Characterizing thin films by RF and DC methods

Tobias Junginger

Outline• Cavity measurements

– Example: Nb on Cu (HiPIMS)

• Sample measurements with the Quadrupole Resonator– Example: Nb on Cu (ECR)

• Point contact tunneling– Example: Nb on Cu (HiPIMS)

• Muon Spin Rotation – Example 1: Nb on Cu (HiPIMS)

– Example 2: NbTiN/Nb

T. Junginger - Characterizing thin films by RF and DC methods 2/17

Cavity measurements

• The Q vs E measurement of an accelerating cavity is a critical milestone for any thin film development

• Sputter coated Nb films on Cu cavities have been used in the past for LEP-II and currently for LHC and HIE-Isolde

• In the 1990 the dcMS technology has been investigated in depth for 1.5 GHz cavities

• More recently energetic condensatio techniques such as HiPIMS are investigated as an alternative approach to overcome the current limitation of this technology, i.e. the field dependent residual surface resistance

T. Junginger - Characterizing thin films by RF and DC methods 3/17

HiPIMS: Motivation

T. Junginger - Characterizing thin films by RF and DC methods 4/17

By applying pulses of high power the sputtered target material atoms are ionized and applying biasing potential we can achieve:• Target material ions can be accelerated towards the substrate, higher kinetic energy upon

arrival• Ions are directed to the surface, thus non-flat surfaces can be sputtered with good

uniformity of the film

Rs(B) depends on the ion impact angle

352 MHz

HiPIMS: Status at CERN

T. Junginger - Characterizing thin films by RF and DC methods 5/17

1μm

Magnetron HiPIMS

• So far only tests without biasing• SEM images show more ordered surface

structure• RF performance equal to dcms with same

surface preparation • Biasing could be the key

• RRR of about 20-30

dcMS: 1.5 GHzHiPIMS: 1.3 GHz

Electron Cyclotron Resonance (ECR) at JLAB

• Another technique to create Nb films with high RRR and large grain size

• Deposition in vacuum with no sputter gas

• Deployement for deposition on cavities has just been initiated

• Tests on flat samples are necessary to probe the RF performance

T. Junginger - Characterizing thin films by RF and DC methods 6/17

Quadrupole Resonator (CERN/HZB)

T. Junginger - Characterizing thin films by RF and DC methods 7/17

dSHRPPPSample

SurfaceDCDCRF 2

2,1, 21

dSH

PPR

Sample

DCDC

Surface

2

2,1, )(2 Measured directly

• Measurement of transmitted power Pt

• Pt=c∫H2ds, c from computer code

Quadrupole Resonator (CERN)

T. Junginger - Characterizing thin films by RF and DC methods

400 MHz,4K

Features:• Sample tests over a parameter

range inaccessible to elliptical cavities

• Niobium and copper substrates can be used

• Three different RF frequencies, with almost identical magnetic field configuration

• Wide temperature range (2-20K)

• Precise calorimetric measurement (Accuracy about 0.05 nΩ)

Results on an ECR sampleRRR=53 Q-slope mitigated

S. Aull et al. SRF17 TUBA03

8/17

DC measurement techniques• There are several thin film coatings which might

be potentially interesting for SRF application but cannot yet be deposited on cavities

• Additional to RF sample tests there are several DC techniques, which can give invaluable information to optimize coating parameters and provide an understanding of loss mechanism

• Point contact tunneling can be used to measure the density of states directly

• Low energy muon spin rotation can be used to directly probe the field penetration

T. Junginger - Characterizing thin films by RF and DC methods 9/17

Point contact tunneling

T. Junginger - Characterizing thin films by RF and DC methods 10/17

dI

dV= -s N N(E)

¶ f (E + eV )

¶(eV )ò dE

Smeared BCS DOS

N(E) = Re| E | -iG

(| E | -iG)2 - D2

é

ëêê

ù

ûúú

Some zero bias peaks:Magnetic impurities?

Measurement on a HiPIMS sample

Muon Spin Rotation (μSR)

T. Junginger - Characterizing thin films by RF and DC methods 11/17

• Muons are deposited one at a time in a sample

• Muon decays emitting a positron preferentially aligned with the muon spin

• Right and left detectors record positron correlated with time of arrival

• The time evolution of the asymmetry in the two signals gives a measure of the local field in the sample

Left detector

Right detector

Depth dependent (low energy) μSR

T. Junginger - Characterizing thin films by RF and DC methods 12/17

0 1 2 3 4 5 6 7 8 9 10

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Muon S

pin

Pola

risation

Time (s)

0 1 2 3 4 5 6 7 8 9 10

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Muon S

pin

Pola

risation

Time (s)

0 1 2 3 4 5 6 7 8 9 10

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Muon S

pin

Pola

risation

Time (s)

B(z)

z0

Superconductorin the Meissner State

Bext

l

Measuring fluctuating field

T. Junginger - Characterizing thin films by RF and DC methods 13/17

Slow Fluctuations

Main effect is relaxation of

the ⅓ tail at long times,

because 1/3 of the muons

see a field in spin direction

and do not process

Fast Fluctuations

No recovery. For faster

fluctuations slower

depolarization (motional

narrowing)

Can be due to muon

diffusion or paramagnetism

May 10, 2013 CMMS China Tour 17

Thepolarisation functionsfromthefield distributionDynamical field at muon site (2)

Results for theGaussian field distribution:

PZ(t ). Envelope of PX(t ).

The number next to each

curve is the ratio cγµ ∆G

.

I Sensitivity to dynamicsover several decadesofc.I Whendynamicsisfast, i.e. c

γµ∆G> 1, PZ(t) and theenvelope

of PX(t) areexponential functions(motional narrowing limit)

PZ(t) = exp(−λZt) with λZ = 2γ2µ∆

2Gc(23)

env. of PX(t) = exp(−λXt) with λX = γ2µ∆

2Gc (24)

respectively; (c1/c).I Increased sensitivity to slowdynamics in zero field.

Thepolarisation functionsfromthefield distributionDynamical field at muon site (2)

Results for theGaussian field distribution:

PZ(t ). Envelope of PX(t ).

The number next to each

curve is the ratio cγµ ∆G

.

I Sensitivity to dynamicsover several decadesofc.I When dynamicsisfast, i.e. c

γµ∆G> 1, PZ(t) and theenvelope

of PX(t) areexponential functions(motional narrowing limit)

PZ(t) = exp(−λZt) with λZ = 2γ2µ∆

2Gc(23)

env. of PX(t) = exp(−λXt) with λX = γ2µ∆

2Gc (24)

respectively; (c1/c).I Increased sensitivity to slow dynamics in zero field.

Fast FluctuationsRelaxation is exponential

Slow FluctuationsMain effect is relaxation of

the ⅓ tail

at long times

Polarization function for different

fluctuation rates. The “0” function

corresponds to a Gaussian distribution of

random fields.

HiPIMS – PCT+muSR

T. Junginger - Characterizing thin films by RF and DC methods 14/17

Hints for magnetic impurities from PCT Dynamic response

for HiPIMS sample

Growing a N2 overlayer on top of Nband stop muons close to Nb

Dynamic response rules out diffusionCrosscheck with Ni confirms that muon is static in N2

N2 on Nb

N2 on Ni

Co

mb

ined

res

ult

s st

ron

gly

sugg

est

par

amag

net

ic

imp

uri

ties

in

HiP

IMS

sam

ple

T. J

un

gin

ger

et a

l. ar

Xiv

:17

03

.08

63

5

MOTIVATION• T. Kubo suggests the use of a

bilayer system without insulator to reach high accelerating gradients

• He first calculates the penetration profile within London theory with appropriate boundary conditions

• He uses this result to calculate the forces acting on a vortex on the surface

• He concludes: A boundary of two SCs introduces a force that pushes a vortex to the direction of the material with larger penetration depth

T. Junginger - Characterizing thin films by RF and DC methods 15

Depth dependent muSR on NbTiN/Nb

Depth dependent muSR on NbTiN/Nb

T. Junginger - Characterizing thin films by RF and DC methods 16/17

• We observe a single exponential dacaywith λ0=223(7)

• Proximity effect? Dirty Nb due to diffusion?

SIS: λ0=380(100) d=98(13) 8K

SS: λ0=204(18) d=135(11) 11K

Fit: cosh(x/λ0)/cosh(d/(2 λ0))

• Measure above Tc of Nb• Field enters from both sides

• Comparison with SIS sample (no proximity)

• Significantly different λ0 and d support proximity for SS

To benefit from the counter current flow an insulating layer is essential at least for the NbTiN/Nb system

Conclusion

• RF sample tests enable

– Testing materials which are not yet ready for deposition on cavities

– Having a faster turnaround than cavity tests, providing feedback for coating optimization

– Accessing a parameter space inaccessible to cavity tests

• DC methods

– Can measure superconducting and material parameters

– Provide information for coating parameter optimization

– Can give insight which material/structures are potentially useful for SRF application

T. Junginger - Characterizing thin films by RF and DC methods 17/17