Mirco-Structures of RF Surfaces in The Electron-Beam-Weld Regions of Niobium
Mirco-Structures of RF Surfaces in The Electron-Beam-Weld Regions of Niobium
238 TUP021
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
(a)
TUP021 239
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
4.3 Weld surfaces after EP followed by lower side in the picture. The configuration of the other two
repeated BCP linear features is actually undercut, which is a frequently
observed feature for a weld surface etched by BCP.
In this study, the sample S3C1-3-EQUATOR was first
etched by EP until a surface removal of 100m was As can be seen from Fig. 4, the smooth RF surface is
reached. The sample was then etched by BCP in a step by easily destroyed by a light BCP. An extra etching of 12 m
step manner. Fig. 4 illustrates the evolution of its weld sur- by BCP is sufficient to re-establish sharp edges, although
face with an increased amount of surface removal by BCP. the step height is not that great. Again increasing of the
The linear feature starting from the center of the photo to step height with the amount of surface removal by BCP is
right is a going-down step if walking from the upper to the demonstrated.
240 TUP021
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
(a)
(a)
(b)
(b)
TUP021 241
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
-10
-15
-20
(a) 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Horizontal Scan (micron)
(a)
6
4
2
-2
-4
-6
-8
-10
0 50 100 150 200 250 300 350 400 450 500
Horizontal Scan (micron)
(b)
(b) 10
8
Figure 6: The RF surface of the non-affected region of the
Step Height (micron)
-2
Fig. 6(b). -4
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Horizontal Scan (micron)
0.5
ness in the non-affected region.
0
In order to have a better understanding about these sur-
-0.5
face irregularities, data in the vertical dimension are highly
-1
desired. Although the height of a particular step in the weld
-1.5
region can be measured with an SEM by tilting the sample,
-2
statistical data about the steps in the weld region and the 0 50 100 150 200 250 300 350 400 450 500
Horizontal Scan (micron)
roughness of the surface in a non-affected region can not
be yielded effectively in this manner. Yet these data are (d)
crucial in determining the distribution function of the field
enhancement factor discussed in [6]. For this reason, the Figure 7: Surface profiles of the sample TWC-155E. The
surfaces of the samples were measured with a surface pro- weld (a) and non-affected region (b) after a surface removal
filer. The stylus of the profiler used in this study (Alpha- of 117 m by BCP. The weld (c) and non-affected region
step-500 by Tencor) has a curvature radius of 5 m and a (d) after an extra 90 m removal by EP. Note the difference
shank angle of 60 . in horizontal scale for the weld and non-affected region.
242 TUP021
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
Counts
8
pendent to the microscopic photography described in pre-
vious sections. Unfortunately, limited by the curvature ra- 6
4
6.2 Step height statistics
2
In order to compare the surface characteristics after differ-
0
ent surface treatments, statistical data are needed due to the 0 5 10 15 20 25 30 35 40
fact that large number of steps with diverse configurations Step Height (micron)
are involved. Here the distribution of step height is em-
(b)
phasized because of its close relationship with the distri-
35
bution function of the magnetic field enhancement factor
BCP 117 micron + EP 90 micron
discussed in [6]. The RMS step height is cited only for a 30
general comparison purpose.
Fig. 8 depicts histograms of step height in the weld re- 25
TUP021 243
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
14
BCP 117 micron
Table 1: Statistics of step height in the weld region of the
12 sample TWC-155E
surface removal RMS [m] Max. [m]
10 117 m BCP 8.8 17.8
8 250 m BCP 16.2 36.8
117 m BCP + 90 m EP
Counts
5.5 13
6
4
Table 2: Statistics of step height in the non-affected region
2 of the sample TWC-155E
surface removal RMS [m] Max. [m]
117 m BCP
0
0 2 4 6 8 10 12 14 4.8 11.4
Step Height (micron) 250 m BCP 4.3 12.6
(a) 117 m BCP + 90 m EP 2.2 6.7
45
BCP 250 micron
40 ferent from that of the bulk due to either morphological fac-
35 tors or compositional factors. Recent studies have shown
30 that a cavity etched by BCP may suffer from high-field Q
25
drop. whereas, after certain amount of further etching by
Counts
EP, the Q drop could be removed and the quench field could
20
also be raised [3]. Among others, a model based on the
15 magnetic field enhancement has been developed to explain
10 this phenomenon [6]. Studies presented in this paper have
5 important implications in the context of the high-field Q
0
drop of a superconducting cavity.
0 2 4 6 8 10 12 14
Step Height (micron)
7.1 Magnetic field enhancement
(b)
Magnetic field can be enhanced locally on the surface of the
35
BCP 117 micron + EP 90 micron corner of a step. The dependence of the enhancement fac-
30 tor on the curvature radius of the corner of a step has been
dealt with in great detail in [6]. Preliminary simulations
25
in a 3-D fashion with the code MAFIA [9] suggest that the
20 field enhancement factor is also a strong function of the ori-
Counts
244 TUP021
Proceedings of the 1999 Workshop on RF Superconductivity, La Fonda Hotel, Santa Fe, New Mexico, USA
into normal conducting due to magnetic field enhancement. mensions in the lining-up direction, as a result their bound-
Theoretical work regarding the relationship between the aries are highly linear.
breakdown field and the geometry of a defect has shown The step edge is significantly rounded off and step height
that linear defects are more vulnerable to thermal instabili- is appreciably reduced with extra surface etching by EP. A
ties compared with circular ones [10]. As can be seen from surface removal of more than 150 m is needed to smooth
Fig. 1, steps in the weld region of a niobium being well out the weld region of a niobium having been previously
etched by BCP have elongated dimension in the direction well etched by BCP. Reversely, a smooth surface obtained
of about 60 away from the weld joint line, and hence by heavy EP can be easily destroyed by further BCP, even
they are highly linear. As a result, a superconducting cav- as light as a surface removal of 12 m.
ity will be more likely to quench in the EBW region than in
the other regions (here we limit ourselves to a cavity work- 9 REFERENCES
ing in the high gradient regime, > 20 MV/m, assuming it
is free from thermal breakdown and field emission). This [1] H. Padamsee, J. Knobloch, and T. Hays, RF superconduc-
tivity for accelerators, Chapter 6, John Wiley & Sons, Inc.,
claim goes even stronger considering the fact that the mag-
1998.
netic field enhancement factor is higher in the EBW region,
as discussed in previous section. [2] H. Safa, Proc. of the 8th workshop on superconductivity,
Abano Terme (Padova), Italy, 1997, p814.
[3] E. Kako et. al., Proc. of the 1999 Particle Accelerator Con-
7.3 How much etching is needed ference, New York, USA, 1999, p432.
It has been established that a surface removal of more than [4] J. Kirchgessner, Proc. of the 4t h Workshop on RF Supercon-
100 m is needed to reach an ideal accelerating gradient ductivity, Argonne National Lab., Argonne, IL, USA, 1998,
[7][8]. In this study, we found the RF surface in the weld p533.
region is dominated by circular defects until a surface re- [5] These SEM photos were taken in the secondary electron
moval of about 120 m was reached. One could argue that mode with the secondary electron detector located over the
there may be some correlation between these circular de- top of the sample. The sample was titled 80 with respect
fects and cavity performance at gradients < 20 MV/m. to the ground level. Photos presented in the paper have been
More surface etching by BCP is helpful in eliminating converted with tilt correction taken into account. In this con-
these circular defects and pushing the gradient to higher figuration, a bright feature in the photo represents a step
levels in the < 20MV/m regime. However, aggresive sur- facing towards the detector as a result of the combination
face etching by BCP beyond 150 m may be unfavorable, of the emission contrast and detection contrast. Due to the
surface sensitive nature of secondary electron emission, the
as it will trigger the high-field Q drop when the cavity
brightness of a step edge can be in a way translated into its
works at higher gradients ( > 20 MV/m ). Because the step sharpness.
height will increase with more etching by BCP and a higher
[6] J. Knobloch, R.L. Geng, M. Liepe, H. Padamsee, these pro-
field enhancement is resulted. For this reason, we suggest
ceedings, TUA004.
an aggressive etching by BCP of more than 150 m surface
removal should be cautioned. [7] P. Kneisel, Proc. of the 8th workshop on superconductivity,
Abano Terme (Padova), Italy, 1997, p830.
Basically surface etching by EP will round off step edges
and reduce the step height on a niobium surface. Studies [8] K. Saito et. al., Proc. of the 8th workshop on superconduc-
presented in this paper show that even after a surface re- tivity, Abano Terme (Padova), Italy, 1997, p795.
moval of 90 m by EP is reached, the step height is still [9] A step-loaded rectangular cavity, big enough compared with
rather high, 10 m, in the weld region of a surface pre- the dimension of the step, was used to simulate the magnetic
viously well etched by BCP. It is therefore expected that field enhancement effect with different step orientations.
more surface removal, presumably more than 150 m, is [10] N. A. Buznikov and A. A. Pukhov, Supercond. Sci. Technol.,
needed to smooth out a weld surface having received sig- 11(1998), P1201-1208.
nificant etching by BCP.
8 CONCLUSIONS
The RF surface in the EBW region of niobium after suffi-
cient etching by BCP is characterized by micro-steps with
sharp edges near grain boundaries. The step height in-
creases with repeated etching by BCP and can be as high
as more than 30 m. No step height saturation in the weld
region was seen up to a surface removal of 250 m.
The steps in the weld region are lined up in a way that
they form an angle of about 60 with respect to the weld
joint line. The grains in the weld region have elongated di-
TUP021 245