DC Spark Test System for CLIC Trond Ramsvik TS / MME 10 June 05

Outline Introduction Experimental Setup Field Emission Breakdown Field Motivation, Origin of Breakdown, Materials Experimental Setup Field Emission Breakdown Field Local Field Automatic Spark Conditioning Effect of Residual Gas Further work Conclusions Introduction Motivation, Origin of Breakdown, Materials Experimental Setup Field Emission Breakdown Field Local Field Automatic Spark Conditioning Effect of Residual Gas Further work Conclusions 10 June 05

Introduction: Why DC spark experiments? With a DC spark-test system, materials can be subjected to high electric fields, and their properties and responds to different treatments can be examined relatively easy and quickly. Goal: To find materials that withstand the highest field without breakdown or have low level of deterioration even when breakdown events occur. Need to understand the details of the breakdown phenomena 10 June 05

Introduction: Origin of Breakdown The physics of breakdown is still not perfectly understood Commonly accepted that breakdown at small gaps (d ≤ 0.5 mm) is initiated by an electron field emission based mechanism from one or a few microprotrusions* Process towards breakdown. Suggested Models: “Anode-initiated” electron bombardment of anode release of gas and/or anode material through intense localized heating or electron stimulated desorption (ESD) Avalanche ionization of the released species “Cathode-initiated” The microprotrusion on the cathode becomes unstable ohmic heating from high field emission current density fracture of the surface due to the tensile stress produced by the electric field Exchange Avalanche of mutual secondary emission of ions, electrons between the electrodes 10 June 05 *High Voltage Vacuum Insulation, edited by R. V. Latham (Academic, New York, 1995).

Introduction: Choice of materials Criteria: low vapor pressure high tensile strength high melting point high thermal conductivity high electrical conductivity Molybdenum & Tungsten 10 June 05

Breakdown Measurements Experimental Setup Sphere / Plane geometry HV supply 0 to + 12 kV UHV Sample Tip Q-meter Scope C Switch Breakdown Measurements HV supply 0 to + 12 kV UHV Sample Tip A-meter Field Emission Measurements 10 June 05

Field Emission Fowler – Nordheim plot: 2r r h E x ∆ V(x) ~ 3 GV/m M ~ 4.5 eV F B Metal Vacuum e- e- 0 Å 10 Å 5 Å  2r e h E r local Fowler – Nordheim plot: R. A. Millikan and C. C. Lauritsen, Proc. Nat. Acad. Sci. (US), 14, 45-49 (1928) R. H. Fowler and L. Nordheim, Proc. Roy. Soc. A 119 (1928) 173. 10 June 05

Molybdenum Tip and Sample Fowler – Nordheim plot Enhancement factor: Molybdenum Tip and Sample Emission Radius: 10 June 05

Enhancement Factor (β) Results from Fowler – Nordheim plots Molybdenum Tip and Sample Enhancement Factor (β) Emission Radius (re) [nm] Before bake-out: β ~ (40 ± 10) re ~ (2 → 11) nm After bake-out: β ~ (60 ± 10) re ~ (5 → 40) nm High uncertainty 10 June 05

Molybdenum Anode and Cathode Spark Cycle The voltage over the gap is increased until breakdown occurs Molybdenum Anode and Cathode Gap :~ 20 m 10 June 05

Field evaporation ↔ Tensile Strength Measured tensile strength of Mo*: Local Field Field evaporation ↔ Tensile Strength Elocal = β∙Eapplied Measured tensile strength of Mo*: 750 MPa Elocal ≈ 11 GV/m (max 21 GV/m) before-mentioned local field, which is the value one should use when discussing the tolerance of the materials itself. For example, there have been claims that breakdown can be triggered by field evaporation, i. e. microparticles being torn off the material, thereby contributing to the plasma eventually leading to breakdown. For Molybdenum we found that the avarage local field is appr. 11 GV/m with maximum value reaching above 20 GV/m, and using this simple relation between the tensile stress and the field, one finds that it corresponds to around 520 MPa and nearly 2 GPa. The tensile strength of this particular sample has been measured here at MME to be around 750 MPa. So, the conclusion is that there is a possible that field evaporation can occur. However, I will show that this cannot be the whole story behind the initialization of breakdown. tensile ≈ 520 MPa (max ~1950 MPa) 10 June 05

Automatic Spark Conditioning Single measurement up to breakdown Results of Spark Conditioning 10 June 05

Automatic Spark Conditioning Molybdenum (Mo) – Tip and Sample Spark Scan Histogram 10 June 05

Scanning Electron Microscopy Molybdenum surface after ~ 1600 sparks Average energy per spark: 0.8 J 14 m 1.3 mm Depth Profile / Inside Spot Outside spot Inside spot 8 m 8 m Local Melting smoothes out the surface Number of micro-protrusions strongly reduced 10 June 05

Electro Discharge Machined (EDM) Rolled / Chem. cleaned Mo EDM Mo 10 June 05

Gas Experiments From literature: Pressure effect on the breakdown field significant for large gaps Breakdown field deteriorates at pressures typically above 10-2 mbar R. Hackham and L. Altcheh, J. Appl. Phys., 46, 627-36 (1975) 10 June 05

Clear effect on pressure also at high vacuum / small gap (!) Gas Experiments Baked system Laboratory Air Unbaked system Dry Air Clear effect on pressure also at high vacuum / small gap (!) 10 June 05

The saturated breakdown field is reduced by ~150 MV/m Gas Experiments From ~1×10-9 mbar to ~ 1×10-6 mbar The saturated breakdown field is reduced by ~150 MV/m Damaged tip? 500 MV/m k ≈ 6000 MV/m (mbar)-1 n ≈ 1/4 10 June 05

Spark Conditioning vs. Pressure Rise Mo Tip & Sample Pav ~ 4×10-8 mbar At start: Follows the amount of energy deposited over the gap Slight decrease in pressure rise during spark Pressure rise typically in the order of a few percentage per spark 10 June 05

Comparison Molybdenum - Tungsten W: 10 June 05

Comparisons - 30 GHz measurements 420 400‡ Mo 340 350‡ W 260 170† Cu Max. surface field in RF [MV/m] (DC) E breakd sat OBS! Not saturated! DC and RF breakdown measurements give similar breakdown fields Superior behavior of both Mo and W with respect to Cu. †M. Kildemo, S. Calatroni and M. Taborelli, PRST-AB, 7, 092003 (2004) 10 June 05 ‡Saturation breakdown field measured at (2-4)×10-8 mbar

Future Work Installing new sampleholder and manipulator XYZ movements E-beam heating → ~ 1000oC “In-situ” annealing several samples more experiments to study the effect of residual gas on the Mo breakdown field pressure rise studies at UHV (~1×10-9 mbar) other materials: CuZr, Ti, Mo-Re alloys,W-films,… 10 June 05

Conclusions Surface preparation techniques influence the conditioning speed There is a residual gas effect on the breakdown field also for high vacuum and small gaps From 1×10-9 mbar to 1×10-6 mbar a reduction in the saturated breakdown field of ~150 MV/m is seen Molybdenum shows higher DC saturation breakdown field than Tungsten after extensive spark conditioning Similar to the results found for RF 10 June 05

Gonzalo Arnau Izquierdo Ahmed Cherif Ana Sousa E Silva Contributors Morten Kildemo Sergio Calatroni Mauro Taborelli Holger Neupert Gonzalo Arnau Izquierdo Ahmed Cherif Ana Sousa E Silva Alessandra Reginelli 10 June 05

Movement of electrons in RF Phase of RF field z [mm] t [ps] ω = 2π∙33×109 s-1 φ = π/2 q = 1.6×10-19 C E0 = 330×106 MV/m Me = 9.1×10-31 kg φ = 0 z0 = 0 10 June 05

Depth profile / ca. 100 sparks 10 June 05

After ~1600 sparks and exposed to air XPS After ~1600 sparks and exposed to air Carbon reduced Oxygen unchanged Molybdenum peaks stronger 10 June 05 Alessandra Reginelli TS/MME

Vickers ↔ Tensile Strength C. Bourgés Monnier, “Propriétés du molybd è ne et des alliages à base de molybdene”, Ecole des mines, Paris 10 June 05

10 June 05

Details of experimental setup HV supply 0 to + 12 kV HV switching electronics Q-meter UHV Tip Sample PC A-meter Scope 10 June 05