Re: Color Mixing (Pigment) question

[John Denker <jsd@AV8N.COM>]

jmsphys@juno.com wrote:
> Our physics book (h.s.) had a question about what color is obtained
> when pure green and pure blue pigment are mixed.  The book answer was
> "cyan pigment."  Can someone explain this please?

Michael E. rather charitably "explained" it as being a
misprint.

Less-charitable explanations include textbook authors
who, in blissful ignorance, apply additive color-mixing
theory to pigments, without having checked to see
whether this was the conventional wisdom, and without
having done the experiment.

> Why is the result not black?  The blue pigment filters out red &
> green light and the green pigment filters out blue & red light, so
> why isn't the result black?

Why not, indeed.

Vickie Frohne wrote in part:
> If you had a blue filter with a very narrow wavelength bandpass,
> stacked on top of a green filter with a very narrow bandpass, so that
>  the two wavelength ranges didn't overlap, then these filters would
> block the light entirely. But particles in mixed pigments are more
> side-by-side than blocking each other, so the reflected blue & green
> colors average.

That's a nice theory.  But have you done the experiment?

The results of some systematic experiments can be found at:
  http://www.handprint.com/HP/WCL/color8.html
I call particular attention to the figure about 3/4ths of the
way down that page, namely
  "chroma curves of mixtures around the color wheel"
  http://www.handprint.com/HP/WCL/IMG/satcurves.gif

It may take you a couple moments of thought to figure
out how to interpret the curves, but it's worth it.  For
the case at hand, if we take "blue pigment" to mean
ultramarine (PB23) and take "green pigment" to mean
"permanent sap green" then mixing them a very dark
color, less than 20%.  You can see this by looking at
where the sap-green curve passes above the "PB23" chip,
and/or where the ultramarine curve passes above the
"sap" chip.

This result does not even remotely resemble what you
would get from a proper cyan pigment.

This result is what you would expect on the theory that
mixing pigments is approximately like mixing inks or stacking
filters in series, i.e. subtractive color mixing.  The
mixing of real inks is not simple, and the mixing of real
pigments is even less simple, but treating pigments as
subtractive (not additive) is a very reasonable starting
point.

========

sampere wrote:
> We're not talking about overlapping filters here, we're talking about
> a fine, uniform layer of particles that reflect light.

That's making a number of unreliable assumptions, principally
 1) That the particles simply reflect light, and
 2) The light makes one bounce within the paint and
    then leaves.

In fact the particles are too fine to be really efficient
reflectors.  So the light visits quite a number of them
before being scattered out of the paint layer.

========

As usual, Michael E. and Larry W. made a number of good points.
  http://lists.nau.edu/cgi-bin/wa?A2=ind0405&L=phys-l&F=&S=&P=26102
  http://lists.nau.edu/cgi-bin/wa?A2=ind0405&L=phys-l&F=&S=&P=26327