Test For Lead in Pewter

Intro
Pewterware is cheaply available from antique and second hand stores; in some cases is essentially +90% pure tin. Pewterware is useful for casting as well as for use of a source of tin for reductions etc. For casting pieces used to serve food, pewter which contains significant amounts of lead is of course undesirable. Therefore, I have developed and optimized a test for lead in pewter. There are some kinks that could still be worked out (which would only serve to lower the LOD) but as it stands the test is more than sufficient for this application.

Pewter
"Modern pewter is tin hardened by adding a few percent of other metals to make it serviceable. The usual range of composition of English pewter or Britannia metal is tin 90-94%, antimony 5-7% and copper 1-3%. Old pewter generally consisted of an alloy of tin and lead. The Romans used tin 80, lead 20%, and tin lead alloys were in general use until fairly recent times; as however, much of the old pewter was made by melting down earlier pewter vessels the composition gradually became somewhat indefinite." --Thorpe's Dictionary of Applied Chemistry, 1943.

Test
The working idea is to dissolve the pewter in aqua regia and then add an iodide which should produce the insoluble PbI2 as a yellow precipitate.

Experimental

Materials


(everything)

Fuming Nitric acid (sulfate free!)
20% Hydrochloric acid
Decolorized Iodine Solution (sold at pharmacy in first aid department, essentially a solution of ammonium iodide in aqueous ammonia with some added EtOH)
Ca(NO3)2*3.7H2O

60 Sn, 40 Pb Solder (Positive control)
95 Sn, 5 Sb Solder (Negative Control)
Unknown Pewter Platter produced after 1964

Note: The nitric acid must be sulfate free as lead sulfate is insoluble in water, likewise SnSO4 easily hydrolizes to form the insoluble oxide. Since my nitric acid was prepared from H2SO4 and a nitrate, lower bp fuming acid should carry over very little H2SO4

First attempt
Little pieces of each sample were collected and digested with 2-3 mL of 1:4 nitric/hydrochloric acid. This was hastened by heating over open flame. Once the metal had fully dissolved, the samples were allowed to cool and the decolorized iodine was added dropwise; this rapidly reacted with the excess nitric acid forming molecular iodine and little precipitate.


(Approximate sample size)

Second Attempt
The digestion was just like the first, except after the metal dissolved, the excess HNO3 was removed by reduction with sodium metabisulfite. Unfortunately this produced a lot of precipitate (of PbSO3 presumably); upon addition of the iodide the PbI2 precipitate was hard to detect.

Third Attempt
Digestion as the first, but the solutions were concentrated to ca. .3 mL by boiling, more hydrochloric acid was added and the solutions were again concentrated. This was repeated until, in total, 3 concentrations were performed. Upon addition of the iodide, the positive controls showed very clear precipitate but the solution was yellow/red from liberated iodine. The Negative control showed no precipitate however was red/yellow in color. While this test is acceptable, it is not very decisive for very small sample sizes.

Fourth Attempt
Like third, except solutions were concentrated to dryness (without overheating) followed by re-suspension in HCl, concentration to dryness and finally resuspension in HCl.
The results for this were better than the 3rd attempt, but I was still unhappy with the decisiveness of the test. Some samples showed definite liberation of iodine (as could be noticed by boiling the solutions which caused the color to lighten to nearly colorless).

Fifth Attempt
The samples were dissolved in hydrochloric acid containing calcium nitrate, about a spatula tip of calcium nitrate was used and the solution volume was 2mL. The amount of calcium nitrate should be as little as possible to fully dissolve the metal. The digestion was again hastened by heating. After digestion, the sample was allowed to cool and iodide was added. For samples which show no precipitate but still had yellow coloration, an excess of iodide was added, and the solution was boiled for a minute or two. If after boiling the solution is appreciably colorless, the test is a negative. If the solution still has distinct yellow coloration that is not significantly changed by this process, the test is a positive for lead.


(Samples being digested. From left to right: Neg. Control in Aqua regia, Pos Control in Ca(NO3)2/HCl, Pos Control in Aqua Regia, Unkown in Aqua Regia)


(Same sample order as above after digestion, the nitric acid controls have been boiled with HCl. The unknown, is showing PbCl2 precipitate and is ready for another concentration)


(concentrating down an HCl/HNO3 Sample)


(Positives: Left to Right: Pos. Control Ca(NO3)2/HCl, Pos Control aqua regia concentrated to dryness (fourth attempt), unkown aqua regia concentrated (third attempt))


(Left: HCl/Ca(NO3)2 Pos Control (5th attempt). Right: 5th attempt Neg. Control after boiling off I2)

Discussion and Conclusion
The unknown was found to contain a significant proportion of lead.

What I thought was going to be a simple test was complicated by the ease at which iodide is oxidized by nitric acid to molecular iodine and nitrogen oxides. This made it necessary to have as little excess nitric acid as possible in the test solution. Removal of nitric acid, by way of HCl oxidation was an effective way to remove the bulk of the nitric acid, however enough nitric acid remained to greatly decrease the test sensitivity. Concentrating to dryness was fairly effective but upon redissolving in HCl the solution was sometimes faintly turbid, which made positives harder to distinguish from negatives. With both the third and fourth attempts, the results were not very consistent; some showed little iodine liberation while others showed significant free iodine. I attribute this to differences in boiling time for trial 4 and residual nitric acid adhering to the test tube walls for trial 5.

The method with Calcium nitrate suffered from the same problems as that with nitric acid, except calcium nitrate is much more convenient to use. Even without boiling the possible negative tests to remove molecular iodine, the negatives were more decisive than the nitric acid tests probably because I used much less total nitrate in these tests. It seems probable that if the nitric acid tests were performed with the minimum amount of nitric acid to dissolve the sample they would be more accurate.

In the fifth attempt, boiling the solution to drive off molecular iodine is only applicable when there is excess iodide with respect to nitrate in solution. If there is excess nitrate, all the iodide will simply be oxidized and driven off as iodine. I believe that the lingering yellow coloration in my negative controls is mostly due to lead impurity in my negative control.

I am sure someone will ask, why I didn't simply neutralize the solutions with base and then run the tests. I didn't do this because I didn't want even the slight precipitation of salts to interfere with the test. These insolubles make it hard to gauge how much of the precipitate is the lead iodide and how much is simply the basic nitrate or oxide of tin.

The final word is, if you want to see if your sample contains a few percent lead, follow the fifth attempt but don't do the second part to remove molecular iodine. If you want a more accurate test for lead do the step to remove molecular iodine.

I also want to point out that I am a inexperienced inorganic chemist, so be gentle with critique

I carried out my own tests:

0.87 g of tin/lead solder (15 y old 60 Sn / 40 Pb - admittedly this is more than just a lead 'contamination') alloy and 0.91 g of lead (roadkill: from a car wheel balancing weight) where treated with about 10 ml of 38 w% HNO3. With the solder reaction started immediately but the lead was more sluggish. After briefly heating the solder it dissolved in about 15mins, forming a precipitate of SnO2. The lead by contrast had to be steam bathed for about an hour.

At the start:



The solder solution/precipitate was then filtered and slightly diluted and divided into three portions. The third was repeatedly (2 x) boiled to almost nothing to get rid of excess HNO3. The filtrate is slightly yellow because the solder contains a fluxing resin core.

Solder filtrate:



NaCl (left), K2Cr2O7 (middle) and KI (right) were then respectively added to tube 1, 2 and 3:



It's always a joy to see the typically delayed precipitation of PbCl2: in this case it was like snowing inside the tube, with crystals as long as 4 mm needles formed. With KI, no I2 was formed.

The lead nitrate solution had developed a white precipitate but clearly crystalline in nature. The clear liquid was decanted off and some 10 ml cold water added to the precipitate which dissolved quickly and completely upon shaking. That liquid was then added to the rest and divided over three test tubes.

NaCl (left), K2Cr2O7 (middle) and KI (right) were then respectively added to tube 1, 2 and 3:



Here the higher concentration of Pb caused the PbCl2 to drop out more quickly in a 'cheese-like' precipitate. Notice also how the iodide tube did develop some iodine: here the tube hadn't been boiled to expel HNO3.

Some interesting additional observations were also made. Turns out that the SnO2.n H2O precipitate was soluble in hot 22 w% HCl, from which with 5 M NaOH fresh Sn(OH)4 can be precipitated which redissolves effortlessly to Na2Sn(OH)6 with more 5 M NaOH.

That would explain why smuv's experiments the tin did dissolve into 'aqua regia' and a mixture of HNO3 and HCl may indeed be the best solvent for tin. Sn (IV), at least according to the potential series, can be reduced to Sn (II) by H2 (but H2 cannot further reduce Sn (II) to Sn (0), at least not in watery solution). So adding Zn and HCl should make the hydrogen in statu nascendi reduce the Sn (IV) to Sn 2+. Precipitating with ammonia should cause Sn(OH)2.n H2O to form, while the Zn ammonia complexes (Sn(OH) is also amphoteric - see stannites - but I doubt if NH4OH is alkaline enough). This may be a faster route to Sn (II) salts than direct dissolution in HCl which is reported to be very slow...


Some more results with HNO3/HCl mixtures with surprising results to be posted shortly...


[Edited on 24-10-2010 by blogfast25]

Quote: Originally posted by smuv

@blogfast: With dilute solutions of "HNO3" produced from Ca(NO3)2 and 20% HCl, full digestion of the metal could be achieved in less than 15 min, with intermittent heating over open flame. Maybe #1 was slow for you because you had a lower HCl concentration which caused passivation by surface oxide? When I digested only tin containing solders, the HCl/HNO3 solution stayed pretty much colorless (even though the nitric acid was orange to begin with). The lead bearing solders developed orange coloration but, not extremely intense. The pewter samples seemed to develop more orange coloration than the lead ones, I am not sure why. As I stated in one of my replies, according to literature (which is congruent with my experiences) aqua regia will not precipitate tin as the oxide.

Et voila, 2 quick tests with the pewter, #1: 0.4 g with a few ml of 22 % HCl; #2: 0.9 g with 1 ml HNO3 38 w% and a few ml of HCl 22 %.

First up #1 which at RT (but it's cold in my lab) reacted slowly but very noticeably. On heating in bunsen flame the reaction rate becomes entirely acceptable, vigorous I'd say and finished in about 10 min, reheating regularly when needed. On steam bath, dissolving this pewter with conc. HCl should be doddle.

#2 was very similar but faster I'd say (I didn't compare in real time). No Sn oxide was formed. A slight yellow colour developed (NOCl?). The question is what's the oxidation state of the Sn in this run: +II or +IV? Do nitrate ions oxidise Sn +II to +IV? Since as air oxygen does it I'm guessing NO3- can do it too but I saw or smelled no NO/NO2... I'll have to look that up. The oxidation state +IV is what I need for my purpose.

In both cases there was a small amount of unknown black insoluble residue, the solutions were filtered.

Here's the two test tubes after very careful neutralisation with 5 M NaOH. That doesn't tell much at all about the oxidation state of either, although I assume that #1 is pure Sn +II.

Quote: Originally posted by smuv

So if I understand this correctly blogfast, you used HCl alone to digest mug that had lead free pewter? Or did it contain lead? It was a good idea blogfast to buy something meant to be eaten out of, even if it is old there is a much lower chance of it containing lead.

Some more tests...

The oxides in both test tubes were dissolved in HCl, both dissolved well, #2 (very subjectively) a wee bit slower.

The black, insoluble residues of both test were decanted and combined and subjected to 38 % nitric. The residues reacted vigorously, SnO2 and NOx were formed, it would appear the residue is mainly tin. Filtering gave a lime green filtrate:



This did not respond to K2Cr2O7 indicating this pewter is relatively lead-free. Also it didn't respond to NH3 1.5 M: no precipitate and no blue complex: this rules out Cu2+...

Two other, larger scale test were then carried out with the aim of synthesising some more (NH4)2SnCl6.

First a test with 3.85 g pewter and 20 ml HCl 22 %. This is the stoichio amount (+ small excess) for obtaining SnCl4 (even though in the first step only SnCl2 is formed). Dissolution on steam bath took about 20 mins but left some black insoluble residue behind. The solution was filtered and 3 g of HNO3 38 % was added: this is stoichoi + plus small excess for 3 Sn (II) + 2 HNO3 --> 3 Sn (IV) + 2 NO (not fully balanced here). That was simmered for a few minutes. No NOx was observed. Then about 3 g of solid NH4Cl was added and the solution simmered with reflux until boiling behaviour became odd. On cooling to RT (then refrigerator), typical crystalline sheets of (NH4)2SnCl6 appeared. Nice crop too!

The second test (with 5.1 g pewter) was very similar but here the HNO3 was added right from the start to the HCl. Events unfolded quite differently though.

Although this too started out fine at BP, about half way the reaction completely died, due I believe to passivation of the tin which manifests itself in a stubborn black sludge coating the metal. It was removed, the solution set aside and the metal scrubbed with sand paper and dissolved in virgin HCl (w/o HNO3). Some black residue remained. I combined the first solution with the latter and added some more HNO3 (I patently forgot to filter first!). On simmering something strange happened: all of a sudden the solution 'whooshed', threw some outside the beaker (not very much) and dissolved also the residue! On top of that, the solution was now greenish:



This was brought to the boil and solid NH4Cl was added and something strange happened again: a whitish precipitate formed. SnO2? (NH4)2SnCl6? Something else? What ever it was on adding quite a bit of HCl 22 % and heating, the precipitate dissolved completely: by then the volume was about 125 ml. That was boiled down to half way and I got my supper. On returning, in the cool solution quite few white crystals had formed. Very well formed. Except... they didn't look like any of the (NH4)2SnCl6 previously obtained. I collected the crystals on a plastic tea strainer, then washed them with cold DIW: they are essentially white. They dissolve in water and tested negative for Pb2+ (I suspected maybe Pb(NO3)2 because they look like it) but positive for behaviour of (NH4)2SnCl6: the solution sheds SnO2 on heating and also when strong alkali is added.

Finally the rest of the solution was reduced and cooled: more crystals formed...


[Edited on 26-10-2010 by blogfast25]

Quote: Originally posted by smuv

Very cool blogfast. I love how experimental and photo-rich this thread has become. Keep in mind that there are non-negligable amounts of copper and antimony in lead free pewter. Cupric oxide could be responsible for some of the insoluble crap. About the (NH4)2SnCl6, why don't you take a sample of the product you are unsure about and heat it in a test tube with aqueous NaOH. If the smell of ammonia is noticed as well as precipitation of Stannic oxide (that part you already recorded) I think it is pretty fair to say you have ammonium hexachlorostannate, I also would like to warn you, that most 'pyrex' kitchenware is no borosilicate glass. So watch out with heating over naked flame if it isn't. You can check by looking at the edges of the glass, if they are strongly blue/green they are normal soda-lime glass, if they are only faintly green you have borosilicate.

Firstly, it's not the aluminium that does the reducing here but the hydrogen, in particular monoatomic hydrogen (hydrogen radicals, if you prefer), generated by Al + acid ---> Al3+ + H* (2 H* --> H2). This method (also with zinc or iron) is often used as a reduction agent. It reduces for instance Ti (IV)aq to Ti (III)aq. SO2? Worth looking into, IMHO...

Re copper, I'd have to look that up, I think it might be reduced to elemental copper, as antimony. But I have to look both up.

Here's what I did.

1.92 g of pewter was first attacked with 25 ml simmering HCl 22% and when that reaction showed signs of slowing down, 5 ml of HNO3 38 % were added. The reaction that followed was almost instantaneous: large eruption of NOx, remaining residue completely disappeared and a clear, lime green solution formed. This is the first time things were so straight forward with my reagents and pewter.

The solution was then quantitatively transferred to a Nalcene round calibrated 250 ml measuring flask and diluted to the mark, resulting in a stock solution of approx. 0.1 N (=0.05 M) Sn, probably mainly Sn (IV).

The titrant solution was ferric alum 0.1 N (0.1 M) but years old. Kept in the dark and well stoppered it looked in good condition with no precipitation or observable hydrolysis. Of course it would at least have to be re-standardised but here the purpose was to show a principle.

My titration station:



20.0 ml of stock solution were pipetted into a clean conical flask and 25 ml HCl 22 % and 0.6 g of Al added. The Al used is extremely low in C, BTW from similar experience (titration of Ti (IV)aq. ) After a bit of heat the reaction starts. And here's the snag: fairly small amounts of an amorphous looking black powder started to form. After all Al was digested, 3 ml of KSCN (Fe3+ indicator) were added. I decided to titrate with the precipitate present. It's clear the reaction Sn (II) + 2 Fe (III) ---> Sn (IV) + 2 Fe (II) [1] proceeds: each time you add some titrant solution, temporarily the red colour of FeSCN (2+) appears then fades quickly as the Fe (III) is reduced away.

A the start of a reduction: no black residue formed yet:



But there's a snag: endpoint determination was difficult with a slow and 'spongy' endpoint. End point was around the 20 ml area.

The second test was identical, except that I filtered the reduced solution quantitatively (it was then clear as water) and then titrated it. That improved things but endpoint was till difficult to determine, I believe because at near endpoint of [1] the reaction speed had dropped to almost nothing. Near the end, when you add another drop the solution turns, say amber red, but in minutes gets back to completely clear! A remedy would be to increase the concentration of both solutions to about 0.5 N, instead of 0.1 N.

But it's been shown that titrating Sn with Fe3+ after reducing to Sn (II) should be possible in the right conditions.

As regards the nature of the black residue, I'll get some more tomorrow to investigate...

[Edited on 28-10-2010 by blogfast25]

Yes, methylene blue would be one of them. Need to find out what's redox potential at the turn, though. And I haven't got any right now.

The test with CuCl2 + HCl + Al surprised me a little. Here's a pic in mid-reaction: Cu is dropping out like the clappers, much faster than would happen in neutral conditions (all Cu dropped out):



Regards the residue, I'm now more or less convinced that it has to do with the Al and with the acid reserves in the sample, which may have been borderline. These are teething problems that should be easy to iron out...

Residue:



It glistens because it's still wet.

Yesterday I made 250 ml stock solution that was approx. 0.5 N in Sn (IV) and 250 ml of ferric alum ( 0.5 N, as yet not standardised) (in 0.1 M HCl) as titrant solution. This works much, much better than 0.1 N: endpoint is still a bit shaky but much easier to establish. Tomorrow the 0.5 N solution will be standardised.

250 ml of approx. 0.5 N ferric alum in 0.1 M HCl:



Finally, another 10 g pewter was dissolved into HCl/HNO3 and an excess of NH4Cl added. The solution was then gently simmered down until the first crystals appear.

After a bit of standing (solution still hot - about 50 C) and a nice crop of (still presumed (NH4)2SnCl6) crystals has already appeared (yield determination tomorrow):



I'll also try and make the potassium salt, which is likely to be even less soluble at RT and which has been used for the separation of Rb and Cs.

Finally regarding the antimony: it should be detectable as (black) Sb2S3 in not too acidic conditions by saturating with H2S.

O/T: silly chemists' joke:

Two amphoterics at the bar.

Says the first one: hey, you're putting them down fast, watch out or you'll get acid burn!

Number two: you're right, I'll slow down because I've got a date later on and I'm hoping to touch first base!



[Edited on 29-10-2010 by blogfast25]

Quote: Originally posted by smuv

Methylene blue: E0 = 0.53v @ pH = 0. Really odd your mug doesn't have copper in it. Does the metal seem rather soft?

Quote: Originally posted by blogfast25

This strongly points to something in the pewter being dissolved ate the end of the dissolution process and forming a green compound. But what?

They are but do they do any of the following?

Today I made some more K2SnCl6, this time using HCl that is essentially Fe free (previously I thought the strange colours might be due to it).

Well, here’s the pewter, nearly completely dissolved in the HCl:



The solution is a strange amethysty colour…

After adding the nitric and some simmering to mop up any residue and oxidise the Sn (II) to Sn (IV) the solution clears up completely and turns green. The intensity of the green is very similar to that obtained with the Fe-containing HCl:



So it would appear that some moiety is blueish-purple prior to oxidation then turns green upon oxidation… It could be a decent redox indicator if I knew what it was!

Here’s the end product: first crop of K2SnCl6 crystals obtained simply by allowing the solution, after addition of KCl, to cool down:



The second crop was obtained by reducing the supernatant liquor from 4 cups to 3 cups, it was much larger.

K2SnCl6 is more soluble in water and strong HCl than I anticipated: adding cold, saturated KCl solution to a cold, previously oxidised pewter solution (contains SnCl4, or SnCl6 (2-)) yields no precipitate at all. Only further reducing this solution causes potassium hexachlorostannate to crystallise out.

The salt also dissolves in water, into a clear solution that sheds Sn(OH)4 on heating and on adding NH3 solution, both precipitates dissolve back into HCl…

Sn content of homemade (NH4)2SnCl6

250 ml of KMnO4 0.1 N were prepared and standardised against a primary standard of fresh 0.1 N oxalic acid according to the following procedure (*.pdf):

http://goes.flexinet.com.au/download/chem12/exp/redoxEXP5.pd...

About 5 g of homemade (NH4)2SnCl6, already quite dry from drying with CaCl2, were dried in an electric oven at 70 C, until weight was constant. During drying a faint smell of NH4Cl could be observed. Excess NH4Cl sublimating or (NH4)2SnCl6 decomposing?

4.63 g of the dried product was then dissolved in 100 ml of 22 % HCl in a 500 ml conical flask, with cautious warming to help dissolution. 1.1 g of Al was then added to reduce the Sn (IV) to Sn2+, the solution then simmered to flush out any residual hydrogen and cooled. It was then quantitatively transferred into a 250.0 ml measuring flask and diluted to the mark with 1 M HCl.

This solution (20.0 ml) was then titrated thrice with the standardised KMnO4. The three titrations were in excellent agreement. The Sn content of the 4.63 g sample was then calculated to be 32.2 w%. Theoretically the Sn content of pure (NH4)2SnCl6 is 34.0 w%. So quite good agreement considering the homemade product was made from technical grade ingredients (including quite horrible homemade NH4Cl) and hadn’t even been recrystallised once.

But there was also a strange ‘incident’. When I added the crystals of the ammonium hexachlorostannate to the conical flask, a strange ping discolouration beset much of the crystals:



The Erlenmeyer is a used one but clean and in good condition, the last liquid in it was DIW. In Dutch, ammonium hexachlorostannate is known as 'pink salt' although it isn't clear whether this is a reference to colour...

On adding the 100 ml of 22 % HCl and dissolving the pink immediately disappeared.

During reaction of the Al with the hot acid:



I can’t help but thinking that the blackish residue that temporarily forms may be elemental tin (plated out), which then redissolves in the hot HCl to SnCl2. On the older tin salts thread someone claims the same. For all my experiments with tin I do have a stainless steel teapoon plated with tin metal, no kidding...

After the reduction with Al/H2:



The solution turns purple (it’s also ever so slightly turbid), reminiscent of the colour of pewter solutions prior to oxidation…

The K2SnCl6 will be assayed for Sn in the same way…


[Edited on 13-11-2010 by blogfast25]

Some tests with allegedly pure tin (99.9 %). About 2.5 g about to face the acid:



During dissolution – the amethysty colour shows up. Is this coming from the acid or from the tin?



At the end of the dissolution, nitric acid was added to oxidise Sn (II) to Sn (IV): the colour changes to yellow. This is taken after ammonium chloride was added and during simmering down: