Copyright [insert date set by system] by [CH2M HILL Entity] • Company Confidential

Passive Treatment Systems for the Removal of Selenium: Selenium Removal Mechanisms with Biochemical Reactor Substrate

R.C. Thomas, J.J. Tudini, J.S. Bays, M.A. Girts, K.B. Jenkins, L.C. Roop, and T. Cook

June 3 2013

Overview of Presentation

Outline

Barrel-Study Overview Aqueous Results Volumetric/Aerial Loading Rates Substrate Se Removal Volatilization Substrate Se Speciation Modeling Calibration/Removal Rate

Development Secondary Parameters Conclusions

NAMC Selenium Report 2010

http://www.namc.org/docs/00062756.PDF

Case History (2011-2012)Two Concurrent BCR Pilot Studies for Coal Mine Drainage

Outlet A: 389 daysPumped Inflow

Outlet B: 180 daysGravity Inflow

Influent

Effluent ManifoldEffluent Sample Ports

Influent – Gravity Feed

Effluent Sample Ports

TDS 1 g/L Se in 6–17 µg/LNO3-N 0.47 mg/LTP 0.071 mg/L

Source:R. Thomas (2011)

Outlet 022: Variable Flow and Limited Flow Control Resulting in Variable Se Reduction

0.1

1.0

10.0

5/27/2011 6/26/2011 7/26/2011 8/25/2011 9/24/2011 10/24/2011 11/23/2011

Sele

nium

Con

cent

ratio

n (p

pb)

Outlet 022 Pilot: 180 Days

Barrel A Barrel B Barrel C Barrel D Monthly Avg Limit Max Daily Limit Influent

Outlet 033: Consistent Se Reduction Through Winter

8.2 µg/L

Outlet 033 Pilot: 389 days

0.8-0.9 µg/L

Source:CH2MHILL (2012)

8.2°C mean4°C min<0°C air

Focus on Winter Water Minimum Temperature 4°C

Avg Water Temp 8.2°CAvg Min Air Temp 4.2°C

Average Se Mass Removed Per Day – 022

4,106

542

3,564

Outlet 022 - Barrel A, Avg. Se Mass (µg/day)

Se Influent

SeEffluent

Se Removed

87% removed

3,827

338

3,489

Outlet 022 - Barrel BAvg. Se Mass (µg/day)

Se Influent

SeEffluent

Se Removed

92% removed

3,523

678

2,844

Outlet 022 - Barrel CAvg. Se Mass (µg/L)

Se Influent

SeEffluent

Se Removed

83% removed

3,027

288

2,739

Outlet 022 - Barrel DAvg. Se Mass (ug/L)

Se Influent

SeEffluent

Se Removed

90% removed

Average Se Mass Removed Per Day – 033

1,726

139

1,587

Outlet 033 - Barrel A, Avg. Se Mass (µg/day)

Se Influent

SeEffluent

Se Removed

91% removed

2,348

179

2,169

Outlet 033 - Barrel BAvg. Se Mass (µg/day)

Se Influent

SeEffluent

Se Removed

92% removed

1,761

198

1,563

Outlet 033 - Barrel CAvg. Se Mass (µg/L)

Se Influent

SeEffluent

Se Removed

89% removed

1,752

151

1,601

Outlet 033 - Barrel DAvg. Se Mass (ug/L)

Se Influent

SeEffluent

Se Removed

90% removed

Se Mass Removed vs Volumetric Loading

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035

Se R

emov

ed (g

/m3 -d

)

Se Loading (g/m3-d)

022 HRT12 022 HRT18 022 HRT24 033 HRT12 033 HRT18 033 HRT24

Winter Barrel A

Winter Barrel B

Reaction Fronts in BCR Substrates Over Time

Time

Se Mass Removed vs Surface Aerial Loading

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

0.018

0.000 0.005 0.010 0.015 0.020 0.025

Se R

emov

ed (g

/m2 -d

)

Se Loading (g/m2-d)

022 HRT12 022 HRT18 022 HRT24 033 HRT12 033 HRT18 033 HRT24

Winter Barrel A

Winter Barrel B

Barrel Selenium Profile

012345678910

Se Concentration (mg/kg)

Outlet 022 Substrate Total Se Concentration (180 days)

022A

022B

022C

022D

Top

Middle

Bottom

upflo

w

Sampled 11/22/11(180 days)

Flow Initiated 5/26/11

Barrel Selenium Profile

024681012

Se Concentration (mg/kg)

Outlet 033 Substrate Total Se Concentration (180 & 389 days)

033C

033D

033A

033B

Series5

Top

Middle

Bottom

dow

nflow

Surface

Sampled 6/18/12(389 days)

Sampled 11/22/11(180 days)

Flow Initiated 5/26/11

Selenium Mass Balance vs Sediment Core: Gap Suggests Volatilization Loss

B-A B-B B-C B-D A-A A-B0

100

200

300

400

500

600

700

CoreBalance

Core

Se (m

g)

022 - A 022 - B 022 - C 022 - D 033 - C 033 - D

Vertical Distribution and Speciation of Selenium: Reduction, Sorption, Volatilization

Source:CH2MHILL (2012)

0

1

2

3

4

5

6

7

8

9

10

0

1

2

3

4

5

6

7

8

9

10

mg/

kg To

tal S

e

mg/

kg Se

Substrate Selenium Speciation after 180 days

Salt Weakly adsorbed S2Se/Elemental Se Selenides Total Se

123% 88%105%131%87% 88%

Percent Recovery of Sequential Extractions versus the total digests

87% 105% 108% 88% 94%

In In Out In OutIn

Vertical Distribution and Speciation of Selenium: Reduction, Sorption, Volatilization

Source:CH2MHILL (2012)

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0.0

2.0

4.0

6.0

8.0

10.0

12.0

mg/

kg To

tal S

e

mg/

kg Se

Substrate Selenium Speciation after 389 days

Salt Weakly adsorbed S2Se/Elemental Se Selenides Residual Total Se

84% 189%95%121%94%Percent Recovery of Sequential Extractions versus the total digests

167% 99% 67% 80%

In Out In Out

Model Rate Constant Calibration

Empirical data from barrel studies used to calibrate treatment rates in the development of a sizing model for a full-scale treatment system

General terms of mass conservation for a plug flow system are used while maintaining the first order model as the basis

Empirical data was modeled using first order area-based treatment wetland model of Kadlec & Knight (1996) & Kadlec & Wallace (2009)

Model estimates the potential BCR effluent concentration for a given flow rate and concentration for a given BCR area, as adjusted for temperature

J = k(Ci – Ce) where:J = zero-order contaminant removal rate [g/m2/yr])k = first-order, area-based rate constant (m/yr)Ci = influent concentration (g/m3)Ce = effluent concentration (g/m3)

0 A

Cout/Cin

0

1

Wetland Area

First-order Area-based Treatment Model

C*

Kadlec and Knight first-order treatment model

Q dC/dA = -k(C-C*) Calibrate k by pollutant

? ? ?

First-order treatment model was expanded to the P-k-C* model, which is solved by relating: 1) hydraulic loading, 2) removal rate, 3) concentration terms, and 4) hydraulic mixing

Rate Constant vs Hydraulic Loading Rate

0

500

1,000

1,500

2,000

2,500

3,000

0 50 100 150 200 250

k T(m

/yr)

HLR (cm/d)

22A 33A 22B 33B 22C 33C 22D 33D

Barrel Trendline Equation R2

Outlet 022A y=281.32e0.0115x 0.9888Outlet 022B y=413.88e0.0108x 0.9045Outlet 022C y=293.37e0.012x 1Outlet 022D y=331.43e0.0134x 0.8944Outlet 033A y=122.74e0.0283x 0.9122Outlet 033B y=155.06e0.0272x 0.9892Outlet 033C y=201.14e0.019x 0.9872Outlet 033D y=64.395e0.035x 0.9215

First-Order, Area-Based Tanks-in Series Rate Constant Modeling Results

Passive Treatment Systems rarely operate as plug-flow reactors– Heterogeneous substrate,

preferential flow Passive Treatment Systems are

better characterized hydraulically as a series of well-mixed tanks (3 – 6 in series for VFW)

Barrel-study data was configured in a three tank series using the P-k-C* model to assess the variable hydraulic characteristics on overall performance

First-Order, Area-Based Tanks-in Series Rate Constant Modeling Results

1st-order, Area Based, Rate Constants (m/y)

Barrel Outlet 022a Outlet 033a Outlet 033b Outlet 033c

A 1,379 1,498 675 675

B 1,744 1,671 1,177 1,177

C 1,246 1,180 — —

D 1,547 1,115 — —

Note: Target effluent concentration was set at 4 µg/L. a Influent temperatures monitored during the summer and fall operations used in model ( set at 1). b Influent temperatures monitored during the winter operations used in model (10.5°C) ( determined to be 0.95). c Influent temperature set at 9°C in model (θ determined to be 0.96).

Secondary Parameters (Salts & Labile C):Conductivity, TDS, Alkalinity, BOD

0

500

1,000

1,500

2,000

2,500

3,000

5/7/2011 8/15/2011 11/23/2011 3/2/2012 6/10/2012

Cond

. (um

hos)

Outlet 033

Inf A B C D

0

500

1,000

1,500

2,000

2,500

5/7/2011 8/15/2011 11/23/2011 3/2/2012 6/10/2012

TDS

(mg/

L)

Outlet 033

Inf A B C D

1

10

100

1,000

5/7/2011 7/16/2011 9/24/2011 12/3/2011

BOD

(mg/

L)

Outlet 033

Inf A B C D

0

200

400

600

800

1000

1200

5/7/2011 8/15/2011 11/23/2011 3/2/2012 6/10/2012

Alka

linity

(mg/

L)

Outlet 033Alkalinity

Inf A B C D

Secondary Parameters: Nutrients and Color

0.0

0.5

1.0

1.5

2.0

2.5

5/7/2011 7/16/2011 9/24/2011 12/3/2011

Nitr

ate+

Nitr

ite (m

g/L)

Outlet 033

Inf A B C D

0

5

10

15

20

25

30

5/7/2011 7/16/2011 9/24/2011 12/3/2011

Phos

phor

us (m

g/L)

Outlet 033

Inf A B C D

0

10

20

30

40

50

60

70

80

90

5/7/2011 7/16/2011 9/24/2011 12/3/2011

TKN

(mg/

L)

Outlet 033

Inf A B C D

0

500

1,000

1,500

2,000

2,500

3,000

5/7/2011 7/16/2011 9/24/2011 12/3/2011

Colo

r (ηg

/L P

tCc)

Outlet 033

Inf A B C D

Secondary Parameters: Metals

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

5/7/2011 7/16/2011 9/24/2011 12/3/2011

Iron

(mg/

L)

Outlet 033

Inf A B C D

0.0

0.5

1.0

1.5

2.0

2.5

3.0

5/7/2011 7/16/2011 9/24/2011 12/3/2011

Mn

(mg/

L)

Outlet 033

Inf A B C D

Conclusions

Selenium was effectively reduced by ~90% on average in all barrels. Selenium removal was effective at all HRTs tested (12 to 24 hours). Cold temperatures (below freezing) did not affect selenium removal

rates significantly, and effluent selenium concentrations remained below discharge limits following startup.

The organic substrates tested could generate secondary byproducts:– short‐term increase in conductivity from the flushing of excess salts from the

substrate during startup– elevated oxygen demand from excess labile carbon in the effluent– suspended solids (mainly fine particulate organic matter)– excess nutrients in the substrate mixture– release of regulated metals such as iron and manganese, which might also

occur in small amounts in the substrate mixture.

Conclusions Se removal occurs at the influent water‐substrate interface

– consistent with first‐order processes– extent of selenium distribution within the substrate is dependent on the selenium load.

Substrate Se speciation indicates:– dominant mechanism is reduction of selenate to reduced forms of selenium that are weakly adsorbed to

the substrate during early stages of selenium removal– approximately >50% weakly adsorbed Se attributable to selenite– Highly immobile elemental selenium or selenosulfide account for about a quarter of the total selenium

retained– very little metal selenide was found.

Both substrate coring and water-balance mass-balance evaluation methods show selenium retention– higher water balance method suggests a higher overall selenium retention than that observed in the

substrate cores, suggesting loss to volatilization. Barrel media type B appeared to balance high treatment performance with moderate

production of byproducts. Selenium treatment BCRs have been shown to treat mine site discharges effectively in

order to achieve selenium compliance limits.

Acknowledgements

Thanks to Alpha Natural Resources

Thanks to supporting engineering and science staff at CH2M HILL

Questions