EPA/600/R-21/032 | June 2021 | wvwv.epa.gov/research
&EPA
United States
Environmental Protection
Agency
Report: Development of a Disposable AChE Sensor
for As(lll) and Field Analysis Method; Tests with
Groundwater Samples from Shepley's Hill Landfill
Progress for o Stronger Future
Office of Research and Development
Center for Environmental Solutions and Emergency Response
Land Remediation and Technology Division
-------
Report: Development of a Disposable AChE Sensor for As(III)
and Field Analysis Method; Tests with Groundwater Samples
from Shepley's Hill Landfill
by
Tao Li1, Robert Ford1, Jason Berberich2, Endalkachew Sahle-Demessie1, Eunice
Varughese3, Rick Wilkin4
1 Land Remediation & Technology Division (LTRD), Center for Environmental Solutions and
Emergency Response (CESER), Office of Research and Development, US EPA, Andrew W. Breidenbach
Environmental Research Center, 26 W. Martin Luther King Dr., Cincinnati, OH 45268
2 Department of Chemical, Paper, and Biomedical Engineering, Miami University, 650 E. High St.,
Oxford, OH 45056
3Water Infrastructure Division (WID), Center for Environmental Solutions and Emergency Response
(CESER), Office of Research and Development, US EPA, Andrew W. Breidenbach Environmental
Research Center, 26 W. Martin Luther King Dr., Cincinnati, OH 45268
4 Groundwater Characterization & Remediation Division, Center for Environmental Solutions and
Emergency Response (CESER), Office of Research and Development, US EPA, Robert S. Ken-
Environmental Research Center, 919 Kerr Research Dr., Ada, OK 74820
Project Officer: Barbara Butler
Land Remediation and Technology Division
Center for Environmental Solutions and Emergency Response
Cincinnati, Ohio, 45268
ii
-------
Notice/Disclaimer
The views expressed in this presentation are those of the authors and do not reflect the official
policy or position of the Unites States Environmental Protection Agency (EPA). Any mention of
trade names, manufacturers or products does not imply an endorsement by the United States
Government or the EPA. The EPA and its employees do not endorse any commercial products,
services, or enterprises.
The EPA, through its Office of Research and Development, funded and conducted the research
described herein under an approved Quality Assurance Project Plan (Quality Assurance
Identification Number K-WID-00121400-QP-l-l). It has been subjected to the Agency's peer and
administrative review and has been approved for publication as an internal EPA document.
This report covers a period from 2014 to 2019, with all work completed by 2020.
-------
Foreword
The U.S. Environmental Protection Agency (US EPA) is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions leading to a
compatible balance between human activities and the ability of natural systems to support
and nurture life. To meet this mandate, US EPA's research program is providing data and
technical support for solving environmental problems today and building a science
knowledge base necessary to manage our ecological resources wisely, understand how
pollutants affect our health, and prevent or reduce environmental risks in the future.
The Center for Environmental Solutions and Emergency Response (CESER) within the
Office of Research and Development (ORD) conducts applied, stakeholder-driven
research and provides responsive technical support to help solve the Nation's
environmental challenges. The Center's research focuses on innovative approaches to
address environmental challenges associated with the built environment. We develop
technologies and decision-support tools to help safeguard public water systems and
groundwater, guide sustainable materials management, remediate sites from traditional
contamination sources and emerging environmental stressors, and address potential
threats from terrorism and natural disasters. CESER collaborates with both public and
private sector partners to foster technologies that improve the effectiveness and reduce
the cost of compliance, while anticipating emerging problems. We provide technical
support to EPA regions and programs, states, tribal nations, and federal partners, and
serve as the interagency liaison for EPA in homeland security research and technology.
The Center is a leader in providing scientific solutions to protect human health and the
environment.
Gregory Sayles, Director
Center for Environmental Solutions and Emergency Response
iv
-------
Acknowledgements
We would like to acknowledge and thank the following people for reviewing and proofreading
the manuscript.
Mr. Bob Lien, Contaminated Sites & Sediments Branch, Land Remediation and Technology
Division, Center for Environmental Solutions and Emergency, Office of Research and
Development, US EPA, 26 W. Martin Luther King Dr., Cincinnati, OH 45268
Dr. Chunming Su, Surface remediation Branch, Groundwater Characterization and Remediation
Division, Center for Environmental Solutions and Emergency Response, Office of Research and
Development, US EPA, Robert S. Kerr Environmental Research Center, 919 Kerr Research Dr.,
Ada, OK 74820
Professor George Sorial, Department of Chemical and Environmental Engineering,
College of Engineering and Applied Science, University of Cincinnati, 2901 Woodside Dr.,
Cincinnati, OH 45219
Professor Neil Danielson, Department of Chemistry and Biochemistry
Miami University, 651 E High St., Oxford, OH 45056
Dr. John McKernan, Emerging Contaminants and Technologies Branch, Land Remediation and
Technology Division, Center for Environmental Solutions and Emergency Response, Office of
Research and Development, US EPA, Andrew W. Breidenbach Environmental Research Center,
26 W. Martin Luther King Dr., Cincinnati, OH 45268
v
-------
Contents
Notice/Disclaimer iii
Foreword iv
Acknowledgements v
List of Figures vii
List of Tables ix
Acronyms and Abbreviations x
Executive Summary xi
1. The need for an in-field speciation, total As determination tool 1
2. A prototypic enzyme sensor for As(III 2
3. Ferric oxyhydroxide formation and other concerns in sample preparation 7
4. Background of Shepley's Hill Landfill Site, the need for on-site sensors 9
5. Groundwater samples from Red Cove Area, Shepley's Hill, spatial and temporal variations of
groundwater condition and chemistry 10
6. Speciation change in sampling and current state of the art for arsenic speciation preservation 17
7. Using oxalic acid to preserve speciation for samples from Red Cove Area, Shepley's Hill Landfill, a
proposal 20
8. Pre-reduction of As(V) to As(III) for total As determination: current state of the art 23
9. Testing the AChE sensor with groundwater samples from Red Cove area, Shepley's Hill 26
9.1 Test of the field samples arrived on May 14, 2019 26
9.2 Test with Field samples in Oct 24, 25 2019 29
9.2.1 The Day 1 results and impact of oxalic acid on As(III) recovery 29
9.2.2 Storage stability of the samples at room temperature (21-23 °C) 35
9.2.3 Summary and conclusions for the second test of field samples 37
References 43
Publications and reports associated with this project: 49
vi
-------
List of Figures
Figure 1. The transduction reactions to monitor AChE activity 3
Figure 2. Kinetic mechanism for AChE inhibition by As(III) 3
Figure 3. Activity measurement set up (A) and As(III) determination by AChE sensor inhibition (B) 4
Figure 4. Profile of steady state currents before (io) and after inhibition (i;) 4
Figure 5. Comparison of total As determination by ICP-AES and As(III) determination by AChE sensor 5
Figure 6. Manual fabrication of the AChE sensors and the distribution of initial activity (io) 6
Figure 7. Projected life cycle of the disposable sensor 7
Figure 8. The oxidation of Fe2+ through oxygenation and dehydration 8
Figure 9. The location of Shepley's Hill Landfill and Red Cove of Plow Shop Pond 9
Figure 10. The distribution of monitoring wells (RSK) and Piezometers (PZ) in the Red Cove area 10
Figure 11. Clustering of groundwater samples based on pH, DO, and ORP (K=3) 11
Figure 12. Temporal (red box) and spatial (black Box) variations of GSK samples in terms of pH, DO,
and ORP 12
Figure 13. Arsenic (mg/L) concentration in RSK wells (Mar 2006-Oct 2007) 12
Figure 14. Fe in RSK wells (Mar 2006-Oct 2007) 13
Figure 15. Scree plot of PC analysis with the water chemistry of RSK samples 14
Figure 16. Biplot of principal component analysis for RSK water chemistry 14
Figure 17. Clustering of groundwater samples based on water chemistry (K=3) 15
Figure 18. Temporal and spatial variations of water matrix in RSK samples 16
Figure 19. The distribution of As and Fe in groundwater samples 17
Figure 20. Groundwater reactions that can be triggered by oxygen exposure 18
Figure 21. Oxidation of [Fe11 EDTA] under acidic condition 21
Figure 22. Oxidative decomposition of oxalic acid catalyzed by iron (OX, oxalic acid anion ligand) 22
Figure 23. Conceptual flow chart for arsenic assay with the enzyme sensor 23
Figure 24. Reactions involved in thiol-based pre-reduction of As(V) to As(III) 24
Figure 25. Correlation of the average [As(III)] of sensor analysis and total As by ICP- AES (mg/L) 28
Figure 26. Appearance of samples in field (left) and 1 day later in lab 30
Figure 27. Tris consumption to titrate acidified (Y) or untreated (N) samples (t (22) = - 5.98, p < 0.0001)
31
Figure 28. Dependence of relative standard deviation (RSD) on [As(III)] for the AChE sensor 33
Figure 29. Recovery difference between OA modified sample and unmodified sample 34
Figure 30. Flow diagram of glucose meter development (Error Grid Analysis and Accuracy Profile are
vii
-------
quoted from Chang et al104) 38
Figure 31. The design with ars operon for reporting [As(III)]. O/P, operator/promotor; Amp, amplifier for
transcription; GFP, green fluorescent protein 39
viii
-------
List of Tables
Table 1. Speciation composition and results by ICP-AES and AChE-Sensor 5
Table 2. The formulation (in |o,M) of a synthetic groundwater based on the water chemistry in Shepley's
Hill Landfill in Fort Devens, MA 6
Table 3. Condition of the groundwater samples from RSK 1-39 (Mar 2016-Sept 2017) 11
Table 4. Cluster and mean values of physical condition for the RSK groundwater samples 11
Table 5. Chemical composition of the samples from RSK wells (Mar 2006-Oct 2007, high variance in
bold) 13
Table 6. Cluster and mean values (mg/L) of chemicals for the RSK groundwater samples 15
Table 7. Technical options for preserving arsenic speciation 19
Table 8. Equilibrium constants of oxalic acid with H+, Fe3+, Fe2+ and Ca2+ 70 21
Table 9. National Recommended Water Quality Criteria for As- EPA recommendation for water quality
in terms of pollutants 24
Table 10. Variables for As(V) Pre-reduction in selected examples 25
Table 11. Samples in the test of May 2019 27
Table 12. Preparative work for sample analysis with the AChE sensor 27
Table 13. Recovery of the enzyme sensor analysis with ICP-AES as the reference 28
Table 14. Main elements in the water matrix for first field sample test (ICP-AES, mg/L) 29
Table 15. Additional Tris needed to titrate untreated sample and redox species in water matrices 31
Table 16. Summary of the sensor tests for As(III) and total As by ICP-AES 32
Table 17. T-test (n=3) evaluation of OA impact on As(III) stability 33
Table 18. Evaluation of the impact of OA modification with As(III) recovery (%) 34
Table 19. Summary of As(III) determination (|a,M) in samples with different storage times at room
temperature 35
Table 20. T-test (n=3) evaluation of storage time on As(III) stability in OA modified and unmodified
samples 36
Table 21. The impact of storage time on OA modified samples with As(III) recovery (%) 36
Table 22. The impact of OA modification on As(III) recovery (%) over 5-7 days 37
Table 23. Assessment of the arsenic sensor development with Technology Readiness Level (TRL) 41
-------
Acronyms and Abbreviations
AAS
Atomic Absorption Spectroscopy
AchE
Acetylcholine Esterase
AES
Atomic Emission Spectrometry
BSA
Bovine Serum Albumin
DO
Dissolved Oxygen
EDTA
Ethylenediaminetetraacetic Acid
EGA
Error Grid Analysis
GFAAS
Graphite Furnace Atomic Absorption Spectrometry
GFP
Green Fluorescent Protein
GMO
Genetically Modified Organism
GPS
Global Positioning System
HG
Hydride Generation
HPLC
High Performance Liquid Chromatography
ICP
Inductively Coupled Plasma
MS
Mass Spectrometry
OA
Oxalic acid
ORP
Oxidation Reduction Potential
PC
Principal Component
PPT
Push Point Device
PQQ
Pyrroloquinoline quinone
PZ
Piezometers
ROS
Reactive Oxygen Species
SMBG
Self-Monitoring of Blood Glucose
TW
Tube Wells
GPS
Global Positioning System
X
-------
Executive Summary
Mission:
The analysis of arsenic groundwater samples includes arsenic speciation and total arsenic determination.
This is typically carried out in the lab with chromatography and atomic spectrophotometry. This approach
is slow, expensive, and often inadequate to meet the need in field. In-field analysis would improve
efficiency by reducing the need for sample preparation and management. For field analysis, it is necessary
to eliminate the dependence on chromatography and atomic spectrophotometry. An amperometric AChE
sensor was proposed to address the need. In this method, the enzyme activity is reported by coupling the
enzyme reaction to amperometric sensing. As(III) is determined based on its inhibition of AChE. Total As
can be determined when a field method is ready to convert As(V) to As(III).
Management of anthropogenic sites has been identified as the target market for this sensor. This sensor
can be used to characterize the site, monitor arsenic mobility as a function of environmental factors,
evaluate treatment efficacy, and support discharge compliance.
Groundwater
Sampling
Lab or Field
Sample
Preparation
Arsenic Speciation
Total Arsenic
Determination
i
Site mapping: temporal
and
spatial variation of As
As mobility and
sequestration
Water treatment efficacy
Discharge compliance
Process for arsenic analysis: speciation and total arsenic determination.
Research approach:
The work was planned in three stages. The first stage was for sensor prototype design involving survey
and selection of the bioreceptor and a transduction mechanism, and the development of a prototypic
sensor. In the second stage, the lab prototype was tested using field samples to identify gaps for practical
use. The prototype will now be redesigned to give the a-prototype, in which all necessary features are
embedded. Product development is in the third stage.
Each stage involved multiple tasks as shown in the table below. They were created to organize research
activity, justify resource requirement, and benchmark progress. The specific aims were either created by
projected needs or based on the proposals to address the issues identified in development. Tasks were
interrelated, and their priorities changed as research proceeded.
This report covers the activity from Stage 2. Based on findings, three tasks have been identified in the
critical path for prototype design.
XI
-------
Tasks
Stage 1: Prototype design
Stage 2: Prototype development
Stage 3: Product design
1
Establish Enzyme-Inhibition as
As(III) detection mechanism, and
AChE as the bioreceptor for
As(III).
Assess field-test readiness: batch
production of the sensor, product
uniformity, storage stability, and
deployment compatibility.
Sensor production, screen
printing ink formulation, print
process design, mass production
of enzyme sensor.
2
Develop sensing mechanism:
transduction and detection.
Characterize test site: site scale,
water chemistry, temporal and
spatial variation of arsenic.
Cost: battery powered,
miniaturized, single voltage
potentiostat.
3
Integrate bioreceptor with
transducer. Develop sensor
fabrication method. Validate
sensing mechanism.
Propose chemical mechanisms of
arsenic speciation change,
sampling and sample preservation
method for sensor assay.
Sample preparation: standardized
protocols for preserving
speciation and pre-reduction of
As(V) to As(III).
4
Characterize the prototypic
sensor: selectivity, stability,
dynamic range, detection limit.
Evaluate sensor performance in
groundwater matrix, redesign
sensor prototype.
Sensor assessment in application:
error grid analysis and accuracy
profile evaluation, implication on
site modeling and decision
making.
5
Characterize accuracy, and the
precision profile of the sensor.
Develop pre-reduction for total
arsenic determination.
6
Improve biosensor modules: more
selective and durable bioreceptor,
more efficient and selective
transduction, fabrication method
that can be scaled for mass
production, low cost, and efficient
sample preparation.
7
Alternative transduction and high
throughput assay development.
Key Findings
The sensors were prepared at the rate of 100 per batch by one person (4 days/batch). The activity
of the sensor was 18.8-24.8 |oA, with amean of 21.3 ± 1.6 |o,A (N = 39). The sensors were dry
stored for > 7 days or wet stored in 0.1 M Tris-HCl, pH 7.0 for >150 days. Variation in activity
had minimal impact on its sensitivity to 7.5 |o,M As(III). This prototypic sensor was deemed ready
for field sample testing.
Cluster analysis was used to review the historical data of RSK wells at the Red Cove Area, in
Shepley's Hill pond. Temporal and spatial variations were found in the physical condition and
water chemistry. Significant water chemistry change was found in samples of 11-days apart.
Water chemistry change was observed within two well groups, suggesting significant spatial
variations within <50 meters.
In the May 2019 testing, the accuracy of the sensor was evaluated by comparing the results to
those from the ICP-AES. In the range of 0 - 20 |oM As(III), the correlation was 95%, indicating
xii
-------
good agreement between the two methods. However, the sensor had low precision at low
concentrations (< 4 |o,M).
Sample stability was identified as a critical issue. The groundwater was anoxic and rich in Fe.
Exposure to air caused As loss to the precipitate of iron oxide. A new task was created to evaluate
sample preservation with oxalic acid.
In the October 2019 testing, the sensor was tested with 24 groundwater samples from 12 wells.
The objectives included evaluating sensor precision in a wider range of matrices and testing a
hypothesis to preserve sample speciation with oxalic acid. We found the sensor had acceptable
precision (RSD 13 - 54%) when [As(III)] was between 5- 10 |oM, and poor precision (RSD 97 -
280%) when [As(III)] was between 0-3.4 |a,M.
Statistical analysis showed that oxalic acid modification prevented Fe precipitation and slowed
the loss of As(III) from the field samples.
To-do Tasks in Stage 2:
Evaluate oxalic acid modification to preserve arsenic speciation in groundwater.
Develop pre-reduction of As(V) to As(III) for total arsenic determination.
Survey new enzyme immobilization methods for sensor fabrication.
xiii
-------
1. 111" " i i vh in ill 'IK" | >¦ ;iaf i n i i vl 1 • i<• 111i111 vii mi
tool
Arsenic-related health incidents caused by human activity have been documented since the 1900s.''2 Earlier
reports included acute episodes of occupational exposures in chemical manufacture or smelting, and food
poisoning from As contaminated ingredients. Later reports showed that environmental exposures to anthropogenic
arsenic were much worse: they caused chronic hazards to more people in a large area, and the health impacts had
the potential to last a lifetime. Epidemiological studies showed that arsenic could cause health issues for
vulnerable populations at very low doses. Chronic exposures resulted in cancers and systemic health issues such
as damage to cardiovascular, metabolic, and reproductive functions. The exposure also caused neurocognitive
function loss in young children.3
As part of the problem identification, the anthropogenic arsenic sites can be classified by the types of human
activity including: 4' 5
• Unintended release from industrial manufacture, chemical storage, and arsenic waste disposal
• Application of arsenic agrochemicals and leaching of chromated copper arsenate from treated wood
• Unintended release of arsenic from coal burning
• Mining and smelting waste pollution
• Geogenic arsenic mobilization by organic waste or oil spill
The plan to manage an anthropogenic site depends on its pollution risk and the evaluation of associated health
hazards.6, 7 Risk management, involving source reduction and site remediation, relies on understanding the
pollution source and contamination mechanism. Mine waste site characterization stands out as a highly systematic
approach to support risk management in complex environmental settings. 8~'° Understanding site stability involves
evaluating the mobility and sequestration of arsenic. Arsenic is mobilized when dissolved in water and
sequestered by forming minerals with other elements in the water matrix. The underlying mechanism for arsenic
mobility is based on arsenic speciation. In mineral form, As(III) and As(V) have different solubilities. Their
interconversion through redox oscillation depends on environmental settings.9
In natural water, arsenic speciation is coupled to the change of dominant redox active species, including Fe, Mn
and S. "¦12 This has a strong impact on As ionization. Arsenous acid (H3ASO3) is uncharged at common pH range
as its pKi is 9.2. It is projected to be more mobile in water, although in certain circumstances it can form minerals
such as arsenolite/3 Arsenic acid (H3ASO4) is the product from As(III) oxidation. With a pKi of 2.2, it is
negatively charged in most water matrices and behaves like phosphoric acid. Arsenate tends to form secondary
minerals with a wide range of stabilities and solubilities. It is projected to be less mobile/3,14
Mobilization of As involves hosting mineral dissolution through redox reactions. In reductive dissolution, As(V)
is transformed to As(III) by microbial activity with organic matter as the reductant. Seasonal change promotes the
input of organic matter, microbial activity, and arsenic diffusion and fluxes. Arsenic is sequestered in the oxic
zone, typically by precipitation with Fe(III) minerals in sediment." Site topography and the heterogeneity in
geochemistry play critical roles in arsenic spatial distribution/6 The temporal pattern of arsenic mobility is
seasonal/7 Porewater is the medium for arsenic redistribution at sites/5 Surface water and groundwater are the
media for arsenic migration beyond mining waste areas/7_i9
The objective of site characterization for anthropogenic sources is to build our understanding on spatial
distribution, transport, and fate of arsenic. The distribution pattern is generated by mapping the site over a time
period. The sampling plan is dependent on the scale and geochemical complexity of the site, local hydrology, and
the pattern of local weather.
1
-------
Arsenic contamination at anthropogenic sources is usually on a local (100 m - 10 km) scale/ thus creating
environmental threats for communities or regions.7 Even at this scale, the distribution of arsenic can be highly
heterogenous. The mapping of arsenic distribution may require dense sampling. In a survey of gold deposits in the
Tampere area in Finland, arsenic distribution was surveyed with that of Au to determine if there was a
correlation.20 The survey started with mapping a 16 km2 area with 16 samples per km2. Then the sample density
was increased to 400 per km2 for two 1 km2 areas. The focused mapping revealed multiple hot spots in the two
areas, with arsenic varying from < 25 to 13,500 ppm within the 1 km2 areas. Mobility was surveyed by
characterizing the flow path of water (surface or ground) impacted by the pollution site, evaluating the stability of
the minerals that can potentially host arsenic, estimating the discharge and transport of arsenic from the pollution
source, and determining the location and efficiency of attenuation. The evaluations were carried out over a time
period that represents the pattern of seasonal change and weather events such as rainstorms.2' 22
Pollution from legacy mine waste usually comes from both point sources (mine adit, open pit, etc.) and diffuse
sources (contaminated soils). For pollution management, the preferred option is minimizing the formation of
diffuse sources by isolating the point sources from triggering events such as exposure to air or water. The mobility
of arsenic can be reduced by diverting water flow from both point and diffuse sources. A stable catchment is
needed to deposit the drainage from pollution sources, where passive or active remediation can be implemented/4
23 In all events, hydrogeological characteristics were assumed to be the controlling factor for arsenic mobility. An
inherent risk of this assumption is from the dynamics of local water discharge and variability of the water table. A
survey of the Ponderosa mine in the Iberian Pyrite Belt, Spain evaluated the relationship between rainfall and
pollutant discharge (kg/day) by the adit between Dec 2015 and Nov 2016.23 The maximum discharge of major
pollutants was around 15x above average, and maximum arsenic discharge was 16x above average. These results
emphasize the need for management infrastructure capacity to be adequate for extreme weather events.
Sampling is critical for site characterization and management because details of arsenic mobility depend on
sampling density and frequency. This need is yet to be met by current analytical capability. As speciation requires
expensive sample management and instrumental analysis in a lab. The process is slow because sample
management is an inefficient and difficult step for arsenic analysis. The current protocol typically involves in-
field filtration and sample preservation. Filtration and preservation eliminate microorganisms and slow down
chemical reactions during shipping.19,24 The samples are shipped under controlled conditions. Speciation
preservation has been a major challenge, because oxidation is a serious problem for samples from anoxic
groundwater25,26 To avoid the stringent requirements for sample preservation, in-field separation with ion
exchange (solid phase extraction) has been used to separate As(III) from As(V).'5 In this practice, the throughput
is still limited by sample management. Ideally, total As and speciation could be determined in field, with minimal
need for sample preparation, and with no need for specialized personnel or instruments. To meet the need for
frequent and high throughput sample analysis for high density site characterization, we have been developing a
sensor that can analyze As(III) on site. We also plan to use this sensor to determine total As.
| I I I ' | I'. I i III. < 'I l! s I i I < III >
We have developed a prototypic sensor for field determination of arsenic based on an enzyme inhibition
mechanism 27 As(III) is known to inhibit acetylcholine esterase (AChE). The degree of inhibition correlates to the
concentration of As(III), therefore AChE was selected as the bioreceptor. To make this occur in a controlled way,
we started with a transducer, and invented a reaction sequence to report AChE activity amperometrically (Figure
1). The fabrication of the sensor involves immobilization of AChE with bovine serum albumin (BSA) on the
working electrode by glutaraldehyde crosslinking. (Patent Application 16/793455 filed February 18, 2020).
2
-------
Reaction 1
Biochemical
reaction
Reaction 2
Electrochemical
reaction
Figure 1. The transduction reactions to monitor AChE activity
Early development included an immobilization method and a measurement protocol. The objectives in
immobilization included retaining maximal enzyme activity and its sensitivity to As(III), high operational stability
in different media, long storage stability, a streamlined protocol for efficient fabrication, and low variance
between individual sensors. The objectives for measurement protocol development included selective and
efficient signal transduction, high reproducibility between repeated assays, and compatibility with in-field
practice. The two developments were interdependent because enzyme activity, As(III) sensitivity, transduction
efficiency, and operational stability were all controlled by both immobilization method and measurement
protocol.
The protocol to determine As(III) is based on the inhibition mechanism. For the AChE sensor, the inhibition of
AChE by As(III) follows a pseudo-irreversible mechanism, involving a fast association step (k\ = 2.84 X103 M"1
min"1) and slow dissociation step (k.\ = 1.75 X 10"1 M"1) in 0.1 M Tris-HCl at pH 8.0. (Figure 2). Therefore, we
designed a two-step measurement, in which the activity assay and As(III) inhibition were carried out separately.
k, [As(lll)]
AChE - AChE-As (III)
k-1
Figure 2. Kinetic mechanism for AChE inhibition by As(III)
The activity of AChE sensor was reported by the steady state current with working potential at 0.35 V (Figure 3
A). The concentration of As(III) was determined by incubating the sensor in an environmental sample for 1 h at
20-24 °C (Figure 3B) . As(III) in the sample formed a kinetically stable complex with the AChE and inhibits the
enzyme. The remaining AChE activity was determined amperometrically in the next step. Matrix effect was
minimized because only As(III)-AChE complex was carried over from the environmental sample to the assay
solution.
XT'
Hydrolases
OAc (Bioreceptor)
HO-Cf + 'A°H
*»
pi
li.
«»
0^
3
-------
Figure 3. Activity measurement set up (A) and As(III) determination by AChE sensor inhibition (B)
AChE activity was reported as the steady state current in amperometric assay (Figure 4), where the rate of
electrochemical reaction equaled that of the biochemical reaction (Figure 1). The steady state current of the
sensor was measured before and after the incubation with a sample to give h and /, respectively. The inhibition
was calculated by Equation 1.
[
\
Steady State Current
i\
w
io
vV
v.
i i i i
0 20 40 60 SO 100 120
Time (s)
Figure 4. Profile of steady state currents before (io) and after inhibition (i,)
/ = Equation 1
The correlation of /with [As(III)] was consistent with the mechanism in Figure 2. There was a logarithmic
relationship in the range of 2 - 500 liM As(III). In a narrower range of 1- 20 |a,M, the correlation (Equation 2)
was linear (I % = 2.27 [As(III)] + 5.3, R2 = 0.93).
/% = a [As(III)] + h Equation 2
4
-------
As(V) had no impact on the assay of As(III). The speciation with the AChE sensor was validated by comparing
the results with those from ICP-AES analysis. The standard solutions contained As(V) at a fixed concentration of
5.0 |jM, and As(III) concentrations from 2-20 |o,M (Table 1). The correlation of the two methods was a straight
line with a slope of 0.9923, an intercept of 5.13 |oM, and the R2 = 0.9262. The slope suggested that the AChE
sensor reports As(III) with excellent accuracy. The intercept showed that the presence of As(V) did not interfere
the AChE sensor assay (Figure 5).
o
10
c
Cl)
U~l
u
<
>-
-Q
10
<
y = 0.9923x-5.1365
R2 = 0.9262
Total As by ICP-AES (nM)
Figure 5. Comparison of total As determination by ICP-AES and As(III) determination by AChE sensor
In addition, the validation also revealed problems of the sensor at lower concentrations. The recoveries of AChE
assay were 170% and 161% when [As(III)] were 2 |oM, 4 |oM respectively. Variances were larger at lower
concentrations. The recovery and variance were significantly improved when [As(III)] reached 8 |oM.
Table 1. Speciation composition and results by ICP-AES and AChE-Sensor
Test
As(V)
As(III)
ICP-AES
Total As
AChE-Sensor
Average As(III)
(pM)
(pM)
(pM)
Recovery (%)
(|_iM, n = 5)
Recovery (%)
1
5.0
0.0
5.20
104
-3.70 ±3.44
2
5.0
2.0
7.19
102
3.40 ±2.05
170
3
5.0
4.0
9.21
101
6.46 ±2.30
161
4
5.0
8.0
13.17
101
8.73 ± 1.01
109
5
5.0
14.0
19.33
101
14.11± 1.15
100
6
5.0
20.0
25.32
101
19.00+1.15
95
The sensor fabrication was systematically optimized and streamlined. One can make 100 sensors per batch with
reasonable consistency. In atypical batch, the range of in was 18.8 - 24.8 |oA, with Mean = 21.3 ± 1.6 |oA (N =
39) (Figure 6). The sensors were stable in 0.1 M Tris-HCl, pH 7.0 for at least 150 days.
5
-------
To prove this sensor would work in groundwater, we tested the sensor m a synthetic groundwater matrix. The
formula (Table 2) was based on the median values of the main chemicals in water samples from Fort Devens,
MA. 28 The AChE sensor showed similar sensitivity to As(III) in the matrix (I % = 2.22 [As(III)] + 6.11 „ /?2 =
0.95), indicating none of the inorganics in the matrix interfered with sensor sensitivity. This matrix, however, was
unstable under air due to the oxidation of Fe(II). The synthetic groundwater was prepared in deoxygenated water,
and used within 6 hrs. This experience confirmed that sample preparation for groundwater with high Fe(II) would
be critical. It is well known that oxygen exposure causes Fe(II) oxidation and precipitation of ferric
oxyhydroxides, and this event typically preceded As(III) oxidation to As(V) under oxic conditions. Since arsenic
precipitation and removal could be facilitated during Fe(II) oxidation/5,30 preventing ferric oxyhydroxide
precipitation was a primary objective.
Table 2. The formulation (in liM) of a synthetic groundwater based on the water chemistry in Shepley's Hill
Landfill in Fort Devens, MA
Salt
[cation]
COs2"
ci-
SO42-
Silicate PO43"
EeCk
612
lia
MnCb
83
83
CaCOa
1105
1105
K2CO3
213
106
MgSCU
256
256
NaHCOs
847
847
iNIIu-CO'
143
72
NazSiOs
400
400
NazHPCM
100
100
Total
2130
695
The prototypic biosensor would be feasible for in-field test as shown in the projected lifecycle (Figure 7). It
begins with the sensor fabrication in lab. Ideally, the sensor can be stored in dry conditions under ambient
temperature for 1-3 months so that a stock of the sensors is ready to deploy for field tests. There are two tasks in
6
-------
field (in the bracket) to prepare for the assay, including hydration of the electrodes and water sample preparation.
When sample preparation involves chemical modification, sensor performance should be re-evaluated.2 The
single-use sensor is disposed of as general waste.
0
Disposal:
Electrode, reagent,
k and sample
Figure 7. Projected life cycle of the disposable sensor
To justify the specifications of storage and hydration, we evaluated the impact of condition and time on sensor
activity and sensitivity to 8 |o,M As(III). Sensor consistency can be improved by simplifying and streamlining the
fabrication method. A narrow distribution of sensor activity was desirable, because we found sensor activity had a
small but significant impact on its sensitivity to As(III). Based on this consideration, we removed a Tris-curing
step in sensor fabrication because variance of activity increased significantly during curing. The sensor underwent
slow activity loss during dry storage or prolonged hydration at high pH and ambient temperatures. In general, the
sensor can be dry stored for > 7 days, and hydrations should be between 12- 72 h in buffers with pH < 7.0 at
ambient temperature. We have not evaluated the interaction between dry-storage condition and hydration need. In
practice, we discarded the sensors with activity of < 15 |oA because the RSD of As(III) inhibition increased.
3. Ferric oxyhydroxide formation and other concerns in sample
preparation
Analytical error may come from sample instability. Therefore, arsenic sample stability in a realistic matrix
becomes a critical challenge in sensor development.
For synthetic groundwater preparation, we vacuumed, and nitrogen-flushed the groundwater. Solid FeCh was
added as the last chemical. Still, iron oxidation/precipitation took place in several hours after occasional air
exposure in the experiment. The stability would be more difficult to control in field because groundwater would
experience oxygen exposure and temperature shift in sampling. The speciation would then change to reach new
equilibria/' In this change, inorganic arsenics may either precipitate with ferric oxyhydroxides or undergo redox
reaction.
The oxidation of Fe2+ by O2 in aqueous medium involves oxygenation to give ferric oxyhydroxide (Step 2, Figure
8), followed by dehydrations to give goethite (FeO(OH)) and then hematite (Fe2C>3). The oxidation of each mole
of Fe2+ also generates two moles of H+. In addition to dissolved oxygen, the Fe2+ oxidation depends strongly on
the pH.32 The ferrous iron is protonated as pH decreases: Fe(OH)2 is the dominant species in pH 5-8, Fe(OH)+ is
7
-------
the main species in pH 4-5, and almost all ferrous is Fe2+ when pH is < 4. The oxidation of Fe(OH)2 takes less
than a few seconds. The first order rate for oxidation of Fe2+ is 6 x 10 5 min"1, indicating the half-life at pH < 4 is
about 8 days. Ferric oxyhydroxide (Fe(OH)3), (Step 2, Figure 8) is poorly soluble in water (Ksp = 2.79X10"39). In
groundwater, ferric oxyhydroxide hydrates commonly present as suspended particles in the matrix with dynamic
solubility (pQ). The pQ is a function of several factors including [Fe2+], Eh, K„ (activity product of water), pH,
and temperature.33
4 Fe2* + 8H20 4 Fe(OH)2 + 8 H* 1
4 Fe(OH)2 * 02 + 2 H20 —- 4 Fe(OH)3 2
4 Fe(OH)3 4 FeO(OH) + 4 H20 3
4 FeO(OH) 2Fe203 + 2 H20 4
Overall reaction
4 Fe''' + O.. + 4 H.,0 2 Fe203 + 8 H'
Figure 8. The oxidation of Fe2+ through oxygenation and dehydration
The short exposure to oxygen in groundwater sampling essentially creates a circumstance similar to that at
anoxic-oxic boundary, where arsenic is sequestered by newly formed rust. Both As3+ and As5+ can be removed
from water by adsorption to ferrihydrite in < 2 h.30 The adsorption depends on arsenic speciation, ferric oxide
mineral forms, and environmental factors.34 The adsorption capacity is in the order of amorphous iron oxide »
goethite ~ magnetite. For arsenite, the adsorption is stronger at high to neutral pH (pH 7-8). For arsenate, the
adsorption is stronger at low pH (pH 5-6). When organic matter is present, the newly formed iron oxide may form
a colloid in the matrix. The formation and dissolution of the colloid depend on pH, Fe/C ratio (< 0.02-0.1), and
the nature of organics. The dynamics of the complex matrix will cause re-distribution of arsenic between aqueous
solution, colloid, and particles.35
Other considerations in speciation sampling are redox reaction of arsenic coupled with the interconversion of
Fe(III) and Fe(II) and redox reactions mediated by microorganisms. Species preservation focuses on slowing
down the oxidation reaction, stabilization of Fe3+, and microorganism removal.25 Acidification is the first line
option because it can slow the oxidation of Fe(II) and As(111) .7 The choice of acid depends on analytical method
and water matrix. For example, acidification with HNO3 may dissolve ferric particles but also promote As(III)
oxidation because HNO3 is an oxidant. Phosphoric acid can not only acidify the sample, but also form a complex
with Fe3+, although the K,_{ is relatively high (2.7 X10"4 M). EDTA has been used in species preservation because it
forms very stable complex with Fe3+, but the sample also needs to be acidified and preferably stored at 4 °C36
8
-------
4. Background of Shepley's Hill Landfill Site, the need for on-
site sensors
Fort Devens was built in north-central Massachusetts in 1917 as a US Army training camp. Shepley's Hill
Landfill is on the northeast corner of the former main post of Fort Devens. It has a total area of 84 acres. It used to
receive 6,500 tons year"1 of household refuse and constmction debris, with a portion buried below the water
table. A
/¦
Shepleys
Hill
Landfill Area
Red Cove
j Ayter ^
G W ARCHER ^ Q The Billiards Cafe
^ Central Ave
M Ayer (3 **
'Qir> St
Compassionate Care
Ayer District Court ^
^ Plow Shop Pond
Pirone Park
f
G<'
Q\ie Pond
;ns Recycling Center
Disc Golf ^ ©WestRock
Disc bolt ~ MB Kenney
Company, Inc
Figure 9. The location of Shepley's Hill Landfill and Red Cove of Plow Shop Pond
In 1989, the landfill was added to EPA's National Priorities List because arsenic contamination was found in the
groundwater. The groundwater flowed from southwest to northeast, expressing as surface water in Plow Shop
Pond (Figure 9). The groundwater is rich in Fe(II). As the groundwater was discharged to the pond, the Fe(II)
was oxidized to rust, and the area became known as Red Cove. Along with the rust formation, arsenic was
precipitated in the sediment in this area. The groundwater and surface pollution presented health threats for the
people in the town of Ayer through fishing, recreating activities or well water consumption.
The origin of arsenic was found to be mainly from the landfill leachate. The arsenic could come directly from the
landfill waste or be released by reductive dissolution of geogenic source:' The total groundwater flow was
estimated to be 3,800 ft3 day"1 (10.76 m3 day"1), with a mean As concentration of 430 ppb. Therefore, the total
discharge of arsenic from groundwater to Red Cove was estimated to be 17 Kg year"1 ,JS Several efforts were made
to mitigate the groundwater pollution. Capping the landfill increased the level of arsenic. In 2006, a pump-and-
treatment plant was installed to remediate groundwater via collection wells. In 2012, an 850 ft hydraulic barrier
wall was installed to channel the groundwater back to the landfill area. 39 Sediment excavation and treatment
started in 2013 at Red Cove. EPA's Region 1 office planned high-resolution groundwater monitoring to evaluate
9
-------
the efficacy of the installed pump and treat system. In addition, "simple and inexpensive tools" were expected to
play important roles in refining the conceptual site model.is
5. Groundwater samples from Red Cove Area, Shepley's Hill,
spatial and temporal variations of groundwater condition
and chemistry
To characterize the pattern of arsenic discharge to Red Cove via groundwater, groups of monitoring wells (RSK)
with different depths were installed in the vicinity of Red Cove (Figure 10). The distances between the wells
within a group was < 50 m. The groundwater was sampled at scheduled times to determine the water chemistry
and arsenic content. Temporary devices such as a push point device (PPT) or tube wells (TW) were used in
combination with GPS to increase the spatial density of sampling. Piezometers (PZ) were installed along the edge
of the pond to monitor the groundwater entering Red Cove. (Figure 10)"
RSK30 r 4 * _
RSK32 RSK16"21 PZ8
PZ3STAFF1
RSK29 RSK48 * \^] ,pzJ
RSK28
N3-P1.P2
RSK23
13
IRSK36-43I
RSK13-15jj
Figure 10. The distribution of monitoring wells (RSK) and Piezometers (PZ) in the Red Cove area
The survey from March 14, 2006 to September 12, 2007 involved collecting 53 samples on 12 different dates.25
Most samples from RSK wells were reductive, anoxic and had neutral pH. However, there were significant
variations in dissolved oxygen and ORP. (Table 3)
10
-------
Table 3. Condition of the groundwater samples from RSK 1-39 (Mar 2006-Sept 2007)
Parameter
Median
Average
Range
N
Max - Min
PH
6.53
6.53 ±0.03
5.87-6.97
53
1.1
ORP
(mV)
-105
-88 ±6.7
-156.3-43.8
41
200.1 mV
DO
(mg/L)
0.4
0.77 ±0.13
0-4.5
53
4.5 mg/L
In the pH range of 5.87-6.97, the shift of DO and ORP caused rapid interconversion of Fe(II) and Fe(III).
Metastable minerals such as green rust formed and played a significant role in arsenic transportation and natural
attenuation in groundwater/0,41 To characterize the variation of groundwater in terms of pH, DO, and ORP, the
groundwater samples were grouped into three clusters (Figure 11). Most samples (n=34) fell in the first cluster,
with low ORP, moderate pH, and low DO. In the second cluster (n=5), the samples had high DO, but ORP
remained negative. Only two (RSK3, 300ct07 and RSK15 12Sept07) fell in the third cluster, featuring high ORP
and low DO (Table 4).
O2 • '
i i
1 1 1 1 ' 1 ' 1 1 1 1 r
-2-101234
Prin 1
Figure 11. Clustering of groundwater samples based on pH, DO, and ORP (K=3)
Table 4. Cluster and mean values of physical condition for the RSK groundwater samples
Cluster
PH
ORP
DO
N
1
6.56
-102.01
0.42
34
2
6.73
-41.16
3.20
5
3
6.19
22.85
0.055
2
For Fe rich groundwater, water chemistry is dictated by the change of pH and redox conditions. To evaluate the
temporal and spatial variations of groundwater conditions, groundwater samples were designated by cluster
number, and plotted using well number and sampling date (Figure 11). Spatial variations within RSK of the same
location were found in RSK 8-12 (Mar 14, 2006), RSK13-15 (Sept 12, 2007), and RSK 16-21 (Mar 14, 2006).
Out of 9 surveys, a third (3) showed concentration variations over short distances (Green boxes in Figure 12).
Temporal variation was found in RSK 4, 8, 11, 12, 15, 17, and 20 (red boxes in Figure 12). Out of 20 sample
pairs, 7 were cluster-crossing, suggesting temporal variation was more common. To capture variations, sampling
11
-------
frequency should be increased until the temporal and spatial variation patterns stabilize. In our current data set,
the minimum time of cluster-crossover was about 13 months (well numbers 17 and 20). More frequent sampling
would help to correlate the change to the fluctuation of environmental conditions.
10/30/2007
•
•
Cluster
09/12/2007
•
l- •
•
09/11/2007
•
•
•
•
• 3
08/22/2007
08/21/2007
04/25/2007
04/24/2007
08/10/2006
• •
•
* •
08/08/2006
•
•
05/18/2006
•
• •
•
03/14/2006
•
• •
•
51
03/13/2006
•
cnor-f\jmT}-i/ivDr--cocnor--cocno*-rM
Well Number
Figure 12. Temporal (red box) and spatial (black box) variations of GSK samples in terms of pH, DO, and ORP
In general, arsenic was higher in wells RSK 1-7, 8-12, and 16-20 (~ 0.5 - 1.1 mg/L), and lower (< 0.4 mg/L) in
wells RSK 13-15 and 37-42, although ~ 0.5 mg/L arsenic was occasionally found in RSK 40 and 41 (08/22/2007)
(Figure 13). Based on the groundwater flux and concentration, arsenic was largely discharged into Red Cove via
RSK 8-12 area. The discharge from RSK 1-7 and 16-20 areas were less, due to lower groundwater flow.
Therefore, the center of the arsenic plume was expcted to be in the groundwater at RSK 8-12.
10/30/2007
9/12/2007
9/11/2007
8/22/2007
8/21/2007
8/10/2006
8/8/2006
5/18/2006
4/25/2007
4/24/2007
3/14/2006
As(mg/ml)
• • • • •
I
I
«- r- «- r- i-
Figure 13. Arsenic (mg/L) concentration in RSK wells (Mar 2006-Oct 2007)
12
-------
Chemical composition varied widely among different RSK samples, suggesting the groundwater matrix
underwent drastic changes when discharged to the surface water (Table 5). The biggest change was soluble Fe,
which was the most sensitive element to redox conditions. Samples with high Fe were from RSK 8-12. Low Fe
was found in samples from RSK 1-7 and RSK 37-42. Within RSK 8-12, the highest Fe was 81.5 mg/L in RSK 10
(1 lSept07), and lowest was 30 mg/L in RSK 8 (14Mar06) (Figure 14). The spatial variations of Fe within group
RSK 8-12 and RSK 16-20 appeared to be correlated to the redox condition, as shown in Figure 12. Although As
was high in samples from RSK 1-7, samples were highly reductive and low in DO (Cluster 1, Table 4) and Fe
(Figure 14). Apparently, the discharge of As from groundwater involved changes in the redox condition and Fe2+
oxidation.
Table 5. Chemical composition of the samples from RSK wells (Mar 2006-Oct 2007, high variance in bold)
Chemical
Median
(mg/L)
Average
(mg/L)
Range
(mg/L)
N
Max/Min
As
0.96
0.672 ± 0.037
0.007-1.10
53
157
Fe
34.6
39.5 ±2.67
0.02-81.5
53
4075
Mn
2.09
2.26 ±0.18
0.01-7.56
53
756
Ca
42.5
42.1 ±2.1
5.09-65.4
53
13
K
8.4
8.66 ±0.46
1.17-16
53
14
Mg
6.27
6.15 ±0.29
1.19-10.3
53
9
Na
18.8
17.2 ± 1.01
0.96-27
53
28
CI
22.0
20.0 ± 1.31
1.31-35.4
53
27
S04
8.64
8.27 ±0.72
0.36-16.2
45
45
NH3
3.15
4.31 ±0.45
0.03 -10.9
51
363
Alkalinity
213
216.8 ± 14.3
15-487
51
32.4
TOC
2.3
2.44 ±0.26
0.67-14.2
50
21
10/30/2007
9/12/2007
9/11/2007
8/22/2007
8/21/2007
8/10/2006
8/8/2006
5/18/2006
4/25/2007
• • #
• • •
Fe (mg/ml)
i
I
4/24/2007
3/14/2006
t-fNJfOTfi/ivDr^ooCTioi-rvJfo^i/ivDr-coCTi —
, ,
i/ii/ii/ii/ii/iwtflwwyyyyyyyyyyyyyyyy
ttuiiiliiitaBlccavii/iwvivii/iviviviwii/ivivii/ii/ivivi
ceKccKKKcctttttttfccccccccoccc
Figure 14. Fe in RSK wells (Mar 2006-Oct 2007)
13
-------
In the water matrix, the most dynamic elements were those sensitive to redox conditions, including Fe (4075x),
Mn (756x), and As (157x). Other unique chemicals underwent significant changes. These are NH3 (363x), TOC
(21x), SO4 (45x), and alkalinity (32.4x).
The analysis of the water matrix can be simplified because the concentrations of chemicals were dependent on
each other. Dimensionality reduction was carried out with Principal Component (PC) analysis, showing that 4
significant PCs could account for a total of 90.1% variance (Figure 15). To further simplify the evaluation, the
first two PCs were employed as they accounted for 75.8% of the total variation.
Number of Components
Figure 15. Scree plot of PC analysis with the water chemistry of RSK samples
0 Label variables
-2-10 1
Component 1 (47.7 %)
Figure 16. Biplot of principal component analysis for RSK water chemistry
Bi
14
-------
Biplot of the first two PCs with overlapping Loading Plot (Figure 16) generated several findings. It was clear that
there was a highly positive correlation between Fe and NH3. The correlation of Fe with As was positive but weak.
In addition, Fe also had negative correlations with Mn and SO4. The distribution of samples in the biplot appeared
to fall into three groups. The group in upper right quadrant was high in Mn and SO4, but low in Fe and NH3. The
two groups on the right had similar amounts of As. The group in the lower right quadrant was high in Fe, NH3 and
TOC but low in Mn and SO4. The group on the border of the two quadrants had a distinctive feature of low
alkalinity.
2-
1 -
« 0-
C
¦c
"¦ -1-
-2-
-3-
Figure 17. Clustering of groundwater samples based on water chemistry (K=3)
Prin 1
Clustering analysis with water chemistry confirmed the impression from the biplot (Figure 17). The three clusters
had the similar locations to those in the biplot. The cluster mean values of chemicals were consistent with the
chemical compositions from the biplot (Table 6).
Table 6. Cluster and mean values (mg/L) of chemicals for the RSK groundwater samples
Cluster
(N)
As
Fe
Mn
Ca
K
Mg
Na
CI
SO4
NH3
TOC
Alk
1(8)
0.76
55.0
2.17
43.0
10.87
6.22
18.64
24.8
4.06
7.04
2.41
293
2(12)
0.77
27.1
2.90
57.2
7.96
8.69
24.61
24.0
13.0
2.34
2.22
229
3(14)
0.35
26.1
2.73
24.1
4.34
3.50
6.68
6.52
7.44
1.45
1.32
117
15
-------
9/12/2007
9/11/2007
8/8/2006
8/22/2007
8/21/2007
| 8/10/2006
5/18/2006
4/25/2007
4/24/2007
3/14/2006
10/30/2007
Figure 18. Temporal and spatial variations of water matrix in RSK samples
Despite high temporal variations of redox conditions (Figure 12), the water matrices were quite stable for each
well. Amoung 10 pairs of repeat sampling, only RSK 7 underwent cluster crossover from Cluster 1 to Cluster 2
during 10Aug07 and 21Aug07 (Red box in Figure 18). The change involved decreasing of Fe (30%) and NH3
(28%), increasing of SC>4(156%), and moderate changes in Mn (13% decrease) and As (7% decrease). Spatial
variations were observed in RSK 1-7 (10Aug06), and RSK 8-12 (25Apr07). High As samples fell into Cluster 1
(mostly from RSK 8-12 and RSK 16-20), and Cluster 2 (mostly from RSK 1-7). Medium to low As samples fell
into Cluster 3 (mostly from RSK 37-42 and RSK 13-15). The water matrix of RSK 12-13 fell into Cluster 3.
Fe distribution appeared to be path dependent. The highest concentrations were found in RSK 8-12, although the
temporal variations were big in RSK 8, 11 and 12. Fe concentration in RSK 16-21 were highly variable, and the
corresponding As were high (0.5-1.0 mg/L). RSK 1-7 and RSK 13-15 had mid-range Fe, but the As
concentrations were quite different between the two groups. Only RSK 37, 38 and 39 had Fe in the concentrations
of 0.02, 4.94, and 5.93 mg/L, respectively. Interestingly, they also had the lowest As (0-0.1 mg/L) (Figure 19).
Higher As (0.2-0.6 mg/L) were found in RSK 40-42, and the corresponding Fe concentations were between 14.5 -
45 mg/L.
16
-------
80
70-
60-
f 50-
I
HI
40-
30-
20-
10
0
Figure 19. The distribution of As and Fe in groundwater samples
Fe and As exibited different mobility in groundwater. A field survey of a landfill site with As pollution in
southern Maine suggested that Fe was preferably removed based on the finding that the ratio of soluble As vs Fe
increased 60-fold when groundwater moved away from the landfill. 42 Fe(II) was the main redox buffering
chemical. It appeared to be sequestered by oxidation as the groundwater flowed to a more oxic environment.
Speciation analysis assisted interpreting the change of As/Fe ratio from contaminant source to the pond.
The groundwater samples from Red Cove Area in Shepley's Hill Landfill Site were characterized by high Fe
content. Of all 53 samples collected, most (50) had Fe between 12-81 mg L1, suggesting Fe was in the form of
Fe(II). Since Fe(II) is more sensitive to changing of physical conditions, the change of [Fetotai] may precede that of
[AStotal] •
6. Speciation change in sampling and current state of the art
for arsenic speciation preservation
Groundwater sampling involved several preparation steps prior to sample collection.26 These steps included
measuring water level, estimating well volume, and pumping well water untill the physical properties were
stabilized. The monitored properties in this study were temperature, electrical conductivity, pH, dissolved oxygen
and redox potential. It usually took 3-4 wellbore volumes to stabilize concentrations. In sample collection, the
water was exposed to air and light as it was pumped into a sample container. The exposure to oxygen depended
on the total area of the water in contact with air and the time to fill the container. The reactivity of oxygen with
the electron donors (Fe2+, arsenite, etc) depended on UV and many other physical, chemical and biological
factors.
-------
Fe(II) oxldaiton and Rust formation
4Fe2+ +0, + 6 H,0
[Fe{OH)3]
Ferric axyhy dioxide
4 FeOi'OHl + 8H* 1
Geothite
Fe(OH)3 + Fe~+ + otter anions
Green Rust
Arsenic (III) oxidation
JH ;A'.-0 . + O. 1
air^ei:cn? acid
2Fe'+ + HjAsOj + H;0
2H2AsO+* + M+
arsemc acid
hv
:fp:+ - h- v.o,+ ;ir
Adsorption
Arsenous acid + Rust Asflll) complexes
Arsenic acid + Rust —~ As(V) complexes
Arseotu/Arsemc Acids + Rust »- As|IH).'As(V) complexes
Figure 20. Groundwater reactions that can be triggered by oxygen exposure
Oxygen exposure may cause three types of reactions in arsenic polluted groundwater when Fe is the dominant
redox species (Figure 20). The oxidation of Fe generates oxides such as ferric oxyhydroxide, geothite, or green
rusts. These oxides (or rusts) usually have low solubilities and the solids can adsorb both As(III) and As(V) to
create different complexes.3a 31 4143 In addition, oxygen exposure may oxidize As(III). The ferric iron can also
oxidize As(III) when exposed to light/4 The overall process can be complex when exposed to multiple reactive
oxygen species (ROS). Oxygen activation by Fe2+ may generate all ROS between molecular oxygen and hydroxyl
anion, including superoxide, peroxide, and hydroxyl radicals/5, 46 Similarly, As(III) oxidation also generates
ROS/7 Each ROS can oxidize Fe(II) or As(III). These reactions interact with each other through ROS
interconversion along with pH decrease caused by oxidation. The situation is further complicated by the fact that
the oxidations can be catalyzed by activated iron oxide minerals or microorganisms/^5' Given the diversity of
water matrix types and physical conditions, it is not feasible to predict arsenic speciation change or loss to
precipitate caused by oxygen exoposure in the field. Indeed, speciation preservation of inorganic arsenic was
found to be more challenging than the speciation analysis.36
Since air exposure was inevitable, method development for species preservation focused on preventing ferric
mineral from precipitation and slowing down arsenic (III) oxidation. The most common speciation preservation
involved filtration followed by chemical stabilization.52-56 Certain samples needed to be cooled at 4 °C when the
two-step protocol was not adequate.26 57> 58 Chemical stabilizaion involved matrix modification with an acid or
EDTA (Table 7). Common candidates of acids were acetic acid, hydrochloric acid (HC1), and phosphoric acid.
The selection of acid and whether to include EDTA depended on the analysis method, because modification of
matrix might change the efficiency of sample transformation or interfere with detection.24 For hydride generation-
graphite furnance atomic absorption spectrometry (HG-GFAAS), HC1 was adequate, therefore EDTA was not
recommended.53 For inductively coupled plasma mass spectrometry (ICP-MS), HC1 caused inerference through
formation of 40Ar35Cl.5S The need for EDTA depended on the concentration of Fe plus other divalent cations. It
18
-------
was used as persevant for groundwater (1.25 mM) and acid mine drainage (12.5 mM) subsequent to in-line
filtration/6
Table 7. Technical options for preserving arsenic speciation
Technique
Objectives
Note
Sampling
Obtain the most representative sample of
groundwater from well
Minimizing exposure to air and light
Filtration
Remove microorganisms and ferric
oxyhydroxide
Pore size of filter can be significant, may cause
additional air exposure
In-situ cation removal
Remove iron and divalent cations
Suitable for acid mine drainage. Limited stability.
Sample container
Maintain sample integrity, compatible with
chemical modification and storage needs
Minimizing exposure air in sample collecting step,
impermeable to light and air during storage, low
As adsorption or desorption
Chemical Modification:
EDTA
Chelation of Fe2+ and Fe3+ and divalent cations
including Mn2+, Mg2+, Ca2+ to prevent ferric
oxyhydroxide formation
The need for EDTA depends on total di-valent
cations. EDTA and Fe2+ complex have different
kinetic pathways in oxidation.
Chemical Modification:
Acidification
Reduce the rate of Fe(II) oxidation, solubilizing
Fe(III), stabilizing As(III)
May interfere with the analysis
Refrigerating or freezing
Slowing down microbial activity, reducing
reaction rate
Solubility of minerals can be sensitive to
temperature. Cooling may cause irreversible
precipitation.
Typically, acid mine drainage is rich in iron (~ 500 mg/L) and many other cations. Preserving sample speciation is
particularly challenging, as the cations may form precipitate on HPLC column in species separation. Keeping the
cations in solution requires high concentraions of EDTA. Onsite sample cleaning becomes critical. In this
situation, a protocol was developed featuring cation removal with ion exchange prior to analysis/9,60 HC1 was
used to adjust the pH prior to the cleanup. The samples were shown to be stable for at least 48 h.
Filtration has been adopted as a standard step, usually prior to sample collection, to remove microorganisms and
particles including ferric oxyhydroxides. The standard protocol employed a 0.45 |a,m membrane. Sterilized
membranes of 0.1 or 0.2 |a,m were shown to be more effective in stabilizing the sample. 2558 Limiting air exposure
by reducing filtration time (< 10 min) was effective,25 but in-line filtration seemed to better serve this purpose.56
Raw water samples may have particulate or colloidal As in addition to soluble As. Particulate As was removed by
filtration with a 0.45 |a,m membrane. Colloidal As was separated from soluble As by filtration with a 0.02 |a,m
membrane. For total As determination, particulate As in raw water was digested with 5% HNO;,.~v Occasionally,
the amount of particles removed were different between 0.2 |a,m and 0.45 |a,m membranes, causing large
discrepancies in total As determination.55
The order of chemical modification and filtration showed significant impact with high iron samples. It was
observed that chemical stabilization prior to filtration provided improved As results.55 For samples with moderate
iron concentrations (< 5 mg/L), preservation involved EDTA/HAc modification, and no filtration was used.36 56 61
Groundwater exposure to air and light causes reactions, as illustrated in Figure 20. The impacts are challenging to
quantify since they vary greatly. The unrecognized difference in sampling practice may contribute to the wide
discrepancies in literature. Therefore it is desirable to design a simple, streamlined protocol. Ideally, proper
selection of a container plus chemical modifier would be adequate for preserving arsenic speciation. The container
should be non-permeable to air and light, and not leach or adsorb arsenic, and filled to minimize headspace in
sample collection 25,54 When necessary, additional measures can be included in the sample
collection/preservation procedure (Table 7).
19
-------
7. U ih, „ vii.. v hlf i | i i < ¦ || < ¦ i vii in i i vii!| I- ¦ ill iii
Red Cove Area, Shepley's Hill Landfill, a proposal
Based on water matrix analysis in Section 5 and iron chemistry in Section 6, the main risk of sample stability
involves the rust formation and generation of ROS. The former risk depends on the concentration of Fe because
only a small amount of oxygen is required to induce rust formation. The later risk mainly depends on the exposure
to air and light since Fe was found in all the samples. Microbial activity and prescence of colloidal As cannot be
ruled out although they have not been documented for the samples from Shepley's Hill site.
To preserve anoxic samples with high [Fe], the first priority is preventing precipitation of iron oxide. This was
achieved by slowing down Fe2+ oxidation and chelation of Fe3+. The widely-used method invovled modifying a
water matrix with 87 mM acetic acid and 1.34 mM EDTA. This method was developed with samples containing
0.016-1.78 mg/L Fe.36 The receipe has been successfully used to preserve groundwater samples with 0.17-4.2
mg/L of Fe(II),62 synthetic goundwater with 3 mg/L of Fc(II)/' and groundwater with 0-20.36 mg/L Fe(II) 26 HC1
(24 mM) was used in combination with 10 mM EDTA to preserve groundwater samples with 0-14 mg/L Fe.57
EDTA was necessary when tested in combination with HNO3, HC1, Formic acid, or acetic acid to preserve
speciation of different samples, including groundwater with 2.3 mg/L Fe/4 In most cases, samples were stored at
low temperatures26,36 H57 Apparently, chelation and pH needed to be recalibrated as the iron concentrations were
much higher in samples from Shepley's Hill (Figure 19).
As the main redox buffer, Fe(II) may play a complex role in the stability of arsenic speciation. It may quench the
finite amount of oxygen absorbed in sampling. Assuming oxygen reaches 0.25 mM (8 mg/L) in samples due to air
exposure, 1 mM (or 56 mg/L) of Fe2+ is needed to quench the oxygen. For groundwater from Shepley's Hill, the
dissolved oxygen was 0 - 4.5 mg/L (Table 3), therefore the need for Fe2+ would be 0-31.5 mg/L. On the other
hand, the reduction of oxygen always involves the formation of ROS. The reactivity of ROS with As(III) in
relation to Fe(II) depends on the matrix. In addition, iron chemistry changes when forming complexes with
different ligands/3 It is necessary to understand the relative reactivity of Fe(II) in a practical range of conditions
to justify a speciation preservation method.
The autoxidation of Fe2+ was enhanced by chelation with organic ligands.64 These ligands were usually oxygen
rich chelators such as EDTA, oxalic acid, and citric acid. At pH 7.0, the oxidation rates for the complexes were in
the order of EDTA» citrate>oxalate> control. However, the molar ratio of Fe2+consumption to the quenched
oxygen was in the order of oxalic acid (3.4), citrate (2.5), and EDTA (2.1). The amounts of unaccountable
oxidants (ROS) were consistent with the order by radical trapping reactions. In selective oxygen quenching, the
consumption ratio of Fe2+/02 needs to be close to 4.
Importantly, the optimal pH to oxidize ferrous EDTA chelate is acidic. At pH around 3, the reaction was too rapid
to monitor with conventional spectroscopy (time scale in s)/'5 This feature was consistent with the need to slow
the oxidation of As(III) under acidic conditions. A study with stopped-flow and rapid scan spectroscopy
suggested that the process involved a multistep mechanism (Figure 21). 66 The key step involved intramolecular
electron transfer (Step 3, Figure 21). The formation of the oxygen complex required protonation and a molecule
geometry change (Step 1, Figure 21) as the first step. The optimal pH for the protonation was between pH 2.5-3.
In the study, excess amount of EDTA was used so that Fe(II) was completely chelated. The ferric product
remained in solution as EDTA chelate. The product was highly stable because ferric EDTA complex has a K,_\ =
10 24 M, while the ferrous EDTA complex has a KL\ = 10"14 M. 67 EDTA did not play a catalytic role in Fe(II)
oxidation because the dissociation of ferric EDTA complex was slow.65 It was oberved that when [EDTA] <
[Fe(II)], the oxidation began with a fast phase and then entered a slow phase, indicating two reaction mechanisms
were involved.69 Ferric hydroxide precipitate formed eventually.
20
-------
K\
[Fen(edta) H20]2" + H4" ¦. " [Fen(edtaH) H.O]- 1
MCP PB
kj
[Fen(edtaH) H20]" + 02 * [Fen(edtaH) 02]" + H20 2
k. j
[Fen(edtaH) Ch]" ¦¦ [Fenl(cdtaH) (02 )] 3
[Fem(edtaH) (CK)]' + [Fe"(edtaH) H:0]' ^ [(edtaH) Feln(022-)(edtaH) Fera]2 + H20 4
[(edtaH) Fem(022-)(edtaH) Fenl]2-+ 4H20 » 2 [Fem(edtaH) H20] - H202 - 2OH' 5
2 [Fen(edtaH) H20]" + H202 + 2Hh 2 [Fem(edtaH) HiO] + 2H20 6
Figure 21. Oxidation of [Fe11 EDTA] under acidic condition
Oxalic acid is widely found in nature. It is the smallest diacid with pKa values of 1.25 and 4.14 (Table 8).
Although oxalic acid has strong buffering capacity between pH 1-4, it has never been tested in combination with
EDTA for preserving As species. Oxalic acid itself is a chelator for both Fe3+ and Fe2+. 0 Similar to EDTA, the
Fe3+ complexes are more stable. The speciation depends on pH and oxalate concentration. In 10 mM oxalalic acid,
Fe2+ started to form [Fen(C204)2]2" when pH increased from 2 to 3. In 100 mM oxalic acid, this transition
happened at pH 1.5. For Fe3+, the species included [FemHC204]2+, [Fem(C204)2]", and [Fem(C204)3]3". The
transition from 1 to 3 ligands in the complex took place from pH 1.0 to 3.0 when oxalate concentration was 10
mM. This transition shifted to pH 1.5 when oxalate was 100 mM, and pH 0.8 when oxalate was 1 M. Apparently,
oxalate should not only acidify, but also stabilize the Fe3+ of a groundwater sample. The iron contents were from
0.3 |o,M to 1.46 mM (0.02-81.5 mg/L) in the samples from Shepley's Hill (Table 5). It required less than 4.4 mM
(3 x 1.46 mM) of oxalic acid to keep all the iron in solution. Other divalent cations may form complexes with
oxalic acid. Among them, calcium stands out because the solubility of calcium oxalate is low (0.2 - 0.59 mM in
the pH range of 2.2-5.9). '' Since the average concentration of Ca2+ was 1.05 ± 0.05 mM in the samples (Table 5),
arsenic loss in calcium oxalate precipitate was a concern.
Table 8. Equilibrium constants of oxalic acid with H+, Fe3+, Fe2+ and Ca2+ 0
Reaction #
Reaction
Equilibrium constant
1
H2C2O4 = H+ + HC2O4-
Kai = 5.6 x 10"2 (pKai = 1.25)
2
HC2O4- = H+ + C2O42-
A'ai = 6.2 xl0-5(pKa2 = 4.14)
3
fFemC204l+ = Fe3+ + C2O42-
Kd= 3.98 x 10"10 M
4
[Fem(C204)2]- = [FemC204]+ + C2O42-
Kd = 6.31 x 10"21 M
5
[Fenl(C204)3]3- = [Fem(C204)2]-+ C2O42-
Kd = 3x 10"21 M
6
Fe(C204)(s) = Fe2+ + C2O42"
Ksp = 2 x 10"7 M2 (solubility 0.97 g/L 25 °C)
7
[Fen(C204)2]2- = Fe2+ + 2C2O42-
Kd= 2 x 10"8 M2
8
[Fen(C204)3l4-= Fe(C204)22- + C2O42-
Kd= 6.6 x 10"6M
21
-------
For sample preservation, oxalate may prevent ferric precipitation by dissolving ferric oxyhydroxide in particulate
or colloidal arsenic formed after air exposure. Oxalic acid has been used to dissolve rust for metal cleaning. 2 The
mechanisms for iron oxide dissolution involved surface protonation, oxalate complex formation, or reductive
dissolution. 3' 4 The dissolution required acidic conditions, and was catalyzed by Fe2+ or light at ambient
temperature.
When light was excluded, the main impacts of adding oxalic acid were acidification, chelation, and iron oxide
dissolution. At pH 7.0, oxalate increased the rate of Fe autoxidation.w At lower pH, the reactive species
([Fen(C204)2]2") decreased. In 20 mM oxalate, the autoxidation decreased by 8 fold as the pH decreased from 4.75
to 2.5. 5 Apparently, adding < 100 mM oxalic acid to the groundwater samples slowed Fe autoxidation because
the pH was around 2.0 and the chelation with Fe2+ was not significant. If oxygen removal with Fe(II) is desirable,
adding a catalytic amount of EDTA may help.
Under certain conditions, oxalic acid is a reductant. When oxalate-Fe(III) solution was exposed to natural light,
the oxalate ligand was decomposed in [Fem(C204)3]3" and Fe3+ was reduced to give Fe2 . 6 One product of oxalate
decomposition is carbon dioxide radical anion. It is a strong reductant that can activate oxygen to form
superoxide, which is transformed to H2O2 through dismutation and then hydroxyl radical by Fenton reaction
(Figure 22). Overall, one mole of oxalic acid is decomposed to two moles of CO2 by 0.5 mole of oxygen. The
product of the oxygen reduction is one mole of hydroxyl radical. The hydroxyl radical oxidized As(III) to As(IV),
which was further oxidized to As(V) by oxygen. However, the hydroxyl free radical may not be a serious issue
because it can be quenched by oxalic acid, or by additional chemicals such as MeOH or iPrOH, or Tris./& 9
Photolysis of ferric oxalate was significant in pH 1-6 and dependent on many other factors including the
concentration and speciation of Fe, the light wavelength, and sample exposure to light.50 The mechanism
involved a fast intra-molecular electron transfer, giving Fe2+, CO2, and carbon dioxide radical anions as
products.53 The efficiency of the reaction was a function of specific absorption coefficient and quantum yield at
specific wavelengths. [Fem(C204)2]~ and [Fem(C204)3]3" had similar activity at 313 nm. At 436 nm, photo
reactivity decreased 33-fold for [Fem(C204)2]~, and 122-fold for [FeIII(C204)3]3~.'w Ordinary glass bottles would
only slow down the photolysis by ~ 5-fold, as clear glass can only filter UVB light.
FemOx/
Oxidant
Oxalate
FenOx-r~
Net result
C2of + 03 + H+
CO, + cov
CO,
02
o,"
H1"
Superoxide
Dismutation
0.5 H1O-, + 0.5 O,
t-
FeL
Fe
hi
0.5 OH* +05 OH"
Fenton Reaction
2 CO2 + 0.5 O2 + 0 5 OH"+ 0.5 OH"
Figure 22. Oxidative decomposition of oxalic acid catalyzed by iron (OX, oxalic acid anion ligand). The reactive
oxidative species are highlighted in red in the scheme.
Choice of acid has been a concern for instrumental speciation. For example, HC1 caused isobaric interference in
ICP-MS by forming ^Ar^Cl or 4"Ca35Cl in ICP-MS,5,v but it was preferred for samples analyzed with HG-AAS:5'
22
-------
nitric acid was used to digest particulate arsenic, but it caused interference with the reduction in hydride
generation. 24 Similarly, the impacts of oxalic acid on enzyme sensor and the reference method need to be
evaluated.
Method development for As speciation preservation involves evaluating the impact of oxalic acid with/without
EDTA in groundwater matrices. The vairables are Fe, As, pH, and dissolved oxygen. The impact is evaluated in
terms of As speciation, oxygen consumption, Fe speciation, stoichiometry of Fe(II) oxidation, pH, and oxalic acid
consumption. The objective is to understand stoichiometry relationship and kinetics of the redox reactions.
A simple, streamlined speciation preservation is valuable, because it can create a pause between sampling and
analysis (Figure 23). Multiple samples can be collected and preserved in one field trip. Then, analysis can happen
on-site a different day.
=>
Field sampling
day
Pause
=>
Analysis day
Figure 23. Conceptual flow chart for arsenic assay with the enzyme sensor
8. Pre-reduction of As(V) to As(lll) for total As determination:
current state of the art
Measuring total As is a highly desirable objective for a field sensor, because total As is used in the regulation for
water discharge (Table 9). The standard depends on the time-intensity of a pollution. For freshwater discharge,
the total As is 340 |ag/L, and 150 |a,g/L in acute and chronic situations, respectively. The AChE sensor can be a
promising tool for compliance application, because the current detection limit is around 150 ppb of As(III). Once
a simple pre-reduction method for As(V) to As(III) is available, the AChE sensor can be used for compliance
applications. Several other enzymes (e.g., acid phosphatase) have been developed to determine As(V), but they
suffered from performance issues including stability or selectivity.55 The strategy to measure total As with the
AChE sensor appears to have lower risk if As (V) can be reduced to As(III) in sample preparation.
23
-------
Table 9. National Recommended Water Quality Criteria for As - EPA recommendation for water quality in terms
of pollutants
Pollutant
(P = Priority
Pollutant)
CAS
Freshwater
CMC
(acute)
(irg/L)
Freshwater
CCC
(chronic)
(irg/L)
Saltwater
CMC
(acute)
(irg/L)
Saltwater
CCC
(chronic)
(irg/L)
Publication
Year
Arsenic
7440-38-2
340
150
69
36
1995
*CMC: Criterion Maximum Concentration, CCC: Criterion Continuous Concentration; ** This recommended water quality criterion was
derived from data for arsenic (III) but is applied here to total arsenic; *** Freshwater and saltwater criteria for metals are expressed in
terms of the dissolved metal in the water column.
Reductants for As(V) to As(III) have been developed mostly as the pre-reduction step for sample preparation in
hydride generation - atomic spectroscopy analysis (HG-AAS, HG-AES, etc)56 In hydride generation, As (III) or
As(V) is reduced to arsine, a gas that is separated from the liquid matrix and detected by flame photometric
detection (Figure 24, Reaction 2).
Reductions of As(III) with NaBH4 are efficient as long as the pH is < 5, while As(V) reduction is slow and
requires 2 M HC1 to completely protonate arsenate. The reactivity difference was as high as 220-fold, therefore
As(III) was analyzed selectively. 87 To determine total As with HG method with NaBEU reduction, As(V) needs to
be pre-reduced to As(III). Many reductants have been used for pre-reduction; among them, thiochemicals are
popular (Table 10). However, pre-reduction caused problems in arsine generation. Many side reactions took place
in the mixture of pre-reduced sample and NaBEU (Figure 24): NaBEU reacted with water in acidic medium; both
borohydride and arsenite formed complexes with thiochemicals, etc. The overall outcome of the reactions was
hard to predict.57"59 Therefore a custom pre-reduction needs to be developed for specific sample analysis.
2 RSH + AsvCO){OH)3 RS SR + AsIM(OH)3 + H20
1
As(OH)j + 3 NaBH4+ 3 H+ * AsH3 + 3 BH3 + 3 H20 + 3 Na*
2
lit l3 + 3 H20 —- H3B03 + 3H2
3
NaBH, * H30+ —- H2Q-BH3 + H2 + Na+
4
NaBHj + RSlV * RSH lit h ~ H2 » Na+
5
NaBH4 + RS-SR -—RSBH3Na > RSH
i
As(OH)3 + RSH — (RS)nAs(OH)3.n + H^Ori
7
Figure 24. Reactions involved in thiol-based pre-reduction of As(V) to As(III)
In general, the development involves optimization of the reactants and reaction conditions, evaluating matrix
impacts, and validating the reaction with real samples. The main criterion is to equalize the signals between As(V)
and As(III) of equal concentrations. Since the pre-reduction (Figure 24 Reaction 1) is coupled to arsine
generation (Figure 24 Reaction 2) in method development, the concentrations of pre-reductant and acid are
24
-------
related to the concentrations NaBEU and NaOH (for NaBEU stabilization in working solution). Pre-reduction is
usually carried out in batch or stop-flow mode as the reaction condition is different from that of arsine reactions.
When the reduction of As(V) is kinetically comparable to that of As(III), flow reaction is feasible. Cystine was
preferred over potassium iodide (KI) or ascorbic acid because it reduced As(V) at low acid concentrations.
Cystine was also able to reduce interference from other inorganic elements in HG-AAS or AES.90,91 For field
application, cysteine is a promising pre-reductant, but the reaction can be sluggish at room temperature. Water
bath or microwave heating was used to facilitate the pre-reduction. 9ft 92,93 Thioglycolic acid92 and thiosulfate 94,95
were more efficient at ambient temperature, and are therefore attractive candidates for field pre-reduction.
Table 10. Variables for As(V) Pre-reduction in selected examples
Pre-Reductant
Acid
NaBELt/NaOH
Temperature
Time
Matrix impact
Reference
Ascorbic acid
0.5-10%; KI 0.3-3%
HC1
0.1 M
0.4-0.6%/0.5%
NC
5-60 min
Drinking water
Not studied
Stanic 20099tf
Potassium iodide (KI)
(30%)
HC1 (12 M)
1.0%)/ not given
NC
Stop 5-60 s
Removed online by
cation exchange
Tyson 1992s7
Ascorbic acid 0.5-4%,
KI 0.2-2%
HC1 (4 M)
1.5%/0.05M
110-150 °C
< 3 s
Recovery reduced
by Cu(II) Co(II)
Se(IV) Ni(II)
Nielsen 199798
Ascorbic acid
0.1-10%; KI 0.1-1.0%
HC1 (1.5-4.7 M)
0.1-0.3%/not
given
NC
Batch 15 min
Cys used to
suppress Fe
Naykki 2001"
Na2S2O30.1 M and 0.01
MNa2S204
HC1 (0.1 M)
NA
(ES-MS)
RT
5 min
As(III) extracted
into organic by
forming complex
with pyrrolidine
dithiocarbamate
Minakata
200994-95, Lai
2006™
L-Cys 1-20 gL"1
HC1 orHNOs
0.01-0.1 M
20 g L"1/ 8 g L"1
RT to 100 °C
5-140 min
5 g L"1 L-Cys
masked 14
interfering elements
Chen 1992s0
L-Cys 0.2-1%
HC1 0-0.05 M
0.2-l%/0.2-l%
90 ± 5 °C
5 - 100 min
Not evaluated
Cordos 200693
L-Cys 1%
HC1 0.01-0.05
M
0.2-0.8%/0.05%
NC
1 h
Urine sample
Guo \997WI
L-Cys 20 g L"1
HC1 0.01-0.1 M
4%/0
RT or
microwave
>15 min
L-Cys masked 8
interfering elements
Howard 1996s"
L-Cys 2%
HNOs 0.02-0.1
M
2%/0.5%
NC
>15 min
Ground water
acidified (pH 1-1.3)
with HNO3
Shraim 2008J°2
L-Cys (5 %)
HC1 (0.001-0.2
M
l%/0.5
NC
10 min
Mine wastewater
treated and
untreated
Shraim 1999s8
L-Cys (0 -1 %)
HC1 0.1-2.5 M
3%/l%
RT
1 h
Soil and tobacco
samples digested
with HNO3 and
HCIO4 (4:1)
Wietestka
2003s9
Thioglycolic acid (5 %)
HC10.05M
4%/0
RT
< 1 min
Not evaluated
Howard 199892
*For flow injection analysis, the concentrations are those for working solutions before mixing in arsine reaction step.
Strong acids like HC1 and nitric acid are the most common options to acidify matrices for As(V) pre-reduction
because the pKai for arsenate is 2.19. Acetic acid was used for As(III) reduction.57 Oxalic acid has apKai at 1.25.
It is stronger than phosphoric acid (pKai = 2.16), and therefore should be adequate for As(V) pre-reduction.
As(V) pre-reduction was developed in the context of AES or AAS. The As(III) was not isolated from the matrix
for subsequent arsine generation reactions. The pre-reduction efficiency was monitored as AsH3 indirectly by
25
-------
elemental spectroscopies. The impacts of As(III) reduction and atomization of AsH3 were compounded in the
results. For accurate analysis in pre-reduction development, a direct analysis of As(III) is desirable.95
,N I ¦ mi, ihe " lill <'ii s i hiiIII ,i i1111 \ * >.!<-I vii!| IN mi Mii
I 1 ' ' ' -I ' ' v II II 'I :k> , II [ill
The sensors were evaluated with groundwater samples from Red Cove area, Shepley's Hill. Field samples were
collected in clear glass bottles and shipped to Cincinnati overnight. At AWBERC, each sample was split into two
parts. One was tested for As (III) with the enzyme sensor within 2-6 days. The other part was tested with ICP-
AES for total As by the ORD/CESER contractor (Pegasus Technical Services Inc.).
The sensors were fabricated prior to the field sample test. Each batch preparation consisted of 100 sensors
prepared over a 2 day period following a protocol. The quality of the batch was evaluated by testing the activity
with 5 electrodes (5% batch test). Once the batch passed quality assurance checks, the sensors were stored under
dry conditions at room temperature for 6-19 days. They were hydrated for 16-48 h before analysis. This practice
was designed to evalaute the feasibility of field deployment. The sensors are to be sent to users via mail under dry
conditions. The users need a minimum shelf life of one week before the analysis. For users, the preparation of the
sensor for analysis involves hydration in 0.1 M Tris-HCl buffer (pH 7.0) at room temperature for > 16 h
(overnight).
Two field sample tests were carried out with objectives for field application. The test in May 2019 involved
analyzing 4 samples to evaluate the compatibility of the sensor with groundwater matrices and oxalic acid as a
matrix modifier. The test in October 2019 involved analyzing 24 samples to evaluate the sensor in a wider range
of matrices, and evaluating speciation preservation with oxalic acid. Issues with field use compatibility and data
quality have been identified from the test data, and will be used to select additional objectives for future sensor
research and development.
'' ii I i i ill ' ii rll-! ¦ > ii111|¦ II¦"s arrived on 11 i 11 i ''
The first batch of field samples were collected and added to sample bottles that were filled to the rim to minimize
oxygen exposure. The samples were chilled and sent to the EPA's ORD lab in Cincinnati by overnight delivery.
There were four samples in the batch. Two were from a well at the As source and the other two were from a
piezometer in the pond, representing groundwater and surface water, respectively (Table 11). All samples were
visibly clear when received. The groundwater samples had a smell of sulfur, suggesting As was dominated by the
As(III) species.
The objective of this test is to establish the correlation between sensor and spectroscopic method in groundwater
matrices. Additional As(III) (in 5 |oM) was added to each sample, creating 4 more samples for comparison. Two
lab samples (#9, 10, Table 11) were also included in the evaluation as a matrix-free reference.
26
-------
Table 11. Samples in the test of May 2019
Sample #
Sample Name
Description
1
SHP-2016-6B-1
From a groundwater monitor well (SHP-2016-6B) installed in
bedrock (100 ft) collected at the beginning of the pumping process
2
SHP-2016-6B-1A
Sample #1 with 5 |_iM additional As(III)
3
SHP-2016-6B-3
From a ground monitor well (SHP-2016-6B) installed in bedrock
(100 ft) collected at the end of the pumping process
4
SHP-2016-6B-3A
Sample #3 with 5 |_iM additional As(III)
5
PZ5 DSW
Deep Surface Water sample from PZ5
6
PZ5-DSW-A
Sample #5 with 5 |_iM additional As(III)
7
PZ5- SSW
Shallow Surface Water sample from PZ5
8
PZ5-SSW-A
Sample #7 with 5 |_iM additional As(III)
9
Lab reference 8 |_iM
Lab reference in 0.1M Tris-Oxalate, pH 8
10
Lab reference 20 |_iM
Lab reference in 0.1M Tris-Oxalate, pH 8
Several tasks were carried out prior to the test (Table 12). The sensors were fabricated on April 24, 2019 so that
there was time to evaluate the impact of storage and hydration in a buffer. The sensors were stable in dry
conditions for up to 19 days. Hydration in Oxalate-Tris buffer was pH sensitive. When pH was at 8, activity loss
was observed in 3 days. This loss slowed down significantly at pH 7.0. The average rate of activity loss at pH 7.0
was estimated to be 0.38 |o,A per day. In theory, the activity loss does not change the accuracy of inhibition
measurement but should reduce the precision. Therefore, sensors with low activities (Ao < 15 |oA) were rejected.
A standard curve was prepared prior to the field sample analysis. The sample preparation involves adding oxalic
acid to 50 ml of the field sample followed by pH adjustment with Tris. If the enzyme sensor proved to be working
in the sample, oxalic acid was added to groundwater sample in-field as a speciation stabilizer. Sample preparation
involved pH adjustment with Tris.
Table 12. Preparative work for sample analysis with the AChE sensor
Task #
Task
Description
1
Sensor fabrication
and storage
Sensors fabricated on April 24, 2019. Dry stored for 9 day s and hydrated in pH 7.0
Tris-Oxalate buffer for 4 days. Quality check: Ao = 22.8 ± 0.54 |iA .
2
Sensor preparation
for analysis
Dry stored for 19 days and hy drated in Tris-Oxalate buffer pH 7 (0.1 M) for 1-8
days. Sensor activity had significant loss over 8-day hydration (Ao = 16.45 ±1.45 |_iA,
Day 8). However, they were still used in the assays if Ao > 15 |_iA.
3
Standard curve
1% = 1.85 [As(III)] + 4, R2 = 0.9308. /in %, [As(III)] in |iM
4
Sample preparation
Add 0.29 g of oxalic acid to 50 ml of a field sample, then adjust the pH to 8.0 with
Tris base. The estimated amount is 0.873 g/50 ml. Final concentration for the buffer
is oxalic acid = 46 mM, Tris = 144 mM. The samples were prepared prior to the
assay.
Each sample was analyzed in 5 repetitions with the sensor. The average concentration was validated with the
corresponding result of ICP-AES analysis (Table 13). Based on the indication that the groundwater matrix was
sulfidic we expected that the As was all As(III), hence the recovery should be 100%. The result showed that the
errors of recovery were concentration dependent. Out of 10 samples, 6 had recovery between 94-118%. Their
concentrations were between 0.527 - 1.29 mg/L (8-20 |oM). For surface water sample #5 and 7, the arsenic
concentrations were close to zero by ICP-AES. Sample #6 and #8 had concentration around 4 |oM by ICP-AES
(actual values should be 5 |oM). For #5 and 7, corresponding results by the sensor were similar. For #6 and 8,
sensor results were both 7.0 mM, representing recoveries of 167 and 174%, respectively (based on ICP-AES).
27
-------
Table 13. Recovery of the enzyme sensor analysis with ICP-AES as the reference
Sample
#
Sample
ICP-AES
(mg L"1)
AcliE sensor
(MM)
StdDev
(|_iM, n=5)
RSD
(%)
AcliE
sensor
(mg L"1)
Recovery
%
1
SHP-2016-6B-1
0.7261
9.8
1.81
18
0.734
101
2
SHP-2016-6B-1 -A
1.045
13.1
1.87
14
0.982
94
3
SHP-2016-6B-3
0.9031
14.2
2.29
16
1.064
118
4
SHP-2016-6B-3-A
1.222
17.4
2.66
15
1.300
106
5
PZ5-SSW
<0.005
-1.0
2.68
-258
-0.078
n/a
6
PZ5-SSW-A
0.3143
7.0
2.19
31
0.524
167
7
PZ5-DSW
0.0095
-0.59
2.23
-378
-0.044
-463
8
PZ5-DSW-A
0.3008
7.0
2.89
41
0.526
174
9
Lab Ref 8 |_im
0.5272
8.0
2.07 (n=3)
26
0.599
113
10
Lab Ref 20 |_im
1.292
20
2.17(n=3)
11
1.498
115
The overall accuracy of the sensor was evaluated by linear regression with sensor results as a function of the
concentrations obtained by ICP-AES (Figure 25). The regression equation was [As(III)] sensor 1• 07 [As] ICP-AES +
0.034 with R2 = 0.9503, demonstrating close correlation between the two methods. The slope indicated that sensor
results were overall 7% higher than ICP-AES results.
1.5
H 1 1 1 1 1 1
0 0.5 1 1.5
ICP-AES
Figure 25. Correlation of the average [As(III)] of sensor analysis and total As by ICP- AES (mg/L)
The water matrices of the samples from SHP-6B-1 or PZ5-SSW were atypical compared with those reported
earlier.25 Groundwater sample SHP-201606B-1 had high S and high Ca (Table 14). None of them seemed to
interfere with sensor analysis. There was no loss of As to calcium oxalate in high Ca samples. None of the
samples had high Fe, although rust formation took place in non-acidified PZ5-DSW. The variation between
samples from the same location on the same day suggested that water matrix depends on sampling practice.
28
-------
Table 14. Main elements in the water matrix for first field sample test (ICP-AES, mg/L)
SAMPLE
As
Ca
Fe
Mg
Mn
Na
S
SHP-2016-6B-1
0.726
44.2
0.42
4.723
1.34
161.2
89.59
SHP-2016-6B-1 -A
1.045
44.75
0.409
4.781
1.33
162.2
88.37
SHP-2016-6B-3
0.903
23.84
0.052
2.677
0.318
38.51
9.033
SHP-2016-6B-3-A
1.222
24.21
0.06
2.707
0.323
39.18
9.165
PZ5-SSW
<0.005
11.9
0.338
2.048
0.049
35.93
1.914
PZ5-SSW-A
0.314
12.05
0.324
2.057
0.05
35.84
1.895
PZ5-DSW
0.01
27.6
9.77
3.565
1.12
21.78
0.959
PZ5-DSW-A
0.301
27.31
9.9
3.532
1.1
21.96
1.039
Lab Ref 8um
0.527
0.556
<0.037
<0.134
< 0.007
<0.191
1.85
Lab Ref 20um
1.292
0.974
<0.037
<0.134
< 0.007
0.547
1.873
I ¦ ¦ i with Fiiell"! ¦ > iin|-Ik' > nil vrt '« " i ''
The second field sample test was carried out in October 24 - Nov 12, 2019. The samples were collected from 12
monitor wells in Red Cove area, representing a wider range of groundwater matrix. The main objective of this test
was to evaluate speciation stabilization for field samples with oxalic acid.
For site characterization and monitoring, it was desirable to maximize the number of samples in field.
Groundwater sampling involves purging and stabilization for samples, with multiple checks of physical
conditions. The purging rate and volume of water varied between wells because stabilization depended on the
well design and permeability of groundwater through the well/03 Our speciation preservation method did not
disrupt these activities.
In this test, the sample preservation involved adding 200 ml of the groundwater sample to a clear glass bottle pre-
loaded with 1.16 g of oxalic acid dihydrate (9.2 mmol, CAS 6153-56-6). It did not cause significant delay of
sampling. The samples were shipped next day to the lab for testing. There was no deliberate control of bottle head
space or light exposure. All samples were handled at ambient temperature. The control samples were not modified
with oxalic acid but otherwise similarly handled.
9.2.1 The Day 1 results and impact of oxalic acid on As(lll) recovery
The sensors for this test were fabricated on October 17, 2019, and dry stored. They were hydrated on October 21,
2019 in 0.1 M Tris-Oxalate, pH 7.0. The calibration was carried out on the October 22, 2019, giving a standard
curve with regression equation ofI% = 1.58 [As(III)] + 5.8, R2 = 0.8721.
29
-------
Figure 26. Appearance of samples in field (left) and 1 day later in lab
A total of 12 pairs of samples were collected and shipped to EPA ORD's Cincinnati lab. The total time between
sampling and arrival at the laboratory was about 36 h. When received, rust formed in all unmodified samples
except RSK 37, which had low Fe. Acidified samples did not have rust, but some had white precipitate (Figure
26). For sensor analysis, 50 ml of the groundwater was taken from an agitated sample. The OA modified samples
were adjusted to pFI 8.0 with Tris base. On average, 0.839 ± 0.0183 g (6.93 ± 0.15 mmol) of Tris base was needed
for acidified samples. For unmodified groundwater, equal amounts (0.290 g, 2.3 mmol) of oxalic acid dihydrate
were added to 50 ml samples and allowed to stand for 20 min to solubilize the rust. The solution was then
adjusted to pH 8.0 with Tris Base. The average amount of Tris was 0.8835 + 0.0181 g (7.30 ±0.15 mmol, Table
15). The consumption of Tris was significantly different between the two groups as shown with a t-test (Figure
27). The most likely cause was the oxidation of Fe2+ because oxidation of 1 mole of Fe2H oxidation requires !4
mole of O2 and gives 2 moles of H+ (Figure 8). In OA modified samples, Fe2+oxidation was slowed down and
ferric oxyhydroxide formation was prevented. Practically, pH adjustment may not require titration. For future
work, we may adjust pFI by adding a fixed amount of Tris to a fixed sample volume since Tris has strong
buffering capacity at pFI 8.
30
-------
Table 15. Additional Tris needed to titrate untreated sample and redox species in water matrices
Sample
Tris-OA
Tris-Unmodified
Difference
Fe
Mn
(g)
(g)
(nmol)
(UM)
(UM)
RSK1
0.8276
0.8697
468
394
62
RSK6
0.8139
0.8454
350
573
53
RSK7
0.817
0.9127
1063
609
75
RSK37
0.8502
0.8928
473
1
2
RSK39
0.8651
0.8934
314
56
113
RSK41
0.8744
0.8776
36
286
38
RSK8
0.8396
0.887
526
466
80
RSK10
0.8366
0.8703
374
985
51
RSK12
0.8321
0.8855
593
340
29
RSK13
0.8406
0.8835
477
752
80
RSK14
0.8233
0.9089
951
788
60
RSK15
0.8474
0.8749
305
448
62
Average
0.8390
0.8835
494
475
59
StdDev
0.0183
0.0181
279
290
28
0.92 -
0.9-
0.88 --
"Q
"O
<
* 0.86-
£
0,82-
Figure 27. Tris consumption to titrate acidified (Y) or untreated (N) samples (t (22) = - 5.98, p < 0.0001)
The samples were analyzed in 4 samples/group with 3 repeats for each sample. Each sensor analysis needed 1 h
incubation and 5 min for the activity assay. The 4-sample group took about 2 h to complete. All samples were
analyzed on the day of their arrival.
Elemental analyses, including total As analysis, were carried out by ICP-AES. The sensor analysis results are
summarized in Table 16. Total As concentrations by ICP-AES were used as the reference to determine As(III)
recovery.
Oxalic acid
31
-------
Table 16. Summary of the sensor tests for As(III) and total As by ICP-AES
Sample
OA
As(III)-l
(MM)
As(III)-2
(MM)
As(III)-3
(MM)
Total As by ICP-AES
(MM)
RSK1
N
5.81
6.92
5.97
8.28
RSK1A
Y
9.54
12.28
10.67
8.28
RSK6
N
6.39
1.61
2.66
9.88
RSK6A
Y
13.28
9.70
6.56
9.88
RSK7
N
-0.64
1.70
-0.80
4.81
RSK7A
Y
3.26
2.96
4.93
4.81
RSK8
N
4.42
-3.97
2.25
5.47
RSK8A
Y
-0.01
3.77
6.74
5.47
RSK 10
N
2.16
-2.50
-2.43
5.07
RSK10A
Y
-0.99
-4.70
1.01
5.07
RSK 12
N
-3.24
-3.91
-2.69
1.47
RSK12A
Y
2.01
-2.15
0.44
1.47
RSK 13
N
-2.43
-2.78
-0.81
5.61
RSK13A
Y
3.50
-2.45
-0.20
5.61
RSK 14
N
0.85
-2.88
2.19
6.41
RSK14A
Y
7.27
-6.35
9.30
6.41
RSK 15
N
0.10
4.65
-1.04
3.34
RSK15A
Y
0.57
4.06
-1.58
3.34
RSK37
N
-2.75
0.44
-1.93
0.29
RSK37A
Y
-2.03
-2.85
-8.66
0.29
RSK39
N
-1.92
-0.76
0.61
0.59
RSK39A
Y
-3.70
-4.52
3.43
0.59
RSK41
N
-0.46
-2.24
-0.39
5.61
RSK41A
Y
2.18
7.71
5.52
5.61
Total As concentrations were 1.4, 0.29, and 0.59 |o,M in RSK 12, 37, and 39, respectively. These values were all
below the detection limit of the AChE sensor even if all the As was in As(III) form. Sensor results from these
samples were negative and had large variance. The average [As(III)] for these samples reported by the sensor
were between - 4.52 and 0.1 |a,M, confirming that the concentrations were below the detection limit. The impact
of oxalic acid on speciation preservation were evaluated with at-test (Table 17). The [As(III)] in OA-modified
samples from RSK1, 6, 7, and 41 ranged from 3.72-10.83 |a,M. These were significantly higher than those in
unmodified samples. The result indicated that OA modification improved stability of As(III) in these samples.
The [As(III)] in OA-modified samples from RSK 8, 10, 13, 14, and 15 ranged from 0.29 - 3.50 |a,M. The
difference between these results and unmodified samples was not significant. At low concentration of As(III), the
impact of OA modification should be evaluated with a more precise speciation method such as HPLC-ICP-MS.
32
-------
Table 17. T-test (n=3) evaluation of OA impact on As(III) stability
Well
ICP-AES
|_iM
OA modified
|_iM
Unmodified
|_iM
t
P
RSD-OA
modified
%
RSD-
Uiimodified
%
RSK1
8.28
10.83 ± 1.37
6.23 ±0.60
5.30
0.0082
13
10
RSK6
9.88
9.85 ±3.36
3.55 ± 2.51
2.60
0.0325
34
71
RSK7
4.81
3.72 ± 1.06
0.09 ± 1.40
3.58
0.0130
28
RSK8
5.47
3.50 ±3.38
0.90 ±4.36
0.817
0.231
97
RSK 10
5.07
-1.56 ±2.90
-0.92 ±n2.67
-0.280
0.603
RSK 12
1.47
0.10 ± 2.10
-3.28 ±0.61
RSK 13
5.61
0.29 ±3.00
-2.01 ±0.15
1.25
0.159
RSK 14
6.41
3.41 ±8.51
0.05 ±2.62
0.652
0.286
250
RSK 15
3.34
1.02 ±2.85
1.24 ±3.01
0.0920
0.534
280
243
RSK37
0.29
-4.52 ±3.61
-1.41 ± 1.66
RSK39
0.59
-1.60 ±4.38
-0.69 ± 1.27
RSK41
5.61
5.14 ± 2.79
-1.03 ± 1.05
3.59
0.0239
54
300
250
cr>
II
200
c
150
cN
Q
100
U~l
oc
50
0
0.0
5.0
10.0
15.0
[As (III)] |iM
Figure 28. Dependence of relative standard deviation (RSD) on [As(III)] for the AChE sensor
The precision of the sensor decreased rapidly when [As(III)] was <4 DM as shown in the relationship between
relative standard deviation (RSD) and [As(III)] (Figure 28). When [As(III)] > 4 |oM, RSD values were between
13-54%. RSD values rose rapidly in 2 - 4 |o,M of [As(III)]. When [As(III)] was < 2 |a,M, RSDs were around 250%.
OA modification appeared to stabilize speciation in low As samples, but the standard deviations were too great to
evaluate the significance. This hypothesis should be evaluated with a more precise analysis such as a
chromatography-spectroscopy method.
The overall impact of OA modification was evaluated with the samples having concentrations above the detection
limit. The evaluation was based on the hypothesis that if OA had no impact on speciation stability, As(III)
recovery would be the same between the two sample groups. Recovery was determined for total As (analyzed
with ICP-AES) (see Table 18). The difference should be zero if the hypothesis was true. For matched pair tests
for recovery, at least one of the concentrations should have been above the detection limit. Negative results were
hypothetically set at 0% recovery (unmodified sample from RSK 41). Out of 12 pairs of samples, 7 were used for
the evaluation.
33
-------
Table 18. Evaluation of the impact of OA modification with As(III) recovery (%)
Well
Total As
(ICP-AES) |_iM
OA modified
l_iM
Unmodified
l_iM
Recovery
OA modified
Recovery
Unmodified
A Recovery
RSK1
8.28
10.83 ± 1.37
6.23 ±0.60
131
75
56
RSK6
9.88
9.85 ±3.36
3.55 ± 2.51
100
36
64
RSK7
4.81
3.72 ± 1.06
0.09 ± 1.40
77
2
75
RSK8
5.47
3.50 ±3.38
0.90 ±4.36
64
16
48
RSK 10
5.07
-1.56 ±2.90
-0.92 ±n 2.67
<0
<0
RSK 12
1.47
0.10 ± 2.10
-3.28 ±0.61
7
<0
RSK 13
5.61
0.29 ±3.00
-2.01 ±0.15
5
<0
RSK 14
6.41
3.41 ±8.51
0.05 ±2.62
53
1
52
RSK 15
3.34
1.02 ±2.85
1.24 ±3.01
30
37
-7
RSK37
0.29
-4.52 ±3.61
-1.41 ± 1.66
<0
<0
RSK39
0.59
-1.60 ±4.38
-0.69 ± 1.27
<0
<0
RSK41
5.61
5.14 ± 2.79
-1.03 ±1.05
92
<0
92
The distribution showed that 6 out of 7 of the recovery differences were above 40%. Only one was -7% (Figure
29). The mean value of the A-Recovery was 54 ± 31%. With the hypothesis of A-Recovery = 0, t-test gave t (6) =
4.640 and p = 0.0018. Therefore, we concluded that OA modification increased As(III) recovery.
Distribution of recover difference
Figure 29. Recovery difference between OA modified sample and unmodified sample
It has been noted that OA modified RSK1 had a recovery of 131%. The error could either be caused by the
variance (RSD = 13%) in sensor analysis or by the matrix. Based on our results, our working hypothesis is that
high total As (~ 10 |a,M) correlates to high As(III) at this site. As long as the recovery of OA modified RSK1 was
not less than 75% (the recovery of unmodified RSK 1 on day 1), the conclusions would be valid for the matched
pair test.
Sequestration of As took place at the front edge of the As(III) plume, where Fe (II) and/or As(III) were oxidized.
Assuming the results from OA modified samples were accurate, wells at the edge of the As(III) plume could be
identified. These included RSK 7 in well group RSK 1-7, RSK 8 and 10 in RSK 8-12, and RSK 14 and 15 in well
group RSK 13-15.
34
-------
9.2.2 Storage stability of the samples at room temperature (21-23 °C)
Six pairs of field samples received on October 24 and 25, 2019 were left at ambient temperature and analyzed
again after 5-7 days. On Day 1, average As(III) concentrations were 3.41 - 10.83 |o,M in OA modified samples,
and < 0 to 6.23 |o,M in unmodified samples. The total As concentrations were from 4.87-9.88 |o,M. Non-modified
samples were also tested to find out if As(III) underwent further oxidation. The stability of speciation was
evaluated by comparing the results of Day 1 and Day 5-7 (Table 19).
Table 19. Summary of As(III) determination (|a,M) in samples with different storage times at room temperature
Well
OA
Day
#1
#2
#3
Average
Std Dev
RSK1
Y
1
9.54
12.28
10.67
10.83
1.37
RSK1
Y
7
7.12
9.10
3.79
6.67
2.68
RSK6
Y
1
13.28
9.70
6.56
9.85
3.36
RSK6
Y
7
9.06
3.52
7.99
6.86
2.94
RSK7
Y
1
3.26
2.96
4.93
3.72
1.06
RSK7
Y
6
7.09
5.05
2.88
5.01
2.10
RSK8
Y
1
-0.01
3.77
6.74
3.50
3.38
RSK8
Y
5
2.63
1.22
2.62
2.16
0.81
RSK14
Y
1
7.27
-6.35
9.30
3.41
8.51
RSK14
Y
6
1.44
1.86
6.17
3.16
2.62
RSK41
Y
1
2.18
7.71
5.52
5.14
2.79
RSK41
Y
7
6.91
10.49
8.21
8.53
1.81
RSK1
N
1
5.81
6.92
5.97
6.23
0.60
RSK1
N
7
10.97
1.81
8.52
7.10
4.74
RSK6
N
1
6.39
1.61
2.66
3.55
2.51
RSK6
N
7
3.67
0.86
7.44
3.99
3.30
RSK7
N
1
-0.64
1.70
-0.80
0.09
1.40
RSK7
N
6
2.29
-1.60
-3.83
-1.05
3.10
RSK8
N
1
4.42
-3.97
2.25
0.90
4.36
RSK8
N
5
0.65
-1.16
-0.98
-0.50
1.00
RSK14
N
1
0.85
-2.88
2.19
0.05
2.62
RSK14
N
6
-0.75
1.81
3.30
1.45
2.05
RSK41
N
1
-0.46
-2.24
-0.39
-1.03
1.05
RSK41
N
7
3.20
-3.25
-0.29
-0.11
3.23
T-tests results indicated that OA modified RSK1 samples underwent significant decrease of As(III) (Table 20).
This was due to the large variations in the first test, in which the recovery was > 100%. For the rest of the
samples, the changes were insignificant based on t-tests. Apparently, in OA modified samples, As(III) was stable
in this period of time. In unmodified samples, As(III) oxidation took place mostly in Day 1.
35
-------
Table 20. T-test (n=3) evaluation of storage time on As(III) stability in OA modified and unmodified samples
Well
OA
Day 1 (|i.M)
Day 5-7 (jiM)
t
P
RSD Day 1
(%)
RSD Day 5-7
(%)
RSK1
Y
10.83 ±1.37
6.67 ±2.68
2.39
0.0486
13
40
RSK6
Y
9.85 ±3.36
6.86 ±2.94
1.16
0.156
34
43
RSK7
Y
3.72 ± 1.06
5.01 ±2.10
-0.948
0.414
28
42
RSK8
Y
3.50 ±3.38
2.16 ± 0.81
0.669
0.566
97
38
RSK 14
Y
3.41 ±8.51
3.16 ±2.62
0.0486
0.965
250
83
RSK41
Y
5.14 ±2.79
8.53 ± 1.81
-1.77
0.163
54
21
RSK1
N
6.23 ±0.60
7.10 ±4.74
-0.314
0.782
10
67
RSK6
N
3.55 ± 2.51
3.99 ±3.3
-0.18
0.865
71
83
RSK7
N
0.09 ± 1.40
-1.05 ±3.10
0.578
0.607
1574
-295
RSK8
N
0.90 ±4.36
-0.5 ± 1.00
0.541
0.638
484
-202
RSK 14
N
0.05 ±2.62
1.45 ±2.05
-0.728
0.509
5162
141
RSK41
N
-1.03 ± 1.05
-0.11 ±3.23
-0.468
0.679
-102
-2931
Matched pair tests in terms of recovery showed that the differences between Day 1 and Day 5-7 were not
significant for OA modified samples (Table 21). The mean value of the A-Recovery was 3.5 ± 41%. With the
hypothesis of A-Recovery = 0, t-test indicated t (5) = 0.2089 and p = 0.8427.
Table 21. The impact of storage time on OA modified samples with As(III) recovery (%)
Well
Total As
Recovery OA Day 1
Recovery OA Day 5-7
A Recovery
RSK1
8.28
131
81
50
RSK6
9.88
100
69
30
RSK7
4.81
77
104
-27
RSK8
5.47
64
39
25
RSK 14
6.41
53
49
4
RSK41
5.61
92
152
-61
Matched pair analysis with recovery showed that OA modification helped to preserve As(III) after 5-7 days at
ambient temperature. Except RSK 1, all the samples had higher recovery as the result of OA modification. The
mean recovery difference was 57.6 ± 58%. The test of mean value was t (5) = 2.41, p = 0.0302 (Table 22).
36
-------
Table 22. The impact of OA modification on As(III) recovery (%) over 5-7 days
Well
Total As
OA-modified Day 5-7
Unmodified Day 5-7
Recovery Difference (%)
RSK1
8.28
6.67
7.1
-5
RSK6
9.88
6.86
3.99
29
RSK7
4.81
5.01
0*
104
RSK8
5.47
2.16
0*
39
RSK 14
6.41
3.16
1.45
27
RSK41
5.61
8.53
0*
152
* Negative value treated as 0
9.2.3 Siiiiini d conclusions for the second test of field samples
• Rust formation was observed in unmodified samples on the day of sample collection. In OA modified
samples, no rust formed over > 7 days.
• Acidifying unmodified samples with OA dissolved rust in ~ 20 min. Titration of these samples consumed
more Tris than that with OA modified sample. Since oxidation of Fe(II) generates acid, it was concluded
that OA modification slowed Fe(II) oxidation in these samples.
• Each pair of samples (OA modified and unmodified) has been evaluated with t-tests in terms of [As(III)]
(t-test with concentration). However, the precision of the sensor was not adequate when [As(III)] was < 4
|o,M. The RSD increased dramatically when [As(III)] < 4 |o,M.
• OA modification can be evaluated over the entire dynamic range (2-11.4 |o,M) of the sensor by matched
pair analysis, involving comparison of the average As(III) recoveries between OA modified groups and
unmodified groups (t-test with recovery).
• For Day 1 samples, t-tests indicate [As(III)] and overall recoveries were significantly higher in OA
modified group. After 5-7 days, the recoveries were still higher in OA modified group, confirming that
OA helped to stabilize As(III) at ambient temperature with no strict control of air or light exposure.
• No significant change was found between Day 1 and Day 5-7 samples for most samples in terms of
concentration or recovery. Only [As(III)] in RSK1 decreased significantly.
• Assuming results with OA modified samples on Day 1 were accurate, RSK 7 in cluster RSK 1-7, RSK 8
and 10 in RSK 8-12, RSK 14 and 15 in cluster RSK 13-15 appeared to be on the leading edge of the
As(III) plume.
I 'I i s I h ' | III I I | , . I V "ill! »|| I " v "ll i.;> vl i" " || vi II I
forward
Among enzyme-based biosensors, the widely-used glucose meter is most impactful. Historically, it is the most
important reference case in enzyme sensor development. Since the introduction of enzyme-based glucose
meters in 1970s, this technology development has been driven by the need to obtain accurate blood sugar
levels at the point-of-care or in patient's homes. As of 2000, there were 171 million diabetic patients,
accounting for 2.8% of the global population/04 This vast market potential has attracted substantial investment
in the advancement of glucose meter technology for clinical and outpatient use.'05 Technology hurdles have
been identified and overcome to improve sensor stability, accuracy, reliability, measurement time, versatility,
user interface, and cost. As these technologies evolved, the glucose meter market expanded. It was used first in
the diagnosis of glycemic patients in clinics, and later for management of blood sugar both in clinics and at
home. Currently, self-monitoring of blood glucose (SMBG) represents the largest market for glucose meters,
37
-------
and has contributed to the dramatic reduction of complications among diabetic patients/06,10' Apart from
essential applications such as diagnosis and screening for diabetics, SMBG has also extended to education,
protection, and guidance for behavior such as food and exercise. It has helped to achieve glycemic targets for
both Type I and Type II diabetic patients/05 In Type I diabetic (insulin dependent) patients, SMBG has helped
to reduce complications by 60%.'° Yet while the glucose meter's value is proven, it has not reached its full
potential in healthcare. A recent survey in 37 nations showed that the cost of a 10 ml insulin ranged from $3.84
to $34.09, and the cost of a SMBG test was $0.25 to $ 1.65. Personal cost coverage of these tests was always
lower than that of insulin except in three affluent nations/09
c
Supplier
Fabrication
Shipp lingstorage
Specification
Operator
i
Sampling
Calibration
' Analysis
Data management
Glucose analyzer
'
Sample 1 [>
r ¦ 1
Bioreceptor 1—y
r (
Transducer 1—y Data
I J V J
Error Grid Analysis ^ *
V
(0
S 60
3
£ «
k_
0)
s w
50 100 150 ZOO 2SO 300 350 400 *50 500 550
Reference (mg/dl)
• 0
Distribution of Sensor Results
as a function of accuracy
tamar*'mm -I",
t" 5*41
>10% ¦ ii* WiBui
V
Better understanding, prediction/management, decision making
Figure 30. Flow diagram of glucose meter development (Error Grid Analysis and Accuracy Profile are quoted
from Chang et al1"4)
The bio receptors of glucose meters are enzymes involved in the oxidation of glucose or glucose phosphate. A
desirable enzyme would have high selectivity and activity in glucose oxidation, be stable for practical sensor
development, be tolerant to matrix variation, and be compatible with highly efficient and selective transduction.
Over time, three generations of glucose oxidation enzymes have been used as bioreceptors for glucose meters/06,
110 Glucose oxidase (EC 1.1.3.4) was the first enzyme used for glucose sensing. It catalyzes glucose oxidation with
O2, giving glucolactone and H2O2 as the products. The reaction rate was reported by amperometric measurement
of H2O2. This mechanism served to measure glucose in blood samples because glucose oxidase was selective in
38
-------
the complex matrix. However, sensor accuracy suffered from two main issues. One was its dependence on O2
concentration; the other was the transduction interference of oxidizable chemicals (ascorbic acid, uric acid, etc.) in
blood, as high working potential is needed to oxidize H2O2. The second-generation glucose meter launched in the
1980s proved far more successful. It overcame the earlier problems by employing a PQQ glucose dehydrogenase
(EC 1.1.5.2) as the bioreceptor. This enzyme catalyzes the oxidation of glucose with PQQ (pyrroloquinoline
quinone) as the electron acceptor. A ferrocene mediator was incorporated so that low working potential was
adequate to oxidize reduced PQQ in an amperometric assay. This mechanism did not require oxygen, and the
transduction was more specific. Along with accuracy improvement, the sensor was also miniaturized for
outpatient use. Further development in the 1990s-2000s focused on direct charge transfer from glucose oxidation
enzymes to the electrode. The transduction has very high selectivity, and the enzymes were improved with
molecular engineering. This third-generation enzyme has been used to develop implantable sensors for real time
monitoring/ia 111
Several transductions have been developed or commercialized for glucose sensors, including optical sensors in the
1990s,106 112 and surface plasmon resonance in recent years/0 ,111 Electrochemical transduction remains the most
important option because the cost is low, and efficiency and selectivity have been dramatically improved.
Therefore, we prioritized electrochemical transduction in arsenic sensor development. Glucose sensor products
were extensively evaluated in terms of technical accuracy and application accuracy, and automation and error-
proof features were implemented into their design/06/" Application accuracy was profiled by error grid analysis
(EGA, Figure 30).which linked accuracy profiles to specific clinical risks.114 We believe that EGA is a
valuable approach to assess the impact of sensor accuracy on environmental risks, and plan to use it for As sensor
product characterization.
Biosensors for arsenic, including cellular and enzyme sensors, appeared in late 1990s.115,116 The most advanced
version is the cellular sensor based on ars operon (Figure 31)." '118 As(III) is detected by the binding to ArsR, a
promoter for gene expression. As(III) concentration is reported by downstream green fluorescent protein (GFP)
expression.^6 A fluorometer or cell phone camera can be used to measure the optical signal.118 The progress
largely targeted sensitivity improvement (to < 1 ppb) by genetic technology. Cellular sensors were evaluated in
field to determine bioavailable As (not speciation). However, the genetically modified bacteria (a GMO) and the
vessels in contact with the bacteria need to be sterilized before disposal.
ArsR
c
GFP
As3+
O/P
I777TK
| Translation
mRNA
Transcription
t
Figure 31. The design with ars operon for reporting [As(III)]. O/P, operator/promotor; Amp, amplifier for
transcription; GFP, green fluorescent protein
39
-------
Despite efforts since the 1990s, the development of enzyme sensors for arsenic has not progressed beyond
prototype design.27 There are no reports on field tests. Matrix effects and sample preparation needs have not been
evaluated for enzyme sensors. For prototype development, these issues should be studied in practical
applications.107 We surveyed the literature to compare cellular and enzyme sensors when the project started in
2014. Cellular sensors for As was a relatively mature technology, yet limitations remained after >10 years of
intensive research and development. We found opportunities to improve the enzyme sensor by addressing crucial
issues with the technology, and determined that applications of enzyme sensors should be different from those of
cellular sensors.
We redesigned the prototypic sensor based in AChE inhibition by incorporating two improvements.27 The new
transduction mechanism is simple, efficient, and selective. Under current assay condition, Interference found in
thiocholine-based transduction has been eliminated. The anodic oxidation does not require a redox mediator. The
hydroquinone product can be oxidized at 0.35 V with an amplification factor of 57 on carbon electrodes. The
other improvement is a mechanism-based inhibition assay to eliminate interference by reversible inhibitors (such
as Cu(II)) in the sample matrix. Many inhibition-based biosensing systems failed due to interference by
known/unknown components in the matrix."' The inhibition of AChE by As(III) is pseudo-irreversible.
Therefore, activity assay and inhibition are carried out in two separate steps. The binding in inhibition serves as a
purification mechanism. Only kinetically irreversible inhibitors are carried over to activity assay.
Cost, productivity, and reliability are the three most important considerations for sensor fabrication. Screen
printed carbon electrode was selected as the base for enzyme immobilization because it was least expensive,
disposable, and potentially customizable for precise chemical modification. Glutaraldehyde crosslinking was
selected for AChE immobilization because it was tunable to link different amounts of AChE on a fixed area.
Sensor storage condition and stability have been evaluated to ensure their practicality. The chemistry and
procedure of crosslinking has been systematically optimized to maximize sensitivity and minimize variance
between sensors in a batch. The crosslinking reaction has been streamlined so that 100 sensors can be made in a
batch.
The initial application of the AChE sensor is set for anthropogenic sites because the arsenic concentrations at
these sites are usually high, and dense sampling should be valuable to support site cleaning or groundwater
resource management.120 The superfund site at Shepley's Hill Landfill was proposed to guide sensor
development. The pollution was caused by a landfill through reductive dissolution mechanism, therefore As
migration and sequestration depended on the redox oscillation between As(III) and As(V). Importantly, the
arsenic concentrations in groundwater were mostly in the dynamic range of the prototypic AChE sensor (2 to 20
|o,M)2S
We have found that the sensitivity of the AChE-sensor to As(III) was not influenced by main components in
groundwater matrices, including Fe2+(up to 55 mg/L), Ca2+, Mn2+ (up to 4.4 mg/L), Mg2+, SO42, CO32, SiC>32,
PO43", and NH4+. In the first test with groundwater samples, we evaluated the sensor in oxalic acid modified
samples, because we intended to use oxalic acid in sample preparation. The accuracy has been validated with ICP-
AES in the range of 7-20 |oM As(III) in modified groundwater matrices (Table 13).
Readiness for testing on a larger scale became the top priority once the feasibility had been established for As(III)
determination in oxalic acid modified groundwater samples. In the life cycle of the AChE sensor, variance can
compound over successive steps prior to As(III) determination (Figure 7). To manage data quality, the impacts of
variance in fabrication, storage condition, and hydration condition were characterized. The characterization
involved evaluation of variance in terms of activity, and sensitivity to 8 |a,M of As(III). For the second field
sample test, 200 sensors were prepared to test 24 samples on two time points in one week. They were fabricated,
stored, and hydrated by justified specifications.
In the second test of field samples, speciation preservation with oxalic acid was carried out in field sampling. The
preserved samples from 12 wells were compared with corresponding controls. We found oxalic acid slowed Fe(II)
40
-------
oxidation, prevented ferric oxyhydroxide precipitation, and improved recovery of As(III) by 54 ± 31% in the
concentration range of 3.72-10.83 |o,M. For samples with low concentrations (< 4 |o,M) of As(III), the stabilization
was difficult to evaluate due to low precision of the AChE sensor.
Table 23. Assessment of the arsenic sensor development with Technology Readiness Level (TRL)
TRL
/characteristics
Program/Time
Activity
Deliverable
Participants
1
Basic approach
for pollutant
sensing: enzyme
inhibition-based
sensing
PIP3
Developing practical
evaluation of arsenic in
groundwater based on
biological interactions
2014-2015
Arsenic biosensor mechanism selection:
cellular vs enzyme.
Evaluating enzymes for inhibition-based
biosensor development: Pyruvate
dehydrogenase, Acetylcholinesterase
Acid Phosphatase, Cytochrome C,
Proteases, RNAases, Glutathione
Reductase.
Acetylcholinesterase has been
selected as the bioreceptor for
As(III) for further development.
The sensing mechanism based on
pseudo-irreversible inhibition of
AChE has been verified.
Endalkachew
Sahle-Demessie,
Tao Li, Jason
Berberich
2
Technology
concept:
amperometric
AChE sensor for
As (III)
determination
PIP4
Developing a biosensor
prototype for field test of
arsenic in groundwater
2015-2016
Designs of transduction'. ATCI, coupling
with choline oxidase, novel transduction
mechanism.
Enzyme immobilization: covalent (not
tested), crosslinking, entrapment.
Characterization of the transduction
mechanism: efficiency and selectivity of
enzyme reaction and electrochemical
reaction.
Mechanism of immobilized AChE (cross-
lined)'. kinetic studv.
Novel amperometric transduction
mechanism with 4-acetoxyphenol
as the substrate; sensor fabricated
by crosslinking AChE on SPE base.
Activity of immobilized AChE can
be reported efficiently with the
transduction mechanism. Kinetic
mechanism of As(III) inhibition
validated with the sensor.
Tao Li, Jason
Berberich,
Endalkachew
Sahle-Demessie
3
Proof of concept
Prototype
design:
integrating
bioreceptor with
the transducer,
application field
SSWR 6.03, Transformative
Approaches and Technologies
for Water Systems
2017-2018
Protocol development based on the sensor
design', sensor fabrication, storage, sensor
condition before assay, amperometric
assay of sensor activity, calibration and
assay method for [As(III)].
Literature survey on arsenic analysis to
monitor groundwater.
A prototype of the AChE sensor for
As(III) determination.
Characterization of the prototypic
sensor in terms of linear range,
sensitivity, and stability.
A report on field sensor and arsenic
analysis to monitor groundwater.
Tao Li, Jason
Berberich,
Endalkachew
Sahle-Demessie,
Eunice
Varughese
4
Technology
validation in the
lab
SSWR 6.03, Transformative
Approaches and Technologies
for Water Systems
2018-2019
Preparation for validation', sensor
characterization, analytical method
characterization.
Accuracy (systematic error)', recovery of
standard addition, validation with ICP-
AES.
Precision: LOD, RSD.
Source of errors and risk mitigation
impact on senor activity and sensitivity.
Sensor specifications
Accuracy & Precision in the
dynamic range: dose-dependent
precision,
Ao in response to uncontrollable
variations in fabrication,
distribution of A0with current
fabrication, correlation of Ao to
sensitivity to [As(III)], matrix
impact.
Tao Li, Jason
Berberich,
Endalkachew
Sahle-Demessie
5
Technology
validation in
relevant
environment
SHC 2 Site Characterization
and Remediation, Output 5:
Innovative Technologies to
eliminate or control mining
wastes as sources of water
contamination, Product 4:
Arsenic Biosensor
2019-
Impact of groundwater matrix on sensor
performance.
Stability of groundwater sample.
Sample preservation method: validation in
lab, testing with groundwater sample from
field.
Performance of the sensor in
groundwater matrix.
Oxalic acid for sample preparation,
Pre-reduction for total As
determination.
Proposal for streamlining arsenic
analysis with field sampling and
the lifecvcle of the sensor.
Tao Li, Robert
Ford,
Endalkachew
Sahle-Demessie,
Rick Wilkin
6
Technology
demonstration in
relevant
environment
SHC 2 Site Characterization
and Remediation, Output 5:
Innovative Technologies to
eliminate or control mining
wastes as sources of water
contamination, Product 4:
Arsenic Biosensor
2019-
Identify field use barrier, uncontrollable
variations, manufacture readiness for
technical demonstration.
Sensor fabrication readiness, deployment
readiness.
Manufacture readiness, storage
readiness, feasibility for field.
Distribution and randomness of
residual error, (Ao, storage time,
hydration, assay).
Tao Li, Robert
Ford,
Endalkachew
Sahle-Demessie,
Rick Wilkin
7-9
From prototype
to product
41
-------
Developing a low-cost sensor for field use is an evolving process. Decisions in earlier stages have strong impact
on later stages. From lab to field, different development stages involve different activity, participants, and risks.
The roadmap of past development can be characterized with an assessment by Technology Readiness Levels
(TRL, Table 23). Assessment with TRL is a formal approach to evaluate the maturity of a technology, and has
been widely adopted by industry to manage technology development./2/ There are three main stages of
development. The first stage (technology with TRL 1-3) involves designing a feasibility prototype based on an
invention. The work is typically carried out in a lab by academics. The second stage (technology with TRL 4-6)
involves prototype development, aiming to deliver the a-prototype, that embeds all the essential features in the
final product. The work is mostly done by small businesses as they have better access to markets. The final stage
(technology with TRL 7-9) involves commercial product development.
The specific criteria of TRL depends on the area of a product. We are not aware of any publications on TRL for
sensor development. Therefore, the characteristics, activity and deliverables in Table 23 are proposed based on
our experience. Our analysis showed that the current AChE sensor was at TRL 5 and moving towards TRL 6.
Sensor prototype development is guided by improving the information both in quality and its relevance to
decision making. We have identified several issues in the two field sampling tests. We plan to revise the prototype
and its interface with other field activities by addressing the following issues:
• Low precision when [As (III)] < 4 |o,M.
The equilibrium of As3+ binding to AChE was slow to reach at low concentration. We also frequently
observed false positive results with blank samples. This was probably caused by the physical loss of
immobilized enzyme as glutaraldehyde crosslinking was reversible.
We plan to find an irreversible immobilization of AChE for sensor fabrication and improve the mass transfer
by minimizing barriers to the enzyme layer on the electrode.
When feasible, we will study binding kinetics and affinity of As(III) to AChE with surface plasmon
resonance, and how quaternary ligands improve the binding kinetics/22 This would help us to improve the
sensitivity and reduce the error in As (III) analysis with improved assay protocol.
• Sample stability
An oxalate-based speciation preservation for arsenic will be evaluated and optimized.
• Total As determination
A pre-reduction protocol is going to be developed for total As determination.
The vitality of sensor development depends on the sensor's ability to fill the knowledge gap left by laboratory
analysis/23 For arsenic pollution site management and water treatment evaluation, this gap is caused by the low
throughput of lab analysis. Based on this consideration, we expect a microplate reader-based method may also fit
the need.i2¥ A 96-well plate reader is now a generic analytical instrument. The operation does not require
specialized training. We have proven that 4-acetoxyphenol was a suitable substrate for optical transduction to
report AChE activity. The amplification factor at 300 nm was around 75. The AChE from electric eel (current
bioreceptor for the amperometric sensor) has a Km = 5.2 ± 0.8 mM and a V„vix = 81.5 ± 5.0 mmol mg"1 min"1 for
the substrate. Based on preliminary tests, we expect the precision of the colorimetric assay to be better than that of
the current amperometric sensor. In addition, other components in groundwater such as Fe2+ and Fe3+ can be
measured colorimetrically. 125 They can be analyzed with the plate reader at similar throughput. In this study, we
have found that it is feasible to evaluate dynamics of As speciation in relation to that of Fe.
42
-------
References
[1] Rahman, M. M., Sengupta, M. K., Chowdhury, U. K., Lodh, D., Das, B., Ahamed, S., Mandal, D., Hossain,
M. A., Mukheijee, S. C., Pati, S., Saha, K. C., and Chakraborti, D. (2006) Arsenic contamination
incidents around the world. Managing Arsenic in the Environment (Naidu, R., Euan, S., Owens, G.,
Bhattacharya, P., and Nadebaum, P., Eds.), 3-30, CSIRO Publishing.
[2] Mandal, B. K., and Suzuki, K. T. (2002) Arsenic round the world: a review, Talanta 58, 201-235.
[3] Tchounwou, P. B., Patlolla, A. K., and Centeno, J. A. (2003) Carcinogenic and Systemic Health Effects
Associated with Arsenic Exposure-A Critical Review. Toxicol. Pathol. 31, 575-588.
[4] Garelick, H., Jones, H., Dybowska, A., and Valsami-Jones, E. (2008) Arsenic pollution sources. Rev. Environ.
Contam. Toxicol. 197, 17-60.
[5] Polya, D. A., and Lawson, M. (2015) Geogenic and anthropogenic arsenic hazard in groundwaters and soils:
distribution, nature, origin, and human exposure routes. 23-60, John Wiley & Sons, Inc.
[6] Gibson, J., Polya, D. A., Russell, N., and Sauer, J. (2017) Valuing the damage of arsenic consumption:
economic non-market valuation methods. Best Practice Guide on the Control of Arsenic in Drinking
Water (Bhattacharya, P., Polya, D. A., and Jovanovic, D., Eds.), p 0, IWA Publishing.
[7] Mahimairaja, S., Bolan, N. S., Adriano, D. C., and Robinson, B. (2005) Arsenic Contamination and its Risk
Management in Complex Environmental Settings. Advances in Agronomy. 1-82, Academic Press.
[8] Nordstrom, D. K. (2011) Hydrogeochemical processes governing the origin, transport and fate of major and
trace elements from mine wastes and mineralized rock to surface waters. Applied Geochemistry 26, 1777-
1791.
[9] Nordstrom, D. K., Blowes, D. W., and Ptacek, C. J. (2015) Hydrogeochemistry and microbiology of mine
drainage: An update. Applied Geochemistry 57, 3-16.
[10] Anawar, H. M. (2015) Sustainable rehabilitation of mining waste and acid mine drainage using
geochemistry, mine type, mineralogy, texture, ore extraction and climate knowledge. Journal of
Environmental Management 158, 111-121.
[11] Smedley, P. L., and Kinniburgh, D. G. (2002) A review of the source, behaviour and distribution of arsenic
in natural waters. Applied Geochemistry 17, 517-568.
[12] Lu, P., and Zhu, C. (2011) Arsenic Eh-pH diagrams at 25°C and 1 bar. Environmental Earth Sciences 62,
1673-1683.
[13] Drahota, P., and Filippi, M. (2009) Secondary arsenic minerals in the environment: A review. Environment
International 35, 1243-1255.
[14] Magalhaes, M. C. F. (2002) Arsenic. An environmental problem limited by solubility. Pure and Applied
Chemistry, p 1843.
[15] Andrade, C. F., Jamieson, H. E., Kyser, T. K., Praharaj, T., and Fortin, D. (2010) Biogeochemical redox
cycling of arsenic in mine-impacted lake sediments and co-existing pore waters near Giant Mine,
Yellowknife Bay, Canada. Applied Geochemistry 25, 199-211.
[16] Lions, J., Guerin, V., Bataillard, P., van der Lee, J., and Laboudigue, A. (2010) Metal availability in a highly
contaminated, dredged-sediment disposal site: Field measurements and geochemical modeling.
Environmental Pollution 158, 2857-2864.
[17] Brown, B. V., Valett, H. M., and Schreiber, M. E. (2007) Arsenic transport in groundwater, surface water,
and the hyporheic zone of a mine-influenced stream-aquifer system. Water Resources Research 43.
[18] Concas, A., Ardau, C., Cristini, A., Zuddas, P., and Cao, G. (2006) Mobility of heavy metals from tailings to
stream waters in a mining activity contaminated site. Chemosphere 63, 244-253.
[19] Coynel, A., Schafer, J., Dabrin, A., Girardot, N., and Blanc, G. (2007) Groundwater contributions to metal
transport in a small river affected by mining and smelting waste. Water Research 41, 3420-3428.
[20] Reimann, C., Matschullat, J., Birke, M., and Salminen, R. (2009) Arsenic distribution in the environment:
The effects of scale. Applied Geochemistry 24, 1147-1167.
[21] Cidu, R., Frau, F., and Da Pelo, S. (2011) Drainage at Abandoned Mine Sites: Natural Attenuation of
Contaminants in Different Seasons. Mine Water and the Environment 30, 113-126.
[22] Nordstrom, D. K. (2009) Acid rock drainage and climate change. Journal of Geochemical Exploration 100,
97-104.
43
-------
[23] Canovas, C. R., Macias, F., and Olias, M. (2018) Hydrogeochemical behavior of an anthropogenic mine
aquifer: Implications for potential remediation measures. Science of The Total Environment 636, 85-93.
[24] Edwards, M., Patel, S., McNeill, L., Chen, H.-w., Frey, M., Eaton, A. D., Antweiler, R. C., and Taylor, H. E.
(1998) Considerations in As analysis and speciation. Journal - AWWA 90, 103-113.
[25] Kumar, A. R., and Riyazuddin, P. (2010) Preservation of inorganic arsenic species in environmental water
samples for reliable speciation analysis. TrAC Trends in Analytical Chemistry 29, 1212-1223.
[26] Wang, S.-W., Liu, C.-W., Lu, K.-L., Chang, Y.-P., and Chang, T.-W. (2011) Distribution of Inorganic As
Species in Groundwater Samples with the Presence of Fe. Water Quality, Exposure and Health 2, 181-
192.
[27] Li, T., Berberich, J., Sahle-Demessie, E., and Varughese, E. (2019) A disposable acetylcholine esterase
sensor for As(iii) determination in groundwater matrix based on 4-acetoxyphenol hydrolysis. Analytical
Methods 17,5203-5213.
[28] Ford, R. G., Scheckel., K. G., Acree, S., Ross, R., Lien, B., Luxton, T., and Clark, P. (2008) Final Report;
Arsenic Fate, Transport and Stability Study; Groundwater, Surface Water, Soil And Sediment
Investigation, Fort Devens Superfund Site, Devens, Massachusetts. US Environmental Protection
Agency, Washington DC.
[29] Roberts, L. C., Hug, S. J., Ruettimann, T., Billah, M. M., Khan, A. W., and Rahman, M. T. (2004) Arsenic
Removal with Iron(II) and Iron(III) in Waters with High Silicate and Phosphate Concentrations.
Environmental Science & Technology 38, 307-315.
[30] Raven, K. P., Jain, A., and Loeppert, R. H. (1998) Arsenite and Arsenate Adsorption on Ferrihydrite:
Kinetics, Equilibrium, and Adsorption Envelopes. Environmental Science & Technology 32, 344-349.
[31] Wilkin, R. T., and Ptacek, C. J. (2002) Session 4 summary: Field measurement of geochemical redox
parameters. 89-99, National Technical Information Service.
[32] Morgan, B., and Lahav, O. (2007) The effect of pH on the kinetics of spontaneous Fe(II) oxidation by 02 in
aqueous solution - basic principles and a simple heuristic description. Chemosphere 68, 2080-2084.
[33] Whittemore, D. O., and Langmuir, D. (1975) The Solubility of Ferric Oxyhydroxides in Natural Waters.
Groundwater 13, 360-365.
[34] Dixit, S., and Hering, J. G. (2003) Comparison of Arsenic(V) and Arsenic(III) Sorption onto Iron Oxide
Minerals: Implications for Arsenic Mobility. Environmental Science & Technology 37, 4182-4189.
[35] Bauer, M., and Blodau, C. (2009) Arsenic distribution in the dissolved, colloidal and particulate size fraction
of experimental solutions rich in dissolved organic matter and ferric iron. Geochimica et Cosmochimica
Acta 73, 529-542.
[36] Gallagher, P. A., Schwegel, C. A., Parks, A., Gamble, B. M., Wymer, L., and Creed, J. T. (2004)
Preservation of As(III) and As(V) in Drinking Water Supply Samples from Across the United States
Using EDTA and Acetic Acid as a Means of Minimizing Iron-Arsenic Coprecipitation. Environmental
Science & Technology 38, 2919-2927.
[37] Ford, R. G., Acree, S. D., Lien, B. K., Scheckel, K. G., Luxton, T. P., Ross, R. R., Williams, A. G., and
Clark, P. (2011) Delineating landfill leachate discharge to an arsenic contaminated waterway.
Chemosphere 85, 1525-1537.
[38] Brandon, W. (2018) Comprehensive GW/SW evaluations Drive CSM Evolution from Denial to Remedial
Sucess, Red Cove Study Devens, MA. US EPA Region 10 Seattle, Washington.
[39] O'connor, A. (Setember 20, 2016) Army seeks to halt Devens arsenic cleanup. Sentinal & Enterprise, Mark
O'Neil, Fitchburg, Massachusetts.
[40] Christiansen, B. C., Balic-Zunic, T., Dideriksen, K., and Stipp, S. L. S. (2009) Identification of Green Rust in
Groundwater. Environmental Science & Technology 43, 3436-3441.
[41] Jonsson, J., and Sherman, D. M. (2008) Sorption of As(III) and As(V) to siderite, green rust (fougerite) and
magnetite: Implications for arsenic release in anoxic groundwaters. Chemical Geology 255, 173-181.
[42] Keimowitz, A. R., Simpson, H. J., Chillrud, S. N., Stute, M., Tsang, M., Datta, S., and Ross, J. (2005)
Oxidation of Groundwater Arsenic and Iron. Advances in Arsenic Research, 206-219, American
Chemical Society.
[43] Jain, A., Raven, K. P., and Loeppert, R. H. (1999) Arsenite and Arsenate Adsorption on Ferrihydrite:
Surface Charge Reduction and Net OH- Release Stoichiometry. Environmental Science & Technology 33,
1179-1184.
44
-------
[44] Emett, M. T., and Khoe, G. H. (2001) Photochemical oxidation of arsenic by oxygen and iron in acidic
solutions. Water Research 35, 649-656.
[45] Pierre, J. L., and Fontecave, M. (1999) Iron and activated oxygen species in biology: The basic chemistry.
Biometals 12, 195-199.
[46] Collin, F. (2019) Chemical Basis of Reactive Oxygen Species Reactivity and Involvement in
Neurodegenerative Diseases. IntJMol Sci 20.
[47] Liu, S. X., Athar, M., Lippai, I., Waldren, C., and Hei, T. K. (2001) Induction of oxyradicals by arsenic:
Implication for mechanism of genotoxicity. Proceedings of the National Academy of Sciences 98, 1643-
1648.
[48] Amstaetter, K., Borch, T., Larese-Casanova, P., and Kappler, A. (2010) Redox Transformation of Arsenic by
Fe(II)-Activated Goethite (a-FeOOH). Environmental Science & Technology 44, 102-108.
[49] Silver, S., and Phung, L. T. (2005) Genes and Enzymes Involved in Bacterial Oxidation and Reduction of
Inorganic Arsenic. Applied and Environmental Microbiology 71, 599-608.
[50] Osborne, T. H., McArthur, J. M., Sikdar, P. K., and Santini, J. M. (2015) Isolation of an Arsenate-Respiring
Bacterium from a Redox Front in an Arsenic-Polluted Aquifer in West Bengal, Bengal Basin.
Environmental Science & Technology 49, 4193-4199.
[51] Liao, V. H.-C., Chu, Y.-J., Su, Y.-C., Hsiao, S.-Y., Wei, C.-C., Liu, C.-W., Liao, C.-M., Shen, W.-C., and
Chang, F.-J. (2011) Arsenite-oxidizing and arsenate-reducing bacteria associated with arsenic-rich
groundwater in Taiwan. Journal of Contaminant Hydrology 123, 20-29.
[52] Daus, B., Weiss, H., Mattusch, J., and Wennrich, R. (2006) Preservation of arsenic species in water samples
using phosphoric acid - Limitations and long-term stability. Talanta 69, 430-434.
[53] McCleskey, R. B., Nordstrom, D. K., and Maest, A. S. (2004) Preservation of water samples for
arsenic(III/V) determinations: an evaluation of the literature and new analytical results. Applied
Geochemistry 19, 995-1009.
[54] Wu, D., and Pichler, T. (2016) Preservation of co-occurring As, Sb and Se species in water samples with
EDTA and acidification. Geochem.: Explor., Environ., Anal. 16, 117-125.
[55] Kim, Y.-T., Yoon, H., Yoon, C., and Woo, N.-C. (2007) An assessment of sampling, preservation, and
analytical procedures for arsenic speciation in potentially contaminated waters. Environmental
Geochemistry and Health 29, 337-346.
[56] Bednar, A. J., Garbarino, J. R., Ranville, J. F., and Wildeman, T. R. (2002) Preserving the Distribution of
Inorganic Arsenic Species in Groundwater and Acid Mine Drainage Samples. Environmental Science &
Technology 36, 2213-2218.
[57] Gault, A. G., Jana, J., Chakraborty, S., Mukheijee, P., Sarkar, M., Nath, B., Polya, D. A., and Chatteijee, D.
(2005) Preservation strategies for inorganic arsenic species in high iron, low-Eh groundwater from West
Bengal, India. Analytical andBioanalytical Chemistry 381, 347-353.
[58] Polya, D. A., Lythgoe, P. R., Abou-Shakra, F., Gault, A. G., Brydie, J. R., Webster, J. G., Brown, K. L.,
Nimfopoulos, M. K., and Michailidis, K. M. (2003) IC-ICP-MS and IC-ICP-HEX-MS determination of
arsenic speciation in surface and groundwaters: Preservation and analytical issues. Mineral. Mag. 67,
247-261.
[59] Oliveira, V., Sarmiento, A. M., Gomez-Ariza, J. L., Nieto, J. M., and Sanchez-Rodas, D. (2006) New
preservation method for inorganic arsenic speciation in acid mine drainage samples. Talanta 69, 1182-
1189.
[60] Sanchez-Rodas, D., Oliveira, V., Sarmiento, A. M., Gomez-Ariza, J. L., and Nieto, J. M. (2006) Preservation
procedures for arsenic speciation in a stream affected by acid mine drainage in southwestern Spain.
Analytical and Bioanalytical Chemistry 384, 1594-1599.
[61] Samanta, G., and Clifford, D. A. (2006) Preservation and Field Speciation of Inorganic Arsenic Species in
Groundwater. Water Quality Research Journal 41, 107-116.
[62] Samanta, G., and Clifford, D. A. (2005) Preservation of Inorganic Arsenic Species in Groundwater.
Environmental Science & Technology 39, 8877-8882.
[63] Kryatov, S. V., Rybak-Akimova, E. V., and Schindler, S. (2005) Kinetics and Mechanisms of Formation and
Reactivity of Non-heme Iron Oxygen Intermediates. Chemical Reviews 105, 2175-2226.
[64] Welch, K. D., Davis, T. Z., and Aust, S. D. (2002) Iron Autoxidation and Free Radical Generation: Effects of
Buffers, Ligands, and Chelators. Archives of Biochemistry and Biophysics 397, 360-369.
45
-------
[65] Yoshimi, K., Ryoji, O., and Niro, M. (1968) Oxygen Oxidation of Ferrous Ions Induced by Chelation.
Bulletin of the Chemical Society of Japan 41, 2234-2239.
[66] Seibig, S., and van Eldik, R. (1997) Kinetics of [Fell(edta)] Oxidation by Molecular Oxygen Revisited. New
Evidence for a Multistep Mechanism. Inorganic Chemistry 36, 4115-4120.
[67] Jones, S. S., and Long, F. A. (1952) Complex Ions from Iron and Ethylenediaminetetraacetate: General
Properties and Radioactive Exchange. The Journal of Physical Chemistry 56, 25-33.
[68] Sunda, W., and Huntsman, S. (2003) Effect of pH, light, and temperature on Fe-EDTA chelation and Fe
hydrolysis in seawater. Marine Chemistry 84, 35-47.
[69] Jones, A. M., Griffin, P. J., and Waite, T. D. (2015) Ferrous iron oxidation by molecular oxygen under acidic
conditions: The effect of citrate, EDTA and fulvic acid, Geochimica et Cosmochimica Acta 160, 117-131.
[70] Lee, S. O. (2005) Dissolution of iron oxides by oxalic acid. School of Chemical Engineering and Industrial
chemistry, The University of New South Wales Sydney, Austrailia.
[71] McComas, W. H., and Rieman, W. (1942) The Effect of pH on the Solubility of Calcium Oxalate. Journal of
the American Chemical Society 64, 2948-2949.
[72] Panias, D., Taxiarchou, M., Paspaliaris, I., and Kontopoulos, A. (1996) Mechanisms of dissolution of iron
oxides in aqueous oxalic acid solutions. Hydrometallurgy 42, 257-265.
[73] Blesa, M. A., Marinovich, H. A., Baumgartner, E. C., and Maroto, A. J. G. (1987) Mechanism of dissolution
of magnetite by oxalic acid-ferrous ion solutions. Inorganic Chemistry 26, 3713-3717.
[74] Suter, D., Banwart, S., and Stumm, W. (1991) Dissolution of hydrous iron(III) oxides by reductive
mechanisms. Langmuir 7, 809-813.
[75] Park, J. S. B., Wood, P. M., Davies, M. J., Gilbert, B. C., and Whitwood, A. C. (1997) A Kinetic and ESR
Investigation of Iron(II) Oxalate Oxidation by Hydrogen Peroxide and Dioxygen as a Source of Hydroxyl
Radicals. Free Radical Research 27, 447-458.
[76] Zuo, Y., and Hoigne, J. (1992) Formation of hydrogen peroxide and depletion of oxalic acid in atmospheric
water by photolysis of iron(III)-oxalato complexes. Environmental Science & Technology 26, 1014-1022.
[77] Kocar, B. D., Herbel, M. J., Tufano, K. J., and Fendorf, S. (2006) Contrasting Effects of Dissimilatory
Iron(III) and Arsenic(V) Reduction on Arsenic Retention and Transport. Environmental Science &
Technology 40, 6715-6721.
[78] Jeong, J., and Yoon, J. (2005) pH effect on OH radical production in photo/ferrioxalate system. Water
Research 39, 2893-2900.
[79] Hicks, M., and Gebicki, J. M. (1986) Rate constants for reaction of hydroxyl radicals with Tris, Tricine and
Hepes buffers. FEBSLetters 199, 92-94.
[80] Zuo, Y., and Hoigne, J. (1994) Photochemical decomposition of oxalic, glyoxalic and pyruvic acid catalysed
by iron in atmospheric waters. Atmospheric Environment 28, 1231-1239.
[81] Chen, J., Zhang, H., Tomov, I. V., and Rentzepis, P. M. (2008) Electron Transfer Mechanism and
Photochemistry of Ferrioxalate Induced by Excitation in the Charge Transfer Band. Inorganic Chemistry
47, 2024-2032.
[82] Pozdnyakov, I. P., Kel, O. V., Plyusnin, V. F., Grivin, V. P., and Bazhin, N. M. (2008) New Insight into
Photochemistry of Ferrioxalate. The Journal of Physical Chemistry A 112, 8316-8322.
[83] Mangiante, D. M., Schaller, R. D., Zarzycki, P., Banfield, J. F., and Gilbert, B. (2017) Mechanism of Ferric
Oxalate Photolysis. ACS Earth and Space Chemistry 1, 270-276.
[84] Weller, C., Horn, S., and Herrmann, H. (2013) Effects of Fe(III)-concentration, speciation, excitation-
wavelength and light intensity on the quantum yield of iron(III)-oxalato complex photolysis. Journal of
Photochemistry andPhotobiology A: Chemistry 255, 41-49.
[85] Chen, J., and Rosen, B. P. (2014) Biosensors for Inorganic and Organic Arsenicals, Biosensors 4, 494-512.
[86] G. Howard, A. (1997) (Boro)Hydride Techniques in Trace Element Speciation. Journal of Analytical Atomic
Spectrometry 12, 267-272.
[87] Pitzalis, E., Ajala, D., Onor, M., Zamboni, R., and D'Ulivo, A. (2007) Chemical Vapor Generation of Arsane
in the Presence of 1-Cysteine. Mechanistic Studies and Their Analytical Feedback. Analyiical Chemistry
79, 6324-6333.
[88] Shraim, A., Chiswell, B., and Olszowy, H. (1999) Speciation of arsenic by hydride generation-atomic
absorption spectrometry (HG-AAS) in hydrochloric acid reaction medium. Talanta 50, 1109-1127.
46
-------
[89] Wieteska, E., Ziolek, A., and Drzewinska, A. (2003) Effectiveness of L-cysteine as a pre-reducing agent in
the process of arsenic determination in natural materials by batch system hydride generation atomic
absorption spectrometry (HG AAS). Chem. Anal. (Warsaw, Pol.) 48, 313-325.
[90] Chen, H., Brindle, I. D., and Le, X. C. (1992) Prereduction of arsenic(V) to arsenic(III), enhancement of the
signal, and reduction of interferences by L-cysteine in the determination of arsenic by hydride generation.
Analytical Chemistry 64, 667-672.
[91] Howard, A. G., and Salou, C. (1996) Cysteine enhancement of the cryogenic trap hydride AAS
determination of dissolved arsenic species. Analytica Chimica Acta 333, 89-96.
[92] G. Howard, A., and Salou, C. (1998) Arsenic speciation by cryogenic trap hydride generation atomic
absorption spectroscopy: performance enhancement by pre-derivatization. Journal of Analytical Atomic
Spectrometry 13, 683-686.
[93] Cordos, E. A., Frentiua, T., Pontaa, M., Abrahamb, B., and Margineana, 1. (2006) Optimisation of analytical
parameters in inorganic arsenic (III and V) speciation by hydride generation using L-cysteine as
prereducing agent in diluted HC1 medium. Chemical Speciation & Bioavailability 18, 1-9.
[94] Minakata, K., Nozawa, H., Yamagishi, I., Gonmori, K., Kanno, S., Watanabe, K., Suzuki, M., Ahmed, W. H.
A., and Suzuki, O. (2009) Sensitive determination of arsenite and arsenate in plasma by electrospray
ionization tandem mass spectrometry after chelate formation. Forensic Toxicology 27, 37-40.
[95] Minakata, K., Suzuki, M., and Suzuki, O. (2009) Simple and selective determination of arsenite and arsenate
by electrospray ionization mass spectrometry. Analytica Chimica Acta 631, 87-90.
[96] Stanic, A. R., Jovanic, S. I., Maijanovic, N. J., and Suturovic, Z. J. (2009) The use of L-ascorbic acid in
speciation of arsenic compounds in drinking water. Acta Period. Technol. 40, 165-175.
[97] Tyson, J. F., Offley, S. G., Seare, N. J., Kibble, H. A. B., and Fellows, C. (1992) Determination of arsenic in
a nickel-based alloy by flow injection hydride generation atomic absorption spectrometry incorporating
continuous-flow matrix isolation and stopped-flow pre-reduction procedures. Journal of Analytical
Atomic Spectrometry 7, 315-322.
[98] Nielsen, S., and Hansen, E. H. (1997) Determination of As(III) and As(V) by flow injection-hydride
generation-atomic absorption spectrometry via on-line reduction of As(V) by KI. Analytica Chimica Acta
343,5-11.
[99] Naykki, T., Peramaki, P., Kujala, J., and Mikkonen, A. (2001) Optimization of a flow injection hydride
generation atomic absorption spectrometric method for the determination of arsenic, antimony and
selenium in iron chloride/sulfate-based water treatment chemical. Analytica Chimica Acta 439, 229-238.
[100] Lai, G., Chen, G., and Chen, T. (2016) Speciation of AsIII and AsV in fruit juices by dispersive liquid-
liquid microextraction and hydride generation-atomic fluorescence spectrometry. Food Chemistry 190,
158-163.
[101] Guo, T., Baasner, J., and Tsalev, D. L. (1997) Fast automated determination of toxicologically relevant
arsenic in urine by flow injection-hydride generation atomic absorption spectrometry. Analytica Chimica
Acta 349, 313-318.
[102] Shraim, A., Abu-Yousef, I. A., Kanan, S. M., Abdo, N., Olszowy, H., Petry, S., and Ng, J. (2008)
Quantification of total arsenic in groundwater by HG-AAS using low acid concentration and L-cysteine.
J. Int. Environ. Appl. Sci. 3, 215-223.
[103] Barcelona, M. J., Varljen, M. D., Puis, R. W., and Kaminski, D. (2005) Ground water purging and sampling
methods: History vs. hysteria. Groundwater Monitoring & Remediation 25, 52-62.
[104] Chang, A., Orth, A., Le, B., Menchavez, P., and Miller, L. (2009) Performance analysis of the OneTouch
UltraVue blood glucose monitoring system. J Diabetes Sci Technol 3, 1158-1165.
[105] Luong, J. H. T., Male, K. B., and Glennon, J. D. (2008) Biosensor technology: Technology push versus
market pull. Biotechnology Advances 26, 492-500.
[106] Clarke, S. F., and Foster, J. R. (2012) A history of blood glucose meters and their role in self-monitoring of
diabetes mellitus. Br. J. Biomed. Sci. 69, 83-93.
[107] Turner, A. P. F. (2000) Techview: Biochemistry: Biosensors-sense and sensitivity. Science (Washington, D.
C.) 290, 1315-1317.
[108] Klonoff, D. C. (2007) Benefits and Limitations of Self-Monitoring of Blood Glucose. Journal of Diabetes
Science and Technology 1, 130-132.
47
-------
[109] Klatman, E. L., McKee, M., and Ogle, G. D. (2019) Documenting and visualising progress towards
Universal Health Coverage of insulin and blood glucose test strips for people with diabetes. Diabetes
Research and Clinical Practice 157, 107859.
[110] Okuda-Shimazaki, J., Yoshida, H., and Sode, K. (2020) FAD dependent glucose dehydrogenases -
Discovery and engineering of representative glucose sensing enzymes. Bioelectrochemistry 132, 107414.
[111] Yoo, E.-H., and Lee, S.-Y. (2010) Glucose biosensors: an overview of use in clinical practice. Sensors
(Basel) 10, 4558-4576.
[112] Gochman, N., and Schmitz, J. M. (1972) Application of a new peroxide indicator reaction to the specific,
automated determination of glucose with glucose oxidase. Clin. Chem. 18, 943-950.
[113] Tonyushkina, K., and Nichols, J. H. (2009) Glucose meters: a review of technical challenges to obtaining
accurate results. J Diabetes Sci Technol 3, 971-980.
[114] Clarke, W. L., Cox, D., Gonder-Frederick, L. A., Carter, W., and Pohl, S. L. (1987) Evaluating clinical
accuracy of systems for self-monitoring of blood glucose. Diabetes Care 10, 622-628.
[115] Stoytcheva, M., Sharkova, V., and Panayotova, M. (1998) Electrochemical approach in studying the
inhibition of acetylcholinesterase by arsenate(III): analytical characterisation and application for arsenic
determination. Analytica Chimica Acta 364, 195-201.
[116] Ramanathan, S., Shi, W., Rosen, B. P., and Daunert, S. (1997) Sensing Antimonite and Arsenite at the
Subattomole Level with Genetically Engineered Bioluminescent Bacteria. Analytical Chemistry 69, 3380-
3384.
[117] Merulla, D., Buffi, N., Beggah, S., Truffer, F., Geiser, M., Renaud, P., and van der Meer, J. R. (2013)
Bioreporters and biosensors for arsenic detection. Biotechnological solutions for a world-wide pollution
problem. Current Opinion in Biotechnology 24, 534-541.
[118] Wan, X., Volpetti, F., Petrova, E., French, C., Maerkl, S. J., and Wang, B. (2019) Cascaded amplifying
circuits enable ultrasensitive cellular sensors for toxic metals. Nature Chemical Biology 15, 540-548.
[119] Luque de Castro, M. D., and Herrera, M. C. (2003) Enzyme inhibition-based biosensors and biosensing
systems: questionable analytical devices. Biosensors andBioelectronics 18, 279-294.
[120] Berberich, J., Li, T., and Sahle-Demessie, E. (2019) Chapter 11 - Biosensors for Monitoring Water
Pollutants: A Case Study With Arsenic in Groundwater. Separation Science and Technology (Ahuja, S.,
Ed.), 285-328, Academic Press.
[121] Olechowski, A., Eppinger, S. D., and Joglekar, N. (2015) Technology readiness levels at 40: A study of
state-of-the-art use, challenges, and opportunities. 2015 Portland International Conference on
Management of Engineering and Technology (PICMET), 2084-2094.
[122] Wilson, I. B., and Silman, I. (1977) Effects of quaternary ligands on the inhibition of acetylcholinesterase
by arsenite. Biochemistry 16, 2701-2708.
[123] Siontorou, C. G., and Batzias, F. A. (2010) Innovation in biotechnology: moving from academic research to
product development—the case of biosensors. Critical Reviews in Biotechnology 30, 79-98.
[124] Ringuet, S., Sassano, L., and Johnson, Z. I. (2011) A suite of microplate reader-based colorimetric methods
to quantify ammonium, nitrate, orthophosphate and silicate concentrations for aquatic nutrient
monitoring. Journal of Environmental Monitoring 13, 370-376.
[125] Chen, L., Tian, X., Xia, D., Nie, Y., Lu, L., Yang, C., and Zhou, Z. (2019) Novel Colorimetric Method for
Simultaneous Detection and Identification of Multimetal Ions in Water: Sensitivity, Selectivity, and
Recognition Mechanism. ACS Omega 4, 5915-5922.
48
-------
II I'll III', villi li VI I" " I' || si I i v „ ' I Vh ' " llll ill lili | I |[ ' II
• Book chapter: Biosensors for Monitoring Water Pollutants: A Case Study With Arsenic in Groundwater.
Separation Science and Technology, ed. S. Ahuja, Academic Press, Vol 11 (2019): 285-328.
• A disposable acetylcholine esterase sensor for As(III) determination in groundwater matrix based on 4-
acetoxyphenol hydrolysis. Anal. Methods Vol 11 (2019): 5203-5213.
• Patent application submission: U.S. Patent Application for: "USING 4-ACETOXYPHENOL AS A
SUBSTRATE FOR MODULAR HYDROLASE BIOSENSORS" in the name of U.S. Environmental
Protection Agency, Serial No. 16/793455; Filed February 18, 2020.
• Technology Niche Analysis® A New Reaction Sequence for A Hydrolase Sensor to Detect Arsenic.
Foresight Science & Technology, Project Number: WCT0017TN, June 12, 2017.
49
-------
50
-------
-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED
STANDARD POSTAGE
& FEES PAID EPA
PERMIT NO. G-35
Office of Research
and Development
(8101R)
Washington, DC
20460
Official Business
Penalty for Private Use$300
------- |