Embed Size (px)
Citation preview
Ig
HMÍ
a
ARAA
KIKGA
1
atcKeaatdpsvyfeo
bSLtp
0h
Geothermics 49 (2014) 66–75
Contents lists available at ScienceDirect
Geothermics
journa l homepage: www.e lsev ier .com/ locate /geothermics
DDP—The chemistry of the IDDP-01 well fluids in relation to theeochemistry of the Krafla geothermal system
alldór Ármannsson ∗, Thráinn Fridriksson, Gudmundur H. Gudfinnsson,agnús Ólafsson, Finnbogi Óskarsson, Dadi Thorbjörnsson
SOR, Iceland GeoSurvey, Grensásvegur 9, IS 108 Reykjavík, Iceland
r t i c l e i n f o
rticle history:eceived 29 May 2012ccepted 4 August 2013
a b s t r a c t
The Leirbotnar field, where IDDP-01 is situated consists of an upper liquid dominated zone to1000–1400 m depth, 190–220 ◦C, sulphate major anion, and a lower two phase zone, 300 ◦C chloride mainanion. The IDDP-01 fluid is dry steam, local origin, pH 3. The major anion is chloride (20–166 mg/kg), prob-
vailable online 12 September 2013
eywords:DDPraflaeochemistrycid well fluids
ably of magmatic origin. The major metallic cations, Fe (5–100 mg/kg), Cr (0–6 mg/kg), Ni (0–5 mg/kg) andMn (0.1–0.8 mg/kg) seem to be derived from the well casing and sampling equipment. The gas contentis low (about 0.1%) and the gas is apparently not directly emitted from magma.
© 2013 Elsevier Ltd. All rights reserved.
. Introduction
The Iceland Deep Drilling Project (IDDP) plans to drill 5 km inton active mid-ocean ridge hydrothermal system to investigate itsemperatures and pressures, its permeability structure and theomposition of its fluids and rocks (Fridleifsson et al., 2003). Therafla high temperature geothermal system is located in Northeast-rn Iceland (Fig. 1). The geology of the area is characterized by anctive central volcano containing a caldera and a magma chambert 5–8 km depth (Einarsson, 1978). Fig. 2 is a cross-section from easto west for the Leirhnúkur, Leirbotnar, Vesturhlídar and Sudurhlí-ar fields showing the relative positions of selected wells to theroposed magma chamber. The volcano is crosscut by an active fis-ure swarm that extends tens of km to the north and the south. Theolcanic activity at Krafla is episodic, occurring every 250–1000ears, each episode lasting 10–20 years. The last eruptive periodrom 1975 to 1984, resulted in 21 magmatic/tectonic events and 9ruptions (Björnsson, 1985). The magma chamber is the heat sourcef the geothermal system.
The area has been divided into several sub-fields: Leir-otnar (Lower Leirbotnar, Vítismóar), Sudurhlídar, Vesturhlídar,andabotnar, Hvíthólar, Vestursvædi, Leirhnúkur (Fig. 1). In the
eirbotnar field the system is divided into an upper zone downo 1000–1400 m depth, which is liquid dominated with a tem-erature 190–220 ◦C and a lower two phase zone at about 300 ◦C∗ Corresponding author. Tel.: +354 5281500.E-mail addresses: [email protected], [email protected] (H. Ármannsson).
375-6505/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.geothermics.2013.08.005
(on the boiling point curve) both with sulphate as the majoranion. Hvíthólar is a two-phase system following the boiling curvedown to about 1000 m depth where chloride is the main fluidanion but is cooler and liquid dominated below that where sul-phate is the main anion. One well has discharged in Sandabotnarsuggesting a two-phase fluid (boiling point curve) from a reser-voir at about 260 ◦C with chloride as the main fluid anion. Thetemperature in the Sudurhlídar and Vesturhlídar sub-fields fol-lows the boiling point curve and thus they deliver a two phasefluid at about 300 ◦C whose main anion is chloride. The liquidwater phase observed in all these sub-fields is dilute and close toneutral pH. The same characteristics as in Sudurhlídar and Ves-turhlídar were observed for the one well drilled in the Leirhnúkurarea. No well has discharged from the Vestursvædi area and theone well (KV-01) that was drilled there proved cool and unpro-ductive. Three separate upflow channels for geothermal fluidshave been identified, the major one associated with the Hveragilfissure. The recharge is essentially local in origin in Leirbotnar,Vesturhlídar and Leirhnúkur according to isotopic ratios (Darlingand Ármannsson, 1989), the Suðurhlídar and Hvíthólar sub-fieldsmay be recharged from far south and the fluid from the Sand-abotnar well is most likely derived from far south. Fig. 3 is aschematic view of the inferred general groundwater currents inthe area based on isotope (�D, �18O) data (Hjartarson et al., 2004).Magmatic gas (Ármannsson et al., 1989) probably affects the com-
position everywhere but it is more likely that excess gas directlyemanating from magma is only observed in the areas closestto the magmatic inflow where equilibrium has not been estab-lished.
H. Ármannsson et al. / Geothermics 49 (2014) 66–75 67
featur
2
sfltSwdBosop
Fig. 1. Map of the Krafla area showing tectonic
. Acid fluids
Two types of acid fluids are mostly observed in geothermalteam, acid sulphate fluids and acid chloride fluids. Acid sulphateuids are more common in relatively cool liquid-dominated sys-ems (Moya et al., 2005; Villa et al., 2000; Moore et al., 2002;ugiaman et al., 2004; Reyes, 1990, 1991; Bowyer et al., 2008)hereas acid chloride fluids are more prevalent in hot vapourominated systems (Truesdell et al., 1989; D‘Amore et al., 1990;ell, 1989; Hirtz et al., 1991; Izquierdo et al., 2000). The first obvi-
us manifestations of acidity in Krafla well fluids were found in atream formed after a blowout of well K-4 in 1976 in which a pHf 1.86 was observed (Gíslason and Arnórsson, 1976). The com-osition of the stream water suggested sulphuric acid but lateres, the different wellfields and individual wells.
computations with comparison with later nearby acid wells K-10and K-25 and nearby shallow well K-2 suggested that the acid-ity was due to hydrochloric acid but the sulphate originated fromshallower inflows (Ármannsson and Gíslason, 1992). The acidityhas been associated with magmatic activity in the Krafla volcanicsystem 1975–1984 which manifested itself in excess gas in thegeothermal system and deposition of iron sulphides, iron silicates,iron oxides and silica (Ármannsson et al., 1982, 1989, 1987).
Superheated steam from geothermal wells has been reported inat least four geothermal areas apart from Krafla, i.e. The Geysers,
California, Larderello, Italy, Matsukawa, Japan and Tatun, Taiwan(Truesdell and White, 1973; Truesdell et al., 1989). The first drysteam well in Krafla was well K-12, drilled in 1978. Its initial dis-charge was described by Hauksson (1979). The steam contained a
68 H. Ármannsson et al. / Geothermics 49 (2014) 66–75
logic mA
srptitmHf
A
Fig. 2. East-west cross-section showing a geodapted from Hauksson and Guðmundsson (2008).
ubstantial amount of molecular HCl which turned into hydrochlo-ic acid upon condensation causing severe corrosion during whicharticles were formed which in turn caused the erosion of theurbine blades in the power plant. The problem was solved bynsulation of the wellhead and by mixing the flow with that fromhe liquid dominated well K-9. Truesdell et al. (1989) proposed a
echanism involving the existence of a deep brine from which theCl rich steam was boiling at depth. Evidence for this brine was
ound in the flow of a saline liquid from nearby well K-7 whose
Fig. 3. Proposed flow of groundwater from the highlanddapted from Hjartarson et al. (2004).
odel inferred for the Krafla geothermal field.
liquid fraction concentration of Cl increased from 78.7 mg/kg to5295 mg/kg from March 19 to March 29, 1977 without a significantincrease in the magnitude of the liquid fraction. This was inter-preted as being from a deep inflow into the well which at most othertimes was blocked (Truesdell et al., 1989). The fluid from well K-26was specifically reported on by Fridleifsson et al. (2006) at which
time there were plans to drill the IDDP well close by. There fluidsturned out to be acid and the chloride concentration relatively highfor Krafla fluids but gas concentrations were moderate. The sames into the Námafjall and Krafla geothermal areas.
H. Ármannsson et al. / Geothermics 49 (2014) 66–75 69
canicA
ps2ccrwdotfe
dfadt
Fig. 4. Map showing the inferred depth of voldapted from Mortensen et al. (2009).
attern has been followed in subsequent wells where fluids havehown acid character (e.g. K-27, K-29, K-33). Well K-36 drilled in007 turned out to be very powerful with a relatively low total gasoncentration especially the CO2 concentration whereas the H2Soncentration was relatively high and consequently the CO2/H2Satio was low similar to that observed for the IDDP-01 fluid, but theell was soon damaged due to acid fluids and deposits. This wellraws fluids from near to the present IDDP well and a re-evaluationf the location of the drill site was carried out with reference tohese results. It was concluded that this was a suitable drill siterom which a lot could be learnt about fluid handling (Fridleifssont al., 2008).
Additional information has been obtained from well K-35,rilled in 2007 with acid fluids at depth close to Leirhnúkur, also
rom well K-38 which extends far to the north of Víti and containscid fluids, and well K-39 drilled in 2008 with acid fluids at 2.5 kmepth in the Sudurhlídar area i.e. a considerably shallower depthhan previously envisaged. In Fig. 4, adapted from Mortensen et al.gas inflow into the Krafla geothermal system.
(2009), the depth to inflow of magmatic gas is shown and thus anindication of the depth of acid aquifers (in m b.s.l.) in Krafla. Thisfigure clearly illustrates that acid fluids may be expected anywherein the Krafla system but at the most shallow levels in the Leirbotnararea.
Thus recent information available after the drilling of well IDDP-01 is that magma may be encountered at a shallower depth thanpreviously considered possible and that the acid fluids are probablyspread over the whole area at >2000 m depth but are not confinedto a specific part.
The Fluid Handling Group of the IDDP has been studying possi-ble solutions to the acid fluid problem. It has been speculated thatthe acid fluids only occur over a limited depth interval in whichcase they may be cased off with suitable material. If they persist
at all depths, means should be available to transport them to thewellhead without condensation, i.e. without forming an acid, andprotecting the wellhead by the use of resistant materials and fur-thermore by a shallow injection of NaOH or just geothermal water
70 H. Ármannsson et al. / Geothermics 49 (2014) 66–75
Table 1Approximate wellhead temperature, pressure and discharge 2010–2012 (see alsoIngason et al., 2013).
Discharge period P0 (bar g) T0 (◦C) Discharge(kg/s)
From To
22.03.2010 25.03.2010 120 320 625.03.2010 10.04.2010 120 290 3511.05.2010 24.08.2010 20 200–370 25–4517.05.2011 25.05.2011 60 400 38
fgioT(
3
btigsfldiKtiidflatEfaI
4
a(isap2t
tt
CU(tdt
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Jan2010
Apr2010
Jul2010
Oct2010
Jan2011
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
pH
Water Condensate
09.08.2011 11.08.2011 70 360 4528.09.2011 16.11.2011 140 330–420 716.11.2011 25.07.2012 140 440 14
rom plant separators. Pilot tests with such shallow injection haveiven promising results (Hauksson et al., 2014). If all else fails deepnjection of NaOH could be effected taking into account experiencef systems such as those described for Larderello, Italy (Bell, 1989),he Geysers, California (Hirtz et al., 1991) and Tiwi, the PhilippinesGardner et al., 2001)
. Well IDDP-01
Well IDDP-01 (Fridleifsson et al., 2013) is in the Northern Leir-otnar field in the Krafla area (Fig. 1). Its original design downo 4.5 km depth is described by Thórhallsson et al. (2013). Dur-ng drilling a conventional logging programme was followed, i.e.eophysical logs (temperature, pressure) were carried out at eachtage of casing (see Ingason et al., 2010) and the pH of the drillinguids was monitored. As magma was encountered at 2104 mrilling was discontinued and the programme changed accord-
ngly (Thórhallsson et al., 2014). Magma intruded into a well inrafla during the Krafla fires (Gudmundsson, 1979), in an eruption
hrough a well in Námafjall (Larsen et al., 1978) and in well K-39n Krafla in 2008. The only other record of drilling into magmas in Big Island, Hawaii (Wang, 2008). The present well is casedown to 1949 m with a slotted liner below. Upon completion aow test programme including enthalpy and flow measurementsnd chemical sampling was initiated. The main features of the flowest are described below but in more detail by Ingason et al. (2014).nthalpy and flow measurements have been carried out using dif-erential pressure across orifices (James, 1966). Times of dischargend main flow are listed in Table 1 but dealt with in more detail byngason et al. (2014).
. Chemical composition
Sampling and analysis. A special high pressure steam-liquid sep-rator was designed for this project to obtain wellhead samplesIngason et al., 2010). This separator was used to collect the samplesn 2010. As condensate only has been observed in 2011 and 2012amples have been drawn directly through valves, generally fromsampling valve located near the wellhead, after the first orificelate at a pressure of approximately 17 bar-g except on November4, 2011 when samples were obtained from a sampling valve beforehe first orifice plate at a full well-head pressure.
Water and condensate liquids were partitioned into raw, fil-ered, untreated, acidified, precipitated and extracted according tohe constituent to be determined.
Na, K, Ca and Mg, were in most cases determined by AAS,O2 and H2S in liquid water samples by titration, SiO2 and B, byV/vis spectrophotometry, Cl, Br, F and SO4 by Ion Chromatography
IC), pH and conductivity by electrometry and TDS by gravime-ry, in the ÍSOR laboratory. Na, K, Ca, Mg, Fe, Mn, Cr and Ni wereetermined by Inductively Coupled Plasma/Atomic Emission Spec-rometry (ICP/AES), Ba, Cd, Mo and Pb by Inductively Coupled
Fig. 5. Variation in pH of liquid and condensate samples from the IDDP-1 well during2010–2012.
Plasma/Mass Spectrometry (ICP/MS), higher concentrations of Sr,Cu Zn, Al, Sr and As by ICP/AES but lower concentrations by ICP/MS,and Hg by Atomic Fluorescence Spectrometry (AFS) by the ALS lab-oratories, Luleå, Sweden.
Steam samples were collected into a NaOH solution for thedetermination of acid gases by titration and the head space gas wasdetermined by gas chromatography (Ármannsson and Ólafsson,2006).
Precision and limits for ICP are found in ALS Laboratories (2012)For titration, AAS, IC, electrometry, fluorimetry and spectropho-tometry (for B and SiO2) precision and detection limits are foundin Pang and Ármannsson (2006). Isotope ratios were determinedby MS at the British Geological Survey, Wallingford, UK and at theUniversity of Iceland Institute for Earth Sciences.
During the first day, March 22–23, 2010, the flow was through a�10 mm orifice but after March 25, 2010 through �50 and �22 mmorifices and a DN50 pipeline. Equipment for the determination ofenthalpy and flow was not installed at first but Hauksson (pers.comm.) made two estimates from differential pressures over ori-fices and liquid outflow from the gravel silencer. During the initialflow he obtained 6 l/s total, with about 1 l/s liquid flow, but afterthe maximum opening March 25 he obtained 33.5 l/s with about0.4 l/s liquid flow.
From March to August 2010 a total of eighteen samples of water,condensate and steam were collected and analyzed. Thirteen waterand condensate samples, one water (only) sample and two con-densate (only) samples were analyzed for stable isotopes (�D and�18O).
Well IDDP-01 only discharged dry, super-heated steam in 2011when a total of 17 samples were collected. Eleven of these weresubjected to complete analysis of steam and condensate, two areduplicate samples for which different sampling methods or sampletreatments were tested, and in three pH and total dissolved solidsonly were determined. In 2012 two samples for total analysis werecollected.
Results. The composition of a sample from March 2012 whenthe well had discharged superheated steam only, relatively undis-turbed since September 2011, is shown in Table 2 along with themain features of the chemical composition of the dry steam fromtwo earlier superheated wells, KG-12 and K-36. Changes with timefor the most important components in liquid and condensate sam-
ples are shown in Figs. 5–9, 11 and 15.pH. The pH of the condensed vapour ranges from 2.4 to 3.6 butmost samples show between 3.0 and 3.5. There seems to be a sys-tematic trend towards lower condensed vapour pH values with
H. Ármannsson et al. / Geothermics 49 (2014) 66–75 71
Table 2Results of analysis of sample from well IDDP-01, January 2012 with selected dataon samples from wells KG-12 and KJ-36.
Parameter IDDP-08.03.2012 KG-12 1979 KJ-36 2007
T (◦C) 440 176.5P0 (bar g) 138 7.9 9.0H0 (kJ/kg) 3090 2776 2676Condensate
pH (◦C) 2.44/22.7 3.48/19 3.30/21Conductivity (�S/cm/◦C) 977/25 290/25B (mg/kg) 2.2 0.46SiO2 (mg/kg) 6.2 28TDS (mg/kg) 70 81.2�D (‰ SMOW) −85.1�18O (‰ SMOW) −9.77Na (mg/kg) 0.08 0.17K (mg/kg) 0.02Mg (mg/kg) 0.004 0.04Ca (mg/kg) <0.1 0.43F (mg/kg) 8.45 0.24 8Cl (mg/kg) 89.6 112 400SO4 (mg/kg) 5.78 6.5Al (mg/kg) 0.0648As (mg/kg) <0.003Ba (mg/kg) <0.0005Cd (mg/kg) <0.0001Co (mg/kg) 0.0522Cr (mg/kg) 4.90Cu (mg/kg) <0.005Fe (mg/kg) 21.5 26Hg 0.000031Li (mg/kg) 0.612Mn (mg/kg) 0.531Mo (mg/kg) <0.003Ni (mg/kg) 3.18Pb (mg/kg) <0.0005Sr (mg/kg) <0.002Zn (mg/kg) 0.0105
GasCO2 (mg/kg) 560 17,077 6463H2S (mg/kg) 250 1127 3320H2 (mg/kg) 8.77 44.6 32.8N2 (mg/kg) 16.3 0 175
dnOcch
F
0
5
10
15
20
Jan2010
Apr2010
Jul2010
Oct2010
Jan2011
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
F (
mg/
kg)
Water Condensate
Fig. 7. Variation of F in IDDP-1 liquid and condensate samples during 2010–2012.
0
100
200
300
400
500
Jan Apr Jul Oct Jan Apr Jul Oct Jan Apr
SO
4 (m
g/kg
)
Water Condensate
0
5
10
15
20
25
30
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
CH4 (mg/kg) 0.27 6.7 2.7Ar (mg/kg) 0.53 3.6
ischarge time (Fig. 5). Tests have shown that there is not a sig-ificant difference between on site and laboratory determinations.n the other hand there is a significant difference in that samples
ondensed in polypropylene bottles show a lower pH than thoseondensed during flow through a stainless steel cooling coil, and aigher pH is observed if the sample is collected from a small than0
500
1000
1500
2000
Jan2010
Apr2010
Jul2010
Oct2010
Jan2011
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
Cl (
mg/
kg)
Water Condensate
0
50
100
150
200
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
ig. 6. Variation of Cl in IDDP-1 liquid and condensate samples during 2010–2012.
2010 2010 2010 2010 2011 2011 2011 2011 2012 2012
Fig. 8. Variation of SO4 in IDDP-1 liquid and condensate samples during 2010–2012.
from a large flow through the cooling coil. The condensation inthe polypropylene bottles may have been incomplete resulting inan elevated concentration of Cl− and F− ions causing a lowered
pH, whereas metal ions may have been dissolved from the coolingcoil causing a raised pH when flow rate is limited. The optimumdeterminations are obtained from a fast flow through the coolingcoil.0
200
400
600
800
1000
Jan2010
Apr2010
Jul2010
Oct2010
Jan2011
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
SiO
2 (m
g/kg
)
Water Condensate
0
5
10
15
20
25
30
35
40
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
Fig. 9. Variation of SiO2 in IDDP-2 liquid and condensate samples during 2010–2012.
7 eothermics 49 (2014) 66–75
fo
spfmfanfHpbttacAaeccmaqimh
blc(t
afbt
mosddSetanUbed2ire
v22tl
0
5
10
15
20
25
30
35
40
250 300 350 400 450
SiO
2 (m
g/kg
)
Well head temperature (°C)
30 bar
100 bar
140 bar
Condensate
2 H. Ármannsson et al. / G
The pH is apparently controlled by the generation of H+ ionsrom HCl and HF during condensation but modified by dissolutionf iron and other metals from pipes.
Cl. Cl concentrations range from 20 to 166 mg/kg in condensedamples of superheated steam, but from 115 to 1525 mg/kg in sam-les of small liquid fractions. Cl− is the major anion and judgingrom the experience from nearby wells it is likely derived from
agmatic gas which comes into contact with liquid water andorms hydrochloric acid. Generally these concentrations are rel-tively low compared to those of acid fluids from wells in theeighbourhood (see Table 2). The possibility of HCl being derived
rom either a boiling brine or a brine that has boiled dry leavingCl in a vapour state but hydroxides reacting to produce alterationroducts (Bischoff et al., 1996; Fournier and Thompson, 1993) haseen considered. As seen below the isotope composition suggestshat the recharge is present day and local. The local groundwa-er is dilute and the amount of brine postulated would have takenlong time to accumulate and somewhat lower values reflecting
ooler regimes would have been expected for the deuterium ratio.lthough most of the hydroxides formed would have taken part inlteration reactions it seems unlikely that hardly any alkali metalsscape with the vapour giving rise to the very low alkali metaloncentrations observed. It also seems that the significant con-entration of fluoride observed is more likely to be derived fromagma than brine formed from the dilute local groundwater. It is
lso worth considering that the well is cased to a depth which isuite close to the depth at which magma was encountered so that
t seems more likely that it is drawing fluid directly from the mag-atic environment rather than a possible vapour or brine pool that
appens to be present at this depth and location.F. F concentrations in the condensed vapour samples are
etween 2.4 and 8.8 mg/kg, but from 5.94 to 17.9 in samples ofiquid fractions. Its origin is probably magmatic similar to that ofhloride. Cl− concentrations do not show a clear trend with timeFig. 6) but F− concentrations appear to increase with dischargeime (Fig. 7) probably due to increasing temperature.
SO4. SO42− was only determined in the liquid fraction in 2010
nd varied then from 61.5 to 450 mg/kg. In the condensate samplesrom 2011 and 2012 it varied from 1.5 to 25 mg/kg and appears toe rather decreasing with time. This behaviour may be connectedo the changes in well flow in November 2011 (Fig. 8).
As discussed below H2S is present as a gas but attempts at deter-ining other sulphur species by chromatography did not reveal any
ther species (S. Arnórsson, pers. comm.). FXRD analysis and SEMtudies have demonstrated that native sulfur accumulates on filtersuring sampling (Mortensen et al., 2011), at about 100 mg/kg con-ensate (S. Markússon, pers. comm.). Reaction between H2S andO2 in the condensate during sampling has been postulated butarly attempts at determining SO2 in the steam using UV spec-rometry were unsuccessful (E. Ilyinskaya, pers. comm.). Furtherttempts at analysing for SO2 by oxidation to sulphate and determi-ation by ion chromatography (A. Stefánsson, pers. comm.) and byV spectrometry (Burton et al., 2012) showed SO2 concentrationselow detection limits but using a MultiGas sensor system with anlectrochemical sensor for SO2 concentrations up to 0.1 ppm wereetected in the plume from the silencer rock mound (Burton et al.,012). Processes taking place in the surface installations, includ-
ng the rock muffler, might remove SO2 from the steam and thusesult in the relative absence of this gas from the steam plume, orlemental sulfur might be transported with the steam.
SiO2. SiO2 was only determined in the liquid fraction in 2010 andaried then from 244 to 818 mg/kg. In the condensate samples from
011 it varied from 1.5 to 55 mg/kg, and the sample from March012 has a silica concentration of 6.2 mg/kg. Such silica concen-rations are unusual for geothermal steam condensates, even theowest values. It can also be seen from Figs. 9 and 10 that the SiO2Fig. 10. SiO2 concentrations vs wellhead temperature in the IDDP-1 produced fluids.Curves represent silica concentrations calculated from the quartz solubility data ofFournier and Potter (1982).
concentration in condensate from IDDP-01 essentially increaseswith discharge time and increasing well head temperature. Sil-ica was determined by two different methods, spectrophotometrywith ammonium molybdate and ICP/AES which do not yield signif-icantly different results. There is, however a small but significantdifference between concentrations of filtered (0.2 �m) and unfil-tered samples. Still most of the silica carried with the steam isin aqueous form as the condensate passes through the filter. Sil-ica is known to be transported in a gaseous state, especially athigh temperatures and high pressures. Thermodynamic calcula-tions using the composition of fumarole gas from Kudryavy volcanoindicate that silicon tetrachloride (SiCl4) and silicon tetrafluoride(SiF4) could be important transport species and the latter has beendetected in volcanic fumarole gas (Mori et al., 2002). Modellingstudies suggest that either of these species could be dominantalthough HF may be the major F bearing species thus favouringSiCl4 (Churakov et al., 2000; Churakov, 2001; De Hoog et al., 2005).The F and Cl concentrations in the Kudryavy fumarole gas are muchhigher than in the samples from well IDDP-01 and SiCl4 and SiF4are unlikely to form in high enough concentrations to explain thetransport of silica in the IDDP-01 steam, considering that HCl andHF are likely to be the dominant Cl- and F-bearing species. It hasbeen shown that in industrial boilers silica enters the steam as avolatile, with the concentration in the steam strongly dependenton pressure (the density of the steam) and the silica concentra-tion of the boiler water. At a pressure of 140 bar and with 500 ppmsilica in the boiler water, the predicted silica content of the coex-isting steam is about 20 ppm but at about 100 ◦C overheating at thesame pressure, the silica content increases to about 60 ppm (GasProcessors and Suppliers Association (GPSA), 2004; Bahadori andVuthaluru, 2010). Provided that a boiler is a good analogue systemfor well IDDP-01, this suggests that volatility of silica can explainthe concentration of silica found in the condensate and that it willdecrease sharply with lower pressure.
Cations. The most prominent cations in condensate samples areH+ and those of the metals Fe, Cr, Ni, and Mn. The oxidation statesof metal species were not determined. There are variations in con-centrations both in liquid and vapour phases due to opening andclosing the well during the early discharge periods (see Fe, Fig. 11).Data on concentration–time trends for these metals can only be
considered reliable after September 28, 2011 when continuous dis-charge started (Table 1). None of these metals show a systematictrend with time after that although increasing temperature mighthave caused increased dissolution.
H. Ármannsson et al. / Geothermics 49 (2014) 66–75 73
0
100
200
300
400
500
600
700
800
Jan2010
Apr2010
Jul2010
Oct2010
Jan2011
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
Fe
(mg/
kg)
Water Condensate
0
20
40
60
80
100
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
Fd
acFwataa
atodcotastit
Fpss
0
1
2
3
4
5
6
0 1 2 3 4 5 6
Cr
(mg/
kg)
Ni (mg/kg)
Condensate max Cr S−316 max Cr K−55
ig. 11. Variation in Fe concentrations in IDDP-1 liquid and condensate samplesuring 2010–2012.
The Fe concentrations in the liquid phase may seem high butgain the small amount of liquid in which it is dissolved has to beonsidered. Assuming maximum liquid flow of 1 l/s and a maximume concentration of 100 mg/l in the liquid then about 78 kg of Feould have been discharged during liquid flow. The relatively small
mount of Fe discharged and the observed decrease in pH suggesthat the liquid was far from being saturated with dissolved Fe assolution saturated with Fe would have exchanged more H+ ions
nd its pH would have increased.The concentrations of Fe and Mn in the condensate samples
ppear to be strongly correlated and Cr and Ni concentrations seemo have an even stronger correlation. However, the concentrationsf Cr and Ni are not correlated to those of Fe (or Mn) in the con-ensate samples. A near linear correlation between Mn and Feoncentrations (Fig. 12) suggests that these metals have a commonrigin in either the K-55 steel of the liner or the S-316 steel used inhe cooling coil. There is an excellent linear correlation between Crnd Ni in the condensate samples. This linear correlation stronglyuggests a common source of these metals that is consistent with
he Cr/Ni ratio of the S-316 steel (Fig. 13). Corrosion in the well dur-ng discharge and corrosion of the sampling equipment both appearo affect the condensate composition. Samples collected without0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 10 20 30 40 50 60 70 80 90
Mn
(mg/
kg)
Fe (mg/kg)
Condensate max Mn S−316 max Mn K−55
ig. 12. Mn concentration vs Fe concentration in condensate samples in IDDP-1roduced fluids. Lines represent the Mn/Fe ratio in stainless steel (S-316) and mildteel (K-55) assuming the maximum Mn concentrations reported for the respectiveteel types.
Fig. 13. Cr concentration vs Ni concentration in condensate samples. Lines repre-sent the Cr/Ni ratio in stainless steel (S-316) and mild steel (K-55) assuming themaximum concentrations reported for the respective steel types.
the use of a stainless steel cooling coil with the steam condenseddirectly in a PP bottle revealed moderate to high Fe and Mn con-centrations, but the lowest Cr and Ni concentrations found. Thusa substantial part of the dissolved Fe and Mn in the condensate isderived from corrosion of the mild K-55 steel in the well, whereasCr and Ni appear to be derived mostly from the stainless S-316 steelcooling coil.
The concentrations of the alkali and alkaline earth metals in con-densate samples (see Table 2) are for the most part very low. Naconcentrations in most samples are between 0.05 and 1.5 mg/kg,K and Ca concentrations between 0.1 and 1 mg/kg and those ofMg generally between 0.02 and 0.2 mg/kg. In general, these com-ponents do not appear to be important players in the vapourchemistry.
Stable isotopes. �D and �18O for H2O were determined in thir-teen water and condensed steam samples, one water (only) sampleand two condensate (only) samples in 2010, eight of the condensedsteam samples from 2011, and two from 2012. The ratios in thecondensed steam samples have remained very constant. �D val-
ues are between −80‰ and −90‰, typically between −84.9‰ and−85.8‰ and �18O values are with one exception between −9.80‰and −10.49‰ relative to SMOW (Fig. 14). With reference to previ-ous work on stable isotopes this fluid falls in the group of Leirbotnar−100
−90
−80
−70
−60
−50
−13 −12 −11 −10 −9 −8 −7
δD (
%o
SM
OW
)
δ18O (%o SMOW)
Water Condensate Meteroric water line
Fig. 14. �D vs �18O in liquid water and condensate samples. Line represents theglobal average for meteoric water (Craig, 1961).
74 H. Ármannsson et al. / Geothe
0
500
1000
1500
2000
2500
3000
Jan2010
Apr2010
Jul2010
Oct2010
Jan2011
Apr2011
Jul2011
Oct2011
Jan2012
Apr2012
Con
cent
ratio
n (m
g/kg
)
CO2H2S
F2
wao
actio1opoptegcwmtsu1
ahbsutwuepteI
s
5
d
ig. 15. Variation in CO2 and H2S concentrations in steam samples during010–2012.
ater which has its origin in local precipitation according to Darlingnd Ármannsson (1989) but a considerable oxygen isotope shift isbserved as would be expected at such high temperatures (Fig. 14).
Non-condensable gases. The gas contents of the fluid were gener-lly relatively low (Fig. 15) and the CO2/H2S ratio very low. The CO2oncentration in steam seems to have stabilized at 6–900 mg/kg,he CO2/H2S ratio at 1.1–1.7, and H2 at 10–20 mg/kg. This would benterpreted as no direct excessive intrusion of magmatic gas such asccurred during the Krafla fires 1975–1984 (see Ármannsson et al.,989). Using gas geothermometers from Arnórsson et al. (1998), thebserved CO2 concentrations would correspond to a reservoir tem-erature of about 200 ◦C. Gas geothermometer temperatures basedn H2S and H2 show about 250 and 280 ◦C, respectively. These tem-eratures are much lower than even the well-head temperatures inhe IDDP-01 well. These geothermometers are based on presumedquilibria between deep liquid and secondary minerals found ineothermal systems between 230 and 350 ◦C, so that such low gasoncentrations are surprising, particularly the CO2, concentration,hich was very high in steam from wells in Krafla affected by mag-atic gas input (Ármannsson et al., 1989). It must be concluded
hat the silicic magma encountered by the IDDP-01 well does noteem to be emitting CO2 to its surroundings in contrast to the sit-ation with the basaltic magma involved in the Krafla fires in the970s and 80s.
Summary. Thus the composition of the superheated steamppears manageable. The chloride concentration was generallyigher in both well K-12 (112 mg/kg) and K-36 (400 mg/kg) fluidut the pH is probably a little lower in the condensed IDDP-01team. Cl and F are apparently transported to the surface as molec-lar HCl and HF but form acid upon condensation at the surface andhen attack the steel in the surface equipment where it can be dealtith adequately. Thus it seems that this fluid should be suitable forse in power production if due precautions are taken (Haukssont al., 2014). The choice of material for surface equipment is stillroblematic as titanium which is resistant to this type of composi-ion cannot withstand temperatures much higher than 340 ◦C, andven the highest quality stainless steels are prone to cracking in theDDP steam (Á. Einarsson, pers. comm.).
The overall conclusion is that the chemistry of the superheatedteam appears manageable.
. Conclusions and recommendations
As condensed steam from the well can be expected to be acidicue care should be taken to prevent corrosion. In the case of
rmics 49 (2014) 66–75
the present well suitable material has not yet been found butprecautions could involve the use of such material and/or shallowor deep injection of an alkali such as sodium hydroxide. Varioustests to this effect are described by Hauksson et al. (2013). The wellis not likely to produce from supercritical conditions as originallyintended but yields useful information on the composition of hotfluids in proximity to magma.
The next IDDP well is due to be drilled at Reykjanes, where prox-imity to magma is not likely to present problems. However, thesalinity of fluids there may present interesting complications butshould also yield additional information on the chemistry of veryhot geothermal systems.
Acknowledgements
Ms. B. Jónsdóttir, Ms. G.S. Jónsdóttir and Ms. A.M. Mortensenare thanked for assisting with figures, and Dr. R.F. Fournier andProfessor S. Arnórsson for reviews that have much improved thepaper.
References
ALS Laboratories, 2012. http://www.analytica.se/hem2005/eng/miljo/vattennaturliga.asp (accessed 19.12.12).
Ármannsson, H., Gíslason, G., 1992. The occurrence of acidic fluids in the Leirbotnarfield, Krafla Iceland. In: Kharaka, Y.K., Maest, A.S. (Eds.), Water-rock Interaction.Balkema, Rotterdam, pp. 1257–1260.
Ármannsson, H., Ólafsson, M., 2006. Collection of geothermal fluids for chemicalanalysis, Report No. ÍSOR-2006/016 Iceland GeoSurvey, 17pp.
Ármannsson, H., Benjamínsson, J., Jeffrey, A.W.A., 1989. Gas changes in the Kraflageothermal system, Iceland. Chemical Geology 76, 175–196.
Ármannsson, H., Gíslason, G., Hauksson, T., 1982. Magmatic gases in well fluidsaid the mapping of the flow pattern in a geothermal system. Geochimica etCosmochimica Acta 46, 167–177.
Ármannsson, H., Gudmundsson, Á., Steingrímsson, B.S., 1987. Exploration and devel-opment of the Krafla geothermal area. Jökull 37, 13–30.
Arnórsson, S., Fridriksson, Th., Gunnarsson, I., 1998. Gas chemistry of the Kraflageothermal field, Iceland. In: Arehart, G.B., Hulston, J.R. (Eds.), Water Rock Inter-action. Balkema, Rotterdam, Brookfleld, pp. 613–616.
Bahadori, A., Vuthaluru, H.B., 2010. Prediction of silica carry-over and solubility inboilers using simple correlation. Applied Thermal Engineering 30, 250–253.
Bell, D., 1989. Description of an operational desuperheating and chloride scrub sys-tem. Geothermal Resources Council – Transactions 13, 303–307.
Bischoff, J.L., Rosenbauer, R.J., Fournier, R.O., 1996. The generation of HCL in thesystem CaCl2–H2O: vapor–liquid relations from 380–500 ◦C. Geochimica et Cos-mochimica Acta 60, 7–16.
Björnsson, A., 1985. Dynamics of crustal rifting in NE Iceland. Journal of GeophysicalResearch 90, 10151–10162.
Bowyer, D., Bignall, G., Hunt, T., 2008. Formation and neutralization of cor-rosive fluids in the shallow injection aquifer, Rotokawa geothermalfield, New Zealand. Tauhara North No Trust, GNS, Mighty River Power,www.geothermal.org/Powerpoint08/Tuesday/GeochemistryI/BowyerTue08.ppt
Burton, M., Ilinskaya, E., Yeo, R., Markússon, S., 2012. Measurements of geothermalgas emitted from IDDP well, Krafla, Iceland, 19 June 2012. Landsvirkjun, Iceland(Unpublished report), 10 pp.
Churakov, S., 2001. Physical-chemical properties of complex natural fluids. Technis-chen Universität, Berlin, pp. 144 pp (Dr. rer. nat. Dissertation).
Churakov, S.V., Tkachenko, S.I., Korzhinskii, M.A., Bocharnikov, R.E., Shmulovich,K.I., 2000. Evolution of composition of high-temperature fumarolic gases fromKudryavy Volcano, Itutup, Kuril Islands: the thermodynamic modeling. Geo-chemistry International 38, 436–451.
Craig, H., 1961. Isotopic variations in meteroric waters. Science 133, 1702–1703.D‘Amore, F., Truesdell, A.H., Haizlip, J.R., 1990. Production of HCl by mineral reaction
in high temperature geothermal systems. In: Proceedings of the 15th Work-shop on Geothermal Reservoir Engineering, Stanford University, Stanford, CA,pp. 23–25.
Darling, W.G., Ármannsson, H., 1989. Stable isotopic aspects of fluid flow in theKrafla, Námafjall and Theistareykir geothermal systems of Northeast Iceland.Chemical Geology 76, 175–196.
De Hoog, J.C.M., van Bergen, M.J., Jacobs, M.H.G., 2005. Vapour-phase crystallisationof silica from SiF4-bearing volcanic gases. Annals of Geophysics 48, 775–785.
Einarsson, P., 1978. S-wave shadows in the Krafla caldera in NE-Iceland, evidencefor a magma chamber in the crust. Bulletin Volcanologique 41, 1–9.
Fournier, R.O., Potter, R.W., 1982. An equation correlating the solubility of quartz
in water from 25 ◦C to 900 ◦C at pressures up to 10,000 bars. Geochimica etCosmochimica Acta 6, 1969–1973.Fournier, R.O., Thompson, J.M., 1993. Composition of steam in the systemNaCl–KCl–H2O-quartz at 600 ◦C. Geochimica et Cosmochimica Acta 57,4365–4375.
eothe
F
F
F
F
G
G
G
G
H
H
H
H
I
I
I
J
feasibility of acid well utilization: The Philippines’ well MG-9D experience. In:Proceedings of the 20th Annual PNOC EDC Geothermal Conference, Philippines,
H. Ármannsson et al. / G
ridleifsson, G.Ó., Ármannsson, H., Árnason, K., Bjarnason, I., Gíslason, G.Th., 2003.Part I. Geoscience sand Site Selection. In: Fridleifsson, G.Ó. (Ed.), Icelandic DeepDrilling Project, Feasibility Report, Report No. OS-2003/007. Orkustofnun, Reyk-javík, p. 104 pp.
ridleifsson, G.Ó., Ármannsson, H., Árnason, K., Gudmundsson, Á., Sæmundsson, K.,2008. Re-evaluation of the first IDDP drillsite., pp. 13 pp (IDDP report).
ridleifsson, G.Ó., Ármannsson, H., Mortensen, A.K., 2006. Geothermal conditionsin the Krafla caldera with focus on well KG-26. In: A review in relation to theIcelandic Deep Drilling Project. Report No. ÍSOR 2006/030, Iceland Geosurvey,37 pp.
ridleifsson, G.Ó., Ármannsson, H., Gudmundsson, Á.G., Árnason, K., Mortensen, A.K.,Einarsson, G., 2013. Site selection for IDDP-1 at Krafla. Geothermics (this vol-ume).
ardner, A.B., Rivera, N.N., Santos, R.M., Toreja, J.S., Villasenor, L.B., Gallup, D.L., 2001.Proceedings of the 22nd Annual PNOC EDC Geothermal Conference , ManilaPhilippines.
as Processors and Suppliers Association (GPSA), 2004. Engineering Data Book, 12thed. Gas Processors and Suppliers Association, Tulsa, OK, USA, pp. 155–159.
íslason, G., Arnórsson, S., 1976. A progress report on flow and chemical compositionof fluids from boreholes 3 and 4 in Krafla (in Icelandic). Report No. OS-JHD-7640,Orkustofnun, Reykjavík.
udmundsson, Á., 1979. Well letter (in Icelandic) No. 12, Orkustofnun, Reykjavík.,pp. 14 pp.
auksson, T., 1979. Well KG-12 (in Icelandic). Well Letter (in Icelandic) No. 11,Orkustofnun, Reykjavík, 5 pp.
auksson, T., Markússon, S., Einarsson, K., Karlsdóttir, S.N., Einarsson, Á., Möller,A., Sigmarsson, Th., 2014. Pilot testing of handling the fluids from theIDDP-a1 exploratory geothermal well, Krafla, N.E. Iceland. Geothermics 49,76–82.
irtz, P., Buck, C., Kunzman, R., 1991. Current techniques in acid-chloride corrosioncontrol and monitoring at the Geysers. In: Proceedings of the 16th Work-shop on Geothermal Reservoir Engineering, Stanford, CA. Stanford University,pp. 83–95.
jartarson, A., Sigurdsson, Ó., Gudmundsson, Á., Ármannsson, H., Karlsdóttir, R.,2004. A computer simulation of the Námafjall geothermal system and forecastabout its future state upon production of 90 MW electricity from Bjarnarflag (inIcelandic). In: Report No. ÍSOR-2004/009, ÍSOR, Reykjavík, p. 119 pp.
ngason, K., Albertsson, A., Pálsson, B., Ballzus, C., Ármannsson, H., Matthíasson,J., Thórhallsson, S., Gunnarsson, T., 2010. Iceland Deep Drilling Project (IDDP)– Fluid handling and evaluation. In: Proceedings of the World GeothermalCongress 2010, p. 5 pp.
ngason, K., Kristjánsson, V., Einarsson, K., 2014. Design and development of thedischarge system of IDDP-1. Geothermics 49, 58–65.
zquierdo, G., Arellano, V.M., Aragón, A., Portugal, E., Martinez, I., 2000. Fluid acidityand hydrothermal alteration at the Los Humeros geothermal reservoir Puebla,
Mexico. In: Proceedings of the World Geothermal Congress 2000, Kyushu-Tohoku, Japan, pp. 1301–1305.ames, R., 1966. Metering of steam-water two phase flow by sharp-edged orifices.In: Proceedings of the Institute of Mechanical Engineers, vol. 180, London, pp.549–572.
rmics 49 (2014) 66–75 75
Larsen, G., Grönvold, K., Thórarinsson, S., 1978. Volcanic eruption through a geother-malborehole at Námafjall, Iceland. Report No. 78-12 Nordic VolcanologicalInstitute & No. RH-78-10. University of Iceland Science Institute, 22 pp.
Moore, J.N., Christenson, B., Browne, P.R.L., Lutz, S.J.,2002. The mineralogic conse-quences and behaviour of descending acid-sulfate waters: an example from theKaraha-Telaga Boda geothermal system, Indonesia. In: Proceedings of the 27thWorkshop on Geothermal Reservoir Engineering. Stanford University, Stanford,CA, pp. 257–265.
Mori, T., Sato, M., Shimoike, Y., Notsu, K., 2002. High SiF4/HF ratio detected in Sat-suma Iwojima volcano’s plume by remote FT-IR observation. Earth Planets Space54, 249–256.
Mortensen, A.K., Gudmundsson, Á., Steingrímsson, B., Sigmundsson, F., Axelsson,G., Ármannsson, H., Björnsson, H., Ágústsson, K., Saemundsson, K., Ólafsson, M.,Karlsdóttir, R., Halldórsdóttir, S., Hauksson, T., 2009. The Krafla Geothermal Field.Overview of geothermal research and a revised conceptual model. Report No.ÍSOR-2009/057.ÍSOR, Reykjavík (in Icelandic), 208 pp.
Mortensen, A.K., Jónsson, S.S., Fridriksson, Th., 2011. Krafla. Analysis of dust fromwell IDDP-01 at the start of discharge May 17, 2011. Report No. ÍSOR 11052.ÍSOR, Reykjavík (in Icelandic), 19 pp.
Moya, P., Nietzen, F., Sanchez, E., 2005. Development of the neutralization systemfor production wells at the Miravalles geothermal field. In: Proceedings of theWorld Geothermal Congress, Antalya, Turkey, 10 pp.
Pang, Z., Ármannsson, H., 2006. Analytical procedures and quality assurance forgeothermal water chemistry, Report No. 1. United Nations Geothermal TrainingProgramme, Reykjavík, 172 pp.
Reyes, A.G., 1990. Petrology of Philippine geothermal systems and the application ofalteration mineralogy to their assessment. Journal of Volcanology and Geother-mal Research 43, 279–309.
Reyes, A.G., 1991. Mineralogy, distribution and origin of acid alteration in Philippinegeothermal systems. In: Matsuhisa, Y., Masahiro, A., Hedenquist, J. (Eds.), Hightemperature acid fluids and associated alteration and mineralization. Report No.277, Geological Survey of Japan. , pp. 59–65.
Sugiaman, F., Sunio, E., Molling, P., Stimac, J., 2004. Geochemical response to pro-duction of the Tiwi geothermal field, Philippines. Geothermics 33, 57–86.
Thórhallsson, S., Pálsson, B., Hólmgeirsson, S., Ingason, K., Matthíasson, M., Bóas-son, H.Á., Sverrisson, H., 2014. Well design for the Iceland Deep Drilling Project.Geothermics 49, 16–22.
Truesdell, A.H., White, D.E., 1973. Production of superheated steam from vapor-dominated geothermal reservoirs. Geothermics 2, 154–173.
Truesdell, A.H., Haizlip, J.R., Ármannsson, H., D‘Amore, F., 1989. Origin and transportof chloride in superheated steam. Geothermics 18, 295–304.
Villa, R.R., Siega, F.L., Martinez-Oliver, M.M., Salonga, N.D., Ogena, M.S., Garcia, S.E.,Buning, B.C., Lichti, K.A., White, S.P., Sanada, N., 2000. A demonstration of the
pp. 25–32.Wang, U., 2008. First contact with inner earth, http://www.greentechmedia.com/
articles/read/first-contact-with-inner-earth-5385/ (accessed 22.04.13).
