1. Christian Hammill, Wayne State University
Outgassing of Stainless Steel Vacuum Chambers and The Vacuum Pumping Speed and Capacity Evaluation of a Titanium Sublimation Pump
Christian Hammill, Wayne State University
Dr. Yulin Li & Dr. Xianghong Liu, Cornell University

2. My Project at a glance First part; Second part; extremely low =>
Extreme high vacuum
Second part;
Very high pumping => Able to handle very large gas loads
November 11, 2018
REU 2007, LEPP Cornell

3. Contents Outgassing of stainless steel chambers
Introduction and purpose
What is the goal of this project?
Methodology
What methods are used to determine ?
What do these methods involve?
What is the difference in these methods?
Results
Conclusions
Vacuum characteristics of the TiSP
Introduction and purpose
What is a TiSP?
What do we want to determine from the TiSP?
Methodology
How do we determine Q(t) and STi(t)?
Results
Conclusions
November 11, 2018
REU 2007, LEPP Cornell

4. SST Outgassing Introduction
Ultra-High Vacuum: 10-9~10-12 torr
Extreme-High Vacuum: <10-12 torr
XHV must be achieved in parts of ERL (i.e. photo-chamber)
Stainless Steel outgassing hinders this
When put under vacuum
400°C bakeout reduces outgassing
Does this last forever?!
My mission: do stainless steel chambers that have been stored properly in N2 for ~8 months keep low outgassing property?
November 11, 2018
REU 2007, LEPP Cornell

5. SST Outgassing Methodology I
Vacuum system comprised of two parts:
Testing Chamber
Enclosed by oven
Heating gun/ air blower
Water cooling coils/ heating tape
SRG/ sample chamber
Sensor Chamber
Valve to Testing Chamber
RGA
CCG
Ion Pump & Turbo Pump
November 11, 2018
REU 2007, LEPP Cornell

6. SST Outgassing Methodology II
Two methods of determining
Rate of Rise method
Where V is the volume of the chamber (28L) and As is inner surface area of the sample chamber (7500 cm2)
Throughput method
Where is Sip the pumping speed of the ion pump and As is inner surface area of the sample chamber
∆P (pressure increase due to outgassing) is explained later
November 11, 2018
REU 2007, LEPP Cornell

7. SST Outgassing Methodology III
Both methods follow a similar procedure:
Set up and leak check
All vacuum flanges are connected and properly tightened
Then, entire vacuum system is pumped down and checked for gas leaks
Bakeout
The adsorbed water molecules (from air exposure) are eliminated via temperatures >120°C
Measurement
is determined using either ROR method or throughput
Repeat
Baking temperature is increased each time (150°C to 200°C to 250°C)
November 11, 2018
REU 2007, LEPP Cornell

8. The Rate of Rise Method
In RoR measurements, the sample chamber is closed off from any pumping
This allows gas to accumulate in the closed system
With a constant outgassing rate, , a linear pressure rise is expected
The rate of rise in the pressure, , is measured to calculate such that:
November 11, 2018
REU 2007, LEPP Cornell

9. The Spinning Rotor Gauge
SRG consists of a magnetized rotor ball in a gimble tube, and a removable head that contains sets of coils
The Rotor is magnetically levitated and spinning
Molecules in the chamber collide with the rotor, ever slightly slowing down the rotor, the “molecular drag”
The SRG measures the pressure by measuring the slowing down rate of the rotor due to the molecular drag
Rotor
Removable SRG
coil head
Gimble
SRG do not Alter the vacuum – Ideal for RoR
November 11, 2018
REU 2007, LEPP Cornell

10. Throughput method Used as crosscheck for RoR method
The pressure at the CCG is recorded while the sample chamber is still closed off from the sensor chamber => baseline pressure
The chamber is then opened and the pressure is left to settle then recorded
The difference between the settled pressure and the baseline pressure is known as the change in pressure, ∆P
Then, using the pumping speed of the ion pump, Sip (~9L/s), we calculate :
November 11, 2018
REU 2007, LEPP Cornell

11. A Typical RoR Result
Pressure measured by SRG clearly show a linear rise in time. The fitted slope:
dP/dt~ 2x10-12 Torr/s.
Very stable temperature control ΔT ~ 0.17 °C
November 11, 2018
REU 2007, LEPP Cornell

12. A Typical Throughput Result
Sip is determined via a pump down of sample chamber:
Fitting (blue) curve, P(t):
Sip=V/τ1
Where,
Gas composition at point “A”:
In this case, ∆P≈1.5x10-11 Torr,
Sip ≈9 L/s therefore,
=1.9x10-14 Torr•L•s-1•cm-2
November 11, 2018
REU 2007, LEPP Cornell

13. Outgassing Results – Summary
Current project data
Tbake(°C)
(10-15 Torr•L•s-1•cm-2)
Rate of rise
Throughput
120
2.6 30
6.0 19
7.5
n/a
150 25 18 15
200
6.7 28
250
9.3 35 38
Data from ~8 months ago
TBake (°C)
(10-15 Torr•L•s1•cm-2)
150
19.0
200
16.0
250
14.0
November 11, 2018
REU 2007, LEPP Cornell

14. SST Outgassing Conclusions
The outgassing rates measured in this study are comparable with the results from 8 months ago.
This indicates that the extremely low outgassing property can be maintained.
The comparison between the rate of rise method and the throughput results is very good, when one considers at least two factors:
The measurements were done at very different pressure ranges (~10-6 torr for rate of rise, and ~10-10 torr for the throughput).
They both use very different gauges (SRG vs. CCG).
Therefore, in certain places in the ERL (i.e. the photo-cathode) one can vent out the vacuum, properly store the stainless steel components in N2, and not damage the outgassing properties of these components
November 11, 2018
REU 2007, LEPP Cornell

15. TiSP Performance Introduction
After measuring the properties of the electron beam, it must be safely terminated at the beam dump in the Cornell Proto-type Photo-cathode Injector
Very large gas loads of H2 gas are generated at the beam dump when this happens
A large TiSP will be used together with two large ion pumps to control the H2 gas load
Will this do the trick?
My mission: evaluate the pumping performance (particularly the pumping speed and the pumping capacity) of the TiSP to determine if it is suitable for this application
Beam Dump
TiSP
November 11, 2018
REU 2007, LEPP Cornell

16. TiSP Performance methodology I
Experimental Setup
The test system was leak checked and baked at 170°C
Ultra-high purity hydrogen is used to measure the TiSP pumping performance
The flow-rate of H2 was experimentally determined before any Ti-sublimation,
After Ti-sublimation, H2 flows through the TiSP chamber, and the pressures are monitored by two cold cathode gauges
November 11, 2018
REU 2007, LEPP Cornell

17. TiSP performance methodology II
Ti Sublimation
There are 3 Ti cartidges on the end of the chamber
Each filament is heated via voltage power source
This heating causes the Ti to sublimate in the chamber
This process is known as “flashing”
The Ti is flashed in the chamber for a certain period of time at a certain power
Ti cartrige with 3 filaments
November 11, 2018
REU 2007, LEPP Cornell

18. Determining STi(t)
After flashing, we open the H2 valve again and let the gas flow
Because the Ti layer inside the chamber is so reactive, it will capture much of the gas as it enters the chamber
This causes the pressure to decrease at the CCG
At the CCG there are now 2 pumping speeds
The effective speed of the turbo, , and the pumping speed of the chamber, STi(t).
Knowing this, we derive the formula for the pumping speed of the chamber over time
 this follows the Equation, =SP
November 11, 2018
REU 2007, LEPP Cornell

19. [STi(tj)• ]∆tj Q(t)= Determination of Q(t)
We want to determine the total amount of H2 gas pumped away by the TiSP chamber, Q(t).
As the Ti film pumps more and more H2 is pumped away it becomes saturated with H2
As it becomes more saturated, its pumping speed decreases
The accumulated Q(t) is directly related to STi(t) such that:
[STi(tj)• ]∆tj
Q(t)=
November 11, 2018
REU 2007, LEPP Cornell

20. TiSP Performance Results I
STi(t) was measured at 2 flashing settings
(A) 3 minutes at 170W
(B) 5 minutes at 195W
Higher Ti-flashing power & longer duration => much thicker Ti layer
The STi(t)’s are plotted against the Q(t)’s
Limited Ti flashing and H2 saturation cycles were done
Thick, H2 rich, Ti films are known to be “flaky”
November 11, 2018
REU 2007, LEPP Cornell

21. TiSP Performance Results II
Q(t) is arbitrary and depends on its application.
Here, we chose Q(t) to be the point where STi(t) drops to ~100L/s
(i.e. ~10% of its initial pumping speed).
With this definition, one sees from that:
QA≈ 4 Torr•L
QB≈ 25 Torr•L
With maximum pumping speeds of:
STi, A(t) ≈900L/s STi, B(t) ≈1200L/s
As a side note, it took ~80 hours for B to saturate from the beginning to 100L/s
Hydrogen Pumping by Ti
Step 1 – Dissociative adsorption
H2  2Hads
Step 2 – Bulk diffusion
Hads  HBulk
Step 1 depends on surface reactivity of Ti film
Step 2 depends on Ti film thickness and solubility
November 11, 2018
REU 2007, LEPP Cornell

22. TiSP Performance Conclusions
It can be seen that the max STi ~ L/s & that it is capable of pumping out ≤25 Torr•L of H2 gas.
The estimated gas load at the beam dump (A5 section of the Cornell Proto-type Photo-cathode Injector) can be as high as 1.8x10-3 Torr•L/s.
Considering that two other ion pumps will be sharing half of the load, the TiSP will have to handle a gas load of 9.0x10-4 Torr•L/s.
With a little math, we can see that the TiSP can go ~8 hours before needing to go through a flashing cycle again.
This is considered to be sufficient for the injector operations
Therefore the TiSP is suitable for the ERL prototype project
November 11, 2018
REU 2007, LEPP Cornell

23. Acknowledgements Thanks are due to :
My mentors; Dr. Yulin Li and Dr. Xianghong Liu who have assisted me in the writing of this paper, the assembling of the experiments, and the overall guidance though this project.
The lab technical staff; Tim Giles, Tobey Moore, and Brent Johnson.
Dr. Rich Galik and Dr. Claude Pruneau who are the assemblers of this REU project and made it able for me to attend the summer here.
The National Science Foundation who every year make it possible for students much like myself to have experiences like this one.
November 11, 2018
REU 2007, LEPP Cornell