Draft technical document guidelines for Canadian drinking water quality - Antimony: Analytical and treatment considerations

On this page

Analytical methods to detect antimony

Standardized methods

Standardized analytical methods available for the analysis of total antimony in drinking water and their respective method detection limits (MDLs) are summarized in Table 5. MDLs are dependent on the sample matrix, instrumentation and selected operating conditions and will vary between individual laboratories. These methods are subject to a variety of interferences, which are outlined in the respective references. The total concentration of antimony is determined from these methods and not the antimony species. A number of accredited laboratories in Canada were contacted to determine the MDLs and the method reporting limits (MRLs) for total antimony analysis and the MDLs were in the range of those reported in Table 5. The MRLs were between 0.5 and 1 μg/L by methods based on inductively coupled plasma – Mass Spectrometry (ICP-MS) (AGAT Laboratories, 2019a, b, c; Paracel Laboratories Ltd., 2019).

The MDLs or MRLs from provincial data are in the range of 0.001 to 2 μg/L (refer to the section on exposure).

Drinking water utilities should discuss sampling requirements with the accredited laboratory conducting the analysis to ensure that quality control procedures are met and that MRLs are low enough to ensure accurate monitoring at concentrations below the proposed MAC.

Table 5. Standardized analytical methods for the analysis of total antimony in drinking water
Method
(reference)
Methodology MDL (µg/L) Interferences/comments
U.S. EPA methods
EPA 200.5 Rev. 4.2 (U.S. EPA, 2003) Axially viewed inductively coupled plasma-atomic emission spectrometry (wavelength 249.68 nm) 0.9 Spectral, physical, chemical and memory interferences Matrix interferences: Ca, Mg and Na > 125 mg/L and Si > 250 mg/L
EPA 200.8, Rev. 5.4 (U.S. EPA, 1994a) ICP-MS 0.4Footnote a
0.04Footnote b
Isobaric elemental and polyatomic ion and physical interferences Matrix interferences: TDS > 0.2 % (w/v)
EPA 200.9, Rev 2.2 (U.S. EPA, 1994b) Stabilized temperature graphite furnace atomic absorption 0.8 Spectral, matrix and memory interferences The HCl present from the digestion procedure can influence the sensitivity Interference by K2SO4 can be reduced by using H/Ar in char step
EPA 6020B (U.S. EPA, 2014) ICP-MS 0.1 (IDL) Isobaric elemental and molecular and memory interferences Matrix interferences: TDS > 0.2 % Sb concentrations of 50–500μg/L require 1% (v/v) HCl for stability
APHA standard methods (SM)
SM 3113B (APHA et al., 2017) Electrothermal atomic absorption spectrometry 0.8 (estimated detection level) Sb not recovered unless HCl used in digestion Molecular absorption, chemical and matrix interferences
SM 3125 (APHA et al., 2017) ICP-MS 0.07 (IDL) Samples should not contain more than 0.5 % dissolved solids Isobaric, abundance sensitivity, polyatomic ion, physical, memory and ionization interferences
ASTM methods
D5673-16 (ASTM, 2016) ICP-MS 0.08 (IDL) Abundance sensitivity and isobaric elemental, isobaric polyatomic ion, physical and memory interferences

Footnotes

Footnote 1

MDL in scanning mode

Return to footnote a referrer

Footnote 2

MDL in selective ion monitoring mode

Return to footnote b referrer

ICP-MS – inductively coupled plasma mass spectrometry; IDL – instrument detection level; MDL – method detection limit; TDS – total dissolved solids

Sample preparation

Total antimony includes both the dissolved and particulate (suspended) fractions of antimony in a water sample and is analyzed using methods for total recoverable antimony. Analysis of total antimony is needed for comparison to the proposed MAC.

Sample processing considerations for the analysis of antimony in drinking water (that is, sample preservation, storage, digestion and so on) can be found in the references listed in Table 5. Accurate quantification of dissolved, particulate and total antimony in samples is dependent on proper sample preservation and processing steps. SM 3030B and SM 3030D provide guidance on filtration, preservation (acidification) and digestion procedures for the determination of dissolved or particulate metals (APHA et al., 2017). SM 3030D provides guidance for metal digestion, including antimony and the necessity of using hydrochloric acid along with nitric acid for proper digestion (APHA et al., 2017). In order to determine dissolved antimony concentrations, samples should be filtered at the time of collection (not at the laboratory) and the filtrate should be acidified to pH < 2 with concentrated nitric acid.

Currently, EPA methods 200.5, 200.8, 200.9 and SM 3113B do not require hot acid digestion for total recoverable metals, unless turbidity of the sample is greater than 1 nephelometric turbidity unit (NTU). However, research conducted on other metals (for example, lead, chromium, cadmium) has found that this may not accurately quantify the total metal concentration in a drinking water sample for all metals. This approach may underestimate total antimony in drinking water when the particulate form of antimony is present and hot acid digestion may be necessary. Hot acid digestion is described in EPA methods 200.5, 200.8 and 200.9 (U.S. EPA, 2003, 1994a, b). Microwave-assisted digestion, outlined in method SM 3030 K (APHA et al., 2017), can also be used for analysis of total recoverable metals for methods that are based on ICP-MS.

Treatment considerations

Treatment technologies that are available to decrease antimony concentrations in drinking water include:

The effectiveness of these technologies varies depending on water quality parameters such as the species of antimony, the pH and presence of competing ions. At the residential scale, certified treatment devices relying on RO and distillation are expected to be effective for removal of antimony (U.S. EPA, 1998). Operational complexity (for example, pre- and post-treatment pH adjustment and alkalinity adjustment) may need to be considered in the selection of treatment options, particularly for small systems.

The antimony species present in water entering a treatment plant is an important factor in determining the effectiveness of treatment. Smaller, neutral species are generally more difficult to remove than larger, charged species. The two main species of antimony present in natural waters are antimonite (Sb(III)) in the form of Sb(OH)3 and antimonate (Sb(V)) an anion in the form of Sb(OH)6- (U.S. EPA, 2006; Deng et al., 2017). Typically, Sb(V) is the form present in oxic surface water and Sb(III) is the form in anoxic groundwater (He et al., 2015).

The redox chemistry of antimony is important in the treatment and removal of antimony from drinking water. The treatment type for antimony determines whether Sb(III) or Sb(V) is better removed. Conventional treatment bench-scale studies indicate that Sb(III) is better removed than Sb(V) whereas removal through RO exhibited better removal of Sb(V).

Municipal-scale

The selection of an appropriate treatment process will depend on many factors, including the raw water source and its characteristics, the species of antimony present in the water, the operational conditions of the selected treatment method and the water utility's treatment goals. Treatment goals may require that pH be adjusted post-treatment to address corrosion issues in the distribution system (Health Canada, 2015). Pilot- and bench-scale testing is recommended to ensure the source water can be successfully treated and to optimize operating conditions.

Conventional coagulation

Conventional coagulation was evaluated for antimony removal through numerous bench-scale studies, with some using natural waters and others deionized waters (refer to Tables 6, 7 and 8). The first study listed in Table 6 evaluated two source waters, collected in and distributed through two historic mine tunnels (Spiro and Judge tunnels) in Park City, Utah, with high total antimony levels as well as other co-occurring contaminants (CH2M, 2016; Najm et al., 2017). Ferric chloride (FC) was used in jar tests to determine the impact of coagulant dose and pH (see Table 6). Overall, the results showed that the removal of total antimony was:

Table 6. Bench-scale removal of antimony using coagulation on two natural watersFootnote a (Najm et al., 2017)
Water Influent Sb (μg/L) pH Sb effluent (μg/L)Footnote b Water quality Co-occurring contaminants
FeCl3 dose (mg/L)
5 20 40
Spiro Tunnel 9.3 5.5 3 4 2.9

pH 7.43

Alkalinity = 133 mg/L as CaCO3

Total hardness = 480 mg/L as CaCO3

As = 41 μg/L

Cd = 0.19 μg/L

Tl = 3.5 μg/L

Zn = 140 μg/L

Fe = 300 μg/L

6.5 6.8 5.8 4.7
7.5 7.6 7 5.3
Judge Tunnel 6.1 5.5 4.5 3 1.6

pH 7.77

Alkalinity 89 mg/L as CaCO3

Total hardness 166 mg/L as CaCO3

As = 8.2 μg/L

Cd = 2.3 μg/L

Tl = 0.03 μg/L

Zn = 770 μg/L

Fe = 340 μg/L

6.5 5.1 4.5 3.1
7.5 5.1 4.7 3.4

Footnotes

Footnote 1

Jar tests, shaken for 13 days to reach equilibrium, 2 mg/L Cl2

Return to footnote a referrer

Footnote 2

Estimated from graph

Return to footnote b referrer

As – arsenic; Cd – cadmium; Fe – iron; Tl: thallium; Zn – zinc

Other bench-scale studies also evaluated Sb(V) and Sb(III) removal using coagulation and these showed that (refer to Tables 7 and 8):

Table 7. Antimony removal using coagulationFootnote a
Influent (μg/L) % RemovalFootnote b Coagulant type Coagulant dose pH Water References
Aluminum-based coagulants
Sb(V) = 6 10% Polyaluminum chloride 5.4 mg/L 5.1

Reservoir water

Turbidity 17 NTU

DOC 3.9 mg/L

Kang et al. (2003)
Sb(III) = 6 (Sb2O3) 40% 5.3 mg/L 5.1

Stream water

Turbidity 4.5 NTU

DOC 3.3 mg/L

Sb(III) = 4 (SbCl3)

20%

Sb(V) = 50 < 20% Aluminum sulphate 1 - 3 x 10-4 mol/L 3.5–9.8

Spiked deionized water

Alkalinity 4.0 x 10-3 mol/L of NaHCO3

T = 25 ± 1˚C

Guo et al. (2009)
Sb(III) = 50 < 25% 1 - 3 x 10-4 mol/L
Iron-based coagulants
Sb(V) = 6 65% Ferric chloride 10.3 mg/L 5

Reservoir water

Turbidity 17 NTU

DOC 3.9 mg/L

Kang et al. (2003)
90% 20.0 mg/L
Sb(III) = 6 (Sb2O3) 90% 10.3 - 20.0 mg/L 5.1

Stream water

Turbidity 4.5 NTU

DOC 3.3 mg/L

Sb(III) = 4 (SbCl3) > 95%
Sb(V) = 250 > 99% Polymeric ferric sulfate 8 x 10-4 mol/L 4

Deionized water

with 4.0x10-3 mol/L NaHCO3 added

Guo et al. (2018)

78% 6
< 55% > 8.5
95% 4 x 10-4 mol/L 5
50% 7
< 45% > 8
Sb(III) = 100 85% 4 x 10-5 mol/L 4
> 95% > 5.5
70% 2 x 10-5 mol/L 4.5
80% 6
> 90% > 8

Footnotes

Footnote 1

Bench-scale jar tests with rapid mix, slow mix and settling stages

Return to footnote a referrer

Footnote 2

Estimated from graph

Return to footnote b referrer

DOC – dissolved organic carbon

Table 8. Antimony removal using ferric-based coagulants at the bench-scaleFootnote a (Guo et al., 2009)
FC dose (mol/L)

Initial Sb(V) (μg/L)

Initial Sb(III) (μg/L)

Sb(V) = 49.2 Sb(V) = 98.4 Sb(V) = 492 Sb(III) = 50.6 Sb(III) = 101 Sb(III) = 506
pH 6.0 ± 0.2 Treated Sb(V) concentration (μg/L) Treated Sb(III) concentration (μg/L)
2 x 10-4 22.1 47.7 241 6.6 13.8 35.2
6 x 10-4 0.7 10.8 50.5 1.5 4.6 13.2
10 x 10-4 Undetectable 2.9 8.2 0.7 3.5 6.4
pH 7.8 ± 0.2 Treated Sb(V) concentration (μg/L) Treated Sb(III) concentration (μg/L)
2 x 10-4 38.3 73.2 341 11.3 13.5 36.1
6 x 10-4 20.8 31.8 106 3.9 9.8 21.5
10 x 10-4 4.7 25.1 60.0 3.5 7.1 6.5

Footnotes

Footnote 1

Spiked deionized water, with 4.0 x 10-3 mol/L NaHCO3; Jar tests; Rapid mix: 140 rpm 3 min; Slow mix: 40 rpm 20 min; Settling 30 min; T = 25 ± 1˚C

Return to footnote a referrer

FC – ferric chloride

A bench-scale study used to evaluate ferric-based coagulants found that the addition of Fe(III) had better Sb(III) removal than that using Fe(II), in artificially contaminated tap water (Mitrakas et al., 2018).

Sb(III) was shown to be better removed than Sb(V) when using FC-based coagulants, and pre-oxidation may have a negative impact on overall removal since this oxidation step converts the Sb(III) to the Sb(V) form. The impact of pre-chlorination (residual free chlorine of 0.5 mg/L at 10 minutes) was examined in a bench-scale study on antimony removal using FC (dose = 10.3 mg as Fe/L) (Kang et al., 2003). When Sb(III) was present, pre-chlorination resulted in declined removal over the entire range of pH values (pH 5 to 10).

Kang et al., (2003) discussed antimony removals compared to those of arsenic. For arsenic, As(V) is better removed than As(III), which differs from antimony, in which Sb(III) is better removed than Sb(V). The authors stated that for Sb(V) removal, the required FC dose at a pH of 5 is about 9 times higher than that for removal of As(V).

The presence of competing ions had variable effects on removal and was a function of the antimony species, pH, coagulant dose and the concentration of competing ion. A bench-scale study evaluating effects of silicate, bicarbonate (HCO3-), sulfate (SO42-), phosphate (PO43-) and humic acid (HA) on both Sb(V) and Sb(III) removal was conducted by Wu et al., (2010) at various FC doses and pH (refer to Table 9). Silicate was found to have minimal effect on antimony removal. HCO3-, SO42-, HA and PO43- presence had a negative impact on Sb(V) removal. The effect of HA and PO43- on Sb(III) removal was variable depending on pH and coagulant dose. The authors also conducted experiments using a synthetic water containing various anions and cations and indicated that Sb(V) removal was impacted to a larger extent than Sb(III). Other bench-scale studies were conducted that showed similar impacts as a result of competing ions (Guo et al., 2009, 2018).

Table 9. Antimony removal using ferric-based coagulation in presence of competing ions at bench-scaleFootnote a (Wu et al., 2010)
Competing ion Competing ion conc. pH Initial Sb(V) = 100 μg/L Initial Sb(III) = 100 μg/L
FC dose (mol/L) FC dose (mol/L)
2 x 10-4 4 x 10-4 8 x 10-4 2 x 10-4 4 x 10-4
HCO3- 0 7.0 80% 87% N/A 90%Footnote b 90%Footnote b
4 x 10-3 mol/L 37% 63% N/A 89%Footnote b 92%Footnote b
SO42- 0 7.0 78% 93% N/A 78%Footnote b 92%Footnote b
200 mg/L 60% N/A N/A 60%Footnote b 90%Footnote b
300 mg/L N/A 84% N/A 65%Footnote b 85%Footnote b
PO43- 0 6.0 93% 96% N/A 93% 95%Footnote b
1 mg/L 82% 90% N/A 75% 90%Footnote b
0 7.5 92% 96% N/A 95% 96%Footnote b
1 mg/L 63% 90% N/A 84% 88%Footnote b
HA 0 6.0 90%Footnote b 95%Footnote b N/A 95%Footnote b 95%Footnote b
4 mg C/L 85%Footnote b 95%Footnote b N/A 75%Footnote b 95%Footnote b
0 7.5 90%Footnote b 80%Footnote b N/A 95% 97%
4 mg C/L 58% 80% N/A 56% 81%
Synthetic See footnote Footnote c Neutral 27% 54% 72% 66% ~ 90%Footnote b

Footnotes

Footnote 1

Jar tests; Ferric chloride; Deionized water; T = 25 ± 1˚C; Rapid mix: 140 rpm 3 min; Slow mix: 40 rpm 20 min; Settling 30 min

Return to footnote a referrer

Footnote 2

Estimated from graph

Return to footnote b referrer

Footnote 3

Synthetic water: HCO3- = 4 mmol/L; SO42- = 100 mg/L; SiO32- = 5 mg/L; PO43- = 0.5 mg/L; Ca2+ = 100 mg/L; Mg2+ = 40 mg/L; HA = 2 mg C/L

Return to footnote c referrer

FC – ferric chloride; HA – humic acid; N/A – not available

Inam et al. (2019) investigated the impact of various natural organic matter (NOM) concentrations on antimony removal through jar tests. Different NOM were investigated to determine the impact on the optimal FC dose in removing Sb(III) and Sb(V). The NOM types that were investigated included hydrophilic salicylic acid and L-cysteine and hydrophobic HA. The tests included rapid mixing, slow mixing and settling stages and were conducted with either Sb(III) or Sb(V) at 1 mg/L in deionized water, with NOM at a concentration of 10 mg/L. The optimal FC dose was found in the presence of each type of NOM for both Sb(III) and Sb(V), and in all cases was higher for Sb(V) than for Sb(III) with greater than 90% removed. For both Sb(III) and Sb(V), the presence of HA resulted in the highest FC optimal dose over the other types of NOM. The authors stated that the hydrophobic molecules in HA suppressed the antimony adsorption onto the FC.

Adsorption

Adsorption can be used to remove contaminants and effectiveness of this treatment technology depends on the contaminant, the sorbent material, water quality parameters and presence of competing ions (U.S. EPA, 1998). Other ions present in the water may compete with antimony for adsorption sites and impact the number of bed volumes (BVs) to breakthrough and the regeneration or replacement frequency.

The town of Alta in Utah uses the Bay-City Tunnel (an abandoned mine tunnel) to collect and store percolating groundwater which is then used as the source for drinking water (Najm et al., 2010). The tunnel water contains total antimony concentrations between 1.2 to 29 μg/L (average of 13 μg/L), along with many other co-occurring contaminants such as arsenic and cadmium. Pilot-scale testing was conducted using a titanium-based media, operating for 11 hrs/day. At a natural pH of 7.3, antimony reached breakthrough concentration of 6 μg/L between 50,000 and 75,000 BVs. When the pH was lowered to 6.5, the performance improved to approximately 100,000 BVs to reach breakthrough. The full-scale system for this site consisted of 2 columns in a lead-lag configuration, with effluent remaining below 1 μg/L during the evaluation period of 30,000 BVs for the lead column and 15,000 BVs for both columns. Comparison of the pilot- and full-scale results during the first 30,000 BVs show better performance at the full-scale with < 1 μg/L of total antimony in the treated water compared to the pilot-scale with approximately 2 μg/L. Finally, fully treated water was compared against a blend of treated and untreated water (at a ratio of 2:1). The results showed that the treated water antimony concentration remained close to 0 μg/L for > 390 days of operation and that the blended water concentrations were between 3 and 6 μg/L over this same time frame.

The Spiro and Judge Tunnel waters in Park City Utah presented in the section on conventional coagulation were also evaluated for removal of antimony, using various adsorbents, through bench- and pilot-scale studies (CH2M, 2016; Najm et al., 2017; Swaim et al., 2017). Antimony and other co-occurring contaminants had to be removed from the mine tunnel waters. The bench-scale study evaluated a titanium dioxide (TiO2) based adsorbent at pH 8, achieving antimony removal to less than 1 μg/L at a dose of 60 mg/L (Najm et al., 2017). The pilot-scale study was carried out in advance of full-scale implementation to ensure adequate removals of antimony, arsenic, cadmium (Cd), iron (Fe), manganese (Mn), thallium (Tl), zinc (Zn), selenium (Se) and lead (Pb). The treatment train consisted of pre-oxidation, high pH coagulation/settling, MnO2 filtration, adsorption, disinfection and conditioning (CH2M, 2016; Swaim et al., 2017). All of the co-occurring contaminants were removed to satisfactory levels after the MnO2 filtration stage, with the exception of antimony. The removal of antimony was examined through evaluation of 3 different adsorption media including TiO2, ferric oxide and ferric hydroxide. It was estimated that ferric oxide and ferric hydroxide at pH 6.5 would need to be changed every 12,000 to 18,000 BVs and 29,000 to 50,000 BVs, respectively. For the TiO2 adsorbent, it was expected to have approximately 170,000 BVs between media changes at pH 6.5 and 66,000 to 91,000 BVs at pH 7.6. The TiO2 column running at a pH of 7.6 was lowered to 6.5, resulting in an increased antimony removal. When the pH was returned to 7.6, the antimony concentrations returned to previous values. Overall, this pilot-scale study showed that (CH2M, 2016):

A full-scale study carried out as part of the United States Environmental Protection Agency's (U.S. EPA) Arsenic Treatment Technology Demonstration program included investigation of antimony. The study included 3 parallel columns that evaluated adsorption of arsenic and antimony using commercially available granular ferric hydroxide (β-FeOOH + Fe(OH)3) (Cumming et al., 2009). The study ran over approximately 5 and a half months and treated 15,753,000 gallons of groundwater. Breakthrough of antimony at 6 μg/L occurred at 3,000 BVs, which the authors stated was unexpectedly short and may have been due to presence of silica and phosphorus in the source water.

Other pilot- and bench-scale studies provided a range of adsorption capacities for various iron-based adsorbents to remove antimony (refer to Table 10). Overall, these studies showed that:

Table 10. Range of adsorption capacities for iron-based adsorbents
Adsorbent Antimony species Adsorption capacity (mg/g) References
Iron-based Not specified 0.42–1.63 Ilavský et al., (2014, 2015); Sazakli et al., (2015); Barloková et al., (2017)
Sb(III) 19.4–53.5 Xi et al., (2013); Guo et al., (2014); Qi and Pichler (2016)
Sb(V) 6.5–62.5 Guo et al., (2014); Miao et al., (2014); Qi and Pichler (2016)
In-situ FeOxHy Sb(III) 12.77 mmol/g
(1 555 mg/g)
He et al., (2015)
Sb(V) 10.21 mmol/g
(1 243 mg/g)

In-situ FeOxHy – freshly prepared ferric hydroxide

Three pilot-scale studies evaluated the removal of antimony from spring water and the use of commercially available iron-based adsorbents by determining the number of BVs to reach 5 μg/L (refer to Table 11) (Ilavský et al., 2014, 2015; Barloková et al., 2017). Barloková et al., (2017) determined that the number of BVs increased with increased column height. Ilavský et al., (2014, 2015) indicated that CeO2·nH2O was the better adsorbent over 90.1% α-FeOOH, FeOOH and β-FeOOH + Fe(OH)3. In these 3 studies, the authors stated iron-based adsorbents could possibly decrease antimony to drinking water levels.

Table 11. Pilot-scale tests evaluating number of bed volumes to reach breakthrough at 5 μg/L
Adsorbent type Influent (μg/L) EBCT (min) BV HeightFootnote a (cm) Filtration rate (m/h) Description References
β-FeOOH + Fe(OH)3 90.3 (ave) 5.5 1 537 50 5.3–5.4 (Average) Spring water (high antimony due to mining) Barloková et al. (2017)
7.7 3 736 70
10.1 4 659 100
27.7 ± 3.41 5.29 3 236 48–49 5.55–5.58 (average) Spring water pH 8.2; TDS 190 mg/L Ilavský et al. (2015)
CeO2·nH2O 5.16 3 967
β-FeOOH + Fe(OH)3 58.3 (ave) 6.0 1 700 50–52 4.7–5.3 Spring water Ilavský et al. (2014)
90.1% α-FeOOH 6.4 715 4.3–4.9
FeOOH 6.3 790 4.3–5.1

Footnotes

Footnote 1

Column diameter 5.0 cm

Return to footnote a referrer

BV – bed volumes; EBCT – empty bed contact time; TDS: total dissolved solids

He et al. (2015) examined the removal of antimony in deionized water with potassium nitrate solute at 0.01 M using in-situ FeOxHy at the bench-scale. The maximum adsorption capacity of in-situ FeOxHy was shown to be 6.68 mmol/g and 5.34 mmol/g for Sb(III) and Sb(V), respectively. The authors then conducted a pilot-scale study to examine the removal of Sb(V) over 716 hours from 2 columns in series. The influent concentration ranged between 20 and 30 μg/L with the pH gradually being decreased in a step-wise manner from 7.6 to 5.2. The effluent Sb(V) concentration from each column was found to increase slowly over time and then decrease corresponding to the step-wise decline in pH. The authors stated that the pH adjustment was valuable to increase the adsorption capacity and decrease the regeneration frequency.

Membrane filtration

Studies investigating antimony removal by RO illustrate good antimony removal (refer to Table 12). As part of the U.S. EPA's Arsenic Treatment Technology Demonstration program, a point-of-entry RO system was evaluated at a school (Wang et al., 2011). Antimony removal was examined in addition to arsenic removal and had 99% rejection over the 8-month study period.

A bench-scale study examined the removal of Sb(III) and Sb(V) by RO and showed that Sb(V) was better removed than Sb(III) and was less impacted by pH (Kang et al., 2000). The removals of Sb(III) and Sb(V) are almost constant over the pH range of 3 to 10 with both the polyamide and polyvinyl alcohol membranes. The exception was for the removal of Sb(III) with the polyvinyl alcohol membrane, where it decreased sharply from 60.2% to 45.7% when pH changed from 7 to 10.

Table 12. Antimony removal using RO membrane filtration processes
Influent (μg/L) Rejection (%) RO Membrane pH Process description Reference
8.6–13.2 99% 15 membrane modules

7.9 – initial

6.9 – after RO

7.4 – after calcite filter

Full-scale

Point-of-entry RO at a school

Two 2.5-in x 40-in thin-film

Groundwater

recovery rate of 40%

April to December 2009

Wang et al. (2011)
Sb(III) = 10 85%a Polyamide 3–10

Bench-scale

Distilled water

Two RO membranes (flat-sheet type)

Kang et al. (2000)

58%Footnote a Polyvinyl alcohol 3
60.2%Footnote a 7
45.7%Footnote a 10
Sb(V) = 10 95%a Polyamide 3–10
90%a Polyvinyl alcohol 3–10

Footnotes

Footnote 1

Estimated from graph

Return to footnote a referrer

RO – reverse osmosis

Limitations of the RO process include possible membrane scaling, fouling and failure, as well as higher energy use and capital costs. Calcium, barium and silica can cause scaling and decrease membrane efficiency. Since RO completely removes alkalinity in water, it will continually lower product water pH and increase its corrosivity. Therefore, the product water pH must be adjusted and alkalinity may need to be increased to avoid corrosion issues in the distribution system such as the leaching of lead and copper (Schock and Lytle, 2011; U.S. EPA, 2012).

Combined technology

Studies evaluating combined treatment processes (coagulation followed by an ultrafiltration membrane) were reviewed. Two bench-scale studies examined antimony removal using freshly prepared hydrolytic flocs (Du et al., 2014; Ma et al., 2017a, b) (refer to Table 13). The studies that examined Sb(V) (Ma et al., 2017a, b) showed that:

The study that examined Sb(III) (Du et al., 2014) showed that:

Given that only bench-scale studies are available, it is critical to conduct bench- and pilot-scale studies prior to full-scale implementation to ensure the effectiveness of combined treatment and its optimal operating conditions.

Table 13. Coagulation and flocculation followed by ultrafiltration for Sb(V) removal at bench-scale
Influent (μg/L) Coagulant type Coagulant dose Removal Effluent (μg/L)

TMP

(kPa)

pH Conditions Process description
Sb(V) removal (Ma et al., 2017a, b)
5.4–19.8 None 0 10.7% N/A 74.6 (7 days) 7.5

T = 27.2–29.8˚C;

Turbidity =

0.3–1.0 NTU;

Residual chlorine =

0.4–1.0 mg/L Aeration rate 0.1 L/min

Samples taken after 10 days

Surface water

Hydrolytic flocs freshly prepared followed by ultrafiltration

Flocs injected into membrane tank

Hollow-fiber membrane module (polyvinylidene fluoride)

1 L/h flow rate; 30 min filtration and 1 min backwashing; HRT = 2.2 hour

AlCl3·H2O

10 mM/inj

(1 inj every 2 days)

27.2% N/A 61.3 7.5
FeCl3·H2O 92.8% N/A 3.7 6
40.7% N/A 17.9 7.5
15.9% N/A 22.3 9
50 mM over 10 days (Continuous) 53.6% N/A 31.2 7.5

5 mM/inj

(1 inj/day)

45%Footnote a N/A 26.8
4.1–13.7

10 mg/L

(5 day sludge discharge frequency)

N/A 0.32 36.1 7.3-7.6

T = 8.9 - 28.5˚C;

Turbidity = 0.2–2.1 NTU; Residual chlorine =

0.4–1.0 mg/L

Samples taken after 100 days

10 mg/L

(10 day sludge discharge frequency)

N/A 0.26 33.7
Sb(III) removal (Du et al., 2014)

62.5 ±

0.1

Ferric coag. 0 9% N/A Not given 6

Turbidity 0.65–0.89 NTU;

TOC 1.60–1.84 mg/L

T = 28 ± 1˚C

River water:

Coagulation 300 to 400 rpm

Flocculation 50 to 200 rpm

Hollow-fibre membrane with an effective surface area 0.03 m2; flux 30 L/m2·h

12% N/A 8
0.2 mM 76% N/A 6
66% N/A 8
0.4 mM 96% N/A 6
94% N/A 8
> 0.6 mM 96% N/A 6 and 8
100 ± 3.6 0.4 mM 97% N/A 5–8
94% N/A 9
30, 54, 89, 124 and 158 0 5%–7.5%Footnote a N/A

8.5 ±

0.2

0.4 mM > 95%Footnote a N/A

Footnotes

Footnote 1

Estimated from graph

Return to footnote a referrer

HRT – hydraulic retention time; inj – injection; N/A – not available; TMP – transmembrane pressure; TOC – total organic carbon

Residential-scale

In cases where antimony removal is desired at the household level, for example, when a household obtains its drinking water from a private well, a residential drinking water treatment unit may be an option for decreasing antimony concentrations in drinking water. Before a treatment unit is installed, the water should be tested to determine the general water chemistry and antimony concentration in the source water. Periodic testing by an accredited laboratory should be conducted on both the water entering the treatment unit and the treated water, to verify that the treatment unit is effective. Units can lose removal capacity through use and time and need to be maintained and/or replaced. Consumers should verify the expected longevity of the components in the treatment unit according to the manufacturer's recommendations and service it when required. Systems classified as residential scale may have a rated capacity to treat volumes greater than that needed for a single residence, and thus, may also be used in small systems.

Health Canada does not recommend specific brands of drinking water treatment units, but it strongly recommends that consumers use units that have been certified by an accredited certification body as meeting the appropriate NSF International Standard/American National Standards Institute (NSF/ANSI) for drinking water treatment units. The purpose of these standards is to establish minimum requirements for the materials, design and construction of drinking water treatment units that can be tested by a third party. This ensures that materials in the unit do no leach contaminants into the drinking water (that is, material safety). In addition, the standards include performance requirements that specify the removal that must be achieved for specific contaminants (for example, reduction claim) that may be present in water supplies.

Certification organizations (that is, third party) provide assurance that a product conforms to applicable standards and must be accredited by the Standards Council of Canada (SCC). Accredited organizations in Canada (SCC, 2020) include:

An up-to-date list of accredited certification organizations can be obtained from the SCC.

The drinking water treatment units that are expected to be effective for antimony removal at the residential-scale include RO and distillation (U.S. EPA, 1998). Currently, antimony is not included in the performance requirements (e.g., reduction claims) of NSF/ANSI standards. However, use of a treatment unit that is certified to the standards for RO or distillation will ensure that the material safety of the units has been tested. These standards are NSF/ANSI Standard 58 (Reverse Osmosis Drinking Water Treatment Systems) (NSF International, 2021a) and NSF/ANSI Standard 62 (Drinking Water Distillation Systems) (NSF International, 2021b).

The effectiveness of RO units for antimony removal is dependent on the membrane (filter) type and pH of the water and anticipated removals range from 46% to 99% (based on municipal-scale data). Therefore, an RO system will need to be carefully selected in order to achieve treated water concentrations below the proposed MAC. In addition, it may be necessary to pre-treat the water to reduce fouling and extend the service life of the RO membrane. Although there is a lack of data regarding the use of distillation for removal of antimony from drinking water, it is expected to perform adequately because it is effective for the reduction of other inorganic contaminants. However, this process requires a high electrical energy input. Consumers may want to consult a water treatment professional for advice on available treatment systems, as well as installation and maintenance costs, based on their specific water quality.

Water that has been treated using RO and distillation may be corrosive to internal plumbing components. Also, as large quantities of influent water are needed to obtain the required volume of treated water, these devices are generally not practical for point-of-entry installation. Therefore, these devices should be installed only at the point-of-use.

A consideration for limiting exposure to antimony is to specify that drinking water materials (that is, components and treatment chemicals) meet health-based standards. These standards ensure that materials meet health-based requirements and are safe for use in potable water applications. NSF/ANSI Standards 61 (NSF International, 2021c) and 60 (NSF International, 2021d) require that the concentration of antimony not exceed the single product allowable concentration of 0.0006 mg/L in components and treatment chemicals, respectively.

Distribution system considerations

Treated water, accumulation in distribution system and leaching from brass and solder materials are potential sources of antimony in both distribution and household plumbing systems.

Antimony deposition and accumulation

The accumulation of trace inorganic contaminants in the drinking water distribution system is a complex function of numerous factors, including contaminant concentration in treated water, pH, redox conditions and pipe material. Iron oxyhydroxides and hydrous manganese oxides are significant sinks for trace inorganic contaminant accumulation because of their adsorptive affinity for them. Water quality changes or physical disruptions in the distribution system can remobilize contaminants into the bulk water. Indicators of this include the presence of discoloured water or increased turbidity.

Total and dissolved antimony during water flushing trials of distribution systems were examined (Friedman et al., 2016). The results consistently showed that particulate antimony concentrations were higher in flushed samples than bulk water. The study indicates a lack in adsorption and release mechanisms of antimony within the distribution as compared to clear mechanisms for arsenic to iron scale and lead to manganese scale.

Friedman et al. (2010) identified several key water quality conditions that should be controlled in order to maintain water stability for deposited trace inorganic contaminants. These include pH, the oxidation-reduction potential and the corrosion-control measures. It is also important to avoid the uncontrolled blending of surface water with groundwater and of chlorinated water with chloraminated water. Indicators of potential release of trace inorganic contaminants in the distribution system may include discoloured water and increased turbidity.

In a study of scale and sediment samples collected from the distribution systems of 20 U.S. drinking water utilities supplied by groundwater, surface water and blended water sources, antimony was found to be the tenth most concentrated of the 12 inorganics analyzed (Friedman et al., 2010). The authors reported that antimony was found in all solids but that its concentration was significantly lower than other metals. The median antimony concentration of all scale deposits and sediment samples combined was 0.14 μg/g (1.4 x 10-5 weight %), with 10th and 90th percentiles of 0.05 μg/g (5 x 10-6 weight %) and 0.86 μg/g (8.6 x 10-5 weight %), respectively. The median antimony concentrations in scale deposits and hydrant-flush solids were 0.13 μg/g and 0.17 μg/g (1.3 x 10-5 weight % and 1.7 x 10-5 weight %), respectively. It was noted that of the six samples with higher antimony content (> 0.9 μg/g) there were no obvious commonalities to other co-occurring contaminants. The authors also reported an estimated antimony mass of 0.1 lb (0.05 kg) accumulated on a 100-mile pipe length (160 km) (based on a 12-in. diameter pipe [30.5 cm]). Theoretically, 60% to 85% of the scale deposit would need to be released to exceed the U.S. EPA drinking water standard for antimony of 0.006 mg/L. Based on these results, the accumulation of antimony (and its potential release) in distribution systems is not considered significant relative to other inorganic contaminants.

Scale from lead service lines (N = 5) ranged from 2.54 mg/kg to 100 mg/kg and from an iron service line (N = 1) was 0.78 mg/kg (Schock, 2005). Scale from 23 lead pipes from older installations (dating from 1880 to 1947, and 2 unknown dates) had a range of 11.3 mg/kg to 292 mg/kg (Kim and Herrara, 2010). Friedman et al. (2016) analyzed 13 opportunity samples (those that became available including those from water meters, pipes and filters) with a range of 0 to 168 mg/kg. Antimony accumulation is thought to be due to surface adsorption and co-precipitation reaction with soluble Sb(OH)6- which is in anionic form at the pH typically found in the distribution system (Friedman et al., 2010).

Clement and Carlson (2004) investigated the distribution system of the Park City, Utah drinking water system discussed previously (refer to the sections on conventional coagulation and adsorption). At the time of this study, the drinking water system was supplied by several sources, including one of the historic mine tunnels (not specified). The source water had an antimony concentration of 0.0066 mg/L, mainly in soluble form and had no physical treatment but was chlorinated prior to entering the distribution system. Samples from the distribution system were taken during a flushing exercise. It was found that the antimony concentration continued to rise throughout the system and reached 0.027 mg/L after 10 minutes when sampling was discontinued and maximum was not reached. The flushing concentration was 4 times greater than the source water concentration. Scanlan (2003) analyzed sediment samples from one reservoir from this same water system and had an average antimony content of 48 mg/kg. Another study examined 45 homes using random daytime sampling, with presence of lead in at least part of the service line. A total of 135 samples were collected and analyzed for antimony with a detection limit of 0.03 μg/L and results showed a maximum of 0.23 μg/L (Deshommes et al., 2010). A third study investigated 16 water pipes, in which all 401 samples were below the detection limit of 0.5 μg/L (Zietz et al., 2015). Deshommes et al. (2012) investigated 45 taps in a large facility, including 9 basement taps that simulated dead ends due to their location and infrequent use. The water was sampled after 12 hours of stagnation and samples from the basement taps had antimony concentration of 8 μg/L, which authors stated was due to low frequency of use and no flushing prior to sampling. The remaining taps had antimony concentration of 5 μg/L. The authors noted a correlation between lead concentration and antimony that may be attributed to lead release from brass and leaded solder.

Leaching of antimony from non-lead based solder

Non-lead-containing solders are a potential source of antimony in drinking water (Fuge et al., 1992; U.S. EPA, 2006). Two studies assessed the leaching of metals from copper pipes or copper coupons with non-lead-based solders (tin/antimony, tin/silver and tin/copper/silver) (Subramanian et al., 1991, 1994). Different waters were tested and the tests run over long periods to evaluate the leaching under different scenarios. Subramanian et al. (1991) evaluated 3 waters using copper pipes and took samples over a 90-day period. The tests run with high-purity water and well water had antimony concentrations below the detection limit of 1.2 μg/L for all solders tested. The test using tap water had no detected antimony up to day 7, with antimony concentration then increasing from day 14 to day 90 for tin/antimony solder. Another study by these same authors using copper coupons with various non-lead-based solders was run with 4 different test waters over a 28-day period (Subramanian et al., 1994). For all waters tested and all types of solder, the antimony concentration remained below the detection limit of 1.2 μg/L. These studies overall showed no antimony leaching for most waters, with the exception of tap water in the Subramanian et al. (1991) which had some antimony leaching after long contact times.

Accumulation of lead and other heavy metals in 2 distribution systems was studied (Fuge et al., 1992). Three samples had antimony content ranging from 0.4 to 1.7 mg/kg. The authors stated that the antimony in the pipe solid was due to the solder present in the system.

Leaching of antimony from brass materials

Plumbing devices that contain brass used in distribution systems or in premise plumbing have the potential to impact drinking water quality. Non-leaded brass and lead-free brass generally contain a maximum of 0.25% antimony and in some cases as high as 0.5% (Sandvig et al., 2007). A few studies were conducted to determine levels of antimony leaching from different brasses (refer to Table 14) (Sandvig et al., 2007, 2012; Turković et al., 2014). Two of the bench-scale studies showed no antimony leaching with a detection limit of 6 µg/L (Sandvig et al., 2007) and 0.08 µg/L (Turković et al., 2014). Turković et al. (2014) stated that this may be due to the low antimony concentration in the brasses investigated. A study using the NSF/ANSI Standard 61 testing protocol evaluated 10 non-leaded and lead-free brass devices (Sandvig et al., 2012). The tests were conducted using 11 test waters and mainly had results below the detection limit. The 2 test waters with the highest occurrence of antimony leaching, up to 0.4 μg/L, were the NSF/ANSI Standard 61 Section 9 synthetic test water in which 1 was chlorinated and the other was chloraminated.

Table 14. Antimony leaching tests from brasses
Material tested Waters tested Testing Procedure Results Reference
7 brasses

1. NSF/ANSI Standard 61 Section 9 synthetic test waterFootnote a

2. Aggressive leaching solutionFootnote b

Brass rods submersed

"Dump and fill" protocol

Water changed 3x/week

4 week exposure

N = 42

All samples < 6 ppb (6 μg/L)

Sandvig et al. (2007)

6 devices

NSF/ANSI 61 Section 8 devicesFootnote c

11 test waters with varying pH, alkalinity.

Either chlorine or chloramine.

NSF/ANSI Standard 61 testing protocol

Control run (N = 66) – all < 0.5μg/L

Runs with test waters (DL = 0.1μg/L):

- 14 detects/ N = 198

- Max concentration = 0.41 μg/L

- 4 devices all non-detect for all waters

- 2 devices had detects for 4 waters

Sandvig et al. (2012)
4 devices NSF/ANSI 61 Section 9 devicesFootnote d

Control run (N = 32) – all < 0.5μg/L

Runs with test waters (DL = 0.01–0.5μg/L):

- 19 detects/ N = 96

- Max concentration = 0.07 μg/L

- 2 devices all non-detect for all waters

- 2 devices had detects for 6 waters

5 brasses

(4 non-leaded and 1 leaded)

Antimony content: < 0.001% to 0.003%

NSF/ANSI 61 Section 9 synthetic test watera Short-term leaching tests in accordance with NSF/ANSI Standard 61 Section 9 test protocol

Detection limit = 0.08 μg/L

No antimony leaching which the authors stated may be due to low antimony content in the brasses

Turković et al. (2014)

5 "corner waters"

- representative of ~66% of distributed water in U.S. and Canada

Long-term leaching tests (experimental protocol of DIN EN 15664-1)Footnote e

26-week period

Detection limit = 0.08 μg/L

No antimony leaching which the authors stated may be due to low antimony content in the brasses

Footnotes

Footnote 1

NSF/ANSI 61 Section 9 synthetic test water: alkalinity of 500 mg/L (± 25 mg/L) as CaCO3; pH of 8.0 (± 0.5); and free chlorine dosed at a concentration of 2.0 mg/L (± 0.05 mg/L)

Return to footnote a referrer

Footnote 2

Aggressive leaching solution: synthetic; alkalinity of 10 mg/L as CaCO3; pH of 7.4; and chloramines 4.0mg/L Cl2

Return to footnote b referrer

Footnote 3

Section 8 devices are in-line devices (for example, water meters, building valves)

Return to footnote c referrer

Footnote 4

Section 9 devices are end point devices (for example, drinking water fountains, faucets)

Return to footnote d referrer

Footnote 5

Realistic simulation of the operating conditions of a dead-end line and of the consuming behaviour at a kitchen faucet

Return to footnote e referrer

ANSI – American National Standards Institute; DL – detection limit; N – sample size; NSF – NSF International

Residuals management

Treatment technologies may produce a variety of residuals that contain antimony (for example, backwash water, reject water/concentrate, media waste). The appropriate authorities should be consulted to ensure that the disposal of liquid and solid waste residuals from the treatment of drinking water meet applicable regulations. Guidance can be found elsewhere (CCME, 2003, 2007).

Page details

Date modified: