How to choose electrochemical analyzers for low-current measurements

A low-current electrochemical analyzer should be selected by the smallest signal change it can distinguish under the required measurement conditions—not by the number of digits it displays. At picoampere and nanoampere levels, input bias, leakage, drift, cell capacitance, and measurement bandwidth can matter more than nominal current resolution.

The strongest evidence is a repeatable measurement using the intended cables, shielding, acquisition settings, and a representative cell. An analyzer that produces a quiet trace after prolonged averaging may still be unsuitable for a short pulse, a rapid potential scan, or a measurement immediately after a large charging transient.

Define the measurement before comparing instruments

“Picoampere capability” is not a complete requirement. A steady 10 pA signal measured over several seconds presents a different problem from a 10 pA change that must be resolved within milliseconds. Likewise, measuring an absolute current accurately is different from detecting a small change against a larger background.

A useful application specification establishes:

• The smallest meaningful current change: the difference that affects the experimental conclusion, rather than merely the lowest expected current.
• The measurement window: when the current must be captured after a potential step, pulse, or other disturbance.
• The full current excursion: including charging currents, startup events, and unexpected reaction currents.
• The cell conditions: expected impedance, capacitance, reference-electrode characteristics, and applied potential range.
• The allowable uncertainty and drift: over one reading, one experiment, and repeated measurements.

For example, resolving a hypothetical 2 pA change on a 100 pA baseline requires more than a display increment below 2 pA. Noise, baseline drift, range-dependent error, and blank variability must be sufficiently smaller than that change. If only relative changes matter, a stable offset may be tolerable; if absolute current is required, the same offset becomes part of the uncertainty budget.

Separate resolution, noise, and accuracy

Resolution describes the smallest reported increment. It does not establish whether successive increments contain useful information. Additional ADC bits or software averaging cannot remove leakage, correct gain error, or recover a transient that the analog front end has already filtered out.

Supplier specifications become comparable only when their definitions and conditions are aligned.

An accuracy expression containing both a percentage-of-reading term and a percentage-of-range term deserves particular attention. At low currents, the range term can dominate. Calculate the specified error at the actual operating current on the range that will be used—not just at full scale.

Check what an “input bias current” specification refers to. In a potentiostat, reference-electrode input bias and working-electrode current-measurement offset are different quantities. Reference-input bias produces a voltage error through the reference source impedance; current-measurement offset directly affects the reported working-electrode current. A single headline bias value does not characterize both.

Compare noise at the bandwidth the experiment needs

A low noise figure is meaningful only with its measurement bandwidth. Strong filtering can produce an exceptionally smooth trace while obscuring the response of interest. Conversely, a wider-bandwidth instrument may appear noisier while preserving information that a slower configuration removes.

For approximately white current noise, RMS noise scales with the square root of equivalent noise bandwidth. That relationship does not fully describe low-frequency drift, mains interference, or electrochemical fluctuations. It explains, however, why noise figures obtained with different filter settings cannot be ranked directly.

Require the supplier to identify the analog bandwidth, digital filter, integration time, output data rate, and any averaging used. Sampling rate alone is insufficient: recording many samples per second does not guarantee that the analog signal path responds equally quickly. Filter delay also matters when a current reading must correspond to a specific point in a pulse sequence.

The smallest current range commonly uses high transimpedance gain, with a potential trade-off in bandwidth and settling time. For steady-state measurements, this may be acceptable. For pulse methods, verify performance within the intended sampling window rather than accepting a steady-state noise specification.

Treat the cell and cabling as part of the measurement system

The potentiostat’s current front end does not operate independently of the cell. Cell capacitance, cable capacitance, and control-loop settings can affect noise, settling, and stability. Voltage noise can be converted into current noise through the connected impedance, so a quiet input test does not necessarily predict performance with an electrochemical cell.

Charging current also creates a range-selection problem. For a capacitor, i = C × dV/dt. As an illustrative calculation, 1 µF subjected to a 0.1 V/s potential ramp produces 100 nA of capacitive current before any faradaic contribution is considered. An experiment targeting picoampere features may therefore require an instrument to handle currents several orders of magnitude larger during other parts of the waveform.

Check how electrochemical analyzers respond to overload and current-range changes. Relevant characteristics include settling time, missing data, switching artifacts, and recovery after saturation. A sensitive range is of limited use if it cannot recover before the required reading. Fixed-range operation can avoid switching artifacts, but only if the range accommodates the waveform without sacrificing necessary sensitivity.

Control-loop stability must also be demonstrated with representative impedance. Ask which stability or bandwidth settings are required and whether those settings compromise the measurement timing. Reference-electrode impedance and counter-electrode compliance should remain within the instrument’s operating limits throughout the experiment.

Leakage and interference can consume the entire signal budget

At picoampere levels, insulation is an active selection concern. From Ohm’s law, 1 V across a 1 TΩ leakage path produces 1 pA. Connector contamination, moisture, damaged insulation, or an unsuitable switching accessory can therefore create an error comparable to the intended signal.

A low-current specification should identify whether it applies at the instrument connector or includes the supplied cable and cell fixture. It should also state relevant temperature, humidity, and voltage conditions. Performance demonstrated with short, clean laboratory connections may not transfer to a longer installation or a humid enclosure.

Shielding and guarding solve different problems. A conductive shield reduces electric-field interference. A driven guard reduces leakage and effective capacitance along supported high-impedance paths by holding a surrounding conductor near the sensitive node’s potential. A guard is not simply another ground, and its connection must follow the instrument’s architecture. Triaxial cabling is valuable where the analyzer actually supports the corresponding guarded arrangement.

Mechanical handling matters as well. Cable motion can generate triboelectric currents, while vibration and changing capacitance can produce disturbances. Low-noise cables should be secured during verification. A Faraday cage can reduce external electric-field pickup, but it will not eliminate internal leakage, ground-loop currents, or real fluctuations in the cell.

Compare instruments in the intended connection configuration, including any computer, stirrer, temperature controller, or other attached equipment. Adding accessories after a successful bench demonstration can introduce new interference paths.

Use staged verification to distinguish instrument limits from cell behavior

A convincing acceptance test separates electronic performance from electrochemical variability. A quiet trace alone is not enough: the system must also reproduce a known signal and respond correctly to changes.

Establish the electronic baseline

Begin with the manufacturer-recommended low-current test configuration. Record the range, bandwidth, filters, warm-up time, cable arrangement, shielding, temperature, and humidity. An open input, shorted input, and resistive load exercise different conditions; none should be treated as a universal representation of an electrochemical cell.

Evaluate baseline mean, RMS variation, drift, and occasional disturbances over a duration relevant to the experiment. Preserve the time series rather than accepting only a cropped screenshot. Long records can reveal thermal drift or periodic interference that a short demonstration misses.

Verify response with a characterized electrical load

A suitable high-value resistor network or low-current source can test gain, polarity, and response at relevant current levels. The test source must have sufficiently low leakage and known uncertainty. At very high resistance, resistor tolerance, voltage dependence, temperature sensitivity, and fixture insulation can limit the reference itself.

Add representative capacitance or an appropriate dummy-cell network to evaluate settling and stability. Apply a waveform that includes the expected transient and observe the return to the low-current region. This is more revealing than testing only a constant current near the center of the most sensitive range.

Confirm performance with the intended chemistry

Use a blank and a representative sample with the actual electrode configuration. Compare baseline stability, repeatability, and the response to a controlled experimental change. If electrical tests pass but the blank remains unstable, electrode condition, contamination, dissolved gases, temperature, or cell construction may be setting the practical limit.

Do not translate an instrument noise figure directly into an analytical detection limit. Detection capability also depends on blank variability, calibration sensitivity, sample preparation, and the measurement protocol.

Make the purchase decision against an agreed operating configuration

Request verification data for the configuration being purchased, not simply the best result available elsewhere in the product family. That includes the current-measurement module, cable set, shielding, software options, and any multiplexer. Switching hardware is particularly important because it can add leakage, capacitance, and settling requirements.

Acceptance limits should reflect the application: maximum baseline variation over a defined interval, acceptable drift, error at selected current points, recovery after a specified transient, and repeatability with a representative cell. Record the settings needed to meet those limits. Otherwise, a supplier may demonstrate low noise using averaging that makes the required experiment impossible.

Software should preserve the evidence behind the result. Confirm access to numerical data, timestamps, range-change records, acquisition settings, and filter information. Automatic smoothing or zero subtraction should be identifiable and controllable. Calibration documentation should cover the relevant ranges and uncertainty; a calibration certificate alone does not establish performance with the intended cell.

The defensible choice is the analyzer that meets the signal, timing, and stability requirements with a reproducible setup. A lower quoted noise floor offers little benefit if achieving it requires unacceptable averaging, frequent manual correction, or a connection arrangement that cannot be maintained. Once the required performance is demonstrated, compare the complete installed cost—including low-current accessories, calibration, and verification effort—rather than paying for sensitivity the experiment cannot use.