downstream processing of reverse osmosis brine
TRANSCRIPT
Accepted Manuscript
Downstream Processing of Reverse Osmosis Brine: Characterisation of PotentialScaling Compounds
Masuduz Zaman, Greg Birkett, Christopher Pratt, Bruce Stuart, Steven Pratt
PII: S0043-1354(15)00283-3
DOI: 10.1016/j.watres.2015.05.004
Reference: WR 11275
To appear in: Water Research
Received Date: 16 October 2014
Revised Date: 24 February 2015
Accepted Date: 1 May 2015
Please cite this article as: Zaman, M., Birkett, G., Pratt, C., Stuart, B., Pratt, S., Downstream Processingof Reverse Osmosis Brine: Characterisation of Potential Scaling Compounds, Water Research (2015),doi: 10.1016/j.watres.2015.05.004.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
1
Downstream Processing of Reverse Osmosis Brine: 1
Characterisation of Potential Scaling Compounds 2
3
Masuduz Zamana, Greg Birketta, Christopher Prattb, Bruce Stuartc, and Steven Pratta,* 4
5
a School of Chemical Engineering, The University of Queensland, St Lucia, 4072, 6
Queensland, Australia 7
8
b Department of Agriculture, Fisheries and Forestry, 203 Tor St, Toowoomba, 4350, 9
Queensland, Australia 10
11
c Australia Pacific LNG, 135 Coronation Drive, Milton, 4064, Queensland, Australia 12
13
* Corresponding Author(s): Steven Pratt, E-mail: [email protected], Tel: +61 7 33654943 14
Masuduz Zaman, E-mail: [email protected], Tel: +61478975778 15
16
ABSTRACT 17
Reverse osmosis (RO) brine produced at a full-scale coal seam gas (CSG) water treatment 18
facility was characterized with spectroscopic and other analytical techniques. A number of 19
potential scalants including silica, calcium, magnesium, sulphates and carbonates, all of 20
which were present in dissolved and non-dissolved forms, were characterized. The presence 21
of spherical particles with a size range of 10-1000 nm and aggregates of 1 to 10 microns was 22
confirmed by transmission electron microscopy (TEM). Those particulates contained the 23
following metals in decreasing order: K, Si, Sr, Ca, B, Ba, Mg, P, and S. Characterization 24
showed that nearly one-third of the total silicon in the brine was present in the particulates. 25
Further, analysis of the RO brine suggested supersaturation and precipitation of metal 26
carbonates and sulphates during the RO process should take place and could be responsible 27
for subsequently capturing silica in the solid phase. However, the precipitation of crystalline 28
carbonates and sulphates are complex. X-ray diffraction analysis did not confirm the presence 29
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
2
of common calcium carbonates or sulphates but instead showed the presence of a suite of 30
complex minerals, to which amorphous silica and/or silica rich compounds could have 31
adhered. A filtration study showed that majority of the siliceous particles were less than 220 32
nm in size, but could still be potentially captured using a low molecular weight ultrafiltration 33
membrane. 34
Keywords: Reverse osmosis; brine; scaling compound; filtration 35
1. Introduction 36
Groundwater desalination is vital for water supply as well as management of water 37
extracted during mining and agricultural activities in many parts of the world. Reverse 38
osmosis (RO) membrane separation is widely used for this purpose, with the products being 39
desalinated water and brine. The brine is typically 15-20% of the feed water. Near the coast 40
the brine can be discharged to the ocean, but in inland regions management of this brine is 41
one of the key challenges for RO desalination plants (Morillo et al., 2014). 42
One option for sustainable management of RO brine is concentration and recovery of 43
salts, which results in the production of more desalinated water and near zero waste-liquid 44
discharge (ZWLD) (Bond & Veerapaneni, 2007). Salt recovery from brine can be achieved 45
using evaporative separation technology, although such technologies are highly susceptible to 46
the formation of scale (Cipollina et al., 2011; Mericq et al., 2010). Mineral precipitation and 47
scaling occur when the concentrations of scale precursor ions (Ca2+, Mg2+, Sr2+, Ba2+, OH-, 48
SO42-, CO3
2-, H3SiO4-, etc.) exceed the solubility limit of silicates and various mineral salts. 49
Although precipitation and fouling of these compounds on RO membrane surfaces has been 50
studied significantly over the past years (Goosen et al., 2004; Potts et al., 1981), less attention 51
has been paid on their formation in the reject brine during RO post-treatments. 52
Of particular concern for fouling RO membranes and downstream evaporators is 53
silica scale, which, once deposited, is extremely difficult to remove. Silica may be present in 54
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
3
particulate or dissolved form (Zaman et al., 2013). The rate of deposition of particulate silica 55
is one or two orders of magnitude higher than of polymerisation of dissolved silica (Weres & 56
Tsao, 1981). Particulate silica deposits are generally less dense, porous in nature and easier to 57
remove from heat exchangers surface. However, deposits from a mixture of colloidal and 58
dissolved silica are more dense and hard compared to pure colloidal silica deposits (Iler, 59
1979). 60
The effect of temperature, pH, and salinity on silica solubility has been extensively 61
studied in the literature. At 25oC, neutral pH, and zero salinity, solubility is in the order of 62
100 to 120 mgL-1. This increases with increasing temperature, increases sharply at very high 63
pH (above 9.5), but decreases with increasing salinity (Fournier, 1970; Marshall & 64
Warakomski, 1980; Okamoto et al., 1957). The concentrations of silica and other problematic 65
compounds in groundwater are generally far below those that lead to deposition, but their 66
concentrations are elevated in RO brine, generally by a factor of 5 to 10 (Antony et al., 2011; 67
Mi & Elimelech, 2013; Sanciolo et al., 2014; Tomaszewska & Bodzek, 2013; Zhang et al., 68
2014), which could possibly results in scale formation. 69
This work investigates the consequence of RO treatment on the solubility of scaling 70
compounds, with a view to understanding the potential to capture these compounds prior to 71
subsequent brine concentration. Emphasis is placed on silica, its solubility and its interaction 72
with metal precipitates, as silica can co-precipitate with other common minerals and 73
multivalent cations forming complex mineral structures (Badruk & Matsunaga, 2001; Butt et 74
al., 1995; Butt et al., 1997; Greenlee et al., 2009; Hsu et al., 2008; Ji et al., 2010; Mariah et 75
al., 2006; Mericq et al., 2010; Pandey et al., 2012; Qu et al., 2009; Sheikholeslami & Bright, 76
2002). The groundwater considered in this work is water associated with the extraction of 77
coal seam gas (CSG). Coal seam gas (CSG), a growing industry in Australia, anticipated to 78
produce in the order of 30 GL of reverse osmosis brine every year (Klohn Crippen Berger, 79
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
4
2012). Several CSG producers in Australia are considering a zero waste-liquid discharge 80
(ZWLD) strategy for sustainable management of brine, which will require post RO 81
evaporation and salt crystallization. 82
83
2. Experimental 84
Samples from a full-scale CSG water treatment facility were collected for the 85
characterization of chemical composition. Fig. 1 shows the typical treatment train for ZWLD 86
CSG water treatment. The produced CSG water from the coal seams is separated from the 87
gas at the well site and sent to the water treatment facility. The raw water undergoes 88
microfiltration followed by RO desalination. The RO brine is sent to the brine pre-treatment 89
process for removal of scaling compounds. The pre-treated brine is then sent to the brine 90
treatment facility where salt is recovered. 91
Physical and chemical characterization of silica and other potential scalants were 92
conducted using various spectroscopic and analytical techniques. Water quality data obtained 93
from the CSG water treatment plant was used to help interpret the experimental observations 94
about scaling compounds in the reverse osmosis brine. 95
96
2.1. Sample Collection 97
CSG water and brine samples were collected from water treatment facility near Roma 98
in Queensland, Australia. Samples were collected from the input and output of major process 99
units as shown in Fig. 1. All samples were collected on the same day with minimal time delay 100
in order to minimize the fluctuations in water quality. 101
Figure 1 102
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
5
2.2. Quantification of silica 103
Molybdosilicate colorimetric method (HACH 8185) was used to measure the 104
dissolved silica in all water and brine samples (Ali et al., 2004). For brine samples, a five 105
times dilution was applied in order to maintain an ionic strength of less than 0.05 molar. A 106
HACH DR2700 spectrophotometer was used to measure the dissolved silica concentration. 107
The instrument was calibrated with a 100 mgL-1standard solution obtained from HACH, 108
Australia. All the samples were filtered with a 0.45µm filter before analysis to remove 109
suspended particles. Triplicates revealed a standard deviation of ±1mgL-1 Si. 110
The quantification of total silicon (dissolved and particulate) was carried out by the 111
hydrofluoric (HF) acid transformation method (Zeng et al., 2007). HF digestion transformed 112
the non-dissolved silica into dissolve silica. Inductively coupled plasma-optical emission 113
spectroscopy (ICP-OES) was then used to measure the total silicon content in water and brine 114
samples. The digestions were performed by microwave in Teflon vessels; the digested 115
samples were made up in plastic volumetric flasks and a Teflon Sturman-Masters design 116
spray chamber and ceramic nebuliser was used to transport the sample to the ICP torch. 117
Standard silicon solutions of 0, 10, 50, and 100 ppm was used to calibrate the instrument. The 118
particulate silica concentration of a given sample was calculated from the difference between 119
total and dissolved silica concentration. 120
121
2.3. Transmission Electron Microscopy (TEM) Analysis 122
Size and shape of the particulates was analysed by transmission electron microscopy. 123
20 µL of CSG brine with particulate silica concentration of 30 mgL-1 was pipetted onto a 124
carbon coated 200 mesh copper grid. The particles were allowed to settle for 5 min and the 125
excess liquid was removed by a blotting paper, and then examined using a JEOL 1010 126
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
6
scanning transmission electron microscope (STEM), operated at an acceleration voltage of 127
200 keV. 128
129
2.4. Elemental composition of particulate material 130
The elemental composition of the solids was determined from the chemical difference 131
between the filtered and un-filtered sample. The brine sample was filtered using a VivaSpin 132
15R modified regenerated cellulose ultra-filtration membrane (MWCO: 5000) using a fixed 133
angle centrifuge at 6000g. The filtration tube was pre-rinsed with distilled water at the same 134
speed to remove trace amount of glycerine and sodium azide associated with the membrane. 135
The filtrate and concentrate was decanted after filtering the CSG brine. After filtration the 136
brine filtrate and unfiltered brine was analysed for inorganic elements using ICP-OES, and an 137
Flash 2000 CHNS/O elemental analyser (for C, O, N, S). Filtration of particulate solids was 138
also studied by filtering the brine solution through 0.22µm, 0.45µm, and 1µm filter paper. 139
The filtrates were analysed with ICP and elemental analyser. 140
141
2.5. X-ray diffraction study 142
The phase structure of the brine solids were identified using a Bruker D8 Advance X-143
Ray Diffractometer. The dried powdered sample was irradiated with X-rays from Cu Kα 144
radiation (λ=1.5418A°) operating at 40mA and 40kV. The 2θ range was from 10° to 90° at 145
scan rate of 1 sec/step. The fixed divergence slit was 0.26deg and receiving slit width was 146
5.0mm. Sample was rotated at 15rpm. 147
3. Results 148
3.1. Coal seam gas water chemistry 149
The CSG water used in this study was obtained from a full scale water treatment 150
facility in Roma, Queensland. The RO process operates at a desalting ratio of 80%, which in 151
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
7
turn generates a brine stream which is about 5 times concentrated compared to the raw CSG 152
water. The water quality data of RO feed and RO brine is shown in Table 1. In short, CSG 153
water is very rich in sodium and bicarbonate, with potassium, calcium, magnesium and 154
strontium also prevalent. The silica concentration in raw CSG water is relatively low, but the 155
RO water recovery process increases the silica concentration to near saturation. 156
157
158
Table 1. 159
3.2. Silica characterization in coal seam gas water 160
In natural water systems, silica can exist in both dissolved and particulate forms 161
(Belton et al., 2012). Although the literature on particulate silica commonly refers to the 162
polymeric colloidal silica and biogenic silica, it can also refer to the silica-containing mineral 163
precipitates in a groundwater system. Fig. 2 shows the forms of silica in various process 164
streams of the CSG water treatment facility. The difference between the total and dissolved 165
silica reveals the particulate silica in each sample. The figure shows that in all process 166
streams silica exists in both dissolved and particulate forms. The particulate form of silica in 167
the pond water may originate form biological or geochemical processes. The increased total 168
silica concentration in the pond water can be attributed to the unsteady state of operation at 169
the water treatment facility and evaporation of water from the pond. The RO brine shows 170
increased total silica concentration due to a concentration factor of 5 through reverse osmosis. 171
172
It is also evident that the particulate fraction of silica is not the same in all streams. 173
The pond water and RO brine consist of significant fractions of particulate silica compared to 174
other water samples. While, as already mentioned, the origin of particulate silica in pond 175
water could be from biological or geological sources, the particulate silica in RO brine must 176
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
8
be from the precipitation/polymerization of scalants in the brine (Antony et al., 2011; Mi & 177
Elimelech, 2013; Tomaszewska & Bodzek, 2013; Zhang et al., 2014). 178
179
180
Figure 2 181
182
183
3.2. Morphology and size characterization of brine solids 184
Transmission electron microscopy was used to investigate the morphology and state 185
of dispersion of particulates in CSG brine. Fig. 3 shows the morphology of particulate 186
mineral precipitates in CSG brine. The particles were found to be generally spherical. The 187
higher magnification image in Fig.3B shows a core-shell structure of these nanoparticles. 188
Similar structures were observed for precipitates in the RO brine with high silica 189
concentration (Malki & Abbas, 2013). The particle size was found to be in the size range of 190
10 - 1000 nm. Also, some aggregated network-like structures (Fig. 3C) were observed which 191
can be due to the high ionic strength of the brine solution (Belton et al., 2012). Some of the 192
aggregation could also be artefact of the drying process of TEM sample preparation. 193
194
195
Figure 3. 196
3.3. Filtration of brine solids 197
Silica characterization revealed that a significant fraction (nearly one-third) of total 198
silica in RO brine is associated with particulates. Therefore, filtration could potentially 199
remove some silica from the CSG brine. It can be seen from Fig. 4, the total silica 200
concentration in CSG brine was reduced from 120 mgL-1 to about 105 mgL-1in conventional 201
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
9
microfiltration membranes (0.22, 0.45, and 1 µm pore size). On the other hand, with an 202
ultrafiltration membrane (5000 MWCO) the silica concentration was reduced from 203
120 mgL-1 to just 86 mgL-1. 204
205
Figure 4 206
207
208
3.4. Elemental composition of brine solids 209
Elemental composition of the colloidal solids in CSG brine was determined from the 210
water chemistry difference between filtered and un-filtered sample. CSG brine was filtered 211
using a VivaSpin 15 ultrafiltration membrane (MWCO: 5000). Three parallel filtrations were 212
conducted and each filtrate was analysed three times so that the chemical difference could be 213
evaluated. Fig. 5 shows the difference in elemental composition of CSG brine before and 214
after the ultrafiltration. K, Si, Sr, Ca, B, Ba, Mg, P, and S were found to be the key elements 215
of solid phase in CSG brine. The difference in elemental composition for all other elements 216
were found to be below the limit of quantification for ICP-OES, and hence not considered in 217
the solid phase composition. 218
219
Figure 5 220
221
3.4. Supersaturation in RO brine 222
Reverse osmosis concentrates the water in the CSG water treatment plants. The brine 223
is expected to be further concentrated to achieve zero liquid waste discharge. These multi-224
stage concentration processes of CSG brine result in an increase of the saturation index of 225
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
10
various sparingly soluble salts; the saturation index of mineral salts is given by the following 226
equation (Rahardianto et al., 2007): 227
Saturation Index, SIx=IAP/Ksp,x ………………………………………...(1) 228
Where, IAP is the ion activity product and Ksp,x is the solubility product for mineral 229
salt, x. 230
The activity of a given ion in the solution can be calculated by multiplying the given 231
concentration with its activity coefficient in the solution, which can be obtained from Debye-232
Huckel equation (Tissue, 2013). The solubility of the given minerals can be obtained from 233
literature and corrected for ionic strength of the solution (Tissue, 2013). 234
An increase in saturation index above 1 could lead to the precipitation of dissolved 235
minerals and formation of scalants in the CSG brine, which could cause fouling problems on 236
process equipment. This would severely hinder the heat transfer efficiency and reduce the 237
overall process performance. 238
239
240
Table 2 241
242
243
It can be seen from Table 1, the RO concentrates the CSG water by a factor of nearly 244
5 resulting in the saturation index, SIx, to exceed 1 for a number of minerals (See Table 2): 245
4.7 (calcium carbonate), 79.67(calcium phosphate), 3.39 (strontium carbonate), 1.1(silica). 246
Clearly, concentrating this brine in the multiple effect evaporator will further increase the 247
saturation index. 248
Calcium carbonate is a common scaling compound. It is also well known for 249
adsorption/co-precipitation with silica (Badruk & Matsunaga, 2001; Hsu et al., 2008; Qu et 250
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
11
al., 2009). It can be seen from Table 2, RO brine is oversaturated with respect to calcium 251
carbonate and Fig.3 clearly shows the presence of spherical colloidal particles, which could 252
be possibly due to the formation of calcium carbonate precipitate. It should be noted that for 253
precipitation of CaCO3, calcium is the limiting element as there is a significant amount of 254
carbonates in both the RO feed and RO brine. 255
Candidate precipitates were checked for by XRD (Fig. 6). However, the investigation 256
did not confirm the presence of the simple precipitates listed in Table 2, but rather a suite of 257
complex minerals (Table 3). The apparent discrepancy between the XRD data and 258
precipitation based on saturation indices is discussed in the following section. 259
260
261
Figure 6 262
263
264
Table 3 265
266
4. Discussion 267
The characterization of silica in CSG water shows a significant amount of dissolved 268
silica was transformed to solid form in the RO brine. It can be seen from Fig. 2 that the RO 269
feed contains mostly dissolved silica (more than 98%). However, in RO brine nearly one-270
third of the total silica is particulate silica. 271
Silica in solid form could exist as adsorbed/co-precipitated silica with precipitates like 272
CaCO3 in that stream (Qu et al., 2009). The elemental composition study confirms the 273
colloidal particles contained significant amounts of calcium and silicon, as well as K, Sr, B, 274
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
12
Ba, P, and S. Considering the concentration of Ca2+ in RO brine (and the abundance of 275
bicarbonates and carbonates) the saturation index of calcium carbonate confirms precipitation 276
of CaCO3 is certainly possible. The same is true for SrCO3. And both CaCO3 and SrCO3 can 277
co-precipitate with silica and remove dissolved silica from solution (Lauchnor et al., 2013). 278
Concentrating the CSG brine in the multiple effect evaporators will further increase 279
the level of supersaturation of various mineral salts and possibly form scales on heat 280
exchanger tubes due to mineral precipitation. Although CaCO3 and SrCO3 are the two most 281
probable precipitates in RO brine, further concentration may produce other precipitates 282
including BaCO3, BaSO4 and silica polymers. 283
Interestingly, XRD analysis did not confirm the presence of CaCO3 or SrCO3, but 284
instead revealed a suite of complex crystalline candidate minerals, rich in calcium, 285
magnesium and carbonates as well as the other relevant elements. 286
Regardless of the base minerals present, it is likely that the particulates were rich in 287
silicates do to adsorption of silica (and other components) to seed material; amorphous 288
components are not revealed by XRD. For example trace amounts of aluminium based 289
minerals are common in RO brine, even in cases where the aluminium concentration is very 290
low (tens of ppb to several ppm), and it has been reported that alumino silicates form due to 291
the condensation reaction between hydrated aluminium ions (Al(OH)4-) and silicic acid 292
(Gallup, 1997; Iler, 1979). These aluminium silicate anions have the potential to precipitate in 293
the presence of counter ions, such as sodium, potassium, iron, boron, calcium, and 294
magnesium, and significantly, such precipitates can act as seed crystals for silica collection 295
due to silica polymerization on their surface (Malki & Abbas, 2013). The TEM image in 296
Fig.3B shows the core-shell structure of the particulates in brine. The particles were found to 297
be a coated with an amorphous thin layer, which can be due to the polymerization of silica on 298
mineral precipitates surface (Malki & Abbas, 2013). 299
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
13
Removal of particulate silica solids was tested by micro- and ultrafiltration 300
membranes. As can be seen from Fig.4, ultrafiltration performs significantly better than 301
microfiltration and removes nearly all the particulate silica from CSG brine. Nearly 30% of 302
the total silica in brine was removed in ultrafiltration (MWCO:5000); microfiltration (0.22, 303
0.45, 1 µm) was also able to remove about 12% of total silica. Most of the particulate solids 304
were nanoparticles (0-100 nm size range). While microfiltration was able to capture large 305
aggregated networks of solids, the nanoparticles could only be filtered using a low molecular 306
weight ultrafiltration membrane. 307
308
309
5. Conclusions 310
The physical and chemical characterization of reverse osmosis brine produced at a 311
CSG water treatment facility clearly showed the formation of scaling compounds in retentate 312
stream. Due to the enrichment of carbonates in CSG brine, metal carbonates were found to be 313
the key scalants. A significant fraction of total silica was also found to be in particulate form 314
in CSG brine. Colloidal particles with size range of 10 - 1000 nm were found in both 315
dispersed and aggregated form in CSG brine. The solubility analysis of various sparingly 316
soluble salts and silica suggests, instead of polymerization, silica might have been captured or 317
adsorbed on precipitated calcium carbonate and other mineral groups in CSG water during 318
concentration in the RO process. Downstream processing of this brine in concentrators might 319
face the problem of carbonate and silica scales. Filtration study suggests that colloidal 320
precipitates in CSG brine could be almost completely removed by ultrafiltration. 321
322
323
324
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
14
Acknowledgements 325
The authors acknowledge the support from Origin Energy for providing the CSG 326
water samples and water quality analysis data. The authors also like to thank Mr. Richard 327
Webb at Australian Institute for Bioengineering and Nanotechnology (AIBN) for helping 328
with the TEM micrographs. Funding from Australian Research Council (ARC-LP120100664) 329
is greatly appreciated. 330
References 331
Ali, M., Hamrouni, B., Bouguecha, S., & Dhahbi, M. (2004). Silica removal using ion-332
exchange resins. Desalination, 167(1-3), 273-279. 333
Antony, A., Low, J. H., Gray, S., Childress, A. E., Le-Clech, P., & Leslie, G. (2011). Scale 334
formation and control in high pressure membrane water treatment systems: A review. 335
Journal of Membrane Science, 383(1-2), 1-16. 336
Badruk, M., & Matsunaga, S. (2001). Experimental results of silica removal from simulated 337
solutions of geothermal brine of kizildere field, turkey. Geothermics, 30(5), 561-570. 338
Belton, D. J., Deschaume, O., & Perry, C. C. (2012). An overview of the fundamentals of the 339
chemistry of silica with relevance to biosilicification and technological advances. FEBS 340
Journal, 279(10), 1710-1720. 341
Bond, R., & Veerapaneni, S. (2007). Zero liquid discharge for inland desalination. Awwa 342
Research Foundation, 343
Butt, F. H., Rahman, F., & Baduruthamal, U. (1995). Identification of scale deposits through 344
membrane autopsy. Desalination, 101(3), 219-230. 345
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
15
Butt, F. H., Rahman, F., & Baduruthamal, U. (1997). Characterization of foulants by autopsy 346
of RO desalination membranes. Desalination, 114(1), 51-64. 347
Cipollina, A., Micale, G., & Rizzuti, L. (2011). A brine evaporative cooler/concentrator for 348
autonomous thermal desalination units. Desalination and Water Treatment, 31(1-3), 349
269-278. 350
Fournier, R. O. (1970). Silica in thermal waters: Laboratory and field investigations. 351
International Symposium on Hydrogeochemistry and Biochemistry, Tokyo. pp. 122-129. 352
Gallup, D. (1997). Aluminum silicate scale formation and inhibition: Scale characterization 353
and laboratory experiments. Geothermics, 26(4), 483-499. 354
Goosen, M. F. A., Sablani, S. S., Ai-Hinai, H., Ai-Obeidani, S., Al-Belushi, R., & Jackson, 355
D. (2004). Fouling of reverse osmosis and ultrafiltration membranes: A critical review. 356
Separation Science and Technology, 39(10), 2261-2297. 357
Greenlee, L. F., Lawler, D. F., Freeman, B. D., Marrot, B., & Moulin, P. (2009). Reverse 358
osmosis desalination: Water sources, technology, and today's challenges. Water 359
Research, 43(9), 2317-2348. 360
Hsu, H. -., Chen, S. -., Lin, C. -., & Chang, T. -. (2008). Silical pretreatement for RO 361
membrane by softening-adsorption. Journal of Environmental Engineering and 362
Management, 18(2), 99-103. 363
Iler, R. K. (1979). The chemistry of silica. Canada: Wiley. 364
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
16
Ji, X., Curcio, E., Al Obaidani, S., Di Profio, G., Fontananova, E., & Drioli, E. (2010). 365
Membrane distillation-crystallization of seawater reverse osmosis brines. Separation and 366
Purification Technology, 71(1), 76-82. 367
Klohn Crippen Berger. (2012). Forecasting coal seam gas water production in queensland's 368
surat and southern bowen basins (Technical ReportDepartment of Natural Resources 369
and Mines. 370
Lauchnor, E. G., Schultz, L. N., Bugni, S., Mitchell, A. C., Cunningham, A. B., & Gerlach, 371
R. (2013). Bacterially induced calcium carbonate precipitation and strontium 372
coprecipitation in a porous media flow system. Environmental Science & Technology, 373
47(3), 1557-1564. 374
Malki, M., & Abbas, V. (2013). Controlling aluminium silica formation in membrane 375
separation processes. The International Desalination Association (IDA) World Congress 376
on Desalination and Water Reuse, Tianjin, China. 377
Mariah, L., Buckley, C. A., Brouckaert, C. J., Curcio, E., Drioli, E., Jaganyi, D., et al. (2006). 378
Membrane distillation of concentrated brines - role of water activities in the evaluation 379
of driving force. Journal of Membrane Science, 280(1-2), 937-947. 380
Marshall, W., & Warakomski, J. (1980). Amorphous silica solubilities .2. effect of aqueous 381
salt-solutions at 25-degrees-C. Geochimica Et Cosmochimica Acta, 44(7), 915-&. 382
Mericq, J., Laborie, S., & Cabassud, C. (2010). Vacuum membrane distillation of seawater 383
reverse osmosis brines. Water Research, 44(18), 5260-5273. 384
Mi, B., & Elimelech, M. (2013). Silica scaling and scaling reversibility in forward osmosis. 385
Desalination, 312, 75-81. 386
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
17
Morillo, J., Usero, J., Rosado, D., El Bakouri, H., Riaza, A., & Bernaola, F. (2014). 387
Comparative study of brine management technologies for desalination plants. 388
Desalination, 336, 32-49. 389
Okamoto, G., Okura, T., & Goto, K. (1957). Properties of silica in water. Geochimica Et 390
Cosmochimica Acta, 12, 123-132. 391
Pandey, S. R., Jegatheesan, V., Baskaran, K., & Shu, L. (2012). Fouling in reverse osmosis 392
(RO) membrane in water recovery from secondary effluent: A review. Reviews in 393
Environmental Science and Bio-Technology, 11(2), 125-145. 394
Potts, D. E., Ahlert, R. C., & Wang, S. S. (1981). A critical-review of fouling of reverse-395
osmosis membranes. Desalination, 36(3), 235-264. 396
Qu, D., Wang, J., Fan, B., Luan, Z., & Hou, D. (2009). Study on concentrating primary 397
reverse osmosis retentate by direct contact membrane distillation. Desalination, 247(1-398
3), 540-550. 399
Rahardianto, A., Gao, J., Gabelich, C. J., Williams, M. D., & Cohen, Y. (2007). High 400
recovery membrane desalting of low-salinity brackish water: Integration of accelerated 401
precipitation softening with membrane RO. Journal of Membrane Science, 289(1-2), 402
123-137. 403
Sanciolo, P., Milne, N., Taylor, K., Mullet, M., & Gray, S. (2014). Silica scale mitigation for 404
high recovery reverse osmosis of groundwater for a mining process. Desalination, 340, 405
49-58. 406
Sheikholeslami, R., & Bright, J. (2002). Silica and metals removal by pretreatment to prevent 407
fouling of reverse osmosis membranes. Desalination, 143(3), 255-267. 408
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
18
Tissue, B. M. (2013). Basics of analytical chemistry and chemical equilibria. New York: 409
John Wiley. 410
Tomaszewska, B., & Bodzek, M. (2013). Desalination of geothermal waters using a hybrid 411
UF-RO process. part II: Membrane scaling after pilot-scale tests. Desalination, 319, 107-412
114. 413
Weres, O., & Tsao, L. (1981). Chemistry of silica in cerro prieto brines. Geothermics, 10(3-414
4), 255-276. 415
Zaman, M., Birkett, G., Stuart, B., & Pratt, S. (2013). Silica removal from coal seam gas 416
brine using activated alumina. CHEMECA:Australasian Conference on Chemical 417
Engineering, Brisbane, Australia. pp. 700-703. 418
Zeng, Y., Yang, C., Pu, W., & Zhang, X. (2007). Removal of silica from heavy oil 419
wastewater to be reused in a boiler by combining magnesium and zinc compounds with 420
coagulation. Desalination, 216(1-3), 147-159. 421
Zhang, M., Hou, D., She, Q., & Tang, C. Y. (2014). Gypsum scaling in pressure retarded 422
osmosis: Experiments, mechanisms and implications. Water Research, 48, 387-395. 423
424
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Table 1. Water quality data of feed and brine stream from reverse osmosis desalination of
coal seam gas water at a CSG water treatment plant. Standard deviations in brackets (n = 5
over 5 years).
Analyte (mmol/litre) RO feed RO Brine
Carbonate alkalinity as CO32- 3.34 (0.75) 6.02 (1.16)
Aluminium <0.001 <0.004
Barium 0.019 (0.003) 0.100 (0.012)
Boron 0.27 (0.02) 1.31 (0.13)
Calcium 0.21 (0.04) 1.10 (0.14)
Chloride 72.73 (7.56) 381.60 (49.59)
Magnesium 0.10 (0.02) 0.49 (0.08)
Phosphorous (mainly as HPO43-) 0.003 (0.002) 0.025 (0.015)
Potassium 0.37 (0.07) 1.85 (0.25)
Silica 0.75 (0.15) 3.69 (0.32)
Sulphur as Sulphate ND – 0.375 ND – 1.64
Sodium 109.04 (7.89) 561.44 (64.74)
Strontium 0.044 (0.004) 0.208 (0.021)
pH 9.125 (0.095) 8.842 (0.050)
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Table 2. Potential scaling information with respect to the major scale constituent in CSG
water at pH 8.8 and 80% water recovery. IAP determined using activity coefficients
estimated from Pitzer equations, considering ionic strength. Ksp for standard conditions.
Minerals Saturation
Index,
SIx(IAP/Ksp,x)
Ksp (25oC) Cation
Concentration
(mM)
Anion
Concentration
(mM)
CaCO3 93 2.8 x 10-9 1.10 6.02
SrCO3 890 1.1 x 10-10 0.208 6.02
BaSO4 22 1.1 x 10-10 0.100 1.64
BaCO3 170 5.1 x 10-9 0.100 6.02
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Table 3. Minerals in RO brine solids, identified by XRD in the 10-90° theta range
Peaks and 2-theta angles Peaks and d-spacing Candidate minerals
4) 31.7, 6) 45.4, 7) 57
4) 2.83, 6) 2.00, 7) 1.62 Halite: NaCl
1) 17
1) 5.3 Grandidierite:
(Mg,Fe)Al3(BO4)(SiO4)O
2) 27.4
2) 3.25 Florkeite:(K3Ca2Na)(Al8Si8O32).12H2O;
Aragonite (CaCO3)
3) 30.1, 4) 31.7
3) 2.98, 4) 2.83 Omongwaite: Na2Ca5(SO4)6.3H2O
5) 37.9, 3) 30.1, 6) 45.4
5) 2.36, 3) 2.98, 6) 2.00 Hibbingite: (Fe, Mg)2(OH)3Cl
8) 67, 9) 76, 10) 85 8) 1.40, 9) 1.33, 10)
1.31
Halite: minor peaks
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Figure 1. Simplified block flow diagram and sampling layout of a CSG water treatment
facility
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Figure 2. Silica concentrations at various sampling point in the CSG water treatment facility.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Figure 3. TEM micrographs of particulate solids in CSG brine (scale bar: 1µm): (A)
individual nanoparticles(scale bar: 1µm), (B) core-shell morphology of nanoparticles(scale
bar: 200 nm), (C) aggregated network structures(scale bar: 1µm)
(A) (B)
(C)
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Figure 4. . Silica removal using filtration
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Figure 5. Concentration difference in elements in CSG brine before and after ultrafiltration.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Figure 6. XRD scan of RO brine solids
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
• Silica can exist both in dissolved and particulate form in coal seam gas brine
• Mineral precipitation followed by silica polymerization can produce particulate scalant in RO
brine
• Mineral precipitation can be quite complex in multicomponent mixture and produce
minerals other than simple salts.
• Ultrafiltration can be used to remove a significant portion of scalant from the RO brine