This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
23
REMEDIATION Spring 2006
Nanotechnology application to contaminated site remediation, and especially the use of
nanoscale zero-valent iron particles to treat volatile organic compound (VOC)-impacted
groundwater, is now recognized as a promising solution for cost-effective in situ treatment.
Results obtained during numerous pilot tests undertaken by Golder Associates between 2003
and 2005 in North America (United States and Canada) and Europe have been used to present
a synthetic cross-comparison of technology dynamics. The importance of a comprehensive un-
derstanding of the site-specific geological, hydrogeological, and geochemical conditions, the
selection of appropriate nanoscale particles, the importance of monitoring geochemical pa-
rameters during technology application, and the potential of nanoparticle impact on microbial
activity are discussed in this article. The variable technology dynamics obtained during six pilot
Evaluation and remediation of environmental risks arising from contaminated soiland groundwater quality is an issue of worldwide significance. Driven by the dynam-ics of regulatory requirements, transfer of property ownership, site redevelopment,and so on, site subsurface characterization and remediation are often required to beaccomplished in short time frames and at minimal costs. The use of nanotechnologyfor site remediation could potentially provide a solution for faster and more cost-ef-fective site remediation.
Over the last five years, the use of nanoscale zero-valent iron (NZVI) has beentested and finally entered the field of groundwater remediation as an efficient tech-nology for treatment of various compounds (e.g., chlorinated solvents, hexavalentchromium being the main targeted compounds to date, although tests performed onother contaminants appeared successful). Numerous pilot tests undertaken by GolderAssociates in North America (United States and Canada), Europe, and Australia havebeen used to collect a large amount of field data that was interpreted together to es-tablish the first cross-comparison between various field applications. This article pre-sents the results of this comparison, illustrated by the six most representative pilottests undertaken by Golder Associates in various geological and hydrogeological con-ditions, and the steps needed to successfully implement the NZVI field tests, as wellas guidelines that can be used to design full-scale treatments.
Nanotechnology and GroundwaterRemediation: A Step Forward inTechnology Understanding
NZVI TECHNOLOGY
Zero-valent iron is an effective reductant that can treat many contaminants—and it isparticularly effective for treating chlorinated solvents. Chlorinated solvents can be com-pletely reduced to nontoxic compounds such as ethene and ethane, as initially demon-strated by research conducted at the University of Waterloo (1991). Nanoscale zero-va-lent iron particles have been shown to be more reactive and extremely effective becauseof their increased surface area compared to granular iron (Wang & Zhang, 1997). NZVIcan be readily placed in the subsurface in slurry form even by simple means (e.g., viamonitoring wells). However, this rather simplistic implementation concept has to besupported by reliable knowledge and designs, including detailed understanding of thesite conceptual model, selection of appropriate NZVI particles capable of treating thesite-specific constituents within the desired time frame, appropriate design and monitor-ing of the NZVI field test, and data interpretation within the context of the site setting,as further discussed below.
UNDERSTANDING THE SITE SETTING—SITE CONCEPTUAL MODEL
Investigations of geologically complex, high-risk sites can be lengthy and costly, requir-ing comprehensive and detailed investigations of the risks.To minimize or eliminate lia-bilities related to the presence of unacceptable risks to health and/or the environment,remedial measures chosen for implementation should not become cost-prohibitive.Todesign a cost-effective remedy, consultants should focus their remedial approach veryearly in the evaluation on the project site on identifying appropriate risk-based goals,and on the evaluation and design of technologies that will achieve those goals.This ap-proach requires definition of the contaminant mass distribution in the various subsurfaceman-made and geologic units. Remedial technology selection is often driven by the re-duction of the contaminant source mass. An understanding of the contaminant mass dis-tribution from the site characterization or predesign phases allows the selection of themost appropriate delivery options based on factors such as the extent and depth of con-tamination, and contaminant mass.
The site characterization data should be integrated into a conceptual model of thesite (SCM) that is a basic description of how contaminants enter a system, how they aretransported around and within the system, and where routes of exposure to organismsand humans occur. As such, the site characterization data provide an essential frameworkfor assessing risks from contaminants, developing remedial strategies, determiningsource control requirements, and providing solutions for mitigating unacceptable risks.The model can be a three-dimensional representation that integrates information on hy-drogeologic conditions, the contaminant source(s), and the fate and transport of con-taminants.The formation of a conceptual model should begin early in the site character-ization process and be continually refined as new data become available. A well-definedconceptual model, as illustrated in Exhibit 1, will provide a framework for assessing re-medial options, and for testing, early on, the potential efficacy of a certain technology asa remedy for the site.The conceptual model is subsequently relied on to evaluate poten-tial risks from the hazards identified at the site. Common problems encountered duringdevelopment of conceptual models and site characterizations include identification ofthe source, selection of parameter values, and documentation of alternative exposure
Nanoscale zero-valentiron particles have beenshown to be more reactiveand extremely effectivebecause of their increasedsurface area compared togranular iron.
scenarios.Therefore, it is important that simplifications, assumptions, and justificationswithin conceptual model development be clearly documented and refined in such amanner that accounts for the various sources of uncertainty.
SELECTION OF PARTICLES—REACTIVITY
Iron exists universally in the combined state as oxides and carbonates (e.g., magnetite,hematite, limonite, and siderite). Pure (zero-valent) iron is primarily man-made mate-rial derived in a laboratory setting. Chemically pure iron can be produced by hydrogenreduction of iron oxides or hydroxides. For example, zero-valent iron can be made byheating precipitated iron (III) hydroxide in a stream of pure hydrogen.The reduction iscompleted when the temperature is raised to and maintained at 700°C.The iron ob-tained by hydrogen reduction of oxides at relatively low temperatures (360–530°C) istypically in powdered form.
Since 1996, researchers at Lehigh University have produced NZVI using borohy-dride as the reductant for the chemical reduction of ferrous or ferric ion (Fe2�, Fe3�) to
Exhibit 1. Example of comprehensive site conceptual model
zero-valent iron [Fe(0)].The laboratory-prepared iron particles are spherical, with theirsize in the range of 1–200 nm, characteristic of particles formed in solution.
An obvious effect of small particle size is the increased surface area.The smaller theparticle size, the larger the specific surface area (SSA). For a spherical particle with a di-ameter of d, SSA can be calculated by the following equation:
(1)
where � is the density (7,800 kg/m3) of the solid particle.For example, iron powders or iron filings used in permeable reactive barriers
(PRBs) have diameters typically on the order of 0.5 mm and, thus, a theoretical SSA ofabout 1.5 m2/kg. For nanoscale iron particles of 50 nm, the corresponding SSA is about15,000 m2/kg.
As a consequence, NZVI is much more reactive than other iron powders withlarger particle size.The iron reactivity also depends on the percentage of zero-valentiron.The greater the zero-valent iron percentage, the higher the reactivity of thenanoscale iron powder.The presence of ferrous or ferric ion (Fe2�, Fe3�) as a greaterpercentage of the nanoscale iron powder tends to lower the reactivity, since fewer elec-trons are available to participate in the chemical reactions leading to contaminantdegradation.The reactivity of nanoscale iron particles is further increased by adding acoating of a noble metal on the surface of the particles (e.g., palladium), which createsa bimetallic nanoscale particle (BNP).
PILOT TEST LAYOUT DESIGN
Selecting a layout for the injection- and monitoring-well network should be based onconsideration of the conceptual site model as it relates to site stratigraphy and targetcontaminant distribution.The consultant should consider the contaminant distribution,configuration of permeable horizons, and the estimated radius of influence of the injec-tion wells when placing injection and monitoring wells.
The conceptual site model should be a factor in the design of injection and monitor-ing wells. Screened intervals for the injection and monitoring wells should target thestratigraphic intervals selected for treatment to minimize the potential for contaminantdilution or cross-contamination between various depth layers. Accommodating site geo-logical conditions may necessitate monitoring- or injection-well screen lengths that areshorter than those often used for conventional monitoring-well installations. Althoughexisting two-inch monitoring wells can be used for NZVI injection during pilot testing,newly installed wells dedicated to injection should be of sufficient diameter to accom-modate any downhole equipment that may be required for the iron particle injection.
Injection of iron nanoparticles is often accomplished through gravity feed or low-pressure injection using positive displacement pumps (Exhibit 2). Mixing tanks and in-jection equipment need to be sized and configured to reflect the estimated injectionrates and duration of the injection program. Maintenance of iron nanoparticles in sus-pension prior to injection is an important step that allows a larger subsurface spread ofthe nanoparticles.
Selecting a layout for theinjection- and monitoring-well network should bebased on consideration ofthe conceptual site modelas it relates to site stratig-raphy and target contami-nant distribution.
MONITORING INJECTION PROGRAMS
The monitoring program for an NZVI pilot test or full-scale injection program needs toincorporate the parameters used to measure distribution of the iron nanoparticles in thesubsurface. As the quantity of nanoparticles cannot be easily measured in groundwatersamples, the monitoring program often relies on measurements of changes in geochemi-cal parameters (notably, dissolved oxygen [DO], oxidation-reduction potential [ORP],and pH) to infer the presence of nanoparticles in monitoring wells.
Collecting these data, which are used to monitor distribution of iron nanoparticlesin the subsurface during the injection program, can be achieved using conventionalhandheld equipment such as portable pH or ORP meters. However, the labor-intensivenature of field measurements using such equipment often produces a lesser frequency ofmeasurement.This decreased measurement frequency may make it more difficult to ac-
Exhibit 2. Photograph of pressure injection equipment used by Golder
Associates in Canada
curately measure the arrival time of the injected iron nanoparticles, and may decreasethe accuracy of time-of-travel estimates for the injected material.
Another consideration in selecting the method and equipment used to measure geo-chemical parameters during NZVI injection relates to the accuracy as well as the fre-quency of measurement. Accurate ex situ measurements of some geochemical parame-ters, such as temperature, DO, and ORP, may require use of flow-through cells orsimilar devices to minimize atmospheric contact with the groundwater samples and pre-serve sample integrity. Installation and operation of this equipment at multiple monitor-ing locations may prove costly and time-consuming.
As an alternative, groundwater geochemical parameters can be measured using mul-tiparameter data-logging probes (Exhibit 3).These instruments are easily installed inconventional 50 mm (2 inch) diameter monitoring wells and can be configured to auto-matically record measurements of hydraulic head, DO, ORP, temperature, and pH at auser-specified frequency. Although the purchase or rental costs of multiparameter dataloggers often substantially exceeds the cost of conventional handheld meters, the use ofmultiparameter instruments may permit data collection at greater frequency and withless effort than with conventional monitoring equipment, and permit more accurate de-termination of the time of travel of the injected iron nanoparticles from the injection lo-cation to the monitoring wells.
DISCUSSION OF TREATMENT EFFICIENCY
Exhibits 4 and 5 show the percentage reduction of chlorinated compounds measured inthe injection wells (Exhibit 4) and monitoring wells (Exhibit 5) during the six pilot testresults presented (Czech Republic 1—CZ1, Italy—IT, Canada—CN, New Jersey—NJ,New York—NY, and North Carolina—NC).
As shown in Exhibit 4, it is possible to sustain more than 80 percent of total chlori-nated compounds reduction for 300 days under pilot test conditions, and with one sin-
Exhibit 3. Photographs of multiparameter data-logging equipment
used during the NZVI pilot tests
gle injection of a relatively small amount of nanoparticles (only 6 kg of NZVI were usedduring test CZ1).This graph also illustrates the importance of the particles’ reactivityon the long-term effect.Very reactive particles react faster but are consumed rapidly.Less reactive particles reach the same efficiency in the same range of time (between 1and 100 days), but contaminant rebound is much more progressive than when usinghighly reactive particles.
The monitoring wells used to develop Exhibit 5 were located 10 feet downgradientfrom the corresponding injection point.The time delay between the injection event(zero datum) and the reduction observed in the monitoring wells vary with site-specificgeological and hydrogeological conditions (fractured rock or porous media with variablehydraulic conductivity). In general, it is a good assumption to wait at least 100 days toobtain significant results in the 10 feet surrounding the injection point. Results pre-sented for the New York test show a different dynamic, due to the presence of an on-sitepumping well inducing the nanoparticles to flow from the injection point to the moni-toring-well location.
Exhibit 6 uses data presented in Exhibit 4 to summarize the typical dynamic test re-sponse that can be obtained depending on the site-specific conditions, assuming the geo-logical, hydrogeological, and geochemical conditions of the site have been properly iden-
Exhibit 4. % VOC reduction observed in injection wells
tified and that they are compatible with the limits of NZVI technology applicability. Asshown on this figure, a successful pilot test can lead to a range of 40–80 percent VOCreduction persistence one year after a single injection of NZVI particles, the long-termtreatment efficiency being a function of the type of particles used (very reactive oneswith a short lifetime or less reactive ones that have a longer-term effect on contaminantreduction) and contaminant flux.
POTENTIAL NZVI IMPACTS ON MICROBIOLOGICAL ACTIVITY
The application of NZVI is potentially synergistic with certain biological treatments. Manyof the compounds treatable by NZVI are also treatable through biological means, becausethe oxygen-depleting and -reducing conditions developed by NZVI are comparable to theconditions in which anaerobic bacteria develop and proliferate, particularly the bacteriathat facilitate biodegradation of organic compounds, including chlorinated solvents.Whilethe conditions developed by NZVI are oxygen depleting and reducing, in close proximityto the application (i.e., the iron particles), significant reducing conditions are developed(e.g., ORP of –500 to –600 millivolts).The capability of the appropriate bacteria to sur-vive and proliferate under these conditions has not been shown.The most highly reducing
Exhibit 5. % VOC reduction observed in monitoring wells
condition typically found in groundwater environments is methanogenesis, in which theORP is approximately –240 millivolts, substantially lower than the levels developed in thelocalized area of the NZVI particles. However, downgradient of the NZVI application area,less significant reducing conditions will be developed at levels amenable to anaerobic bac-teria capable of degrading chlorinated solvents. In addition, the application of NZVI hasbeen shown to stimulate anaerobic bacteria and eliminate aerobic bacteria, thereby reduc-ing the competition for available electron donors (nutrients).While the geochemical con-ditions are favorable for bacterial growth subsequent to NZVI application, a sufficient elec-tron donor may be required to provide a carbon source to fully support bacterialrespiration and growth. At sites where insufficient electron donor concentrations are pre-sent naturally, this can be accomplished with commonly used amendments such as lactate.
CONCLUSION
As a general conclusion, in all cases, the pilot tests performed by Golder Associates haveshown a dramatic decrease in the concentration of chlorinated solvents over a short timespan after the injection of particles. However, based on various input parameters such asvolume/mass of injected NZVI, type of NZVI (fast, short-term reacting or slow, long-
Exhibit 6. Pilot test data interpretation/optimization
term reacting), concentration of contaminants in the groundwater in the area of the in-jection well(s), hydraulic parameters and type of aquifer, hydrochemical parameters ofthe aquifer, and competitors in the reduction process (e.g., sulfates, nitrates, etc.), theresults of the test vary widely. Some of the more pertinent results of the six pilot testsused for the illustrations are summarized below:
• In the pilot tests, the concentrations of chlorinated solvents have decreased dra-matically during the first hours/days after the injection and have remained lowfor a long period of time that is in agreement with the mass balance of zero-va-lent iron versus the mass of chlorinated hydrocarbons present in and around thetest area, especially in highly conductive aquifers.
• In cases where sulfates and nitrates existed in the aquifer, these have decreased inparallel with the chlorinated solvents; however, this happened at the expense of adiminished effect of the zero-valent iron injection onto the reduction of solvents.
• The testing demonstrated that NZVI particles, when handled and applied prop-erly, move with groundwater away from their injection point and can, thus, beused to treat larger areas of impacted aquifers.
• While a dramatic but short-lasting decrease of VOC concentrations was found infractured bedrock aquifers, under similar injection conditions a slower but steadydecrease of VOC concentrations was measured in primary porosity aquifers.
• The rate of VOC decrease appears to be indirectly proportional to the hydraulicconductivity.The same can be said of the travel velocity; the higher the hydraulicconductivity, the shorter the travel time.
• The persistence of treatment over a long period of time also depends on the typeof particle: BNP reacts quicker but is spent more rapidly; NZVI particles reactslower but have a longer effect.
• Minor, inconclusive changes to the microbial community structure were observedthus far.
REFERENCE
Wang, C. B., & Zhang, W.-X. (1997). Synthesizing nanoscale iron particles for rapid and complete dechlori-
nation of TCE and PCBs. Environmental Science & Technology, 31, 2154–2156.
Christian Macé, Steve Desrocher, Florin Gheorghiu, Allen Kane, Michael Pupeza, and
Ramesh Venkatakrishnan are professional environmental consultants working with Golder Associates
in Canada, Europe, and the United States. Golder Associates is a premier global group of consulting compa-
nies, specializing in ground engineering and environmental science. Golder Associates has experienced
steady growth for more than four decades and has more than 4,500 staff and over 100 offices across Africa,
Asia, Australia, Europe, North America, and South America. www.golder.com
Dr. Wei-xian Zhang is an associate professor of civil and environmental engineering at Lehigh University.
Dr. Zhang is the leading expert on the manufacture, treatment mechanics, and use of the BNP technology and
has led successful field demonstrations of BNP treatment of chlorinated volatile organic compounds at an in-
dustrial site in Trenton, New Jersey. Dr. Zhang is also working on a part-time basis with Golder Associates.