ATLAS Upgrade dissimilar metal joining activities 1 Overview

ATLAS Upgrade dissimilar metal joining activities 1 • Overview • Stainless Steel to Titanium joining (temp connection) • Copper to Titanium (pressure drop measurements) Richard French, Paul Kemp-Russell: Sheffield Keith Birmingham: Aerobraze Europe Neil Austin: VBC Group Ltd (brazing division) Trevor Smith: Firmachrome Ltd Peter Cookson: Bodycote

Stainless VCR connector to Ti tube 2 Used to make temporary test fitting stubs for cooling circuits………… Method A = cheap idea • Electroless Nickel Plating. • • • Electroless Nickel coating is an alloy of nickel and phosphorous. Ability to work to close tolerances without post-plating grinding, whilst holding the original surface finish. Electroless Nickel can improve corrosion resistance, wear resistance, lubricity, solderability or be used to rectify and recover close tolerance undersize parts. The big advantage of elecroless (chemical) plating over electrolytic plating is it will adhere to Titanium, in a very controlled manner. We do not need to plate the entire circuit, just local areas as wherever the solution touches it will plate. Method B = proven but expensive • Vacuum Brazing using an aerospace proven method • Using a silver copper eutectic braze alloy, coat the Ti with the alloy, assemble the Stainless VCR fitting to the tube (post electro-polishing) and place in furnace at a lower temperature somewhere around 850°C. • http: //www. vbcgroup. com/focus/Brazing-Division/Brazing-Alloy-selection-tables/Braze. Precious. htm • Excellent joint, clean and pressure handling proven up to 250 bar. • 3. 175 mm Ti has heavy oxidisation that is proving difficult to remove. This is needed for the Cusil alloy to adhere to the Ti tube. • Once tube is cleaned (glass bead blasting) good adhesion is found. • Both methods are the “active” or direct joining of dissimilar materials with a braze filler metal (BFM). This is ideal for Ti as the BFM forms a strong permanent joint with the base materials. • What we did not realise is that at certain temperatures the Ti can suddenly start taking on alloying abilities with the BFM. This should not have happened when correctly controlled. • Titanium is a strong oxygen-getter, and thus will react with any oxygen that it can as it is heated from room temperature up to brazing temperature, therefore, "reacting" too early. • This is with free oxygen or water-vapor in the furnace atmosphere, or with metal-oxides on the metal surfaces during heat-up (such as when the metals are not properly cleaned prior to brazing), then the so called “brazing” (joining/bonding) of alloy-to-metal may be completely prevented from happening.

Method A • During this process, the bonding was be very successful but, the difference in thermal expansion between the 6 LV stainless steel – Nickel – Titanium which it is being joined caused premature cracking to develop in the brazed joint upon cooling. If we did not see the cracks at this stage they would present themselves during subsequent use in service. • It is very important to try to match the expansion characteristics of the metals to filler to metal joint, so that huge stresses in the joints are not built up. • During the furnace cycle (1100 C) something really odd happened. High temp was down to a miscommunication. • The braze alloy has a initial melting temp of around 400 C. The furnace temp was in the region of 1100 C. Therefore we managed to alloy Ni with Ti: 3 Stress cracking CTE mismatched Ni-Ti alloyed tube As the Ti reaches 600 C the oxygen in the Ti starts getting thirsty and in the resulting exchange drags the Ni into the microstructure. As the assembly cools, it falls apart as the CTE mismatch is beyond what the structure can cope with. GOOD NEWS – we don’t necessarily need to vac braze all our components and can do this with any induction furnace (have small tube furnace in lab ready). Ni plating works fine so will drop the cost of the heater block joining for pressure drop work.

4 Method B • For mechanical tolerance, achieve a good push fit in the Ti tube to VCR fitting. Micropolish fitting. • Using Cusil braze alloy from VBC simply clean components and plate the Stainless Steel component. 1/8” Ti to VCR • Mechanically clean Ti, chemically clean the Ti, assemble components and remember your nuts. • Place in furnace at 850 C and cycle once allowing time to cool. • Bingo – one joint. • Items of weirdness to note: • The VCR nut threads are silver plated to prevent gauling during assembly. This silver is reflowed during the furnace cycle. NEW REFLOWED 2. 275 mm OD Ti to VCR 1/8” Ti to VCR

5 Costs to date (A) Electroless nickel plating of Ti Cost of initial trials on 1 off previously de-scaled tube Background research £ 60 Materials £ 16 10 litre solution make up £ 13 processing trials £ 30 Sub total £ 119 Cost of initial trials pickling heavily scaled tube Background research £ 30 Materials £ 8 10 litre solution make up £ 13 processing trials £ 20 Sub total £ 71 Total to date £ 190 Immediate cost to incur in Feb Further cost of pickling trials (expected costs) Processing trials £ 100 If the trials are successful no further costs are expected for trial purposes Best guess prices for processing based on tube length SCALE UP COSTS inc bends upto 100 mm £ 3. 50 100 to 250 mm £ 5. 00 260 mm + £ 2. 50 per 100 mm electropolishing small stainless steel components @£ 0. 25 each This is the preferred method for joining the Cu heater blocks to the Ti and 316 L tubes as we can carry out the brazing in-house using our mini tube furnace that will cycle at 700 C to 1100 C so 850 C is attainable.

Costs to date (B) Cusil vac brazing of Ti 6 No costs incurred. All R&D free through good will. Braze alloy: 50. 00 gms @ £ 8. 86 per gramme 0. 9 mm Dia VBC ALLOY 4010 CUSIL BS EN ISO 17672: 2010 AG 272 V 1 £ 443. 00 Delivery Charge £ 15. 00 Total £ 458. 00 Brazing and assembly: Furnace cycle charged at £ 500 per cycle (we will process all parts in one cycle) £ 10 per joint for assembly and cleaning. Copper plating MOC of £ 359 (currently a 7 -10 day lead time) SCALE UP COSTS Preferred solution for manufacture ot Ti to VCR test stubs for welding. One furnace cycle will make enough parts for 2012 to 2014 when correct OD tube selected. 50 g of alloy equates to ~8 m length. 200 mm length completes 30 joints = 1200 joints possible per furnace cycle. = ~ £ 3500 for more than we need!