Thin Films for SRF What Might the Future Hold?

Charles Reece

Oct 4, 2010

2SRF Thin Film Wkshp – Oct 2010 cer

SRF Thin Films• The principal driver for pursuit of materials and surfaces with

vanishingly low surface resistance to rf fields (SRF) comes from the particle accelerator field.

• So, what serves the customer’s needs?

• Low losses and high fields – of course

• Really, it is low integrated system cost per unit acceleration

{Resonator system, cryostat, rf drive, cryo system }

• The great majority of future applications for SRF will be CW

• Realizing lowest Rs is imperative

• Only with low Rs will high fields become usable.

• So what are the credible opportunities for SRF thin films to make dramatic impact?

3SRF Thin Film Wkshp – Oct 2010 cer

The Stage for SRF Thin Films

• For the customer, only the net price of each unit of acceleration matters. The rest is “just details.”

• As a heuristic tool to guide us in dreaming not too far off the real plane, I create a (simplified) model which includes the key cost contributions for providing 100 MV of CW linac suitable for accelerating a ~ 1 mA beam.• For several applications, this is either the approximate need or a

suitable building block for much larger systems.

• I seek to link RF surface resistance, accelerating gradient, and cryogenic and rf capital and 10-year operating costs required to provide 100 MV acceleration. Increasing beam current would mainly increase RF costs.

• So where are the opportunities for SRF thin films to create a significantly different future?

• Let’s dream of being limited by BCS losses:

4SRF Thin Film Wkshp – Oct 2010 cer

The Landscape @ 1300 MHz

BCS surface resistance

BCS Rs calculated by adaptation of Halbritter program my Gigi Ciovati

It’s not yet clear where MgB2 goes on this plot. Stay tuned.

5SRF Thin Film Wkshp – Oct 2010 cer

Cost Model for 100 MV• Cost100 MV (f, T, Eacc, Rs(mat’l), G, R/Q)

• I take the JLab upgrade cryomodule zone as a baseline from which to scale.

• Model assumptions:• Real estate is free.• Static cryo load is linear with active length.• RF power is constant cost, controls and distribution cost scale linearly

with length.• Cryomodule cost is proportional to active length, i.e. once we know

what to do, doubling the accelerating gradient costs negligibly more.• Cryo capacity designed to target load. • Cryo capital cost and operating efficiency is per JLab expert model. (Rao

Ganni)• Rs is BCS limited• Cavity geometry fixed to clarify scaling patterns (would be tailored to

specific application).

6SRF Thin Film Wkshp – Oct 2010 cer

Net Cost of 100 MV @ 2 K

Surface resistance has a dominant influence on system cost

7SRF Thin Film Wkshp – Oct 2010 cer

The Landscape @ 650 MHzBCS surface resistance @ 650 MHz

BCS Rs calculated by adaptation of Halbritter program my Gigi Ciovati

It’s not yet clear where MgB2 goes on this plot. Stay tuned.

8SRF Thin Film Wkshp – Oct 2010 cer

Net Cost of 100 MV @ 4.2 K

Conceivable economic opportunities for future SRF films at 4.2K

9SRF Thin Film Wkshp – Oct 2010 cer

100 MV Dream SolutionsSeven Dream Objectives - caveat emptor (cost/incremental 100 MV)

R1 – 20 MV/m @ 1497 MHz, 2.07 K bulk Nb – (10M$) CEBAF UpgradeD1 – 40 MV/m @ 650 MHz, 2 K via Nb Film – (5.5M$)

Cheapest Nb solution, ~half the price of CEBAF acceleration D2 – 20 MV/m @ 1300 MHz, 2 K, via NbN SIS Film – (5.9M$)

1st incremental application of SIS, 70% reduction in dynamic loadD3 – 50 MV/m @ 1300 MHz, 2 K, via Nb3Sn SIS Film – (8.9M$)

Compact FELs ? ILC ? 4GLS ?D4 – 20 MV/m @ 650 MHz, 4.2 K , via NbN SIS Film – (7.1M$)

Driver for ADS?D5 – 20 MV/m @ 650 MHz, 4.2 K, via Nb3Sn thick or SIS Film – (5.2M$)

Cheaper driver for ADS?D6 – 60 MV/m @ 1300 MHz, 4.2 K, via Nb3Sn SIS Film – (4.1M$)

Cheaper ILC ? Commercial THz sources? Industrial processing?D7 – 105 MV/m @ 650 MHz, 4.2 K , via Nb3Sn SIS Film – (4.2M$)

Compact, portable interrogation systems ?

Add MgB2 ? on the map?