 |
| Richard C. Kullberg
|
A major concern for hermetic packages is that they will "spring a leak" and acquire more moisture from external ambient than is safe for components and devices inside. Leak physics have been mathematically modeled to quantitate leak rates. An array of leak test methods has been standardized within the MIL SPEC system to test package leak rates as low as 10-11 cc-atm/sec.
Mass spectrometry, intended to measure moisture and other volatiles in an enclosure, is
 |
| Robert K. Lowry
|
used to infer leak rates based on increases in moisture from amounts assumed nil at seal (a dangerous assumption) to amounts measured after some elapsed time from seal to analysis.
One characteristic of atoms and molecules is that they are not always round and hard like billiard balls. They are really "clouds of probability" defined by electronic orbitals surrounding their nuclei. When atoms combine in various proportions to form molecules the resulting molecular shapes and their physical and chemical behavior can be complex and unpredictable.
Four gases of great interest to hermeticians are oxygen, argon, moisture, and helium. Presence of any of the first three suggests a leak, while helium is used both with inert fill gas and as a leak test substance. Behavioral properties of these molecules are not always fully captured by the leak models and can significantly modify leak characteristics, for better or worse.
Argon is a minor component of air always present at a fixed concentration of 0.934 volume percent. Helium is a negligible component of natural air. Both have an electronic configuration of eight outer shell electrons, making argon and helium chemically stable and essentially spherical monatomic molecules.
Oxygen is a major component of air always present at a fixed concentration of 20.95 volume percent. It is a diatomic molecule consisting of two oxygen atoms covalently bonded (sharing electrons). As a molecule oxygen has a two-lobed linear shape. With six outer shell electrons and thus electronic vacancies on each of the atoms oxygen is chemically reactive.
Water is a minor component of air varying in concentration with temperature, pressure, and relative humidity. Depending on local conditions moisture content of air ranges from 0.1 to 3.0+% volume percent. Water is a triatomic molecule with two hydrogen atoms bonded to an oxygen atom at an angle of 104.5º.
It has a complex "cloud of probability" of orbiting electrons spending less time around the hydrogen atoms (partial positive charge) and more time around the oxygen atom in the form of two sets of unpaired electrons (partial negative charge). This charge configuration makes water an oddly-shaped polar and chemically reactive molecule.
Some chemical and physical properties of hermeticity-relevant gas molecules.
|
|
He |
Ar |
O2 |
H2O |
|
Molecular weight |
4.00 |
39.95 |
32.00 |
18.00 |
|
Shape |
Sphere |
Sphere |
Barbell |
3-lobe, V shape |
|
Type of chemical bond |
- |
- |
Nonpolar covalent |
Polar covalent |
|
Bond length |
- |
- |
O−O, 1.208A |
O−H, 0.958A |
|
Bond angle |
- |
- |
180º |
104.5º |
|
Bond energy |
- |
- |
O-O, 116 kcal/mole |
O−H, 110 kcal/mole |
|
Charge distribution |
uniform |
Uniform |
linear, uniform |
+ on H atoms, - on O atom |
|
Molecular diametera |
1.90A |
2.88A |
2.98A |
- |
|
Molecular diameterb |
2.65A |
2.94A |
2.92A |
2.78A |
a. from viscosity b. from van der Waal’s equation
Clearly these molecules are not the same chemically and physically. They have subjective properties not reflected in the models. On the molecular level the inerts collide elastically without chemical reaction, but can physisorb on surfaces. In contrast the water molecule is sort of "squishy."
Water and oxygen physisorb, chemisorb, and readily react to form hydroxides and oxides. They behave quite unpredictably especially in long tortuous leak paths of varying surface physics and chemistry and/or with physical constrictions. Such leaks partition components of air, causing differential leak rates, so that internal gas analysis of units with air ingress seldom shows the stoichiometric O2/Ar ratio of 22.43.
A leak in a glass-to-metal seal transits each gas differently than an identical leak in a weld seal. The "finer" and more tortuous the path and the more chemically reactive its surfaces, the less reliable the leak rate models for molecular flow. A path that transits gases at one rate early may transit them at a different rate later as surfaces hydrate or oxidize.
Many enclosures already contain moisture immediately post-seal, plus dynamically increasing amounts of it as package materials volatilize. Keep all this in mind when working with leak rate results. Inferred times to moisture noncompliance via leak could be dramatically shorter or longer than the numbers suggest. Later in this discussion we will look at what Mother Nature actually gives us in terms of leaks and molecules, but next time we will discuss the classic leak models built around the simple case of He.
See previous articles written for this series in Semiconductor Packaging News
Leak Rate Implications posted on December 15, 2009
It's Still Way To Much Water posted on October 27, 2009
Way Too Much Water posted on August 26, 2009
Removing Water posted on July 23
Where Does Water Come From? posted on June 15, 2009
STICTION posted on May 13, 2009
