High Performance, Low Cost…

XMOS Galvanostat

Here, an XMOS Startkit shield (see image below) is used to build a very low cost trace metal analyzer, employing the electrochemical technique known as potentiometric stripping. Akin to anodic stripping voltammetry, this involves pre-concentrating metal ions from a stirred solution onto a metal film electrode (MFE). A negative potential is applied to a glassy carbon working electrode (WE) and a thin film of bismuth forms over the electrode surface (n.b. bismuth is added prior to the start of the experiment by the analyst to give a concentration of 1 ppm) . During the deposition period metal ions from solution are reduced and the metals amalgam into the Bi film.

After the deposition period (which typically lasts several minutes) the control voltage is disconnected from the WE and the potential of this electrode is monitored over time. Oxidants present in solution (O2 and Bi3+) cause sequential re-oxidation of the metals held in the MFE, resulting in a series of waves in the potential vs time trace (for an example, see here).
To conduct an experiment, the shield pictured above connects to the users PC via an FT232R usb-to-serial converter at J5 and to two electrodes, a Ag|AgCl reference electrode (RE) and a glassy carbon working electrode (WE) via connections on the mini-DIN6 connector. Stir and purge logic control signals are also brought out to this connector.

The on-board circuitry is organised to hold the WE at ground while the potential at the RE is controlled by the galvanostat. During the deposition phase the WE must be taken negative relative to the RE and so a positive potential is applied at the RE.

The XMOS processor on the Startkit generates a pulse width modulation (PWM) signal that is fed to an instrumentation amplifier (AD623) via an RC filter. This amplifier provides a gain of 2 but also offsets the signal so that both positive and negative potentials can be applied at the RE. The application of these potentials is controlled by a solid state relay (Goodsky DIR-S8-105A), with a software timer determining the duration of the deposition step.

The RE is also connected to a JFET op-amp (TL-072) configured as a low gain voltage amplifier. The output of this amplifier appears on a two pin IDC header at the bottom right of the shield (J2). The voltage present on this header is connected to the Startkit’s on-board high speed, 12 bit analog-to-digital converter (ADC) via a short, two wire ribbon cable.

Each ADC conversion result (one sample every 10 microseconds, in the range 0-4095) is used to increment the corresponding element in a 4096 element buffer array. By recording data in this manner for a period of 10-20 seconds, a history is accumulated of the amount of time that is spent at each potential. The resulting histogram shows a series of peaks, one for each metal re-oxidized.

The XMOS shield is controlled from a LabVIEWTM front panel.
The potentiometric stripping trace above was obtained using a pH 4.5 acetate buffer solution containing 1 ppm Bi and 100 ppb each of Zn, Pb and Cd. During a 4 minute deposition period the solution was stirred. Near the end of this period, stirring was turned off, the control potential disconnected and the potential of the reference electrode monitored over time. Three potentiometric stripping peaks seen here are left: Zn, middle: Cd and right: Pb. Note the excellent signal-to-noise ratio achieved in this experiment.

The rise in counts at the right edge of the trace is due to stripping of Bi. At the conclusion of a “scan” the RE is taken negative (relative to the WE) to completely remove the Bi film. This ensures good reproducibility run-to-run.

The versatility of potentiometric stripping is nicely illustrated in the black and red traces shown opposite.

In the black trace, the XMOS galvanostat described above is first used to measure a solution containing 100 ppb each of the elements indium (In), tin (Sn) and thallium (Tl). Only two peaks are seen here as the stripping potentials for In and Sn are too close for their peaks to be separately resolved.

An additional 100 ppb each of Zn, Cd and Pb are then added to the electrochemical cell and the run is repeated (red trace). The peaks from these added metals appear at the same potentials as were observed in the blue trace shown above, together with the In/Sn and Tl peaks that were seen in the black trace.

Notice that in the red trace five different metal signatures are visible in the one scan! It is noteworthy that in a square wave stripping voltammetry (SWSV) experiment on the In,Sn,Tl sample studied here that only a single broad peak was observed.

A second galvanostat XMOS-based shield has been developed and this is shown in the image below. This employs an Analog Devices AD5542A 16 bit DAC and a Texas Instruments ADS8681 16 bit ADC, potentially offering higher performance than the previous design, both in terms of analog input resolution and speed (1 usec/point).
Galvanostat waveforms can be recorded in one of two ways with this second system. The picture below shows a voltage vs time trace after the deposition voltage has been disconnected. Three waves are seen in this trace, corresponding to the oxidation of first Zn, then Cd and then Pb.

Alternatively, one can collect data in a “histogram binning” mode. When operating in this manner a memory buffer space is first allocated. Once the deposition voltage has been disconnected, the ADC is polled at full speed and after each reading the ADC count is used to index and then increment a single bin in this buffer. By doing so, the bins record counts proportional to the amount of time that has been spent at each potential during the V versus t trace. This is the method that was used in recording the blue trace with the first instrument, described earlier.