Redox Processes in Polluted Groundwater: Measuring and Interpreting Redox Potentials, Essays (high school) of American literature

Download Redox Processes in Polluted Groundwater: Measuring and Interpreting Redox Potentials and more Essays (high school) American literature in PDF only on Docsity! Background Literature for the PhD course Risk Assessment of Contaminated Sites Vejle, August 2008 Please note that this document includes 3 of the suggested background readings for the course. The remaining suggested background reading is in the form of books, and must be acquired at your local university library or similar… Background reading included in this document: REDOX CHARACTERIZATION 1. Christensen, T.H.; Bjerg, P.L.; Banwart, S; Jakobsen, R.; Heron, G.; Albrechtsen, H.J. (2000): Characterization of Redox Conditions in Groundwater Contaminant Plumes. Journal of Contaminant Hydrology, Vol. 45, pp. 165-241. ECOTOXICOLOGY AND EXPOSURE 2. Baun, A.; Nyholm, N.; Kusk, K.O. (2008): Environmental Risk Assessment of Chemicals. DTU course material. Department of Environmental Engineering, Technical University of Denmark. 3. Swartjes, F.A. (2007): Human Health Risks Due to the Consumption of Vegetables from Contaminated Sites. RIVM report 711701040. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 167 This paper reviews the different approaches described in the literature for determina- tion of redox conditions in contaminant plumes. After an introductory presentation of the concepts of electron activity, redox potentials, redox conditions and redox buffering, the various approaches are described: electrochemical redox potentials, groundwater sample composition with respect to redox-sensitive parameters, hydrogen concentrations in groundwater, volatile fatty acid concentrations in groundwater, sediment characteristics Ž Ž . Ž .iron species, sulphur species, oxidation capacity OXC , reduction capacity RDC , .extractants , microbial characteristics in terms of biomass composition, biomarkers Ž .PLFA , and redox bioassays. 2. Electron activity, redox potential, redox capacity and redox conditions ( ) ( )2.1. Electron actiÕity p´ and redox potential EH Aqueous systems contain no free electrons, but the relative electron activity, as an Ž .intensity parameter, can still be defined Stumm and Morgan, 1996 .  y4p´sylog e 1Ž .  y4p´ gives the hypothetical electron activity, e , and measures the tendency of a system to accept or transfer electrons. In a highly reducing system, the tendency to donate electrons, that is the hypothetical Aelectron pressureB, or electron activity, is relatively large and p´ is low. In contrast, high p´ values indicate a relatively low electron activity and a relatively oxidized system. ŽAny reduction reaction an oxidized species Ox reacts with n electrons to form a .reduced species Red can be written as: Oxqney™Red 2Ž . The Law of Mass Action defines the constant K ) as:  4  4 yn )Red r Ox e sK 3Ž . This leads to: )  4  4p´s 1rn logK q 1rn log Ox r Red 4Ž . Ž . Ž . Since no free electrons are present in the system, the shown reduction reaction must be linked to an oxidation reaction. For reference purposes, the oxidation of hydrogen is used leading to the equation: 0  4  4p´sp´ q 1rn log Ox r Red 5Ž . Ž . Here, p´0 is the standard electron activity of the actual reduction half reaction when coupled to the oxidation of hydrogen under standard conditions. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241168 The electron activity is, via the Nernst equation, linked to the redox potential, E :H p´sE r 2.3RTFy1 6Ž . Ž .H or: 0  4  4E sE q 2.3RTrnF log Ox r Red 7Ž . Ž .H H T is the absolute temperature in Kelvin, R is the gas constant, and F is Faraday’s number. At 258C, 2.3RTFy1s0.059 V moly1. Another useful relation is DGsynFE 8Ž .H where DG is Gibb’s energy of the reaction. Ž . Ž .Eqs. 7 and 8 can be combined yielding an equation for calculating the actual energy yield, DG , for a given redox reaction at in situ conditions:r 0  4  4DG sDG q2.3RT log Ox r Red 9Ž .r where DG0 is Gibb’s energy of the reaction at standard conditions. As an example, the oxidation of H with the reduction of the crystalline FeOOH,2 goethite is given: H q2FeOOHq4Hq™2Fe2qq4H O2 2 Gibb’s energy of this reaction is calculated from: 2 40 2q q 4  4  4DGsDG q2.3RT log Fe r H HŽ .r 2 If the activities of H , Hq, and Fe2q in the system are known, the energy available2 Ž .syDG for the reaction can be calculated and the feasibility of the reaction beingr microbiologically mediated can be evaluated if minimum energy requirements are Ž .known see Section 5 . 2.2. Aquifer-releÕant redox processes Several redox processes are relevant for contaminant plumes in aquifers. Fig. 1 shows Ž .the most relevant half reaction for reduction electron-accepting and for oxidation Ž . 0Ž . Želectron-donating and their corresponding p´ W values standard electron activity of . Ž .the half reaction at pH 7 , as adapted from Stumm and Morgan 1996 . A reduction half Ž .reaction upper part of Fig. 1 can, from a thermodynamical point of view, be combined 0Ž .with any oxidation reaction, if the p´ W of the reduction half reaction is higher than 0Ž .the p´ W of the oxidation half reaction. Note that the redox system of contaminant Ž . Ž y qplumes may involve gases O , N , CH , CO , dissolved components NO , NH ,2 2 4 2 3 4 2q 2q 2y y q. ŽCH O, Fe , Mn , SO , HS , H as well as solids FeOOH, MnO , FeCO ,2 4 2 3 . Ž 2q 2q q.MnCO and components Fe , Mn , NH associated with the solids by ion3 4 exchange. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 169 Ž 0 Ž .. Ž . Fi g. 1. R ep re se nt at iv e re do x ha lf re ac tio ns fo rc on ta m in an tp lu m es an d th ei rs ta nd ar d el ec tro n ac tiv ity at pH s 7 p´ W af te rS tu m m an d M or ga n, 19 96 . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241172 ide minerals requires much greater changes in the coordination environment for iron. Ž .Structural Fe III at the mineral surface must form chemical bonds with adsorbed reductants, organic compounds, reduced metal complexes or with extracellular iron Ž .reductase Quigley and Banwart, 1999 . Subsequent to electron transfer, multiple metal-oxygen bonds in the mineral crystal must then be broken to release structural Ž . Ž . 2qFe II to solution. As a consequence, the reduction of structural Fe III to form Fe ions is a much slower reaction than analogous reaction between the aquo ions. As a further example, the configurational changes that are required in oxidizing aromatic organic compounds to carbon dioxide are even more complex. The ring structure of benzene, e.g., is very different from the structure of the CO gas molecule. The reaction2 mechanism proceeds by multiple electron transfers with substantial activation energies corresponding to the significant structural rearrangements that must take place as electrons are removed from the carbon atoms in the benzene ring. There are a large number of chemical, mineralogical and microbiological factors that play a role in electron-transfer kinetics of redox processes in groundwater. Slow abiotic kinetics due to the high activation energies for structural rearrangement are an important reason that many redox processes in contaminant plumes require microbial catalysis to proceed at appreciable rates. Enzyme systems act to stabilise intermediates, and facilitate the step-wise transfer of electrons to and from intermediates, with stable reaction products as the end result. Abiotic redox reactions involving dissolved iron often proceed at measurable rates in the presence of suitable oxidants and reductants. Redox processes involving C, S and N, often require microbial catalysis due to the multiple electron transfers required to reach stable products, and the extensive structural rear- rangement between reactants and products. 2.4. Redox capacitiesrredox buffering Ž .In addition to the intensity parameter, the electron activity p´ or the theoretical Ž .redox potential E , the redox properties of the system also include its capacity toH Žbuffer the effects of entrance of reduced or oxidized components Scott and Morgan, .1990 . In theory, the OXC of a volume of an aquifer can be expressed by the sum of oxidized equivalencies, that potentially could be reduced within a contaminant plume, minus the sum of reduced equivalencies, that could be oxidized in an aquifer system Ž .Scott and Morgan, 1990 . This definition would for the redox system be an analogue to alkalinity for acid-base systems. However, since internal redox equilibrium is uncom- mon it is more practical to define the OXC as the sum of oxidized equivalencies and the RDC as the sum of reduced equivalencies within a volume of the plume: OXCsÝ Oxidized equiv. 10Ž . Ž . RDCsÝ Reduced equiv. 11Ž . Ž . The relevant redox species are defined by their presence as redox species in pristine or strongly contaminated aquifers. Tables 3 and 4 show sample calculations of OXC in two ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 173 Table 3 Ž 3 Ž . .Calculated OXC milliequivalents per dm litre of aquifer for two aerobic aquifers Ž . Ž .Species Reduction Vejen DK Sand Ridge IL, USA Content OXC Content OXC 3 3Ž . Ž .meqrdm meqrdm Ž . Ž . Ž .O aq O 0 ™O yII 10 mgrl 0.44 9 mgrl 0.392 y Ž . Ž . Ž .NO aq N V ™N 0 15 mgrl 1.9 1 mgrl 0.123 Ž . Ž . Ž . Ž .Mn IV s Mn IV ™Mn II 0.2 mgrg 12 0.4 mgrg 23 Ž . Ž . Ž . Ž .Fe III s Fe III ™Fe II 6 mgrg 175 6.8 mgrg 200 2y Ž . Ž . Ž .SO aq S VI ™S yII 40 mgrl 1.2 36 mgrl 1.14 Ž . Ž . Ž .CH O s C 0 ™C yIV 0.2 mgrg 110 0.4 mgrg 2202 The contents of the oxidized species are per litre of groundwater and grams of sediment, respectively, and converted to dm3 of aquifer assuming a porosity of 0.35 and a dry bulk density of 1.6 kgrdm3. Partly based Ž . Ž .on Barcelona and Holm 1991 and Heron et al. 1994a . aerobic, uncontaminated aquifers and of RDC in two strongly polluted anaerobic plumes, respectively. The lack of internal equilibrium allows an aquifer volume to have both OXC and RDC at measurable levels, although this is not illustrated by the cases in Tables 3 and 4. Of the inorganic species, iron and, to some extent, manganese contribute significantly to the OXC in the pristine aquifer, while contributions by dissolved species are very minor. In the contaminant plume, reduced iron and sulphur species as well as ammo- nium contribute significantly. Organic matter, as the model component CH O with a2 carbon oxidation state of zero, is included in both the OXC and the RDC calculated for Table 4 Ž 3 Ž . .Calculated RDC milliequivalents per dm litre of aquifer for two anaerobic contaminant plumes Ž . Ž .Species Oxidation Leachate plume Vejen, DK Leachate plume Grindsted, DK Content RDC Content RDC 3 3Ž . Ž .meqrdm meqrdm Ž . Ž . Ž .CH aq C yIV ™C IV 20 mgrl 3.5 15 mgrl 2.64 Ž . Ž . Ž .CH O s C 0 ™C IV 600 mgrkg 320 200 mgrkg 1062 Ž . Ž . Ž .CH O aq C 0 ™C IV 200 mgrl 23 75 mgrl 8.82 Ž . Ž . Ž .Reduced S s S yI ™S VI 500 mgrkg 200 100 mgrkg 40 y Ž . Ž . Ž .HS aq S yII ™S VI 0.5 mgrl 0.04 0.2 mgrl 0.02 Ž . Ž . Ž .Reduced Fe s Fe II ™Fe III 1000 mgrkg 29 250 mgrkg 7.3 2q Ž . Ž . Ž .Fe aq Fe II ™Fe III 50 mgrl 0.3 200 mgrl 1.3 q a b bŽ . Ž . Ž .NH s, iec. N yIII ™N V 20 mgrkg 14 20 mgrkg 144 q Ž . Ž . Ž .NH aq N yIII ™N V 40 mgrl 6 25 mgrl 44 Ž . Ž . Ž .Reduced Mn s Mn II ™Mn IV 50 mgrkg 3 15 mgrkg 0.9 2q Ž . Ž . Ž .Mn aq Mn II ™Mn IV 10 mgrl 0.1 5 mgrl 0.06 The contents of the reduced species are per litre of groundwater and grams of sediment, respectively, and converted to dm3 of aquifer assuming a porosity of 0.35 and a dry bulk density of 1.6 kgrdm3. Data is partly Ž . Ž .based on Heron et al. 1994b, 1998 , Bjerg et al. 1995 , and partly assumed by the authors. a iec., Ion exchanged onto aquifer sediment. bThe content is assumed by the authors. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241174 the samples. Actually, organic carbon may have an oxidation state varying from Ž . Ž .C yIV in methane to C III in formic acids, so the organic model component in the solids represents an average well-matured organic pool. It is, however, questionable to which extent this pool is reactive and significant in the context of a contaminant plume. Ž .Griffioen et al. 1999 showed that the bulk organic matter of aquifer sediments had a Ž .measurable RDC against oxygen, while Jakobsen and Postma 1994 showed that organic matter in aquifers, in general, had a very low reactivity. Likewise, some uncertainty exists about the availability of the solid iron as well as sulphur components. To illustrate the development in p´ and pH as a volume of aerobic aquifer is titrated Ž .by dissolved organic matter, Scott and Morgan 1990 carried out a series of equilibrium calculations, as shown in Fig. 3. All the redox buffering was gone after about 500 mm carbon was oxidized, suggesting that the model redox system had very little buffer capacity. For example, compared to the actual sandy aquifers characterised in Table 3, Ž .the iron oxide content of the Vejen aquifer solids 6 mg Ferg of dry sediment would provide a buffer capacity per litre of aquifer that would match on the order of 120 000 mm carbon corresponding to 1400 mg Crl in the porewater contained in 1 l of the aquifer. However, Fig. 3 still shows how the p´ of the system develops in a stepwise fashion, assuming the redox reactions establish their equilibria fast. The redox buffering takes place on the horizontal plateaux of the p´ diagram as p´ does not change as DOC is being oxidized by a given electron acceptor. When the electron acceptor is depleted, the p´ drops rapidly until the level of the next redox couple is met. The main drop in p´ happens between manganese and iron reduction as expected from the difference in 0Ž . Ž .p´ W values of the two redox couples see Fig. 1 . The redox buffering plateaux for Ž . Ž .Fig. 3. Model calculated developments in p´ a and pH b as a function of dissolved organic matter titrating Ž . Ž . Ž .a fictitious groundwater volume: Numbered segments correspond to sequential reduction of 1 O aq , 22 y Ž . Ž . Ž . Ž . Ž . Ž . Ž . 2y Ž . Ž .NO aq , 3 MnO s , 4 Fe OH s and 5 SO aq from Scott and Morgan, 1990, with permission .3 2 3 4 ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 177 3. Electrochemical measurement 3.1. Background The definition on electron activity presented in Section 2 and the link to the redox potential, via the Nernst equation, may suggest that immersing two electrical leads in a groundwater sample and connecting them to the input of a voltmeter may well result in a measurable electrical potential that is associated with the redox condition of the sample. Electrochemical redox measurements are easily performed and have been successfully applied in analytical chemistry to aquatic solutions of pure chemicals. However, prior to application of electrochemical redox measurements to assessing complex biological and geochemical processes in a plume of groundwater contamination, a number of difficult questions must be addressed. To what solutes and processes do such measurements respond? Do measured potentials correspond to rigorously defined electrochemical potentials and, thus, free energies of relevant reactions? If so, is the information useful to assessment of groundwater redox conditions? There is an enormous difference between the thermodynamic concept of an electro- chemical potential, and a measured electrochemical potential. The first is a theoretical construct, while the second results from a practical exercise in analytical chemistry. Under certain conditions, the two can be associated in a meaningful way to help assess groundwater redox conditions. Often, electrochemical redox measurements, as referred in the literature, are either considered useless or interpreted beyond their reliability. This probably is due to lack of understanding of the basic concepts and, therefore, of the possibilities and limitations of the measurements. The classic paper by Lindberg and Ž .Runnells 1984 abolished meaningful interpretation of measurements of electrochemical redox potentials due to lack of internal equilibrium. Still, redox potentials have been helpful in field monitoring to identify strongly reducing conditions and, when measured with care as discussed in the following sections, can be interpreted quantitatively with respect to the redox speciation of dissolved iron. 3.2. Principles The electrochemical measurement involves two electrodes, a nonpolarisable reference electrode and a polarisable working electrode. If a small potential is applied, current will flow at the working electrode as redox-active ions undergo oxidation and reduction at the electrode surface. The working electrode will respond to kinetically fast redox couples and the potential between the working electrode and the reference electrode, measured when no current flows, corresponds to electrochemical equilibrium for the redox couple. A review of the basic electrochemical principles behind direct measure- ments of redox potentials can be found at www.shef.ac.ukr;gprg. The main problem in applying these principles to measuring redox potentials in polluted groundwater is the very slow chemical kinetics of redox reactions involving C, S and N species. Actually, the only redox couple that may respond kinetically fast and have sufficient ion activity is the Fe3qrFe2q redox couple. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241178 Although a reading of electrochemical redox potential can be obtained for most anaerobic groundwater samples, the lack of internal equilibrium and the slow kinetics of most redox couples, may make the reading meaningless unless it is based on measurable ion activities within the Fe3qrFe2q redox couple and this redox couple is at equilib- rium. Pitfalls and consequences in electrochemical measurements of redox potentials are summarized in Table 5. 3.3. Methods Redox measurements are not valid in aerobic groundwater due to the presence of O2 Ž .www.shef.ac.ukr;gprg . Although most contaminant plumes are anaerobic, it is important for measuring reliable redox potentials that the sample is unchanged when pumped from the well to the measuring cell by using gas impermeable pumps and tubing. Specifications for measurement cell and electrode configurations are available in Žthe literature Walton-Day et al., 1990; Grenthe et al., 1992; Lyngkilde and Christensen, .1992 . A standard configuration is a gas-tight cylinder through which groundwater can pass, and into which electrodes are inserted while maintaining a gas-tight seal around openings for flow connections and electrodes. Measurement must take place in a gas impermeable cell since diffusion of oxygen and slow oxidation of ferrous iron species can result in mixed potentials that drift slowly. Since it is essentially impossible for the cell to be completely gas impermeable, insurance against mixed potentials is provided by a continuous flow of anaerobic groundwater past the measuring electrode such that the concentration of dissolved ferrous iron species is representative of the groundwater. An important precaution against mixed or otherwise unrepresentative potentials is to record measured potentials with time, until a stable potential is obtained. This ensures Table 5 Pitfalls and their consequences in determination of electrochemical redox potentials in contaminant plumes Ž .after Grenthe et al., 1992 Pitfalls and consequences Measurement of non-equilibrium potentials Incorrect application of measured potential in thermodynamic calculations Measurement of mixed potentials Incorrect assignment of measured potential to a particular redox pair Incorrect application of measured potential in thermodynamic calculations Measurement in oxic or sub-oxic groundwaters Drifting or fluctuating potentials Incorrect assumption that measured potential reflects O rH O equilibrium2 2 Incorrect assumption that measured potential reflects redox status of groundwater Measurement below detection limits due to low ion actiÕities Drifting or fluctuating potentials Incorrect assumption that measured potential reflects redox status of groundwater Incorrect assignment of measured potential to a particular redox pair Incorrect application of measured potential in thermodynamic calculations ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 179 that there is no drift due to traces of O , and it indicates that the composition of the2 flowing groundwater is not changing dramatically and the measured potential is repre- sentative for the groundwater in the well capture zone. Fig. 5 shows data obtained using a standard flow cell configuration. Data were recorded after a purge time that was suitable to flush the tubing and pumps of any residual air or groundwater. Stable readings in conductivity, pH and E were then obtained by the electrodes in the cellH within about 30 min. Shorter or longer times are possible, depending on pumping rates, groundwater composition and its variability and the design and performance of the sampling system, measurement cell and electrodes. An additional refinement for obtaining reliable potentials is to simultaneously record potentials obtained with two different electrodes with time. Fig. 5 compares potentials for simultaneous measurements using platinum and glass carbon electrodes housed in a Ž .Fig. 5. a Establishment of stable redox potential reading with time using a flow through cell receiving Žpumped discharge from a well within a phenolic contaminant plume in a sandstone aquifer unpublished data . Ž .courtesy of S. Thornton, The University of Sheffield, UK . b Long-term continuous measurement of redox Žpotentials using comparative measurements from a glassy carbon and a platinum electrode unpublished data .courtesy of Swedish Nuclear Fuels and Waste Management, Stockholm . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241182 Fig. 7. Redox potentials measured by a platinum electrode in a transversal section of the plume at the Ž . Ž .Wurtsmith site MI, USA from Chapelle et al., 1996, with permission . Žand hydrogen concentrations, showed values in the rangey64 toy158 mV except one .value of y32 mV . Four sampling points showing oxygen above 0.5 mgrl revealed Ž .positive redox potentials. The authors, Chapelle et al. 1996 , found some inconsisten- cies among the different redox-indicating parameters and warned against interzonal mixing by pumping wells. Ž .Lendvay et al. 1998a,b combined flow-cell measurement of redox potentials with Ž .hydrochemical data for depth profiles at the St. Joseph site MI, USA to elucidate redox environments and biological processes within a shallow contaminant plume and their seasonal dynamic behaviour. They observed a sharp drop from positive to negative redox potentials across the dissolved O gradient at the plume fringe. Due to the2 abundance of iron mineral phases in the aquifer sediment and the existence of significant Žconcentrations of dissolved iron, redox potentials below y40 mV read from the .published illustration seem to indicate strongly reducing conditions, important for the reduction of chlorinated solvents in the plume. The data shown in Fig. 5 were applied to model the aqueous iron speciation, and the Ž .solubility and redox equilibria of ferrous clays sampled at that site, as a sink for Fe II Ž .produced during microbial iron reduction by organic carbon species Banwart, 1999 . The measured potentials proved essential to understanding the iron mass balance at the site, which helped quantify the extent and rate of biodegradation processes. Because iron oxide, carbonate, silicate and sulphide mineral phases are potential reactants and products during anaerobic biodegradation processes, understanding and quantifying the iron system can be of tremendous importance when attempting to close electron and mass balances based on field data, and to quantify the associated in situ rates of Ž .degradation processes. Jakobsen et al. 1998 recently outlined how the equilibrium redox potential for TEAPs can help quantify conditions where simultaneous TEAP occur ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 183 Ž .overlapping iron- and sulphate-reducing and methanogenic zones . In this case, redox potential measurements coupled with hydrochemical data provides the only possibility to assess the mineral solubility and associated redox equilibria of the phases present. 3.5. EÕaluation In the applications described above, redox potentials were used in two different senses. In the first three cases, indicative measurements were used qualitatively to identify strongly reducing zones within a plume. Electrochemical measurements in contaminant plumes are likely to respond primarily to the iron redox couple and, at neutral pH values, measurements belowy50 mV seem to suggest that strongly reducing Ž .conditions iron-reducing, sulphate-reducing or methanogenic prevail in the plume. However, lower values cannot be interpreted as representative of more reducing conditions, such as sulphate-reducing or methanogenic conditions. Depending on the methodology, electrochemical measurements may also yield information on iron specia- tion in the plume. In the fourth case, considerable effort went into obtaining reliable, long-term and continuously measured redox potentials at a single discharging well. Because the objectives for that research project focused strongly on the thermochemical stability and redox behaviour of iron-bearing mineral phases within the well capture zone, such measurements were essential to interpretation of relevant processes at the site. It seems clear that redox potential measurements can be routinely carried out at field sites, within a standard sampling protocol, and may help to at least rapidly distinguish Žbetween strongly reducing zones iron-reducing, sulphate-reducing and methanogenic .conditions and zones representing higher redox conditions. In some cases, where measurements have been made carefully and methods documented, a more quantitative interpretation may be possible, although limited to helping interpret iron speciation in the groundwater. Because of the difficulties in obtaining reliable quantitative redox potential data, the cost in time and equipment needs to be weighed against the potential value of the data to a particular field problem. 4. Groundwater composition 4.1. Background The redox half reactions presented in Fig. 1 and the overall redox reactions exemplified in Tables 1 and 2 involve reactants and products that are present in groundwater as dissolved ions or as dissolved gases. Outside a narrow interval contain- ing the standard redox potential of a given reaction, either the reactants or the products dominate and their presence, therefore, reflects the current redox conditions. The primary redox-sensitive species in groundwater are the dissolved ions SO2y, HSy, Fe2q,4 Mn2q, NHq, NOy, NOy and the dissolved gasses CH , N O and O , but also other4 2 3 4 2 2 Ž .components in groundwater e.g., DOC and organic N may reflect the redox levels in ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241184 contaminant plumes. Since groundwater sampling is used extensively to characterize the Ž .plumes in terms of pollutants e.g., BTEX and chlorinated aliphatic compounds , it is obvious, also to sample and analyse for redox-sensitive species in the same locations. 4.2. Principles Assuming that both the oxidized and the reduced species of the important redox processes in groundwater could be quantified analytically with good accuracy and that the solute species were in equilibrium with the solid species, a redox potential could be Ž .calculated for each redox couple according to Eq. 7 . However, this is usually not Ž .possible for various reasons. 1 Analytical methods may have detection limits too high to quantify both species of a redox couple due to a very low concentration of one of Ž Ž . Ž .. Ž .them e.g., Fe II rFe III . 2 Analytical methods may not exist for both redox species Ž Ž . Ž ..e.g., Mn II rMn IV . Lack of analytical detection of one species of a redox couple Ž .may be circumvented by measuring total concentration e.g., Fe and one of the species Ž Ž ..e.g., Fe II in the redox couple and then assuming that the other species makes up the Ž .difference Nicholson et al., 1983 . However, this introduces significant uncertainty for Ž .low concentrations. 3 Several redox-sensitive species are also precipitates and the dissolutionrprecipitation reactions are not necessarily at equilibrium or the actual Ž 3q .precipitate is not known e.g., Fe rFeOOHrFe O . Therefore, the quantification of2 3 redox-sensitive species usually cannot be used to calculate a redox potential, but the redox levels can still be deduced paying attention to the geochemistry of species, as done below for the hierarchy of electron acceptors. Issues related to sampling and behaviour of the most commonly used redox-sensitive species are summarized in Table 7. Aerobic conditions require that free O is present2 Ž .and O can in practice standard wells and sampling equipment with some caution be2 measured reliably down to at least 0.5–1 mgrl. Lower concentrations can be measured, but it is hard to tell whether the reading represents the oxygen concentration in the Žplume or is caused by intrusion of oxygen during pumping and sampling e.g., by .diffusion through plastic tubing, Kjeldsen, 1993 . Clearly aerobic conditions are, thus, Ž .fairly easy to identify, but microaerophilic low oxygen concentrations are hard to document. Denitrifying conditions do not necessarily require that measurable nitrate is present, since the process may limit the nitrate concentration below detection limits. Denitrification may, thus, be determined as decreased nitrate concentrations andror the presence of reduced intermediates as, e.g., N O. Manganese reduction is recognized by2 Žthe presence of dissolved manganese. Oxidized manganese is present as a solid e.g., .MnO and has such a low solubility at near neutral groundwater pH that all soluble2 manganese above background concentrations, for all practical purposes, can be assumed to be Mn2q. Presence of dissolved iron may indicate iron reduction as dissolved iron can Ž . Ž .be assumed to be Fe II at neutral pH. Analytical methods do exist for Fe II and the Ž .validity of this assumption has been proven Postma and Lyngkilde, 1990 . Sulfate reduction may be detected as decreased sulphate concentrations andror the presence of hydrogen sulphide. However, sulphide has very low solubility if Fe2q is present and analytical detection may be difficult. Methanogenesis results in production of CH ,4 ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 187 atmosphere. Similarly, samples used for analysis of compounds where solid species may Ž 2q . Žinterfere Fe , Fe, Mn should be properly filtered. Dissolved gases H S and CH ,2 4 .and O when measured by the Winkler method should be sampled in an unbroken2 streamline and transferred to gas-tight containers without contacting air until analysis is complete. Samples should not be filtered prior to analysis of these compounds as the filtration process may cause significant losses. Preservation of samples to be analysed in Ž 2q .the laboratory should be done in the field to avoid precipitation Fe , Mn , oxidation Ž 2q q. Ž y .e.g., H S, Fe , Mn, NH or conversion of the compounds NO , CH . Preferably,2 4 3 4 analysis of Fe2q and H S should be performed in the field, as storage of such labile2 compounds may be very difficult. 4.4. Applications The use of redox-sensitive compounds for deducing redox conditions is, as indicated Žabove, no easy task; however, the advantages easy analysis and low cost, assuming that .the groundwater is sampled anyway call for use in most cases. Redox-sensitive compounds have been used in at least three different ways: Ø Identification of reduced and oxidized conditions, Ø Assignment of redox zones, Ø Determination of predominant redox reactions. Identification of reduced and oxidized conditions in a plume follows closely the lines Ž .presented in the section on principles Section 4.2 . This has been presented in many different ways for contaminated sites. A comprehensive discussion of redox couples in Ž . Ž .the Borden Landfill CND leachate plume is presented in Nicholson et al. 1983 . A recent example presents the redox conditions in a groundwaterrsurface water interface Ž .contaminated by chlorinated ethenes Lendvay et al., 1998a,b . Amirbahman et al. Ž .1998 described aqueous phase data as indicators of redox processes beneath and Ž .downgradient the Winterthur Landfill CH . Evaluation of dynamic changes is of great interest and here the redox-sensitive compounds may be a good approach, because Ž . 2qFig. 8. Concentration mM of Fe , Mn and CH in groundwater downgradient an oil spill at the Bemidji4 Ž . Ž .MN, USA site after Baedecker et al., 1993 and Baedecker and Cozzarelli, 1994, with permission . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241188 Ž .repeated sampling in the same point is possible opposed to sediment samples . Detailed monitoring of temporal or spatial changes has been done during an anaerobic field Ž . Ž .injection experiments at the Grindsted Landfill site DK Albrechtsen et al., 1995 . At Ž .the Bemidji site MN, USA , contaminated by crude oil, the spatial and temporal trends Žin groundwater chemistry were investigated in the downgradient wells e.g., Fig. 8, . Ž .Baedecker et al., 1993 . Significant increases in methane and Fe II were observed indicating a change in the overall redox conditions from aerobic to anaerobic conditions due to the oil spill. In general, the data presented in the literature support the usefulness of characterizing groundwater samples in terms of redox-sensitive species, but the limitations with respect to species interacting with solid phases are apparent. Generally, evaluation of redox conditions based on measurements of only a few redox-sensitive compounds in a few wells should only be taken as indicative. Assignment of redox zones to leachate plumes has been done in several cases. The early cases separated leachate plumes, on the basis of nitrogen compounds and oxygen, Žinto an anaerobic zone, a transition zone and an aerobic zone Golwer et al., 1969; .Baedecker and Back, 1979a,b . In the latter case, the ratio between Kjeldahl N and Table 8 Ž .Criteria used for assigning redox conditions to groundwater samples from the Vejen Landfill DK leachate Ž . Ž . Ž .plume Lyngkilde and Christensen, 1992 and Grindsted Landfill DK leachate plume Bjerg et al., 1995 Parameter Methanogenic Sulfate- Iron- Manganese- Nitrate- Aerobic reducing reducing reducing reducing Methane )1 -1 -1 -1 -1 -1 )25 – – – – -1 Sulfide – )0.2 -0.1 -0.1 -0.1 -0.1 – )0.1 – – -0.1 -0.1 Sulfate -40 – – – – – – – – – – – Dissolved iron – – )1.5 -1.5 -1.5 -1.5 -150 -150 )150 -10 -10 -1.5 Dissolved manganese – – – )0.2 -0.2 -0.2 -5 -5 -5 )5 -0.2 -0.2 Ammonium – – – – – -1 – – – – – -1 Dinitrogene oxide NI NI NI NI NI NI -1 -1 -1 -1 )1 – Nitrite -0.1 -0.1 -0.1 -0.1 )0.1 -0.1 -0.1 -0.1 -0.1 -0.1 )0.1 -0.1 Nitrate -0.2 -0.2 -0.2 -0.2 – – -0.2 -0.2 -0.2 -0.2 – – Oxygen -1 -1 -1 -1 -1 )1 -1 -1 -1 -1 -1 )1 The upper criteria are for the Vejen Landfill and the lower criteria are for the Grindsted Landfill. All concentrations in mgrl, except dinitrogenoxide in mgrl, nitrate, nitrite and ammonium in mg Nrl and sulfate in mg Srl. –, No criterion defined. NI, Not included in the criteria for the Vejen Landfill. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 189 nitrate was also used for characterising the plume. The concept of assigning redox zones on groundwater redox-sensitive compounds was fully developed by Lyngkilde and Ž . Ž .Christensen 1992 , but Chapelle et al. 1995 also used redox-sensitive compounds for characterization of redox conditions. Lyngkilde and Christensen analysed the data on redox-sensitive compounds from 366 groundwater samples from the Vejen Landfill Ž .DK leachate plume. Each sample was given a specific redox label based on the criteria given in Table 8. The concept was based on the ideas presented Table 7, but the actual concentration levels were selected from an evaluation of leachate and plume composi- Ž . Ž .Fig. 9. Identified redox zones at the Vejen Landfill DK Lyngkilde and Christensen, 1992 and the Grindsted Ž . Ž .Landfill DK from Bjerg et al., 1995, with permission . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241192 redox process in a specific section of the plume or the actual rate of the process can be Ž .estimated Amirbahman et al., 1998; Ludvigsen et al., 1998 . The major drawback of this method is a clear underestimation of the iron-reduction and manganese-reduction processes as the presence of ferrous iron and manganese in the aqueous phase only Ž .represent a minor fraction of the total reduced pools see discussion in Section 7 . Likewise, sulphate reduction can be underestimated if sulphide is used for the calcula- Ž .tion sulphide is likely to precipitate ; thus, dissolved sulphate should be the preferred parameter unless there are indications of sulphate sources, such as dissolving barite or gypsum or oxidizing sulphides. 4.5. EÕaluation Quantifying redox-sensitive species in groundwater samples is a simple and useful tool for identifying the redox conditions of a groundwater contaminant plume. The basic principles are thermodynamically sound; but the actual criteria for assigning the redox status depend, to some extent, on local conditions, such as the natural groundwater geochemistry, compounds leaching from the source and quality of sampling and analytical equipment. The major limitation in using the concept is related to migration of redox-sensitive species away from active zones, geochemical processes involving pre- cipitation of compounds and the actual overlap between different redox zones. The transport of reduced redox-sensitive species, e.g., methane and ferrous ion, will smear out the redox zones identified by groundwater sample composition. Presumably, the strongly reduced zones will be slightly overestimated. Precipitation and ion exchange of iron and manganese and precipitation of sulphur species will also affect the identifica- tion of the redox state in a given point. For example, iron and sulphate reduction, based on the groundwater composition, may be suggested in areas where the presence of reduced sulphur and ferrous iron primarily is due to dissolution of solid precipitates rather than actually on-going redox processes. However, precipitation of reduced sulphur, iron and manganese may also cause underestimation of significant redox processes if solely quantified in the groundwater. The above limitations cannot be avoided, but a general awareness of the potential problems can limit mistakes and put the interpretation and use of the identified redox zones into the correct perspective. Redox-sensitive species have been successfully applied for characterization of redox environments in contaminant plumes in sandy aquifers. Identifying a full sequence of redox zones requires many high-resolution samples from short screens. The actual assignment of redox zones will be influenced by the specific criteria set in each case and the zones identified are, therefore, only indicative of the actual redox conditions. However, despite these limitations, the approach seems useful and further applications in different types of contaminant plumes would be interesting. Changes in concentrations of redox-sensitive species in groundwater have been widely used for calculating the dominant electron acceptor processes. However, a simple calculation based on the groundwater composition alone will often be erroneous, because the solid phases are neglected. This is, in particular, problematic with respect to iron reduction, because high amounts of reduced iron may have precipitated on to the solids. This should be taken into account in the future. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 193 5. Hydrogen 5.1. Background The use of H , as a general indicator of the predominant redox process, was2 Ž .introduced by Lovley and Goodwin 1988 . The paper indicated that each anaerobic redox process was characterized by a well-defined range of H concentrations. The2 proposed theory suggested that in a steady-state system, limited by the availability of organic matter, the H concentration is constant and controlled by parameters related to2 the physiology of the mediating bacteria, which again is related to the redox process they are mediating. The physiological parameters were also assumed to be related to the Ž .energy available at AstandardB pHs7 conditions. This, in turn, suggested that redox zonation was a result of the competition for the limited substrate, where microorganisms able to use lower H concentrations, e.g., Fe-oxide-reducing bacteria, out-compete2 microorganisms which demand higher H concentrations, e.g., sulphate-reducing bacte-2 ria. This competition would lead to characteristic hydrogen concentrations for different redox zones. The concept was applied to pristine and polluted groundwater by, e.g., Ž . Ž .Chapelle and McMahon 1991 and Chapelle et al. 1995 . However, the concept of characteristic H levels neglects the influence of solute concentrations, the stability of2 iron oxides and temperature, shown later, and now seems of less general value. Recent studies have shown that H values reflect the energetics of the system studied2 Žmuch more precisely e.g., Hoehler et al., 1998; Jakobsen et al., 1998; Jakobsen and . Ž .Postma, 1999 than assumed by Lovley and Goodwin 1988 . As discussed below, this implies that a given measured H concentration does not necessarily indicate a given2 redox process. Theoretically, two different redox processes may, depending on substrate and product activities, lead to the same H concentration. Still, the measurement of H2 2 is an extremely valuable tool in analyzing the energetics of microbiological processes Ž .Hoehler, 1998 , and it appears to enable the integration of redox processes in a Ž .thermodynamically based description of a given system Jakobsen and Postma, 1999 . 5.2. Principles As shown in Fig. 2, the anaerobic oxidation of organic matter from complex organic compounds generally goes through a three step process. After an initial hydrolyzation of the organic substances, the evolving substances are fermented into smaller organic molecules, such as lactate, propionate, butyrate, acetate and formate, as well as CO and2 H . These fermentation products are then used as electron donors in the TEAPs by the2 bacteria mediating the TEAPs. One of the dominant products is H . The relative2 distribution of products is a function of the concentration of these, so at low H2 Ž .concentrations relatively more H is produced Dolfing, 1988 . The described stepwise2 Ž .degradation is not necessary in aerobic systems Fig. 2 ; exactly what hydrogen measurements from the aerobic zone relate to is currently not clear. The initial hydrolyzation has experimentally been shown to be overall rate limiting Ž .for the rate of methanogenesis Gujer and Zehnder, 1983 , implying that the supply of ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241194 fermentation products is limited. This is also indicated in a more general sense by the extremely low concentrations, in the nM range for H and mM range for acetate2 measured in groundwater, compared to the mM changes in concentrations observed for Ž .the other reactants Postma and Jakobsen, 1996 . This implies that the concentration of the intermediates is under the control of the bacteria carrying out the TEAPs. H is2 special in this context in several ways. The nM concentrations of H in groundwater2 makes the pool of H extremely small compared to the rate of production and2 consumption of this very important intermediate. The residence time is, thus, on the Ž .order of seconds at high TEAP rates )30 mMryear , to minutes in systems with low Ž .TEAP rates -0.5 mMryear . Therefore, H concentrations must reflect ongoing2 processes, as transport is extremely limited within this time frame. Even in a groundwa- ter system with a high flow rate of 1 mrday, a residence time of H of 1 min, the2 advective transport distance for H is -1 mm. As described in Section 6.2, there may2 be short transient phases, when shifting TEA, where the microbiology itself, and not the H concentration, is rate limiting. The low H concentrations otherwise observed2 2 confirms that the supply of intermediates is normally rate limiting. The representativity of measured H concentrations may depend on the system2 generating the H . In systems where particulate organic matter is fermented and the2 Ž .TEA being reduced is dissolved in the porewater e.g., sulphate , H -producing fer-2 menters are presumably associated with the organic matter and are surrounded by TEAP bacteria. In this case, there will be a short gradient in the H concentration from the2 Ž .fermenting to the juxtaposed closely associated or even directly connected TEAP Ž .bacteria Conrad et al., 1985; Hoehler et al., 1998 . The TEAP bacteria will lose whatever H they do not use to the bulk water by diffusion. This should mean that the2 bulk water ideally reflects the intracellular concentration of H of the TEAP bacteria2 Ž .Hoehler et al., 1998 . Generally, when the aquifer is sampled by taking water samples, the H concentration measured will reflect the bulk water. If the organic matter being2 fermented is dissolved, as in most pollution plumes, the fermenting and the TEAP bacteria may be less closely associated. In this case, there would be a gradient in the H2 concentration from the fermenting to the TEAP bacteria, and sampling of bulk water might give an intermediate concentration. In the case of a solid electron acceptor, such as Fe oxides, the bulk water would also have an intermediate concentration with the Ž .detailed distribution depending on the rate of Fe-oxide reduction Hoehler, 1998 . In systems with internal gradients and several TEAPs occurring in different parts of the sediment, a correct interpretation of a measured concentration in a pumped water sample is not straightforward. The H concentration, controlled by the TEAP mediating bacteria, is related to the2 Žminimum energy required in order that the microorganisms can store through ATP . Žsynthesis the energy released by the TEAP for later use in their life processes e.g., .Westermann, 1994; Hoehler et al., 1998 . The energy available, to a microorganism for Ž .a given TEAP, is a part of the Gibb’s energy of the reaction DG . The more negativer the DG the higher the energy available. Table 9 lists estimates of the required energyr for the different redox processes. If the H activity is isolated in equations for2 Žcalculating Gibb’s free energy of reaction for a TEAP involving oxidation of H see2 Ž . .Eq. 9 and the example associated , we can calculate the H activity as a function of2 ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 197 concentration. An alternative explanation, proposed here, is that when lowering the H2 concentration as much as possible, by using the TEA allowing the lowest H concentra-2 tion at the given in situ conditions, the fermentation is shifted towards a more complete oxidation of the organic carbon to CO . This would imply a larger supply of H to the2 2 H oxidizing bacteria, and perhaps an overall more efficient use of the total energy2 available in the organic carbon. The tendency of the bacteria to always minimize the energy of the ongoing TEAPs to a fixed value close to equilibrium suggests that the system can be viewed as being in a Žstate of partial equilibrium Postma and Jakobsen, 1996; Jakobsen et al., 1998, Jakobsen .and Postma, 1999 . The term partial equilibrium was originally used by Helgeson et al. Ž .1968 for describing irreversible weathering processes. In terms of redox processes, partial implies that the rate of hydrolyzation and fermentation of the organic carbon controls the overall kinetics of the system, while the TEAPs are described by equilib- rium chemistry by taking into account the energy needed for the bacteria. Based on this, the redox state of the system is evaluated by calculating the DG of the TEAPs using ther expressions shown in Table 9. The calculated DG will indicate the relative potential forr the given TEAP, just as calculating the E for a given process would. However, byH using the DG it is possible to compare the calculated value with the minimum energyr yield for the given TEAP, ideally enabling a judgement of whether a process may or may not occur. With this approach, the interpretation does not rely on the assumption of Ž .steady state implied in the Lovley and Goodwin 1988 approach. Especially in Žcontaminant plumes, an assumption of steady state may not be valid Vroblesky and .Chapelle, 1994 . Calculating the DG ’s, however, has other limitations. Evidently, ther Ž .relevant parameters must be available to calculate the relative potential DG for ther different TEAPs, but to precisely evaluate whether a given process may or may not occur requires a value for the minimum energy yield for that process. The values of Ž . Ž .Hoehler et al. 1998 are well documented Fig. 11 , but relate to a marine environment q Ž .where the high Na concentration might have an influence Hoehler, 1998 . It could Ž .explain why the minimum energy yield observed by Schulz and Conrad 1996 in limnic Ž . Ž .Lake Constance CH sediments are lower Table 9 . This would imply that H2 concentrations in pollution plumes might vary, depending on the ionic strength of the plume, related to the plume source. For Fe oxides, it is close to impossible to know exactly which Fe oxides are being reduced. However, the measured H concentration2 should, under circumstances where there are indications of Fe-oxide reduction, enable an estimate of which Fe oxide apparently controls the H concentration. Currently, the2 database is too small to address the various limitations and aspects related to the use of DG values in detail.r An advantage of the DG approach is that it can be used to directly address ther potential for degradation of xenobiotic compounds, if thermodynamic parameters and minimum energy yields are known. Fortunately, these values are starting to appear Ž .Yang and McCarty, 1998 . The approach may also be used the other way around. In cases where a downgradient increase in methane indicates methanogenesis, but the calculated DG is too high for methanogenesis to be feasible, methanogenesis must ber occurring in stagnant subdomains of the plume. Knowing the minimum energy yield for methanogenesis enables a calculation of the minimum H concentration in these2 ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241198 subdomains. With this, the degradation potential within the subdomains for a compound for which thermodynamic parameters and minimum energy yields are known may be evaluated. 5.3. Methods The nanomolar concentrations of H found in groundwater are quantified by gas2 Ž .chromatography with a reduced gas detector RGD2 or 3, TraceAnalytical , calibrated with standards of H diluted in N . The principle is that reduced gases react with a2 2 heated bed of HgO producing free Hg vapour, quantified by a UV spectrophotometer. Measurements of groundwater concentrations are usually made using Abubble-strip- Ž .pingB Chapelle and McMahon, 1991 where a bubble of N is equilibrated with the2 groundwater by letting groundwater flow continuously through the bubble. The method Ž .can be scaled down as described in Bjerg et al. 1997 , and current tests indicate that further downscaling is possible. When the bubble is equilibrated a gas sample is taken from the bubble with a syringe and injected into the gas chromatograph. The equilibra- tion time is 10–30 min depending on bubble size, water flux, agitation intensity, and temperature. Care should be taken to limit flow velocities in the sampled aquifer Žvolume, as high flow rates tend to lower the measured H concentration Jakobsen and2 .Postma, 1999 . The measurement must be made in the field since H is generally not2 stable in the mixed gas phase, which develops in the bubble, and H is extremely2 volatile and tends to escape. An alternative is to let the gas sample equilibrate with the groundwater by diffusion through a gas permeable wall. This can be done by lowering a teflon-tubing spiral into Ž .the well Chapelle et al., 1997 , requiring that the water in the well is continuously renewed either naturally or by pumping during the equilibration. In small-scale tracer tests, prolonged continuous pumping can be a problem, but the spiral can be inserted in Ž .direct contact with the sediment Frank, 1996 . Removal of water is then not needed, but the teflon tubing must be connected directly to the surface by, e.g., copper tubing, limiting the depth to which the method can be applied. The known air volume of a spiral system is sampled by mixing it with a known volume of N using a syringe.2 Equilibration times are on the order of 12–24 h, depending on sample volume, tubing surface area, wall thickness, and temperature. Other alternatives are described in Ž .Chapelle et al. 1997 . Ž .Generally, reactive metal surfaces Bjerg et al., 1997 and electrical DC pumps Žshould be avoided as they may reduce the protons in water to hydrogen Chapelle et al., .1997 . Wells should not be used for sampling until the elevated H concentrations2 induced by the well installation is gone, and working close to the well with heavy Ž .equipment prior to sampling should be avoided Bjerg et al., 1997 . Ž .As discussed by Hoehler 1998 , measurements of concentrations in sediment sam- ples cannot be made by transferring the sediment to an incubation flask, waiting for 1–2 h for equilibrium between the pore water and the gas, and then sampling the overlying headspace. It is necessary to wait until the entire system is reequilibrated. Due to the very dynamic nature of the H pool, the physical pore waterrgas equilibration with the2 ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 199 headspace will remove H from the sediment, but H will rapidly be replenished by2 2 fermentation in an attempt to reach equilibrium. More H will go into the headspace and2 so on, until the headspace is finally in equilibrium with the sediment H concentration.2 It is impossible to calculate the sediment concentration, until this new equilibrium is obtained because the total amount of H in the system is increasing over time at an2 unknown rate. However, once equilibrium is reached, incubation flasks may give very good results, if great care is taken with regards to keeping the temperature constant and if flasks are not resampled until the system has reequilibrated. Equilibration times of headspace flasks are on the order of days and weeks depending on sedimentrheadspace ratios and volumes and the H production rates.2 5.4. Applications There is a limited number of studies of H distributions in plumes. Three studies2 Ž .describe specific pollution plumes: Hanahan SC, USA dominated by JP-4 jet fuel Ž . Ž .Vroblesky and Chapelle, 1994 , Kingsbay GA, USA dominated by chlorinated Ž . Ž .solvents Chapelle and Bradley, 1998 , and Wurtsmith MI, USA dominated by Ž .petroleum products and solvents Chapelle et al., 1996 . Two studies describe mixed Ž . Ž .landfill plumes: the Grindsted Landfill DK leachate plume Jakobsen et al., 1998 , and Ž . Ž .the Norman Landfill OK, USA Harris et al., 1999 . Finally, there is a study from a Ž .plume of MAHs and PAHs entering a complicated fill-aquifer in Charleston SC, USA Ž .Landmeyer et al., 1998 . Fig. 13 shows the H levels and the concentrations of redox-sensitive parameters2 Ž . Ž .from a volume at the Hanahan site SC, USA . Chapelle et al. 1995 used a combination of geochemical data, in terms of changes in solute concentrations along the flowpath, and the H concentrations in each well to describe zones predominated by2 methanogenesis or sulphate reduction. The geochemical data, however, suggest that several redox processes may well be occurring concomitantly, at comparable rates, along several paths in this system. A pronounced increase in methane is seen along the Ž .proposed sulphate reduction flow path Well 41B™Well 33B . This increase is 35 times larger on a molar basis, than the decrease in sulphate. The flow path from Well 31A to Well 34A also has a larger increase in methane compared to the sulphate decrease and furthermore shows an increase in the Fe2q concentration. Along the Ž .proposed predominantly methanogenic flow path Well 31B™Well 34B , a pronounced increase in Fe2q, not noted in the paper, is seen. The increase might be due to the reduction of a rather stable Fe oxide, or it could be related to a decrease in pH Ž .Chapelle, personal communication due to acid-producing reactions, leading to dissolu- tion of FeS phases. The latter is supported by a similar increase in the total sulphide concentration. It is noted that even if Fe-oxide reduction is occurring, there is no doubt that the flow path from Well 31B to Well 34B represents the most reduced conditions within the shown volume of the aquifer. Ž . Ž .The Kingsbay GA, USA Chapelle and Bradley, 1998 study is similar in the way that at a closer look several redox processes appear to take place within the same aquifer volume. Though this shows that the H levels alone should not be used for identifying2 ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241202 could not be confirmed by microcosm measurements. So though H concentrations were2 not clear in diagnosing the redox processes, the highest H values were clearly found in2 what was expected to be the most reduced part of the plume. 5.5. EÕaluation H measurements are a potentially powerful tool for studying microbial processes, in2 general, and more specifically TEA degradation processes involving H . This is a2 consequence of the close link, through the thermodynamics of these processes, between H concentrations and the microbiological processes actually occurring in the plume.2 However, a number of issues need to be addressed before this approach to quantification of redox levels can be fully exploited. One of the difficulties is to know precisely what a given H sample represents.2 Depending on aquifer conditions, sampling method, as well as electron donor and acceptor distributions, bulk and stagnant water may be represented differently and the H concentrations in the two might well be different. This has implications for the2 prediction of degradation potentials. Also needed, is a larger database on what minimum energies are necessary for the different processes to occur under different conditions. There are indications that the minimum energy required is related to environmental Ž . q Žfactors, such as temperature Schulz and Conrad, 1996 and Na concentration Hoehler .et al., 1998 . Further complications are related to TEA threshold concentrations. This is important for sulphate reduction, which appears to more or less stop when the sulphate Ž .concentration is lower than 5–20 mM Ingvorsen et al., 1984; Lovley and Klug, 1986 . Ž .Jakobsen and Postma 1999 found sulphate reduction to cease even though DG wasr still low enough for the reaction to be feasible. How TEA threshold concentrations affect degradation of TEAs, in general, is not known. Finally, insight is still lacking on precisely how TEAPs, involving the reduction of solid phases, such as Fe oxides and Mn oxides, are related to in situ H concentrations.2 Ž .Specific steady-state H levels as suggested by Lovley and Goodwin 1988 are not2 applicable to pollution plumes. The main shortcomings are that fixed H levels do not2 account for the effects of actual concentrations of dissolved redox species, variability in iron oxides being reduced, actual temperature of the aquifer and the fact that redox processes may be overlapping and not exclusive. Also, the necessity of assuming steady state within the studied system is a limitation. The concept of specific H levels could2 be viewed as a special case of a more general Apartial-equilibriumB approach where the in situ activities of reactants and products controlling the redox processes are taken into account. The H concentration will, regardless of the chosen approach, indicate the redox2 level in a relative sense. High H concentrations are found in systems that are more2 reduced than systems with low H concentrations. In terms of degradation of specific2 compounds, designating the presence of highly reducing conditions may be more important than knowing the specific inorganic TEAP. Repeated monitoring of H2 concentrations, may be a sensitive tool in evaluating the stability of plume conditions, which is of importance in relation to in situ bioremediation. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 203 ( )6. Volatile fatty acids VFAs 6.1. Background VFAs are produced as fermentation products through the degradation of organic Ž .matter Fig. 2 and specific levels of volatile fatty acids, such as acetate, have been reported for different redox processes in different types of sediments, e.g., by Lovley Ž . Ž .and Phillips 1987a and McMahon and Chapelle 1991 . Of the VFAs, acetate and formate are observed in the highest concentrations, whereas propionate and butyrate are measured in lower concentrations. So far, only little data have been published regarding the relation between the VFAs and the redox conditions in aquifers. 6.2. Principles VFAs are intermediates produced under anaerobic conditions, as H is, and it has2 been observed that the concentration of VFAs often increase as electron acceptors become depleted. This has raised the hypothesis, that the concentration level of VFA can Ž .be used as an indicator of redox processes Vroblesky et al., 1997 . However, the dependency of the available energy for the oxidation of VFA is much lower than it is for H , because more electrons are transferred per mole of VFA and, therefore, the use of2 VFAs as indicators of redox processes is inherently more difficult. Recent results from Ž .Hoehler et al. 1999 indicate that transient high peak concentrations of VFAs may be related to a shift in the TEAP, rather than to a substantially different steady state concentration for the new TEAP. Data both from an in situ marine sediment and an Ž .incubation experiment Fig. 15 show how high concentrations of VFA occurred in connection to a shift in the predominant redox process from sulphate reduction to methanogenesis as sulphate was exhausted. In the lag time from the exhaustion of sulphate until full methane production was reached, the buildup of acetate resulted in a pool that lasted for some time — especially in the natural setting. A similar phe- nomenon was seen for H in the incubation. An intermittently high, but relatively2 long-lasting concentration peak in the acetate concentration, may have been misinter- preted as plateaus in short term experiments and in noncontinuous studies of various systems. In other words, the higher concentrations are observed when the TEAP changed may have erroneously been attributed to the TEAP taking over, and not to a transient phenomena related to the TEAP shift itself. 6.3. Methods Groundwater samples collected for analysis of VFAs should be protected from loss of the volatile acids through evaporation during, and after sampling. Since the VFAs are readily degradable under aerobic, as well as anaerobic conditions, the samples have to be preserved immediately. Addition of 0.2% chloroform and immediate freezing in polypropylene vials has been found to be a very simple and yet efficient preservation ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241204 Fig. 15. The transition from sulfate reduction to methanogenesis and the peak in the acetate concentration Ž . Ž .related to this transition. a An incubation experiment, b At 13.5 cm depth in Cape Lookout Bight sediments Ž .from Hoehler et al., 1999, with permission . Ž .method Hansen, 1998; Albrechtsen et al., 1999 . This method of preservation also prevents generation of additional VFAs, which may occur in samples containing dissolved organic matter, e.g., humic and fulvic acids. The VFAs are generally analyzed Ž .for by Ion Exclusion Chromatography Bradley et al., 1993 . Using a Dionex AS-10 Ž .column and suppressed EC detection, Hansen 1998 obtained a detection limit of 0.2 mM formate or acetate. 6.4. Applications In spite of the inherent difficulties, different levels of VFAs have been reported for different redox conditions. Formate has been observed at 5–60 mM in sulphate-reducing Ž .environments McMahon and Chapelle, 1991; Chapelle and Lovley, 1992 and at 0–6 Ž . Ž .mM in Fe III -reducing environments Chapelle and Lovley, 1992 . Acetate has been Žobserved at 2–50 mM in sulphate-reducing environments Lovley and Phillips, 1987a; .McMahon and Chapelle, 1991; Chapelle and Lovley, 1992 and at 0.5–3 mM in Ž . Ž .Fe III -reducing environments Lovley and Phillips, 1987a; Chapelle and Lovley, 1992 . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 207 known about the reactivity, and usually the reactivity of the sediment-bound organic Ž .carbon is considered to be low see Section 2.4 . Redox potentials and activity relate to on-going processes but do not reflect the history of redox processes in terms of pools of species reacted over time, nor do they reflect possible future changes due to limitations of pools. To learn about plume development with respect to redox and to assess the pools of redox species that have reacted or may react, the sediment pools must be addressed. This may be done by quantifying solid pools of primarily organic carbon, iron, sulphur, and manganese Ž . Ž .species or as capacities in terms of OXC and RDC as defined in Eqs. 10 and 11 . 7.2. Principles Sediment redox characterisation may involve: Ø Identification of the nature of individual sediment minerals; ŽØ Species capacity quantification of the bulk content of individual species, such as a .total iron content ; ŽØ Reactive fraction measurements an operationally defined fraction of a species that is available for a given reaction or the kinetics of a given reaction at a given capacity .level ; Ž .Ø Redox capacity measurements quantification of the available RDC and OXC . Identification of individual sediment minerals is done by traditional mineral analysis. These methods identify minerals at the grain scale level and are mostly used for direct confirmation of minerals anticipated from other observations, e.g., from capacity measurements. The methods are primarily used qualitatively, since quantitative measure- ments are very demanding and fairly uncertain. Species capacity measurements are used to determine bulk contents of species, such Ž .as iron, manganese, sulphur, and organic carbon Ball et al., 1991 . These elements, however, are found in a variety of different solid compounds; specific extraction and quantification of individual compounds is cumbersome, uncertain and for some com- pounds not feasible at all. Table 11 summarizes forms of iron, manganese, sulphur and carbon found in aquifers and possible extractants. Consequently, most bulk methods neglect the individual species and their redox state and determine only the element content. This can be done by a variety of methods, such as strong acid digestion and analysis of the digest, combustion of the sample followed by quantification of the Ž .evolved gases, and X-ray diffraction XRD . Redox reactions in groundwater often are surface-related reactions, and since only a minor fraction of the bulk content of a Žmineral participates in these reactions 1–10%; Heron and Christensen, 1995; Kennedy .et al., 1998 , capacity measurements have limited value for identifying temporal changes and spatial gradients in sediment composition caused by redox reactions. Several wet chemical methods for obtaining a more useful indication of the Areactive Ž . ŽfractionB or the expressed capacity have been developed Lovley and Phillips, 1986; .Heron et al., 1994b; Kennedy et al., 1998 . These relatively mild extractions preferen- ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241208 Table 11 Ž .Major solid species of Fe, Mn, S and C found in aquifers and contaminant plumes based on Heron, 1994 Species Formula Oxidation state Extraction Iron, Fe 2q aŽ .Ion-exchangeable Fe II Fe q2 Anaerobic neutral salt solution Siderite FeCO q2 No good method3 bTrolite, greigite, FeS, Fe S q2 AVS extraction3 4 Mackinawite cPyrite FeS q2 Sequential extraction2 Ž . Ž .Silicates, clays Fe II , Fe III q2,q3 No empirical methods dŽ .Ferrihydrite Fe OH q3 Reductive dissolution3 dGoethite FeOOH q3 Reductive dissolution dHematite Fe O q3 Reductive dissolution2 3 Magnetite Fe O q8r3 No good method3 4 Ž .Green rust Fe OH q8r3 No good method3 8 Manganese, Mn 2q aŽ .Ion-exchangeable Mn II Mn q2 Anaerobic neutral salt solution dPyrosulite MnO q4 Reductive dissolution4 dMixed Mn oxides MnO H q2,q3,q4 Reductive dissolutionx y Rhodochrosite MnCO q2 No good method3 Sulphur, S 2y bAdsorbed sulfate SO q6 NaH PO extraction4 2 4 bSolid sulfate CaSO , FeSO q6 NaH PO extraction4 4 2 4 Ž .Elemental sulfur S 0 0 Chemical extraction bAcid volatile sulfur FeS, Fe S y2 Hot HCl extraction3 4 cPyrite, marcasite FeS y1 Sequential extraction2 Carbon, C eŽ .Organic carbon, TOC Humic and fulvic ? assumed 0 Standardized TOC analysis acids and humin eInorganic carbon, TIC CaCO , FeCO q4 Acid extraction3 3 a Ž .For example, Hodgson 1960 . b Ž .Landers et al. 1987 . c Ž .Canfield et al. 1986 . d Ž .For example, Heron et al. 1994b . e Ž .Heron et al. 1997 . tially address surface-related or very reactive minerals and leave the bulk and more stable minerals behind. A standard reaction time is chosen for a given method, but due to differences in sediment composition and the possibility of this interfering, a given method does not necessarily extract the same Areactive fractionB from different sedi- ments. The analysis of extracts and evolved gases is used for determining the composi- Ž .tion of the reacted phases. In the kinetic approach described by Postma 1993 , the reaction is followed over time, by monitoring the extract concentration, allowing assessment of initial reactivities and a description of how the reactivity will change as a function of how much of the phase has reacted. Results for Fe oxides using this method give continuos curves for the release of Fe, showing that there are no well-defined pools of distinct reactivities, no Areactive fractionsB. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 209 Redox capacity measurements have been suggested as an overall quantification of Žsediment redox properties Barcelona and Holm, 1991; Pedersen et al., 1991, Heron et .al., 1994a . OXC and RDC are lumped variables containing contributions from multiple electron acceptors or donors and determined by the chemistry of the proposed reagent. These capacities fall in a category somewhat in between the bulk capacity measurements and the reactive fraction measurements, since the nature and strength of the chemicals, Žand the duration of the extraction affects the measured capacities Barcelona and Holm, .1991; Heron et al., 1994a,b . 7.3. Methods 7.3.1. Sampling Reliable sediment analysis heavily depends on obtaining a representative solid sample and keeping the reactive species intact during sampling, handling and storage. The current experiences include: Ø The outer layers of soil cores should be pared of and discarded in the laboratory in order to avoid interference from the collection sleeve and from contamination. Ø Drying of sediment can lead to oxidation of reduced species and loss of volatile components. Ø Wet sample splitting is difficult and the reproducibility is usually poor. Ø Storage in inert atmosphere is essential. Storage in an oxygen-containing atmosphere leads to oxidation of, e.g., ferrous sulphides, while storage in atmospheres containing hydrogen may result in reduction of electron-acceptors. Ø No matter how samples are stored, microbial activity can change the composition; so short holding times and cooling are preferred before solids analyses. In conclusion, rapid sample handling in the field, flushing with an inert gas, sealing and cold storage in an inert atmosphere, such as nitrogen or argon for only a limited Ž .time is recommended see also Section 8.3.1 . 7.3.2. Identification of indiÕidual sediment minerals Direct identification of individual solid phases in reduced contaminant plumes has Ž . Žbeen done primarily using Scanning Electron Microscopy SEM Baedecker et al., .1992; Heron et al., 1994b, 1998; Tuccillo et al., 1999 and energy-dispersive X-ray Ž . Ž .analysis EDS Cozzarelli et al., 1999a . For SEM, the sediment must be dried Ž . Žanaerobically e.g., by freeze-drying and sieved to remove larger grains e.g., )0.2 . Ž .mm to allow for coating e.g., carbon andror gold prior to the analysis. Samples should preferably be run the same day as prepared. SEM and EDS analysis allow for qualitative description of the sample with respect to abundance of minerals, the mineral Ž .shape and size, and bulk atomic composition. Heron et al. 1994b identified iron oxides, Ž .Fe II -rich carbonates, and framboidal pyrite crystals in a landfill leachate-affected Ž . Ž .aquifer. Baedecker et al. 1992 and Tuccillo et al. 1999 identified iron-oxide coatings on pristine sediments and ferroan calcite crystals in solids from the reducing plume at ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241212 Ž . Ž . Fi g. 16 .I ro n m in er al og y an d re do x bu ff er ca pa ci ty ch an ge s al on g a st re am lin e in th e V ej en La nd fil l D K le ac ha te pl um e. A V er tic al tra ns ec ts ho w in g th e m aj or Ž . re do x en vi ro nm en ts in fe rr ed fr om gr ou nd w at er sa m pl es an d th e ce nt ra lf lo w -li ne . B Th e sc he m at ic di st rib ut io n of iro n sp ec ie s al on g th e ce nt ra lf lo w -li ne in fe rr ed Ž . Ž . fr om an al ys is of 45 se di m en ts am pl es . C R ed uc tio n ca pa ci ty he re TR C ,e qu al to R D C of th e se di m en td et er m in ed by a m od ifi ed ch em ic al ox yg en de m an d m et ho d, Ž . Ž . Ž . Ž .Ž an d D O xi da tio n ca pa ci ty O X C m ai nl y re la te d to iro n ox id es de te rm in ed by th e Ti II I -E D TA m et ho d H er on et al ., 19 94 a m od ifi ed fr om H er on an d C hr is te ns en , . 19 95 ,w ith pe rm is si on . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 213 summary of the findings is presented in Fig. 16. It was shown by OXC determination Ž .and HCl extractions that Fe III oxides, such as goethite, were depleted in the methanogenic and sulphate-reducing zones close to the landfill. The distribution of iron and sulphur species along a central flow line in the plume strongly indicated that Ž .reduced precipitates included pyrite identified positively by SEM; Fig. 17 , amorphous Ž .sulphides seen as elevated AVS contents in the reduced zones , and iron carbonates Ž Ž .increased Fe II contents extracted by HCl as well as Fe-rich carbonates observed by .SEM . The findings of reduced precipitates were supported by elevated RDC in these sediments. Further downgradient, in the transition zone dominated by nitrate reduction Ž .and aerobic respiration, elevated Fe III contents as well as higher OXC indicated that part of the reduced iron had travelled downstream and then precipitated as ferrihydrite Ž . Ž .when oxidized Fig. 16 . This was supported by high Fe III contents extracted by 0.5 N Ž .HCl, indicating a relatively amorphous iron hydroxide Heron et al., 1994b . Analysis of Ž . Ž .sediment-bound Fe II formed during microbial degradation of organics under Fe III -re- ducing conditions provided evidence of the iron reduction not revealed by groundwater Ž .sampling alone Nielsen et al., 1995 . Ž . Ž .Baedecker et al. 1992, 1993 , and Tuccillo et al. 1999 studied the iron chemistry in Ž .sediments in the proximity of the Bemidji oil spill MN, USA with a recently developed redox gradient caused by hydrocarbons leaching from an oil body. Close to the source, Ž .Fig. 17. Scanning Electron Microscopy photos of redox-sensitive iron and sulfur minerals. A Ferroan calcite Ž . Ž . Ž .identified in reduced sediments at the Bemidji MN, USA oil spill site Tuccillo et al., 1999 . B Framboidal Ž . Ž . Ž .pyrite crystals found in the methanogenic zone downgradient of Vejen Landfill DK Heron et al., 1994b . C Ž .Iron oxiderhydroxides appearing as bright coatings on a quartz grain from the Fe III -reducingrprecipitating Ž . Ž . Ž .zone at Vejen Landfill DK Heron et al., 1994b . D Weathered iron-dominated grain found in the Ž .methanogenicrsulfate reducing zone at Grindsted Landfill DK . This crystalline iron oxide grain constitutes a Ž .large capacity of iron, but its reactivity is low from Heron et al., 1998, with permission . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241214 anoxic conditions prevailed, and a 24–32% reduction in the 0.5 N HCl-extractable Ž . Ž . Ž .Fe III was observed. Changing ratios of Fe II to Fe III along the longitudinal plume axis and a significant increase in the total iron content in the anoxicroxic transition zone Ž .indicated downstream iron transport and reoxidation Fig. 18 . The identification of Ž . Ž .ferroan calcite in the reduced samples Fig. 17 and the abundance of Fe III oxides on feldspar grains from the transition zone presented a direct evidence of the contrasts in Ž .mineralogy along the plume transect Tuccillo et al., 1999 . The finding that iron reduction was a very important redox process in the plume was somewhat surprising, Ž .since dissolved iron concentrations were low in the entire plume Bennet et al., 1993 . ŽPrecipitation of ferrous iron carbonate siderite and ferroan calcite, Fig. 17, Tuccillo et . Žal., 1999 was identified as the reason for the low dissolved iron Baedecker et al., Ž . Ž . Ž .Fig. 18. Distribution of Fe III and Fe II downgradient from the oil spill at the Bemidji MN, USA site. The distance of 160 m corresponds to the toe of the reduced plume, where the aquifer is still under aerobic Ž .conditions. Fe III is lowered close to the source, and elevated downgradient in the anoxicroxic transition Ž . Ž .zone. Fe II is elevated in the reducing environments from Tuccillo et al., 1999, with permission . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 217 and sulphur species intact, and extracting amorphous minerals, such as ferrihydrite and ferrous sulphide. Hydrochloric acid and ascorbic acid extractions are the most promising approaches for reactive fraction measurements. Direct identification of mineral forms and crystal sizes by SEM has added confidence to geochemical interpretations and helped explain different levels of reactivity for comparable bulk iron and sulphur contents. Key parameters are crystallinity, size, and surface area of the minerals. Ž .Sediment analysis revealing depletion of OXC and Fe III as well as increased RDC Ž .and reduced Fe II and sulphide contents has been shown to describe the importance of past processes in older plumes and may also contribute to identification of currently limiting factors for sediment associated redox processes. Quantification of redox capaci- ties is, in particular, important for remediation approaches considering changes in redox levels, since the dominating capacities are associated with sediments. 8. Microbial measurements 8.1. Background Most redox processes in contaminant plumes are microbially mediated; and 30 years Ž .back, Farkasdi et al. 1969 distinguished between reduction, transition and oxidation zones in leachate plumes at German landfills. This investigation related the observed Ž yzones to microbial activity by enumerating sulphate- and nitrate-reducing NO to3 q. Ž y .NH as well as denitrifying NO to N bacteria in the plumes. On this background,4 3 2 it seems obvious that characterization of redox environments in contaminant plumes should include the microorganisms performing the redox processes. Different approaches are used to identify, enumerate and quantify the different microorganisms carrying out each of the redox processes. This may involve traditional techniques for culturing of the organisms or modern chemical techniques based on direct Ž .detection of biomarkers, e.g., ester linked PhosphoLipid Fatty Acids PLFA for specific processes or organisms, and the use of molecular DNA or RNA probes. Instead of focusing on the number of the individual microorganisms or groups of Žmicroorganisms, specific types of processes e.g., sulphate reduction, denitrification, .etc. can be investigated by TEAP bioassays, where each of the processes can be detected and quantified. Furthermore, the TEAP bioassays harbor a unique opportunity to detect ongoing processes. Albeit the TEAP bioassay approach implies incubation of field samples in the laboratory, this approach is currently the only approach used in plume studies providing a direct estimate of field-related redox processes and their rates. 8.2. Principles The presence of microorganisms able to conduct specific redox processes are traditionally quantified by growth on synthetic media, e.g., agar plate-spreading tech- niques where each single bacteria is allowed to form colonies, which then can be ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241218 counted. This technique is mainly suitable for aerobic conditions, but poor for investiga- Ž .tions of contaminated anaerobic aquifers. Instead the most probable number MPN enumeration technique is used, since this technique allows to a higher degree for controlled, anaerobic conditions. The MPN enumeration technique is based on the principle of preparing serial dilutions of a sample, and from each of these dilutions, a set of replicate tubes with defined media are inoculated. The activity in each replicate tube is registered, e.g., in terms of production of reduced end product from the electron acceptor of concern. At a certain dilution, the density of microorganisms becomes so low that activity is detected only in some of the replicates and by means of a probability table, the pattern of positive and negative scores is converted to the MPN of the investigated type of organisms in the original sample. An alternative to the classical culturing methods is a chemical characterization of microbial communities. A significant advantage of the chemical approaches is that the samples can be preserved immediately after sampling, thus avoiding artificial changes in the microbial population during transport and handling. Furthermore, the samples can be stored for long periods, which reduces the peak in the workload immediately after the sampling. One promising approach is the measurement of PLFAs, which is ubiquitous in cell membranes. The PLFAs can be extracted from an aquifer sample and the content and composition of the PLFAs can be measured by GC-MS, and from the composition of the PLFAs the presence and relative occurrence of different types of microorganisms Ž . Ž .White et al., 1983 can be detected, e.g., in an aquifer sample Ludvigsen et al., 1999 . Some PLFAs occur in the lipid membranes of sulphate-reducing bacteria, including 15:1 Ž . ŽDesulfobulbus sp., D.C. White, personal communication , i17:0 and i17:1v7c De- . ŽsulfoÕibrio sp., Edlund et al., 1985; Kohring et al., 1994 , and 10me16:0 Desulfobacter .sp., Dowling et al., 1986 , and the presence of some of these biomarkers in a sample may, thus, indicate the presence of sulphate reducers. The monoenoic 16:1v7c is the most abundant PLFA in cell membrane of Geobacter metallireducens and the second Ž .most abundant PLFA in Shewanella sp. Coleman et al., 1993 , which both are iron reducers. On the other hand, 16:1v7c is also common in the cell membranes of all Ž .Gram-negative bacteria as well as many microeucaryotes Ludvigsen et al., 1999 . Thus, the suggested biomarkers for different specific redox groups may not be that specific, implying a risk of false positive results. The interpretation of PLFA analyses is further complicated by the fact that more than 50 different PLFAs have been detected in aquifer Ž .samples e.g., Ludvigsen et al., 1997 Other emerging alternatives are the RNA and DNA probes. The principle in these Ž .techniques is to establish a fragment or a piece of DNA or RNA the probe correspond- ing to the gene analyzed for, and to label this probe by, e.g., a stain. When added to the Ž .sample the probe will bind hybridize to the gene, allowing for detection of the gene. Ž .Smith and Tiedje 1992 identified specific denitrifying strains in aquifer samples by DNA probes. Sulfate-reducing bacteria have been identified by 16S rRNA probes in Žother environments, e.g., marine sediments Devereux et al., 1996; Ramsing et al., 1996; .Teske et al., 1996 . RNA probes may even be used to estimate the activity of a certain process, such as sulphate reduction. An important strength of the technique is that it is not necessary to grow or culture the investigated organisms, a procedure step that always ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 219 causes some selectivity. An important limitation is that probes can only be prepared for known genes and, thus, for organisms that have been isolated. DNArRNA techniques Ž .may also be combined with the MPN approach MPN-PCR , which gives information about the dominant species cultured in the MPN tubes as done for iron reducers by Ž .Anderson et al. 1998 . These techniques may prove of interest, but they will not be discussed further because of their limited use in pollution plumes to date. The above-mentioned methods focus mainly on the presence and enumeration of the microorganisms, but the actual processes are more important in the characterization of the redox environment. The TEAP bioassay is a useful approach to investigate redox processes by enclosing a sample and over time following the consumption of the electron acceptor of concern or the production of intermediates or end products. The assays are set up without any amendments, but sampling and handling of the samples may, in the short term, increase the availability of some sediment associated electron- Ž . Ž . Ž .acceptors, such as Fe III and Mn IV or of electron donors organic matter or H . This2 may influence the rates depending on what is the limiting factor in the bioassay. One of the major limitations of this technique is that the change over time in concentration of the measured variable is often very small compared to the background concentration of the variable. This is especially problematic for determining low rates of iron reduction Ž .Ludvigsen et al., 1998 . To overcome this, long incubation times are often required before the changes in the concentrations are detectable. Besides indicating if a certain redox process occurs, TEAP bioassays also allow for estimation of redox rates. 8.3. Methods It is a prerequisite for meaningful microbial investigations that the investigated sediment and groundwater have been collected intact with respect to redox conditions. Microbiological contamination during sampling and handling must also be avoided. Thus, sterile techniques have to be used although the investigation concerns mixed populations. 8.3.1. Sampling and handling The sampling requirements are similar to the requirements for groundwater and sediment sampling discussed in the previous section; but on top of these requirements, all sampling equipment and containers should ideally be handled by sterile techniques. The samples should be processed as soon as possible, since just a few days of storage even at low temperature may alter the composition and activity of the microbial Ž .population Brockman et al., 1998 . The outer 2–5 mm of sediment cores should be pared off in the laboratory by a paring device to remove potentially contaminated sediment. The use of this procedure reduces the demand for strictly sterile work in the field. Unfortunately, similar ap- proaches are not available for water samples, unless the samples are sterilized by, e.g., filtration, which removes the bacteria to be investigated. The sediment samples are often handled in an anaerobic box to prevent contamination by oxygen in the laboratory. During the setup in the laboratory, the choice of materials should minimize the release of potential substrates to the samples and, therefore, glass and stainless steel are ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241222 An indirect approach for determining the TEAP is to add 14C acetate and follow the production of 14CO and 14CH . When this is combined with addition of molybdate,2 4 which selectively inhibits sulphate reduction, it is possible to distinguish between Ž 14 . Ž 14methane production by increase of CH , sulphate reduction by increase of CO ,4 2 . Ž 14when molybdate was not added and iron reduction by increase of CO , when2 14 . Ž .molybdate was added and by absence of CH Anderson et al., 1998 .4 8.4. Applications 8.4.1. MPN The enumeration of redox-specific microorganisms has been used as a characteriza- tion and to support that the redox processes observed are biologically mediated. Essaid Ž .et al. 1995 counted aerobic heterotrophic, anaerobic fermentative, denitrifying, iron reducing, sulphate reducing and methanogenic bacteria in a crude oil spill site at Bemidji Ž .MN, USA . They found iron reducers in high numbers close to the oil body. Denitrifiers and sulphate reducers were present in lower numbers than all other types of organisms in accordance with the low availability of nitrate and sulphate in the groundwater. These counts were in accordance with the conceptual model of an aerobic, MnrFe-reducing, and methanogenic degradation sequence in the plume. However, the MPN numbers of acetate-oxidizing iron reducers were high, and more or less even, in both the iron-reducing part of the plume area and an anaerobic, uncontaminated Ž . Žreference site Anderson et al., 1998 . An investigation by molecular techniques 16S .rRNA probes of the iron reducers in the MNP-tubes revealed that only a sequence closely related to Geothrix fermentans was recovered from the uncontaminated site, whereas the sediments from the iron-reducing zone also contained a sequence character- istic of Geobacter species. Geothrix fermentans are not known to oxidize aromatic Ž .compounds with the reduction of Fe III , in contradiction to Geobacter. Their occur- rence seemed to be stimulated by the presence of aromatic contaminants, which may indicate benzene degradation by iron reduction. High MPN numbers of iron reducers Žand Geothrix fermentans were also observed at two other oil spill sites: at Hanahan SC, . Ž .USA and at Rocky Point NC, USA , but no significant anaerobic oxidation of benzene was observed and Geobacter species were not found. This suggests that comparative studies on the size of the Geobacteraceae community in petroleum-contaminated aquifers might aid in the location of zones in which benzene degradation coupled to iron Ž .reduction is taking place Anderson et al., 1998 . Different physiological groups of sulphate reducers and methanogens using different carbon substrates have been observed in the leachate-contaminated aquifer at Norman Ž . Ž .Landfill OK, USA by the MPN technique Beeman and Suflita, 1987, 1990 . The composition of the microbial population changed along the plume downstream of the landfill with the main occurrence of methanogens and sulphate reducers close to the landfill and decreasing numbers in the more distant part of the plume. Later investiga- tions confirmed the occurrence of relatively high numbers of methanogenic bacteria in the center of the leachate plume where the methane concentration was the highest and the sulphate concentration was relatively low. Higher numbers of sulphate reducers were observed outside the center of the plume in the shallow and the deep portion of the Ž .aquifer Harris et al., 1999 . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 223 The MPN technique has also been used in the leachate plume at the Grindsted Ž .Landfill DK to enumerate nitrate, manganese, iron, and sulphate reducers and Ž . Ž .methanogens Fig. 19 Ludvigsen et al., 1999 . The iron, manganese, and nitrate reducers occurred in surprisingly high numbers compared to the total cell numbers and varied only little with distance. The number of sulphate reducers was higher in areas where sulphate reduction occurred and the number of methanogens was higher where methane production was observed. However, the different bacteria groups were present in nearly all the samples investigated without any clear relation to the dominant redox processes. This may reflect that the MPN technique enumerates all the bacteria with a potential for conducting the investigated process, no matter whether this potential is expressed or not. The same type of organisms may also be enumerated in several groups, e.g., as iron, sulphate and manganese reducers. The density of a specific group of microorganisms has been tested as a measure of the capacity of the specific redox environment, but with little success. Kao and Borden Ž .1997 found no correlation between the number of denitrifiers enumerated by the MPN technique and the degradation rates for toluene, ethylbenzene or xylene under denitri- Žfying conditions in a range of fuel oil-contaminated sites including Rocky Point, Fort .Bragg, and Chapel Hill in NC; Traverse City and Sleeping Bear Dunes in MI . Finally, quantification of redox-specific bacterial groups in modelling of redox Ž .processes in the oil-contaminated aquifer at Bemidji MN, USA has been attempted Ž .with limited success Essaid et al., 1995 . The MPN technique is useful as a confirma- tion of the presence of a redox-specific microbial population and a potential for certain redox processes, despite the fact that the results obtained with the MPN technique may be limited by the growth conditions defined by the choice of medium, substrate, etc. However, the presence of certain bacteria does not necessary mean that these processes predominate. Ž . Ž .Fig. 19. The log number of methanogens Archaea , sulfate-, iron-, manganese- and nitrate-reducing bacteria measured by MPN enumerations of aquifer sediments at different locations down gradient of the Grindsted Ž . Ž .Landfill DK plume 95% confidence limits shown . Samples were pooled in sets of two or three per distance. When more than one set of samples were collected per distance, 2 bars are shown. ‘b.d.’ indicates below Ž .detection limits, ‘n.m.’ indicates not measured. modified from Ludvigsen et al., 1999, with permission . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241224 8.4.2. PLFA The presence of PLFA biomarkers has been used as indication for the presence of Ž .specific bacteria in various environments e.g., Coleman et al., 1993 , including landfill Ž .leachate plumes Albrechtsen et al., 1995; Ludvigsen et al., 1999 . Only a few results from analysis of PLFAs are available and PLFAs seem of limited value in identifying specific microbial populations. This was shown in a multivariate statistical analysis of the relation between the PLFA profile and different microbial redox processes in a Ž .landfill leachate plume Ludvigsen et al., 1997 . Some specific lipid biomarkers, however, suggested that the proportion of sulphate- and iron-reducing bacteria was Ž .increased at the border of the landfill Ludvigsen et al., 1999 . Also, in a jet fuel-con- taminated aquifer, elevated concentrations of terminal-branched and cyclopropan fatty Žacids were coincident with sulphate-reducing conditions Fang et al., 1997; Fang and .Barcelona, 1998 . 8.4.3. TEAP bioassays TEAP bioassays are useful for identification of ongoing microbial redox processes, for verifying redox sequences, for geographical location of the activities resulting in observed distribution of dissolved redox-sensitive parameters, and for estimating rates of the investigated redox processes. TEAP bioassays with unamended groundwater and sediment samples verified the Ž .presence of the following metabolic redox activities in the Grindsted Landfill DK Ž .plume Ludvigsen et al., 1998 : denitrification, iron, manganese, sulphate reduction, and Ž .methane production Fig. 20 . Similar results have been obtained from the Vejen Ž . Ž .Landfill plume DK Albrechtsen and Christensen, 1994; Christensen et al., 1994 . At Ž .the Norman Landfill site OK, USA , the TEAP bioassays demonstrated that sulphate reduction is a dominant electron-accepting process in a narrow zone just below the water table where the sulphate concentrations were the highest, but in this zone also some iron reduction was detected by TEAP bioassays. No iron reduction could be detected in a Ž .deeper sulphate-depleted zone, although the Fe II concentration was very high, making Ž .it unlikely that the dissolved Fe II was generated at this location. Maximum methane production was observed in this zone in accordance with high methane concentrations Ž .Cozzarelli et al., 1999b; Harris et al., 1999 . When testing for different processes, TEAP bioassays set up with material from the same sediment sample, the Grindsted Ž .Landfill DK plume showed simultaneous occurrence of several different microbially Ž . Ž .mediated redox processes Ludvigsen et al., 1998 Fig. 20 . Thus, the TEAP bioassay has the ability, as no other current redox characterizing approach, to identify not only the dominant redox process, but also less significant redox processes in the same sample. This more differentiated description may have implications for the potential of a given redox zone, as identified by the dominant redox process, to degrade organic chemicals. Since the TEAP bioassays allow for estimating rates for each of the Ž .electron-accepting processes, the organic matter assuming oxidation level zero mineral- ized to carbon dioxide can be calculated from the measured rates. The calculations, Ž .presented in Table 12 for the Grindsted landfill DK plume, showed that one electron- accepting process clearly dominated each sample accounting for more than 70% of the Ž .equivalent carbon conversion Ludvigsen et al., 1998 . ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 227 Fig. 21. Unamended and amended bioassays with sediment from the iron reducing zone of the Vejen Landfill Ž . ŽDK plume illustrating limiting availability of iron and dissolved organic matter from Albrechtsen et al., .1995, with permission . not by sulphate reducers. Also, the occurrence of different volatile fatty acids differed between the locations, a fact that might influence the competition between sulphate Ž .reduction and methane production Beeman and Suflita, 1987, 1990 . TEAP bioassays conducted with different 14C-labelled substrates were also used to show the dominance Ž .of acetoclastic methane production from acetate over autotrophic methane production Ž . Ž .from bicarbonate Harris et al., 1999 . TEAP bioassays can also be used to confirm the actual ongoing microbial redox Ž .processes during degradation studies. Nielsen et al. 1995 used TEAP bioassays to Ž .evaluate the redox processes methane production, iron reduction and denitrification ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241228 during degradation of phenolic, aromatic and chlorinated aliphatic compounds investi- gated by in situ microcosms. The TEAP bioassays showed that the processes changed Ž .during the experiment in the in situ microcosms. Baedecker et al. 1993 used TEAP microcosms in connection with degradations studies at a crude-oil aquifer at Bemidji Ž .MN., USA to verify that iron and manganese reduction were connected with oxidation Ž .of toluene and benzene. Another example is the increase of Fe II in TEAP bioassays, Ž .suggesting that the mineralization of vinyl chloride was coupled to Fe III reduction Ž . Ž .Bradley and Chapelle, 1996 . In a similar way, Hunt et al. 1997 related mineralization Ž . Ž .of BTEXs with an increase of Fe II , and Adrian and Suflita 1994 related mineraliza- tion of halogenated N-, S-, and O-heterocyclic compounds with an increase in methane. Finally, investigations by TEAP bioassays in a field injection experiment at Grindsted Ž . Ž .Landfill DK Rugge et al., 1999 revealed the occurrence of methane production in¨ some locations in the plume dominated by iron-reducing conditions according to the Ž .water chemistry Albrechtsen et al., 1999 . The small methane production may reveal the presence of strongly reduced areas explaining the observed reductive dechlorination Ž .of some chlorinated aliphatic compounds Rugge et al., 1999 .¨ 8.5. EÕaluation The enumeration of redox-specific bacteria by the MPN technique is resource-de- manding and has some flaws since the technique is biased toward culturable organisms. Since some organisms are able to use several electron-acceptors, they may contribute to the number of bacteria in several redox groups. Thus, the technique primarily confirms the presence of a certain microbial potential for specific microbial redox processes. There is no growth requirement for measuring PLFA and biomarkers, which is a strong advantage. This technique may confirm on-going redox processes, but presently the technique cannot stand alone regarding quantifying or identifying dominant redox- Ž .specific populations. Molecular probes DNA and RNA also do not require growing or culturing of organisms. On the other hand, a major limitation is that probes can only be prepared for known genes and, thus, for organisms that have been isolated. These techniques may prove of interest but, so far, their use in aquifers has been limited. Generally, although present or detectable under optimum growth conditions, the Ž .different redox-specific bacteria sulphate reducers, denitrifiers, etc. are not necessarily active in situ, and significant correlation between the abundance of any of the different redox-specific bacteria measured by MPN or estimated from specific PLFA biomarkers, and quantitative measurements of redox reaction rates has not been observed. The TEAP bioassays, although tedious and laborious, provide the most valuable information, such as identification of on-going processes, physical location of on-going processes in a redox sequence, identification of spatial variation of redox processes, which cannot be observed from dissolved redox-sensitive parameters, and verification of the occurrence of simultaneous redox processes. Transformation rates can be estimated from the TEAP bioassays and a dominant redox process can be appointed by transform- ing these rates to carbon conversion. The TEAP bioassays can be manipulated by amendments and, thus, provide insight into the complexity of processes in a contami- nated aquifer and identify factors limiting or controlling the redox processes. Further- ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 229 more, TEAP bioassays can verify redox conditions during experiments and relate ongoing redox processes to mineralization processes. When assessing TEAP bioassays, however, it is important to bare in mind that the results are limited by the electron acceptors, which were investigated, and consequently by the TEAP bioassay setup. Also, the detection limit is a problem, especially in relation to iron reduction. Iron reduction may be underestimated, since even small rates, which may not be measurable or be considered significant in the iron-reduction bioassay, may convert more carbon than methane production rates, which can be detected with a lower detection limit. A way to lower the detection limits is to increase the incubation time. However, long incubation time is a problem from a practical point of view, and since the TEAP bioassays are set up as batches, limitations or successions in microbial populations may develop over time in the batches. The approach would benefit from further development in terms of improving the sensitivity, especially with regard to iron-reduction, and by including fermentation — which could be important in the context of landfill leachate plumes. However, so far, TEAP bioassays may be considered as the most useful and ultimate approach for characterizing the on-going microbial redox processes in contaminated aquifers. 9. Discussion and conclusion Redox conditions in a contaminant plume must be addressed to understand the biogeochemistry of the plume, and much progress, as reviewed in the previous sections, have been made in this regard in the last decade. However, it should be emphasised that our current documentation and experience on measuring redox conditions in pollution plumes is still rather limited. Ø The number of actual plumes, where assessment of redox has been addressed in Žany elaborated way e.g., in more than 20 sampling points and by more than one .method , is currently less than ten. Apparently, no investigation has used all the approaches discussed in this review on a plume containing several redox environments. Ø Most of our current experiences are related to plumes in relatively shallow, sandy and, prior to contamination, originally aerobic aquifers. Ø In nearly all current cases, the plumes have been present for more than a decade suggesting that the central part of the plume and maybe the full plume has reached a Ž .pseudo steady state. To our knowledge, only the Bemidji oil spill site MN, USA Ž .represents a newly formed progressing plume, and only the Winterthur CH landfill leachate plume represents a significantly receding plume. Ø Most, if not all, current plumes assessed with respect to redox have been addressed within the framework of reduced, dissolved organic carbon being oxidised by inorganic Ž .electron acceptors including carbon dioxide present naturally in the aquifer. Oxidation of, e.g., methane, sulphides, ferrous iron, and ammonium has been dealt with only rudimentarily in the context of plumes. This is in spite of the fact that recent modelling Ž .of redox conditions in plumes Hunter et al., 1998 suggests that such secondary oxidation reactions may be very important in contaminant plumes. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241232 and in applying monitored natural attenuation as a remediation strategy. A pseudo-sta- tionary plume often will have changed significantly the redox capacities of the aquifer as measured on the sediment, which seldom will be the case in a recent or temporary plume. Ø Is the apparent attenuation of an organic chemical in a plume consistent with our current knowledge on the degradability of that chemical at the redox conditions in the plume? Naturally, this requires that we can determine which redox conditions are present, not necessarily dominating, in the zone of the plume, where the organic chemical apparently is degrading. This may also require that the size of the zone and the retention time of the organic chemical in the zone are estimated to provide apparent degradation rates for comparison with the current database. Ž .Ø Are the redox conditions in a degradation experiment laboratory or field regarding a critical organic chemical comparable to the redox conditions in the plume? If the redox conditions in an experimental system change over time, e.g., due to leaks or intrusion of air during sampling or due to depletion of a critical electron acceptor caused by the spiking of the system with the organic chemical, the degradation patterns and rates observed in the experiment may not apply to the plume. Monitoring of both redox environments may be the only way to assure compatibility. Ø How easily can the redox conditions in the plume be altered in order to enhance the degradation of a critical pollutant? In the context of an engineered remediation, it may be of interest, e.g., to reduce the redox conditions to enhance dehalogenation of chlorinated solvents or to provide aerobic conditions to enhance benzene degradation. Evaluation of the feasibility of such approaches must involve assessment of the oxidation and RDC of the plume and the availability of redox buffers at the aimed redox condition beneficial for the enhancement of degradation. Ø What is the capacity of the aquifer for buffering against accidental spills of fuels from planned or existing storage tanks located on the land overlaying the aquifer? Aquifers with a high OXC are likely better at limiting the pollution of the groundwater from an accidental spill and, therefore, the land above more suited for hosting the storage tanks. The above-mentioned issues that redox measurements can help address show large diversity and remind us that the approach to measuring redox conditions must be selected according to the purpose. However, consistently linking approaches and pur- poses based on the current limited experiences would be premature. But it is important in each case, in view of the purpose, to pay attention to the following issues: Ø How detailed should the redox characterization be? The resolution wanted in the redox characterization should of course be balanced with the number of sampling points. However, even if a high resolution is not needed, the sample volume should ideally be smaller than sediment heterogeneities. Otherwise, samples may contain contributions from aquifer volumes of different redox conditions, resulting in averaged, nonrepresenta- tive results. The redox mapping must be sufficiently detailed that hydraulic retention time in a given redox zone can be estimated, if degradation rates are to be linked to redox conditions. Ø Is it sufficient to determine dominating redox environments, e.g., in terms of redox zones defined by the dominating redox process, or is the diversity of redox processes in ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 233 a given zone also of importance? If the pollutants of concern also are the pollutants that Ž .drive the redox processes e.g., petroleum hydrocarbons as dominant electron donors , it may suffice to know only the dominating redox condition. However, if the pollutant of concern is a trace component and the redox zones are primarily created by the transformation of other secondary pollutants, e.g., nonspecific dissolved organic carbon, it may be important also to know that other redox processes take place. Ž .Ø Are current redox conditions or processes to be identified only do they happen or Ž .are estimates of the reaction rates needed how fast do they happen ? If the pollutants of concern also are the pollutants driving the redox processes, estimates of redox rates may supplement observed rates of pollutant degradation, but with respect to trace pollutants, major redox process rates will not provide much information. Ø Are redox capacities important? Measurements of redox capacities may, in some cases, help in identifying limiting factors controlling current redox processes, remaining potentials, or redox equivalents needed to change current redox conditions. The capacity measurements, in contrast to the intensity measurements, are needed in understanding past plume history and predicting future natural or engineered changes. 9.3. Promising and less promising approaches to measuring redox conditions In view of the current experience, the diversity of purposes for measuring redox conditions, and the variable resources available for the measurements, specific recom- mendations on how to measure redox conditions in groundwater pollution plumes are currently not warranted. However, some approaches seem more promising than others, as summarised below: Ø Electrochemical redox potentials are easy to measure with a combination electrode and a potentiometer, but any elaborate interpretation is not defendable. In strictly anaerobic redox environments, the electrode is likely to respond to the iron redox couple, but the readings may not reflect the iron couple at equilibrium and cannot be used for meaningful general electrochemical calculations. Such simple measurements of redox potential, performed in-line in a flow cell, are primarily useful in the field for quickly assessing whether the groundwater is sampled from strongly reducing conditions Ž .typical readings: E -y50 mV , i.e., in the central part of the plume, or from theH Žoutskirts of the plume at higher redox conditions showing higher, typically positive .values . Experience from unpolluted groundwater environments may suggest, however, Žthat a better performed redox potential measurement two working electrodes of .different material, sufficient stability, etc., as specified in Table 6 in certain redox environments dominated by iron reduction may provide a redox potential useful for electrochemical calculations regarding the iron system. Ø Groundwater composition with respect to key pollutants is determined in most pollution plume studies. By ensuring redox intact sampling and handling of the samples, and analysing for redox-sensitive inorganic compounds in the samples, dominant redox conditions can be assessed from the relative distribution of the compounds. The strength of this approach is the use of conventional technology, but the mobility of some of the Ž .redox-sensitive species in particular, methane and the site-specific empirical criteria used for assigning the redox labels add some uncertainty to the precise physical ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241234 positioning of the redox zones. However, it is very likely that the relative groundwater composition with respect to redox-sensitive species will reflect the redox processes dominating in the different zones. We suggest that redox-sensitive species always be measured in groundwater samples and used, together with other approaches, in assessing redox conditions. Ø Hydrogen concentrations in groundwater samples can be fairly easily determined, but special precautions in well design, sampling and analysis are needed. Measurement of hydrogen concentrations has great potential for identification of redox processes under strongly reducing conditions. It is uncertain whether the technique is diagnostic at manganese and denitrifying redox conditions. The strength of the approach is that it reflects on-going redox processes precisely at the point of sampling. The direct use of hydrogen concentrations as a diagnostic tool for assigning redox zones is not appropri- ate. Temperature, concentrations of dissolved species as well as type of solid iron oxides involved in iron reduction influence the energetics of the system and actually may allow both iron reduction and sulphate reduction to take place at the same hydrogen concentra- tion. However, the combination of measured hydrogen concentrations with groundwater composition in thermodynamic calculation of Gibb’s energy of reactions may provide valuable insight into which redox processes may occur and, in the future, which chlorinated or other organic compounds may act as electron acceptors. Full exploitation of the potential of hydrogen measurements for determination of redox conditions requires improvements in the current thermodynamic data bases as well as further experience from real plumes. So far, experiences from just six plumes have been reported in scientific journals. Ø VFA concentrations in groundwater samples might theoretically, in analogy to the use of hydrogen measurements, reflect ongoing redox processes, but the few attempts to apply this to real plumes have so far provided no useful insight into redox conditions in plumes. Measurement of VFAs, however, may potentially turn out to be useful for evaluating short-term plume stability. Ø Aquifer sediment characterization primarily addresses iron and sulphur com- pounds. Manganese contents may also be determined, but since analytical distinction Ž . Ž .between Mn IV and Mn II is not easily accomplished and manganese is usually a less significant redox components, the focus is on iron and sulphur. Aquifer sediment has a far larger redox buffer capacity than the groundwater, and any evaluation of capacities must include sediment analysis. Applications may include, but not be limited to, assessing plume stability, limiting factors or capacities for iron reduction, quantification of the significance of iron reduction, and documentation of a receding plume by identification of pyrite accumulation in previous, sulphate-reducing zones of the plume. The strength of sediment redox characterization is that only the sediment accumulates information about past processes and provides a basis for evaluating future capacities. These aspects are, in particular, important in the context of plume remediation. The limitation of the sediment characterization is the costly sampling, extraction and analysis. Since iron reduction seems to be a very important redox process in many Ž . Ž .plumes and both Fe III and Fe II accumulates in the sediment, characterization of sediments is suggested in all plume investigations, where plume stability is a key issue or manipulation of the plume is considered. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 237 measurements for delineating redox processes in a contaminated aquifer. Environ. Sci. Technol. 30, 3565–3569. Chapelle, F.H., Vroblesky, D.A., Woodward, J.C., Lovley, D.R., 1997. Practical considerations for measuring hydrogen concentrations in groundwater. Environ. Sci. Technol. 31, 2873–2877. Cherry, J.A., Shaikh, A.U., Tallmann, D.E., Nicholson, R.V., 1979. Arsenic species as an indicator of redox conditions in ground water. J. Hydrol. 43, 373–392. Christensen, T.H., Kjeldsen, P., Albrechtsen, H.-J., Heron, G., Nielsen, P.H., Bjerg, P.L., Holm, P.E., 1994. Attenuation of landfill leachate pollutants in aquifers. Crit. Rev. Environ. Sci. Technol. 24, 119–202. Ž .Coleman, M.L., Hedrick, D.B., Lovley, D.R., White, D.C., Pye, K., 1993. Reduction of Fe III in sediments by sulphate-reducing bacteria. Nature 361, 436–438. Conrad, R., Phelps, T.J., Zeikus, J.G., 1985. Gas metabolism evidence in support of the juxtaposition of hydrogen-producing and methanogenic bacteria in sewage sludge and lake sediments. Appl. Environ. Microb. 50, 595–601. Cozzarelli, I.M., Herman, J.S., Baedecker, M.J., Fischer, J.M., 1999a. Geochemical heterogeneity of a gasoline-contaminated aquifer. J. Contam. Hydrol. 40, 261–284. Cozzarelli, I.M., Suflita, J.M., Ulrich, G.A., Harris, S.H., Scholl, M.A., Schlottmann, J.L., Jaeschke, J.B., 1999. Biogeochemical processes in a contaminant plume downgradient from a landfill, Norman, Okla- homa. U.S. Geological Survey Water Resources Investigations Report a 99-4018C, Reston, VA. Devereux, R., Hines, M.E., Stahl, D.A., 1996. S cycling: characterization of natural communities of sulfate-reducing bacteria by 16S-rRNA sequence comparisons. Microb. Ecol. 32, 283–292. Ž .Dolfing, J., 1988. Acetogenesis. In: Zehnder, A.J.B. Ed. , Biology of Anaerobic Microorganisms. Wiley. Dowling, N.J.E., Widdel, F., White, D.C., 1986. Phospholipid ester-linked fatty acid biomarkers of acetate- oxidizing sulfate reducers and other sulfide forming bacteria. J. Gen. Microbiol. 132, 1815–1825. Edlund, A., Nichols, P.D., Roffey, R., White, D.C., 1985. Extractable and lipopolysaccharide fatty acid and hydroxy acid profiles from DesulfoÕibrio species. J. Lipid Res. 26, 982–998. Essaid, H.I., Bekins, B.E., Godsy, E.M., Warren, E., Baedecker, M.J., Cozzarelli, I.M., 1995. Simulation of aerobic and anaerobic biodegradation processes at a crude oil spill site. Water Resourc. Res. 31, 3309–3327. Fang, J., Barcelona, M.J., 1998. Biogeochemical evidence for microbial community change in a jet fuel hydrocarbon-contaminated aquifer. J. Org. Geochem. 29, 899–907. Fang, J., Barcelona, M.J., West, C., 1997. The use of aromatic acids and phospholipid-ester-linked fatty acids Ž .for delineation of processes affecting an aquifer contaminated with JP-4 fuel. In: Eganhouse, R.P. Ed. , Molecular Markers in Environmental Geochemistry. American Chemical Society, Washington, pp. 65–76. Farkasdi, G., Golwer, A., Knoll, K.-H., Matthess, G., Schneider, W., 1969. Mikrobiologische und hygienische Untersuchungen von Grundwasserverureinigung im Unterstrom von Abfallplatzen. Stadtehygiene 20,¨ ¨ 25–31. Frank, S., 1996. A new concept for measuring rates in aquifers. MSc thesis, Dept. of Geology and Geotechnical Engineering, Techn. Univ. of Denmark, Lyngby. Golwer, A., Matthess, G., Schneider, W., 1969. Selbstreinigungsvorgange im aeroben und anaeroben Grund-¨ wasserbereich. Vom Wasser 36, 65–92. Grenthe, I., Stumm, W., Laaksohaarju, M., Nilsson, A.-C., Wikberg, P., 1992. Redox potentials and redox reactions in deep groundwater systems. Chem. Geol. 98, 131–150. Griffioen, J., van der Grift, B., Buijs, A., Hartog, N., 1999. Oxygen consumption of natural reductants in Ž .aquifer sediment related to in situ bioremediation. In: Alleman, B.C., Leeson, A. Eds. , In Situ Bioremediation of Petroleum Hydrocarbon and Other Organic Compounds. Battelle Press, Columbus, OH, USA, pp. 463–468. Guckert, J.B., Ringelberg, D.B., White, D.C., Hanson, R.S., Bratina, B.J., 1991. Membrane fatty acids as phenotypic markers in the polyphasic taxonomy of methylotrophs within the Proteobacteria. J. Gen. Microbiol. 137, 2631–2641. Gujer, W., Zehnder, A.J.B., 1983. Conversion processes in anaerobic digestion. Water Sci. Technol. 15, 127–167. Hansen, L.K., 1998. Biogeochemistry of methane in a shallow sandy aquifer. PhD Dissertation, Technical University of Denmark, Lyngby, Denmark. Harris, S.H., Ulrich, G.A., Suflita, J.M., 999. Dominant terminal electron accepting processes occurring at a ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241238 landfill leachate-impacted site as indicated by field and laboratory measurements. U.S. Geological Survey Water Resources Investigations Report a 99-4018C, Reston, VA. Helgeson, H.C., Garrels, R.M., MacKenzie, F.T., 1968. Evaluation of irreversible reactions in geochemical processes involving minerals and aqueous solutions: II. Applications. Geochim. Cosmochim. Acta 33, 455–481. Heron, G., 1994. Redox buffering in landfill leachate contaminated aquifers. PhD Thesis, Dept. of Environ- mental Science and Engineering, Technical University of Denmark, Lyngby, 49 pp. Heron, G., Christensen, T.H., 1995. Impact of sediment-bound iron on redox buffering in a landfill leachate Ž .polluted aquifer Vejen, Denmark . Environ. Sci. Technol. 29, 187–192. Heron, G., Christensen, T.H., Tjell, J.C., 1994a. Oxidation capacity of aquifer sediments. Environ. Sci. Technol. 28, 153–158. Ž . Ž .Heron, G., Crouzet, C., Bourg, A.C.M., Christensen, T.H., 1994b. Speciation of Fe II and Fe III in contaminated aquifer sediments using chemical extraction techniques. Environ. Sci. Technol. 28, 1698– 1705. Heron, G., Barcelona, M.J., Andersen, M.L., Christensen, T.H., 1997. Determination of nonvolatile organic carbon in aquifer solids after carbonate removal by sulfurous acid. Ground Water 35, 6–11. Heron, G., Bjerg, P.L., Gravesen, P., Ludvigsen, L., Christensen, T.H., 1998. Geology and sediment Ž .geochemistry of a landfill leachate contaminated aquifer Grindsted, Denmark . J. Contam. Hydrol. 29, 301–317. Hodgson, J.F., 1960. Cobalt reactions with montmorillonite. Soil Sci. Soc. Am. Proc. 24, 165–168. Hoehler, T.M., 1998. Thermodynamics and the role of hydrogen in anoxic sediments. PhD dissertation, University of North Carolina, Chapel Hill. Hoehler, T.M., Alperin, M.J., Albert, D.B., Martens, C.S., 1998. Thermodynamic control on H concentra-2 tions in anoxic sediments. Geochim. Cosmochim. Acta 62, 1745–1756. Hoehler, T.M., Albert, D.B., Alperin, M.J., Martens, C.S., 1999. Acetogenesis from CO in an anoxic marine2 sediment. Limnol. Oceanogr. 44, 662–667. Hostettler, J.D., 1984. Electrode electrons, aqueous electrons and redox potentials in natural waters. Am. J. Sci. 284, 734–759. Hunt, M.J., Shafer, M.B., Barlaz, M.A., Borden, R.C., 1997. Anaerobic biodegradation of alkylbenzenes in laboratory microcosms representing ambient conditions. Biorem. J. 1, 53–64. Hunter, K.S., Wang, Y., Van Cappellen, P., 1998. Kinetic modeling of microbially-driven redox chemistry of subsurface environments: coupling transport, microbial metabolism and geochemistry. J. Hydrol. 209, 53–80. Ingvorsen, K., Zehnder, A.J.B., Jørgensen, B.B., 1984. Kinetics of sulfate and acetate uptake by Desulfobacter postgatei. Appl. Environ. Microbiol. 47, 403–408. Ž .Jakobsen, R., Postma, D., 1994. In situ rates of sulfate reduction in an aquifer Rømø, Denmark and implications for the reactivity of organic matter. Geology 22, 1103–1106. Jakobsen, R., Postma, D., 1999. Redox zoning, rates of sulfate reduction and interactions with Fe-reduction and methanogenesis in a shallow sandy aquifer, Rømø - Denmark. Geochim. Cosmochim. Acta 63, 137–151. Jakobsen, R., Albrechtsen, H.-J., Rasmussen, M., Bay, H., Bjerg, P.L., Christensen, T.H., 1998. H2 Ž .concentrations in a landfill leachate plume Grindsted, Denmark : in situ energetics of terminal electron acceptor processes. Environ. Sci. Technol. 32, 2142–2148. Jørgensen, B.B., 1978. A comparison of methods for the quantification of bacterial sulfate reduction in coastal marine sediments: III. Estimation from chemical and bacteriological field data. Geomicrobiol. J. 1, 49–64. Kao, C.-M., Borden, R.C., 1997. Site-specific variability in BTEX biodegradation under denitrifying condi- tions. Ground Water 35, 305–311. Kennedy, L.G., Everett, J.W., Ware, K.J., Parsons, R., Green, V., 1998. Iron and sulfur mineral analysis methods for natural attenuation assessments. Biorem. J. 2, 259–276. Kjeldsen, P., 1993. Evaluation of gas diffusion through plastic materials used in experimental and sampling equipment. Water Res. 27, 121–131. Kohring, L.L., Ringelberg, D.B., Devereux, R., Stahl, D.A., Mittelman, M.W., White, D.C., 1994. Comparison of phylogenetic relationships based on phospholipid fatty acid profiles and ribosomal RNA sequence similarities among dissimilatory sulfate-reducing bacteria. FEMS Microb. Lett. 199, 303–308. ( )T.H. Christensen et al.rJournal of Contaminant Hydrology 45 2000 165–241 239 Ž . Ž .Kota, S., Barlaz, M.A., Borden, R.C., 1999. Influence of Fe II Fouling on Fe III Bioavailability in Aquifer Sediment. Submitted. Landers, D.H., David, M.B., Mitchell, M.J., 1987. Analysis of organic and inorganic sulfur constituents in sediments, soil and water. Int. J. Environ. Anal. Chem. 14, 245–256. Landmeyer, J.E., Chapelle, F.H., Petkewich, M.D., Bradley, P.M., 1998. Assesment of natural attenuation of aromatic hydrocarbons in groundwater near a former manufactured-gas plant, South Carolina, USA. Environ. Geol. 34, 279–292. Langmuir, T., 1997. Aqueous Environmental Geochemistry. Prentice Hall, London, 600 pp. Lendvay, J.M., Sauck, W.A., McCormick, M.L., Barcelona, M.J., Kampbell, D.H., Wilson, J.T., Adriaens, P., 1998a. Geophysical characterisation, redox zonation, and contaminant distribution at a groundwater–surface water interface. Water Resour. Res. 34, 3545–3559. Lendvay, J.M., Dean, S.M., Adriaens, P., 1998b. Temporal and spatial trends in biogeochemical conditions at a groundwater–surface water interface: implications for natural attenuation. Environ. Sci. Technol. 32, 3472–3478. Light, T.S., 1972. Standard solution for redox potential measurement. Anal. Chem. 44, 1038–1039. Lindberg, R.D., Runnells, D.D., 1984. Ground water redox reactions: an analysis of equilibrium state applied to Eh measurements and geochemical modelling. Science 225, 925–927. Lovley, D.R., Chapelle, F.H., 1995. Deep subsurface microbial processes. Rev. Geophys. 33, 365–381. Lovley, D.R., Goodwin, S., 1988. Hydrogen concentrations as an indicator of the predominant terminal electron-accepting reactions in aquatic sediments. Geochim. Cosmochim. Acta 52, 2993–3003. Lovley, D.R., Klug, M.J., 1986. Model for the distribution of sulfate reduction and methanogenesis in freshwater sediments. Geochim. Cosmochim. Acta 50, 11–18. Lovley, D.R., Phillips, E.J.P., 1986. Availability of ferric iron for microbial reduction in bottom sediments of the freshwater tidal Potomac River. Appl. Environ. Microbiol. 52, 751–757. Lovley, D.R., Phillips, E.J.P., 1987a. Competitive mechanisms for inhibition of sulfate reduction and methane production in the zone of ferric iron reduction in sediments. Appl. Environ. Microbiol. 53, 2636–2641. Lovley, D.R., Phillips, E.J.P., 1987b. Rapid assay for microbially reducible ferric iron in aquatic sediments. Appl. Environ. Microbiol. 53, 1536–1540. Lovley, D.R., Baedecker, M.J., Lonergan, D.J., Cozzarelli, I.M., Phillips, E.J.P., Siegel, D.I., 1989. Oxidation of aromatic contaminants coupled to microbial iron reduction. Nature 339, 297–300. Lovley, D.R., Chapelle, F.H., Woodward, J.C., 1994. Use of dissolved H concentrations to determine2 distribution of microbially catalyzed redox reactions in anoxic groundwater. Environ. Sci. Technol. 28, 1205–1210. Ludvigsen, L., Albrechtsen, H.-J., Holst, H., Christensen, T.H., 1997. Correlating phospholipid fatty acids PLFA in a landfill leachate polluted aquifer with biogeochemical factors by multivariate statistical methods. FEMS Microbiol. Rev. 20, 447–460. Ludvigsen, L., Albrechtsen, H.-J., Heron, G., Bjerg, P.L., Christensen, T.H., 1998. Anaerobic microbial redox Ž .processes in a landfill leachate contaminated aquifer Grindsted, Denmark . J. Contam. Hydrol. 33, 273–291. Ludvigsen, L., Albrechtsen, H.-J., Ringelberg, D.B., Ekelund, F., Christensen, T.H., 1999. Composition and Ž .distribution of microbial biomass in a landfill leachate contaminated aquifer Grindsted, Denmark . Microbial Ecol. 37, 197–207. Ž .Lyngkilde, J., Christensen, T.H., 1992. Redox zones of a landfill leachate pollution plume Vejen, Denmark . J. Contam. Hydrol. 10, 273–289. Lyon, W.G., Glass, R.L., 1997. Anaerobic HCl extraction of soils or sediments: rapid sample preparation prior to a screening test for microbially available iron. Report US EPA, Robert S. Kerr Environmental Research Center, Ada, OK, 13 pp. McMahon, P.B., Chapelle, F.H., 1991. Microbial production of organic acids in aquitard sediments and its role in aquifer geochemistry. Nature 349, 233–235. Nicholson, R.V., Cherry, J.A., Reardon, E.J., 1983. Migration of contaminants in groundwater at a landfill: a case study. 6. Hydrogeochemistry. J. Hydrol. 63, 131–176. Nielsen, P.H., Albrechtsen, H.-J., Heron, G., Christensen, T.H., 1995. In situ and laboratory studies on the fate of specific organic compounds in an anaerobic landfill leachate plume: 1. Experimental conditions and fate of phenolic compounds. J. Contam. Hydrol. 20, 27–50. Course 12230 Environmental Chemistry and Ecotoxicology Course 12235 Environmental Risk Assessment of Chemicals Environmental Risk Assessment of Chemicals Anders Baun, Niels Nyholm and K. Ole Kusk Department of Environmental Engineering Technical University of Denmark January 2008 2 Anders Baun, Niels Nyholm and K. Ole Kusk Department of Environmental Engineering Technical University of Denmark Building 115, DK-2800 Kgs. Lyngby, Denmark 5 the substance should be registered with the national environmental pollution agency in question with details concerning the substance’s physical-chemical properties as well as information on the substance’s fate and effect in the environment and possible health related effects and a suggestion as to the substance’s classification. One of the main problems about the directive was that only “new substances” were subject to this procedure, while there were no demands to the 100,000 existing substances. Another main problem was that it was the responsibility of the authorities to assess if there was a risk to humans and environment – not only for the 4.000 new substances, but also for the many existing substances. This has not only been an enormous and time consuming, but also in reality an impossible task, since the data required for the risk assessments simply doesn’t exist for most of the substances. The new EU act on chemicals, REACH – Registration, Evaluation, and Authorization of CHemicals, which stepped into force in July 2007 (EU Parliament and Council, 2007) sets all chemicals equal. All chemicals, independently of how long they have been on the market, have to be registered in the new EU Chemical Agency and their properties documented by information provided by the producers/importers. They also have to make self classification and risk assessment of their substances. The most hazardous of the substances may be subject to a ban or may need an authorization for still being on the market for specific purposes. It is expected that the new law will help to reduce the data gaps by forcing producer/importers to deliver the information necessary to perform risk assessments on their substances. Besides the responsibility for making the risk assessments is now the producers’/importers’ responsibility to document that their chemicals are safe to use and do not harm humans or environment. The amount of information needed for the registration depends on the production/import volume. Thus, more information is needed at larger production volumes than at smaller volumes and most information is needed for the so-called High Production Volume chemicals (HPVC). These are the chemicals which are produced above 1000 tons per year per producer. Besides production volumes special characteristics about a substance can make it a “special case” and increase the information demand. This is the for example the case for highly toxic or carcinogenic chemicals. If the data demand is not fulfilled the product can not be marketed (the “No Data – No Market” principle). In this paper we will focus on xenobiotic (anthropogenic) organic chemicals that will be referred to as “chemicals” or “substances”. Human health risk assessment is not covered as we focus on the chemical substance’s fate and effects in the external environment. Many of the principles and procedures used are similar for human and environmental risk assessments. However, one very important difference between these two types of risk assessments is that the aim of environmental risk assessments is the protection of thousands of different species, while we consider only one species (Homo sapiens) when dealing with human health risk assessments. In environmental risk assessments all spheres have to be included.. While assessments of e.g effects on soil or sediment dwelling organisms are equally important to risk assessments in the aquatic environment, it is still so that the work with standardisation and determination of principles for environmental evaluation has reached the highest level of 6 development for the aquatic environment. Therefore, the aquatic environment has been chosen as the main point of focus for this paper. Table 1.1 Timeline for regulatory actions taken to improve chemical safety in EEC/EU Year Regulatory action 2018 REACH shall be fully implemented 2007 REACH: ”New” and ”existing” substances are set equal. All substances > 1 ton/year shall be assessed with 3, 6 and 11 years time Level of assessment depend on production volume 1992 Environmental classification required for “new” substances 100.000 “existing” substances are not covered by this regulation 1981 The EINECS list (EEC) is formed. 100,000 substances on the European market are registered as ”existing” substances 1979 OECD: Chemical testing program provide guidelines for testing. EEC: Demands for environmental data on substances introduce to the European market (“new” substances) 1.1 Environmental evaluation of chemicals: Risk assessment and risk management The evaluation of a chemical environmental effect can be split into to parts: Risk assessment and risk management. Risk assessment delivers the information that (mainly) on a scientific basis creates the foundation for a possible implementation of risk management. In connection with risk management social and political factors amongst others as well as technical and economic possibilities play a roll. This could for example include weighting of which risks we are willing to accept in relation to which sacrifices we are willing to suffer. In relation to this it is important to underline that the concepts, risk and accept, are always mutually integrated in a complete chemical evaluation. As mentioned, risk/benefit-analysis cannot be solved in a technical-scientific fashion alone, but to a high degree will be controlled by society’s needs and desires/wishes. Even if risk assessment and risk management are two different areas which conceptually should be strictly kept separate - in practice this will often not be possible. The collective risk assessment of chemical substances with regard to the external environment is carried out in 3 steps: Step 1: Hazard identification Step 2: Hazard assessment Step 3: Risk characterisation As illustrated in Figure 1.1 the risk assessment begins with a hazard identification after which one “works his way outwards” and utilises the knowledge that has been gained at the previous levels. 7 The hazard identification serves to map the chemical substances inherent properties - that is the properties a substance has regardless of its production, use and disposal. Thereby a comparison and evaluation of physical-chemical data can be facilitated (i.e. melting point, vapour pressure, solubility, degradability, bioaccumulative ability and ecotoxicological data). With the help of analogue conclusions, rules-of-thumb, QSAR (Quantitative Structure Activity Relationships) and possibly simple mathematical models, the first (rough) environmental profile for a chemical substance can be assembled. The environmental profile describes the substance’s possible behaviour in the environment and especially important element in this environmental profile is the evaluation of the chemical substance’s distribution in the various compartments of the environment (water, soil and air). It is common practice that the connection between exposure and effect should be researched when environmental assessments of chemicals are performed. This results specifically in the hazard assessment, which is divided in an exposure assessment and an effect assessment. Figure 1.1 The three parts of environmental risk assessment: Hazard identification, hazard assessment and risk characterisation. In the effect assessment one of the main purposes is to obtain a quantitative estimate of which substance concentration can be expected not to have any effect on the ecosystem’s species structure and function. This concentration is called PNEC (Predicted No-Effect Concentration) and is set by ecotoxicological tests with regard to uncertainties in extrapolation from known to expected effects. By exposure analysis general scenarios are applied for realistic exposure pathways for the substance. Knowledge gained through the hazard identification is now utilised, but also information about the released amounts and concentrations, transport routes 10 2 HAZARD IDENTIFICATION Hazard identification at an introductory level is a comparison and assessment of the easily accessible data of a substance’s inherent properties. The individual parts, which together make up the hazard identification, will be described in this chapter. As illustrated in Figure 2.1 a long series of processes and reactions are of vital importance for an organic chemical’s distribution and dispersion in the environment and therefore focus will be directed on these connections in the following. At the end of this chapter a summary - “Checklist for hazard identification” - contains information which is designed to help with the completion of hazard identification in concrete cases. Figure 2.1. Processes and reactions of vital importance for a chemical’s distribution and dispersal in the aquatic environment. The typical sources for illustration of a substance’s inherent properties are handbooks and databases available through libraries or via the Internet. In the following table the most important types of data for completion of the hazard identification are given. It should be underlined that it is seldom possible to collect all the data listed in Table 2.1 for a given substance for the simple reason that the data is either non-existent or is not publicly available. Therefore, it may be necessary to estimate unknown values based on substance characteristics for substances, which have structural similarities with the substance in question. In such cases the so-called QSAR-estimations (Quantitative Structure Activity Relationships) can be used. Based on the chemical structure of substances one can use QSAR to estimate for example the physical- chemical parameters on the basis of knowledge gained from other structurally related substances. 11 Table 2.1 Data for implementation of hazard identification of chemicals Physical-chemical Molecular weight (Mw) Melting point (m.p.) Boiling point (b.p.) Vapour pressure (pa) Solubility (S) Acid dissociation constant (pKa) Henry’s Law constant (KH) Octanol-water partitioning constant (Kow) Distribution constant between organic carbon and water (Koc) Accumulation Bio-concentration factors for various organisms (BCF) Degradability Hydrolysis Photolysis Biodegradability evaluated from results of standardised test methods Ecotoxicity Results from standardised ecotoxicological tests (LC50, EC50, NOEC, LOEC) for a minimum of 3 species (algae, crustaceans, and fish) 2.1 Volatilisation Henry’s law can approximately describe the distribution of a substance between the gaseous and aqueous phase (valid for Cw <approx. 1 mol/l). The Henry’s law constant, KH, is the central parameter that describes the substance’s tendency to escape from an aqueous solution to the air. Figure 2.2 shows the phase transition schematically as it is assumed that the two-film theory is valid for volatilisation of a chemical substance from the aqueous phase. Figure 2.2 Phase transfer of a chemical from liquid to gas illustrated with the two- film model 12 If one considers the volatilisation of a substance dissolved in the aqueous phase (concentration Cw), based on this model the substance must first pass through a thin liquid film (see Figure 2.2). A concentration gradient will arise here due to diffusion limits. The dissolved substance thereafter volatilises and must pass through a thin gas film where the mass transfer again is limited by the rate of diffusion. After this the substance actually reaches the gas phase with the partial pressure pi. The transport through the liquid and gas film is described by two mass transfer constants kl and kg, respectively which express the rate of a molecules movement through the film. For the entire border surface the rate of a molecule’s movement is described by the total mass transfer coefficient, KL. The dependency of the concentration in the water and the partial pressure of the substance in the air at equilibrium are given by Henry’s law as: pi =KH  Cw - where, pi is the substance’s partial pressure in the gaseous phase (atm) Cw is the concentration of the substance in the aqueous phase (mol·m-3) KH is Henry’s law constant (atm·m3·mol-1) It should be noted that Henry’s law constant thereby expresses the relationship between the concentrations in the two phases, in that the partial pressure via the ideal gas law can be considered as a concentration (pi= R·T·(n/V)air = R·T·Cair). Henry’s law constant can be looked up in handbooks/databases for a number of substances as well as estimated from the relation: K p SH a - where, pa: The substance’s vapour pressure at temperature T (atm) S: The substance’s water solubility at temperature T (mol·m-3) The unit for Henry’s law constant in the present text is given as: atm·m3·mol-1, but it is important to notice that Henry’s law constant is often dimensionless (denoted K’H or Kair-water) in the literature: RT KKK HwaterairH   ' - where, R: The ideal gas constant (8.2·10-5 atm·m3·mol-1·K-1) T: Temperature (K) The volatilisation rate of a substance is a function of the overall mass transfer coefficient and the concentration difference between water and air given by the expression: 15 Table 2.2 Volatility of a chemical substance from aqueous solution. Volatility KH (atm·m³·mol-1) Phase transfer Less volatile than water < 3·10-7 Gas film controls the rate of volatilisation, which depends on the size of KH Slow rate of volatilisation 3·10-7< KH < 10-5 Gas film controls the rate of volatilisation, which depends on the size of KH Significant rate of volatilisation 10-5< KH < 10-3 Both the gas- and liquid film contribute to the rate of volatilisation dependent on the size of KH Volatile in aqueous solution > 10-3 The liquid film controls the rate of volatilisation independent on the size of KH ------------------------------------------------------- Example: The volatility of toluene from an aqueous solution. The volatility and half-life with respect to evaporation of the substance toluene from a lake is to be determined. The lake has a depth of 1.0 m. The following data is found in a reference on physical-chemical data: Toluene: C7H8 Molecular weight: 92.1 g/mol Vapour pressure: 0.0380 atm Water solubility: 518 mg/l = 518 g/m³ Henrys constant is determined: mol matm mg molgatm S pK aH 3 3 3 1076.6/518 /1.920380.0      For toluene it applies that KH>10-3 atm·m³·mol-1 and the substance can be considered as volatile from an aqueous solution (see Table 2.2). The liquid film dominates in the phase transfer and therefore the overall mass transfer coefficient can be estimated as follows: 1 KL = 1 k l , where k Ml w = 20 32 cm h    1 4/ For substances where KH>10-3 atm·m³·mol-1 KL is no longer dependent on Henry’s constant, but only on the substance’s molecular weight. The half-life in the lake is given by the expression: h 5.4 / 1.92 3220 693.0101 K ln2Z = 4/1 2 L ½       hcm cmt ------------------------------------------------------- Finally, it should be noted in connection with the hazard identification of volatile 16 substances - which have a vapour pressure larger than 100 Pa and which contain fluorine, chlorine or bromine - that a more thorough investigation must be undertaken with regard to the assessment of the substance’s ODP (Ozone Depletion Potential). 2.2 Sorption When assessing an organic chemical’s sorption to a solid phase (soil, sediment or suspended solid) during hazard identification, it will often be sufficient to consider the sorption process as a linear relation between the concentration in the solid phase and the concentration in the liquid phase: Cs = Kd · Cw -where, Cs: The concentration of the sorbed substance (mg · (kg solid phase)-1) Cw: The concentration in the aqueous phase (mg · (litre water)-1) Kd: Partitioning coefficient between the solid phase and water (l·kg-1) The partitioning coefficient, Kd, can be determined experimentally, however when considering neutral organic substances Kd can be estimated as well on the basis of the solid phase’s content of organic carbon (foc, i.e. fraction of organic carbon) and the substance’s adsorption coefficient between organic carbon and water (Koc): Kd = foc · Koc The value, Koc can be found in handbooks or calculated as a rough estimate using the octanol/water partitioning coefficient in the following equation: Koc = 0.41 · Kow (Karickhoff, 1981) The assumption allowing the use of Koc for the assessment of a chemical substance’s sorption is that they bind to the solid phase as a consequence of the substance’s hydrophobic characteristics. This is valid for a number of neutral compounds, while for other substances it can be hydrogen bonds and ion exchange, which dominate the substance’s partitioning between the aqueous solution and the solid phase. As an approximation of a substance’s tendency to be sorbed to a solid phase, the Koc value can be used for neutral compounds as long as the solid phase has a certain content of organic carbon (more that 2%). In soil the substance’s mobility is defined as the distance, a given chemical “travels” in a column of water saturated soil in comparison to the distance the water has travelled. Thus, a large mobility means that the substance has a poor tendency to sorb to soil and it will therefore be transported with the water. McCall et al. (1980) has proposed the following Table (Table 2.3) for assessment of a substance’s mobility in soil based on the substance’s known or estimated Koc value. 17 Table 2.3 Classification of a substance’s mobility in soil based on the Koc value (according to McCall et al., 1980) Mobility in soil Adsorption coefficient (Koc) Very large 0-50 Large 50-150 Medium 150-500 Low 500-2000 Very small 2000-5000 Immobile >5000 ------------------------------------------------------- Example: Sorption of Atrazine in an activated sludge reactor Task: Determine the distribution of the pesticide Atrazine in an activated sludge reactor with a content of suspended solids of 3 g SS/l (Css), where 30% is organic carbon (foc=0,30). A Koc value of 145 for Atrazine is found in the literature. Atrazine actually sorbs to the solid phase by hydrophobic binding but also by hydrogen binding to the surface of the solid matter. In this example it is assumed that only the hydrophobic binding is significant for the sorption. A mass balance is set up for Atrazine: Total amount of substance = Amount in aqueous phase + Sorbed amount of substance mtot,chem = mw,chem + ms,chem Ctot·Vtot = Cw·Vw + Cs·ms Where: mtot,chem = The total mass of the chemical in water and suspended solids (g) mw,chem = The total mass of the chemical in the aqueous phase (g) ms,chem = The total mass of the chemical in the suspended solids (g) Ctot = The total concentration of the chemical (g/m3). Vtot = The volume of water and suspended solids (m3) Cw = The concentration of the chemical in the aqueous phase (g/m3) Vw = Volume of the water (m3) Cs = The concentration of the chemical in the suspended solids (g chemical/g suspended solids) ms = The mass of suspended solids (g). Since the concentration in the sorbed phase is given as Cs = Kd·Cw, the mass balance can be written as follows: Ctot·Vtot = Cw·Vw + Kd·Cw·ms The mass of the suspended solids is given by: ms=Css·Vtot. Kd is determined from the