Anatomy of an IRMS Lecture 3 – Vacuum Systems

Extra Reading: pdf’s from The Vacuum Lab

1) Introduction
a) Last time we talked about how to create a vacuum, ie types of pumps. Today talk about all the
other technology surrounding vacuum systems.
b) Flow in a vacuum. Our intuition based on viscous flow through pipes is wrong. Affects
dimensions and shapes of system components.
c) Materials. Compatibility, outgassing, strength, cost, etc.
d) Fittings. What are the standard ways of connecting vacuum components to ensure a good seal?
e) Gauges. How do we measure the vacuum.
2) Flow in a vacuum
a) Recall that at high vacuum, collisions between molecules and walls are more common than
between molecules. So flow is motion-driven (“molecular flow”), not pressure-driven (viscous
flow). More analogous to diffusion.
b) Fundamental Relationship: Q = SP
i) Q = mass flow rate (mole/time; also P*V/time, eg Torr-L/sec)
ii) P = pressure (units of P, eg Torr)
iii) S = conductance (units of vol/time, eg L/sec)
c) Analogy to Ohm’s law (V = IR)
i) V (voltage) analogous to P
ii) I (current) analogous to Q
iii) R (resistance) analogous to 1/S
iv) Substitution into Ohm’s law would give P=Q/S, rearrange to Q=SP
d) Conductance.
i) Ability of a physical component to allow molecules to pass. It is the inverse of resistance. For
a simple tube (pipe), conductance is proportional to cross-sectional area. For a pump,
conductance is the pumping speed.
(1) Why is this? When a molecule hits a surface, it does not reflect like light off a mirror.
Instead it momentarily sticks, and is released at a random angle. Thus the probability of
making it through a tube is inversely proportional to the number of wall collisions that
you experience.
(2) Bigger tubes have a wider “acceptance angle”, ie direction in which molecules could be
traveling and will make it through with no collisions.
ii) In general, conductance is a fixed property of the vacuum system. Note that it must be a
combination of pumps, tubing, fittings, etc.
(1) For simple components like tubes, can estimate conductance based on size and shape (see
table).
(2) When we have multiple components in series (tubes then pump), calculate the effective
conductance just as we would for resistance (Reff = R1 + R2 +…; 1/Seff = 1/S1 + 1/S2+…)
(3) Anything you can do to help increase conductance of tubing, etc will help overall system.
Note futility of putting larger and larger pumps on a system whose conductance is
controlled by other components, like tubing.
e) Flow rates. Need to distinguish between volumetric and mass flow rates, which are both in the
fundamental equation
i) Mass flow rate (Q) is measured in molecules per time, and is independent of pressure. Makes
intuitive sense when describing gas load (ie, input from a capillary), leak rate, etc.

1
Anatomy of an IRMS Lecture 3 – Vacuum Systems

ii) Volumetric flow rate is measured in volume per time, and is an intuitive way of describing
pumps (ie, pumping speed; units are vol/time). Recall that they have constant S, and so mass
flow decreases with P.
f) Behavior of the system
i) Pumpdown. Conductance is fixed by system components, and is generally constant wrt
pressure (ie, pumping speed does not vary with pressure). Gas inputs to the system (leaks,
load) are a constant mass flow rate. If SP > Q, then pressure will continue to drop until SP=Q.
That P is the ultimate vacuum of the system for given gas load.
(1) Remember that S may change as a function of P, depending on where in the pumping
curve you sit.
ii) At fixed mass flow rate (determined by inputs), higher pumping speed and/or conductance
gives lower pressure. Bigger pumps yield better vacuum (up to a point…)
iii) At fixed mass flow rate, lower conductance (smaller tubes, etc) yields higher pressure.
Importance of proper design of vacuum system components.
iv) If mass flow rate and pressure are known, pumping speed can be calculated. These
calculations are often very useful in diagnosing vacuum system problems.
v) Ionization efficiency/pressure. The volume in which molecules are ionized is not in perfect
contact with the surrounding vacuum system. In many cases, it is purposefully enclosed
(“tight” source) to yield higher pressure and thus more efficient ionization.
(1) If we want mass flow rate in ion source, can use measured pressure and pumping speed to
estimate
(2) If we know mass flow rate and limiting conductance, can calculate pressure inside the ion
source.
3) Vacuum system design
a) Need to maximize conductance of all system components. Note that because they are adding in
inverse, doesn’t help to increase something far above the “limiting” conductance, which ideally
should be the pump.
b) Tubing size.
i) Generally want wider diameter and shorter length. This is why vacuum connections are big
(typically >1” ID), and we never put vacuum pumps across the hall.
ii) Point of diminishing returns as surface area of system (and thus desorption) increases.
c) Geometry. Straight flow paths better than turns.
d) Surfaces. Smooth better than rough, both for increasing conductance and minimizing surface
desorption.
e) Virtual Leaks
i) Imagine a screw driven into a blind threaded hole. Air is trapped at the end of the screw, and
must now flow out around the screw. Even though its not leak-tight, that very small &
tortuous path has very low conductance, so air will come back out very slowly. Ends up
looking just like a leak, that takes forever to get rid of.
4) Materials
a) Primary considerations are
i) Strength to resist external forces (for chamber)
ii) water adsorption
iii) degassing rate
iv) heat tolerance
v) Leak rate (really only for seals)
b) Metals.
i) Stainless steel is common choice
(1) type 304 is best (303 has too much S)

2
Anatomy of an IRMS Lecture 3 – Vacuum Systems

(2) for UHV, type 304L has less carbon, even lower degassing
(3) Electropolished surface is preferred to minimize water adsorption; fine machined surface
(4) cleaning tool oils is difficult, requires care
ii) Aluminum is also OK
(1) cannot be anodized (adsorbs too much water)
(2) subject to corrosion
(3) easy to strip threads on flanges, etc.
(4) less heat tolerant (weakens quickly)
c) glass
i) Pyrex is fine, but slow diffusion of He (special glasses available)
ii) glass capillaries are also fine, just remove polyimide coating
d) ceramics
i) Need to make sure it is high-density sintered ceramic, otherwise much outgassing
ii) best material for making insulating feed-throughs
e) Rubber.
i) OK to use thick rubber vacuum hose for rough pumps, but nowhere else.
f) Things to avoid
i) Plastics. Virtually all outgas large amounts, will coat the vacuum system and be there forever.
Need to be very rigorous about this.
ii) Porous materials (ceramic, glass, etc). Will take forever to reach adequate vacuum.
iii) Epoxy. Even vacuum epoxy degasses. If you have to use it, make sure surface area is minimal
(on outside of pinhole, not inside).
iv) Avoid anything containing Zn, Cd, especially plated metal screws, etc. They will degas.
5) Cleaning
a) Common activity; needs to happen every time something is taken out of a vacuum system.
Fingerprints can destroy a vacuum.
i) Mechanical abrasion, polishing; diamond grit, sandpaper, glass fiber pens, etc
ii) Degreasing (DCM, etc)
iii) Washing: soap and water, water
iv) Drying: acetone/ethanol, 60°C oven
v) Wrapped in Al foil if practical.
b) Common to wear lint-free gloves while working in vacuum system that won’t be cleaned.
6) Fittings
a) Typically a vacuum system consists of a chamber where all the action happens; multiple holes for
instrumentation, gas flow, meters, pumps, etc. Every connection is accomplished by a fitting, and
every one leaks a little bit.
i) First goal is to minimize number of connections. Leak rate is roughly proportional to length
of seals in system.
ii) Second goal is to choose fitting types based on vacuum needs (low versus high)..
b) Quick-release flanges, used for connecting rough pumps/hoses/etc
i) Can be designated as QF, KF, NW, or DN, depending on details of construction (chamfered
here or there, etc)
ii) All should be interchangeable. Mating facies are symmetric (‘not sexed’).
iii) Seal is by polymer o-ring, generally Viton
iv) limited by o-ring to 120°C, 10-8 mbar
c) Large flange standard, using for permanent connections to vacuum chamber (source, pump, etc)
i) Often designated ISO, but also LF, LFB, MF
ii) Seal can be provided by Viton o-ring, metal knife-edge seal (Al), soft metal gasket (Au, Ag)
iii) connection made either by bolts or claw-clamps

3
Anatomy of an IRMS Lecture 3 – Vacuum Systems

iv) sized by inner dimension in mm

v) temp and pressure determined by gasket material; with metal seals, can reach 300C and 10-10
mbar.
d) Conflat standard; uses metal-metal knifedge seals for ultra-high vacuum
i) Name is patented by Varian; most makers call these “CF flanges”
ii) flange faces have machined knife edge (both sides), which cut into soft copper gasket
between them
iii) Very good at filling minor surface imperfections
iv) baking to 450C, vacuum to 10-13 mbar
e) Feedthroughs; how do you get high voltage electrical signals through a metal wall while keeping
vacuum?
i) A variety of technologies, including both mechanical and electrical feedthroughs
ii) Our requirements are the most stringent: very high voltage, high vacuum, bakeable
iii) Requires direct ceramic-to-metal sealing; you just buy these, and try not to break them
7) Gauges
a) Review gauge pressure versus absolute pressure. Gauge pressure depends on local atmospheric
pressure, really only useful for high-P measurements.
b) Mechanical – work by measuring displacement or strain under pressure; useful for ~1 Torr to
high pressure
i) Manometer- measure the height of liquid displaced (water or mercury)
(1) McLeod gauge; compresses gas into a liquid manometer; very accurate, down to 10-6 torr,
but inconvenient; often used for calibration of other gauges
ii) Anaeroid – metal pressure plate pushes against a spring and displacement is measured
iii) Bourdon tube – a flattened tube tends to become more circular as pressure differential
increases; when tube is bent into C shape, this causes it to straighten out at pressure drops;
can be surprisingly sensitive
iv) Diaphragm gauge – measures deflection of a flexible membrane. Often implemented as a
‘capacitance manometer”, in which system measures capacitance of the membrane moving
relative to a fixed plate of known capacitance. Baratron gauge is one example. Useful down
to 10-2 Torr (mbar)
c) Thermal conductivity (useful range: 10-3 to 10 Torr)
i) Basic principle is that the rate of heat loss of a wire in vacuum depends on number of
molecular collisions, which depend on density of molecules. Drawback of all these is that
they depend on heat capacity of gas, must be calibrated for specific gases.
ii) Thermocouple gauge; constant current applied to wire, separate thermocouple junction heated
by conduction measures temperature of the wire. This is a 2-wire system.
iii) Pirani gauge. Uses two platinum wires as legs of a Wheatstone bridge, one of them held at
sample pressure other at fixed reference pressure. Each wire is held at constant temperature,
measures change in power required to measure this temp (heat dissipation by collisions with
molecules, proportional to pressure).
d) Ionization gauges (useful range: 10-10 to 10-3 Torr)
i) Ionize gas and force it to migrate to a cathode down a voltage gradient. Measure resulting
current, proportional to pressure. Only works at low pressure to sustain ions.
ii) Hot cathode design – uses a heated filament to boil off electrons (thermionic emission) and
ionize gas, then accelerated between grid and cathode by voltage gradient
iii) Cold cathode design (Penning gauge, magnetron) – In a high voltage field, ionization process
is self-sustaining (electron released strikes other atoms and generates new ions). Magnetic
field confines electrons to tight spiral to make them more likely to strike other atoms.
Difficulty comes in initiating the self-sustaining ionization cascade.

4
Anatomy of an IRMS Lecture 3 – Vacuum Systems

e)