0
$\begingroup$

I've heard that a surface roughness needs to be within a defined range, but in theory, is it possible to seal a perfectly smooth surface? E.g. with an o-ring, static or dynamic?

Intuition tells me that this is the perfect surface to seal against, but am I missing something here?

EDIT:

For concreteness, let's say I'm sealing a polymer o-ring against stainless steel :)

$\endgroup$
3
  • 1
    $\begingroup$ Depends what it has to seal, if there is movement due to vibration or heat, possibilities of distortion etc. but face to face sealing is possible, but also things like Wills rings... $\endgroup$
    – Solar Mike
    Commented Nov 18, 2022 at 11:28
  • $\begingroup$ Good point! edited to make the question more concrete. $\endgroup$ Commented Nov 18, 2022 at 11:31
  • $\begingroup$ There are many ( a hundred ?) high pressure seals in oilfield connections, all as smooth as cutting tools can make the surfaces. Most metal to metal, some with elastomers. Generally tolerate 10,000 psi, some much more. Also Swage-lok compression fittings ; the sealing ferrules .etc, looked smooth to me. Where did you get the idea of roughness? $\endgroup$ Commented Dec 18, 2022 at 19:23

2 Answers 2

0
$\begingroup$

I think I've figured this out (through some discussion with a helpful physicist). Yes, of course it's possible to seal a perfectly smooth surface.

The purpose of roughness in o-ring sealing is twofold:

  1. Retention-- roughness helps o-rings remained retained in their groove or other retention feature via friction.

  2. Stress concentration -- roughness provides local stress concentrations to produce a better seal. A perfectly flat surface will seal, but may require higher sealing force than an appropriately rough surface due to lack of local stress concentrations in the o-ring. There is an optimal roughness which is nonzero-- too rough and there's a chance you will have leaks across your sealing surface. Too smooth and you're suboptimal with respect to sealing force.

$\endgroup$
4
  • 2
    $\begingroup$ Things like hydrogen can pass right through the crystal lattice of steel, and there is no such thing as perfectly smooth, atoms and molecules being blurry discrete objects that are mostly empty space. $\endgroup$
    – DKNguyen
    Commented Nov 18, 2022 at 16:33
  • 1
    $\begingroup$ I wouldn't be sure about the greater force required to seal smooth surface. Sealing rough surface requires great initial force to plasticise the gasket so it can match the rough surface. If this is not achieved, you may end up with micro-channels due to the roughness. The basic requirement is of course to have the gasket contact pressure higher than fluid pressure. The higher it is, the lower the leak rate. You can check out EN 13555 standard for measuring gaskets with associated website gasketdata.org with test results submitted by manufacturers. $\endgroup$ Commented Nov 19, 2022 at 11:40
  • 1
    $\begingroup$ @TomášLétal thanks! super helpful. Do you disagree with my thinking about stress concentrations meaning a lower sealing force requirement? I can imagine the required force vs. roughness curve being nonlinear with a positive second derivative. $\endgroup$ Commented Nov 19, 2022 at 22:20
  • 1
    $\begingroup$ @spencerwilson Yes, I would disagree. In initial contact the stress concentrations may increase contact pressure, but only locally. In order to make a good seal, you need the whole surface in the contact (or at least major part of it). $\endgroup$ Commented Nov 20, 2022 at 9:30
0
$\begingroup$

Please note that it is possible to get a perfect seal with a smooth surface if you use a chemically reactive glue that binds to both surfaces in the joint as it sets up.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.