Sergio Calatroni, Fritz Caspers, Hugo Day, Emissivity Studies for the LHC Injection Kickers M.J. Barnes, Z. Sobiech and W. Weterings 21/01/2014 Acknowledgements: Sergio Calatroni, Fritz Caspers, Hugo Day, Marco Garlasche, Mauro Taborelli, Yves Thurel M.J. Barnes Radiative cooling of ferrite – 21/01/2014

LHC Transmission Line Kicker Magnets LHC Injection Kicker Magnets 4 MKIs in RA87 MKI8D MKI8B MKI8A MKI8C LHC Injection Kicker Magnets Kicker magnets for injecting beam into LHC; 4 magnets at each of Point 2 and Point 8; 2.7m long magnet with FERRITE yoke; Ferrite Curie temperature ~130°C; Baked out to 300°C to be ultra-high vacuum compatible (~10-11 mbar in tank); Despite shielding between beam and ferrite, yoke can heat significantly during physics and delay re-injection; Further improved shielding not presently feasible; High voltage: P2  49.6kV PFN. P8  51.3kV PFN. M.J. Barnes Radiative cooling of ferrite – 21/01/2014 02/10/2012 M.J. Barnes 2

Ferrite Temperature (˚C) MKI Tank: Emissivity Courtesy: M. Garlasche and Z. Sobiech Ferrite yoke Tank Capacitor Emissivity of inside of MKI tank Ferrite Temperature (˚C) 45W/m Beam induced power deposition results in heating of the ferrite yoke: The MKI magnet is in vacuum: cooling of the ferrite yoke is mainly due to thermal radiation; Thus the emissivity of the inside of the MKI tank greatly influences the ferrite yoke temperature, for a given power deposition. Electro-polished MKI Tank The tanks housing the MKI magnets are reused from the LEP accelerator: they were used for 300 kV electrostatic separators and thus were electro-polished  low-emissivity. A method of measuring emissivity was required. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Example of MKI8 Measured Temperatures MKI8D: 19 screens. Jacket MKI8D: 15 screens. No jacket from TS2, 2012 MKI8B: no jacket from TS2, 2012. MKI8A: 24 screens (9 short) Initial permeability of CMD5005 increases to a max. at ~100°C, then starts to rapidly reduce. Δ1.4ns Δ2.8ns  Δ3.8% inductance @ 0.74ns/% (~Δ1% of kick for 4 magnets). Δ3.8%  relative permeability of ~100 for ferrite. ~130˚C actual? As of TS3: MKI8B does NOT have a bake-out jacket. MKI2B & MKI2D do NOT have a bake-out jackets. Bake-out jackets removed from MKI8B & MKI8D during TS2, 2012. MKI8D exchanges for new magnet, with jacket, during TS3, 2012. Bake-out jacket of MKI2B & MKI2D removed 11/09/2012 ! Δ2.8ns Radiative cooling of ferrite – 21/01/2014 M.J. Barnes

Summary of Expected Power Depositions in Ferrite Yoke Courtesy: Hugo Day > Curie Temperature Post-LS1 the expected power deposition in the ferrite yoke, with 24 screen conductors, is ~55W/m (based on the worst-case measurement of 6 MKIs) --- i.e. similar to pre-LS1 deposition levels for “normal” (not MKI8D pre-TS3) MKIs. Based on the above, the post-LS1 ferrite yoke temperatures are expected to be similar to (slightly less than) the pre-LS1 temperatures (except MKI8D pre-TS3, 2012): thus the ferrite yoke will be below the Curie temperature. For the HL-LHC 25ns parameters assumed the expected power deposition in the ferrite yoke, with 24 screen conductors, is up to ~190W/m (based on the measurement of 6 MKIs) --- i.e. 20% greater than MKI8D pre-TS3. Based on the above, if no additional cooling is achieved, the post-LS1 ferrite yoke temperatures are expected to be greater than MKI8D pre-TS3, 2012: thus the ferrite yoke will be pass the Curie temperature, limiting LH-LHC operation. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Radiative cooling of ferrite – 21/01/2014 Predicted Temperature of Ferrite versus Power, Emissivity and External Cooling Curie temperature (24 conductors) Post-LS1 (15 conductors) Pre- LS1 (15 conductors with 90˚ twist) Pre- LS1 [MKI8D] M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Emissivity measurements: tank heated externally to 65˚C Procedure for measurement: Measure inside of tank with temperature probe; (2) Adjust emissivity of IR camera to give same temperature reading. Could see our own thermal reflections – emissivities of ~0.6 to ~0.9 measured. Emissivity with 50˚C external was 0.7 to 0.75. Actual emissivity probably ~0.15 ! This method of emissivity measurement provided unreliable results. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI Tank: Emissivity Measurement Heat lamp (Ts) Infra-red camera (Tm) Surface of sample under test (Tamb) Courtesy: M. Garlasche and Z. Sobiech An infra-red camera and heat lamp did not give reliable measurement, e.g. due to concave surface of tank and high reflectivity. A new method of measurement has been developed where 4 conductors, housed in a ceramic tube, are heated and the temperature of the tube is measured with both an IR camera and PT100 sensors: M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI Tank: Treatment and Results A sample SS flange, treated using ion-bombardment, had a post-treatment emissivity of 0.6 ! Thus MKI tanks were treated using the same method, and the emissivity measured before and after treatment: During ion-bombardment (in an Ar-O2 atmosphere) MKI Tank after Ion Bombardment Tank no. 13 8 11 7 6 10 Comment treated outside untreated treated properly 380ᵒC for 10hrs 350ᵒC for 7hrs 387ᵒC for 10hrs Emissivity 0.13 0.12 0.14 0.15 Conclusion: no real gain in emissivity following ion-bombardment. Hence routine treatment of tanks cancelled. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI Tank: Confirmation of Tank Emissivity Courtesy: M. Garlasche and Z. Sobiech The emissivity of the tank greatly influences both the rate of heating and cooling of the ferrite yoke. Transient thermal simulations of an MKI have been carried out and predicted temperatures compared with measured temperatures during oven bake-out. The predictions confirm the emissivity measured using the new setup. Tank no. 8: emissivity measured with new setup  0.14 Studies for improved emissivity and/or liquid cooling are ongoing. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI Tank: New Emissivity Technique with an IR Camera Tank no. 2 Untreated Method With ceramics tube IR camera (held in approximate centre of tank) Emissivity 0.15 0.1 Sticky tape to measure reference material e=0.95 Method Cover most of inside of tank with paper to reduce internal reflections; Place sticky tape on inside the tank; Heat up tank locally, from outside; Use IR camera to measure temperature of sticky tape and adjacent tank (with “close-up” lens, down to 79mm distance) – beware of angle of camera w.r.t. tank surface!; Measure reflected temperature in same area as sticky tape (crumpled Al foil); Post-process with “FLIR Tools”: adjust emissivity value so that indicated temperature of tank, immediately next to the sticky tape, is equal to the indicated temperature of the stick tape. Paper cover to reduce internal reflections Future improvements: verify emissivity of tape; use tripod for IR camera. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Improving Emissivity of SS Components 12 The emissivity of clamps and corona shields was successfully increased by oxidation in air: 950ᵒC Average emissivity of samples after treatment at 900ᵒC [Courtesy: Sergio] Untreated clamps Untreated corona shields After treatment, in air, at 800˚C After treatment, in air, at 900˚C Measured emissivity was not significantly improved by oxidizing at 800ᵒC but a significant increase after treatment at 900ᵒC. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Radiative cooling of ferrite – 21/01/2014 Conclusions High emissivity (≥0.6) of the inside of the tank would provide acceptable radiative cooling of the tank: Ion bombardment of an electropolished flange gave an emissivity of ~0.6; Ion bombardment of the inside of the MKI tank, at ~380˚C, did not have any significantly influence on the emissivity; Oxidation of stainless steel components at 900˚C increased emissivity to ~0.6. BUT any coating (e.g. carbon) of the tank must be stable, not peel, and an electric field must not “pull-off” dust particles and must be of a reasonable thickness (see talk by Sune Jakobsen). Emissivity of the inside of the tank is difficult to measure because of high reflectivity (low emissivity). New techniques have been developed to make the measurement: With a heated ceramic tube whose temperature is measured and compared with analytical calculations and FEM predictions; With an infra-red camera, external heating, and appropriate lining of the tank, e.g. with paper, to reduce the reflectivity of surfaces not being measured; These results of these new techniques have been validated by comparing with transient thermal measurements, during an oven bake-out, and comparison with transient FEM predictions. High Curie temperature ferrites are under consideration – but beware of high vacuum pressure at elevated temperatures! M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Radiative cooling of ferrite – 21/01/2014 Spare Slides M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Summary of Expected Power Depositions in Ferrite Yoke Courtesy: Hugo Assumed Beam Parameters: Pre-LS1 operation - 1380 bunches, bunch separation 50 ns with an average bunch population of 1.6e11 ppb and an average bunch length of 1.2 ns. Post-LS1 operation - 2808 bunches, bunch separation 25 ns with an average bunch population of 1.15e11 ppb and an average bunch length of 1.0 ns (i.e. nominal LHC parameters). HL-LHC 50 ns - 1380 bunches, bunch separation 50 ns with an average bunch population of 3.5e11 ppb and an average bunch length of 1.0 ns. HL-LHC 25 ns - 2808 bunches, bunch separation 25 ns with an average bunch population of 2.2e11 ppb and an average bunch length of 1.0 ns. a: Did not limit LHC operation: maximum measured temperature 53˚C. b: Occasionally necessary to wait for yoke to cool-down (yoke approaching Curie temperature), maximum measured temperature 63˚C M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI2 Temperatures: 01/07/12-23/10/12 MKI2B & MKI2D: No jackets As of TS3: MKI8B does NOT have a bake-out jacket. Bake-out jackets removed from MKI8B & MKI8D during TS2, 2012. MKI8D exchanges for new magnet, with jacket, during TS3, 2012. Bake-out jacket of MKI2B & MKI2D removed 11/09/2012 ! All magnets with15 screen conductors: maximum measured temperature ~55˚C. Analysis of “SoftStart” data does not show any signs of having reached Curie Temperature of Ferrite yoke. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI8D Removed from LHC during TS3, 2012 Capacitively coupled end of beam screen (and outline of U-ferrite) Directly connected end of beam screen (and outline of U-ferrite). Highest heating at downstream end, where LHS ferrite leg is unshielded from circulating beam, because of twisted stripes. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Near-Field Effects Classical radiation regime C.M. Hargreaves, “Electromagnetic proximity effects and their consequences for radiation shielding”, Fifth Cryogenic Engineering Conference. Plot shows: Red area: classical radiation regime, for spacing between black surfaces greater than one wavelength. Blue area: energy transfer with both evanescent waves AND classical travelling waves. One can describe this phenomenon as infra-red tunnelling – which works well even between shiny surfaces where classical heat transfer by radiation would be very low (due to low emissivity). There is a smooth transition between radiation and contact regimes – by applying mechanical pressure, size of the gap is simply reduced. “Contact” 6 November 2012 LRFF Meeting: M. Barnes & F. Caspers

Curie Temperature for various NiZn Ferrites from Ceramic Magnetics CMD5005 used in spare MKI pre-LS1 CMD10 looks interesting (higher Curie Temperature (Tc) and saturation flux-density (Bs) than CMD5005, but lower u’) CMD5005 CMD10 M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI Tank Pressure Versus Ferrite Measured Temperature To limit pressure, due to out-gassing, to 5E-9 mbar “measured” temperature should be limited to ~150˚C (>150˚C actual) Extrapolation of fitted trendline, during cool-down of ferrite yoke of MKI8D Oct. 2011, following a beam dump, gives: 1E-9 mbar at 100˚C measured; 2E-8 mbar at 200˚C measured; 7E-7 mbar at 315˚C measured (note: at the end of a 72 hour, 315˚C, plateau, during MKI oven bake-out, the measured tank pressure is ~2E-5 mbar – hence these extrapolations may be optimistic! A higher Curie temperature ferrite, combined with good cooling, would give headroom during HL-LHC operation. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

MKI8BCD: Temp_Tube_Up. May 1, 2012 to 21 Nov. 2012 TS#3 MKI8C_TEMP_TUBE_UP MKI8D_TEMP_TUBE_UP MKI8B_TEMP_TUBE_UP Maximum value of Temp_Tube_Up of MKI8B, since May 1st, has been fairly consistent; Maximum value of Temp_Tube_Up of MKI8D, since May 1st, has been fairly consistent, but is slightly lower for the new MKI8D (changed TS#3); Maximum value of Temp_Tube_Up of MKI8C, until TS#3, was generally < 120˚C, with one excursion to ~150˚C on 27/08. Since TS#3 the temperature has approached 180˚C at times (but not always). M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Position of PT100 Probes, pre-LS1 M.J. Barnes Radiative cooling of ferrite – 21/01/2014

PT100’s new position, post-LS1 The MKIs upgraded during LS1 will have the “ferrite yoke PT100’s” moved from the end ground plate to the side-plates. Advantages include: The PT100’s on the side plate will give a measurement temperature more representative of the ferrite yoke temperature. Ferrite yoke PT100’s will be distant from the ferrite toroids and should thus not be directly influenced by a high temperature of the toroids. M.J. Barnes Radiative cooling of ferrite – 21/01/2014

Radiative cooling of ferrite – 21/01/2014 2D analysis: results From presentation 11/07/2012 Max Ferrite Temperature sensitivity to external convection coefficient @ different power levels Temperature at tip of tank P [W] α [W/m2K] εTANK TFERRITE [°C] TTANK [°C] 150 1.9 0.1 113 40 2.9 110 34 5 107 29 7 106 27 25 105 23.4 100 88 86 30 84 26.7 83 0.3 66 25.3 72 Courtesy: M. Garlasche Case 100W 2.9W/m2K Main parameters: Power (…) and emissivity; not conv. coeff. M.J. Barnes Radiative cooling of ferrite – 21/01/2014