5

I'm trying to measure small electrochemical signals using a trans-impedance amplifier fed into a lock-in amplifier.

I have an electrochemical cell that poses a very high impedance (100s of MegaOhms) that I feed a triangle wave into (5V P-P) the excitation frequency ranges from 1Hz to 100kHz.

The current across the cell is amplified with a trans-impedance amplifier with variable gain. One problem is that there is a lot of noise in the measurement, mainly due to the strong RF fields around (due to the experimental setup). I have been RF shielding and passively low pass filtering the output to get rid of some noise, but the output waveform is still very noisy, especially at high gain.

I have a feeling that lock-in amplification might be the solution but I am not sure how to set it up. I have a model 5210, Princeton lock-in amplifier at my disposal. I have tried splitting the driving signal and feeding one end as reference to the lock-in. I then feed the output of the trans-impedance amplifier into input channel A and record the X and Y outputs. With various filter settings and phase adjustments, I wasn't able to recover any signal successfully. The resulting I-V curve from similar experiments look a bit like this:

Cyclic voltammetry representative I/V curve

I'd appreciate any help with setting up the amplifier, am I using it correctly or is this approach fundamentally flawed? I have little understanding of the theory behind the instrument so I'm feeling slightly lost at the moment.

zzlmes
  • 53
  • 5
  • 1
    How about a diagram of your grounding approach? Is the RF mainly Efields? – analogsystemsrf Oct 26 '17 at 15:44
  • Somebody correct me if I'm wrong, but don't lock in amplifiers only work on sine waves? – JRE Oct 26 '17 at 15:54
  • What frequency did you modulate at? With 100 Meg ohm and a bit of stray capacitance you are going to have a low pass filter maybe in the 10's of Hertz range. What's the capacitance? – George Herold Oct 26 '17 at 16:28
  • @analogsystemsrf Sorry I'm unable to post a diagram at the moment. The RF is due to capacitive plasma generation upstream. The 'ground' excitation electrode for the plasma is more or less directly above my measurement spot. There is a separate RF ground that is hooked up to the RF generator. The generator pumps about 200W of RF into the feed gas (unbalanced). The trans-impedance amp is battery powered, grounded through BNC outer. I do not think that the problem lies with the setup but rather my lack of experience! – zzlmes Oct 26 '17 at 16:43
  • @GeorgeHerold I was modulating at low-frequencies. 1-20 Hz. I assumed that I'll initially have an easier time at low-freq. The intent is to move to higher frequencies if I can get rid of the noise. I am able to see the overall waveform (despite the noise) using a potentiostat, so I'm sure that the problem is with my use of the lock-in rather than any experimental variable. I'll post scope images when I can! – zzlmes Oct 26 '17 at 16:47
  • 2
    @zzlmes, so you can see a noisy signal on the 'scope? Is it a digital 'scope? You can make a 'poor man's lockin' with a DSO. Send the signal into one channel, and modulation into the other. Trigger on the modulation and hit the average button. – George Herold Oct 26 '17 at 17:31
  • @JRE I've ran sine input too. Didn't notice any improvement. You're right though, as far as I know most lock-ins work with sine or square wave inputs. – zzlmes Oct 26 '17 at 17:51
  • @GeorgeHerold Yes I can, I've actually done exactly what you suggested with the Tektronix DPO7054 that I have. The overall shape of the V/I curve is visible. The problem is that the signal I'm looking for is a minute peak in the cyclic voltammogram (with variable baseline). In order to detect it, I probably need the derivative which doesn't get along well with noise. – zzlmes Oct 26 '17 at 17:54
  • @zzlmes, if you are seeing the whole I/V curve on the 'scope then you are not using the lockin the way I pictured. What you'd like to do (I think) is modulate the current (slightly) and use the lockin to measure this signal, the amount of current, as you then slowly sweep the voltage. I need to think about it a bit more. Can you post a pic/ scribble of how you hooked up the lockin? – George Herold Oct 26 '17 at 20:21

1 Answers1

2

OK I think I see now, you're sweeping the voltage and measuring current. In that case to use the lockin you need to modulate the voltage slightly. And then sweep the DC voltage slowly through your signal. In this configuration the signal you will see is the slope of the I/V, the derivative you were seeking.

You have to adjust the modulation amplitude, depending on the 'size' (in volts) of the feature you are looking for.

For adding a modulation voltage onto the slow sweep, I might first try an opamp summing circuit. You could also do it with an audio transformer.

George Herold
  • 4,754
  • 1
  • 18
  • 25
  • 1
    Yes, this was the issue. Summed a tiny sinusoidal signal over my driving waveform and after fiddling with the timebase, voila! Thanks a bunch George! – zzlmes Oct 27 '17 at 17:20