Significance of Microporosity to Reactive
Transport Modeling at DOE Sites
James A. DavisU. S. Geological Survey
Menlo Park, CA
Collaborators:
USGS: Gary Curtis, Kate Campbell, Deb Stoliker, Patricia Fox
LBNL: Carl Steefel, Li Li, Ken Williams
Johns Hopkins University: Joanne Stubbs, Dave Elbert,
Linda Veblen, David Veblen
PNNL: John Zachara, Phil Long, Steve Yabusaki
Uranium-Contaminated DOE Sites
Naturita UMTRA Rifle UMTRA
O
U
2.02Å2.40Å
3.43Å
O
O
O
O
O
Fe
C
O
O
C
O
O
Fe
Fe
2.88Å
+
Reactive Transport Modeling
=
Hanford 300 Area
Reactive transport models are commonly based on the
continuum representation of porous media, in which the
physical, chemical, and biological variables describing the
system vary continuously in space.
Reactive Transport Modeling: Continuum Models
An REV has average values
of physical, chemical, and
microbiological variables
REV1 REV2
River
No-flow
boundary
Discretization of
modeling domainREV
An REV is “well mixed”.
There are no sub-grid
gradients, e.g., in physical
structure, chemical
concentrations, surface area,
or biological properties.
5-100 m
2 km
Naturita
UMTRA site:
Alluvial
Aquifer
Sediment
Texture
50% cobbles,
>6.4 cm;
15% <3mm,
~85% of U(VI)
sorption
Slide 5
Cobble
Sand grain
Silt grain
Preferred
Groundwater
Flowpath
Physical heterogeneity within REVPore
Scale
Faster flow
Mass transfer
RTM simulations by L. Li and C. Steefel
Layer of fine-grained
sediment
Effect of Subgrid Physical
Heterogeneity with Local
Chemical Equilibrium
Slide 7
<3 mm Naturita
composite
5.5 6.0 6.5 7.0 7.5 8.0
pH
10
9
8
7
6
5
4p
C
UO22+
UO2OH+
UO2(CO3)34-
UO2(CO3)22-
(UO2)2CO3(OH)3-
U(VI)tot = 1 x 10-6M; 430 ppm CO2
5 mM NaNO3Ca2UO2(CO3)3
o
1.78 mM Ca(NO3)2
UO2CO3o
CaUO2(CO3)32-
5.5 6.0 6.5 7.0 7.5
pH
10
9
8
7
6
5
4
pC UO2
2+
UO2(CO3)34-
UO2(CO3)22-
(UO2)2CO3(OH)3-
U(VI)tot = 1 x 10-6M; 2% CO2
5 mM NaNO3
Ca2UO2(CO3)3o
1.78 mM Ca(NO3)2
UO2CO3o
Calciteprecipitation
CaUO2(CO3)32-
Fox, Davis, and Zachara, 2006, GCA
Importance of
Ca2UO2(CO3)30 aqueous
species at all 3 DOE sites
At 2% pCO2 typical of groundwater
conditions at all 3 sites
At equilibrium with 430 ppm CO2
Naturita; Rifle: pH 7
Hanford: pH 7.5-8
0 5 10 15 20 25 30 35 40 45 50
Time (hr)
0
10
20
30
40
50
60
70
80
90
100
% U
(VI)
ad
so
rbed
Naturita uncontaminated
U(VI)tot = 10-6M
Ca(NO3)2 = 2.3 mM
NaNO3 = 2 mM
500 ppm CO2
pH = 7.95
sediment (250-500 um) = 77 m2/L
0 5 10 15 20 25 30 35 40 45 50
Time (hr)
0
10
20
30
40
50
60
70
80
90
100
% U
(VI)
ad
so
rbe
d
U(VI)tot = 10-6M
430 ppm CO2
Ca(NO3)2 = 8.9 mM
NaNO3 = 5 mM
pH = 7.3
Fox, Davis, and Zachara, GCA, 70, 1379-1387, 2006
Quartz = 8 m2/L Adsorption/desorption of U(VI)
reaches equilibrium quickly in well-
stirred batch reactors with non-
porous single mineral phase with
Ca2UO2(CO3)30 as the predominant
aqueous species
Adsorption/desorption of U(VI)
approaches equilibrium slowly in
well-stirred batch reactors with
aquifer sediments from all 3 DOE
sites with Ca2UO2(CO3)30 as the
predominant aqueous species,
taking weeks to months to reach a
steady-state U(VI) concentration
(Example: Naturita adsorption)
0 200 400 600 800 1000 1200 1400
Time (hr)
0.00
0.04
0.08
0.12
0.16
0.20
0.24D
es
orb
ed
U(V
I),
mic
rom
ol/
L
Hanford 300 Area contaminated
sediment (sample SPP2-18)= 1530 m2/L
Desorbable U(VI) = 2.4 x 10-7M
Ca = 0.73 mM
0.17% CO2
pH = 8.1
Desorption of U(VI) approaches
equilibrium very slowly in well-
stirred batch reactors with Hanford
aquifer sediments (sample SPP2-18)
(Bond, Davis, and Zachara, 2008)
Flow interruption in column
experiments with sample
SPP2-18 show that the rate
of U(VI) desorption is
rate-limited
(Qafoku et al, ES&T, 2005)
Slide 10
Nanoporosity and surface areas of Hanford sample NPP1-16 (<2 mm)
Hg porosimetry:
Porosity (pore size <300 nm): 12.6%
Surface area in pores <300 nm = 12.5 m2/g
N2 gas adsorption/desorption (BET):
Porosity (pore size <300 nm): 10.9%
Surface area in pores <300 nm = 29.9 m2/g
Total surface area of sample = 30.7 m2/g
Magnified Particle
Scale Showing
Intraparticle Pore
- - - - - - - - - - - - -
- - - - - - - - - - - - -
- - - -
- -
- - -
Temporally-variant bulk water
concentrations,
(H+, U(VI), Ca2+, HCO3-)b
Iron oxide
precipitateCalcite
HCO3-
HCO3-
AlO O
UO O
Ca2UO2(CO3)30
Local Chemical
equilibrium
Ca2UO2(CO3)30
Surface Charge Diffusive flux
Slide 12
Column Experiment: Pore Scale
Diffusive flux
Diffusion plus
dispersion?
0 200 400 600 800 1000 1200 1400
Time (hr)
0.00
0.04
0.08
0.12
0.16
0.20
0.24D
es
orb
ed
U(V
I),
mic
rom
ol/
L
Hanford 300 Area contaminated
sediment (sample SPP2-18)= 1530 m2/L
Desorbable U(VI) = 2.4 x 10-7M
Ca = 0.73 mM
0.17% CO2
pH = 8.1
Approximately 50% of U(VI) desorbs from intragranular porosity??
Bond, Davis, and Zachara, 2008, Adsorption in Geomedia II
Advection, dispersion, and diffusion of tritium out
of a column packed with Hanford sample NPP2-4
c) Elution stage 2
0
0.02
0.04
0.06
0.08
0.1
1.5 2 2.5 3 3.5
sf 1
sf 2
4 hrs
C/Co
Pore volumes
b) Elution stage 1
0
0.2
0.4
0.6
0.8
1
0 0.5 1 1.5 2
sf 1
2 hrs
d) Elution stage 3
0
0.001
0.002
0.003
0.004
0.005
0.006
3 3.5 4 4.5 5
sf 2
sf 3
16 hrs
e) Elution stage 4
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
4.5 5 5.5 6 6.5
sf 3
sf 4
10.75 days
f) Elution stage 5
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
6 6.5 7 7.5 8 8.5
sf 4
sf 5
30 min
g) Elution stage 6
0.E+00
1.E-04
2.E-04
3.E-04
4.E-04
7.5 9.5 11.5
sf 5
sf 6
7.66 days
i) Elution stage 8
0.0E+00
4.0E-05
8.0E-05
1.2E-04
1.6E-04
2.0E-04
14 15 16 17
sf 7
h) Elution stage 7
0.E+00
2.E-04
4.E-04
6.E-04
8.E-04
1.E-03
12 12.5 13 13.5 14 14.5
sf 6
sf 7
4 hrs
Procedure: Pack
sample in
column; let sit for
4 months in water
with high HTO
Solid curve shows
model with HTO
diffusion from
two immobile
zones with a total
intragranular
porosity of 1.05%
Slide 14
Naturita sediment quartz grain coatings
Relative abundances of Al and Fe in grain coatings
(μm thickness)
20 μm 20 μm
Grain coatings: Another type of microporosity?
Slide 15
Bright-field TEM images
showing needle-like
goethite (G) crystals
immersed in illite/smectite
clay matrix.
Davis et al., GCA, 68, 3621 (2004)
Naturita sediment grain coatings
U(VI) diffusive flux
Hanford uncontaminated vadose
zone sample: C5001-67B
Hanford contaminated vadose
zone sample: NPP2-2
20 μm
20 μm
Coating consists of micron-sized
mineral fragments.
Coating has much finer texture,
with a fine-grained clay coating
several microns thick at the outer
ridge of the grain. Probably
influenced by infiltration of low and
high pH pond water containing high
concentrations of Al and Si.
Stubbs et al., 2008
metatorbernite
coating
Lithic fragment
Hanford contaminated vadose
zone sample: NPP2-2
Metatorbernite precipitate [Cu(UO2)2(PO4)2·8H2O] is encapsulated within coatings on
contaminated grains
Stubbs et al., 2008Slide 18
0 20 40 60 80 100 120 140 160 180 200
Time (hr)
0.00E+000
1.50E-008
3.00E-008
4.50E-008
U(V
I) r
ele
as
ed
(m
ole
s/L
)
No grinding
Grinding (mortar/pestle)
Hanford sample NPP2-4
250-500 μm size fraction
Artificial groundwater; pH 7.5; HCO3- = 4 mM
Hanford contaminated vadose
zone sample: NPP2-4
Hanford contaminated vadose
zone sample: NPP2-2
Backscattering image of a 60 μm
wide, fine-grained clay coating.
Outer edge of coating contains
very high concentrations of Zr
and U, presumably from cladding
waste. Electron microprobe WDS
linescans show gradients in U
concentration across the coating.
Stubbs et al., 2008
Zones of Natural Bioreduction in
Rifle Aquifer Sediments
Slide 21
50 g/L air-dried sediment
Extractions performed in air; pH 9.4; 17.2 mM total carbonate
U(V
I) e
xtr
act
ed (
mo
les/
g)
Time (hr)
Sodium carbonate extractions of Rifle
sediment samples (BKG-A and RABS)
Bi/carbonate – H2O
2
Bi/carbonate + 1% H2O
2
RABS total U = 1.76E-8 moles/g
U(V
I) e
xtr
act
ed (
mo
les/
g)
Time (hr)
Sodium carbonate extractions of RABS
sediment sample with H2O2
1% H2O2 added
No H2O2 added
Slide 23
Sodium carbonate extractions of BKG-A
sediment sample with H2O2
Time (hr)
U(V
I) e
xtr
act
ed (
mo
les/
g)
1% H2O2 added
No H2O2 added
- - - - - - - - - - - - - - - - - - -
- - - -
- -
- - -
Iron oxide
precipitate
Fe2+AlO O
UO O
Ca2UO2(CO3)30
Ca2UO2(CO3)30
Diffusive flux
Acetate
Fe2+Fe2+
UO2(ppt)- - - - - - - - - - - - - - - - -
Slide 25
Acetate
Nanoparticulate UO2(s)
1020 1030 10351010 10251015
Field remediationField research site
Column studies
Spectroscopy
Batch studies
Total Surface Sites in System(Fine-grained porous media,0.1 moles sites/m3)
Experimental Upscaling
No bulk spatial
gradients in well-
mixed reactor
Chemical
gradients at pore
scale as f(flow)
and along
reaction fronts
Subsurface
heterogeneity; spatial
and temporal gradients
Lichtner and Kang, Water Resources Research, 43, W12S15, 2007
Multiscale Continuum Models
Primary Continuum (3-dimensional)
Sub-grid scale domains
(1-dimensional)
One of the largest problems for single continuum RTM at the field
scale is heterogeneities of physical, chemical, and biological properties
at the sub-grid scale and the non-linear scale dependence of coupled
processes. Intragranular pore space and mineral grain coatings may be
an important physical regime for many U-contaminated sites.
Conclusion
Research Priorities:
1) Multiscale continuum models (requires high performance computing)
2) Multiscale experimental studies
3) Methods for field characterization of significant parameters
4) Improved but simplified conceptual models for coupled processes (e.g.,
sorption and aqueous speciation)
5) Better understanding of parameter and conceptual model uncertainties
Cobble
Sand grain
Silt grain
Faster flow
Groundwater sampling: What
mixture of water is sampled?
Slide 12
Well
Batch/column Intermediate-scale studies Field-scale predictions
Need for multiscale experiments!
