Electrochemistry in Geophysics
In a simplistic way, crustal rocks can be thought of as consisting of three phases: mineral(s), pore fluid(s) and a boundary phase between the minerals and pore fluids. This boundary electrochemical layer has unique properties. These unique chemical surface properties affect many physical rock properties of interest to geophysicists. This special session on ``Induced Polarization, Self Potentials and Seismic-Electro Coupling: Opportunities and Challenges for the 21st Century'' is acutely aimed at these surface effects and their importance. Induced Polarization is mostly related to adsorbed surface species. Self Potentials arise when gradients in temperature, chemical potentials or hydraulic gradients (producing fluid flow) exist in the earth. Seismo-Electro Coupling refers to the production of electrical signals when a fluid flows transiently, due for example, to the passage of a seismic wave. Electro-Seismic Coupling is the reverse process. What is exciting about these processes is that there is, embedded in their signatures, detailed knowledge of the surface adsorbed chemistry and the pore fluid. The challenge is to be able to measure, understand and decipher these signatures in a robust manner. In my view, this influence of Electrochemistry in Geophysics represents the most exciting area of geophysics and my expectation is that it will yield abundant results in such diverse areas as earthquake prediction, near-surface contaminant monitoring and resource assessment and development. In this presentation I will review the broad importance of chemistry in geophysics and comment critically on some of the concerns and issues that need to be addressed for progress in the fields of Induced Polarization, Self Potentials and the Seismo-Electric and Electro-Seismic phenomena.
Applications of the seismoelectric method: Insights from field experiments and numerical studies
Application of the seismoelectric method (seismic-to-electric conversions) to problems of interest is hindered by: 1) strong coherent noise (much of it created by the seismic source) that obscures the signal of interest, 2) difficulty interpreting with certainty the observed data, and 3) uncertainty regarding the problems to which the method is best-suited. We address all three of these problems with examples from field experimentation and from numerical simulations. Physical separation of the seismic source from the electrode receivers can separate, in time, the signal of interest from coherent source-generated noise (the signal arrives before the noise), eliminating the need for more complex and generally less effective signal/noise separation techniques. We illustrate the effectiveness of this approach with field data and simulated examples that each provide an image of a pair of embedded thin layers that would be difficult to image with conventional seismic methods or other geophysical techniques. In a set of numerical simulation results, we show that the amplitude of the seismoelectric response from a thin layer does not fall off as rapidly with decreasing layer thickness as the seismic reflection from such a layer. These results demonstrate the sensitivity of the seismoelectric method to thin layers, and show that even for a layer that is one twentieth the seismic wavelength, the seismoelectric interface response is stronger than for a single interface. In a numerically-simulated, down-hole, time-lapse, study we show that such time-lapse data can be directly subtracted to remove noise (which is constant) and reveal the signal (which in this case changes with salinity of groundwater in a remote sand lens). This example demonstrates how such experiments could be used to remotely monitor temporal variation in groundwater chemistry. Together, our results indicate that the seismoelectric method is best employed using survey geometries with source and receivers separated (e.g. down-hole and/or cross-well experiments), and that time-lapse studies are particularly promising.
Borehole Measurements of Interfacial and Co-seismic Seismoelectric Effects
We have recently carried out a series of seismoelectric field experiments employing various hammer seismic sources on surface and a multi-electrode `eel' lowered into slotted PVC-cased boreholes penetrating porous sediments. Deploying grounded dipole receivers in boreholes has a number of advantages over surface-based measurements. Ambient noise levels are reduced because earth currents from power lines and other sources tend to flow horizontally, especially near the surface. The earth also provides natural shielding from higher frequency spherics and radio frequency interference while the water-filled borehole significantly decreases the electrode contact impedance which in turn reduces Johnson noise and increases resilience to capacitively- coupled noise sources. From a phenomenological point of view, the potential for measuring seismoelectric conversions from various geological or pore fluid contacts at depth can be assessed by lowering antennas directly through those interfaces. Furthermore, co-seismic seismoelectric signals that are normally considered to be noise in surface measurements are of interest for well logging in the borehole environment. At Fredericton, Canada, broadband co-seismic effects, having a dominant frequency of 350-400 Hz were measured at quarter meter intervals in a borehole penetrating glacial sediments including tills, sands, and a silt/clay aquitard. Observed signal strengths of a few microvolts/m were found to be consistent with the predictions of a simplified theoretical model for the co-seismic effect expected to accompany the regular `fast' P- wave. In Australia we have carried out similar vertical profiling experiments in hydrogeological monitoring boreholes that pass through predominantly sandy sediments containing fresh to saline water near Ayr, QLD and Perth, WA. While co-seismic effects are generally seen to accompany P-wave and other seismic arrivals, the most interesting result has been the observation, at three sites, of interfacial seismoelectric effects that appear to be caused by the arrival of the P-wave at the water table located 3 to 14 m below surface. The signals can be observed arriving simultaneously across dipoles located up to 20 m below the water table and are also detected by dipoles located at surface. Polarity reversals are observed 5 to 11 m below the water table at the three sites. These reversals may be explained by our 2 m dipole receivers passing inside a vertical electric dipole produced by seismoelectric conversion, and thereby provide compelling new evidence in support of existing models for the generation of seismoelectric effects at interfaces.
The Origins of the Self-Potential Response During Hydraulic Fracturing
The self-potential (SP) response during hydraulic fracturing of intact granite specimens was investigated in the laboratory. Excellent correlation of pressure drop and SP suggests that the SP response is created primarily by electrokinetic coupling. For low injection pressures, the variation of SP with pressure drop is linear, revealing a constant coupling coefficient (Cc) of -200 mV/MPa. However for radial pressure gradients > 80 MPa/m, the magnitude of the Cc increases in an exponential trend by up to 80% preceding hydraulic fracturing. This increasing Cc is related to increasing permeability at high pore pressures caused by dilatancy of micro-cracks, and is explained by a decrease in the hydraulic tortuosity. Other variables which may also effect the Cc include: a) decreasing specimen resistivity, b) increasing porosity at high pore pressure, c) increasing zeta potential of fresh microcrack surfaces, d) increased effective viscosity by electroviscous effects, and e) flow separation at high fluid velocity. Additional source mechanisms such as piezoelectricity and co-seismic electrokinetics are considered and are shown not to contribute significantly to the SP response during hydraulic fracturing. At the moment of fracture initiation, injectate rushes into the new fracture area where the zeta potential is likely greater than in the preexisting rock porosity, and an anomalous SP spike is observed that may represent the transient SP response to a changing Cc. Finally, during tensile cracking of wet granite specimens in a point load device with no water flow, a SP transient is created by contact electrification. However, the time constant of this event is much less than that for transients observed during hydraulic fracturing, suggesting that SP created solely from material fracture does not contribute to the SP response during hydraulic fracturing.
Bacterial Nanowires Facilitate Electron Transfer in Saturated Porous Media
Bacterial nanowires are electrically conductive appendages produced by bacteria in response to electron acceptor limitation and may contribute to biogeophysical signatures in saturated subsurface sediments. Using Shewanella oneidensis strain MR-1 as a model organism, we performed controlled laboratory column experiments that conclusively demonstrated that nanowires are necessary to develop self potential (SP) signals spanning more than 500mV. Oxygen, which served as the sole terminal electron acceptor and diffused into the upper portion of the open column, supported cell respiration and viability. Scanning electron microscopy verified the presence of nanowire network that physically and presumably electrically connected cells throughout the column. Sterile control columns and columns inoculated with a mutant strain of S. oneidensis that produced significantly less conductive nanowires developed minor SP signals (up to 10mV). Our results suggest that microbial activity and nanowires greatly impact the electrical properties of porous materials and contribute to our understanding of the mechanisms that underlie geophysical methods for mapping microbial activity in near subsurface environments.
The Role of Geoelectrical Methods in Monitoring Stimulated Sulfate-Reduction: Insights Gained From Field-Scale Experiments
Understanding how microorganisms influence the physical and chemical properties of the subsurface is hindered by our inability to observe microbial dynamics in real time and with high spatial resolution. Here we investigate the use of time-lapse geoelectrical methods to monitor stimulated sulfate-reduction at the field scale during in-situ acetate amendment at the Rifle, Colorado uranium mill tailings site. Modification of the pore fluid and sediment composition as a result of bisulfide production and mineral precipitation was concomitant with changes in induced polarization (IP) and self-potential (SP) signals. With data collected from both the surface and between boreholes, temporal variations in the IP response were characterized by the development of pronounced phase anomalies related to the precipitation of disordered mackinawite (FeS). Sediment samples recovered from the aquifer showed a close correlation between the location of the IP phase anomalies and the enrichment of acid volatile sulfides. Variations in borehole SP signals closely tracked the onset of sulfate- reduction and primarily resulted from an increase in the concentration of bisulfide adjacent to the measurement electrodes. The magnitude of the SP response was dominated by the galvanic interaction of metallic copper and bisulfide, and it closely approximated the electrochemical cell potential of the anodic and cathodic reactions occurring at the electrode surfaces. Both geolectrical techniques delineated spatially discrete anomalies that appear to reflect the interaction of biostimulation with lithological variability within the aquifer.
Relaxation Time Distribution From Time Domain Induced Polarization: Synthetic and lab Data
The main goal in use of Induced Polarization method (IP) in hydrogeophysics is to find links between parameters of the electrical polarization and lithologic (textural) parameters. Useful links between the characteristic time of the polarization (relaxation time) and characteristic sizes of textural elements (pore or throat size, grain size) have been found on the basis of frequency domain (FD) measurements. Earth materials contain textural elements of different origin, of different size, and of different chemical composition, and, therefore, they could produce different relaxation time of the electrical polarization. It follows to the relaxation time distribution (RTD), which characterizes rock/soil texture. RTD are usually extracted from FD measurements. Here we present an approach allowed extracting RTD from time domain (TD) measurements. First we discuss an algorithm to invert TD IP data to RTD. We based the algorithm on the (Tikhonov) regularized solution of the 2nd kind Fredholm integral equation. We test the algorithm on synthetic data, and we show its robustness for reasonable levels of noise, typical of TD laboratory measurements. Then we used the algorithm to invert experimental data obtained on mixtures of sieved sands. We show that peaks in RTD are related to sizes of grains. We also discuss data obtained on sandstone samples and we compare RTD with other petrophysical data.
Systematic Errors in Resistivity and IP Data Acquisition: Are We Interpreting the Earth or the Instrument?
For decades, resistivity and induced polarization (IP) measurements have been important tools for near-surface geophysical investigations. Recently, sophisticated, multi-channel, multi-electrode, acquisition systems have displaced older, simpler, systems allowing collection of large, complex, three-dimensional data series. Generally, these new digital acquisition systems are better than their analog ancestors at dealing with noise from external sources. However, they are prone to a number of systematic errors. Since these errors are non-random and repeatable, the field geophysicist may be blissfully unaware that while his/her field data may be very precise, they may not be particularly accurate. We have begun the second phase of research project to improve our understanding of these types of errors. The objective research is not to indict any particular manufacturer's instrument but to understand the magnitude of systematic errors in typical, modern, data acquisition. One important source of noise, results from the tendency for these systems to both send the source current, and monitor potentials through common multiplexer circuits and along the same cable bundle. Often, the source current is transmitted at hundreds of volts and the potentials measure few tens of millivolts. Thus, even tiny amounts of leakage from the transmitter wires/circuits to the receiver wire/circuits can corrupt or overwhelm the data. For example, in a recent survey, we found that a number of substantial anomalies correlated better to the multi-conductor cable used than to the subsurface. Leakage errors in cables are roughly proportional to the length of the cable and the contact impedance of the electrodes but vary dramatically the construction and type of wire insulation. Polyvinylchloride, (PVC) insulation, the type used in most inexpensive wire and cables, is extremely noisy. Not only does PVC tend to leak current from conductor to conductor, but the leakage currents tend to have large phase shifts/time lags that mimic IP effects. A second source of substantial systematic errors is the tendency of these systems to use the same, simple metal electrodes as current sources for some data and receiver points at other times. Using the electrode as a current source results in the electrode retaining substantial voltage (often hundreds of millivolts) that decays over time. The form of this decay voltage can be fairly complex making it difficult to remove even with long periods of signal averaging. Finally, there are a number of other, smaller but potentially significant systematic errors such as errors due to the limited common-mode rejection of the multi-channel receivers and even leakage of potential from receiver to receiver when electrodes are shared between adjacent measurement channels.