2014 9:25 PM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi all,

Just to muddy the waters a little more, or perhaps, to add a practical
point:

We need to remember that most objects delineated by a change in RI in the XY
plane also have changes in RI in the z plane (i.e., these XY RI patterns
seldom extend without change very far in the z direction).

Imagine a pattern etched in a 1µm thick layer of RI 1.4 plastic on the
surface of RI 1.5 glass slide. If we look at this through immersion oil,
light passing through the top and bottom surfaces of the plastic bits will
be reflected back towards the light source (reflection is proportional to
the square of the difference in RI). Light passing through only glass and
oil will not be so reflected

At high NA (large phase rings), this effect is enhanced because the conical
illumination then approaches these horizontal surfaces at a fairly oblique
angle, a second factor that increases reflection).

Consequently, regardless of what happens with the phase and intensity shifts
induced by the phase ring, plastic areas will look darker than "non-plastic"
(oil-filled?) areas of the specimen.

In the real world, things are seldom as simple as the math.

JP

--
****************************************
James and Christine Pawley, 5446 Burley Place (PO Box 2348), Sechelt, BC,
Canada, V0N3A0, Phone 604-885-0840, email <[hidden email]> NEW! NEW! AND
DIFFERENT Cell (when I remember to turn it on!) 1-604-989-6146

mmodel

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Philippe - you are right, they are doing it using some kind of phase contrast. Thank you!

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Philippe Clemenceau
Sent: Tuesday, March 25, 2014 1:22 PM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Mike,

I think Prof. Popescu SLIM method is what you are calling quantitative phase contrast.
It is basically equivalent to a microscope equipped with a modulated phase contrast objective, meaning that the phase ring can take 4 different phase values.

4 phase contrast images are then recorded , and then the optical path length is reconstructed everywhere in the image. The path length sensitivity is sub-nanometer.

I invite the list to visit the following links for more info :
http://light.ece.illinois.edu/index.html/archives/portfolio-item/spatial-lig
ht-interference-microscopy-slim

A company was even created called PhiOptics which commercializes the SLIM attachment.
http://www.phioptics.com

Regards,

Philippe Clémenceau, Division Manager, MS in Optical Science

Imagine Optic Inc./Axiom Optics
Ph:+1 (617) 401 2198
Cell: + 1 (310) 597 1347
1 Broadway, 14th floor
Fax: +1(425) 930 9818
Cambridge, MA 02142
www.axiomoptics.com

Metrology, Adaptive Optics, Scientific Imaging, Laser Measurements

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of MODEL, MICHAEL
Sent: Tuesday, March 25, 2014 1:09 PM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

I know there is quantitative DIC and quantitative defocused brightfield and various interferometric methods, digital holography, etc. But I never read about quantitative phase contrast. People must have realized that it's not that simple.

Mike Model

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of James Pawley
Sent: Monday, March 24, 2014 9:25 PM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi all,

Just to muddy the waters a little more, or perhaps, to add a practical
point:

We need to remember that most objects delineated by a change in RI in the XY plane also have changes in RI in the z plane (i.e., these XY RI patterns seldom extend without change very far in the z direction).

Imagine a pattern etched in a 1µm thick layer of RI 1.4 plastic on the surface of RI 1.5 glass slide. If we look at this through immersion oil, light passing through the top and bottom surfaces of the plastic bits will be reflected back towards the light source (reflection is proportional to the square of the difference in RI). Light passing through only glass and oil will not be so reflected

At high NA (large phase rings), this effect is enhanced because the conical illumination then approaches these horizontal surfaces at a fairly oblique angle, a second factor that increases reflection).

Consequently, regardless of what happens with the phase and intensity shifts induced by the phase ring, plastic areas will look darker than "non-plastic"
(oil-filled?) areas of the specimen.

In the real world, things are seldom as simple as the math.

JP

--
****************************************
James and Christine Pawley, 5446 Burley Place (PO Box 2348), Sechelt, BC, Canada, V0N3A0, Phone 604-885-0840, email <[hidden email]> NEW! NEW! AND DIFFERENT Cell (when I remember to turn it on!) 1-604-989-6146

Sergey Tauger

Re: Phase contrast microscopy

In reply to this post by Andrew York

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi,

Mike, Phillippe, could you please share the article names of quantitative phase
contrast, brightfield and DIC? I cannot find any articles neither that prove the
methods are quantitative, nor PSF for the methods.

Best,
Sergey

mmodel

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Sergey -
Quantitative DIC has been described in

Kou et al. "Transport-of-intensity approach to differential interference contrast (TI-DIC) microscopy for quantitative phase imaging." Optics letters 35.3 (2010): 447-449.

For quantitative brightfield also look for "transport of intensity".

There is also a method called "defocusing microscopy", I don't know much about it.

Mike

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Sergey Tauger
Sent: Wednesday, March 26, 2014 10:45 AM
To: [hidden email]
Subject: Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi,

Mike, Phillippe, could you please share the article names of quantitative phase contrast, brightfield and DIC? I cannot find any articles neither that prove the methods are quantitative, nor PSF for the methods.

Best,
Sergey

Andrew York

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Some excellent points have been raised:
* Deflection and optical path length differences are often two sides of the
same coin (Zdenek)
* Image brightness in phase microscopy is sensitive to reflection, in
addition to absorption and deflection (James)
...and I'm not sure we've reached a consensus on my original question.

I've tried to illustrate the simplest realistic case which addresses my
question and avoids these complications. In the diagram below:
http://goo.gl/2u6DDe
...my phase object is a homogenous glass slab with two raised regions. The
two regions have different thicknesses, (let's say 200 nm and 400 nm
thick). The two raised regions are wide compared to the lateral resolution
of the imaging system (let's say tens of microns). The raised regions are
perfectly smooth, with the same index of refraction as the slab. The
immersion medium is air. The illumination wavelength is visible (let's say
500 nm).

My expectations:
The phase contrast image shows the same brightness everywhere in the image,
except at the edges of the raised regions.
The interferometric image shows a brightness which depends on the local
thickness of the object, three different brightnesses in this case.

I believe Zdenek and Shalin share my expectations. I sounds like Tobias
does not share my expectations, but is equally curious about the answer.
James, what's your expectation?

My questions:
1. Do my expectations match reality?
2. Is there any confusion or disagreement about what to expect?
3. Does anyone think the phase contrast image intensity near the centers of
the raised regions shows which of the two regions is thicker?

Bonus questions: (Assuming 1, 2, and 3 are straightforward)
4. Does varying the phase shift induced by the phase ring (as described by
Phillippe) allow image intensity near the centers of the raised regions to
tell us local thickness of the raised regions?

On Wed, Mar 26, 2014 at 11:01 AM, MODEL, MICHAEL <[hidden email]> wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Sergey -
> Quantitative DIC has been described in
>
> Kou et al. "Transport-of-intensity approach to differential interference
> contrast (TI-DIC) microscopy for quantitative phase imaging." Optics
> letters 35.3 (2010): 447-449.
>
> For quantitative brightfield also look for "transport of intensity".
>
> There is also a method called "defocusing microscopy", I don't know much
> about it.
>
> Mike
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]]
> On Behalf Of Sergey Tauger
> Sent: Wednesday, March 26, 2014 10:45 AM
> To: [hidden email]
> Subject: Re: Phase contrast microscopy
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hi,
>
> Mike, Phillippe, could you please share the article names of quantitative
> phase contrast, brightfield and DIC? I cannot find any articles neither
> that prove the methods are quantitative, nor PSF for the methods.
>
> Best,
> Sergey
>

Andrew York

Re: Phase contrast microscopy

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hm, I should have drawn the rays differently between the annular aperture
and the condenser lens in the phase contrast illustration. Oops.

I think this is irrelevant to my questions, though. Hope it's not a
distraction.
On Mar 26, 2014 3:59 PM, "Andrew York" <
[hidden email]> wrote:

> Some excellent points have been raised:
> * Deflection and optical path length differences are often two sides of
> the same coin (Zdenek)
> * Image brightness in phase microscopy is sensitive to reflection, in
> addition to absorption and deflection (James)
> ...and I'm not sure we've reached a consensus on my original question.
>
> I've tried to illustrate the simplest realistic case which addresses my
> question and avoids these complications. In the diagram below:
> http://goo.gl/2u6DDe
> ...my phase object is a homogenous glass slab with two raised regions. The
> two regions have different thicknesses, (let's say 200 nm and 400 nm
> thick). The two raised regions are wide compared to the lateral resolution
> of the imaging system (let's say tens of microns). The raised regions are
> perfectly smooth, with the same index of refraction as the slab. The
> immersion medium is air. The illumination wavelength is visible (let's say
> 500 nm).
>
> My expectations:
> The phase contrast image shows the same brightness everywhere in the
> image, except at the edges of the raised regions.
> The interferometric image shows a brightness which depends on the local
> thickness of the object, three different brightnesses in this case.
>
> I believe Zdenek and Shalin share my expectations. I sounds like Tobias
> does not share my expectations, but is equally curious about the answer.
> James, what's your expectation?
>
> My questions:
> 1. Do my expectations match reality?
> 2. Is there any confusion or disagreement about what to expect?
> 3. Does anyone think the phase contrast image intensity near the centers
> of the raised regions shows which of the two regions is thicker?
>
> Bonus questions: (Assuming 1, 2, and 3 are straightforward)
> 4. Does varying the phase shift induced by the phase ring (as described by
> Phillippe) allow image intensity near the centers of the raised regions to
> tell us local thickness of the raised regions?
>
> On Wed, Mar 26, 2014 at 11:01 AM, MODEL, MICHAEL <[hidden email]> wrote:
>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> Post images on http://www.imgur.com and include the link in your posting.
>> *****
>>
>> Sergey -
>> Quantitative DIC has been described in
>>
>> Kou et al. "Transport-of-intensity approach to differential interference
>> contrast (TI-DIC) microscopy for quantitative phase imaging." Optics
>> letters 35.3 (2010): 447-449.
>>
>> For quantitative brightfield also look for "transport of intensity".
>>
>> There is also a method called "defocusing microscopy", I don't know much
>> about it.
>>
>> Mike
>>
>> -----Original Message-----
>> From: Confocal Microscopy List [mailto:[hidden email]]
>> On Behalf Of Sergey Tauger
>> Sent: Wednesday, March 26, 2014 10:45 AM
>> To: [hidden email]
>> Subject: Re: Phase contrast microscopy
>>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> Post images on http://www.imgur.com and include the link in your posting.
>> *****
>>
>> Hi,
>>
>> Mike, Phillippe, could you please share the article names of quantitative
>> phase contrast, brightfield and DIC? I cannot find any articles neither
>> that prove the methods are quantitative, nor PSF for the methods.
>>
>> Best,
>> Sergey
>>
>
>

George McNamara

Re: Phase contrast microscopy

In reply to this post by Philippe clemenceau

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

QPi / QPm ...

The Popescu et al is related to Michael Feld's (deceased) - uses
hardware. See recent paper

Pham ... Popescu 2013 Real Time Blood Testing Using Quantitative Phase
Imaging
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0055676
they used a GPU to speed up calculations.

Hamamatsu's uses hardware (and is Michael Feld's approach):
http://www.hamamatsu.com/us/en/technology/innovation/qpm/index.html

"QPm" is software only (paper is on confocal, but widefield was reported
by the same authors:

Cody SH, Xiang SD, Layton MJ, Handman E, Lam MH, Layton JE, Nice EC, Heath JK. A simple method allowing DIC imaging in conjunction with confocal microscopy. J Microsc. 2005 Mar;217(Pt 3):265-74. PubMed PMID: 15725130.

http://www.ncbi.nlm.nih.gov/pubmed/15725130
http://aups.org.au/Proceedings/34/121-127/121-127.pdf
Iatia is now Ultima Capital
http://www.ultimacapital.net/iatiaimaging/Publications/Iatia%20Imaging/applicationNotes/comparisonWithOpticalPhaseContrastModalities.pdf

Unfortunately, they emphasize emulating DIC and phase contrast images,
rather than the "Q" maps.
The QPm software is how the GE InCell 2000 and 6000 high content
screening instruments produced "phase" contrast images with widefield
objective lenses (but apparently with no option for "Q" maps). One of
these was on the Applied Precision / GE Healthcare DeltaVision bus a
couple of years ago in Miami. My thanks to Janine at API/GEH for
arranging that visit.

Enjoy,

George

On 3/25/2014 12:22 PM, Philippe Clemenceau wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi Mike,
>
> I think Prof. Popescu SLIM method is what you are calling quantitative phase
> contrast.
> It is basically equivalent to a microscope equipped with a modulated phase
> contrast objective, meaning that the phase ring can take 4 different phase
> values.
>
> 4 phase contrast images are then recorded , and then the optical path length
> is reconstructed everywhere in the image. The path length sensitivity is
> sub-nanometer.
>
>
> I invite the list to visit the following links for more info :
> http://light.ece.illinois.edu/index.html/archives/portfolio-item/spatial-lig
> ht-interference-microscopy-slim
>
> A company was even created called PhiOptics which commercializes the SLIM
> attachment.
> http://www.phioptics.com
>
> Regards,
>
> Philippe Clémenceau, Division Manager, MS in Optical Science
>
> Imagine Optic Inc./Axiom Optics
> Ph:+1 (617) 401 2198
> Cell: + 1 (310) 597 1347
> 1 Broadway, 14th floor
> Fax: +1(425) 930 9818
> Cambridge, MA 02142
> www.axiomoptics.com
>
> Metrology, Adaptive Optics, Scientific Imaging, Laser Measurements
>
>
>
>
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]] On
> Behalf Of MODEL, MICHAEL
> Sent: Tuesday, March 25, 2014 1:09 PM
> To: [hidden email]
> Subject: Re: Phase contrast microscopy
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> I know there is quantitative DIC and quantitative defocused brightfield and
> various interferometric methods, digital holography, etc. But I never read
> about quantitative phase contrast. People must have realized that it's not
> that simple.
>
> Mike Model
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]] On
> Behalf Of James Pawley
> Sent: Monday, March 24, 2014 9:25 PM
> To: [hidden email]
> Subject: Re: Phase contrast microscopy
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi all,
>
> Just to muddy the waters a little more, or perhaps, to add a practical
> point:
>
> We need to remember that most objects delineated by a change in RI in the XY
> plane also have changes in RI in the z plane (i.e., these XY RI patterns
> seldom extend without change very far in the z direction).
>
> Imagine a pattern etched in a 1µm thick layer of RI 1.4 plastic on the
> surface of RI 1.5 glass slide. If we look at this through immersion oil,
> light passing through the top and bottom surfaces of the plastic bits will
> be reflected back towards the light source (reflection is proportional to
> the square of the difference in RI). Light passing through only glass and
> oil will not be so reflected
>
> At high NA (large phase rings), this effect is enhanced because the conical
> illumination then approaches these horizontal surfaces at a fairly oblique
> angle, a second factor that increases reflection).
>
> Consequently, regardless of what happens with the phase and intensity shifts
> induced by the phase ring, plastic areas will look darker than "non-plastic"
> (oil-filled?) areas of the specimen.
>
> In the real world, things are seldom as simple as the math.
>
> JP
>
>
>> Hi Andrew, the distinction between phase shift and deflection is not
>> quite clear (at least to me). You can clearly get phase shift without
>> deflection and probably the other way round, mut most often you get
>> both (even in your simple model cases). But most often you get both: an
>> irregular refractive index distribution leads to phase shifts in near
>> field and deflection in far field. And, indeed, phase contrast relies
>> both on phase shift and on scattering (deflection). I think the
>> description given on Wikipedia is quite informative. I agree that one
>> definitely does not measure path length by phase contrast imaging. You
>> can only measure path length difference (either locally, interpreting
>> DIC or phase contrast images; or globally, with some sort of
>> interferometric design). You can measure the thickness difference of
>> two films (preferably reasonably thin, with no gap between
>> them) by just observing the transition across the edge where they
>> meet... The matter is by no means trivial and the models predicting the
>> image of a given structure are complicated (some links were given
>> earlier in this thread). After all, Zernike got a Nobel Prize for
>> this...
>> (well... alright, there was also Nobel Prize for photo-electric effect,
>> but that's a different story :-). A final note to your Alaxa 488
>> molecule: angle-dependent phase shift may still be spatially invariant
>> if it happens 'far away from the sample plane' (that means 'near the
>> aperture plane', like all tradidional objective abberations). Then you
>> can define (crippled)
>> PSF. Best, zdenek svindrych ---------- PÛvodní
>> zpráva ---------- Od: Andrew York
>> <[hidden email]> Komu:
>> [hidden email] Datum: 24. 3.
>> 2014 22:02:10 PÞedmût: Re: Phase contrast microscopy "***** To join,
>> leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> ***** It sounds like we agree on my primary
>> point: *Many high-quality, authoritative sources imply phase-contrast
>> imaging measures path length, which is incorrect and misleading*.
>> Statements like this: It would normally be expected that the image of a
>> large phase specimen> having a constant optical path length across the
>> diameter would appear> uniformly dark or light in the [phase contrast]
>> microscope. ...or this: ...optical path differences produce a linear
>> reduction in intensity with> increasing phase shift (the image grows
>> progressively darker) up to a point> ...only make sense if you think
>> phase contrast microscopy measures path length (like interferometric
>> imaging) rather than measuring deflection. If you believe this, you
>> might use phase contrast imaging to measure which of two thin films is
>> thicker, and you would get nonsense results. * Why is this
>> misconception so widespread? What can we do to correct it?* **** Minor
>> details following up on Shalin's points:
>> **** I agree with your refinement to my
>> description: if deflected light is re-deflected, or the different
>> deflection directions encounter substantially different optical path
>> lengths on the way to the detector, this will alter a phase contrast
>> image. I think we agree that this doesn't apply to either of the simple
>> examples I initially asked about: 1. A single, isolated, pointlike bead
>> (very little path length delta, so almost invisible in interferometric
>> imaging; high-angle scattering, so visible in phase contrast imaging)
>> 2. A wide, ~micron thick, uniform slab of glass (huge path length
>> delta, so visible in interferometric imaging; no scattering except at
>> the edges so almost invisible to phase contrast imaging) I think we
>> also agree that if light is re-deflected, or different deflection
>> directions encounter substantially different optical path lengths, this
>> will alter the image in any form of microscopy that depends on
>> emission, and we usually neglect these corrections. For example, if the
>> fluorescent light from a single Alexa 488 encounters optical path
>> lengths which vary substantially for different emission directions on
>> its way to a CCD, the image of that Alexa molecule will be distorted,
>> and we can't model the imaging system as having a spatially invariant
>> PSF. On Thu, Mar 20, 2014 at 7:53 PM, Shalin Mehta
>> <[hidden email]>wrote:>
>> *****> To join, leave or search the confocal microscopy listserv, go
>> to:> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy>
>> *****> > On Thu, Mar 20, 2014 at 12:13 PM, Andrew York>
>> <[hidden email]>
>> wrote:> > It seems to me that phase contrast image intensity is
>> *completely> > unrelated* to optical path difference produced by the
>> specimen.> > Interferometric imaging (like
>> this: http://goo.gl/UuE5Bz ), depends on> > optical path differences
>> directly (assuming the system is aligned with> > perfect constructive
>> interference at output 1 and perfect destructive> > interference at
>> output 2 before the sample goes in). Phase contrast> imaging> > (like
>> this: http://goo.gl/NDtckE ), seems to depend only on the sample> >
>> changing the direction of the light; the sample-induced 1/4-wave phase
>>
>>>> shift is purely an effect of dipole scattering, and has nothing to
>>>>
>> do> with> > optical path differences, right?> > I almost agree with
>> you Andy on your first statement above, but I will> soften it by
>> saying phase contrast image intensity is 'only partially> related' to
>> optical thickness of the specimen along the optical axis> of the
>> microscope.> > You are absolutely correct that the 'phase' in
>> phase-contrast is the> 1/4-wave phase shift *due to dielectric dipole
>> scattering*. But, the> optical path length does affect the amplitude
>> and phase of the> scattered light on top of the 1/4-wave phase shift
>> due to dielectric> interaction of light and material.> > If we read
>> the phasor diagrams (Fig. 1, right, MD') in Zernike's> Nobel lecture,
>> he suggests that there is pi/2 phase shift between> direct and
>> scattered light and THEN SOME. As is correctly documented> at most
>> places, the phase-ring either advances or retards the direct> light by
>> pi/2 so that its superposition with scattered light produces>
>> measurable intensity variation when they come together in the image>
>> plane.> > My approach to understanding this is to ask: if I have a
>> pure phase> grating (say a saw tooth or sinusoidal variation in
>> optical> thickness), at what angles and with what amplitudes does the
>> scattered> light (in this case diffraction
>> orders) propagate?> > The answer can be obtained by taking 1D Fourier
>> transform of periodic> complex transmission
>> [Exp(i*2pi/lambda*refractive index*> thicknessprofile)].> > The
>> spatial frequencies in the spectrum are the sine of the angle at>
>> which diffraction orders leave the specimen, and also the location of>
>> the diffraction orders in the back focal plane of the objective. The>
>> amplitudes of the diffraction orders does change as you increase the>
>> thickness of the grating.
>> If we start with a very small bead or an> extended feature change its
>> size, its diffraction spectrum will change> a little-bit with change
>> in size.> > So the optical path length affects the amplitude of the
>> scattered> light (on top of the pi/4 shift due to scattering), which
>> in turn> produces intensity modulation.> > I tried to construct a
>> more intuitive (non-Fourier) diffraction model> of phase-contrast
>> using Huygen's wavelets, but didn't get very far.> > Shalin>"
>>
>
>

--

George McNamara, Ph.D.
Single Cells Analyst
L.J.N. Cooper Lab
University of Texas M.D. Anderson Cancer Center
Houston, TX 77054
Tattletales http://works.bepress.com/gmcnamara/26/

John Oreopoulos

Re: Phase contrast microscopy

In reply to this post by Andrew York

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Andrew, this paper might be of interest to you:

Title:
Structured illumination diffraction phase microscopy for broadband, subdiffraction resolution, quantitative phase imaging

Authors:
Chowdhury, S; Izatt, J

Author Full Names:
Chowdhury, Shwetadwip; Izatt, Joseph

Source:
OPTICS LETTERS, 39 (4):1015-1018; 10.1364/OL.39.001015 FEB 15 2014

Language:
English

Document Type:
Article

KeyWords Plus:
DIGITAL HOLOGRAPHIC MICROSCOPY; FLUORESCENCE MICROSCOPY; LIMIT; SUPERRESOLUTION

Abstract:
Structured illumination microscopy (SIM) is an established technique
that allows subdiffraction resolution imaging by heterodyning high
sample frequencies into the system's passband via structured
illumination. However, until now, SIM has been typically used to achieve
subdiffraction resolution for intensity-based imaging. Here, we present
a novel optical setup that uses structured illumination with a broadband
light source to obtain noise-reduced, subdiffraction resolution,
quantitative phase imaging (QPM) of cells. We compare this with a
previous work for subdiffraction QPM imaging via SIM that used a laser
source, and was thus still corrupted by coherent noise. (C) 2014 Optical
Society of America

On 2014-03-26, at 6:12 PM, Andrew York wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hm, I should have drawn the rays differently between the annular aperture
> and the condenser lens in the phase contrast illustration. Oops.
>
> I think this is irrelevant to my questions, though. Hope it's not a
> distraction.
> On Mar 26, 2014 3:59 PM, "Andrew York" <
> [hidden email]> wrote:
>
>> Some excellent points have been raised:
>> * Deflection and optical path length differences are often two sides of
>> the same coin (Zdenek)
>> * Image brightness in phase microscopy is sensitive to reflection, in
>> addition to absorption and deflection (James)
>> ...and I'm not sure we've reached a consensus on my original question.
>>
>> I've tried to illustrate the simplest realistic case which addresses my
>> question and avoids these complications. In the diagram below:
>> http://goo.gl/2u6DDe
>> ...my phase object is a homogenous glass slab with two raised regions. The
>> two regions have different thicknesses, (let's say 200 nm and 400 nm
>> thick). The two raised regions are wide compared to the lateral resolution
>> of the imaging system (let's say tens of microns). The raised regions are
>> perfectly smooth, with the same index of refraction as the slab. The
>> immersion medium is air. The illumination wavelength is visible (let's say
>> 500 nm).
>>
>> My expectations:
>> The phase contrast image shows the same brightness everywhere in the
>> image, except at the edges of the raised regions.
>> The interferometric image shows a brightness which depends on the local
>> thickness of the object, three different brightnesses in this case.
>>
>> I believe Zdenek and Shalin share my expectations. I sounds like Tobias
>> does not share my expectations, but is equally curious about the answer.
>> James, what's your expectation?
>>
>> My questions:
>> 1. Do my expectations match reality?
>> 2. Is there any confusion or disagreement about what to expect?
>> 3. Does anyone think the phase contrast image intensity near the centers
>> of the raised regions shows which of the two regions is thicker?
>>
>> Bonus questions: (Assuming 1, 2, and 3 are straightforward)
>> 4. Does varying the phase shift induced by the phase ring (as described by
>> Phillippe) allow image intensity near the centers of the raised regions to
>> tell us local thickness of the raised regions?
>>
>> On Wed, Mar 26, 2014 at 11:01 AM, MODEL, MICHAEL <[hidden email]> wrote:
>>
>>> *****
>>> To join, leave or search the confocal microscopy listserv, go to:
>>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>>> Post images on http://www.imgur.com and include the link in your posting.
>>> *****
>>>
>>> Sergey -
>>> Quantitative DIC has been described in
>>>
>>> Kou et al. "Transport-of-intensity approach to differential interference
>>> contrast (TI-DIC) microscopy for quantitative phase imaging." Optics
>>> letters 35.3 (2010): 447-449.
>>>
>>> For quantitative brightfield also look for "transport of intensity".
>>>
>>> There is also a method called "defocusing microscopy", I don't know much
>>> about it.
>>>
>>> Mike
>>>
>>> -----Original Message-----
>>> From: Confocal Microscopy List [mailto:[hidden email]]
>>> On Behalf Of Sergey Tauger
>>> Sent: Wednesday, March 26, 2014 10:45 AM
>>> To: [hidden email]
>>> Subject: Re: Phase contrast microscopy
>>>
>>> *****
>>> To join, leave or search the confocal microscopy listserv, go to:
>>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>>> Post images on http://www.imgur.com and include the link in your posting.
>>> *****
>>>
>>> Hi,
>>>
>>> Mike, Phillippe, could you please share the article names of quantitative
>>> phase contrast, brightfield and DIC? I cannot find any articles neither
>>> that prove the methods are quantitative, nor PSF for the methods.
>>>
>>> Best,
>>> Sergey
>>>
>>
>>