Chemical Profiles of Formation Waters from Potash Mine Shafts, Saskatchewan

Transcription

1 Chemical Profiles of Formation Waters from Potash Mine Shafts, Saskatchewan G.K.S. Jensen, B.J. Rostron, M.J.M. Duke 2, and C. Holmden 3 Jensen, G.K.S., Rostron, B.J., Duke, M.J.M., and Holmden, C. (26): Chemical profiles of formation waters from potash mine shafts, Saskatchewan; in Summary of Investigations 26, Volume, Saskatchewan Geological Survey, Sask. Industry Resources, Misc. Rep , CD-ROM, Paper A-7, 8p. Abstract Sixty-three inflow samples from access shafts were collected at three separate potash mines in order to construct three m-deep hydrochemical profiles. Total dissolved solids, bromine concentrations, and δd and δ 8 O stable isotopic compositions increase with depth in each of the three cases. Measured isotopic ratios have not changed in 5+ years since the mine inflows were first sampled, implying little change in the hydraulic regimes at the mines over time. However, the bromine concentrations are typically a factor of five lower than previously reported. Newer analytical techniques have improved the accuracy, precision, and resolution of the hydrochemical profiles. Results indicate that the salinity of the inflow waters originated as mixtures of evaporatively concentrated seawater, meteoric water, and brine derived from halite dissolution. Extremely concentrated brines (TDS >525 g/l) were found at the Cory and Allan potash mines, which are some 55 km apart, but their role in the paleohydrogeology of the basin remains uncertain. Keywords: potash, brine, isotopes, bromide.. Introduction The Prairie Evaporite Formation in the Elk Point Basin in Saskatchewan, Canada, contains some of the largest potash deposits in the world. Problematic mine-level flooding has threatened the longevity of potash mines in Saskatchewan almost since their construction (Wittrup et al., 987). It is, therefore, essential to determine the origin of inflows in order to remediate them or to plan preventative maintenance. Previously, hydrochemical tracers such as chloride or major ions have had limited success because the inflows pass through, and dissolve, evaporites of the Prairie Evaporite Formation. Isotopic tracers (e.g., δ 8 O, δd) have met with more success at fingerprinting waters in the potash mines (Wittrup et al., 987; Wittrup and Kyser, 99) and oil fields in the basin (Rostron and Holmden, 2). In addition, although bromine has worked well in oilfield applications (Iampen and Rostron, 2), its use in potash mine-inflow fingerprinting has been hampered by analytical interferences. A recently developed analytical technique using Epithermal Neutron Activation Analysis (ENAA) for determining bromine concentrations in brines (Duke and Rostron, in press) facilitates the re-examination of bromine as a tracer for fingerprinting inflows in potash mines. This, combined with newer continuous-flow stable-isotope methods that allow measurements on small (<5 ml) samples, and the more than 5+ years that have elapsed since the pioneering fingerprinting work by Wittrup and Kyser (99), prompted a re-examination of the chemistry of the inflows into potash mine shafts in Saskatchewan. Thus, the primary objective of this study was to obtain new hydrochemical and stable-isotope vertical profiles at the Potash Corporation of Saskatchewan (PCS) Rocanville, Cory, and Allan potash mines in Saskatchewan (Figure ), and to compare these results to previously published data. Although a paper describing this study has previously been published (Jensen et al., 26), the current work mainly differs in that it contains the analytical data. 2. Methods Sixty-three inflow samples from access shafts were collected at three separate mines (Figure 2) in order to construct three m deep hydrochemical profiles for each mine. Sample points were identified by visual identification of Department of Earth and Atmospheric Sciences, -26 Earth Sciences Building, University of Alberta, Edmonton, AB T6G 2E3; gjensen@ualberta.ca 2 SLOWPOKE Nuclear Reactor Facility, 326 Dentistry/Pharmacy, University of Alberta, Edmonton, AB T6G 2N8. 3 Department of Geological Sciences, University of Saskatchewan, 4 Science Place, Saskatoon, SK S7N 5E2. Saskatchewan Geological Survey Summary of Investigations 26, Volume

2 53 o 49 o AB o PCS-Cory Canada U.S.A. PCS-Allan km PCS-Rocanville SK MT ND Figure - Location of sampled potash mines. water seeping into the dry mine shafts. Samples were collected in sealed plastic containers, and the depth below ground was calculated from the length of cable dispensed to support the elevator car. Field filtration was done using.45 µm PES filters to remove any suspended solids. Oxygen and hydrogen stable isotopes were determined at the University of Saskatchewan. An aliquot of brine was injected directly into a continuous flow (CF) Delta plus XL isotope ratio mass spectrometer. Isotopic values were reported in delta (δ) notation as 8 O/ 6 O and D/H relative to Vienna Standard Mean Ocean Water (VSMOW). Isotopic measurements have an uncertainty of ±3 and ±.3 for δd and δ 8 O, respectively. ENAA was completed at the SLOWPOKE Reactor at the University of Alberta to determine the sodium, chloride, bromide, and iodine concentrations of each sample. Bromide concentrations varied from 2 to 652 mg/l with uncertainties ranging from ±3.9% and ±.6% for this concentration range. Ion chromatography was not used for bromide analysis as a dilution factor of ~, would have been required, due to the high TDS contents of the samples. Major ions, trace metals, and alkalinity data were determined at a commercial laboratory in Edmonton, Alberta using ICP-OES, ICP-MS, and titration techniques. Analytical uncertainties for the commercially determined ions varied with the different techniques, and details are available upon request to the authors. 2 o 3. Results and Discussion Each of the 63 samples collected for this study were fully analyzed for their major and minor ion concentration and isotopic composition (Table ). There is a large variation in the chemical composition of the water samples from the three mines. In general, the total dissolved solids (TDS) increases with depth at each mine (Table ). The largest variation is observed at the Cory Mine, where TDS ranges from approximately 32 mg/l at shallow depths (~58 m) to over 54 mg/l at the mine level (25 m). Similar patterns are observed at the Allan Mine with TDS increasing from approximately 5 mg/l near the surface (~98 m) to almost 36 mg/l just above the mine level (~95 m). Insufficient sample volume of the deepest sample from the Allan Mine precluded a full analysis, so data in Table for sample 2-3 represent minimum TDS values for that depth. TDS data from the Rocanville Mine show a pattern that is slightly different to that at the other two mines: the samples start off at higher values (>39 mg/l) because of their deeper initial sample depth (~475 m), and at greater depths, the shaft inflows do not attain the very high TDS values seen at the Cory and Allan mines (the maximum observed TDS values at Rocanville are approximately 35 mg/l). Major ion concentrations that make up the observed TDS patterns also vary by mine in a systematic pattern. Chloride is the dominant ion in the dataset (Table ), and ranges in concentration from 35 mg/l at a depth of 58 m (Cory Mine) to mg/l at 95 m in the Allan Mine. Sodium is the second most abundant ion and the most prevalent cation and its concentration increases with depth. However, the highest concentrations of sodium do not always occur in the deepest samples. For example, in the Cory and Allan mines, sodium concentration peaks at 8 to mg/l, and then decreases with depth. This decrease in sodium is attributed to the precipitation of halite. Calcium concentrations range from almost nothing (~25 mg/l) to values of over 3 mg/l at the Cory and Allan mines. These calcium chloride brines are of scientific and economic interest having been sold commercially from the Cory mine (Buchinski, 988). Magnesium and potassium make up lesser components of the water samples. Magnesium values range from 6 mg/l to 22 3 mg/l and are essentially subdued versions of calcium concentrations. Potassium values range from ~ mg/l to approximately 2 mg/l (Cory Mine) and for the most Saskatchewan Geological Survey 2 Summary of Investigations 26, Volume

3 MESOZOIC PALEOZOIC UPPER LOWER UPPER Cory Allan Rocanville QUATERNARY Colorado Group CRETACEOUS M.S.L. Miss. Madison Grp. Mannville Group Birdbear Fm. Watrous Fm. Three Forks Grp. Duperow Fm. DEVONIAN Souris River Fm. Dawson Bay Fm. MIDDLE Prairie Fm. 55 km 33 km Sample location Figure 2 - Sample locations in the potash mine shafts, modified after Wittrup and Kyser (99). part track the chloride concentrations. Sulphate was not measured directly in this study. Instead, measured sulphur was converted stoichiometrically to provide a proxy for sulphate, values of which ranged from essentially none (~ mg/l) to ~9 mg/l at the mid levels of the Rocanville Mine. Sulphate follows a similar pattern to sodium in that values increase with depth to the Duperow aquifer, then decrease with further depth down to the mine level. Maximum values of sulphate attained at Rocanville (~9 mg/l) are significantly higher than maximum values attained at either Cory or Allan (~5 to 6 mg/l). Finally, for bicarbonate there are insufficient data to draw any conclusions. Bicarbonate has previously been shown to be essentially invariant with depth in the Williston Basin (Iampen, 23). Furthermore, analytical methods (i.e., small sample volumes) precluded analysis of this ion in most samples. Major ion data reveal there are three end-member types of waters present in the potash mine shafts (Figure 3). They are: i) shallow, low TDS samples (<3 mg/l); ii) deeper, sodium chloride brines (Na-Cl) with TDS between and 3 mg/l; and iii) deepest, calcium chloride brines (Ca-Na-Cl) with TDS >45 mg/l. Waters that fall in between these end members are interpreted to be mixtures of the three types. The two most important parameters for fingerprinting potash shaft water inflows are the bromine and the stable isotopic compositions. m Saskatchewan Geological Survey 3 Summary of Investigations 26, Volume

4 Table - Sample depth, chemical composition, and stable isotope data for the Cory, Allan, and Rocanville mines. Note: Sulphur was reported in its elemental form. Sulphur present is assumed to be in the form of SO4. Ag, Al, Be, Bi, Cd, Co, Mo, Ni, Th, Sn, and V were all below detection limits. Cu, Sb, Se, and Pb were detected in trace amounts in the most saline samples. Allan Mine d 8 O dd Cl Na K Ca Mg Br I SO 4 HCO3 TDS Arsenic Barium Boron Chromium Iron Lithium Manganese Silicon Strontium Titanium Zinc Sample Number Depth (m) (, SMOW) (, SMOW) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) Montana Group n.a. n.a. n.a n.a. n.a. 555 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a Montana Group n.a < <..28 < < Montana Group n.a < < < Montana Group n.a. 66 < < < Montana Group n.a < < Montana Group n.a < <.2.99 < < Colorado Group n.a. n.a. n.a n.a. n.a n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a Watrous n.a. n.a. n.a n.a. n.a. 68 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a U Duperow <.2 < <..6 < < U. Duperow <.2 < <..46 <.5 < < U. Duperow <.3 66 n.a <.2 < <.5 <..44 <.5 < < U. Duperow <.4 65 n.a <.2 < <..44 <.5 < < U. Duperow n.a. n.a. n.a. 5.7 n.a. n.a n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a Duperow (Wymark) n.a. 998 < <. <2 2. < < Duperow (Wymark) <.4 47 n.a. 827 < <.5 < < < L. Duperow n.a. 385 < <.5 <..77 <.5 < < L. Duperow n.a < <.5 <. 2.8 < < U. Souris River n.a < <. < < < U Souris River n.a < <. <2.4 <. 38 < U Souris River n.a < < < U Souris River n.a < < < L. Souris River n.a < <2 2.5 <. 9 < L. Souris River n.a < L. Souris River n.a. n.a. n.a n.a. n.a n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. Cory Mine Montana Group (Belly River) n.a < Montana Group (Lea Park) n.a < <..82 < < Lower Colorado n.a < < < Birdbear < <. < < Birdbear n.a <. < < < < U. Duperow n.a. 49 <. < < < < U. Duperow <. < < < U. Duperow n.a <. < < <.2 < < Duperow n.a <. < < < < L. Duperow n.a <. < < < < L. Duperow <.3 57 n.a <. < < < L. Duperow <.3 59 n.a <. < < < < U. Souris River n.a < < <. < 459 < U. Souris River n.a < <. < <. < 397 <. < U. Souris River n.a < < <. < 39 < U. Souris River n.a < <. <2 4.7 <. < 47 < L. Souris River n.a <.4 <.2 52 <. <2 7.4 <. < 42 <. < L. Souris River n.a <. <2 8.2 <. < 7. < Dawson Bay n.a < < Dawson Bay n.a < < Mine inflow < < Rocanville Mine Lodgepole n.a < < < Three Forks <. < < < < Three Forks - Birdbear <.2 < < < < Birdbear <.2 < < <.5 < < U. Duperow n.a <.4 < < <. < < U. Duperow n.a <.4 < <. < <. < < U. Duperow n.a <.4 < <. <2 8.3 <. < < Duperow <.4 < <. < < 72.. < L.Duperow n.a < <. <2 2 <. < 85.3 <. < L. Duperow n.a <.4 < <. < <. < 69.5 < L. Duperow n.a < <. < <. < 7.3 <. < L. Duperow n.a < <. < <. < 67. <. < L. Duperow n.a <.4 < <. <2 2.3 <. < < Souris River <.4 < <. <2 9. <. < 79.7 <. < Souris River n.a < < < 24 < Dawson Bay n.a <.4 < <. <2 9 <. < 98. <. < Dawson Bay <.4 < <. < < 94. <. < Dawson Bay n.a. 34 <.4 < <. <2 9.2 <. < 98.5 <. <.2 Saskatchewan Geological Survey 4 Summary of Investigations 26, Volume

5 K + Na Ca Mg Meteoric Fe CO3. K + Na Ca Mg Na-Cl Bromine concentrations versus depth for the Cory, Allan, and Rocanville potash mines are shown in Figure 4. Bromine increases with depth at all three mines and ranges from approximately 2 mg/l in nearsurface aquifers (58 m depth), to approximately 55 mg/l in Devonian carbonate aquifers (25 m depth) just above the mining level. Inflow samples with bromine concentrations of >54 mg/l are found at the Cory and Allan mines, which are some 55 km apart. Their wide spacing implies that they are more than just an isolated occurrence at each mine. The origin and implication of these extremely bromine-rich brines are subjects of on-going research. There are significant differences K + Na Cl between the newly determined bromine profiles and those previously published. Firstly, Ca HCO3 there are more samples in these new profiles because the Mg SO analytical techniques require less sample (i.e.,.25 ml), hence Fe CO3 smaller inflows can be analyzed. This gives our new profiles a. finer resolution. Secondly, at any meq/l given depth, the bromine concentrations in this study are Figure 3 - The three end-member types of water found in this study. significantly lower than measured in the previous study, with the difference becoming larger at greater depths (i.e., at higher concentrations). This difference is attributed to the analytical method used by Wittrup et al. (987) (ion selective electrode) being affected by interference with increasingly high chloride concentrations with increasing depth. In some cases, the bromine concentrations obtained using ENAA in this study were an order of magnitude less than previously determined. Thirdly, samples with such elevated bromine (>5 mg/l) were not previously reported at the Allan Mine. Isotopic compositions of the mine inflows are illustrated by means of plots of δd versus depth for each mine (Figure 5). Isotopic compositions of δ 8 O are not shown because they follow a similar pattern to δd and previous work has demonstrated a linear relationship between δd and δ 8 O for the Williston Basin (Wittrup and Kyser, 99). At all three mines, δd compositions in the mine inflows increase with depth (Figure 5). At shallow depths, water samples have an isotopic composition (-6 ) similar to that of meteoric water in Saskatchewan (Wittrup et al., 987). At the greatest depths in the mines, isotopic compositions in Devonian strata reach -45, similar to that reported previously (Wittrup and Kyser, 99; Rostron and Holmden, 2). Variations in the isotopic compositions of the inflows as a function of depth are interpreted to reflect differences caused by the origin and/or mixing history of the fluids in different aquifers at each level of the shaft. Using these data, an isotopic fingerprint has been obtained for the aquifers above the mine, providing excellent tracers of the origin(s) of the shaft inflows. Current isotopic compositions versus depth compare very closely to those published previously (Wittrup et al., 987; Wittrup and Kyser, 99), with a few differences. Firstly, as mentioned above, the isotopic profiles of this study are more detailed than previously shown as a result of an increased number of samples made possible by the CF analytical techniques. Continuous flow can measure the isotopic composition on small (<2 ml) brine samples, Cl HCO3 SO4 Fe CO3. Ca-Cl Cl HCO3 SO Saskatchewan Geological Survey 5 Summary of Investigations 26, Volume

6 Br (mg/l) Br (mg/l) Wittrup et al. (987) This Study Br (mg/l) Depth (km) Cory Allan Rocanville Figure 4 - Bromine concentrations versus depth for Cory, Allan, and Rocanville potash mines; note that bromine is a logarithmic scale. δd ( VSMOW) δd ( VSMOW) δd ( VSMOW) Wittrup et al. (987) This Study Depth (km) Cory Allan Rocanville Figure 5 - δd (VSMOW) versus depth for the Cory, Allan, and Rocanville potash mines. Saskatchewan Geological Survey 6 Summary of Investigations 26, Volume

7 and thus small shaft inflows can be included in the vertical profiles. This has resulted, for example, in data for the deep aquifers above the mining level in Cory and Allan mines (Figure 4) and for other levels in the stratigraphic profiles that were not previously available. Secondly, with the exception of data from 6 to 8 m in Rocanville Mine, the isotopic data from Wittrup et al. (987) overlap those collected in this study. This indicates there has been no change over time in the isotopic composition of the shaft inflows. Observed differences in isotopic compositions at the Rocanville Mine are attributed to sample collection at or near pumping stations where inflows are collected for pumping to the surface, and hence are not considered representative. Revised bromine concentration data from the three mine shafts provide additional insight into the origin of the formation waters in the basin. A plot of Na/Br versus Cl/Br (Figure 6) illustrates a linear relationship in the data, but with a wide variation compared to seawater. Variations in the Na/Br ratios can be used to infer the origin of salinity in these samples (Walter, 99). Samples from Rocanville all plot above and to the right of seawater on the Na/Br versus Cl/Br plot, indicating the salinity in these samples originated from the dissolution of halite. In contrast, samples from Cory and Allan plot over a wide range on the Na/Br versus Cl/Br plot, both above and to the right of seawater (indicating dissolution of halite) and below and to the left of seawater indicating the presence of an evaporated end-member brine. Samples with Br >5 mg/l found at and above the potash mining level at Cory and Allan are thought to represent the evaporated end-member brine. The recognition of the evaporatively concentrated end-member brine at the Allan Potash Mine, and the fact that it is found 55 km away from the previously known occurrence at the Cory Mine, together suggest that this formation-water is more widespread than previously known. These observations bear further investigation. 4. Conclusions Sixty-three samples of inflows into access shafts at PCS Cory, Allan, and Rocanville potash mines were analyzed and used to construct hydrochemical profiles through aquifers in the Williston Basin. There have been no significant changes in the isotopic compositions of the shaft inflows in the5+ years since they were initially studied, indicating that the origin of the inflows have not changed. Newer analytical techniques for both bromine (ENAA) and stable isotopes (CF) enabled smaller samples to be analyzed than in previous studies, allowing creation of higher resolution depth profiles which are both more accurate and more precise for bromine. Sampling identified an evaporatively concentrated brine at the Allan Mine, previously only recorded at the Cory Mine. In addition to stable isotopes of oxygen and hydrogen, bromine concentration data can be used to better define the source of mine-level flooding brines. 2 Cl/Br 5 Na/Br Figure 6 - Na/Br versus Cl/Br (mg/l) of shaft inflow samples. Allan Rocanville Cory Seawater Saskatchewan Geological Survey 7 Summary of Investigations 26, Volume

8 5. References Buchinski, K.W. (988): The occurrence, recovery, and commercial application of calcium chloride brine from a Saskatchewan potash mine; in 9th Annual Meeting of CIM, Edmonton, Paper No. 22, 6p. Duke, M.J.M. and Rostron, B.J. (in press): Determination of bromine, iodine, chlorine and sodium in highly saline formation waters by Epithermal NAA; J. Radio. Nuc. Chem. Iampen, H.T. (23): The genesis and evolution of pre-mississippian brines in the Williston Basin, Canada-USA; unpubl. M.Sc. thesis, Univ. Alberta, Edmonton, 24p. Iampen, H.T. and Rostron, B.J. (2): Hydrochemistry of pre-mississippian brines, Williston Basin, Canada-USA; J. Geochem. Explor., v69-7, p Jensen, G.K.S., Rostron, B.J., Duke, M.J.M., and Holmden, C. (26): Bromine and stable isotopic profiles of formation waters from potash mine-shafts, Saskatchewan, Canada; J. Geochem. Expl., v89, p7-73. Rostron, B.J. and Holmden, C. (2): Fingerprinting formation-waters using stable isotopes, Midale Area, Williston Basin; Can. J. Geochem. Expl., v69-7, p Walter, L. (99): Br-Cl-Na systematics in Illinois basin fluids: Constraints on fluid origin and evolution; Geol., v8, p Wittrup, M.B. and Kyser, T.K. (99): The petrogenesis of brines in Devonian potash deposits of western Canada; Chem. Geol., v82, p3-28. Wittrup, M.B., Kyser, T.K., and Danyluk, T. (987): The use of stable isotopes to determine the source of brines in Saskatchewan potash mines; in Gilboy, C.F. and Vigrass, L.W. (eds.), Economic Minerals of Saskatchewan, Sask. Geol. Soc., Spec. Publ. No. 8, p Saskatchewan Geological Survey 8 Summary of Investigations 26, Volume