Filament Building

 

Hello Again, 

 

            A few more weeks have past since my first actual coating, and the resulting small disaster, which took just 4 days to regrind starting back at 60 grit, polish and figure with the help of Paul Kreisle from our ATM lab.  BTW if you’ve never had the pleasure of “Tag Teaming” a mirror, you should find someone to try it with.  Having two people taking turns grinding and polishing on a mirror makes quick work of the process and gives incredible results!

 

            During the process of “Erasing” my mistake on the mirror, I was dreaming up ways to prevent Al from dripping on the mirror, as well as optimize the coating thickness for uniformity across the mirror.  My initial design for the tungsten filament was just a piece of tungsten used for TIG welding, placed at the top and center of the chamber.  This resulted in two problems.

  1. The filament was too close the mirror surface to give uniform coating thickness.
  2. Since my chamber does the mirrors face up, in the event some Al should drip off the filament, it caused sever damage to the mirror surface.

 

My first thought to prevent damage to the mirror, was to add a shutter, to protect the mirror while the Al was heated to evaporating temperature.   Since I had a vacuum rated mechanical feed thru, I figured this wouldn’t be too difficult of a challenge to overcome.  So I started thinking about how to optimize the evaporator layout to get the most uniform coating possible.  Obviously a single evaporator on center won’t achieve this goal, so several evaporators in a ring were the obvious answer.  The problem was, what diameter ring, and how far above the mirror?  Time for more research J

 

Good old Google showed me the way, actually it was a site on the Large Binocular Telescope coating methodology written by John Hill.  Once I understood the process, which was pretty simple, I put together a brute force program that allowed me to play with the variables and display the results.  Below is a screen shot of the resulting best-fit evaporator layout.

 

 

            The program created some interesting results!  First It showed that the best layout resulted from putting the evaporative filaments out near the edge of the chamber.  Secondly it showed that the best coating could be obtained by having a relatively short distance between the mirror and the filaments.  In the above screen shot, the simulation was ran with a 6 point filament radius of 140mm (11 inch diameter) and only 130mm above the surface of the mirror.  Suddenly not only was my chamber of sufficient height, but with the filaments out near the edge of the chamber, I no longer needed to worry about adding a shutter to the system, as any drips of Al would simply miss the mirror! Whooo Hoooo J 

 

            Another interesting side effect of this layout was I could change the distance between the mirror and the evaporators, which would change the coating layout, either adding additional thickness to the center, midzone, or edge of the mirror.  In effect this meant I could “Tweak” a mirror’s figure by roughly 6nm with a 100nm thick coating.  This isn’t much of a change, but is enough to push a 1/6 wave PV mirror towards 1/10 wave PV, In theory at least!   (If interested you can play with the program here: http://lerch.no-ip.com/atm/nearfieldsim.zip (7kb))

 

            So, all exited I went about the mental exercise of actually building this system into my chamber.  Here are the results:

 

Click for 640x480 image @ 60Kb

Click for 1024x768 image @ 141Kb

Click for 2560x 1920 image @ 727Kb

 

Click for 640x480 image @ 59Kb

Click for 1024x768 image @ 143Kb

Click for 2560x 1920 image @ 753Kb

 

Here’s the details on the construction:

 

1.      Drilled 6 holes around the perimeter of the chamber right before the top of the chamber starts to curve in.

2.      Brazed 5 brass 6-32 x ½” long bolts into place, and one 2” long bolt for a ground point.

3.      Found Purchased 10 6-32 x 1” long ceramic stand offs from http://smallparts.com (Very cool site!)

4.      The bars are ¼” square cold rolled steel from Home Depot

5.      After two failed attempts to drill and tap the ends of the steel bars for 6-32 threads, I gave up and just brazed a brass nut to the back side of the bar (sure wish I’d thought of that before breaking two $5 taps!

6.      The tungsten filaments are held in place with a stainless steel wash and bolt (first attempts were to drill holes into the end of the steel bars, and use the bolt to pinch in place, and after many broken drill bits, I just gave up and used the bolt and washer method!)

 

All that was left to do now, was to create some filaments, which wasn’t as easy as I thought!  My first try involved trying to bend a tig rod around a jig I made.  90% of the time the tig rod would just break into multiple pieces.  Figuring the Thorium in the TIG rods was the culprit I purchased some pure tungsten rods from http://smallparts.com (5’ section of 0.023” was only $10)  This material worked a little better,  only 50% of my attempts would break into multiple pieces, Damn tungsten in BRITTLE stuff!

 

I sent a call for help to rec.crafts.metalworking (another amazing group of people!)  As usual a solution was presented.  I was told to find the “Brittle to Ductile Transition Tempature” of Tungsten, then heat the Tungsten to this temp prior to trying to work it.  Considering how brittle Tungsten is, I had nightmares of never being able to make my own filaments!  So after some research I found that the BDTT of Tungsten is an amazingly low 100 – 400C!  As an experiment I took one of the TIG rods, heated it with a 1200watt heat gun and an “I’ll Be” happened!  The previously brittle tungsten was now easily coiled around a screw driver!  Whoda thunk it J

 

After doing some thinking, I realized a coiled tungsten filament was probably overkill for my application, and a simple loop filament was probably best.  So I modified a pair of pliers to create a forming jig, cut the Tungsten into 2” pieces, placed in the pliers, heated gently, then squeezed, here are the results:

 

Click for 640x480 image @ 52Kb

Click for 1024x768 image @ 134Kb

Click for 2560x 1920 image @ 572Kb

 

 

The Bottom piece of Tungsten is a standard Tig rod, which unheated snapped into three pieces in my forming tool!  The top pieces are from the 5’ section from smallparts.com.

 

Here’s an image of the tool I built to form the tungsten filaments (Probably should have wire brushed it prior to the picture, but oh well J )

 

Click for 640x480 image @ 51Kb

Click for 1024x768 image @ 113Kb

Click for 2560x 1920 image @ 572Kb

 

            Now that I had some filaments to play with, I went about running a bunch of test coating on some scrap window glass.  My first lesson was that its possible to get to a good vacuum (2x10-5torr) while having WD-40 hanging out on the ends of my filament clamps (left over from machining)!  So my first test piece came out really WEIRD as the WD-40 quickly evaporated along with the Al, Ooops J

 

            After a good cleaning, I ran some more tests and learned the following:

1.      The Tungsten filaments are good for 1 – 2 runs, after which they fracture.  The good part is the fracture is an un-eventful event, meaning the filament just blinks off with no hot bits flung off!  The down side of course is I loose half my evaporative ring!  So I now just replace the filaments on each coating run.  Total cost of filament replacement is ~$2.00, which I can live with!

2.      There is a relationship between the diameter of the filament and the amount of Al it can hold!  Here’s what I observed

a.      If you get Greedy, and load too much Al onto the filament, the filament will fracture and little to no Al will be deposited.

b.     It took some experimenting to figure out why it did this, and the results was:

                                                              i.      Tungsten and Al alloy together when heated

                                                            ii.      The alloying is a good thing, as this causes the Al to wet the tungsten

                                                          iii.      During the alloy process some tungsten dissolves into the Al

                                                          iv.      As the Al evaporates, the dissolved tungsten returns to the filament

                                                            v.      If too much Al is hanging on the filament, this causes too much tungsten to dissolve off the filament, causing a thin spot that rapidly overheats and fractures!

 

Well that pretty much brings us to a close for this page, any questions feel free to contact me at Jlerch@tampabay.rr.com

 

Take Care,

James