Hyphenated Techniques for Determining pH
Dependent Pore-Scale Uranium (VI) Speciation:
FFF-ICP-MS
PIs
James F. Ranville and Bruce D. Honeyman
Department of Chemistry and Geochemistry
&
Division of Environmental Science and Engineering
Colorado School of Mines
Golden, CO 80401
Graduate Student Investigator
Emily K. Lesher
Presented at DOE-ERSP PI Meeting
April 9, 2008
Lansdowne,VA
Talk Outline•Introduction
–Project relevance
–Aqueous speciation
•Field Flow Fractionation
–Theory & instrumentation
–Application for U characterization (previous work)
•Laboratory bacteria U sorption
•SREL soil leachate
•Aquitard pore water
–Quantitative application for U speciation
• Well-defined ligands
•Future work
INTRODUCTION: Project Relevance
INTRODUCTION:Project Hypotheses
• Geochemical
– Uranium solution speciation in groundwater willdepend on solution composition and will respondto changes in composition
– Geochemical process such as sorption andbiotransformation will be affected by U speciation
• Analytical
– Hyphenated techniques that combine separation(field flow fractionation) and detection (ICP-MS)can provide a means of speciation measurement
– Developed techniques, which utilize small volumesamples (micro liter), will allow examination of Usolution phase speciation with high spatialresolution in heterogeneous systems
Solution
Speciation
Influences
Contaminant
Behavior
Complex
formation with
aqueous and
surface ligands
Macromolecular,
nanoparticulate,
and colloidal
ligands
INTRODUCTION:Speciation
Field flow fractionation
Theory & Instrumentation
FFF is a separation method that when
combined (hyphenated) with ICP-MS will
allow measurement of aqueous phase
U(VI) speciation
Analogous to chromatography
(no stationary phase)
Field flow fractionation - ICP-MS
Cross flow HPLC pump
Channel flow HPLC
pump Carrier
solution
UV detector
Fluorescence
detector
ICP – sample
introduction,
plasma
generation
Mass spec:
elemental
detection
Calibration standards
(connected at T forstandardization runs)
T valve
T connector
Pump
Internal standard:
50 ppb Bi in 4% HNO3waste
Injection port
Flow FFF: Channel ConfigurationFlow FFF: Channel Configuration
Plexiglas blocksPlexiglas blocks
PorousPorous
ceramicceramic
fritsfrits
••Separates colloids (inorganic or organic solids from ~2 nm Separates colloids (inorganic or organic solids from ~2 nm –– 1 1 m size range)m size range)
••Supra-micron particles can also be analyzed using alternate Fl FFF modes (1-20 Supra-micron particles can also be analyzed using alternate Fl FFF modes (1-20 m)m)
1 K dalton
- - - - - -Porous ceramic frits on
top and bottom
Recirculating fluid crossflow pushes particles
against accumulation wall during initial
relaxation time
Brownian
diffusion, f(d)
Carrier fluid
channel flow
d1 > d2 > d3
v1 < v2 < v3
Parabolic
velocity vectors
w
l
High fieldHigh field Fl FFF for separation of DOC Fl FFF for separation of DOC
Separation of PSS molecular weight standards (Separation of PSS molecular weight standards (DaDa))
elution time (sec)elution time (sec)
UV
absorb
ance
UV
absorb
ance
840840
14301430 65006500
48004800
1750017500
High field: 3.0 ml/min Carrier flow:1.0 ml/min
•Low field: 0.9 ml/min
•Carrier flow:1.0 ml/min
•PSS standards (Duke)•20 μL injection
•Fl detector
Time (sec.)
0 1000 2000 3000
Rela
tive F
L
0.0
0.2
0.4
0.6
0.8
1.0
26 nm 90 nm
138 nm
Low field for nanoparticles
Rt
tR
0
=
Experimental Measurement
t0 = retention time for void volumetr = retention time for sample component
FFF Theory
=l
w=D
Uw
D = diffusion coefficientU= field velocityw = channel thickness (0.25 mm)
=DV 0
Vcw2
=kTV 0
3 Vcw2d
Flow FFF
Vo = void volumeVc= volumetric cross-flow rate
= viscosity
=6kT
d 3 ( 2ro)w
d = diameter= density difference
= rpmr0 = radius of centrifuge
Sedimentation FFF
Stokes-Einstein
Normal-Mode FFF Theory
Computing d from retention time
R = 6 coth 12( ) 2[ ]
Field flow fractionation:
Uranium characterization
applications
• Culture of Shewanella oneidensis (~106 cells/mL)
• pH 5 linear isotherm by FFF-ICP-MS
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
500000
0 100 200 300 400 500 600
50 ppb
500 ppb
950 ppb
1800
4000 ppb y = 14271x - 335639
R2 = 0.9993
0
10000000
20000000
30000000
40000000
50000000
60000000
0 1000 2000 3000 4000 5000
U peak area vs. spiking conc
Laboratory Investigation of U(IV)
Sorption to Bacteria
Jackson et al., Anal. Chem., 2005
U solution
U s
orb
ed
Investigation of U Sorption to Bacteria• pH dependence of sorption
0
20000
40000
60000
80000
100000
120000
140000
160000
0 200 400 600 800
pH 9
pH 7.9
pH 6
pH 5
• Suspected high MW
exopolymer
competes with cells
for U
•Proton competition
is different for cells
vs exopolymer
•FFF can be useful in
studies of mixtures
Jackson et al., Anal. Chem., 2005
U and Ni U and Ni porewater porewater characterization atcharacterization at
Savannah RiverSavannah River Site, Aiken SCSite, Aiken SC
•• Examined filtered (0.2 Examined filtered (0.2 μμm)m)soil extracts (water-soil extracts (water-dispersible colloids) fromdispersible colloids) fromSRELSREL
•• High levels of DOC suggestHigh levels of DOC suggestthat metal-DOC bindingthat metal-DOC bindingmight be important in Ni andmight be important in Ni andU speciationU speciation
•• Mildly acidic pHMildly acidic pH
•• Interface Fl-FFF on-line withInterface Fl-FFF on-line withICP-MS for (multi-) element-ICP-MS for (multi-) element-specific chromatogramsspecific chromatograms
Jackson et al., ES&T, 2005
SP5 water extract:SP5 water extract:
0
100000
200000
300000
400000
500000
600000
700000
800000
0 0.05 0.1 0.15 0.2 0.25
0
1000
2000
3000
4000
5000
6000
Al
UFe
UV
Particle size ( m)
Al,
Fe,
Ni,
U,
ICP
-MS
cp
s
UV
res
po
nse
UV
res
po
nse
low fieldlow field Fl-FFF ICP-MS Fl-FFF ICP-MS
U (( g l-1)
237
Ni ( g l-1)
407
Fe ( g l-1)
2060
Al ( g l-1)
4370
DOC (mg l-1)
200
pH
5.94
Sediment
SP5
Jackson et al., ES&T, 2005
SP5 water extract:SP5 water extract:
0
10000
20000
30000
40000
50000
0.000 0.001 0.002 0.003 0.004 0.005
0
500
1000
1500
2000
U
Ni
Al
UV
Al,
Fe,
Ni,
U,
ICP
-MS
cp
s
Particle size ( m)
UV
res
po
nse
UV
res
po
nse
High FieldHigh Field Fl FFF Fl FFF
ICP-MSICP-MS
Jackson et al., ES&T, 2005
SREL Study SummarySREL Study Summary
12.49.69.122.6HumicHigh Field
30.517.320.338.4HumicLow Field
14.826.717.66.6ColloidFFF
21.82226.829.4Humic
0.152.818.93.9ColloidSEC
TB6SP5B1SP2B1SP1B1U
•• Environmental significance at SRSEnvironmental significance at SRS
–– ‘‘dissolveddissolved’’ U may be partially non-available, U may be partially non-available,associated with colloids and DOCassociated with colloids and DOC
–– Soil/groundwater reactive transport modelsSoil/groundwater reactive transport models shouldshouldinclude speciationinclude speciation
% U distribution
• Located in western
Canada
• 80 m thick
• Mildly alkaline
• High in carbonate
• High in DOC
• Well-instrumented
with
peizometers
• Natural U source
Compare to: Clay-rich glacial till aquitard
Ranville et al., J. Contam Hydrol, 2007
Depth = 2.3 mU = 460 μg/L
DOC = 136 mg/L
Depth = 4.5mU = 430 μg/L
DOC = 77 mg/L
Depth = 11.9 mU = 38 μg/L
DOC = 19 mg/L
Ranville et al., J. Contam Hydrol, 2007
Comparison of speciation model and
FFF results
Water composition (mM)
Ca
Mg
Na
K
Cl
SO4
CO3
U
DOC
(mg/l)
pH
10 150 275 1.5 1.0 300 4.8 1.48E-3 77 7.6
Uranyl species percentage distribution
Ca2UO2(CO3)30 CaUO2(CO3)3
= UO2(CO3)2
= UO2(CO3)3
4- -FA2UO2
66.2 2.2 0.8 28.9 1.9
FFF results
• DOC was low in MW
• 50% recovery of DOC
• % Organic bound U measured = 0.62 %
• Assuming lost DOC contained U
• 0.62 % x 2 = 1.24 %
Field flow fractionation
Quantitative applications
Why develop measurement methods when we
have models?
• Are the models accurate for complex watercompositions ?
– Compare predictions to measurements
• What about complex ligands for whichthermodynamic data are lacking?
– Uncharacterized NOM
– Natural nanoparticles
• Separation-based approach
– Differentiation of “free” vs complexed
– Mixtures
• Competitive reactions (e.g. cells and mineralparticles)
• Development begins with well-characterized ligands
U(VI) solution complexes can be predicted by
computer modeling
• For example-Visual MINTEQ (Gustafsson, 2006)
• Uses thermodynamic data on uranium association with
• Database likely “good” for complexes with dissolved
inorganic and simple organic ligands (subject of new
ERSP project: K. Hatfield PI)
-fulvic acid, etc.-acetate, -citrate, etc.Organic
-FeOx surface sites-OH, -CO3, etc.Inorganic
ColloidalDissolvedLigand
U(VI) solution complexes predicted by computer
modeling: DOC vs carbonate
U=10-6 M, DOC = 1 mg C/L
Ca2+ = 10-3 M, atmospheric CO2,
U=10-6 M, DOC = 10 mg C/L
Ca2+ = 10-3 M, atmospheric CO2,
Ca2+ = 2 x 10-3 M, 10 X atmospheric CO2,
Increase organic complexes
Decrease organic complexes
U(VI) solution complexes predicted by computer
modeling: Soil pH variations
Well-characterized ligand:
Nanoparticulate Hematite
Well-characterized
ligand:
Nanoparticulate
Hematite
Existing thermodynamic
surface complexation
sufficient to allow
predictions
What about natural
nanoparticles ?
Experimental results and FITEQL model simulations of fraction uraniumsorbed onto 0.09 g/l, 0.9 g/l and 9.0 g/l hematite at variable pH. Ionicstrength = 0.1, U(VI)T = 10-6 M and atmospheric concentrations of CO2.Lenhart and Honeyman (1999).
Uranium sorption onto hematite: [hematite]=0.09
g/L, [U]=1 uM (238 ppb), I=0.1 M, atmospheric CO2
-50
0
50
100
150
200
250
300
350
0 2 4 6 8 10 12 14
pH
[U],
pp
b
meas. conc. U, ppb
conc U sorbed
(calculated), ppb
Batch Sorption Experiments
Phase separation using filtration
Volume for ICP-MS ~10 ml
U, Fe fractograms for pH 3.4 sorption experiment
-100
0
100
200
300
400
500
600
700
0 500 1000 1500 2000
time, seconds
[U],
[F
e];
pp
t U
, p
pb
Fe
[Fe], ppb
[U], ppt
U, Fe fractograms for pH 6.1 sorption experiment
-100
0
100
200
300
400
500
600
700
0 500 1000 1500 2000 2500
time, sec
[U],
[F
e];
pp
t U
, p
pb
Fe
[Fe], ppb
[U], ppt
Sorption Experiment:Fl FFF-ICP-MS
20 uL injected into Fl FFF-ICP-MS
U, Fe fractograms for pH 3.4 sorption experiment
0
100
200
300
400
500
600
700
0 500 1000 1500 2000
time, seconds
U a
nd
Fe
co
nc
en
tra
tio
ns
, p
pt
U,
pp
b F
e
0
5
10
15
20
25
30
35
40
45
50
Fe57
Void [Fe]/[U]
Peak [Fe]/[U]
U238
• Smaller [Fe]/[U] ratio in void peak. Indicates somedissolved U.
• Increasing [Fe]/[U] ratio over peak. Function of sorbed [U]being surface area dependant, [Fe] being volumedependant
[Fe]/[U] ratio increases
Over peak because [Fe] is
Volume dependant while
[U] is surface area dependant
U, Fe fractograms for pH 5.3 sorption experiment
-100
0
100
200
300
400
500
600
700
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time, seconds
U a
nd
Fe c
on
cen
trati
on
s, p
pt
U,
pp
b F
e
-10
-5
0
5
10
15
20
25
[U], ppt
[Fe], ppb
peak [Fe]/[U]
void [Fe]/[U]
• Void peak ratio close to peak ratio, indicatesU present in hematite that was not fullyinjected/equilibrated
Comparison of sorption results: FFF vs batch
FFF Filtered: aqueous
analysis pH % Sorbed Log Kd % Sorbed Log Kd
3.4 4.5 1.7 5.1 1.8
4.2 58.4 3.2 29.2 2.7
4.2 50.4 3.1 29.2 2.7
5.3 96.1 4.4 99.1 5.1
6.1 100 5.3 99.6 5.4
•Method Validation
• Complete Hematite work
• Examine IHSS HA
•Up-scaled Lab Experiments
•Construct small tank with layered
heterogeneous materials
•Carbonate, organic matter
•Sample at high spatial resolution, utilizing the
small sample volume requirements of FFF-
ICP-MS to examine U speciation
•Field-scale
•Examine U speciation at field sites (part of
new ERSP project)
Future Work
AcknowledgementsFunding
Student
Investigators
DOE ERSP Grant: ER64419
Edna Bailey Sussman Fellowship
RD-83332401-0
Emily Lesher: Poster S4: Optimization of Flow-Field Flow Fractionation-Inductively
Coupled Plasma Mass Spectrometry for U(VI) Characterization
