Can diamond now be used for telescope mirrors?
Robert Clark
maintain their shape because of its strength and lightness. The
reflectivity and smoothness comes from a thin layer of metal applied
to the surface.
A big problem with multi-meter telescopes is the mirror starts to
deform under it's own weight. However, there are several methods
available now to create diamond in large amounts:
Large Diamonds Made From Gas Are The Hardest Yet.
http://www.sciencedaily.com/releases/2004/02/040226070311.htm
Bob Clark
Uncle Al
>
> Note that the glass used for large telescope mirrors is used to
> maintain their shape because of its strength and lightness.
No. Modern mirrors are membranes (look at diameter to thickness and
compare with a contact lens) with actuators underneath to dynamically
correct the wavefront. Zerodur (very good indeed) or titanium
silicate ULE (better).
> The
> reflectivity and smoothness comes from a thin layer of metal applied
> to the surface.
Almost atomically smooth by plasma polishing.
> A big problem with multi-meter telescopes is the mirror starts to
> deform under it's own weight. However, there are several methods
> available now to create diamond in large amounts:
Diamond has the CTE of Invar. It isn't nearly good enough for a
telescope mirror.
[snip]
--
Uncle Al
http://www.mazepath.com/uncleal/
(Toxic URL! Unsafe for children and most mammals)
http://www.mazepath.com/uncleal/qz.pdf
Boris Mohar
Large is a relative term. Ten millimeters is three zeros short but I will
take the shop floor rejects.
Regards,
Boris Mohar
Got Knock? - see:
Viatrack Printed Circuit Designs http://www3.sympatico.ca/borism/
void _-void-_ in the obvious place
kai-martin knaak
> Modern mirrors are membranes (look at diameter to thickness and compare
> with a contact lens) with actuators underneath to dynamically correct
> the wavefront. Zerodur (very good indeed) or titanium silicate ULE
> (better).
The actuators also compensate for fluctuations of the atmosphere. With
terrestrial telescopes even an ideally rigid mirror would have to be tuned
continuously.
>> However, there are several methods
>> available now to create diamond in large amounts:
>
> Diamond has the CTE of Invar. It isn't nearly good enough for a
> telescope mirror.
But diamond sets the record in thermal conductivity. A diamond mirror may
compensate for the bad CTE by exceptionally small thermal gradients.
Diamond: 2300 W/(m*K)
Invar: 13 W/(m*K)
Zerodur: 1.5 W/(m*K)
---<(kaimartin)>---
--
Kai-Martin Knaak
privat: kai-m...@familieknaak.de
dienst: km...@tem-messtechnik.de
gpg-key: http://pgp.mit.edu:11371/pks/lookup?search=kai-martin&op=index&exact=on
Mark Thorson
> But diamond sets the record in thermal conductivity. A diamond mirror may
> compensate for the bad CTE by exceptionally small thermal gradients.
>
> Diamond: 2300 W/(m*K)
> Invar: 13 W/(m*K)
> Zerodur: 1.5 W/(m*K)
Maybe the ideal mirror would be like corrugated cardboard,
with a flat Zerodur optical surface and a compliant diamond
backing to move the heat around.
John Savard
wrote, in part:
>Modern mirrors are membranes (look at diameter to thickness and
>compare with a contact lens) with actuators underneath to dynamically
>correct the wavefront.
Even in modern telescopes with adaptive optics, quite possibly only one
of the modern mirrors in the modern telescope fits that description.
John Savard
http://home.ecn.ab.ca/~jsavard/index.html
John Savard
<borism_...@sympatico.ca> wrote, in part:
>On 5 Dec 2004 16:08:00 -0800, rgrego...@yahoo.com (Robert Clark) wrote:
>>However, there are several methods
>>available now to create diamond in large amounts:
>>Large Diamonds Made From Gas Are The Hardest Yet.
>>http://www.sciencedaily.com/releases/2004/02/040226070311.htm
>Large is a relative term. Ten millimeters is three zeros short but I will
>take the shop floor rejects.
Ten millimeters is still pretty large for a diamond. But I suspect that
they're not _yet_ ready for mass production as something significantly
cheaper than natural diamond. Maybe for next Christmas season.
In the meantime, even at that size, perhaps they might find optical use
in microscope objectives, just as fluorite originally did.
John Savard
http://home.ecn.ab.ca/~jsavard/index.html
mme...@cars3.uchicago.edu
then way more than just telescope mirror manufacturing will be
revolutionized. We're still few orders of magnirude away from there,
though.
Mati Meron | "When you argue with a fool,
me...@cars.uchicago.edu | chances are he is doing just the same"
Michael Koch
> Almost atomically smooth by plasma polishing.
Is that the same as ion beam polishing? I thought that it
increases the surface roughness.
Michael
mjle...@search26.com
rgrego...@yahoo.com
The coefficient of thermal expansion of diamond and of Invar is still
very good. It's better than the Pyrex used in the Hubble 200" mirror
and comparable to the fused silica used in the Hubble Space Telescope.
But a key factor in diamonds thermal properties is its thermal
conductivity: it's about 1000 times better than Zerodur or ULE.
As this page discusses this is important in the time it takes for a
mirror to stabilize after a temperature change:
Glass For Mirrors.
http://www.oldham-optical.co.uk/Glass.htm
The largest practical single glass mirror now is about 8 meters wide.
The Young's modulus of diamond is about 5 times larger than the glass
used in telescope mirrors. Could we make a diamond mirror 40 meters
wide?
Bob Clark
Steve Taylor
>>Modern mirrors are membranes (look at diameter to thickness and
>>compare with a contact lens) with actuators underneath to dynamically
>>correct the wavefront.
>
>
> Even in modern telescopes with adaptive optics, quite possibly only one
> of the modern mirrors in the modern telescope fits that description.
>
I don;t think modern large scopes usually correct atmospheric effects
on the primary, its done with multiple element secondaries.
As I understand it, the Primary mirror IS actively supported to keep its
figure as it is pointed around the sky, but this is done at maybe 0.2 Hz
type rates, instead of the 25++ required for AO on the secondary, which
is correcting for atmospheric aberrations.
Steve
John Savard
<st...@astronomycentre.org.uk> wrote, in part:
>As I understand it, the Primary mirror IS actively supported to keep its
>figure as it is pointed around the sky, but this is done at maybe 0.2 Hz
>type rates, instead of the 25++ required for AO on the secondary, which
>is correcting for atmospheric aberrations.
Oh, yes, you're quite right about that; this is true, by mechanical
means, even for the Mount Palomar 200 inch.
John Savard
http://home.ecn.ab.ca/~jsavard/index.html
mme...@cars3.uchicago.edu
25 Hz rate. A tad too big for this.
jbuch
What an astounding engagement ring you could present.
Jim
--
...............................
Keepsake gift for young girls.
Unique and personal one-of-a-kind.
Builds strong minds 12 ways.
Guaranteed satisfaction
- courteous money back
- keep bonus gifts
Steve Willner
> maintain their shape because of its strength and lightness.
Glass is used because it can be polished to a smooth and accurate
finish. It is both weak and heavy compared to other structural
materials. Unfortunately other materials all have worse problems, at
least so far, although metal mirrors have been used in telescopes
from several hundred years ago until now (e.g., SST).
> reflectivity and smoothness comes from a thin layer of metal applied
> to the surface.
Reflectivity yes, smoothness no. The metal coating is typically
about 1/1000 of a wavelength thick, far too thin to affect the
smoothness.
> A big problem with multi-meter telescopes is the mirror starts to
> deform under it's own weight. However, there are several methods
> available now to create diamond in large amounts:
Diamond would be a great material if it could be produced in large
sizes and polished to an acceptable shape and finish. I don't expect
either production or polishing will be easy.
Silicon carbide has many of the same advantages as diamond. I
understand SiC mirror blanks have been produced in meter sizes, and
there are claims that people have polished mirrors, but I don't know
offhand of any examples in use. Supposedly a company in Russia was a
source of SiC mirrors, but I don't know which company or what their
capabilities might be.
--
Steve Willner Phone 617-495-7123 swil...@cfa.harvard.edu
Cambridge, MA 02138 USA
(Please email your reply if you want to be sure I see it; include a
valid Reply-To address to receive an acknowledgement. Commercial
email may be sent to your ISP.)
Ian Stirling
<snip>
> Diamond has the CTE of Invar. It isn't nearly good enough for a
> telescope mirror.
Thermal conductivity goes some way to counteracting this, it's what,
several hundred times better than glass.
John Savard
wrote, in part:
>Silicon carbide has many of the same advantages as diamond. I
>understand SiC mirror blanks have been produced in meter sizes, and
>there are claims that people have polished mirrors, but I don't know
>offhand of any examples in use.
The Europeans are orbiting a space telescope with one.
John Savard
http://home.ecn.ab.ca/~jsavard/index.html
John Savard
(John Savard) wrote, in part:
>On 8 Dec 2004 16:45:28 -0400, wil...@cfa.harvard.edu (Steve Willner)
>wrote, in part:
>
>>Silicon carbide has many of the same advantages as diamond. I
>>understand SiC mirror blanks have been produced in meter sizes, and
>>there are claims that people have polished mirrors, but I don't know
>>offhand of any examples in use.
>The Europeans are orbiting a space telescope with one.
I should have been more clear: preparing to orbit a space telescope with
one, the Herschel space telescope.
John Savard
http://home.ecn.ab.ca/~jsavard/index.html
mme...@cars3.uchicago.edu
Steve Willner
Thanks. Details at
http://sci.esa.int/science-e/www/object/index.cfm?fobjectid=34705
As far as I can tell, the shortest wavelength Herschel will observe
is 60 microns. This means the mirror polishing can be about ten
times worse than would be necessary for visible observations. That
doesn't mean better polishing can't be done, of course, only that it
is unlikely to be demonstrated by this telescope.
Robert Clark
Thanks for the response. I've seen different values for the Young's
modulus of diamond. The highest I've seen is 1200 gigapascals in the
[111] direction. I'll take this value. BTW, the modulus for silicon
carbide is 420 GPa.
So at 90 GPa for the Zerodur low exansion glass used in optics,
diamond is better by a factor of 13.3. Yan et.al. who demonstated the
high growth rate CVD method for diamond say their spectrospopic
measurements suggest the 50% increase in hardness extends through the
entire diamond, not just the surface. This suggests there should be an
accompanying increase in strength of 50%, to 1800 GPa. I'll take this
value. So this new CVD diamond is 20 times as strong as Zerodur.
This report compares the physical properties of some materials used
for mirrors:
Primary Mirror Substrate Materials for the XLT Telescope:
A comparison of various options including Silicon Carbide
http://www.hia-iha.nrc-cnrc.gc.ca/VLOT/Reports/XLT-SiC.pdf
On page 10 is given a formula for calculating the root-mean square
deflection for a mirror from its own weight, based on size, material,
and number of supports for the mirror. From this formula, we can
conclude it's proportional to Density*(1-Poisson's ratio^2)/Young's
Modulus. Then the deflection for diamond is smaller than Zerodur by a
factor of (3.52/2.52)*(1-.2^2)/(1-.24^2)*(1/20) = .071. It's also
smaller than the deflection of a SiC mirror by a factor of .26.
We can also see from the formula that if a mirror is scaled up by a
constant factor k in radius and thickness, then the deflection is
changed by a factor of k^2. Then since .2664^2 =.071, we can get the
same level of stability from a diamond mirror as a Zerodur one that is
.2664 times as big. So a diamond mirror 8*.2664 = 30 meters wide would
have comparable stability against deformation to a current Zerodur
mirror 8 meters wide.
A diamond mirror this size would be quite heavy. Note though that
assuming diamond material can be made in arbitrarily large sizes, then
for the support we could also use diamond pillars to support the
mirror. This would be several times stronger than steel for the weight.
It would also have the advantage that the thermal expansion for the
support structure would match that of the mirror.
In regards to increasing the size of the CVD grown diamonds, this
review article suggests the growth rate scales linearly with energy of
the microwaves used:
CVD Diamond - a new Technology for the Future.
"One of the great challenges facing researchers in CVD diamond
technology is to increase the growth rates to economically viable
rates, (hundreds of µm/h), or even mm/hr) without compromising film
quality. Progress is being made using microwave deposition reactors,
since the deposition rate has been found to scale approximately
linearly with applied microwave power. Currently, the typical power
rating for a microwave reactor is ~5 kW, but the next generation of
such reactors have power ratings up to 50-80 kW. This gives a much more
realistic deposition rate for the diamond, but for a much greater cost,
of course."
http://www.me.berkeley.edu/diamond/submissions/diam_review/review.htm
The paper by Yan et.al. discusses using a 6 kW microwave oven for
their CVD process:
Very high growth rate chemical vapor deposition of single-crystal
diamond.
PNAS | October 1, 2002 | vol. 99 | no. 20 | 12523-12525
http://www.pnas.org/cgi/content/full/99/20/12523
Using this they were able to get up to 150 micron/hour growth rates.
If the Yan et.al. process also scales linearly as other microwave CVD
methods, then a 6 megawatt microwave reactor would give a 150
millimeter/hour growth rate. So production of a 30 meter mirror would
require 30,000/150 = 200 hours, less than 9 days.
Another method would be to use several microwave ovens of the same 6kW
size used by Yan et.al. simultaneously, each working on its own seed
diamond. If we used a hundred of these we could get an equivalent total
size of a 30 meter mirror in 90 days.
As each segment approached the desired size, we would want them to
connect to form a single mirror. We could do this by sending a plasma
gas between two formed segments to get a single crystal diamond, just
as the original process forms a single crystal diamond on a single
surface. We would have to carefully match up the crystalline directions
in the separate segments so that the plasma could form a single crystal
consistently on both surfaces. We might insure this by cutting the
separate seeds from a single crystal.
The CVD method also makes it easier to form the shape of the final
mirror. We could cut the diamond seed(s) into the desired parabolic
shape and the CVD deposition would follow this shape. To get the fine
smoothing of the mirror surface, we could control the deposition of the
plasma using electrostatic or magnetic fields, as used for example with
Penning traps.
Another method might be to use laser deposition to get the final
mirror surface. This method produces polycrystalline diamond rather
than single crystal diamond, so it is not strong as the Yan et.al. CVD
method, but it allows finer control by directing the laser. However,
since this would be used to only deposit a thin layer on the top it
would not have to support much weight:
BRILLIANT DISCOVERIES
DIAMONDS ARE A PART'S BEST FRIEND
"A diamond coating breakthrough
"A major breakthrough in diamond deposition technology occurred when
Pravin Mistry, a metallurgist, was doing independent materials research
and consulting to industries requiring better tooling for metal forming
and extrusion. He was working towards fabricating hard materials using
lasers to synthesize ceramics and metal-matrix composites (MMC) on
aluminum extrusion dies to improve their performance and longevity. In
a fortunate misstep during laser synthesis of titanium diboride, Mistry
switched carbon dioxide for nitrogen and produced a coating speckled
with some black particulate inclusions.
"Analysis of the coating's surface indicated the presence of
poly-crystalline diamond. Retracing the steps of his experiment, Mistry
conceived a radical method for synthesizing polycrystalline diamond
films. The QQC Diamond coating process uses the carbon dioxide from the
atmosphere as the carbon source and subjects it to multiplexed lasers
to produce diamond film that can be deposited onto almost any
material."
http://www.advancedmanufacturing.com/November02/exploringamt.htm
Bob Clark
Robert Clark
Robert Clark wrote:
>
> ...
> The paper by Yan et.al. discusses using a 6 kW microwave oven for
> their CVD process:
>
> Very high growth rate chemical vapor deposition of single-crystal
> diamond.
> PNAS | October 1, 2002 | vol. 99 | no. 20 | 12523-12525
> http://www.pnas.org/cgi/content/full/99/20/12523
>
> Using this they were able to get up to 150 micron/hour growth rates.
> If the Yan et.al. process also scales linearly as other microwave CVD
> methods, then a 6 megawatt microwave reactor would give a 150
> millimeter/hour growth rate. So production of a 30 meter mirror would
> require 30,000/150 = 200 hours, less than 9 days.
> Another method would be to use several microwave ovens of the same
6kW
> size used by Yan et.al. simultaneously, each working on its own seed
> diamond. If we used a hundred of these we could get an equivalent
total
> size of a 30 meter mirror in 90 days.
Correction. If several 6 kw ovens were used to make separate segments
it would take more ovens than this or a much longer time. If 900 ovens
were used, you would you get 900 segments each 1 m wide. At 150
microns/hour this would take 1000 mm/.150 mm/hr = 6667 hours, or 278
days.
Bob Clark
Mark Thorson
> Robert Clark wrote:
> >
> > Another method would be to use several microwave ovens of the
> > same 6kW size used by Yan et.al. simultaneously, each working
> > on its own seed diamond. If we used a hundred of these we could
> > get an equivalent total size of a 30 meter mirror in 90 days.
> > ...
>
> Correction. If several 6 kw ovens were used to make separate
> segments it would take more ovens than this or a much longer time.
> If 900 ovens were used, you would you get 900 segments each 1 m
> wide. At 150 microns/hour this would take 1000 mm/.150 mm/hr
> = 6667 hours, or 278 days.
Oh, nuts. NOW you tell me. After I bought the ovens! :-)
John C. Polasek
wrote:
Don't want to get involved, but your extrapolation of 1/2.664 = 3.75
to get a 30 m mirror from an 8m is incorrect. The drooping is
proportional to the square of the diameter. The stiffness is
calculated using moment of inertia mr^2 etc. I believe you can only
go by square root = 1.93 to get a large mirror 15.5 m. Of course,
there's the saving in ovens :).
John Polasek
If you have something to say, write an equation.
If you have nothing to say, write an essay
John C. Polasek
<rgrego...@yahoo.com> wrote:
>I'm a little confused by what you quoted above. That was only used to
>say if you have 30m by 30m mirror (taken square for simplicity.) Then
>this could be made up of 900 segments each 1 meter wide.
>I assume you were actually referring to this earlier passage:
>
>"We can also see from the formula that if a mirror is scaled up by a
>constant factor k in radius and thickness, then the deflection is
>changed by a factor of k^2. Then since .2664^2 =.071, we can get the
>same level of stability from a diamond mirror as a Zerodur one that is
>.2664 times as big. So a diamond mirror 8*.2664 = 30 meters wide would
>have comparable stability against deformation to a current Zerodur
>mirror 8 meters wide."
>
>Several references give the deflection amount according to the
>material and size. Here's one:
>
>Mirror Structural Design.
>http://astron.berkeley.edu/~jrg/MirrorStructure/node1.html
>
>It shows the deflection is proportional to (diameter)^4/(thickness)^2.
>So if the dimensions are increased uniformly by a factor k, the
>deflection goes up by k^2. By replacing low expansion glass by diamond
>you want to see how much you can increase the size when taking into
>account diamonds greater strength ratios so that you don't incur
>greater deformation. You therefore take the square-root of the increase
>in material strength to see by what factor you can uniformly scale up
>the size of the mirror.
>
>
>
> Bob Clark
According to Marks' Handbook, the deflection of a circular plate of
stiffness Y supported at the edge and subject to a uniform pressure P
is
Defl = (r^2/t^2)*P/Y
But the gravity pressure goes up as the weight, g* rho*(area*t)/area =
rho*t
so Defl = (r^2/t^2)*rho*t/Y = (rho/Y)*(r^2/t)
From this if you double the radius the stiffness must go up by 4 and
you have to take into account the density rho of the material. Y/rho
is a sort of specific stiffness, being c^2 for a particular material
(m/l).
But you don't have to have the same deflection, you could allow
yourself a double deflection for doubling the radius, since
%error = defl/radius.
John C. Polasek
<rgrego...@yahoo.com> wrote:
>Can you give me the bibl. ref. for that handbook you mentioned? Every
>reference I've seen gives the deflection for a telescope mirror as
>proportional to (diameter)^4/(thickness)^2.
>But reading that passage again I'm chagrined to see I wrote "So a
>diamond mirror 8*.2664 = 30 meters wide would have comparable stability
>against deformation to a current Zerodur mirror 8 meters wide." Ack!
>Obviously if I'm saying the diamond mirror will be bigger I should
>*divide* by .2664! That equation should be 8/.2664 = 30.
>Bob Clark
It's Mechanical Engineer's Handbook, Lionel S. Marks, Fourth Edition,
McGraw-Hill pg. 477 Fig. 43. It's an old one.
I also made a mistake. He has two formulas, the first being for S
stress that I quoted you and next to it, the one for deflection being
proportional to r^4:
defl = r^4*P/t^3*Y
P = g*rho*(area*t)/area = g*rho*t
so
defl = g(rho/Y)*r^4/t^2
So in addition to picking out stiffness Y, you have to compare it to
its density rho. If they each increase by n%, you haven't gained
anything. Steel has (about) 28e6/7 and aluminum 9e6/3 so it's about
4:3 advantage for steel.
Sorry for giving you that scare. But if it's still diamonds you're
talking about, cash in the diamonds, buy a new Buick and retire! :)
Robert Clark
say if you have 30m by 30m mirror (taken square for simplicity.) Then
this could be made up of 900 segments each 1 meter wide.
I assume you were actually referring to this earlier passage:
"We can also see from the formula that if a mirror is scaled up by a
constant factor k in radius and thickness, then the deflection is
changed by a factor of k^2. Then since .2664^2 =.071, we can get the
same level of stability from a diamond mirror as a Zerodur one that is
.2664 times as big. So a diamond mirror 8*.2664 = 30 meters wide would
have comparable stability against deformation to a current Zerodur
mirror 8 meters wide."
Several references give the deflection amount according to the
Robert Clark
Can you give me the bibl. ref. for that handbook you mentioned? Every
reference I've seen gives the deflection for a telescope mirror as
proportional to (diameter)^4/(thickness)^2.
But reading that passage again I'm chagrined to see I wrote "So a
diamond mirror 8*.2664 = 30 meters wide would have comparable stability