Debugging an Experiment-- Chem 128

lab date: September 27, 2002

web notes: September 27, 2002

• The experiment: determination of iron by spectrophotometry
  • reaction with 1,10 phenanthroline
  • calibration curve with known iron samples
  • finally, analyze for iron in the legendary "green crystal"
    • K₃Fe(C₂O₄)₃ 3 H₂O
• The calibration curves from the morning class looked pretty good
  • (nice straight lines with few points that didn't fit fairly well.)
  • so it looks like the procedure andour technique is good

• However, the determination of Iron in the green crystal didn't fare so well
  • we expect 1 mole of Fe per mole of salt
  • formula weight is 491 for this salt
  • so we expect a ratio of 1/491 = 0.00204 moles/gram
• The results I saw were about 30-40% of that value
  • so what's wrong?
  • So we decided to let the afternoon lab do some detective work-- see if we could solve the puzzle.

• Trouble shhooting in the lab
• Typically we look at the various steps of the experiment and see what might cause an error
  • we plans some tests that can confirm of eliminate some of the suggestions
  • we may find ways to modify the experiment and overcome the original problem.

• Some possible explanations
  • 1. The standard Fe solution was incorrectly prepared
    • all measurements would then be off by the same ratio
    • we'd still get great looking calibration graphs
    • they'd have good precision, lousy accuracy
    • (that would be a massive error-- we are trying to account for afactor of 2-3)
  • 2. One of the other reagents was prepared incorrectly
    • for example, if there wan't enough phenathroline, we wouldn't convert all the iron to colored species and we'd get a low value for iron.
    • we can reject this because we had higher readings for the standard solutions, so we couldn't have run out of reagent.
    • we could also have run out of reducing agent (hydroxylamine hydrochloride)
    • this wouldn't happen for the standards, since they had very little Fe³⁺
    • in practice we add much more reducing agent than sample, so this doesn't sound like a good answer.
• 3. The green crystal was treated with extra acid
  • it's possible that the ammonium sulfate buffer was insufficient to compensate for the acid.
  • this means that both the reduction and the reaction with phenathroline were done in situations that might be too acidic to be effective
  • notice that the extra acid was added to the larger flask, so we only ended up with 1% of that acid in our final flask.The buffer should be able to handle it.
  • 4. It's possible that the oxalate is just too good a ligand-- it never let's go of the iron
    • this would give low values for the iron analysis
    • we assume that the phenathroline is able to bind Fe stronger and would soon force out the oxalate-- but that might take too long to be effective.

Some tests we ran in the afternoon lab

• 1 Some samples were treated with hot, concentrated sulfuric acid
  • this is a dangerous process
  • but it should be able to oxidize(and destroy) the oxalates
  • these samples were otherwise treated the same as the green salt
• 2. Some samples were boiled in concentrated hydrochloric acid
  • at the concentration levels, Cl- should act as ligands and should be able to compete for the iron. This might displace the oxalate.
• 3. Some samples were treated with MnO₄- (titrated)
  • at this point there should be no oxalate remaining
• 4. One sample was heated to high temperatures (600-800^oC)
  • hopefully enough to allow the carbon in the oxalate to burn
• 5. we also had two groups work with pure ferrous ammonium sulfate
  • if they get a correct answer we can rule out errors in the concentration fo the Fe standards.
• 6. We did not do this but we might watch the pH
  • increase the amount of buffer
  • measure pH and compare to the expected range

• 0. The analysis of ferrous ammonium sulfate gave a results quite close the the expected result
  • we can conlude there is no serious error in the iron standard solution or the calibration curve
  • the molar extinction coefficent computer from the data was approximatley 11500-12200 (1.15-1.22 x 10⁴.)
    • I hope to find an accepted value for comparison .
• 1. A direct analysis of the green crystal gave a result that was about 50% of the epxected value.
• 2. Samples boiled in HCl gave values that we about 90% of what's expected.
• 3. The sample treated with concetrated H₂SO₄10:11 PM 9/27/02 gave result nearly identical to the expected result.
• 4. The sample that was heated to high temperatures did not produce useful results.
• 5. The sample treated with permanganate-- results are not known at this time

• heating with concentrated sulfuric acid effectively destroys the complex ion, making all the iron available to analysis.
• heating with conc. HCl helps, but is not sufficient
• We are hoping the permanganate treatment proves useful
  • It is a much easier, safer technique.