Dear everyone,
I am trying to perform a CSV for the quantification of the additives from a of a commercial copper electroplating bath. For now I'm interested in the quantification of the brightener. I have seen that the analysis is performed using an excess of suppressor and then a linear regression is done (i.e. attached article). Someone uses the linear approximation technique (LAT) or a modified linear approximation technique (MLAT), the latter I don't understand what it consists of. Unfortunately, with this method I greatly overestimate the amount of the additive.
I have attached a graph of my results. I used a commercial bath with the following characteristics: CuSO4*5HO2 210 g/L, H2SO4 65 g/L, Cl- 100 ppm. From the technical data sheet the optimal conditions are with 6 mL/L of make up (suppressor?) and 0.3 mL/L of brightener.
Initially, I measured the virgin make-up solution (VMS) containing only the suppressor and not the brightener. The working electrode was a rotating disc electrode (RDE) in platinum with a diameter of 3 mm, the voltammetry was performed from +1.5 V to -0.25 V vs Ag/AgCl with a scan rate of 100 mV/s and a rotation speed of 2000 rpm. I integrated the anodic peak area of the second scan. The measure corresponds to the blue diamond in the graph. Then I made three measurements adding 75 uL/L of the brightener each time (the blue dots). As you can see from the figure the linear regression does not pass through the first point. I have seen that someone subtract or divide the first point from the others for a better reproducibility but it doesn't work anyway.
Then I measured also a new solution containing 60, 135 and 210 uL/L of the brightener (orange dots) which it does not appear to be reproducible.
Has anyone experience with this technique or can help me understand why I can't make a correct quantification?
Thank you
Clearly something went wrong calculating the regression line.
You might try this easy to use program: www.lerenisplezant.be/fitting.htm, also with different models.
The fact that the second series of points is on a different line, indicates there might be some forgotten parameter in the experiment that changed.
Dear Koen,
I explained the problem wrongly: the fitting is done only on the blue points. Reading in the literature, the blue diamond should be the intercept of the line but it is far below. So I was wondering if there was any precaution to follow when measuring the CSV.
I edited the original question: the voltammetries were performed on an RDE with 2000 rpm rotation speed.
Walter Giurlani Yes, but even if the line is fitted through the blue points, it is clearly wrong; it should be steeper. If you can give us the raw data points (and the measurement precisions), I can show you what I mean.
Of course, these are the data points:
blue spots
x y
0.075 12.180
0.150 13.346
0.225 14.125
blue diamond
x y
0.000 9.879
orange spots
x y
0.060 13.848
0.135 14.492
0.210 15.233
Thank you
* Isn't the blue diamond supposed to be on the same line as the other blue points? It seems you left it out for the fitting.
* What changed in the circumstances with the second series?
* Are the precisions really 0.001? Hard to believe...
* An exponential curve like the charging of a capacitor fits much better, but you would need much more data to confirm that. Can you do some extra measurements with many different concentrations?