Viscosity - speed diffusion - temperature dependency

David Jonsson

unread,

Mar 1, 2010, 5:24:24 PM3/1/10

to

Viscosity is explained as speed diffusion. Diffusion increases with
temperature but viscosity decreases with temperature. How is this
explained?

David

Oliver Kaufhold

unread,

Mar 2, 2010, 2:44:00 AM3/2/10

to

David Jonsson schrieb:

> Viscosity is explained as speed diffusion.

Where do you read this?

> Diffusion increases with temperature but viscosity decreases with temperature.

This may be valid for many liquids but viscosity of most gases increases with higher
temperature.

Greetings

Oliver

Message has been deleted

Olin Perry Norton

unread,

Mar 2, 2010, 2:49:26 PM3/2/10

to

Perhaps it might be better to say that that viscosity represents
the diffusion of momentum, or the diffusion of vorticity.

Others have already pointed out that, in gases, viscosity
increases with temperature.

Olin Perry Norton

jfh

unread,

Mar 2, 2010, 3:41:53 PM3/2/10

to

On Mar 2, 11:24 am, David Jonsson <davidjonssonswe...@gmail.com>
wrote:

In a liquid the Sutherland-Einstein law (often mislabelled the Stokes-
Einstein law, but Sutherland's paper was submitted 2 months before
Einstein's) is sometimes a good approximation: it says
diffusivity*viscosity = kT/(6*pi*density*radius of molecule). Think of
molecular motion increasing with temperature (hence kT where k =
Boltzmann's constant, T = temperature) but impeded by viscosity.

John Harper

David Jonsson

unread,

Mar 11, 2010, 7:28:55 AM3/11/10

to

Thanks everyone for clearing out.

My general impression of viscosity is from fluids and practical
experience of warm versus cold oil.

It seems to me that Sutherland-Einstein law does not explain
viscosity's temperature dependency. Can't diffusivity also be
explained?
What if the diffusivity description according to
http://en.wikipedia.org/wiki/Mass_diffusivity is used:
D=D0*exp(EA/R/T)
leading to
viscosity = kT/(6*pi*density*radius of molecule*D0*exp(EA/R/T)) which
seems like a more complex temperature dependency.

Since fluids are partially solid I suppose the molecule dimension in
the Sutherland-Einstein law should be replaced with the size of the
solid particle in the fluid?

I am not specifically interested in viscosity. I asked since I suspect
that retarded waves in fluids carry away energy and since viscosity is
usually determined by measuring the energy dissipation in flows I
suspect that viscosity is measured too high since it also includes the
retarded waves losses. I need to do some more calculations before I
can determine the size of the effect. The coordinate transformation
between retarded and non retarded flows seems to be the critical
point. They can be done in different ways. When they are determined I
just need to integrate the kinetic energy difference between the
inflow and outflow. Does anyone know about similar calculations?

David

Olin Perry Norton

unread,

Mar 11, 2010, 1:51:57 PM3/11/10

to

On 3/11/2010 6:28 AM, David Jonsson wrote:
> [snip]

> What if the diffusivity description according to
> http://en.wikipedia.org/wiki/Mass_diffusivity is used:
> D=D0*exp(EA/R/T)
> leading to
> viscosity = kT/(6*pi*density*radius of molecule*D0*exp(EA/R/T)) which
> seems like a more complex temperature dependency.
>

> [snip]
>

That wikipedia article says that the formula D=D0*exp(EA/R/T) is
for solids.

OPN