I endorse just about everything you wrote. Subtle variations in crystallinity, grain size etc and ability to react / dissolve makes such separations liable to larger than desired uncertainties, even if they are quantified well. 

Many decades ago we used a heavy liquid density separation, before the liquids were banned because of their carcinogenic effects. I used a mixture of tetrabromoethane and acetone. Starting with pure tetrabrome, I floated a tame pure dolomite on the surface and then added acetone until the dolomite just sank. We ran a lot of tests of known calcite-dolomite mixtures and were satisfied we could  separate them perfectly. 

We never bothered to do so, but if we were to have been very picky, we could have taken the  separated, floating fraction and added more acetone until the mixed calcite-dolomite particles, too fine to be separated, sank.

The point of this is not to air a now-unusable technique. We also tried preferential reaction / solublization of calcite but could never get as satisfactory a separation and without proving it, suspected that grinding would always produce some finer-than-average particles; which would also be inevitable in natural finely-intergrown samples.

What a pity we never wrote it up for publication.

Another technique that we did not use was that described by Blatt and Brown (1974). I think this is one of the best examples of a perfectly written, very short paper – well worth reading.


Blatt, H., and Brown, V. M., 1974, Prophylactic separation of heavy minerals: Journal of Sedimentary Petrology, v. 44, p. 260-261.

Dr. Max Coleman
Senior Research Scientist and Principal Scientist
Planetary Surface Instruments Group
NASA Jet Propulsion Lab, California Institute of Technology

(818) 393-6353 Voice
(818) 393-8887 Fax
(818) 687-7704 Cell
[log in to unmask]
My fun home lab work in lockdown

On 30 Oct 2020, at 04:50, Crowley, Stephen <[log in to unmask]> wrote:

Dear Greg,
I would urge you to read a paper by Yui & Gong (2003) Chemical Geology, 201, 359-368. Unfortunately most natural dolomites are relatively poorly crystalline, non-stoichiometric and fine-grained and are thus quite reactive, even with buffered acetic acid.
For chemical isolation of dolomite the use of Na-EDTA (pH 6.3) is more effective than acetic acid in removing calcite/aragonite while minimising dolomite decomposition (see Glover 1961, J. Sedimentary Petrology, 31, 622-626). This due to the fact that EDTA forms Ca complexes preferentially. Again, unfortunately this is a time-consuming process. All samples must be crushed to a standard grain size (<37 microns is a reasonable size) in order to expose “all” calcite in intergrowths of calcite and dolomite and to ensure that grain-size dependent effects on the rate of calcite decomposition are standardised. If you fail to crush to a consistent grain size then the occurrence of a few coarser calcite grains will significantly extend the reaction time required to remove all calcite.
Several messages have advocated the use of sequential phosphoric acid dissolution techniques which have the benefit of being relatively easy to use. In my opinion these methodologies are misleading. Authors reporting the development of these methods validate their procedures based on experiments using artificial mixtures of calcite and dolomite in which the dolomite component is well crystalline, near-stoichiometric and coarse-grained. This in no way reflects the reality of most natural (particularly sedimentary) samples where dolomite is much more reactive. These experiments generally lead to an over-optimistic impression of performance.
I’ve never tried to use a density separation technique. I would imagine that similar sample dependent (particularly textural) factors would greatly influence the effectiveness of this method.
A further consideration is the relative amounts (and the magnitude of the difference in isotopic composition) of calcite and dolomite in any mixture. Once the minor component occur in amounts below 20 wt.% there will be very little chance of obtaining a reliable isotopic composition. Obviously if you’re only interested in measuring the major component then your chances of obtaining a good outcome greatly increase as the minor component heads towards zero.
Steve Crowley
From: Stable Isotope Geochemistry <[log in to unmask]> On Behalf Of Cane, Gregory A
Sent: 29 October 2020 16:25
To: [log in to unmask]
Subject: Calcite-Dolomite mixed samples
Would anyone have literature regarding the separation of Calcite and Dolomite in samples for analysis of stable isotopes of O and C? I recall doing something of the sort years ago using acetic acid to dissolve the calcite fraction first, then a stronger acid to react the dolomitic fraction.