"Bowed" spectral lines
Ioannis
possession of a small hand-held double Amici prism professional
spectroscope, which was bought 10 years ago.
As far as I am concerned my design plans were correct for the larger
spectroscope (bottom of page). Collimation is also correct and has been
tested extensively. The device performs excellently.
http://users.forthnet.gr/ath/jgal/spectroscope/phasmaplan.html
Does anyone have any idea why on both my prism spectroscopes the images of
emission lines appear more "curved/bowed" as you go towards the purple part
of the spectrum? This effect can be seen on pictures I've taken with both my
large spectroscope and my small one.
Here are some spectra photos with both spectroscopes. Note that as one moves
towards the purple Mercury lines, the lines curve more. The effect is
visible on the small spectroscope as well, but is less pronounced, because
it has much smaller resolution. (Photos which show the whole spectrum are
taken with the small spectroscope. Photos which show part of the spectrum
are taken with the large one).
http://users.forthnet.gr/ath/jgal/spectroscope/Hg.html
I know lower wavelengths are bent more, but what *exactly* is the optics
explanation for the fact that the lines appear "bowed" more towards lower
wavelengths?
Thanks much in advance.
--
I. N. Galidakis
http://users.forthnet.gr/ath/jgal/
Eventually, _everything_ is understandable
Bob May
bowing of the lines.
The thing here is that the pictures of the lines don't always have the same
amount or direction of bowing.
--
Why do penguins walk so far to get to their nesting grounds?
Ioannis
news:11268113...@news-1.nethere.net...
>
> It looks to me like the optics of the telescopes is what is causing the
> bowing of the lines.
There are no telescope components in the small spectroscope, except one lens
to collimate the slit:
slit--lens--amici prism--eye
> The thing here is that the pictures of the lines don't always have the
same
> amount or direction of bowing.
The different direction is because collimation is slightly off and the
telescopes are not EXACTLY centered on the optical axis. The problem is
still unexplainable to me.
> --
> Why do penguins walk so far to get to their nesting grounds?
Jim Klein
Jim Klein
Charles Attends
Will give you an explanation later.
It occurs with prisms or gratings.
"Ioannis" <morp...@olympus.mons> wrote in message
news:1126808927.239502@athnrd02...
Sam Goldwasser
> The bowing of spectral lines is absolutely and entirely normal.
> Will give you an explanation later.
> It occurs with prisms or gratings.
I've seen monochromators where the slits were curved.
--- sam | Sci.Electronics.Repair FAQ Mirror: http://repairfaq.ece.drexel.edu/
Repair | Main Table of Contents: http://repairfaq.ece.drexel.edu/REPAIR/
+Lasers | Sam's Laser FAQ: http://repairfaq.ece.drexel.edu/sam/lasersam.htm
| Mirror Sites: http://repairfaq.ece.drexel.edu/REPAIR/F_mirror.html
Note: These links are hopefully temporary until we can sort out the excessive
traffic on Repairfaq.org.
Important: Anything sent to the email address in the message header above is
ignored unless my full name is included in the subject line. Or, you can
contact me via the Feedback Form in the FAQs.
Marvin
> Ο "Bob May" <bob...@nethere.com> έγραψε στο μήνυμα
> news:11268113...@news-1.nethere.net...
>
>>It looks to me like the optics of the telescopes is what is causing the
>>bowing of the lines.
>
>
> There are no telescope components in the small spectroscope, except one lens
> to collimate the slit:
>
> slit--lens--amici prism--eye
Unless the lens has mutiple components to correct for chromatic aberration, the
col;imation is not done well at all wavelngths, which will cause what you see. Most
spectrometers use front-surface mirrors.
Ioannis
news:11ilthg...@corp.supernews.com...
>
> Ioannis wrote:
> > Ο "Bob May" <bob...@nethere.com> έγραψε στο μήνυμα
> > news:11268113...@news-1.nethere.net...
> >
> >>It looks to me like the optics of the telescopes is what is causing the
> >>bowing of the lines.
> >
> >
> > There are no telescope components in the small spectroscope, except one
lens
> > to collimate the slit:
> >
> > slit--lens--amici prism--eye
>
> Unless the lens has mutiple components to correct for chromatic
aberration, the
> col;imation is not done well at all wavelngths, which will cause what you
see.
Sorry, doesn't make sense to me. It's true that the small spectroscope has a
single lens collimator, but it produces less bowing than the big
spectroscope which has two element objectives on both the collimator and
viewer.
If what you say were true, then a two element lens would collimate better,
thus the big spectroscope would have displayed less bowing.
> Most
> spectrometers use front-surface mirrors.
Charles Attends
"Marvin" <phys...@cloud9.net> wrote
>
> Unless the lens has mutiple components to correct for chromatic
aberration, the
> collimation is not done well at all wavelngths, which will cause what you
see. Most
> spectrometers use front-surface mirrors.
The bowing of spectral lines is inherent to the use of prisms.
Collimation has very little if anything to do with it.
Marvin
the prism. The smaller spectrograph presumably has a lower f-number, which tends to
exaggerate problems like this. Prhaps some ray-tracing will convince you.
Ioannis
[snip]
> > The bowing of spectral lines is inherent to the use of prisms.
> > Collimation has very little if anything to do with it.
> >
> Absent collimation, not all of the light of any wavelength follows
parallel paths through
> the prism. The smaller spectrograph presumably has a lower f-number,
which tends to
> exaggerate problems like this. Prhaps some ray-tracing will convince you.
I think you are still missing the point: The small spectroscope produces
LESS of this effect.
I am still waiting for Charles' explanation, though.
Charles Attends
"Marvin" <phys...@cloud9.net> wrote
> Absent collimation, not all of the light of any wavelength follows
parallel paths through
> the prism. The smaller spectrograph presumably has a lower f-number,
which tends to
> exaggerate problems like this. Perhaps some ray-tracing will convince
you.
Even if collimation were perfect, before and after the prism, curvature of
the monochromatic image of the entrance slit would still occur.
A parallel beam will exit the prism as expected as a parallel beam.
In other a prism is indeed stigmatic for infinity.
But refraction has subtle effects (distortion and anamorphosis, as these are
not strictly speaking aberrations).
To obtain the formula that gives the expression of the curvature requires
four (4) 8'' by 11" pages of manual (optical and spherical trigonometry)
calculations and explanations!
Absent a newsgroup entitled alt.binaries.sci.optics (*) and not having a
home page yet I do not know where to post the result of my efforts!
Or one can indeed do some ray tracing to convince oneself of the fact.
(*) alt.binaries.sci.optics would be very useful for many reasons, one of
them, perhaps the foremost, being that ASCII art has its limits and they are
pretty low! Any comments?
Danny Rich
this effect but it is usually only noticeable for very high resolution
spectrometers or those with very tall slits (mm tall compared to microns
wide). Here is a clip from a McPherson catalog that indicates that the
curvature is due to coma in the spectrometer that they offer curved slits to
correct this curvature in the image.
Bilateral Slits - All instruments come with bilateral slits that are
continuously adjustable by precision micrometer from 5 to 4000-痠.
Micrometer readout is in 10-micron thimble divisions. Interpolation may
allow reading and setting of smaller intervals. Slit height can be varied by
a stepped aperture from 2 to 20-mm. To meet your configuration requirement
monochromator slits may be positioned at the axial (end) or lateral (side)
ports of the monochromator. Optional stepper motorized slit micrometers can
be provided. Also, optional curved slits can be provided for coma correction
even at extreme slit heights.
Since the two spectrometers inquestion have different imaging systems, one
larger and more complex than the other, but neither seem to be designed to
correct for coma then one might expect the system with two imaging element
to either have more coma due to each element contributing something to the
coma or less coma if the second element was selected to cancel the coma of
the first element (does not seem to be the case).
The height to width ratio of the slits may be a better indicator of the
magnitude of the coma and curvature than the number optical elements but
that level of detail was not included in any post that I have read.
Danny
"Marvin" <phys...@cloud9.net> wrote in message
news:11ioebl...@corp.supernews.com...
Ioannis
news:j7%We.3299$zN6....@fe10.lga...
>
> I believe that Charles Attends is correct that most spectrometers suffer
> this effect but it is usually only noticeable for very high resolution
> spectrometers or those with very tall slits (mm tall compared to microns
> wide). Here is a clip from a McPherson catalog that indicates that the
> curvature is due to coma in the spectrometer that they offer curved slits
to
> correct this curvature in the image.
A fixed curvature curved slit won't do a high resolution prism spectroscope
much justice. The curvature is variable depending on the spectra area one is
looking at. The large spectroscope in question goes from an R~7,000-10,000
in the extreme red, to over 75,000 in the extreme purple. As you can see
from the images, the bowing of the yellow mercury doublet is signifactly
less than that of the mercury purple doublet.
I don't see how a constant curvature curved slit could account for the
variable resolution throughout such an instrument's spectral range.
[snip]
> Danny
Peltio
>Absent a newsgroup entitled alt.binaries.sci.optics (*) and not having a
>home page yet I do not know where to post the result of my efforts!
You might want to try one of the following free services:
http://imageshack.us/
http://tinypic.com/
http://www.zippyimages.com/
you can upload any image that is on your hard disk in a matter of seconds.
Obviously a direct scan might need to be resized and saved in a compressed
format (such as jpg or png) in order not to waste to much space. Free tools
such as irfanview can do this in a jiffy (resize, then save as... pick the
file extension and you are done).
cheers,
Peltio
Skywise
news:1126990932.428058@athnrd02:
I'm not an optics expert, but this thread has been interesting.
I have a couple of right angle prisms, and looking through them
at the diffracted image shows this curvature being talked about.
I think if one just concetrates on why the diffracted image is
curved without any external optics, then the answer may become
clear. As I said, I'm no optics expert, but to me it's obvious
the curvature is not caused by any external optics.
Brian
--
http://www.skywise711.com - Lasers, Seismology, Astronomy, Skepticism
Seismic FAQ: http://www.skywise711.com/SeismicFAQ/SeismicFAQ.html
Sed quis custodiet ipsos Custodes?
Ioannis
news:11ip9is...@corp.supernews.com...
>
> "Ioannis" <morp...@olympus.mons> wrote in
> news:1126990932.428058@athnrd02:
>
> > Ο "Danny Rich" <Dann...@softhome.net> έγραψε στο μήνυμα
> > news:j7%We.3299$zN6....@fe10.lga...
> > A fixed curvature curved slit won't do a high resolution prism
> > spectroscope much justice. The curvature is variable depending on the
> > spectra area one is looking at. The large spectroscope in question goes
> > from an R~7,000-10,000 in the extreme red, to over 75,000 in the extreme
> > purple. As you can see from the images, the bowing of the yellow mercury
> > doublet is signifactly less than that of the mercury purple doublet.
> >
> > I don't see how a constant curvature curved slit could account for the
> > variable resolution throughout such an instrument's spectral range.
> >
> > [snip]
> >
> >> Danny
>
> I'm not an optics expert, but this thread has been interesting.
Oh, it most certainly is interesting for me, because I paid $2,500 to the
engineer to built it, and even though I designed it myself, I had never seen
this effect on professional spectroscopic images, so I wasn't anticipating
it. It was a huge disappointment because I thought that my design was for
some reason flawed.
I suspect the images I had looked at, throughout my years of training in
Physics and Optics books were almost always taken with low resolution
spectroscopes or were always graphic reps, and not actual images.
> I have a couple of right angle prisms, and looking through them
> at the diffracted image shows this curvature being talked about.
You mean the "refracted" image. I could disassemble the prism dome and check
this on my SF10 prisms, but because I don't want to handle them with bare
hands, I'll take your word for it. Now that you mention it, I have an old
star diagonal somewhere, so I might check it with this one.
My guess is that one solution to correcting the bowing would be if the prism
was simultaneously a prism AND a concave lens. In other words, if it was a
prism with concave faces. This would counteract the bowing which probably
comes from the combo: Spherical lens + flat prism face. That's just a guess,
though.
> I think if one just concetrates on why the diffracted image is
> curved without any external optics, then the answer may become
> clear. As I said, I'm no optics expert, but to me it's obvious
> the curvature is not caused by any external optics.
Well, you can't really have total absense of all optical devices, even
without the collimator and viewer, since you are still collimating the image
with your eye when you look through the prism, so that's why the effect is
still there, since the eye, cornea, lens, retina and all, is still
spherical:
Spherical eye (collimator) + flat face prism = bowed image.
After thinking about this throughout this thread and for a little while
more, the correct answer to all this, is that the image of the slit would be
perfectly straight if and only if the collimator and viewer (and the
viewer's eyepiece) consisted entirely of cylindrical lenses, with all the
cylinder axes parallel to the slit.
I am surprised nobody thought about it sooner and I have no idea why it's
not already implemented in professional spectroscopes. I guess the cost of
making a cylindrical achromat is prohibitive. And by "cylindrical achromat"
I mean achromat in the direction of the cylinder's axis only.
Manufacturing such a spectroscope, would require 2 cylindrical achromats
plus one cylindrical eyepiece, with however many elements required. This
would be crazy in terms of cost. For starters, a slight rotation of any of
the elements would completely ruin the image.
> Brian
Marvin
I finally searched through some texts that I have at hand, and found only a footnote in
"Principles and Practice of Spectrochmeical Analysis, N. H. Nachtrieb, 1950, p. 24: "Field
curvature is not to be confused with the curvature of spectral lines. The latter is due
to the passage of rays through the prism with an inclination to the principle plane of the
prism. The more oblique rays pass through a greater prism thickness, and are consequently
deviated through a larger angle. As a result, the ends of tall spectral lines are
observed to lie toward shorter wavelength than the middle of the lines; i.e., the lines
are concave toward shorter wavelengths." It is what I had in mind whenI posted my first
answer in this thread.
My best optics reference, "Fundamentals of Optics", F. A. Jenkins & H. E. White, 1950, has
nothing on this topic that I could find.
Ioannis
[snip]
> I finally searched through some texts that I have at hand, and found only
a footnote in
> "Principles and Practice of Spectrochmeical Analysis, N. H. Nachtrieb,
1950, p. 24: "Field
> curvature is not to be confused with the curvature of spectral lines. The
latter is due
> to the passage of rays through the prism with an inclination to the
principle plane of the
> prism. The more oblique rays pass through a greater prism thickness, and
are consequently
> deviated through a larger angle. As a result, the ends of tall spectral
lines are
> observed to lie toward shorter wavelength than the middle of the lines;
i.e., the lines
> are concave toward shorter wavelengths." It is what I had in mind whenI
posted my first
> answer in this thread.
>
> My best optics reference, "Fundamentals of Optics", F. A. Jenkins & H. E.
White, 1950, has
> nothing on this topic that I could find.
Fair enough and logical. Thanks Marvin.
Jim Klein
imaging systems or not, the optimization of the optical system can be
targeted so as to minimize the distortion.
In imaging, hyperspectral spectrometers, we call the errors "cross
track spectral" and "spatial co-registration" and sometimes they need
to be controlled to less than a few percent of a pixel.
The control is achieved by the specific choice of optical form for the
optics which image the slit on to a row of detectors and the optics
which map the slit out on to the target scene.
Jim Klein
Danny Rich
"Ioannis" <morp...@olympus.mons> wrote in message
> A fixed curvature curved slit won't do a high resolution prism
> spectroscope
> much justice. The curvature is variable depending on the spectra area one
> is
> looking at. The large spectroscope in question goes from an R~7,000-10,000
> in the extreme red, to over 75,000 in the extreme purple. As you can see
> from the images, the bowing of the yellow mercury doublet is signifactly
> less than that of the mercury purple doublet.
>
> I don't see how a constant curvature curved slit could account for the
> variable resolution throughout such an instrument's spectral range.
>
If my memory is correct, I believe that curved slits are one form of the
continuously variable mechanical slits which are driven by micrometers and
can be either manual - for sitting at one wavelength or by stepper controled
micrometers for computer controlled scanning.
In terms of the comments from Skywise - I believe that his eye which he used
to view the image of the spectra through the prisms has a lens-based imaging
system and thus becomes a part of the spectrometer. The basic elements of a
spectrometer are collimator, disperser, camera. The two imaging elements
can be combined into one element in some clever designs - such as the
Ebert-Fastie grating design or the Littrow prism which uses a single lens.
The trick is to project an image of the input slit onto the output slit
while passing the rays through the disperser.
Ioannis
[snip]
> In terms of the comments from Skywise - I believe that his eye which he
used
> to view the image of the spectra through the prisms has a lens-based
imaging
> system and thus becomes a part of the spectrometer.
Which is exactly what I pointed out to him in one of the previous responses
of mine, as being the reason for the presence of bowed lines, even without
external optical devices:
Ioannis wrote:
> Well, you can't really have total absense of all optical devices, even
> without the collimator and viewer, since you are still collimating the
image
> with your eye when you look through the prism, so that's why the effect is
> still there, since the eye, cornea, lens, retina and all, is still
> spherical:
> Spherical eye (collimator) + flat face prism = bowed image.
[snip]
Skywise
news:1127086718.330941@athnrd02:
Right. I see your point.
So then, my next idea, is to have no imaging optics AFTER the
dispering prism. Would that eliminate the bowed lines?
After all, if I shine sunlight (nearly collimated) through a
prism, or a collimated light source, and view the refracted
light on a wall the spectrum is not bowed.
Then there's diffraction gratings.
Ioannis
[snip]
> Right. I see your point.
>
> So then, my next idea, is to have no imaging optics AFTER the
> dispering prism. Would that eliminate the bowed lines?
>
> After all, if I shine sunlight (nearly collimated) through a
> prism, or a collimated light source, and view the refracted
> light on a wall the spectrum is not bowed.
Will work, if you "pencil-collimate" the incident beam. Where
"pencil-collimation" refers to two consecutive slits, parallel to the
prism's face.
(Excuse the crude ASCII graphics. The incident/emergent beam angles are not
correct)
Slit1 Slit2
| | /\
Light Source O --------->/ \-----> Projection on Wall
| | /____\
> Then there's diffraction gratings.
>
> Brian
--
Bob May
It was wuilt with the slit being a polished rod about 10' from the
diffrection grating and the telescope (binocular field lens) used to
collimate the beam. The diffraction grating was applied to the end of the
telescope and the telescope was pointed at the angle needed to see the
spectral lines. The device worked from the deep red where my eyes could no
longer see the light to the violet where the same thing happened. All of
the lines (thousands of them!) were all quite straight from one end to the
other.The working aperture was close to the 2" of the lens.
I see no reason why a prism would cause differential bending of the liight
as longer paths through the prism will cause the spreading of the spectrum
line rather than the distortion that you're getting. Putting the prism at
an angle should make the lines merget to a common focus point some distance
above or below the direction of the spectrum which is something else than
curved lines but that can be (and should be) tuned out by the correct
orentation of the prism.
On the other hand, optics outside of the prism itself can produce a
distortion which is curved and this can be seen by putting the lenses so
that you run the white light from the slit through them and you will see
defocusing at the ends of the line whee there will be a color difference
when they are off of alignment.
Danny Rich
This sounds like a great piece of kit. At the lengths of the arms of the
spectrometer it would seem that this had a very high f/# and thus the rays
were very much paraxial and near to the axis. According to the McPherson
page, the coma problem arises when the slits are very tall relative to the
width. The would create rays far from the axis and, I think, subject to a
first order aberation like coma.
Danny Rich
"Bob May" <bob...@nethere.com> wrote in message
news:11272461...@news-1.nethere.net...