Understanding Arsenic Removal Mechanisms in Iron Media Filters

1
Understanding Arsenic Removal Mechanisms in Iron
Media Filters
Prepared for presentation at "Arsenic in Drinking
Water An International Conference at Columbia
University", New York,26-27 November2001.

• James Farrell1, Nikos Melitas1, Martha Conklin2
  and Peggy ODay3
• 1Department of Chemical and Environmental
  Engineering, University of Arizona
• 2Department of Hydrology and Water Resources,
  University of Arizona
• 3Department of Geological Sciences, Arizona State
  University

2
Iron Permeable Barriers
CrO42-
ethene 3Cl-
Fe0
TCE
Fe0 CrO42- 4 H2O ?(Fex Cr1-x)(OH)3 2 OH-
Fe0 TCE 3 H ? ethene 3Cl- 3Fe2
Iron barriers have been installed at more than 65
field sites.
3
Iron Media Filters
Effluent

• The majority of systems affected by the new 10
  µg/L MCL are small systems serving fewer than
  10,000 customers.

4
Research Objectives

• 1. Develop a kinetic model,
• 2. Determine the mechanism, and
• 3. Determine the products
• for removal of As(V) by iron media.

5
Possible Removal Mechanisms

• Reductive Precipitation

• Adsorption to Corrosion Products

• Mineral Formation

6
(No Transcript)
7
Experimental Set-Up
Iron wire
Reference electrode
reference electrode
s.s. counter electrode
N2 purge
vent
Iron filings
Iron wire electrode
Sampling port
Anaerobic solutions with HAsO42-
8
Experimental Conditions

• As(V) concentrations ranging from 100 to 50,000
  µg/L in N2 purged 3 mM CaSO4 background
  electrolyte solutions.
• (Note As(III) 0.01As(V))
• Initial and final pH values 7.
• Analysis atomic absorption and ion
  chromatography.
• Batch experiments used iron wires and columns
  were packed with Master Builders Supply iron
  filings.
• Arsenic and iron K-edge X-ray Absorption
  Spectroscopy (XAS) performed at Stanford
  Synchrotron Radiation Lab.

9
Arsenate Removal by Iron Wire in Batch Reactors
1
0.8
0.6
5000 mg/L
C/Co
0.4
100 mg/L
0.2
0
0
10
20
30
40
Elapsed Time (Days)

• Approaches 1st order removal at low
  concentrations k1 0.1 d-1
• Approaches 0th order removal at high
  concentrations ko 120 mgL-1d-1

10
Iron Wire Tafel Plots
Oxidation
Reduction

• The similar bc slope between the blank and the
  arsenate solutions indicates that water is the
  primary oxidant and As(V) reduction is
  negligible.
• The lower Ic observed in the presence of As(V)
  indicates that As blocks water reduction.
• The lower Ia observed in the presence of As
  indicates inhibition of iron oxidation.
• The increase in ba in the As(V) solutions
  indicates the formation of complexes that alter
  the oxidation properties of iron.

11
Corrosion Rates at Different As Concentrations
OH

• The presence of As(V) decreased the corrosion
  rate.
• No concentration effect for As(V) gt 100 µg/L
  suggests saturation of adsorption sites
• at low As(V) concentration.

12
Effect of Corrosion Rate on As(V) Removal
Rate Initial As(V) 100 ?g/L
Cathodically protected electrode Imposed cathodic
current 2 mA Anodic current 0.002 mA
1
0.8
0.6
C/Co
Freely corroding electrode Icorr 0.2 mA
0.4
0.2
0
0
5
10
15
20
25
30
Elapsed Time (Days)

• Similar initial removal rates due to adsorption
  on pre-existing oxide.
• Higher removal is associated with the freely
  corroding electrode.

13
Kinetic Mechanism for Arsenate Removal

• At low HAsO42-
• Excess of adsorption sites ?FeOH results in
  1st order kinetics.
• Removal is limited by mass transfer rate.
• At high HAsO42-
• Shortage of adsorption sites ?FeOH results in
  0th order kinetics.
• Removal is limited by the rate of ?FeOH
  generation.

14
High Concentration Column Experiments
k10.09 min-1 ko92 mg L-1 min-1

• Faster removal near column inlet due to faster
  corrosion by dissolved O2.

15
Apparent Reaction Order
10000
Influent - Port 1
1000
Slope 0.69
Removal Rate (µg/Lmin)
100
Slope 0.38
Port 1- Effluent
10
100
1000
10000
100000
Arsenate Concentration (µg/L)

• Apparent reaction order depends on As, Icorr,
  other ligands.

16
Low Concentration Column Experiments
Initially, the column was operated at a 20 sec
residence time for 1,000 pore volumes
1
Influent Concentration (mg/L)
0.8
100
300
600
1000
1500
5000
0.6
C/Co
0.4
18,000 pore volumes
0.2
0
0
20
40
60
80
100
120
140
160
180
200
220
240
Elapsed Time (Days)

• Effluent As lt 0.5 ?g/L for Influent As100
  ?g/L. ?? 20 sec
• Removal lt 0.5 ?g/L for As lt 1500 ?g/L . ??
  20 min
• Effluent Fe2 40 µg/L (depends on ionic
  strength and O2).

17
Arsenic XAS

• All arsenic is As(V).
• As-Fe distances indicate bidentate corner
  sharing among Fe-octahedra and
• As-tetrahedra.
• No evidence for a separate scorodite phase.

18
Iron XAS

• Presence of three phases metallic iron,
  magnetite, and ferric oxyhydroxide.
• Average iron oxidation state is greater near
  column inlet.

19
Reduction of As(V) to As(III)
No reduction (column studies)

• 1. Lackovic, J. A. Nikolaidis, N. P. Dobbs, G.
  M. Environ. Eng. Sci. 2000, 17, 29.
• 2. Farrell, J. Wang, J. ODay, P. Conklin, M.
  Environ. Sci. Technol. 2001, 35, 2026.
• No reduction of aqueous As(V).
• No detectable As(III) on filings after more than
  1 year elapsed.

Reduction (batch studies)

• 1. Su, C. Puls, R. W. Environ. Sci. Technol.
  2001, 35, 1487.
• After 60 days elapsed, ratio of As(III) to As(V)
  was 31.
• Ratio was independent of whether As(V) or As(III)
  was added to vial.

20
Equilibrium Potential for As(V)/As(III)

• Reduction may occur to adsorbed or solution phase
  As(V).
• Eo for reduction of bound As(V) depends on the
  relative binding strengths of As(V) and As(III)
  to cathodic sites.
• Differences in binding strength of As(V) and
  As(III) will affect the ratio of As(V) to
  As(III) on the iron surface.

21
Electrical Double Layer Affects As(V)/As(III)
Ratio at the Iron Surface

• Reduction of water leads to a negatively charged
  solution adjacent to the iron surface.
• Dissolution of Fe2 and loss of protons leads to
  a negative charge on the iron surface.

22
Iron Wire Tafel Plots
Oxidation
Reduction

• No measurable reduction at -740 mV/SHE and
  As(V) 10 mg/L.
• Test for reduction at lower potentials and
  higher As(V).

23
Current from Iron Wire at Fixed Potential

• Increase in ionic strength allows reduction gtgt
  double layer effects are important.
• Reduction of bound As(V) occurs at a higher
  potential than aqueous As(V) gtgt As(III) is
  bound more strongly at cathodic sites (alkaline
  pH conditions).

24
Ferrihydrite Adsorption Edge
Raven, K. P. et. al. Environ. Sci. Technol. 1998,
32, 344-349.

• Relative binding strength of As(III)/As(V) is
  dependent on the arsenic concentration and
• type of iron oxide.

25
Batch versus Column Conditions

• Elevated pH at iron surface decreases Eeq for
  As(V) reduction.
• Buildup of H2 quenches water reduction reaction
  in sealed vials and lowers surface pH.
• y effect decreases after water reduction ceases.
• As(III)/As(V) ratio should be close to its bulk
  solution value at equilibrium in sealed vials.
• E at iron surface is lower in sealed vials.
• As(V) reduction not likely in open systems.

26
Comparison with Other Technologies

• Recent pilot testing in Tucson showed GFH much
  more effective than activated alumina,
  Fe3-modified alumina, ion exchange, and ferric
  chloride coprecipitation.

27
Conclusions

• Batch
• Arsenic adsorption decreases the rate of iron
  corrosion.
• Removal kinetics should be 1st order for
  environmentally relevant As(V) concentrations.
• Column
• For HAsO42- 100 µg/L, removal to HAsO42- lt
  0.5 µg/L in lt 20 seconds.
• Dissolved oxygen aids removal.
• XAS
• Complex formation is the removal mechanism.
• No detectable reduction to As(III).

28
Acknowledgements

• National Institutes for Environmental Health
  Sciences (P42-ES04940).