The quality of a DLE control system is limited by the quality of its sensors. A control model that is sophisticated in its electrode physics and brine chemistry representation generates setpoints that are only as good as the sensor readings feeding it. Choosing the right sensors and placing them in the right locations in the process are design decisions that affect control performance throughout the plant's operating life.
What You Need to Measure and Why
The minimum viable ionic monitoring stack for electrochemical DLE control consists of four measurements: Li+ concentration at the cell feed inlet, Mg2+ (or at minimum, a proxy for total divalent cation load) at the cell feed inlet, conductivity inline at the cell feed inlet, and temperature inline at the cell feed inlet. These four measurements, combined with electrode potential and current from the power supply, give the control model enough information to run adaptive setpoint control for most continental brine applications.
Adding Li+ at the cell outlet (the effluent stream) enables a second critical function: real-time electrode state-of-charge estimation, and early detection of breakthrough. When outlet Li+ starts rising from near-zero (which indicates the electrode is removing nearly all feed Li+) toward the feed concentration (which indicates the electrode is saturated and no longer removing Li+), the rate of rise and the timing relative to the start of adsorption give the control model a direct measurement of electrode behavior that can be compared against the predicted behavior. Discrepancies between predicted and actual outlet Li+ during adsorption are the first indication that the electrode condition has changed: fouling is reducing effective area, capacity has faded, or feed composition is different than expected.
The more complete sensor stack adds individual Mg2+ and Ca2+ measurements (rather than just divalent proxy), pH inline, and silica (periodic, typically by grab sample and lab analysis rather than continuous inline for routine operations). Each additional measurement improves model accuracy at the cost of sensor procurement, installation, and maintenance.
Ion-Selective Electrode Sensors: Capabilities and Constraints
Ion-selective electrode (ISE) sensors are the primary measurement technology for inline Li+ and Mg2+ monitoring in brine processing. An ISE consists of a membrane that is selectively permeable to the target ion, generating a Nernstian potential difference proportional to the logarithm of the target ion activity. In practice, the relationship is modified by the Nikolsky-Eisenman equation, which quantifies interference from other ions present at concentrations high enough to compete with the target ion for the membrane selectivity mechanism.
For Li+ measurement in a brine with 500 to 2,000 mg/L Li+ and 20,000 to 80,000 mg/L Na+ (a typical continental brine), the Na/Li molar ratio is in the range of 20:1 to 60:1. Most commercial Li+ ISE membranes (chalcogenide glass or lithium silicate membranes) have Na+/Li+ selectivity coefficients in the range of 0.001 to 0.01, meaning Na+ interference at these concentration ratios produces an apparent Li+ reading error of 2 to 6% without correction. This level of error is acceptable for many monitoring applications, but in control applications where the control model is sensitive to small changes in Li+ concentration, the correction should be applied using a simultaneously measured conductivity or Na+ value.
For Mg2+ measurement, the ISE interference situation is more complex because Ca2+ and Na+ both interfere with Mg2+ ISE membranes. At typical salar brine Mg/Ca ratios of 5:1 to 20:1, the Ca2+ interference on a conventional magnesium ISE produces errors of 5 to 15% in the Mg2+ reading. For control purposes, this is usually acceptable because the control model uses Mg2+ as a proxy for total divalent cation pressure on the electrode, and Ca2+ is contributing to that same pressure (albeit via a different mechanism than Mg2+). A biased Mg2+ reading that reflects both Mg2+ and Ca2+ activity may actually be a more relevant proxy than a pure Mg2+ reading in high-Ca brines.
Inline Conductivity: The Workhorse Sensor
Inline conductivity measurement is the most robust and cost-effective continuous monitoring technology available for brine processing control. Electromagnetic conductivity sensors (toroidal or contacting two-electrode) provide a continuous signal proportional to total dissolved ionic solids with response times under 1 second and minimal maintenance requirements compared to ISE sensors.
The limitation is that conductivity is not specific to any individual ionic species. It reflects the combined ion load. For DLE control, conductivity provides two useful functions: detection of rapid brine composition shifts (any significant change in total ionic content produces a detectable conductivity change faster than the ion-selective sensors respond), and correction of ionic strength effects on ISE sensor readings (the high-ionic-strength correction applied to ISE readings uses conductivity as an input).
In dilute produced water service, where Li+ concentrations may be below 100 mg/L, the conductivity signal is dominated by Na+, Ca2+, and other major ions at levels that are typically 100 to 1,000 times higher than Li+. Conductivity alone cannot detect the small composition changes relevant to Li+ monitoring in this application. For produced water, continuous ISE measurement at the cell inlet is more important, and more challenging, than for high-Li+ salar brine service.
Sensor Placement Logic
Where a sensor is placed in the process determines what it measures and what it can contribute to the control model. The key placement decisions are: upstream of pre-treatment (characterization only, not used for real-time control because what enters pre-treatment does not equal what exits), downstream of pre-treatment and upstream of the DLE cell inlet (the primary control sensor location), within the cell outlet or effluent stream (electrode state monitoring), and at the eluate product stream (product quality monitoring).
The primary control sensor at the pre-treatment outlet/cell inlet needs to be as close to the cell inlet as practical, to minimize the hydraulic transit time between the measurement point and the electrode face. A sensor placed 10 meters upstream of the cell inlet on a 3-cm diameter pipe at 3 m3/hr flow has approximately 95 seconds of transit time before its reading reaches the electrode. A sensor placed 2 meters upstream has approximately 19 seconds of transit. The difference matters when brine composition can change over minutes and the feedforward control model is using the inlet reading to predict what the electrode will see.
Multi-cell stacks should have a separate outlet sensor for each cell, not a common outlet monitoring point. The reason: if cells are at different points in their adsorption cycles (which is normal in a multi-cell configuration), a common outlet sensor measuring the combined effluent cannot distinguish which cell is approaching breakthrough and which is operating normally. Per-cell outlet monitoring is more expensive but enables independent cell state estimation, which is necessary for the independent cell cycle scheduling function described in our closed-loop control article.
Sensor Maintenance in Field Conditions
The most common operational failure mode for ISE sensors in field brine service is membrane fouling and degradation. ISE membrane materials are designed for laboratory conditions and do not always perform as specified in continuous industrial service. Salar brines with high TDS saturate some membrane formulations over time, shifting the calibration baseline. High-silica geothermal brines coat ISE membranes with silica films that attenuate the sensor response. Produced water with residual hydrocarbons or suspended solids blocks membrane pores.
The practical maintenance design response is to choose ISE sensors with field-replaceable membrane assemblies (rather than sealed sensor bodies that require complete sensor replacement when the membrane fails), to establish regular calibration intervals appropriate for the brine chemistry (every 2 to 4 weeks for typical continental brine, every 1 to 2 weeks for geothermal or produced water), and to implement sensor validation in the control system that flags readings that deviate beyond expected bounds from the conductivity-based baseline estimate. That last function, in particular, prevents a degraded ISE reading from driving incorrect setpoint adjustments without alerting the operator.
Sensor redundancy (duplicate ISE sensors at the same measurement point, with cross-validation logic) adds capital cost but dramatically reduces the control system's sensitivity to individual sensor failures. For the primary Li+ inlet control sensor, where a sensor failure would cause the control model to operate on stale or invalid data until maintenance can replace the sensor, redundancy is worth considering for projects where unplanned DLE cell downtime has significant production cost implications.