[Big-Daddy]

For one of my chemistry exams I've got a databooklet that lists effective pH ranges of indicators. This means the pH values, approximately, between which the indicator changes colour (so, you want this to fit inside the equivalence point region of the titration curve, so that end-point coincides).

From acid-base theory, I learnt that typically 1 pH unit either side of pK_ind will correspond to the extent of the visible range of the indicator (at 1 pH unit below, [HIn]=10[In^-] and if [HIn] dominates by any greater concentration ratio, you can't see the colour of In^- etc.).

However the ranges in the data-booklet don't correspond at all to this. I'll pick three indicators at complete random:

methyl orange-xylene cyanole solution: pKin=3.5, pH range 3.2-4.2
bromocresol green: pKin=4.7, pH range 3.8-5.4
alizarin yellow R: pKin=12.5, pH range 10.1-13.0

These ranges aren't even symmetrical about pKin (much less ±1 on either side). Not only that but they differ from pKin by different amounts from indicator to indicator.

Is there any way to make a theoretical treatment of this, or do I just have to accept these as experimental values and treat the theoretical modelling attempt (pKind ±1) unfit for practical application?

[Borek]

±1 is a good approximation. However, eye has different sensitivity to different colors, no wonder experimental values don't follow exactly.

[Big-Daddy]

I see. So in a particular case, it would be best to worry about experimentally-found pH range of colour change instead of pKind, because at pH=pKind the observed colour of the solution might still be nearly indistinguishable from that of just HIn or In^- (because that colour dominates the colour of the other)?

[Borek]

When you have a reliable experimental data it should be better than the general rule.