Subject: CTE non-uniformity of Hextek Blanks


Submitted: Tue, 16 Jul 1996 11:45:36 -0700

Message number: 75 (previous: 74, next: 76 up: Index)

Concern had recently surfaced over the thermal properties of Hextek blanks.
Please read Ed Mannery's summary of his investigation into this current
generation of rumors.

Ed's note reminded me of some items I have in a SIRTF file, and I found a
reference for the AMES paper (Proc. SPIE 571 (1985) pp 101-114 by Melugin,
Miller, Angel, Wangsness, Parks, & Ketelsen).  In developing lightwt blanks
for LDR (Large Deploy. Reflect.) the group built a 38-cm borosilicate
(Pyrex) mirror using the "sand-hexing" process, polished it, and looked at
the change in optical figure between temperatures of 298 K and 79K.
Surface figure distortions of 40 nm rms were seen with the 219 deg delta T.
No information was present regarding subsequent measurements to indicate
hysteresis.  I have not looked into whether borosilicate glass cte is
quasi-linear over this regime, but it appears to first order that optic
surface quality variations with temperature should be negligible over the
thermal range we process and use these materials.

Note also that this was done with the older Corning material, which, as Ed
describes, Rich has seen to be less well matched as the Schott material
which he is now using.

From Ed:
>The "rumor" that surfaced at the July 8 APO user's meeting came to Alan from
>Jim Kimbrell at Contraves who apparently heard it from someone else.  The
>story goes that when you cool a Hextek mirror quilting appears.
>Wortley buys Tempax for all his mirrors and only makes sure that he orders
>tubes and plates that have been made from the same glass reciepe.  He makes
>no special attempt to buy plates from the same glass batch nor tubes from a
>common batch.  His method of demonstrating adequate CTE matching is by
>bonding a tube to a plate and annealing it, then inspecting the joint using
>crossed polaroids.  He tests each batch of material he receives in this way.
>Serious problems were encountered matching Corning tubes with Corning sheets
>in early days of Hextek experimenting so Corning glass was abandoned.  He
>now uses only Schott Tempax and rarely detects any optical effects in this
>test that would indicate serious CTE non-uniformity.  Rich's information is
>quantitative in the sense that no observed colors in crossed polaroids puts
>an upper limit on the frozen in stress.  A discussion of the interpretation
>of polaroid observations and the actual values of stress is available but
>not from me.
>Hextek blanks have been sold to customers who had them figured and used or
>are using them at cryogenic temperatures, apparently successfully each in
>its own context.  (might this be relevant re: coating vendors who must heat
>the mirror).  According to Worelty one customer (Ames) actually tested a
>small Hextek mirror at cryo temps and published results in SPIE.  I have not
>yet tracked this down.  Wortley knows of no other reported experience or
>measurements which attempt to compare the performance of  Hextek blank at
>different temperatures.
>Finally, lets please do what we can to prevent propagating a new unfounded
>rumor that there is something wrong with Hextek blanks.  ANY piece of glass
>has a spatial structure of variations in CTE which will excite changes in
>shape when the temperature of the blank is changed.  If we need to know
>anything about Hextek blanks, it is not that they DO this, but how much do
>they change shape with temperature and how much can we tolerate.  I'm sure
>we considered all this before choosing a Hextek blank but I can no longer
>reconstruct the line of investigtigation or reasoning behind it.  It
>undoubtedly involved Roger Angel at the time.  If it makes one feel warmer
>and fuzzier, the MMTO has recently taken delivery of a Hextek blank for use
>as a secondary in the MMT upgrade.
>em 12 July 1996

APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 75 in the apo35-general archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to