Main Menu Programme Author Index

EXPERIMENTAL SIMULATION OF FLUID-MINERAL INTERACTION USING A

CONTINUOUS FLOW HYDROTHERMAL APPARATUS
Bruce W. Mountain1 and Romain Sonney1
1
GNS Science, Wairakei Research Centre, 114 Karetoto Rd., Taupo 3377, New Zealand
[email protected]

Keywords: Experimental geochemistry, flow-through In this contribution, we described results from a set of fluid-
apparatus, water-rock interaction, silica scaling, re- mineral interaction experiments undertaken using a
injection continuous-flow hydrothermal apparatus located at GNS
Wairakei.
ABSTRACT
2. EXPERIMENTAL APPROACH
Experimental simulations of fluid-mineral interactions have
been conducted using a continuous flow-through The high T,P flow-through apparatus is located at GNS
hydrothermal apparatus capable of temperatures up to Science, Wairakei Research Centre. It was constructed by
400oC, pressures of 500 bars and flow rates as low as 1.5 Coretest Systems Ltd., San Jose, California. It is able to
ml/day. simulate fluid-mineral interactions under hydrothermal
conditions reaching up to 400oC and 500 bars at a flow rate
In the experiments described, New Zealand greywackes ranging from 0.06 – 900 ml/hr (Fig. 1)
were reacted with distilled water and re-injection brines to
study the effects on brine chemistry and secondary mineral
formation.

Experiments using unaltered greywacke, reacted with

distilled water and re-injection brine, showed the formation
of clay minerals and possible zeolite phases. Those
conducted at re-injection temperatures using re-injection
brines showed the formation of amorphous silica deposits
whose textural characteristics were affected by the amount
of silica polymerisation.

The results from these experiments demonstrated the

considerable possibilities for this method. Define your
question, we can simulate it.
Figure 1: Schematic diagram of flow-through apparatus
1. INTRODUCTION Continuous flow is achieved by a double-piston metering
Fluid-mineral interaction studies are undertaken to elucidate pump that pumps distilled water into the titanium
the process through the use of computational models or accumulator. This contains a diaphragm that separates the
experimental analogs. These can simulate fluid-mineral distilled water from the study fluid. In this way corrosive
reactions which are occurring in the natural environment and/or saline fluids can to be used without damaging the
allowing us to predict or explain processes occurring in stainless steel metering pump. The mineral/rock material is
infrastructure and in the subsurface. held inside a titanium sample holder located inside the
autoclaves. Pressure is maintained by an electronic back
Computational methods to study fluid-mineral interactions pressure regulator. The pore fluids samples are collected on
include: static, closed-system thermodynamic models such the low pressure side of the back pressure regulator.
as activity-activity diagrams; more complex, open-system
models that have become available through the use of Pore fluid samples were analysed by ICP-OES for cations,
personal computers and even more complex computer IC for anions, and pH by potentiometry (all at GNS
modeling software that can incorporate kinetic data, e.g. Wairakei). Run products were studied microscopically by
GWB. Although computational models are convenient to SEM at Victoria University Wellington and the University
use, they are based on the premise that the thermodynamic of Auckland.
data used is comprehensive and correct. This is an erroneous
assumption. 3. RESULTS

The alternative to computational models is the use of direct 3.1 Waotu1 (Greywacke – Distilled Water)
experimentation to reproduce fluid-mineral interactions in
the laboratory. This approach simulates the processes This experiment reacted Waotu greywacke (Late
occurring in a hydrothermal system using actual rock and Jurassic/Early Cretaceous Manaia Hill Group) with distilled
fluid allowing observation of the chemical and mineralogical water at 25 - 210oC, 35 bar. This rock was chosen as it is
changes occurring subsurface. It allows intrinsic similar to the basement rock in the geothermal areas of the
incorporation of thermochemical parameters that may not be TVZ. The greywacke consists of clasts of siltstone, basalt,
available to modeling programs. It’s the closest that a quartz, plagioclase, pyroxene and hornblende in a matrix of
geochemist can get to actually being there. quartz, feldspar, chlorite, illite, calcite and pyrite. Distilled
water was used to see the effects of the rock on a pristine
1
New Zealand Geothermal Workshop 2011 Proceedings
21 - 23 November 2011
Auckland, New Zealand
fluid. The initial flow rate for this experiment was 1 ml/hr. monomeric silica content of the input brine showed that it
Effluent samples were collected daily. had polymerised and that only ~180 ppm of the SiO2 was
monomeric.
During the initial 25oC period elevated concentrations of Ca
and Mg and alkaline pH values were found. This is due to Samples collected during the inital room temperature period
the dissolution of calcite from the greywacke matrix. Trace showed increases in pH, Ca, Mg and Sr, consistent with the
amounts of Na, K, Al, Sr and SiO2 were also detected. Upon dissolution of the calcite from the greywacke. SiO2
the temperature shift to 210oC, a sharp increase in the concentration decreased by 70 ppm during reaction with the
concentrations of K, Na and SiO2 were measured. As, B, Fe, greywacke at room temperature. After the temperature shift
Li and SO4 also became detectable but only at trace levels. to 154oC, pH decreased to 6.9 and then reached a steady
Ca and Mg dropped to below detection limit. Quench pH state value of approximately 7.5. Other major elements,
stabilise at a value of 7.5. Quartz saturation temperature of excluding SiO2, remained relatively unchanged from brine
216oC was calculated based on the stabilise SiO2 concentrations. SiO2 decreased to a steady state value of 640
concentration agreeing well with the experimental ppm. This concentration represents a Tquartz = 276oC,
temperature. After six days Ca concentrations increase again Tchalcedony = 275oC and Tamorph SiO2 = 153oC. Although the last
to a steady state value of ~1.9 ppm. A decrease in flow rate temperature agrees with the experimental temperature, it
to 0.5 ml/hr did not affect dissolved component must be remembered that the silica in the input brine is
concentrations or pH. This suggests equilibrium with partially polymerised and thus measured silica is not
secondary mineral phases had been reached. representative of the actual monomer present in the
experiment.
SEM examination of fresh unreacted greywacke showed
clean unreacted surfaces of quartz, feldspar, pyroxene clasts The sample holder contains about 25 g of crushed, sieved
surrounded by the fine-grained matrix (Fig. 2a). In contrast, (0.5 – 2.0 mm) and cleaned greywacke. The input fluid
run products from this experiment showed significant enters the sample holder at the bottom. The fluid is furthest
dissolution of quartz and feldspar. Greywacke clasts also from equilibrium with the rock at the entry point. At the
had a semi-continuous coating of secondary phases. EDS slow flow rate utilised, the flow is not turbulent and mixing
examination of a platy, mucaceous phase indicates a Ca - Na is mininal. This means that the brine and rock see the
- K-bearing aluminosilicate interpreted to be smectite (Fig. maximum exchange of components at the bottom of the
2b). Also present was a euhedral phase that EDS sample holder while at the exit the brine and rock are at their
examination showed to be a Ca-Al-Si bearing mineral closest equilibrium.
(zeolite?). Rare secondary calcite was also identified.
Run products from the K50-1 experiment were split into
3.2 Waotu2 (Greywacke – Reinjection Brine) portions based on their position in the sample holder. SEM
examination showed that those at the bottom were coated
To contrast and compare with the simulation with distilled with a continuous covering of amorphous silica (Fig. 4a). No
water, in this experiment, Waotu greywacke was reacted other secondary minerals were identified. The coating
with reinjection brine from Wairakei Power Station. The T,P covered all clast types and the matrix material. It consisted
conditions were 25 – 204oC, 35 bar. At room temperature, of a continous very fine-grained silica layer encompassing
elevated pH and higher concentrations of Ca, Mg and Sr larger spherulitic shaped bodies (Fig. 4b). Greywacke
were found consistent with dissolution of calcite. Upon the fragments from the top of the sample holder were free of this
temperature shift to 204oC, sharp declines in pH, Mg and As coating and appear similar to the unreacted material.
occurred. Other components, excluding SiO2, remained at
their re-injection brine concentrations and appeared to be 3.4 K50-2 (Re-injection Aquifer – Depolymerised re-
unaffected by fluid-mineral interactions. SiO2 concentration injection brine)
in the input brine is 567 ppm and at room temperature this
concentration dropped to 530 ppm. After the temperature The experimental set-up for this experiment was similar to
shift to 204oC, a steady state SiO2 concentration of 440 ppm K50-1, however, in this case, two ovens were used. The re-
was reached. This concentration represents a Tquartz = 241oC, injection brine was pumped into the first oven at 350oC. The
Tchalcedony = 231oC and Tamorph SiO2 = 116oC. residence time in the empty sample holder in this oven is
approximately 50 hours. This is sufficient time for the
The surface of reacted grains from this experiment showed a polymeric silica in the brine to be depolymerised afterwhich
different assemblage of secondary minerals. Rosettes of a it flows in the second oven containing the re-injection
Mg-Fe-Al bearing silicate, most likely chlorite, were found aquifer greywacke. The tubing between the two ovens is
in abundance (Fig. 3a). Also common were coatings of a heated to minimise re-polymerisation of the silica. Although
anhedral flaky/mucaceous aluminosilicate phase composed the apparatus is capable of operating the ovens at different
of varying amounts of a Ca, Na, and K, most likely a clay temperatures, it must be isobaric throughout. As a result, the
mineral (smectite?) (Fig. 3b). Less common phases included pressure in both sample holders was maintained at 209 bars.
a euhedral Ca-Al-Si phase (Fig. 3c) resembling a zeolite The temperature of reaction in the second sample holder for
(clinoptilolite?) and overgrowths of potassium feldspar (Fig. this experiment was 157oC.
3d).
The effluent chemistry from this experiment showed an
3.3 K50-1 (Re-injection Aquifer – Re-injection Brine) increase of pH from 6.0 to 7.8 within three days. During the
same period, SiO2 concentration dropped to 400 ppm and
In this experiment, re-injection aquifer greywacke from a then rose to a steady value of 600 ppm. This represents a
New Zealand geothermal area was reacted at 154oC, 27 bars, Tquartz = 270oC, Tchalcedony = 267oC and Tamorph SiO2 = 147oC.
to simulate a re-injection scenario. The flow rate for this All other major components in the brine remain at close to
experiment was initially 1 ml/hr, then it was decreased to 0.5 their input values. In this case, it is assumed that all the
ml/hr near the end of the experiment. The initial SiO2 silica was monomeric when it entered the sample holder and
concentration in the brine was 980 ppm. Measurement of thus the fluid appears to have been close to equilibrium
2
New Zealand Geothermal Workshop 2011 Proceedings
21 - 23 November 2011
Auckland, New Zealand
Figure 2: Photomicrographs of: a) unreacted greywacke showing the matrix and adjacent quartz clast; b) reacted
greywacke with a coating of secondary minerals (smectite); c) euhedral Ca-Al-Si phase (zeolite?); and d) possible
zeolite phase on the matrix and adjacent quartz clast (right).

Figure 3: Photomicrographs of: a) chlorite rosettes; b) coating of smectite (?) adjacent to a quartz clast; c) Ca-Al-Si phase
(clinoptilolite?); d) secondary K-feldspar overgrowths.

3
New Zealand Geothermal Workshop 2011 Proceedings
21 - 23 November 2011
Auckland, New Zealand
Figure 4: Photomicrographs of: a) re-injection aquifer greywacke with coating of amorphous silica from experiment K50-1;
b) close-up of silica coating from K50-1; c) amorphous silica coating from experiment K50-2; and d) close-up of silica
coating from K50-2.

solubility with amorphous silica as it exited. the problems of the lack of thermodynamic and kinetic data
required for mathematical models. Although these
SEM examination of the run products from the bottom of the experiments may be site-specific, they can provide useful
sample holder shows a deposit of amorphous silica (Fig. 4c). insights into what fluid-mineral interaction processes are
No other secondary phases were identified. The amorphous occurring. Our apparratus at GNS can be used to study many
silica consists of a discontinous coating of small spherules questions relevant to the geothermal industry including:
10 – 20 nm in diameter (Fig. 4d). This contrasts with the
silica coating in the previous experiment (K50-1) where the 1) the process of fluid-mineral interactions at temperature
coating is continuous and there are silica spherules that are up to 400oC and pressures of 500 bars;
much larger and more diverse in size. Greywacke fragments 2) the thermodynamic stability of metal complexes and
from the top of the sample holder in K50-2 are clean and alteration minerals in geothermal systems and power
have no amorphous silica coating. Assuming that pressure stations;
has a minimal effect on silica precipitation, it is evident that 3) the temperature stability and chromatography of tracer
the depolymerisation of silica has had a marked effect on the chemicals (e.g. NDS tracers) in the geothermal
texture of the precipitate. environment;
4) the modification of rock properties due to fluid-
2. SUMMARY mineral interaction;
5) the formation of scaling and corrosion products in re-
Experimental hydrothermal apparati provide the opportunity injection aquifers and power station infrastructure.
to study fluid-mineral interactions in a wide variety of
natural/industrial situations. The experiments described Laboratory simulations are currently the closest we have to
above represent a small selection of possible simulations that simulating, studying and understanding the geothermal
can be conducted. Unlike geochemical modelling, real time environment.
experiments with real fluids and actual host rocks overcome

4
New Zealand Geothermal Workshop 2011 Proceedings
21 - 23 November 2011
Auckland, New Zealand