Applied mineral exploration methods, hydrothermal fluids, baro-acoustic decrepitation, CO2 rich fluids
Newest Topics:
For the latest news, see the NEWEST TOPICS page.

Google is too dumb to let me put the list of news in this column and falsely claims that all my pages are self-duplicates.


Google's so-called "Artificial Intelligence" is an abuse of the concept of intelligence!

Disproportionate inclusion trapping from heterogeneous fluids

An explanation of CO2 fluid inclusion populations lacking aqueous inclusions within a quartz host mineral.

Kingsley Burlinson, September 2013

The Problem

In 1997  Schmidt-Mumm et. al. published research of quartz from gold deposits in Ghana in which they found abundant pure CO2 fluid inclusions, but failed to identify any associated aqueous inclusions. They surmised that the gold must be transported in pure CO2 fluids by an unknown chemical mechanism. However, such an assertion completely fails to recognize or explain the presence of quartz, which comprises some 99.999% of the mineral material transported in the hydrothermal fluids. (Discussed here) It is incorrect to assert without any supporting evidence whatsoever that the gold, and by inference the quartz also, must have been transported in pure CO2 fluids lacking any water. This paper is being referenced by other research work in the literature as proof that gold can be transported in pure CO2 fluids. But this paper alone is completely inadequate to support the hypothesis of gold transport by pure CO2.

The observations can be explained as merely disproportionate trapping of CO2 and aqueous phases from a very normal heterogeneous, aqueous dominated fluid with some CO2 present as an immiscible gas phase. Both the quartz and the gold are probably transported in the dominant aqueous fluids which, due to inclusion trapping conditions, were only poorly trapped as inclusions, while the co-existing separate immiscible CO2 fluid phase was efficiently trapped as fluid inclusions. Heterogeneous trapping is a well known phenomenon and has been discussed at length by E. Roedder (1984). But it has been forgotten or ignored in some research papers which can lead to serious misinterpretation of fluid compositions.

Skip to the summary


Special note: This discussion page focuses on heterogeneous fluids which contain an immiscible gas phase, specifically CO2 here. But another common heterogeneous system is of water vapour within liquid water, found in boiling epithermal deposits and this very different heterogeneous fluid system is discussed here.

CO2 rich non-boiling heterogeneous systems

Most fluid inclusions are trapped under conditions where the parent fluid is a single homogeneous phase due to the high temperatures and pressures during mineral formation. The study of fluid inclusions usually relies on this homogeneity to be able to deduce the original formation conditions from a fluid inclusion which has subsequently separated into 2 or more phases at room temperature. Frequently this is a liquid water phase with a low pressure vapour bubble left behind due to condensation of the water. There may also be some CO2 in the vapour bubble or even a separate CO2 liquid phase. This condition is so common that it is easy to overlook the cases in which the parent fluid was heterogeneous. In particular, CO2 and water have a large field of immiscibility and will exist as a heterogeneous mixture at lower temperatures as shown in this P-T-X diagram.

PTX diagram for CO2 and water

Fluid inclusions formed within the 2 phase immiscibility region will have widely differing compositions depending upon which phase they have trapped, the gas alone, the liquid alone or  a combination of both phases. Trapping from such heterogeneous fluids is not uncommon and often occurs in fluid inclusion populations. But unless recognized, studies of such fluid inclusion assemblages will give misleading results as the inclusion assemblages do not preserve the original component proportions of the fluid system.

E. Roedder (1984, page 29) states:

"The inclusions that are trapped in crystals growing from a heterogeneous system of two fluid phases may reflect both nucleation and wetting phenomena, and very erroneous conclusions can be drawn if these and other features are not considered. As an example, consider bubbles of CO2 floating as the dispersed phase from a crystallizing basaltic melt. The growing crystals of one (or all) of the phases may become coated with tiny gas bubbles, each of which shields that part of the surface with which it is in contact from further growth, and so it becomes enclosed. Note first that this process may yield a large number of primary gas inclusions, without any melt, in a crystal that actually grew from a melt. This would hardly mislead one into assuming that the silicate crystals grew from a CO2 fluid. But the identical phenomenon, involving steam bubbles forming on crystals growing from a hydrothermal fluid and trapped without any of 
the fluid, may easily be misinterpreted as evidence that the crystals grew from a low-density fluid and hence were one more example of "pneumatolysis." The distinction can be of more than just academic importance. Such a nucleation process should be particularly effective if sudden small pressure drops cause periodic effervescence of dissolved gas, or boiling of the fluid, as has been suggested by Barabanov (1958) for some so-called "pneumatolytic" tungsten deposits."

Inclusions trapped from multi-phase inhomogeneous fluids do not preserve the ratio of the original phases. If crystal growth rates were slow, there may be only very few crystal defects in which to trap inclusions of the aqueous phase. It is entirely possible that even the major phase of the parent fluid may not be trapped in inclusions at all and only a lesser phase, such as CO2 gas bubbles, is trapped from a dominantly aqueous system. Failure to understand this can lead to incorrect conclusions as stated by E. Roedder, 1984, page 34:

"Incorrect inferences are frequently drawn in the inclusion literature that the amount of vapour phase trapped, or the ratio of the number of inclusions of vapour to those of liquid phase, give an indication of the relative amounts of the two phases present in the original heterogeneous system. A single tiny inclusion of gas phase indicates gas saturation just as well as a million such inclusions, but the amount and number of such inclusions in a sample are merely a result of the vagaries of the inclusion-trapping processes discussed above, and do not give any valid indication of the phase ratios in the original two-fluid system at any given time. The trapping of gas inclusions may even give misleading data on the sequence of phase changes with time. Thus if one of a series of zones of primary inclusions in a given crystal is made up of gas-rich inclusions, the common interpretation given in the literature is that gases "played a major role" in the deposition of that zone. All it may really mean, however, is that some minor amounts of gas bubbles formed during the growth of that zone and were preferentially trapped."
The process of trapping fluid inclusions is not necessarily simple and is frequently ignored when studying them. However these processes are particularly important when dealing with trapping from a heterogeneous fluid.

An explanation of disproportional inclusion trapping from heterogeneous fluids.

The processes of forming fluid inclusions can be classified into 2 major mechanisms.
  1. Crystal defect type inclusions are growth irregularities or defects in the crystal, in which some fluid is accidentally trapped and sealed in by continued crystal growth around the defect. Such irregularities or defects are predominantly random and depend on growth rate, saturation levels, flow rates, impurities and other factors. The fluids trapped in such inclusions will indicate the composition of the parent fluid component from which the host crystal grew.
  2. Interference type inclusions are obstructions to the crystal growth, typically caused by an inhomogeneous phase such as a bubble of an immiscible gas within a liquid. Such an obstruction can become attached to the growing crystal surface, forcing the crystal to grow around it. The obstructing phase will become trapped as a fluid inclusion, but its composition is not indicative of the fluid component from which the host crystal is growing. There may or may not be additional defect type inclusions which contain the other component(s) of the heterogeneous fluid system. As stated by Roedder (quoted above), the bulk fluid composition of the parental heterogeneous fluid cannot be deduced from such inclusion assemblages because the components of a multiple-phase system are not trapped in proportion to their abundances.
Conditions which favour the formation of inclusions by one of these mechanisms may simultaneously suppress inclusion formation by the other mechanism, which results in extreme disproportional trapping of the heterogeneous fluid phases. This can be mistaken for deposition from a single phase fluid as there may be almost no evidence of inclusions from the other phase! Quartz veins growing from slowly moving fluids will facilitate the attachment of bubbles and at low silica supersaturation concentrations, will also result in very few crystal defects. The resulting inclusion assemblage will be dominated by the subordinate inhomogeneous gas phase with very few or no inclusions of the dominant aqueous phase from which the quartz is actually depositing. High flow rates and turbulence would have the opposite effect as bubbles are swept away and crystal defects become common. The resulting inclusion assemblage will consist of only the quartz depositing aqueous phase (in crystal defect type inclusions) with few or no inclusions of the inhomogeneous gas phase. It is quite possible that inclusion populations represent only one of the compositional extremities of the heterogeneous fluid system, rather than  representing all of the heterogeneous fluid phases or approximating the bulk fluid composition of the multiple-phase fluid system.

It is quite possible to grow crystals with almost zero growth defects. An example is the semiconductor industry which routinely grows very large silicon crystals which are free of defects, a necessity for electronic chips. Although this is done from a melt, it would be possible to grow quartz with very few defects (and thus very few fluid inclusions) from a hydrothermal fluid if conditions were right. Such crystals with very low inclusion abundances do occur in nature (eg in Brazil). But it would be wrong to infer that these crystals grew without fluid involvement merely because of the lack of inclusions.

During quartz vein formation, pressure reductions (for example, during earthquakes) could cause the effervescence of bubbles of CO2 from a homogeneous aqueous-CO2 fluid whose pressure and temperature was close to the immiscibility field. If the fluid flow was gentle and silica deposition slow, these bubbles would be efficiently trapped as interference type inclusions, but at the same time crystal defects would be rare, resulting in very few inclusions of the aqueous phase as they only occur in the rarely formed crystal defect type inclusions. This disproportionate trapping of the 2 fluid phases present could lead to misinterpretation of the bulk fluid composition.

Experimental limitations

The failure to observe additional fluid phases in a heterogeneous fluid system is exacerbated by the standard protocol for the study of fluid inclusions. It is normal to completely ignore very small inclusions (less than about 5 microns across) as it is very difficult or impossible to make precise observations of the fluid phases within such small inclusions. And the many inclusions whose primary / secondary paragenetic origin cannot be ascertained are by default assumed to be secondary and are ignored. Only large, demonstrably primary inclusions are used in most fluid inclusion studies. But this results in the vast majority of inclusions in a sample being ignored for experimental practicality reasons. These ignored inclusions are not necessarily irrelevant and may actually prove that the fluid was heterogeneous.

Sealing up the inclusion

There is so much focus on the fluids within the inclusion that the mechanism of sealing the inclusion to actually trap the fluids being observed is usually overlooked. But this is a critical process without which the fluid inclusion would not even exist. This requires simultaneous deposition of the host mineral, usually quartz, either from the fluid being observed, or from one which co-existed with that fluid and which may not be trapped in inclusions or recognized. This fluid must have been transporting and depositing the host mineral phase at the time the inclusion was trapped. It is logically incorrect to propose a fluid in which the host mineral is insoluble and the observation of only such fluids implies the existence of additional fluids which have been overlooked and ignored. (Secondary or pseudo-secondary inclusions might form by solid state re-crystallization of a pre-existing host mineral with little or no fluid phase transport.)

E. Roedder (1984, page 35) briefly mentions the problem of inclusion fluids in which the host mineral is insoluble:
"The enigma of inclusions of fluid having zero solubility for the host. One fascinating but seldom considered paradox in the formation of inclusions from immiscible fluid pairs concerns the exact mechanism whereby fluid inclusions can become sealed even though they apparently contain only a fluid in which the host crystal is essentially insoluble. The usual explanation is that the fluid that is eventually trapped was present as the dispersed phase (i.e., globules) in a continuous phase, and that crystallization of the  host mineral took place only from the continuous phase."
Clearly, there must be a fluid present which is capable of transporting and depositing the host mineral phase in order for a fluid inclusion to be enclosed and trapped. Proposing a pure CO2 fluid as the only fluid present in a quartz host mineral, as Schmidt-Mumm et. al. have done, requires that a mechanism for the transport and deposition of silica from this same fluid must be explained. As silica is not known to be transported in pure CO2  fluids, then there must be another (probably aqueous) phase present and the fluid system must be heterogeneous. The failure to identify or consider this additional aqueous phase is a serious experimental oversight.


Quartz containing essentially pure CO2 inclusions and lacking aqueous inclusions can be explained as the product of deposition from a heterogeneous fluid comprised of mostly water with a few CO2 bubbles. Slow growth from a heterogeneous, aqueous, silica-bearing fluid which also carried CO2 bubbles would facilitate attachment of the small CO2 bubbles to the quartz surfaces while simultaneously suppressing the formation of crystal growth defects. This would produce quartz with many CO2 inclusions of the subordinate gas phase which are formed by obstruction of the crystal growth, but very few aqueous phase inclusions of the dominant silica bearing fluid because of the lack of crystal defects necessary to trap them. The gas inclusions are likely to be spherical in shape as their shape is controlled by the surface tension of the liquid host fluid.

The quartz observed in Ghana by Schmidt-Mumm et. al. (1997) is best explained as the product of disproportionate trapping from a heterogeneous fluid. There is no need to suggest the transport of gold and silica by pure CO2 fluids as the system was actually dominated by normal aqueous fluids, which simply happen to be poorly trapped in inclusions. This has mislead the authors into thinking that no aqueous fluids were present.  Although Schmidt-Mumm et. al. did consider the possibility of heterogeneous trapping, they wrongly presumed that this could not explain the lack of aqueous inclusions. But as Roedder (quoted above) has stated, fluids trapped from heterogeneous fluids are not trapped in proportion to their phase abundances. It is entirely possible that aqueous fluid inclusions are absent simply because of gentle growth conditions and lack of crystal defect type inclusions, while interference type inclusions (the CO2 gas phase in this case) will simultaneously be efficiently trapped.

In their paper, Schmidt-Mumm et. al. published a photograph of the fluid inclusion assemblage they studied. This shows abundant spherical or near spherical inclusions which is in fact clear evidence that this assemblage of inclusions was trapped from a liquid dominant host phase as heterogeneous bubbles, as discussed in detail here. (Spherical inclusions are formed due to the surface tension of a host liquid phase and cannot be formed from a purely vapour phase or pure CO2 fluid.)

In addition, the failure to explain the transport and deposition of the host mineral (quartz) in pure CO2 fluids confirms that a concurrent silica-bearing aqueous phase must have been present. This aqueous phase quite probably also transported the gold and there is no need whatsoever to propose the transport of gold in pure CO2 fluids.

In the absence of proper consideration of the effects of disproportionate trapping from a heterogeneous fluid and failure to explain the critically important transport of the host mineral quartz, this paper fails to provide any reasonable evidence for the "new category of ore forming fluids" that it postulates.

Fluid inclusion populations trapped from heterogeneous fluids can easily be unusual. The fluids trapped are in no way constrained to be trapped in proportion to their abundance in the parent fluid. In fact it is entirely possible that disproportionate trapping is so extreme that only one of the 2 (or more) fluid phases can be recognized in the inclusion assemblage. Some work (Bodnar et. al, 1985) has proposed that heterogeneous fluids can sometimes be recognized from a bimodal distribution of gas-rich and liquid-rich phase ratios in fluid inclusions within a single inclusion assemblage. However that study was specific to boiling aqueous epithermal systems with essentially no CO2 and is not relevant to the gas rich systems discussed here. Gas rich heterogeneous fluids containing an immiscible gas component phase are unlikely to behave the same as boiling aqueous systems, even though both are technically classified as heterogeneous fluid systems and they should not be expected to give the bimodal phase ratio distributions which can occur in boiling aqueous fluid systems.

Considerable care needs to be taken when interpreting such heterogeneous fluid inclusion assemblages to avoid serious misinterpretation of the composition of the inferred parent fluid systems.


An additional discussion disputing the existance of auriferous quartz formed from only non-aqueous pure CO2 fluids.

Discussion of the differences between boiling and immiscible gas heterogeneous systems.

Edwin Roedder (1984)  Fluid Incusions Reviews in Mineralogy, Volume 12, Mineralogical Society of America.

A. Schmidt-Mumm et al. (1997)   High CO2 content of fluid inclusions in gold mineralisations in the Ashanti Belt, Ghana: a new category of ore forming fluids? Mineralium Deposita, 1997, V32, p107-118

R. J. Bodnar, T.J. Reynolds and C.A. Kuehn, 1985. Fluid inclusion systematics in epithermal systems,   In: Geology and geochemistry of epithermal systems, Reviews in Economic Geology, Volume 2,  Society of Economic Geologists.

Back to main contents