SP corresponds to in situ electrical signals associated with charge polarization mechanisms occurring at depth. Two main sources are generally discussed. One source is electrokinetic coupling [2,32], which is controlled by the relative motion between the charged mineral surface and the excess of charges located in the electrical diffuse layer of the pore fluid [6]. In the shallow subsurface, the motion of fluids forced by a pressure gradient through a porous medium can generate electrokinetic effects. Amplifications due to the presence of a two-phase flow may also be considered [1,12,24], as surface conductivity effects tend to be enhanced by a reduction of the conducting volume. While the role of two-phase flow is still being discussed in the literature, Guichet et al. [8] have shown that the elektrokinetic effect tends to decrease when the water content is decreased.

The diffusion of ionic species creates another SP source associated with electrochemical coupling [20]. More precisely, the electrochemical potential is the sum of (1) liquid junction potential due to ionic movements at the boundary between two different electrolytes and (2) membrane potential related to the presence of clay. Redox potential might also contribute to the overall signal in the presence of metals or “appropriate” contaminants [25]. Both laboratory and field experiments suggest that electrokinetic potentials generated by fluid flows into porous geological materials are often the main source of SP signals.

Once diffusion processes are taken into account, the SP profiles provide an electrical signature of fluid flow in the subsurface or in the borehole. In particular, this method may help to determine zones either of infiltration, and water recharge, or of runoff.

MuSET was designed in the framework of ALIANCE as a joint project between CNRS (Montpellier, France) and ALT (Redange, Luxembourg) in order to investigate downhole hydro-electrical coupling phenomena. The aim was to measure the downhole SP as a function of depth simultaneously with other sensors (p, T, pH, Eh, Cw) traditionally used to characterize fluid movements in boreholes penetrating the shallow subsurface. Simultaneous recording of fluid parameters may permit a control of the one involved in the analysis of surface electrical processes. Such a tool had never been built, although several experiments involving a single SP sensor were made in the KTB deep exploration hole [38].

With MuSET, the electrical spontaneous signal is measured by two unpolarizable electrodes, for example, one of them being located at the surface far from the borehole (distance > 10 meters) which reduces the possibility of measuring effects due only to the fluid in the borehole. MuSET is an extrapolated version of the ALT Idronaut hydrochemical downhole sensor, to which a standard unpolarizable Pb/PbCl₂ electrode [27] with a resolution range of 0.5 μV (Fig. 1) has been added. This electrode is isolated from the metallic cable connected to the tool. The operating range of MuSET covers temperatures up to 70 °C and pressures up to 20 MPa. This sensor may be deployed either for the exploration of natural settings or during active experiments such as pumping tests.

MuSET was first used at the Lavalette site, which is located 3 km to the northeast off-campus, on the outskirt of Montpellier (France). This site serves as a logistical and experimental base for tool development, testing and calibration. The Lavalette site is equipped with three nearby boreholes 70 to 100 m in depth, including a fully-cored one (LAV1). The holes are aligned along a north–south direction (Fig. 2). Below an 8- to 12-m-thick, irregular Holocene cover, the three boreholes cross-fractured marly limestone of Valanginian age (Fig. 2). The Quaternary sediments result from deposits of the nearby Lez River. The groundwater level is stable all year round at about 4 m depth.

MuSET was deployed in LAV1 to document spontaneous electrical potentials in the absence of nearby active pumping. The analysis of MuSET records required a detailed subsurface characterization. To achieve this, an exhaustive geophysical and hydrological logging program [7,9] was conducted, including borehole wall images, electrical resistivity, spectral natural gamma (with total and U/Th/K profiles), full waveform acoustic velocities, fluid temperature and fluid electrical conductivity. The millimetre to centimetre scale structures, mostly fractures and bed boundaries, were described and oriented from logs and high-resolution images [19] respectively. Geophysical data were also used to characterize the penetrated lithology, to identify the presence of clays (or alteration) and to detect discontinuities at depth, such as fault zones. Finally, borehole fluid measurements (p, T, Eh, pH, Cw, SP) contributed to relate fluid movements to the structure.

In particular, the natural gamma record illustrates downhole variations in clay content between marl-rich and limestone layers of Valanginian age (Fig. 3), with imbricated high and low frequency downhole changes. The optical images confirm the presence of marly beds, mostly dipping 30 to 60° to the south-west (above 40.8 m), and to the south (below) (Fig. 3). From borehole wall images, open fractures crossing the borehole are identified and differentiated, either completely or partially filled with calcite (Fig. 3a). The dip angle and the dip direction of fractures are measured from each sinusoid. Although the analysis of these structural data documents large deformations for the Valanginian layer, the fractures do not show a preferential orientation. This is attributed to the complex geological structure in this area resulting from two main tectonic phases: the south–north Pyrenean compression, followed by a NW–SE Oligocene extension. Note that fractures appearing as open from borehole wall images are not necessarily hydraulically transmissive since they may be connected only in the very near vicinity of the borehole. Changes in borehole fluid measurements indicate the presence of fluid flows with the identification of a main transmissive zone close to 60 m depth. The SP record obtained in LAV1 (Fig. 3c; black curve) is found to almost perfectly overlay the natural gamma profile (Fig. 3c; green curve). This suggests that the presence of clay is responsible for most of the SP signal at Lavalette, related to membrane diffusion in the clay.

The ALIANCE saline wedge experimental site is located at Ses Sitjoles, 8 km southwest of the city of Campos, in the southeastern part of the island of Mallorca (Fig. 4a,b), and 6 km from the Mediterranean coast. Subsurface tomography from surface electrical investigations performed in this area indicate that seawater runs more than 15 km inland, resulting in an important pollution of the shallow groundwater. The experimental site is now equipped with 12 holes, all 100 m in depth, including 7 fully-cored boreholes, distributed over a nearly perfect 100-m side square (Fig. 4c). The cores have been fully characterized in petrological, petrophysical, hydrological and geochemical terms, especially a chosen “reference hole” (MC2) located at the southern tip of the site [14,21]. All boreholes cross the Miocene carbonate platform of South-East Mallorca. The structure is rather classical for reef deposits, with shallow platform deposits near the top (surface to 28 m in MC8), reef constructions below (28 to 63 m), and slope deposits underneath (63 m to bottom). The groundwater table lies at a depth of ∼ 37 m, which corresponds to a meter above the present sea level at the coast.

While a detailed description of the site was obtained from cores and by downhole geophysical methods, the fluid conductivity profiles confirm the presence of a saline wedge at depth (Fig. 5b). The top of the transition zone between fresh and saline water is located in MC8 at ∼ 63 m depth, precisely at the reef to slope transition. Active pumping in the fresh water body for agricultural and domestic reasons appears to further enhance the saline intrusion into the southeastern part of Mallorca. Flowmeter tests were also used to identify potential vertical flow in the holes [17,37]. In particular, a heat-pulse flowmeter [10] was deployed in MC8, under ambient and stimulated (pumping) regime [26], at several measurement points along the borehole (Fig. 5c). These measurements demonstrate the existence of an upward vertical flow, ranging from 0.4 to 0.6 mm/s, mostly from 57 m to 45 m depth. The flow is found to be relatively constant between 57 m and 45 m. This vertical flow may be related to the connection, within the vertical borehole, between horizontal fluid flow in the freshwater zone due to nearby agricultural pumping (at ∼ 100 m scale).

Within the Ses Sitjoles boreholes, the geological structure is particularly well described from millimetre-scale optical images of the borehole surface, recorded in water for the fluid-saturated section, or in air through the vadose zone. While mineralised zones appear to correspond to past/present fluid interfaces, several karstified intervals are found to dominate the overall transmissivity of the structure (Fig. 5a). In the fluid-saturated section, the natural gamma counts are very low, except for two zones at 48 and 60 m depth (Fig. 5b). This signal is related exclusively to changes in U content, with high U concentrations (and low Th/U ratios) associated with the presence of organic matter and the presence of coralline algae.

The SP profile obtained in borehole MC8 is presented next to heat-pulse flowmeter results (Fig. 5c). In the absence of clays for the fluid-saturated section of this geological setting, the absence of fit between the SP and natural gamma profiles exclude here a contribution of the membrane component to the natural electrical signal. A good correlation between the SP signal and the flowmeter data between depths of 55 m and 45 m is observed. This corresponds to the vertical flow zone detected with the flowmeter test, indicating a strong hydro-electrical coupling at these locations.

However, the SP profile does not match the heat-pulse flowmeter data at 57 m depth, above 45 m and just below 60 m depth. This difference might be explained by horizontal fluid flows in the freshwater zone. At 57 m depth, while the heat-pulse flowmeter is indicating a relatively constant vertical flow, the peak observed in the SP signal suggests an additional horizontal component associated to a karstified zone (Fig. 5a). Above 45 m depth, while the heat-pulse flowmeter describes an absence of vertical fluid flow in the hole, the SP response appears to point to a different process and horizontal flow in the hole. Below 60 m, Fickian diffusion associated with a “liquid-junction” component in the transition zone between fresh (above) and saline water (below) is a candidate to explain the recorded electrical signal, while the heat-pulse shows the local absence of vertical flow.