Heterogeneous inclusion trapping explains assemblages of pure CO2 , non-aqueous inclusions.
Spherical fluid inclusion shape can indicate heterogeneous trapping.
Kingsley Burlinson : June 2014,
September 2015
The Problem:
Many authors who study fluid inclusions and find samples with only non-aqueous CO2 containing fluid inclusions propose formation models with deposition from gas phase pure CO2 fluids. But they forget that the quartz or other silicate, in which the inclusions are merely a trace occurrence, must have been transported by this fluid. Because there is no scientific data to support such silicate solubility in CO2 fluids, unsubstantiated transport theories are often proposed. These authors invariably ignore or misunderstand the effects of fluid inclusion trapping from heterogeneous fluids, such as CO2 bubble effervescence in a dominantly aqueous fluid, which can easily explain these observations. In particular, spherical non-aqueous CO2 filled inclusions indicate formation as bubbles within a dominant liquid phase, a clear case of heterogeneous trapping. In addition, the apparent absence of aqueous inclusions is not unusual for heterogeneous trapping and does NOT prove the absence of aqueous fluids during deposition.
There is considerable failure to recognize or understand the
effect of heterogeneous trapping in fluid inclusion studies. There
is some general discussion of this problem here on this site
and also an additional discussion relating
to the misinterpretation of pure CO2 filled
inclusions from the Ashanti gold belt, Ghana. The trapping
of inclusions from heterogeneous fluid systems such as an aqueous
fluid with some bubbles of CO2 gas, does not preserve the original
proportions of the fluid system components. The absence of
aqueous inclusions and overwhelming dominance of CO2
filled inclusions is merely a normal product of heterogeneous
trapping. Many authors propose deposition from a gas phase system
to explain the lack of aqueous inclusions, despite the fact that
no-one can properly explain the critically important transport of
quartz or other silicates in such pure CO2 fluids. An
alternative proposal, that the original water has diffused out of
the inclusions, is also often considered. But the most likely
explanation is heterogeneous trapping.
Identification of heterogeneously trapped inclusion assemblages
The recognition of heterogeneous trapping based solely on
temperature and composition measurements of inclusions is
difficult. But observation of the shape of inclusions (the
inclusion morphology) can provide important evidence of
heterogeneously trapped inclusions. If spherical (or distorted
spherical) gas-filled inclusions are observed, then it is
certain that heterogeneous trapping as a bubble within a
dominant liquid phase has occurred. This is because the only
way to form such a carefully ordered spherical shape is due to the
surface tension of the liquid in which the bubble exists. This
surface tension of the liquid minimizes the surface area of the
gas bubble, and this minimum configuration is a sphere. There is
no mechanism by which spherical gas-filled inclusions can form
during deposition from a homogeneous gas phase. Consequently the
existence of gas filled fluid inclusions of spherical, or
distorted spherical form proves that the system was dominated by a
liquid. This is in complete contradiction of the assumptions seen
in the literature that gas filled inclusions imply an entirely gas
phase deposition. In fact, if such rounded, gas-filled inclusions
occur, it actually disproves that assumption. (See some SEM images of opened fluid inclusions here.
They are quite angular)
Studies of fluids at the Ashanti gold mine, Ghana
This image of the fluid inclusions in the Ashanti gold deposits,
Ghana was published by A. Schmidt-Mumm, Oberthur et. al. in
their paper in Mineralium Deposita.

(From Schmidt-Mumm et. al., Mineralium Deposita, V32, 1997, p110)
Original caption: Example from the mineralisation at Ashanti (Ashanti fissure) showing the commonly high number of exclusively gaseous fluid inclusions in quartz.
Note the abundant round (spherical) gas-filled inclusions in this photograph, some of which I have indicated by red arrows. And many other inclusions show a strongly rounded shape. These spherical inclusions cannot have formed from a homogeneous gas phase and were trapped as bubbles within a heterogeneous, liquid dominant fluid. It is certain that the fluids at Ashanti were in fact liquid (probably aqueous) dominant and inhomogeneous with an unknown, but probably only small, amount of CO2 gas bubbles. The quartz and the gold in this deposit was not transported in pure CO2 as the authors suggest, but almost certainly in the abundant co-existing water. The original inference of transport and deposition from pure CO2 fluids made by these authors is incorrect because they failed to recognize that the fluid system was in fact heterogeneous. A PTX plot showing the heterogeneous deposition field is here. The authors report low formation temperatures, confirming that the fluids were within the heterogeneous field when deposition occurred. Their concern about the lack of aqueous inclusions is misguided as this is not unusual for inclusion assemblages trapped from heterogeneous fluids, as explained by Roedder and elaborated upon here.
This image of fluid inclusions in the same Ashanti gold deposit, Ghana was published by Oberthur, Schmidt-Mumm et.al. in Economic Geology.

(From Oberthur et.al., Economic Geology, V91, 1996, p292)
Original caption: Primary gaseous CO2 ± N2 ± CH4 inclusions in quartz, Ashanti mine.
These two papers are closely related variants of the same research with mutual co-authors.
Although far less noticeable in this second image, there seems to be a bias towards rounded inclusions, albeit misshapen. The inclusion morphology does not provide conclusive evidence of heterogeneous trapping in this sample, but does suggest the possibility of heterogeneous trapping.
Yao et.al also studied the same Ashanti gold deposit and these images are from their paper in Economic Geology, 2001.

(From Yao et. al., Economic Geology, V96, Nov. 2001, p1620)
Original caption: One of the major fluid inclusion types in vein quartz from the Sansu mine at Ashanti. 2P=primary monophase liquid CO2 - N2 ± CH4 fluid inclusions.

(From Yao et. al., Economic Geology, V96, Nov. 2001, p1617)
Original caption: One of the major fluid inclusion types in vein quartz from basin type granitoid-hosted gold deposits and prospects at Nhyiaso mine at Obuasi, in the Kumasi basin. 2P=primary monophase liquid CO2 - N2 ± CH4 fluid inclusions.
The single phase gas-filled inclusions in these 2 photographs are quite angular. Only a few photographs are available in this paper and perhaps the lack of spherical inclusions is due to observational bias in choosing large, easily studied inclusions. The morphology of these few inclusions alone cannot resolve the question of heterogeneous trapping. Yao et. al. also found plenty of aqueous fluid inclusions and concluded from their work that the CO2 rich inclusions are the result of entrapment from unmixed fluids resulting from phase separation of an originally homogeneous aqueous-CO2 fluid. That is to say, trapping from a heterogeneous system of liquid water containing exsolved CO2 bubbles.
A study of fluids at the Campbell-Red Lake mine, Ontario, Canada
Guoxiang Chi et. al. studied the Campbell-Red lake gold deposit (Ontario, Canada) and found H2O poor, CO2 dominated fluids.

(From Guoxiang Chi et. al., Mineralium Deposita, 2006, V40 #6-7, p726)
Abbreviated original caption: Various clusters of carbonic fluid inclusions in quartz with homogenization temperatures to liquid. From the Red lake mine, Ontario, Canada.
The authors state that aqueous-carbonic inclusions were only rarely observed and are limited to the low grade gold zones. There is only the above single suitably scaled photograph of fluid inclusions in this paper, and all the inclusions shown are wholly CO2. These inclusions are quite angular and their morphology does not help to determine if the parent fluid was homogeneous or heterogeneous. But this single photograph may be a specially chosen case and may not be representative of the deposit. The angular shape may be the result of post entrapment re-crystallization caused by an invisible thin layer of water in the inclusion, as discussed here.
However the authors did carefully consider 3 possible methods of origin.
- Deposition from a pure CO2 gas phase
- Heterogeneous trapping from an aqueous phase, with inhomogeneous CO2 bubbles.
- Trapping of aqueous - CO2 inclusions followed by preferential leakage of the water to give CO2 only inclusions.
However, both of these concerns are misguided as the authors have not understood that heterogeneous trapping does NOT trap the fluid components in proportion to their abundance, as discussed here. It is not only possible but quite likely that the aqueous phase may not be trapped at all. The trapping of an inhomogeneous gas bubble and the parent aqueous fluid happen by completely different, independent mechanisms which do not act in unison and may in fact act in mutual opposition, excluding the aqueous inclusions while simultaneously efficiently trapping the CO2 gas inclusions.
In their discussion of the proposed formation origin by homogeneous deposition from a pure CO2 fluid, the authors completely ignore the extremely serious problem of transporting and depositing large quantities of silica from a pure carbonic fluid. Diabolically, this silica deposition is a far more serious issue than the unsurprising lack of aqueous inclusions in heterogeneously trapped fluid inclusion assemblages, which they (wrongly) think is the major problem!
Rounded gas inclusions in a petroleum reservoir
This image shows both gas filled and aqueous fluid inclusions from
the fracture filling cement in a gas reservoir. Note that many of
the gas filled inclusions, labeled g, are well rounded, near
spherical. Such gas inclusions would certainly have been trapped
from a heterogeneous fluid system which has resulted in near
spherical inclusion morphology.
(From András Fall et. al., AAPG Bull., V96 #12, 2012)
Heterogeneously trapped gas inclusions within fracture cement in sandstone, formed at about 160 C.
Fluid inclusions are usually angular, not spherical in shape. Spherical gas-filled inclusions indicate heterogeneous trapping.
Many people argue that spherical fluid inclusions are normal, but
this is not so. The numerous photographs of fluid inclusions in
the literature almost invariably show angular irregular or
negative crystal shapes. Authors nearly always comment that the
inclusions they studied were of "negative crystal" shape.
Spherical shaped inclusions are quite uncommon and are the result
of unusual fluid conditions. They are not formed from homogeneous
fluids under normal conditions. The most reasonable explanation
for spherical inclusion shapes of gas-filled inclusions is that
they are the result of trapping of gas bubbles from within a
heterogeneous fluid. Think of the gas bubbles in your champagne -
a heterogeneous system in which liquid dominates. These bubbles
are trapped by interference with the growth of the host mineral, a
completely different mechanism to the formation of inclusions from
homogeneous fluids. (Discussed
here)
These are scanning electron microscope images of opened fluid
inclusions within fluorite and their extremely angular shapes are
quite obvious. Note also the fine striations and steps on the
inclusion walls, resulting from crystal growth irregularities.
Even these minute irregularities are not eliminated by subsequent
re-growth or metamorphism. Inclusions do not become "more
spherical" with aging, but perhaps more angular under the
influence of crystal lattice forces.
Figure 6. Fluid inclusions present in fluorite at the Blockspruit
prospect. SEM image of opened fluid inclusion of type 1 (V+L+S).
FROM: REE Mineralization of the Blockspruit Fluorite Prospect, Bushveld Granitic Complex, South Africa: Geochemical, Mineralogical and Fluid Inclusion Studies
BY: Mihoko Hoshino*, Robert Moritz, Maria Ovtcharova, Yasushi Watanabe, Jorge Spangenberg & Benita Putlitz.
(Poster presentation and abstract at SGA conference, Nancy, France, 2015)
The following 4 SEM images of opened fluid inclusions in quartz
are from Lawrence et al.
Note the preservation of the quartz crystal shape of the
inclusion in C.


These all show good preservation of angular features of the fluid inclusions, with little or no post-entrapment "rounding".
FIG .6. Secondary electron microprobe (SEM) images of opened, quartz-hosted, multi-phase H2O-CO2-NaCl-FeCl2 type 4-e inclusions from Gara and Yalea North. A) to D) Images of the various mineral phases observed in open inclusion cavities.
FROM: A Fluid Inclusion and Stable Isotope Study at the Loulo Mining District, Mali, West Africa: Implications for Multifluid Sources in the Generation of Orogenic Gold Deposits
BY: DAVID M. LAWRENCE, PETER J. TRELOAR, ANDREW H. RANKIN, ADRIAN BOYCE, & PAUL HARBIDGE
IN: Economic Geology, V108 #2 March 2013, p238 (membership required for full text)
The following SEM image shows opened fluid inclusions as negative crystal shapes in sphalerite. Note that the inclusions are angular.

FROM: Fluid inclusions in sphalerite as negative crystals - a case study.Other SEM images of opened fluid inclusions have also been published and these also show angular, not spherical inclusion shapes.
BY: Ivan K. BONEV and Kalin KOUZMANOV
IN: European Journal of Mineralogy(June 2002), 14(3):607-620 (membership required for full text)
*** This is a small scale copy as the journal is "restricted access" to members only.
Discussion
The recognition of heterogeneous inclusion trapping is an important issue. Boiling, single component (H2O) fluids are one case of heterogeneous trapping, which is well studied and can be recognized from the coexistence of inclusions of very different liquid-vapour ratios in the same fluid inclusion assemblage. But immiscible gas effervescence (usually of CO2) from a multi-component fluid (H2O + CO2) is also a heterogeneous system and recognition that this is a heterogeneous fluid is more difficult. However, the gas bubbles in such heterogeneous systems will be spherical or very rounded because of the surface tension of the host liquid phase. When these gas bubbles are trapped, the morphology of the gas-rich fluid inclusions will reflect their origin as spherical gas bubbles within a liquid host phase. The resulting fluid inclusions may not be perfectly spherical because various other processes, including post-entrapment changes, may affect the shape. Consequently, well rounded and/or spherically shaped gas-filled inclusions are an indicator that heterogeneous trapping has occurred.But not all gas filled inclusions are spherical, as seen in some of the images above. Are such angular gas filled inclusions evidence of formation from a homogeneous gas phase fluid, or could they also form from heterogeneous trapping of a liquid - gas mixture? I suggest that it is also possible for heterogeneous trapping to form sub-rounded or somewhat angular gas filled inclusions and that even these inclusions may be formed in a liquid rather than a pure gas phase. The process of trapping an inclusion is not instantaneous and it takes a long time for the enclosing silicate mineral to deposit around the gas bubble until it is completely sealed. The silicate molecules will preferentially deposit in locations controlled by the crystal structure, but are blocked by the presence of the gas bubble. Depending of the silicate saturation level of the liquid, diffusion rates, fluid flow rates and numerous other variables, both the crystal structure and bubble interference will control deposition and affect the overall shape of the inclusion. In addition, the gas bubble may not remain in place depending on fluid turbulence and bubble coalescence. Bubbles may move about or leave completely to be later replaced by another bubble. This would allow some crystal faces to partially develop before the final sealing of the inclusion occurs, leading to a mixture of angular and rounded shapes to the inclusion cavity.
Inclusions which appear to be filled with a single gas phase most likely contain an invisible layer of liquid water adhering to the inclusion walls. The film of water in these inclusions will allow recrystallization of the silica over geological time which will lead to increasing angularity of the inclusions. Instead of expecting inclusions to be rounded, the real issue is to understand why the inclusions are not angular, because angular cavities have lower crystal surface energy than spherical cavities. The lowest surface energy of crystals (or cavities within crystals for fluid inclusion shapes) occurs with surfaces conforming to the crystal shape. Spherical fluid inclusions have higher surface energies and will re-crystallize to give crystal structure controlled surfaces, becoming more angular, if there is enough water in the inclusion to allow silica dissolution, transport and re-precipitation. This mechanism cannot happen in non-aqueous inclusions, which will therefore remain spherical.
In contrast, homogeneous deposition would normally give rise to negative crystal shaped or irregular inclusions, but not spherical or strongly rounded inclusions. The above SEM images of opened fluid inclusions indicate that any post-entrapment changes are dominated by the crystal structure, with the inclusions retaining angular and euhedral (negative crystal) shapes and they do not become rounded.
Conclusions
The shape of fluid inclusions can indicate the occurrence of
heterogeneous trapping. Spherical or well rounded gas-filled
fluid inclusions indicate that the inclusions were
trapped as a gas bubble within a dominantly liquid host phase.
Gas filled inclusions do not necessarily imply that they formed
from a solely gas phase fluid. The most common source for these
inclusions is in fact from heterogeneous trapping of subordinate
gas bubbles within a dominant liquid phase. Such heterogeneous
trapping can be difficult to recognize during fluid inclusion
studies and is frequently not recognized and misunderstood. The
shape of the inclusions may assist in recognition of heterogeneous
trapping because spherical or well rounded inclusions must have
formed by the trapping of a gas bubble within a liquid host phase.
The spherical shape of inclusions is caused by the surface tension
of a liquid which hosts a heterogeneous gas bubble phase, and
which constrains the gas bubble to have the minimum surface area
for its volume, i.e. a sphere. This bubble is then trapped by
interference with silicate deposition of the host mineral,
typically the quartz being studied. Other silicate deposition
mechanisms do not produce spherically shaped inclusions and
instead give inclusion shapes controlled by the crystal structure
of the host mineral, giving rise to negative crystal shaped
inclusions, or else completely random shapes if deposition is
fast, or some combination of these shapes.
A common misconception is that heterogeneous trapping should trap
fluids from both of the phases present. But the trapping mechanism
of the separate phases are different and completely independent as explained here, and there is
absolutely no need for both phases to be represented in the
resulting fluid inclusion assemblage. In addition,
experimental restrictions such as the need to work with large
inclusions and to only study inclusions whose paragenesis can be
determined also lead to the failure to recognize a coexisting
aqueous phase even when those inclusions do exist in the
inclusion assemblage.
Disproportionate inclusion trapping from heterogeneous fluids
Discussion of the differences between boiling and immiscible gas heterogeneous systems.
Thermodynamic experiments show that gold is NOT transported in anhydrous CO2 fluids
Back to the Viewpoints menu