Sciencemadness Discussion Board

The trouble with neodymium...

 Pages:  1    3  ..  6

blogfast25 - 11-7-2010 at 09:38

Wanting to make some small amount of a water soluble neodymium salt from a neodymium (Nd2Fe14B) magnet according to the 'German method' (http://www.versuchschemie.de/topic,9178,0,-Neodym%28III%29-s...) I ran into unforeseen and rather strange problems.

The magnet, 45.4 g, was first thermally demagnetised, stripped (partially) of it's Ni cover, broken up into a few chunks and attempted to dissolve into 100 mL water + 100 mL 50 % H2SO4 (these are very much the ratios used by the versuchschemie experimenter 'Bariumnitrat', adjusted for magnet quantity). I need to specify that my 50 % H2SO4 is in fact toilet cleaner and is slightly contaminated with Fe3+, if not other things.

I met with problems dissolving the magnet which I may elaborate on later on. Having a short fuse patience wise I switched to 32 % HCl as a solvent and used about 350 mL to achieve full dissolution of the magnet although much of the Ni cover was found back undissolved and removed from the solution, now a greenish colour mainly from Fe2+.

I wanted to exploit the inverted solubility/temperature curve of Nd2(SO4)3 like the 'Bariunnitrat' did, so I needed to convert the chlorides to sulphates. I guess I could have tried to saturate the solution with Na2SO4 or (NH4)2SO4 and hoped the Nd sulphate would precipitate out at higher temperature but I didn't fancy the multitude of spectator cations in a saturated solution.

So I went all the way by neutralising a diluted solution of the chlorides with strong NaOH, filtering and washing the precipitate and redissolving it in H2SO4. By now, due to exposure to air oxygen much of the Fe had gone to Fe [+III] and the precipitate was a mixture of ferric hydroxide, hematite and Nd(OH)3.n H2O. A little water was added to the precipitate slurry to improve stirrability and then gradually 100 mL of 50 % H2SO4 was added, under constant stirring. Much heat was generated and most of the precipitate dissolved into a yellow solution with an off white precipitate forming. By then the solution was about 95 C, so the precipitate (which I need to add wasn't lavender coloured at all) seemed consistent with Nd2(SO4)2 having dropped out at high temperature. The stuff settled really well and was washed by means of decanting off boiling water several times and then dried at 130 C for about 2 hours. About 15 g was the weight. A light salmon pink, dry, sandy solid was obtained, still not lavender at all.

And it isn't Nd2(SO4)3 at all, I believe. For one it's neither soluble in hot or cold water. A hydroxy sulphate? Nd(OH)SO4? (Nd(OH)2)2SO4? Something like that?

In 32 % HCl there was some solubility of this product but not much, resulting in the supernatant liquid to take on a greenish-yellow, the colour of NdCl(H2O)7 (2+) perhaps? This liquid was decanted off and set aside. Testing a small amount by adding strong NaOH caused a lavenderish gel like precipitate to form: Nd(OH)3.n H2O?

The insoluble part was then (after washing) subjected to 100 mL of 50 % H2SO4 into which it dissolved effortlessly and spontaneously. But on adding water for dilution a white precipitate formed again, leaving a clear supernatant liquid! This suspension was cooled down to about 7 C but id didn't dissolve (as you'd expect Nd2(SO4)3 to do).

Washed again, another 100 mL H2SO4 was added, this time the precipitate only dissolved with stirring and heating: by about 80 C all had dissolved and the solution was lavender coloured again. This time I cooled it down to about 5 C and added water in the form of ice (the temperature never exceeded 17 C): but to no avail; a white precipitate formed in a clear supernatant liquid. Further cooling did not dissolve it. The low temperature at which the precipitation occurred indicates it's really hydrolysis that's causing the problem.

That's how far I've gotten now.

I'm certainly using far more H2SO4 than is strictly necessary: I believe the amount of Nd in the precipitate to be about 0.08 mol, tops. That would requite about 0.15 mol H2SO4 to convert to Nd2(SO4)3, or 15 g H2SO4, or 30 mL 50 % H2SO4. So, I'll try and see what happens using far less H2SO4, perhaps starting with 20 m 50 % H2SO4 + 20 mL water.

Alternatively I could try HNO3 of which I have 30something %.

Anyone has any ideas as to what is the white, clearly sulphate and Nd based compound might be?

[Edited on 11-7-2010 by blogfast25]

The WiZard is In - 11-7-2010 at 09:51

Quote: Originally posted by blogfast25  
Wanting to make some small amount of a water soluble neodymium salt from a neodymium (Nd2Fe14B) magnet according to the 'German method' (http://www.versuchschemie.de/topic,9178,0,-Neodym%28III%29-s...) I ran into unforeseen and rather strange problems.


[snip]


Some neodymium magnets are bonded (polymer bonded)
... you are going to have to remove the bonding....

[Edited on 11-7-2010 by The WiZard is In]

[Edited on 11-7-2010 by The WiZard is In]

[Edited on 11-7-2010 by The WiZard is In]

blogfast25 - 11-7-2010 at 12:00

That might explain the initial problems I had with dissolving the magnet into 25 % H2SO4: hot H2SO4 is still strong enough to carbonise polymeric materials. Twice the reaction started briskly, then halted abruptly after about 5 mins. Only a few gram had dissolved. That's why I switched to HCl.

But it doesn't explain why I'm having trouble keeping a Nd-sulphate based compound in solution...

IrC - 11-7-2010 at 14:05

What is the boron becoming in this process? Curious as I did not see any mention of it or its reactions in this experiment whether the acid used was sulfuric or hydrochloric.

JohnWW - 11-7-2010 at 20:35

The boron present will not react with any non-oxidizing acid. Under extreme conditions it may be oxidized by HNO3, though.

12AX7 - 12-7-2010 at 00:01

Seems to me I've observed a tan colored precipitate around Fe(III) sulfate. Insoluble, dissolves slowly in H2SO4. It seems to be a compound of Fe (what else would it be?), but it's not nearly as yellow as most Fe(III) compounds are.

Tim

kmno4 - 12-7-2010 at 02:04

To check if your magnet contain only Nd, not Nd+Pr mixture you should take a sample and precipitate lanthanides as oxalates. All Fe(II) should be converted to Fe(III) first.
Recrystalize it from hot 50% HNO3 and burn to oxide.
If oxide is white blue ( may be a little grey) - you have only Nd in your magnet. If oxide is brown, you have Nd+Pr.
Another test: pure Nd oxalte is bright violet under tungsten lamp and white under fluorescent lamp. Pr contaminated oxalate is also bright violet under tungsten lamp but bright green under fluorescent lamp.
Unfortunately, all new magnets I have now seems to be made from Nd-Pr mixture :(

blogfast25 - 12-7-2010 at 04:58

Quote: Originally posted by 12AX7  
Seems to me I've observed a tan colored precipitate around Fe(III) sulfate. Insoluble, dissolves slowly in H2SO4. It seems to be a compound of Fe (what else would it be?), but it's not nearly as yellow as most Fe(III) compounds are.

Tim


I believe that when I redissolved the hydroxides, the separation between Nd and Fe was complete. It's the only way to explain why when I do get the Nd based precipitate to dissolve it dissolves into a neat lavender solution, presumably Nd2(SO4)3 only.

Quote: Originally posted by kmno4  
To check if your magnet contain only Nd, not Nd+Pr mixture you should take a sample and precipitate lanthanides as oxalates. All Fe(II) should be converted to Fe(III) first.
Recrystalize it from hot 50% HNO3 and burn to oxide.
If oxide is white blue ( may be a little grey) - you have only Nd in your magnet. If oxide is brown, you have Nd+Pr.
Another test: pure Nd oxalte is bright violet under tungsten lamp and white under fluorescent lamp. Pr contaminated oxalate is also bright violet under tungsten lamp but bright green under fluorescent lamp.
Unfortunately, all new magnets I have now seems to be made from Nd-Pr mixture :(


Believe me kmno4, if I had oxalic acid I'd have used the oxalate separation route, alas I have none.

I don't believe this magnet contained any appreciable amount of Pr: see the precipitate obtained from the chloride (that bit of the mysterious precipitate that did dissolve in strong HCl); that hydroxide was whitish-lavender, not brown.

Do you have any evidence for your last assertion?

&*((*&**(*&*((*&*(*&*((

Because the latest 'version' of the sulphate based precipitate was obtained in cold conditions I've a gut feeling it might be more soluble. Today I'll try dissolving it in a smaller quantity of about 25 % H2SO4. My instinct tells me too high sulphate ion concentration is somehow causing the problem.

Polverone - 12-7-2010 at 08:35

Is praseodymium a useful additive in NdFeB magnet compositions? I did a cursory web search and didn't turn up anything about that. Is it just left in because it's more trouble than it's worth to separate the praseodymium?

IrC - 12-7-2010 at 09:06

Quote: Originally posted by JohnWW  
The boron present will not react with any non-oxidizing acid. Under extreme conditions it may be oxidized by HNO3, though.


So it precipitates? What if you add H2O2 to the acid?

blogfast25 - 12-7-2010 at 11:54

Treatment with less concentrated H2SO4 and later with 38 % HNO3 yielded nothing.

It must be said though that the product is scarcely soluble but not completely insoluble: every time I treat some of the supernatant liquid with strong alkali a small amount of whitish precipitate forms.

I'm beginning to think that rather than a basic sulphate it's something akin to a bisulphate that may be causing the problem here.

Some 5 M NaOH was poured on the product which was then very briefly simmered: the precipitate became notably 'fluffier'. The supernatant liquid was decanted off and 100 mL 32 w% HCl dissolved nearly all of it. But upon boiling the precipitate showed up again...

Because of the poor solubility of the product the approx. displacement reaction:

Ndsomething + n NaOH ---> Nd(OH)3 + n Nasomething

... should be possible because of the insolubility of the hydroxide.

The precipitate was boiled up for 15 mins with 5 M NaOH and is now standing overnight. If I'm right the washed precipitate should dissolve irreversibly in strong HCl.

IrC - 12-7-2010 at 14:02

I know some bonding methods use epoxy type materials. I wonder if you would do better if you grind the magnets to powder first. Smashing them with a hammer first, then grinding to powder. Seems like a small ball mill with the right choice of grinding material would turn it into powder.



12AX7 - 12-7-2010 at 14:35

Or fire.

Fire, fire fire, Fire!...

Tim

kmno4 - 13-7-2010 at 00:34

Quote: Originally posted by blogfast25  


I don't believe this magnet contained any appreciable amount of Pr: see the precipitate obtained from the chloride (that bit of the mysterious precipitate that did dissolve in strong HCl); that hydroxide was whitish-lavender, not brown.

Do you have any evidence for your last assertion?


I also did not believe in Pr contamination.
You may wish to see a picture (in attachment) of my "Nd2(SO4)3x*8H2O" in this thread :
http://www.sciencemadness.org/talk/viewthread.php?tid=12934
Oxide prepared (rather mixture with carbonates + oxides) from this sulfate looks like cacao or powdered chocolate :(
Evidence ?
Brown oxide and green colour of solution (in fluorescent light).
Brown colour is very strong evidence of Pr contamination according to literature.
Of course, precipitated hydroxides are of lavenda colour, but very light green under fluorescent lamp.
If you do not see this colours - you are lucky and have magnet without Pr.

If monazite ore is used for production of Ln's, it gives mixture Nd : Pr (as one of products) in about 4 : 1 (m/m) ratio
Smart people , probably from China noticed that it is not worth to lose time (weeks and months in extractive methods) to separate Nd - Pr because Nd can be partially replaced with Pr in magnets.
Pair Nd-Pr is not easy to separate. In the past this mixture were called "didymium"
http://en.wikipedia.org/wiki/Didymium
... and from there:
In the late 1920s, Leo Moser recombined praseodymium and neodymium in a 1:1 ratio to create his "Heliolite" glass, which has color-changing properties between amber, reddish, and green depending on the light source.

JohnWW - 13-7-2010 at 01:29

In the Nd2Fe14B magnets, surplus 4s and 3d electrons from the Fe enter the vacant 4f orbitals in Nd (which has only 3 4f electrons of its own), to provide the maximum number possible of unpaired electrons in 3d and 4f orbitals. Pr and Ce could also do this job, but less effectively due to their lesser nuclear charges and having fewer 4f electrons of their own. Because Nd is an even-numbered element, it is likely to be more abundant than Pr.

Uranium-238 (depleted U), which is the homolog of neodymium, and of which there are huge stockpiles sitting around in storage as the byproduct of uranium enrichment (in fissionable U-235) just waiting for uses to be found, could theoretically serve as a substitute for Nd. However, this has not yet been explored by magnet makers for hard drives. Although U has the same numbers of 5f electrons and vacant 5f orbitals as Nd has 4f electrons and orbitals, because U is more electropositive than Nd (in fact able to form hexavalent compounds) the surplus 4s and 3d electrons from Fe may not enter its 5f orbitals to the same extent.

Pr may possibly be chemically separated somehow from Nd through its forming, with strong oxidation, the tetravalent compounds Pr6O11 and PrF4, not formed by Nd, but there are no water-soluble Pr(IV) compounds, unlike Ce(IV).

blogfast25 - 13-7-2010 at 05:35

@ kmno4:

From Wiki:

"In its compounds, praseodymium occurs in oxidation states +2, +3 and/or +4. Praseodymium(IV) is a strong oxidant, instantly oxidizing water to elemental oxygen (O2), or hydrochloric acid to elemental chlorine. Thus, in aqueous solution, only the +3 oxidation state is encountered. Praseodymium(III) salts are yellow-green and, in solution, present a fairly simple absorption spectrum in the visible region, with a band in the yellow-orange at 589-590 nm (which coincides with the sodium emission doublet), and three bands in the blue/violet region, at 444, 468, and 482 nm, approximately. These positions vary slightly with the counter-ion. Praseodymium oxide, as obtained by the ignition of salts such as the oxalate or carbonate in air, is essentially black in color (with a hint of brown or green) and contains +3 and +4 praseodymium in a somewhat variable ratio, depending upon the conditions of formation. Its formula is conventionally rendered as Pr6O11."

As it so happens, the first time I treated the mysterious precipitate with 32 % HCl (by adding 100 mL and boiling in, then repeating this another 2 times) I obtained a yellow-green solution. I'm sure it contains no Fe3+: it precipitates from NaOH as a grayish hydroxide, so that could be Pr3+. I will try and convert it to carbonate and fire it. Is it possible that inadvertently this procedure leached out the Pr? Subsequent leachings with HCl didn't yield a yellow green solution at all.

Now I'm going to check if my displacement reaction yielded a soluble (in HCl) product...

blogfast25 - 13-7-2010 at 12:54

The displacement product dissolved effortlessly in strong HCl, after decanting off the alkaline supernatant liquid. I now have three solutions of NdCl3 in HCl to neutralise, to recover reasonably pure NdCl3.6 H2O.

But right now I'm more interested in the green-yellow solution obtained by leaching the unusual precipitate with strong HCl. Here it is before any treatment. The photo doesn't really do the real colour much justice:



The same solution after dilution:



And after precipitating the hydroxide(s) with 5 M NaOH. The reddish-amberish colour is due to Fe3+ as independently verified with KSCN:



After filtering, washing, and redissolving into 50 mL 32 % HCl, this solution was obtained (the darkish patches are soot on the outside of the Pyrex receptacle):



This solution was then reduced in volume, first in direct flame, then au bain marie. This is what left. Despite the rubbish reflection, patches of green can be seen, next to red/amber. This cannot be Fe2+, or Nd3+ but PrCl3 would fit the bill, as kmno4 suggests.



Unfortunately, in spite of careful drying the FeCl3 had completely hydrolysed. I had hoped to extract it with dry acetone (acc. a contributor on this forum the RE chlorides are insoluble in acetone) but because of the hydrolysis that wasn't possible. Tomorrow I hope to selectively dissolve all the Pr(?) but not the Fe2O3...

blogfast25 - 15-7-2010 at 06:00

Due to beaker breakage the investigation into the presence of Pr had to be abandoned (evidence lost!). Shame.

Below is another pic of the suspected mix of FeCl3 and PrCl3, with prominent green patches between 10 and 11 o'clock:





Nd(OH)3.n H2O suspension about to be filtered, nice lavender colour:





Nd(OH)3.n H2O on the filter:





NdCl3.6 H2O (photo distorts colour), after dissolving Nd(OH)3 in HCl, crystallizing, washing with acetone and drying. Predictably perhaps, when viewed under a 'saver bulb' it takes on a decidedly greenish hue!



condennnsa - 9-7-2011 at 12:17

I dissolved a NdFeB magnet in HCl, and filtered the solution free of boron. What would be the best way to isolate the neodymium from FeCl3? I do not have oxalic acid. thanks

blogfast25 - 9-7-2011 at 13:01

Precipitate the iron and neodymium as their hydroxides and wash carefully. Use sulphuric acid to separate the two, based on the sulphate method outlined high up in this thread?

condennnsa - 10-7-2011 at 00:17

thanks, my problem is that i don't fully understand your process further up...

I precipitated the hydroxides with NaOH and filtered the mud. It is dark green and forms a brown layer on the surface.

If I treat this with sulfuric acid, won't the neodymium also go in solution? Or is it just that Nd sulfate is so insoluble that most of it will precipitate?

I figured after I precipitated the hydroxides, that I could've just used the chloride solution. Wiki says in the NdCl3 article that its melting point is 758C without decomposition.
So I thought I could've just boiled down the chloride solution, and keep heating until all the fecl3 decomposes to insoluble iron (hydr)oxides and HCl, and then a wash and filtering should give me a pretty pure NdCl3 solution, right?

However the same article says that the hydrate does decompose NdCl3 + H2O → NdOCl + 2 HCl what is the truth? Is this NdOCl insoluble?

unionised - 10-7-2011 at 01:06

The classic separation of iron is to oxidise it to Fe+++ then add lots of HCl and extract it with ether.
That's only useful f you have ether, but if you do it's the easy way to do it.

The other simple way is to add a little NaOH slowly to the solution of the chlorides. The Fe+++ will ppt first and leave the Nd in solution.

blogfast25 - 10-7-2011 at 04:33

Quote: Originally posted by condennnsa  
thanks, my problem is that i don't fully understand your process further up...

I precipitated the hydroxides with NaOH and filtered the mud. It is dark green and forms a brown layer on the surface.

If I treat this with sulfuric acid, won't the neodymium also go in solution? Or is it just that Nd sulfate is so insoluble that most of it will precipitate?

I figured after I precipitated the hydroxides, that I could've just used the chloride solution. Wiki says in the NdCl3 article that its melting point is 758C without decomposition.
So I thought I could've just boiled down the chloride solution, and keep heating until all the fecl3 decomposes to insoluble iron (hydr)oxides and HCl, and then a wash and filtering should give me a pretty pure NdCl3 solution, right?

However the same article says that the hydrate does decompose NdCl3 + H2O → NdOCl + 2 HCl what is the truth? Is this NdOCl insoluble?


By treating the Nd(OH)3/iron hydroxides mix with H2SO4, the iron should go into solution, the Nd stays behind as a sulfato complex (apparently)

So I thought I could've just boiled down the chloride solution, and keep heating until all the fecl3 decomposes to insoluble iron (hydr)oxides and HCl, and then a wash and filtering should give me a pretty pure NdCl3 solution, right?

I wouldn’t do that: I think you’ll end up with a goo. Also, FeCl3 is quite resistant to hydrolysis, as long as there’s acid left.

Apart from ether, FeCl3 is also very soluble in acetone. But these extractions aren’t as easy as that when there’s LOTS of iron present. You need to grind your chloride mix quite finely too. I see solvent extraction more as a last purification, to get the last minute bits of iron out of your product.

condennnsa - 12-7-2011 at 01:44

Quote: Originally posted by unionised  
The other simple way is to add a little NaOH slowly to the solution of the chlorides. The Fe+++ will ppt first and leave the Nd in solution.


How can you tell? Is neodymium more reactive than iron? Sorry for my noobness ...

blogfast25 - 12-7-2011 at 03:04

Quote: Originally posted by condennnsa  
How can you tell? Is neodymium more reactive than iron? Sorry for my noobness ...


The solubilities tell us that: in the given conditions Nd precipitates, Fe3+/Fe2+ doesn't. A very common separation technique. If you go to the top of the thread, where I link to the German experimenter and google translate that page, things might become clearer (no pun intended)...

condennnsa - 12-7-2011 at 09:32

Blogfast you say that Nd precipitates first, and Fe doesn't? unionised claimed the other way around...

unionised - 12-7-2011 at 11:10

Fe+++ will ppt first, then Nd+++ then Fe++ so it depends on whether or not you oxidise the stuff.
I'd go with oxidising it with peroxide or some such. That way you can take out the last of the Fe with ether. Also, the Fe++ will oxidise in air so you might as well go along with that.

blogfast25 - 12-7-2011 at 11:41

Quote: Originally posted by unionised  
Fe+++ will ppt first, then Nd+++ then Fe++ so it depends on whether or not you oxidise the stuff.
I'd go with oxidising it with peroxide or some such. That way you can take out the last of the Fe with ether. Also, the Fe++ will oxidise in air so you might as well go along with that.


Unionised, what makes you think Fe(III) falls first? Fe (III) sulphate is very water soluble. And we are talking about a sulphate based separation method (based on the poor cold solubility of Nd sulphate), not a hydroxide based method...

As regards washing with ether, FeCl3 is highly soluble in acetone too: a lot cheaper than ether!

[Edited on 12-7-2011 by blogfast25]

condennnsa - 12-7-2011 at 14:47

When I originally asked unionised which falls first I was referring to adding NaOH to the solution of chlorides...
I thought that if what he said was indeed true, it would be very easy to obtain very pure neodymium hydroxide.

So which precipitates first? and why?

Megamarko94 - 12-7-2011 at 14:55

is NdCl3 soluble in methanol and acetone



[Edited on 12-7-2011 by Megamarko94]

blogfast25 - 13-7-2011 at 03:45

Condennsa:

I doubt very much if selective precipitation of the hydroxides of Fe (III), Nd and Fe (II) is really practical. Unionised is probably going by solubility constants (Ksp) of these substances but I doubt if they are substantially different enough to allow selective precipitation. In any case selective precipitation of hydroxides requires tight control of solution pH.

Here's the deal: you look up the Ksp values for Fe(OH)3, Nd(OH)3 and Fe(OH)2 and I'll do the calculation for ya. Howzat?

With sulphates, you KNOW a cold water insoluble Nd sulfato complex separates out.

Quote: Originally posted by Megamarko94  
is NdCl3 soluble in methanol and acetone

[Edited on 12-7-2011 by Megamarko94]


Oooopsie.... :(

Edit: I believe Fe3+ and Fe2+ would precipitate together as hydrated Magnetite (Fe3O4), see also preparation of ferrofluids.

The K<sub>sp</sub> = [Fe3+]x[OH-]<sup>3</sup> of Fe(OH)3 is 2.8 x 10<sup>-39</sup>, very low indeed. The precipitation behaviour of Fe3+ in alkaline conditions may also be more complex due to various hydroxy species. And I can't find a value for Nd(OH)3...

[Edited on 13-7-2011 by blogfast25]

MrHomeScientist - 13-7-2011 at 12:50

blogfast25,

Farther up the thread you mentioned an oxalate separation route - what does that entail? I know iron oxalate is insoluble, as I've prepared it before, but is Nd oxalate soluble? Is that how you separate the two?

I've got some oxalic acid (boat wood cleaning product), and I'd like to try this procedure to ultimately arrive at Nd metal for my element collection.

blogfast25 - 14-7-2011 at 04:43

Quote: Originally posted by MrHomeScientist  
blogfast25,

Farther up the thread you mentioned an oxalate separation route - what does that entail? I know iron oxalate is insoluble, as I've prepared it before, but is Nd oxalate soluble? Is that how you separate the two?

I've got some oxalic acid (boat wood cleaning product), and I'd like to try this procedure to ultimately arrive at Nd metal for my element collection.


I've never tried this method but it's in fact neodymium oxalate that’s insoluble, both ferrous and ferric oxalates being soluble. So the separation is based on that. Neodymium oxalate can then be calcined to Nd2O3 and redissolved into a strong acid.

Oxalic acid is very easy to purify by means of recrystallisation. Get rid of the .2H2O by heating to about 100 C.

MrHomeScientist - 14-7-2011 at 05:19

Quote: Originally posted by blogfast25  

I've never tried this method but it's in fact neodymium oxalate that’s insoluble, both ferrous and ferric oxalates being soluble. So the separation is based on that. Neodymium oxalate can then be calcined to Nd2O3 and redissolved into a strong acid.

Oxalic acid is very easy to purify by means of recrystallisation. Get rid of the .2H2O by heating to about 100 C.

Really? I could have sworn iron oxalate precipitated when I did it. Ah well, that was over a year ago and I didn't keep notes back then, so I'm sure I've forgotten much of the procedure.

I'll definitely give this a try, and report back with what I find.
The (rough) plan is:
- HD magnets + H<sub>2</sub>SO<sub>4</sub> to get sulfates
- filter off the boron
- sulfates + NaOH to precipitate hydroxides
- hydroxides + oxalic to precipitate & separate Nd oxalate
- heat the oxalate to form Nd<sub>2</sub>O<sub>3</sub>
- use this in a thermite-like reaction with Mg powder to finally end in Nd metal (or convert to NdCl<sub>3</sub> with acid, then to NdF<sub>3</sub> with NaF, and thermite that, as you suggested in another thread)

Hm, seems like there should be an easier route. I'll think about it.

[Edited on 7-14-2011 by MrHomeScientist]

blogfast25 - 14-7-2011 at 07:17

Quote: Originally posted by MrHomeScientist  

I'll definitely give this a try, and report back with what I find.
The (rough) plan is:
- HD magnets + H<sub>2</sub>SO<sub>4</sub> to get sulfates
- filter off the boron
- sulfates + NaOH to precipitate hydroxides
- hydroxides + oxalic to precipitate & separate Nd oxalate
- heat the oxalate to form Nd<sub>2</sub>O<sub>3</sub>
- use this in a thermite-like reaction with Mg powder to finally end in Nd metal (or convert to NdCl<sub>3</sub> with acid, then to NdF<sub>3</sub> with NaF, and thermite that, as you suggested in another thread)

Hm, seems like there should be an easier route. I'll think about it.

[Edited on 7-14-2011 by MrHomeScientist]


Couple of comments:

Sulphates should already separate Nd (insoluble sulfato complex) from iron.

Magnesiothermy on Nd2O3 is unlikely to yield enough energy to obtain a molten mass of Nd metal + liquid MgO: you end up with a useless sintered mass of Nd metal in MgO sinter. Al cannot reduce Nd2O3 (ΔG > 0).

Only reduction of NdF3 with Mg is the best option (or electrolysis of a molten Nd salt).

NaF is very poorly water soluble: better to use NH4HF2 (CAUTION! Highly soluble fluorides are dangerous!)

MrHomeScientist - 14-7-2011 at 08:08

Quote: Originally posted by blogfast25  

Couple of comments:

Sulphates should already separate Nd (insoluble sulfato complex) from iron.

Magnesiothermy on Nd2O3 is unlikely to yield enough energy to obtain a molten mass of Nd metal + liquid MgO: you end up with a useless sintered mass of Nd metal in MgO sinter. Al cannot reduce Nd2O3 (ΔG > 0).

Only reduction of NdF3 with Mg is the best option (or electrolysis of a molten Nd salt).

NaF is very poorly water soluble: better to use NH4HF2 (CAUTION! Highly soluble fluorides are dangerous!)

Hrm. If that's the case, then what's the benefit of using the oxalate separation vs sulfuric acid? Either way you'd have to get rid of the elemental boron somehow.

Thanks a lot for the tips on the thermite. I wanted to use NaF because it's cheaper, but I see now you're right about solubility. I'll take maximum precautions when using the ammonium salt. Would you recommend using NH<sub>4</sub>HF<sub>2</sub> or NH<sub>4</sub>F?

[Edited on 7-14-2011 by MrHomeScientist]

blogfast25 - 14-7-2011 at 13:08

Quote: Originally posted by MrHomeScientist  
Hrm. If that's the case, then what's the benefit of using the oxalate separation vs sulfuric acid? Either way you'd have to get rid of the elemental boron somehow.

Thanks a lot for the tips on the thermite. I wanted to use NaF because it's cheaper, but I see now you're right about solubility. I'll take maximum precautions when using the ammonium salt. Would you recommend using NH<sub>4</sub>HF<sub>2</sub> or NH<sub>4</sub>F?

[Edited on 7-14-2011 by MrHomeScientist]


NH4HF2 packs more F<sup>-</sup> per mol. It's probably proportionally and arguably more dangerous than NH4F. Still, they're usable, just take good precautions and keep yer wits about at all times. Have a back up plan in case of emergency... And use plasticware, not glass!

[Edited on 14-7-2011 by blogfast25]

kmno4 - 14-7-2011 at 14:34

Quote: Originally posted by blogfast25  

I've never tried this method but it's in fact neodymium oxalate that’s insoluble, both ferrous and ferric oxalates being soluble.

As I mentioned, Fe must be in Fe(III) form. In another case, Fe(II) oxalate will precipitate. You can try this with FeSO4 and (COOH)2 - reaction takes some time, but soon you will get Fe(COO)2 sediment.
Nd oxalate can be recrystallized from hot 50% HNO3. This is quite unusuall, and shows how stable is Nd oxalate.

blogfast25 - 15-7-2011 at 09:50

Thanks kmno4, very interesting...

MrHomeScientist - 21-7-2011 at 16:07

Well I did this experiment as far as getting neodymium sulfate, and here are my observations. Unfortunately, I didn't take any pictures until the end.

I started by demagnetizing a single hard drive magnet by directly heating it in a propane torch flame. After this I broke it into chunks and peeled off as much of the Ni coating as I could. These pieces weighed 9.4g.

I reacted this with 40mL of 4.4M battery-grade sulfuric acid. It immediately started reacting to form the sulfates via:
Nd<sub>2</sub>Fe<sub>14</sub>B + 17H<sub>2</sub>SO<sub>4</sub> == 14FeSO<sub>4</sub> + Nd<sub>2</sub>(SO<sub>4</sub>;)<sub>3</sub> + B + 17H<sub>2</sub>
(it turns out that boron isn't as unreactive as I thought, which I'll get to later)

After the reaction, the solution was very dark with a whitish crystalline precipitate. I diluted this down to 150mL to get everything into solution. Then I went through a number of heating and cooling cycles to exploit neodymium sulfates inverse solubility behavior (less soluble in hot soln.) and separate it from the other products. First I cooled it in an ice bath to filter off any unreacted B, leftover Ni coating, and any FeSO<sub>4</sub> that happened to precipitate. Then, I heated the solution to near boiling. When it had evaporated down to ~60mL, pink crystals started forming. These were filtered off while still hot.

After the solution cooled, a lot of iron sulfate had crystallized into this really neat looking little mountain of crystals.
iron sulfate.jpg - 123kB

After a few more cycles of heating and filtering the neodymium sulfate to get it out of solution and then finally recrystallize and purify it, I ended up with 1.4g total product. As reported before, it looks different under different lighting which is pretty cool to see. The first picture is in fluorescent tube lighting, and the second is outside in daylight.
Nd2SO4 under FL tubes.jpg - 94kB

Nd2SO4 under daylight.jpg - 109kB

The bottom vial contains 0.2g of what appeared to be somewhat more pure material. Overall, yield was pretty meager considering stoichiometrically I should have gotten 5g (based on that equation above).

While I was collecting the FeSO<sub>4</sub> crystals, I noticed some thin, clear , flaky crystals clinging to them. These looked pretty different, so I recovered maybe 50mg of them. I suspect these are boric acid, so I dissolved them in methanol and set it on fire to look for the characteristic green of trimethyl borate. It didn't show up that well, but you can see it definitely has a green tint to it.
greenish flame.jpg - 72kB

That strongly suggests it's boric acid, which I didn't think would have been formed. Checking around, I did find that the wiki page for boron states "When finely divided, it [boron] is attacked slowly by ... hot sulfuric acid..." so that seems to support my hypothesis. I'm not exactly sure what the products of B + H2SO4 are, but the flame test closely follows my experience with boric acid. At the very least we can say that some of the boron reacts into some soluble white compound.

I've got some ammonium bifluoride on order, and once that arrives I'll continue with the next step towards elemental neodymium metal. The goal is to make NdF<sub>3</sub>, then thermite that with magnesium dust to obtain the metal. Would I be able to react the bifluoride straight with Nd<sub>2</sub>SO<sub>4</sub> or do I need to convert it to the hydroxide or something? What would that reaction's equation be? I'm not too familiar with fluoride chemistry (I've seen it mentioned often that it's different than the other halides). I already asked this question in beginnings, but I hadn't gotten any replies yet. Thanks, and I hope this post was informative.

unionised - 22-7-2011 at 04:25

Quote: Originally posted by blogfast25  
Condennsa:

I doubt very much if selective precipitation of the hydroxides of Fe (III), Nd and Fe (II) is really practical. Unionised is probably going by solubility constants (Ksp) of these substances but I doubt if they are substantially different enough to allow selective precipitation. In any case selective precipitation of hydroxides requires tight control of solution pH.

Here's the deal: you look up the Ksp values for Fe(OH)3, Nd(OH)3 and Fe(OH)2 and I'll do the calculation for ya. Howzat?

With sulphates, you KNOW a cold water insoluble Nd sulfato complex separates out.

Quote: Originally posted by Megamarko94  
is NdCl3 soluble in methanol and acetone

[Edited on 12-7-2011 by Megamarko94]


Oooopsie.... :(

Edit: I believe Fe3+ and Fe2+ would precipitate together as hydrated Magnetite (Fe3O4), see also preparation of ferrofluids.

The K<sub>sp</sub> = [Fe3+]x[OH-]<sup>3</sup> of Fe(OH)3 is 2.8 x 10<sup>-39</sup>, very low indeed. The precipitation behaviour of Fe3+ in alkaline conditions may also be more complex due to various hydroxy species. And I can't find a value for Nd(OH)3...

[Edited on 13-7-2011 by blogfast25]


Actually, I was relying on the fact that I have done the experiment; it's quite easy to ppt the Fe+++ selectively and leave the Nd in solution. I found a value for the solubility product here
http://pubs.acs.org/doi/abs/10.1021/ja01547a011
where they say it's of the order of 10^-21
The difference in solubility product is about 18 orders of magnitude; I found that to be sufficient. The fact that one product is brown and the other is white also helps.

blogfast25 - 22-7-2011 at 05:24

MrHomeChemist:

Nice work.

Can you describe the heating/cooling cycles a little more?

Boron probably reacts with concentrated sulphuric acid, forming boric acid or boric oxide, depending on water being present. But with relatively weak sulphuric acid? I doubt that. The green does seem to suggest it but ideally you’d have to identify a determining spectral line to confirm the green is boron. In my case the boron was found as a brown amorphous powder.

For making the NdF3, I’d suggest adding the right amount of NH4HF2 to a solution of Nd<sup>3+</sup> salt.

But actually making such small magnesiothermic reactions work is difficult. I suggest you buy some Nd salt to take this further. Or repeat the procedure but looking after yield better…

Unionised:

I cannot read the value of Ks for Nd(OH)3 in the introduction of that paper but lower down it says the solubility of Nd(OH)3 is 2.7 x 10<sup>-4</sup> and that would suggest a Ks of 1.4 x 10<sup>-13</sup>. That would require a pH of between 9 and 10 for precipitation of Nd(OH)3 to start, way after Fe(OH)3 would drop out. As a separation method between Nd3+ and Fe3+ I’ve never heard of it before though…

How did you control pH? Because go a little too far and your Nd will fall. Buffer?


kmno4 - 22-7-2011 at 06:05

As it was mentioned in other posts - if Nd salt is not purely white (colourless) in fluorescent lamp, then it is not pure (Pr gives green colour). Also strong heating should give blue powder of Nd2O3. The same - if it is not blue, it not pure. My samples of magnets contain Nd+Pr.

MrHomeScientist - 22-7-2011 at 07:15

Quote: Originally posted by blogfast25  
MrHomeChemist:

Nice work.

Can you describe the heating/cooling cycles a little more?

Boron probably reacts with concentrated sulphuric acid, forming boric acid or boric oxide, depending on water being present. But with relatively weak sulphuric acid? I doubt that. The green does seem to suggest it but ideally you’d have to identify a determining spectral line to confirm the green is boron. In my case the boron was found as a brown amorphous powder.

For making the NdF3, I’d suggest adding the right amount of NH4HF2 to a solution of Nd<sup>3+</sup> salt.

But actually making such small magnesiothermic reactions work is difficult. I suggest you buy some Nd salt to take this further. Or repeat the procedure but looking after yield better…


Sure, always happy to provide more detail! After the pieces were done reacting with the sulfuric acid, the solution was quite dark and somewhat lavender (in outside lighting of course). I first chilled this in an ice bath to get the Nd into solution, and filtered off the remaining pieces of Ni coating and some boron powder. The boron was very fine and brown, as you described. After letting the solution sit overnight, nearly all of the 'darkness' had settled out of the solution, presumably very fine boron powder. There was only a tiny amount of this. I carefully decanted to get rid of the remaining B.
Next, I heated the solution to near boiling. Once it had reduced in volume to about 60mL, a pink powder precipitated. After it had went down to 50mL, I stopped heating and filtered this off while still hot. After the solution cooled, that mound of iron sulfate had crystallized.
Decanting the solution from that, I heated it again to near boiling. After another small reduction in volume, more pink precipitate formed. I hot filtered this again. I repeated this process about 4 times total to try to squeeze out as much Nd as possible, topping it up with some extra water when the volume got low and I was worried Fe would start dropping out too.
After all this, my product was heavily contaminated with FeSO<sub>4</sub> because I didn't think to wash the powder with hot water while it was on the filter. So, I once again immersed my product (about 5g) in about 50mL of water, heated, filtered, and rinsed with hot water. That's how I arrived at my final amount of 1.4g of much cleaner pink crystals.

As far as yield goes, I feel like I did quite a bit to get as much out as possible. I suppose some could be trapped in the FeSO<sub>4</sub> crystals, as I did notice some pink clinging to some of them. It might be worth another trip into solution.

As for the B, as I mentioned above I did get some in elemental form and also some in those white flaky crystals (at least, I suspect it's a B compound). I wonder if one of those cheapo CD spectrometers would be enough to tell me something about the composition? It could be worthwhile to build one. I did just finish a box of cereal this morning :)

Finally, concerning the final steps: If I do need Nd in solution, I'll plan on using this route: Nd<sub>2</sub>(SO<sub>4</sub>;)<sub>3</sub>(s) --> Nd(OH)<sub>3</sub>(s) --> NdCl<sub>3</sub>(aq) --> NdF<sub>3</sub>(s)
That's assuming the bifluoride will replace chloride, via something like
2NdCl<sub>3</sub> + 3NH<sub>4</sub>HF<sub>2</sub> == 2NdF<sub>3</sub> + 3NH<sub>4</sub>Cl + 3HCl
Don't quote me on that reaction, I don't know if it's really reasonable. I suppose I could just try to dissolve the sulfate in cold water and just use that instead.

I wonder, perhaps I should have just started with hydrochloric acid instead of sulfuric? So I could go straight to NdCl<sub>3</sub>. It might be more difficult to separate from iron chloride, though. In any case, I have a bunch of HD magnets donated by coworkers to the cause so I hope to get a decent amount of product from all that. Hopefully enough to make a 'thermite' large enough to work.

unionised - 22-7-2011 at 07:17

I didn't control the pH as such. I added NaOH soln slowly and stopped when the ppt stopped being brown.
I don't know if this image will work, but it's the Ksp data

sol prod.bmp - 64kB

[Edited on 22-7-11 by unionised]

blogfast25 - 22-7-2011 at 08:48

MrHomeChemist:

Thanks. You seem to have done everything right, despite the low yields.

Precipitating the NdF3 from cold sulphate should be fine. Use fairly dilute solutions to avoid occlusion (co-precipitation).

Unionised:

Something doesn't tally with their Ksp and solubility: a quoted solubility of Nd(OH)3 of 2.7 x 10<sup>-4</sup> mol/l works out as a Ksp of 1.4 x 10<sup>-13</sup>.

And a Ksp of 1.9 x 10<sup>-21</sup> means Nd(OH)3 would start precipitating (from a 1 M solution of Nd3+) from about pH = 7. That doesn't sound right to me: Nd salts aren't very prone to hydrolysis, IIRW...

I'm guessing your method worked and that a value of Ksp = 1.4 x 10<sup>-13</sup> is closer to reality...

A solubility of 2.7 x 10<sup>-4</sup> mol/l means [Nd] = 2.7 x 10<sup>-4</sup> and [OH-] = 3 x 2.7 x 10<sup>-4</sup> (at that concentration dissociation of water can be ignored), so Ksp = 27 x 2.7<sup>4</sup> x 10<sup>-16</sup> = 1.4 x 10<sup>-13</sup>.

MrHomeScientist - 22-7-2011 at 09:15

blogfast:
Thanks for the encouragement, tips, and for checking my procedure. I don't have much formal education in chemistry (my degree was in physics), so I'm always concerned I'm missing some crucial step or doing something incorrectly. Because of that I also don't have as much background in the theory as I'd like. So once again, thanks for the support.

This weekend I plan to dissolve the rest of my magnets and isolate more Nd2(SO4)3. I'll post here if anything interesting happens.

MrHomeScientist - 22-7-2011 at 10:31

I read kmno4's reply concerning Pr contamination, and did some searching to see what could be done about that. I found some really interesting procedures on google books in The metals of the rare earths, by James Frederick Spencer.
Double nitrate salts are what interested me, to separate didymium into its constituent Nd and Pr fractions. Different types of double nitrates are discussed: p.26 - ammonium, p.28 - magnesium, and p.31 - manganese.
Manganese nitrates sound like the best for our project here, but of course has the least information on it. I assume it to be similar to the magnesium nitrate procedure, which I'll quote here:

Quote:

ii) Double Magnesium Nitrates.—The crystallisation of the double magnesium nitrates of the rare earths was introduced by Drossbach227 for the separation of neodymium and praseodymium. The cerium earths form double nitrates with magnesium nitrate of the general formula 2M1n(N03)3,3Mg(N03)2,24H20. The material used by him for this separation was a didymium lanthanum fraction obtained by the oxychloride method (II. (/), p. 44). The mixture was converted into oxides which were then dissolved in nitric acid; the same quantity of nitric acid was treated with as much magnesia as it would dissolve, and the two solutions were mixed. This ensures the addition of exactly the correct amount of magnesium nitrate for the formation of the double salts. The mixed solution was then heated and small quantities of magnesia added until the solution was alkaline to congo paper. 0n diluting the solution, impurities such as iron, manganese, aluminium, silica, and phosphoric acid were precipitated by the excess of magnesia. The solution was then filtered and concentrated until practically the whole of the double nitrates had separated out, The crystals were removed from the mother liquor, placed in a dish and treated with boiling water until only a small portion remained undissolved. Cooling was then brought about byNfloating the dish in water at 30°-40°, and a small quantity of cold water was sprayed on to the surface of the solution. A quantity of large crystals separated on the bottom of the dish, and as soon as about one-tenth of the total earths had crystallised out, the liquor was poured from the crystals into a second dish which was floated on water at 10°-15° and cooled on the surface as before when a second crop of crystals was obtained. The mother liquor from the second crop of crystals was then treated at o° in the same way, and a third crop of crystals obtained. The mother liquor from the third crop of crystals was concentrated and the process repeated. The final mother liquor being rejected for this purpose was used for the isolation of erbium, samarium, yttrium, and gadolinium which it contained. The various crops of crystals were then recrystallised systematically when the following materials were obtained: neodymium containing traces of praseodymium, a little pure neodymium, and much lanthanum containing praseodymium. The double magnesium nitrates of the rare earths crystallise from water in the order228, 229 lanthanum, praseodymium, neodymium, excess magnesium nitrate, samarium, europium, erbium, gadolinium, and yttrium, and the remaining yttrium earths. By using this method, Feit230 was able to concentrate terbium from a mixture of monazite earths in the least soluble fraction, and James 225 uses it for the initial separation of the cerium earths.

Since I don't know the extent of Pr contamination in these magnets, and I'll be working with fairly small quantities, multiple recrystallizations like this might be overkill. Still, very interesting to read about.

blogfast25 - 23-7-2011 at 08:55

MrHomeScientist:

While these methods are interesting from a chemical point of view, they are difficult to carry out by hobbyists because most of us have no easy way to monitor progress of separation: that really requires decent spectroscopy (of one form or another).

Today the separation is mainly carried out by means of tailored ion exchange resins, by exploiting the lanthanide contraction, I believe…

For my own sample of NdCl3 I have some circumstantial evidence that Pr is present. Time allowing (one fine day!) I’ll push on with making more of this stuff from neomagnets (they are really cheap nowadays) and ignore the Pr impurity and try and make the metal by reduction of (Nd,Pr)F3 with Mg. But ample experience with thermites shows that reaction batches of less than 20 g (in open crucible conditions) yield little usable metal.

If you push on and succeed it’ll be a physicist to prepare the first metallic Nd on this forum!

MrHomeScientist - 31-7-2011 at 18:35

I did quite a bit of work over the weekend on this. It takes a very long time to get to neodymium sulfate - peeling off the nickel coating takes forever, dissolving in acid takes at least a few days, and trying to recover the sulfate is extremely slow because of my multiple heating and cooling cycles. I never want to boil away too much water because it reaches saturation of the FeSO4 and that could start dropping out as well. It also tends to bump pretty violently once some powder precipitates out.

I'm almost done processing another batch of magnets, and the neodymium sulfate I'm getting is quite a bit different than in my initial batch. Overall it is much, much more lavender than before and doesn't change colors nearly as much in different lighting. It's also coming out in many different shades each time I recover a bit more from the solution. Each time I would heat the solution and collect the crystals that formed, they would come out different colors. Here's a picture: the two on the left are the ones from my first run as seen above, here for comparison. This is under fluorescent tube lighting in my kitchen.

Nd sulfate colors.jpg - 105kB

Any ideas to why the different shades of color? I wonder if it's just due to different crystal sizes? Larger grains seem to have a more rich color. Regardless, all the colors are quite different from the initial batch and they aren't nearly as affected by lighting conditions either.

Something else interesting: I got a bunch of magnets donated by some coworkers, including some really thick old hard drive magnets. While just as powerful, I suspected that these might not be neodymium so I dissolved one separately from the thinner magnets I've been using. It dissolved with about the same amount of precipitated gunk as the other magnets, but the solution was green instead of lavender. Here's what it looks like, compared to the purple of a "normal" run:

magnets dissolving.jpg - 37kB

I suspect it's samarium cobalt (when the coating is peeled off it looks identical to Nd magnets), but I have yet to process it further to investigate. I spent enough time on the one on the left (it's what produced the vials from above).

Finally, I wasn't kidding when I said I had a bunch of magnets donated:

HD magnets.jpg - 109kB

Many of them are the thicker variety, so I don't know if they will work for this experiment. In any case, I've got a lot of material to work with!

watson.fawkes - 1-8-2011 at 03:40

Quote: Originally posted by MrHomeScientist  
Any ideas to why the different shades of color?
I'd guess it's due to variations in the rare earth mineral sources for the Nd content. I'd be surprised if the purity of Nd in these magnets is higher than technical grade.

MrHomeScientist - 1-8-2011 at 05:39

Quote: Originally posted by watson.fawkes  
I'd guess it's due to variations in the rare earth mineral sources for the Nd content. I'd be surprised if the purity of Nd in these magnets is higher than technical grade.


I agree with that, but all the vials except the left two were from the same magnets. I probably didn't make that clear enough. After dissolving 4 hard drive magnets that all looked identical, I heated the solution until some Nd sulfate precipitated. I decanted this to collect the crystals in a vial, let the solution cool down, and removed any FeSO4 that precipitated. I then heated it again until more crystals dropped out and collected those in another vial. I repeated this many times and got all the different shades in the picture. So all those colors (except the left two, from another batch) originated from the same magnets.

MrHomeScientist - 31-8-2011 at 18:52

It's been a long, slow process, but I've finally got more results to post. I tried out the next step in the process, adding the ammonium bifluoride to the neodymium sulfate, to obtain neodymium fluoride.

The probable reaction for this step is:
Nd<sub>2</sub>(SO<sub>4</sub>;)<sub>3</sub> + 3NH<sub>4</sub>HF<sub>2</sub> == 2NdF<sub>3</sub> (s) + 3NH<sub>4</sub>HSO<sub>4</sub>

The first step was to dissolve the 2.7g of my first batch of product into 50mL of distilled water. That was the first two vials in my previous post. I'm still amazed by how different lighting causes the color to change.

The initial Nd-sulfate
Nd2SO4 under daylight.jpg - 109kB
Sulfate dissolved in water - from top to bottom: under lab (small) fluorescent tubes, under kitchen (large) fluorescent tubes, and under tungsten bulbs
Nd2(SO4)3 different lighting.jpg - 54kB

According to the above equation, 2.7g of Nd-sulfate requires 1.2g NH<sub>4</sub>HF<sub>2</sub>. This was weighed out and dissolved in 20mL water, in a plastic container (CAREFULLY, and with maximum precautions). I added this dropwise, slowly, to the sulfate solution. A very fine, gelatinous off-white precipitate formed. It was pretty neat - a drop would fall into solution, and about 1/2 second later the precip. appeared. I ended up only adding 13.9mL of the bifluoride solution before additional drops produced no more precipitate. That means that, interestingly, only 0.83g bifluoride was actually needed to complete the reaction.

<i>EDIT: I apparently made a math error above, and 0.8g of bifluoride was in fact the correct amount that was needed. Score one for experimentation!</i>

This precipitate was again sensitive to lighting conditions.
Under lab FL lights
NdF3 precip FL light.jpg - 82kB

Under camera flash
NdF3 precip camera flash.jpg - 105kB

I allowed this to settle for a few days, which compacted it significantly. Then I filtered this through some coffee filters, and let it dry in the filter paper. I had the solution drip into a beaker of calcium chloride solution, to precipitate out any leftover bifluoride as the insoluble calcium fluoride.

Strangely, after letting this dry the color had changed to a lime green under the lab fluorescents! It still changes to pink under the tungsten bulbs though.

The green isn't quite as lime in the picture as it is in person.
NdF3 FL light.jpg - 86kB

NdF3 tungsten light.jpg - 91kB

The total yield of NdF<sub>3</sub> was 1.7g.

I'm puzzled by the color change. This could be some iron contamination - iron(III) fluoride is listed as a pale green solid, while neodymium(III) fluoride is listed as purple (though lighting conditions aren't specified). I'd hypothesize that any iron(II) sulfate contamination was initially reacted into iron(II) fluoride (which is white), then upon exposure to air it converted to iron(III) fluoride (green).

So in conclusion, this appears to have been a success in precipitating the fluoride. Apparently there's quite a bit of iron contamination that I wasn't aware of. That means that I should probably recrystallize my sulfates to get a more acceptable purity.

Once I process more of my sulfate material into fluoride, I can move on to the final step and (hopefully) get my Nd metal!

[Edited on 9-1-2011 by MrHomeScientist]

blogfast25 - 1-9-2011 at 04:17

Well done that man! Very interesting...

Good idea also to catch the filtrate in CaCl2.

The iron fluorides are both water soluble, so unless some got 'caught' (occlusion) the iron content of the Nd should actually have decreased further when precipitating the NdF3.

Can you specify the drying procedure a bit?

Are the ampoules made of glass? Residual moisture could still cause very small (and safe) amounts of HF to attack the glass, if so.

1.7 g of NdF<sub>3</sub> isn’t much of course. To convert it to Nd metal, assuming that’s still your goal, could be tricky. Such small batches of NdF<sub>3</sub> + Mg are likely to fizzle early because the crucible walls act as a heat sink and stop the reaction it its tracks, even if you can get it to go initially. I see as only possibility for such a mini-reduction to heat the mixture by heating the crucible with Bunsen, until reaction starts, then stop heating. You need to protect against fumes!

So far, so good!

Chemistry Alchemist - 1-9-2011 at 04:55

I recently dismantled my old play station HDD and took out the 2 small magnets (2.7 grams all up), the magnets have a silverish coating on the actual magnet, is this nickel or what, when i peel some off the magnet, the under side is shiny, but about an hour later its darkened from the tarnish, could i just put it in some boiling water? Neodymium Reacts fast with Hot water, and the Iron and Boron dont react with water so can i separate the Neodymium this way and then separate tthe other 2 a different way?

Fell free to correct me if im wrong

blogfast25 - 1-9-2011 at 05:17

CA:

The coating is indeed nickel, you need to peel off as much as you can (demagnetise the magnet first by heating it in a gas flame for a short time). Nickel really only dissolves in nitric acid.

Dissolving neodymium in water is virtually impossible. Also this material isn't neodymium, it's an alloy (look up 'neodymium magnets'). To dissolve the magnet you need either fairly strong HCl (20 % or higher) or dilute H2SO4 (at least 20 %). Note that hydrogen will evolve, which is highly flammable: ventilation is a must and open flames a no-no...

READ the relevant threads before doing ANYTHING ELSE. :o

Chemistry Alchemist - 1-9-2011 at 05:25

So once i have dissolved the magnet in HCl, i have among other things Neodymium Chloride, How do i separate Boron and Iron from this?

blogfast25 - 1-9-2011 at 05:31

Quote: Originally posted by Chemistry Alchemist  
So once i have dissolved the magnet in HCl, i have among other things Neodymium Chloride, How do i separate Boron and Iron from this?


The boron doesn't dissolve in acid.

To separate the Nd from Fe, see the relevant threads. There are different ways and they're all explained to death in these threads. Use the search facility and yee shall find. :)

[Edited on 1-9-2011 by blogfast25]

Chemistry Alchemist - 1-9-2011 at 05:45

what if i cant get all the nickel off the magnet? how brittle is the actual magnet, could i crush it with mortar and pastel

Endimion17 - 1-9-2011 at 06:08

MrHomeScientist, well done! It's a treat to see the final compound at the end. :)


Chemistry Alchemist, take some emery paper and start scraping. The scrapings will cling to the magnet, but remove them by pinching with fingers until they're gone.

MrHomeScientist - 1-9-2011 at 06:16

Quote:

Quote: Originally posted by blogfast25  
Well done that man! Very interesting...

Good idea also to catch the filtrate in CaCl2.

Thanks!

[rquote]The iron fluorides are both water soluble, so unless some got 'caught' (occlusion) the iron content of the Nd should actually have decreased further when precipitating the NdF3.

That's interesting, I must have missed that. I can try washing the product with water a few times and see if that helps. Looks like it's only slightly soluble though.

Quote: Originally posted by blogfast25  
Can you specify the drying procedure a bit?

I waited a few days after taking the picture of the precipitate under water to allow it to settle, and it compacted to about 1/3 of that volume. Then I decanted most of the liquid through my filter paper, let that run through, and then poured out the rest that held the precipitate. Since it was so gelatinous, it took quite a while to stop dripping. I left everything alone for another few days, and when I came back it had turned green. So, it was very moist and exposed to air. Once it was dry, it had compacted even further and is now a nice powder.

Quote: Originally posted by blogfast25  
Are the ampoules made of glass? Residual moisture could still cause very small (and safe) amounts of HF to attack the glass, if so.

Yes they are - unfortunately that's all I have. But I do have quite a few, so it's fine if I frost a couple up. Thanks for the tip though, I'll keep that in mind.

Quote: Originally posted by blogfast25  
1.7 g of NdF<sub>3</sub> isn’t much of course. To convert it to Nd metal, assuming that’s still your goal, could be tricky. Such small batches of NdF<sub>3</sub> + Mg are likely to fizzle early because the crucible walls act as a heat sink and stop the reaction it its tracks, even if you can get it to go initially. I see as only possibility for such a mini-reduction to heat the mixture by heating the crucible with Bunsen, until reaction starts, then stop heating. You need to protect against fumes!

So far, so good!

Indeed. Nd metal is still my goal, for my element collection (and the achievement of making it myself!). I remember you mentioning that small 'thermites' were difficult to work with, thanks for the excellent explanation of why. I have a ton of material to work with - see the picture of that huge brick of magnets a coworker gave me - so the plan is to process those and make more of the fluoride before moving on. That's whats taking so long; the magnets take ages to dissolve, and separating the sulfates is also time-intensive. I hope to use labor day weekend for more experimenting!

MrHomeScientist - 1-9-2011 at 06:21

Quote: Originally posted by Chemistry Alchemist  
what if i cant get all the nickel off the magnet? how brittle is the actual magnet, could i crush it with mortar and pastel

It doesn't matter if you can't get all the coating off. When I was doing it I found that it can be really tenacious. All you want to do is expose the actual magnet material to the acid - the more nickel you remove, the faster the reaction will go since more surface area is available. Any leftover nickel just won't react.

Try smashing the magnets yourself (after demagnetizing), they are very brittle. Grinding might be possible, but it'll be a chore and is really unnecessary since you're dissolving in acid anyway. I just broke mine into chunks.

I'd also recommend you read this whole thread and others about neodymium (do a forum search) - a lot of your questions can be answered there.

blogfast25 - 1-9-2011 at 07:14

Mr Home Scientist:

My experience with thermites is that 20 g batches of mixture are 'viable', i.o.w. can be lit the usual way (here just Mg ribbon should get it going) and will burn completely through. Yield of obtained metal does increase with batch size but 20 g should be ok for pilots (you DON'T want to put all your eggs in one basket, considering the effort it takes to get the NdF3). Anything below 20 g may become more problematic.

Your NdF3 will also need to be bone-dry: drying at 200 C or more should do it. Moist powder is difficult to light and can lead to porous slag and poor metal separation.

[Edited on 1-9-2011 by blogfast25]

MrHomeScientist - 24-10-2011 at 19:20

It's been a while, but I finally have time to start work on this again. I got very busy at work and haven't been able to do hardly any hobby chemistry for a few months.

I last left off at dissolving 5 large Nd magnets (about 200g worth) in acid, to make about 1.5 liters of "magnet soup." That's been sitting for a couple months, and it looks like a lot of the iron has oxidized to insoluble brown iron(III) sulfate.

Oxidized magnet soln.jpg - 84kB

It's tough to see, but there's bits of it floating around in there and a large fluffy mass at the bottom of the flask.

Anyways tonight I tried out the oxalate separation method on a sample of this, just to see how it went. I started with 20mL of magnet solution and made a dilute solution of 5g oxalic acid dihydrate in 80mL of water. Oxalic is on the left.

Oxalate start.jpg - 112kB

Slowly adding this to the magnet solution immediately yielded a very compact yellowish precipitate that floated at first, but after stirring quickly settled to the bottom.

NdOx precip.jpg - 102kB

After adding 20mL of the oxalic acid solution, no more precipitation was observed. This corresponds to 1.25g worth of oxalic acid dihydrate used. This was then filtered off. Here is the leftover solution, which has turned very yellow. Encouragingly, it shows no color change in different lighting, suggesting nearly all the Nd has been separated out.

NdOx leftover soln.jpg - 83kB

The precipitate also looks distinctly yellow under lab fluorescents, but changes back to pink (with a yellow tinge) under tungsten bulbs.

NdOx fluorescent lighting.jpg - 76kB NdOx tungsten bulb.jpg - 68kB

Now the yellow color suggests to me that some iron(II) oxalate is contaminating my Nd oxalate. Perhaps when it's dry and weighed, I'll try dissolving the iron back into solution? I do like how quick and effective the separation of Nd from Fe appears to be. Much easier than the repeated cooling and heating cycles I've been doing to separate the sulfates, especially with the amount of solution I now have to process.

So assuming this is indeed NdOx (and that I can remove the iron impurities likely present), where do I go from here? I was thinking calcine it to Nd<sub>2</sub>O<sub>3</sub>, then react with sulfuric or hydrochloric acid to get it into solution, then precipitate it back out as NdF<sub>3</sub> with my ammonium bifluoride. Does this sound reasonable?

Thanks for any feedback. I hope to really get moving on this project soon.

DerAlte - 24-10-2011 at 21:51

Re this thread, see also my earlier posts on RE metals Ce, Nd, Sm:

http://www.sciencemadness.org/talk/viewthread.php?tid=12934#...
http://www.sciencemadness.org/talk/viewthread.php?tid=8758&a...

All references I have seen say Nd sulphate is red and Sm is light yellow, Ce white (colourless); and these were the colors I got. All these sulphates are very difficult to dissolve at 0C, the most soluble point. Once dissolved, heating to 100C usually will precipitate them. Better crystals can be obtained by slow evaporation.

Regards, Der Alte

blogfast25 - 25-10-2011 at 05:26

@Mr Home scientist:

If your iron hadn’t been completely converted to Fe (III) then your Nd oxalate precipitate is likely to be contaminated with insoluble Fe (II) oxalate (yellow). Check the mother solution with K3Fe(CN)6 for Fe2+ presence (it gives Turnbull’s blue with Fe2+, quite a sensitive test).

Now try adding 9 % (or better) H2O2 to the mother solution to oxidise any remaining Fe2+ to Fe3+. Do this VERY SLOWLY, aliquot by aliquot, cooling if needed, as it’s highly exothermic and effervescent. You can also use nitric acid, if preferred. Then simmer gently for a bit to get rid of excess H2O2. Cool and then repeat what you did and compare colours.

If your current Nd2Ox3 really is contaminated with FeOx, I think I know a trick to get the latter out of there… A very small amount of Nd2Ox3 I precipitated recently looked just plain white to me, to be honest. I might publish a photo here…

Der Alte: thanks for the links. My Nd2(SO4)3 was more like a salmon-pink colour…

The trouble with the oxalate route is that Nd2Ox3 is so damn insoluble (Ksp = 1.3 x 10<sup>-31</sup>;), so that once you’ve obtained it there’s a ‘what now?’ moment. You can fire it to Nd2O3 but that’s pretty insoluble too. So, shortly I’ll be making some ‘magnet soup’ too and I think I’ll ‘go sulphate’, to be honest…

But I wonder also if use could be made of a lesser known Fe (III) complex, FeOx<sub>3</sub><sup>3-</sup> (ferrioxalate or trioxaloferrate (III) anions) which has a complexation constant of Kf = 2 x 10<sup>20</sup>.

The trick would be to obtain the ‘magnet soup’ with iron as ferric ions, then adding a sufficiency of K<sub>2</sub>Ox and excess oxalic acid. This should the complex the ferric ions to ferrioxalate complex, while precipitating the Nd as oxalate, with near perfect separation. Some have reported difficulties with straight ferric oxalate crystallising out, thereby making separation difficult, this trick could circumvent that from happening.

Some work on ferrioxalate here:

http://www.sciencemadness.org/talk/viewthread.php?tid=17831


[Edited on 25-10-2011 by blogfast25]

blogfast25 - 26-10-2011 at 11:40

Scanning my old Holleman (‘inorganic Chemistry’) for a few more useful clues on RE chemistry, I came across a notion I’ve not seen posted on this forum yet.

Holleman claims that lanthanum, cerium, praseo- and neodymium salts can be separated from the others because the former are insoluble in saturated K2SO4, due to the formation of double sulphates (RE alums?)

I decided to put that to the test with my NdCl3 stock solution (est. about 1 M) by saturating about 2 ml in a test tube with K2SO4. Sure enough, a thick precipitate of presumably (Nd,K)n(SO4)m resulted. This was washed and then excess KOH was added to the slurry with a bit of water. The precipitate became fluffier at once. I heated it for a bit to ensure the supposed conversion to Nd(OH)3.nH2O was complete, then cooled and washed and decanted about 3 times.

The moment of truth: the precipitate dissolves effortlessly and without heat in a couple of ml of 36 % HCl.

So this would appear to provide another oxalate free route: saturate the solution of ‘neomagnet chloride’ with K2SO4, filter and wash the filter cake with sat. K2SO4, then convert the alleged double sulphate with strong alkali hydroxide to Nd hydroxide, wash profusely with hot clean water.

MrHomeScientist - 26-10-2011 at 19:31

Blogfast25,

I tried out your hydrogen peroxide suggestion, and my oxalate came out looking very nice! I also used your idea of testing for iron(II) with potassium ferricyanide - the mother solution has a strong positive test (lots of immediate blue). I tested small amounts during the addition of peroxide, and stopped adding it when the test only showed a faint blue, which still took a few minutes to appear. Writeup follows.


I again started with 20mL of 'magnet soup' and made up a roughly 9% H<sub>2</sub>O<sub>2</sub> solution by diluting 2mL of 28% peroxide to 6mL. Peroxide is on the left. I ended up making another 6mL of this and using all of that as well.

Peroxide start.jpg - 113kB

I added this dropwise, and it rapidly heated up with lots of bubbling. The solution quickly turned very dark brown, almost black.

Peroxide added.jpg - 96kB

I then left to go eat dinner, and when I came back it had changed color to brown! I imagine this is iron(III) sulfate coming out of solution? I tried filtering it but it didn't clear it up at all (coffee filters probably aren't good enough).

Peroxide added, some time elapsed.jpg - 110kB

I decided to press on anyway. I heated it for a short while to decompose the peroxide (as suggested). I then used the rest of my oxalate solution I made for the previous test plus 50mL more of a new batch, for a total of 110mL of solution or about 6.9g worth of oxalic acid. This one took <u>much</u> more oxalic acid to complete. I would add some, the precipitate would form just like last time, but on stirring it would promptly redissolve. The solution first cleared of the brown precipitate and returned to a light brown color (darker than the original solution), then ended up lime green once the Nd finally started precipitating out. I filtered this off and collected it. Here's a comparison between the two methods - the stuff from my previous post is on the left.

Peroxide vs Oxalate.jpg - 117kB

And here's a comparison of the Nd salts under lab fluorescents (left) and tungsten bulb (right). The sample from this test is on the bottom, and the one from my previous post is on top.

Oxalate comparison fluorescent.jpg - 96kB Oxalate comparison tungsten.jpg - 91kB

As you can see, the color is <i>much</i> nicer this time around. I'm sure you were correct in saying some iron(II) oxalate was contaminating my other sample. I really like how pure white the bottom vial (from this run) looks! It sounds like it matches your own observations.

I ended up with 1.2g of Nd<sub>2</sub>Ox<sub>3</sub> this time, and 1.7g last time. I optimistically imagine the extra mass from last trial was the iron contamination.

Quote: Originally posted by blogfast25  
The trouble with the oxalate route is that Nd2Ox3 is so damn insoluble (Ksp = 1.3 x 10<sup>-31</sup>;), so that once you’ve obtained it there’s a ‘what now?’ moment. You can fire it to Nd2O3 but that’s pretty insoluble too. So, shortly I’ll be making some ‘magnet soup’ too and I think I’ll ‘go sulphate’, to be honest…


The idea here is to obtain the oxalate, heat it to decomposition to the oxide, then react this with acid (either sulfuric or HCl) to make the Nd soluble again. Then, this can be reacted with my bifluoride as per the original plan to get NdF<sub>3</sub>. So solubility shouldn't be an issue, unless you see a problem with this strategy?

Quote: Originally posted by blogfast25  
So this would appear to provide another oxalate free route: saturate the solution of ‘neomagnet chloride’ with K2SO4, filter and wash the filter cake with sat. K2SO4, then convert the alleged double sulphate with strong alkali hydroxide to Nd hydroxide, wash profusely with hot clean water.


I like this idea too, I'll give that a try soon. Thanks for doing some extra research on this! That's interesting to me because I just recently got started with another project trying to make alum from coke cans, so it's neat to see double salts here too.

For this route, once I've obtained Nd(OH)<sub>3</sub> would it be possible to react that with bifluoride to get NdF<sub>3</sub>? I'm not sure if it would work all that well since the hydroxide is also insoluble. Seems like no matter what separation technique is used, I'd want my Nd in a soluble form before moving on to the next step.

The reason I'm looking at alternative separation techniques, by the way, is because of the volume of solution I have. In order to squeeze out all the Nd from my magnet soup, I've had to run the solution through 5 or 6 heating and crystal collecting cycles. That takes ages, and it will be very tedious to do that for so much liquid. It'd be nice if I could just mix in some oxalic acid or potassium sulfate and see immediate and total separation. I feel like these would yield higher overall purity too, considering the different colors I got from different crops of crystals from the same solution (see one of my earlier posts). I suppose that could be more or less solved by a recrystallization, though.

blogfast25 - 27-10-2011 at 05:08

MrHomeChemist, I know you’re a physicist but now it’s official: you’re also a chemist!

This is great work.

A few comments.

The ferrioxalate method is a little ‘hairy scary’ in the sense that both times I’ve made this complex it appeared touch and go as to whether the conversion to complex will proceed. In actual fact it simply takes enough oxalate/oxalic acid for it to happen, that’s all. Keep adding oxalate slowly, with time between additions, keeping hot until the red turns green.

The red precipitate (pict. 3) MUST be ferric oxalate: ferric sulphate is too soluble to precipitate like that and ferric hydroxide can mathematically be shown to be too insoluble to re-enter solution as ferrioxalate. So it must be Fe2Ox3. In the actual synth. of K3FeOx3 oxalate is added to yellow FeOx, then peroxidised: it turns red/brown. On adding extra oxalic acid eventually the complex forms over an interval of a few seconds.

In your salts comparison, aren’t you comparing sulphate with oxalate? That would be an unfair comparison: we know that because of Nd’s ‘dangling f-electrons’ Nd salt colours are affected by the type of anion.

Yes, Nd oxalate fires to Nd2O3 (allegedly at about 700 C) and the calcined oxide will dissolve in hot, concentrated sulphuric acid (but not in conc. HCl, I would think). It just adds another step, that’s all. That’s why I’ll be testing the pot. sulphate double salts idea. I should be back in the race soon!

Very good question about reacting the Nd(OH)3 with bifluoride. I’ve been pondering that myself. Only the solubility products (Ksp) of both substances allows to make a judgement. In short, if Ksp,hydroxide > Ksp,fluoride, then conversion should be possible. I have no value for Ksp,fluoride and was planning on re-dissolving the hydroxide back to chloride first, just to be sure.

Volumes? Blimey, you really are thinking like a chemist now. Sure, when processing larger batches, volumes that need to be processed can be prohibitively large!

Al cans to alum? I know the text you’re referring to. Works very well. A bit boring though (done it myself). Spice it up by scaling up a bit. Double salts can be quite interesting.


[Edited on 27-10-2011 by blogfast25]

[Edited on 27-10-2011 by blogfast25]

plante1999 - 27-10-2011 at 08:37

Very nice Synth. of NdF3 here. From wath I read you are planning to reduce it to Nd the element , how are you planning to do that , in wath condition?

blogfast25 - 27-10-2011 at 09:37

Quote: Originally posted by plante1999  
Very nice Synth. of NdF3 here. From wath I read you are planning to reduce it to Nd the element , how are you planning to do that , in wath condition?


Magnesiothermic reduction with magnesium powder:

NdF3 + 3/2 Mg === > Nd + 3/2 MgF2

Very similar to thermite. All thermochemical calculations show that enough heat will be generated to exceed the MP of MgF2 (and of course the MP of Nd). It's similar to the industrial reduction of UF4 with Mg for the production of metallic uranium.

Aluminothermy (on the oxide) doesn't work here because of the high HoF of Nd2O3. And AlF3 is far too volatille...

MrHomeScientist - 27-10-2011 at 10:03

Blogfast25: Thanks so much for your support! That really brought a smile to my face :)

Some clarification on two points:

Quote: Originally posted by blogfast25  
The red precipitate (pict. 3) MUST be ferric oxalate: ferric sulphate is too soluble to precipitate like that and ferric hydroxide can mathematically be shown to be too insoluble to re-enter solution as ferrioxalate. So it must be Fe2Ox3. In the actual synth. of K3FeOx3 oxalate is added to yellow FeOx, then peroxidised: it turns red/brown.

Actually, that picture was taken <i>before</i> adding any oxalic acid. I added the peroxide to the mother solution until it barely tested positive any more for iron(II), then went to eat dinner, and upon returning it had changed to that brown color. I heated it for a few minutes, and only then added the oxalic acid. So that photo should only have iron(III) and Nd salts in it.

Quote: Originally posted by blogfast25  
In your salts comparison, aren’t you comparing sulphate with oxalate? That would be an unfair comparison: we know that because of Nd’s ‘dangling f-electrons’ Nd salt colours are affected by the type of anion.

No, those final two pictures are both Nd oxalate. The top vial in both pics is the yield from my post before that one (where I just added oxalic acid to the mother solution) and the bottom vial is the yield from this last trial. The photo on the left is both samples under fluorescent lighting, and the photo on the right is under a regular tungsten bulb. The sample on top is undoubtedly contaminated with some yellow iron(II) oxalate, whereas the one on bottom from this latest trial looks much better.

Quote: Originally posted by blogfast25  
Very good question about reacting the Nd(OH)3 with bifluoride. I’ve been pondering that myself. Only the solubility products (Ksp) of both substances allows to make a judgement. In short, if Ksp,hydroxide > Ksp,fluoride, then conversion should be possible. I have no value for Ksp,fluoride and was planning on re-dissolving the hydroxide back to chloride first, just to be sure.

That seems like the safest way to me. Even if the hydroxide does directly react, it may be hard to distinguish the two compounds. When I made the fluoride it looked quite fluffy too, so you may not be able to tell when the reaction is complete.

Quote: Originally posted by blogfast25  
Al cans to alum? I know the text you’re referring to. Works very well. A bit boring though (done it myself). Spice it up by scaling up a bit. Double salts can be quite interesting.

Yeah it was quite simple and is looking great so far. I'm just waiting for my solutions to evaporate down a bit more before collecting the crystals. I'd never seen it done before so I thought it would be fun to try. I also thought it would be neat to have a big single alum crystal and say it used to be a soda can! That would boggle the minds of many of my friends :)

blogfast25 - 27-10-2011 at 11:22

Quote: Originally posted by MrHomeScientist  

Actually, that picture was taken <i>before</i> adding any oxalic acid. I added the peroxide to the mother solution until it barely tested positive any more for iron(II), then went to eat dinner, and upon returning it had changed to that brown color. I heated it for a few minutes, and only then added the oxalic acid. So that photo should only have iron(III) and Nd salts in it.

No, those final two pictures are both Nd oxalate. The top vial in both pics is the yield from my post before that one (where I just added oxalic acid to the mother solution) and the bottom vial is the yield from this last trial. The photo on the left is both samples under fluorescent lighting, and the photo on the right is under a regular tungsten bulb. The sample on top is undoubtedly contaminated with some yellow iron(II) oxalate, whereas the one on bottom from this latest trial looks much better.



Oh well, ‘Bang goes my theory’. The brown solution was essentially strong FeCl3 and on cooling something like a ferri hydroxichloride must have precipitated, due to high concentration. Presumably aided by the low pH that was then able to get complexed with the oxalate to the trioxaloferrate anion.

So both photos are Nd2Ox3, huh? Crickey, the second run clearly gives a cleaner product. It’d be great to compare them to a sample obtained with potassium sulphate as a precipitant. It does also seem to confirm that with an excess Fe the oxalate based method (without complexing) may not have the right resolution to achieve full separation between these elements.

It might be useful to pool all your samples together for a final purification.

Re. these alum crystals: as long as you know that growing large, near perfect ones is a labour of love!

MrHomeScientist - 3-11-2011 at 10:14

While trying to figure out what exactly Nd2Ox3 decomposes into so I could do some stoichiometry, I came across this interesting little side note at the end of a short article on neodymium chloride (from 1901, found here ):

Quote:
I may finally add that I have been able to isolate the [Nd] metal itself by the action of sodium on the anhydrous chloride.


That's all the detail he goes into, but it does sound interesting. I know you have the chloride blogfast- got any sodium? The vague word 'action' makes me wonder what all is actually involved. Making it anhydrous may also be difficult.

Quote: Originally posted by blogfast25  
Oh well, ‘Bang goes my theory’. The brown solution was essentially strong FeCl3 and on cooling something like a ferri hydroxichloride must have precipitated, due to high concentration. Presumably aided by the low pH that was then able to get complexed with the oxalate to the trioxaloferrate anion.


Remember I used sulfuric acid here, so all that should be in the mother solution at this point is Nd2(SO4)3, Fe(SO4), and Fe2(SO4)3. No chloride involved!

blogfast25 - 3-11-2011 at 10:59

Reduction of anh. NdCl3 with sodium is probably possible but you'll get Nd sponge riddled with NaCl: hard to separate because Nd is very reactive. Also: hard to get a lump of metal, unless you heat the assembly strongly, vacuumed and backfilled with low pressure Ar. Above my paygrade... Anhydrous NdCl3: heat the hydrate in inert armoshere, mixed intimately with an excess of NH4Cl or heat hydrate in a stream of pure, dry HCl.

For Mn, reduction of the anh. dichloride dissolved in molten KCl with Mg, At the end the crucible is then heated to above the MP of Mn and a puddle of the metal then forms. Could work for NdCl3 too...


barley81 - 3-11-2011 at 11:44

I think this happened:
The brown solution formed by addition of H<sub>2</sub>O<sub>2</sub> is Fenton's reagent (Iron IV) which decomposed to ferric ion and oxygen. Since the overall oxidation from ferrous ion to ferric ion with H2O2 consumes H<sup>+</sup>, the ferric ions precipitated as basic sulfate when the pH grew high enough.

blogfast25 - 4-11-2011 at 06:34

Fenton's reagent doesn't involve Fe (IV).

You're probably right about basic ferric sulphate being formed, though...

Wizzard - 6-11-2011 at 22:02

First, a thought- If we could get iron acetate, and Nd acetate, I read that Nd(III)Ac is non-soluable, while Fe(II)Ac is, and also NdAc decomposes at temp :)

But today for me was a bad day. I found in the back of my storage area a rack previously used for long-term growth of crystals by evaporation... 3 bottles, one grey with white crust (still with some liquid) one green with white crust, and 1 labelled as ammonium nitrate, but bone dry and empty.

Trying to scrape out the green, thick crust of the first bottle failed with my tool at hand, so I poured out the liquid of the grey one into my kitty-litter disposal tub. Taking the bottle with me into the other room (where I had a good plastic scraper), I grabbed my tool and went to jam it in the bottle at the large, noncrystalline mass in the bottom of the bottle- But it was now distinctly purple, with a white crust. I quickly ran to the next room, bathing the sample under flouresent light- Back to grey! The sunlight revealed my mystery chemical.

So now I have a gram or two of mostly neodymium sulfate (the chemical I had been toying with more than a year ago) with some minor iron sulfate impurity, which was partially cleaned out physically by removing the white crust. But I wish I had not thrown out the thick, opaque 5ccs of liquid :(

Pic related.

IMG_20111106_221814.jpg - 109kB

[Edited on 11-7-2011 by Wizzard]

blogfast25 - 7-11-2011 at 10:25

Wizz:

I bet you good money that stuff is badly hydrolysed, it looks sooooo like an NdCl3 solution that fell foul due to my neglect, a while back. Try resuscitation with strong HCl, folloed by filtering. Treat the filter residue with strong alkali to convert to Nd(OH)3, then reapply HCl, filter and combine with first filtrate. To get rid of the Fe, this thread is now jam packed full of ideas.

My next post will be on exploiting the insolubility of a(n unknown to me) double salt of Nd2(SO4)3 and K2SO4 in sat. K2SO4 solution as a means to separate Nd from Fe, Ni and Cu.

Wizzard - 7-11-2011 at 11:29

I'm dessicating it right now, I'm not really worried about it- I have a 125g (weight of magnets) batch running right now :) Maybe I'll hit with some good 34% HCl when I get home, see what happens- If I can get it clean enough as either a sulfate or chloride, I'm going to react it with some zinc acetate I have lying around. Thank God for my recent find of an old photographer's chemical closet :)

blogfast25 - 7-11-2011 at 13:09

So I set out to try and quantitatively isolate the neodymium from neomagnets - Nd2Fe14B. 35 g of magnets were demagnetised by heating and cracked with a hammer. Only part of the Ni-Cu-Ni protective layer was removed and the magnets dissolved in 155 ml of 36 % HCl, on a low setting hot plate and covered with an hour glass. The cover material was later found back but the nickel had dissolved, possible small amounts of copper too. Strong, meaty smell, some suggest of boranes (?)

After filtering and washing the filter cake (mainly undissolved boron) the solution was simmered back to about 250 ml of emerald green solution, covered to prevent loss of salts. To this hot solution, 25 g of crushed (and previously recrystallised) K2SO4 was added, a quantity corresponding to approx. saturation at 20 C. Immediately a quite large amount of precipitate formed, presumably the unspecified double salt of Nd2(SO4)3 and K2SO4 and the solution/slurry simmered for a bit, the allowed to cool. No further precipitation on cooling was observed. With flash on, the precipitate can be clearly seen:



This was filtered off on a coffee filter (which with hindsight was oversized) and the filter cake washed with about 200 ml of sat. (at RT) K2SO4. At the end the filtrate still tested very slightly positive with oxalic acid solution, showing a slight and slightly yellow precipitate of FeOx. Further rinsing reduced this signal but only slowly. I believe it due to Fe2+ solution in the large part of the filter slowly seeping into the filtrate.

To remedy this, the filter cake, which is typically lavenderish in colour, was transferred as quantitatively as possible into 100 ml of saturated (at 20 C) K2SO4 and allowed to cool and decant. The supernatant liquid showed very slight white turbidity with oxalic acid solution, suggesting the unspecified Nd/K/SO4 double salt may be very poorly soluble rather than totally insoluble in sat. K2SO4. Here it is:



The supernatant liquid was decanted off (the double salt is clearly ‘heavy’) and to the thick slurry a solution of 13.5 g of KOH in 100 ml of water was added and the precipitate changed appearance slightly. The slurry was kept warm for about ½ hour, stirring occasionally. Supernatant liquid is strongly alkaline. The precipitate, presumably iron, nickel and copper free neodymium trihydroxide is clearly light pink:



Tomorrow this will be washed with warm water until the filtrate runs almost neutral and sulphate free. The purified hydroxide will then be dissolved in an appropriate amount of hydrochloric acid.

Edit:

I said the hydroxide was ‘lightly pink’ (on the photo) but to the naked eye the precipitate is actually very light blue! Which is what I remember form a first attempt. Those dangling f-electrons…

And here’s a reference to Na, K, and NH4 neodymium sulphate double salts in the service of recycling NdFeB scrap.

Update:

To cut a long story short after washing the hydroxide till the filtrate tested negative for sulphate and ran at pH = 10.5, it was then transferred quantitatively into a clean beaker, the volume made up to 50 ml and 17 ml of 36 % HCl (approx. the least amount) was added. Prior to adding HCl the filter cake was a nice light blue. The resulting solution however, after heating was slightly turbid still, urine yellow in colour and… tested quite positive for ferric ions! The separation hadn’t been 100 %. Quite disappointing…

This solution was then filtered to clarity and the K2SO4 separation method will be applied once again to see if complete elimination of iron is possible.

It appears the double sulphates are of the formula M2SO4.Nd2(SO4)3.2H2O with M an alkali metal (http://books.google.co.uk/books?id=walWQrTRNK4C&pg=PA47&...)


[Edited on 8-11-2011 by blogfast25]

MrHomeScientist - 8-11-2011 at 09:42

Quote: Originally posted by blogfast25  
So I set out to try and quantitatively isolate the neodymium from neomagnets - Nd2Fe14B. 35 g of magnets were demagnetised by heating and cracked with a hammer. Only part of the Ni-Cu-Ni protective layer was removed and the magnets dissolved in 155 ml of 36 % HCl, on a low setting hot plate and covered with an hour glass. The cover material was later found back but the nickel had dissolved, possible small amounts of copper too. Strong, meaty smell, some suggest of boranes (?)


You know, that's something I noticed too. I was under the impression that Ni and B wouldn't dissolve in sulfuric (in your case hydrochloric) acid under these conditions, but if left in there long enough it looks like they do. I've had some magnet pieces with small bits of Ni coating still on them when going into the acid, and couldn't find them after the magnet itself had dissolved. I also noticed that the boron initially separates out as a black powder, but over time also disappears if not filtered off. Weird.

Quote: Originally posted by blogfast25  
To cut a long story short after washing the hydroxide till the filtrate tested negative for sulphate and ran at pH = 10.5, it was then transferred quantitatively into a clean beaker, the volume made up to 50 ml and 17 ml of 36 % HCl (approx. the least amount) was added. Prior to adding HCl the filter cake was a nice light blue. The resulting solution however, after heating was slightly turbid still, urine yellow in colour and… tested quite positive for ferric ions! The separation hadn’t been 100 %. Quite disappointing…

This solution was then filtered to clarity and the K2SO4 separation method will be applied once again to see if complete elimination of iron is possible.


Disappointing indeed :( Everything was looking great up until the end there - you did everything I could think of to get the best product possible. I was excited about this route because of its simplicity and use of a cheap chemical. Hopefully you have more luck with precipitation round 2.


In other news, I took my less pure sample of Nd2Ox3 (the yellowish top vial from the last picture of my last post) and tried heating it to decomposition last night. I used a fused silica crucible heated with a blowtorch from below. I placed the 1.7g of oxalate into it, and after about 5 minutes of strong heating the color started darkening quite a bit. To cut to the chase, the final product is now completely grey and was reduced in mass to 0.9g. This pretty much confirms that iron(II) oxalate made it into my precipitate, and was decomposed to a fine iron powder upon heating. So that failed as well :(

It's possible I didn't heat the crucible enough to reach the decomposition temperature of the Nd-oxalate, since I stopped after it was clear that it was contaminated with iron. I couldn't find a reference for the temperature at which it decomposes - did you ever find anything mentioning this? I'll try the same procedure on my other, pure white Nd-oxalate sample soon and see what I come up with. It would probably help to loosely cover the crucible next time, to help keep the temperature up.


Quote: Originally posted by blogfast25  
And here’s a reference to Na, K, and NH4 neodymium sulphate double salts in the service of recycling NdFeB scrap.


I wish I could see the rest of that paper, looks like it's chock full of ideas. I'd be really interested in reading it.

==Edit==

On further reading, I came across the following reference which I don't really understand:

"To separated only Nd from the solution the U.S. Bureau of Mines has developed a technique using H2SO4 dissolution followed by precipitation of the Nd using an addition of NH4OH. In this process the Fe remains in solution..."

Source: http://www.osti.gov/bridge/servlets/purl/10190438-P7QwV3/web...

Are they saying aq. ammonia supposedly selectively precipitates out the Nd? Would it not also drop out iron hydroxide at the same time?

[Edited on 11-8-2011 by MrHomeScientist]

Wizzard - 8-11-2011 at 11:43

That might be worth trying. They mention the iron impurity is taken left as jarosite- That's got potassium hydroxide written all over it.

blogfast25 - 8-11-2011 at 13:10

Ni and Cu are indeed slightly soluble in HCl or H2SO4, although as solvents for these metals they're hardly ideal (HNO3 is). If the alloy contains some borides and/or intermetallics involving boron, then some dissolution of the boron must take place.

I may have been asking a bit much of the double salts method, considering the initial molar ratio Fe/Nd is about 7! Wait and see with round 2. BTW, Glauber salt, Na2SO4, can also be used, as per that linked reference.

I read that the minimum firing temp. for Nd oxalate to oxide is about 700 C, attainable with a decent crucible (with lid) and a hot blowtorch. Not sure about firing times, though. I'm not keen on the oxide because the only method to dissolve it is with conc. H2SO4 and that leads back to the treacherous Nd sulphate.

I can't wholly see the *.pdf theyr're referring )'i/O error has occurred!') to but I can assure you that selective precipitation of Nd from Fe (or vice versa) is NOT possible. I believe instead they're referring to the ammonium sulphate/neodymium sulphate double salt, by creating (NH4)2SO4 in situ. But when I tried to saturate a bit of NdCL3 solution with (NH4)2SO4 I didn't get any precipitate...

kmno4 - 8-11-2011 at 14:06

Quote: Originally posted by blogfast25  
I'm not keen on the oxide because the only method to dissolve it is with conc. H2SO4 and that leads back to the treacherous Nd sulphate.

It is of course not true. Nd2O3, even heated to 1000 C and cooled is easily soluble in dilute acids (this is typical for lanthanides)
Quote: Originally posted by blogfast25  
I can't wholly see the *.pdf theyr're referring )'i/O error has occurred!') to but I can assure you that selective precipitation of Nd from Fe (or vice versa) is NOT possible.

Also not true. It is easy to remove contamination of Fe as Fe2O3xH2O from Nd salts.

blogfast25 - 8-11-2011 at 14:47

Quote: Originally posted by kmno4  
It is of course not true. Nd2O3, even heated to 1000 C and cooled is easily soluble in dilute acids (this is typical for lanthanides)

(snip)

Also not true. It is easy to remove contamination of Fe as Fe2O3xH2O from Nd salts.


Re. the first statement, how easy/difficult Re2O3 dissolves in strong acids must depend on the degree of calcination and concentration of the acid, surely?

RE. the second point, what do you mean? That Nid(III) can be separated from Fe(II) without oxidising Fe(II) to Fe(III)? That sounds like a hat trick to me! By selectively precipitating Nd(OH)3? Similar to the removal of Fe2O3.nH2O from MnSO4 solutions? Like here: http://www.sciencemadness.org/talk/viewthread.php?tid=17931#... ? That would be very interesting kmno4! I would indeed stand corrected...

Edit:

This paper puts the solubility product (Ksp) of Nd(OH)3 at about 2 E-19. The Ksp of Fe(OH)3 is about 2.8 E-39, that’s a lot of orders of magnitudes lower and would indicate that a mixture of Nd3+ and Fe3+ salts could be separated by adding freshly separated Fe(OH)3 to the solution. I’ll test tube test that today…



[Edited on 9-11-2011 by blogfast25]

blogfast25 - 9-11-2011 at 09:27

I test tubed the idea of separating Nd3+ from Fe3+ by adding freshly precipitated Fe(OH)3.nH2O to such a solution on a few ml of the Fe 3+ contaminated NdCl3 solution described above. After allowing equibrium to be attained and plenty ferric hydroxide was left as precipitate, that solution was filtered (several times - some ferric hydroxide kept coming through) and the clear filtrate was that typical colour of a fairly strong Nd solution and tested very positive with oxalic acid but completely negative for Fe3+!

This opens up another avenue for this separation, as I can’t see any impediment to using the principle on a more concentrated Fe3+ solution. One would dissolve the magnet material in HCl, oxidise all Fe to Fe [+III], adjust the pH with ammonia to about 4 - 5, add an adequate amount of fresh, clean ferric hydroxide, allow to equilibrate, then filter. The filtrate should contain all the Nd and no Fe. Then precipitate Nd as hydroxide or oxalate.

I also repeated the double salt method on the rest of the contaminated sample and this time obtained a nice blue Nd(OH)3 which dissolved in about 10 ml 36 % HCl (suggesting there is no more than 5.6 g Nd there) and that solution showed the typical colours of an NdCl3 solution: greenish under strip light, kind of orangy under incandescent light. But the solution still tested positive for Fe3+!

Arthur Dent - 9-11-2011 at 10:45

How volatile is NdCl3? Iron contamination can be removed from Manganese Chloride by the fact that Ferric Chloride is highly volatile at very high temperature. I was able to refine some dark tea-colored Manganese Chloride solution contaminated with iron simply by boiling it until very little liquid was left.

As the solution was boiling away, it slowly became lighter and lighter until it was nearly colorless. Upon cooling, the resulting pink-ish gunk looked very much free of iron impurities and was dissoved in distilled water. After evaporation, it produced very clean, beautiful Manganese Chloride crystals.

Could this process be applied to a solution of Neodymium Chloride/Ferric Chloride ? Perhaps this substance would act in a similar way? Maybe something to be attempted?

Robert


blogfast25 - 9-11-2011 at 13:05

Quote: Originally posted by Arthur Dent  
How volatile is NdCl3? Iron contamination can be removed from Manganese Chloride by the fact that Ferric Chloride is highly volatile at very high temperature. I was able to refine some dark tea-colored Manganese Chloride solution contaminated with iron simply by boiling it until very little liquid was left.

As the solution was boiling away, it slowly became lighter and lighter until it was nearly colorless. Upon cooling, the resulting pink-ish gunk looked very much free of iron impurities and was dissoved in distilled water. After evaporation, it produced very clean, beautiful Manganese Chloride crystals.

Could this process be applied to a solution of Neodymium Chloride/Ferric Chloride ? Perhaps this substance would act in a similar way? Maybe something to be attempted?

Robert



Perhaps Robert but remember that we're talking Nd2Fe14B here: that's a lot of iron. NdCl3 isn't particularly volatile but the hydrate is likely to hydrolyse severely on strong heating.

Wizzard - 9-11-2011 at 13:28

I'm a bit worried about the excess sulfuric acid reacting with whatever gunk is left over after it's made salts of the iron and Nd- Does anybody know what could possibly be left to react? Does boron react? Is there some slow-reacting, previous unmentioned impurity?

I took the 'slush' from my large batch (it looked like silver/blueish snow slush) and transferred it to a new vessel- I then added 98% Sulfuric, and it's been slowly bubbling for a few days now... I'm going to re-filter and crystallize by evaporation.

Of note- A great way to get the nickel plating from the magnets is crack them into a few pieces (2-3-4 per magnet) and heat with direct heat from a torch- The nickel with mostly peel itsself off, the magnet material will take on a blue hue, making pieces without coating easier to spot.

Poppy - 9-11-2011 at 18:11

As extensive research have been done with various chemicals in an attempt to isolate pure neodimium based compounds, all seemed unsuccesful and as there must had been noticed, all attempts were based on stoichometric quantities to the reagents into account (I so did I in the following) I took the previous information and added to some studies performed with iron compounds to create a theory that stochoimetry is only appliable to foresee reaction routes but if not treated with equilibrium constants they fail at determining the behaviour of non-abc elements (the old ones everyone knows). Thus Nd holds to merit for being long undiscovered rare earth.

First, some iron compounds considerations (stoichometric), since much of the reactions with magnets indeed involves iron!!
Fe + H2SO4 - >FeSO4 (green solution which easily provides beautyful crystals when crystalized)
ferrous sulfate recrystalized.JPG - 18kB
Then picking recrystalized salt:
FeSO4 + 2 NaOH --> Fe(OH)2 + NaSO4
First of all, stoichometric obtention of the products is already not likely to occur, althogh that deviation just to a little extent (runs unnoticeable)
The hint here is to add much more NaOH than necessary stoichometrically to keep pH high enough to prevent formation of hybrids which absolutly occur, unnoticeable.
By the Schirkorr reaction mechanisms, the one which occurs at low temperatures in water
2Fe(OH)2 + 2H2O --> 2Fe(OH)3 + H2
FeII hidroxide is supposed to be a green-white gel. But stoichometric obtention of the same, from chloride or sulphate yield a greyish blue gel, with some very brillant blue micro (visible) incrustations which I assume are hybrids form the precipitation method.
The oxidation to FeIII in water is aided by atmospheric oxygen but this route seems not to form FeII hidroxides only.
When heated in oxygen free water, Fe(OH)2 goes into Fe(OH)3 and rapidly changes into magnetite (Fe3O4 - not Fe2O3). Thus the color of the hidroxide gel goes from greyish blue to greyish black.
But now the important part: does FeII hidroxide only turns into FeIII at high temperatures?
Reacting FeII produced before with sulfuric acid yields a dark brown solution just like a much certain product from the reaction oh hematite (Fe2O3) with sulphuric acid. It should be green FeSO4, but the multistep reaction with sulphuric acid gives ferric sulphate.
Reacting Fe2(SO4) with NaOH to precipitate iron hidroxide again gives a very dark brown iron III hidroxide as product - oxidation really occurred to a great extent.
Notice in the Schirkorr reaction route into magnetite described above it is not seen intermediary brown colorations, as verified experimentally.
By the last into considerations, reacting a hot solution of brown ferric sulphate with 98% concentrated sulfuric acid gave back again a green solution!!!
reduced back to green.JPG - 23kB
Maybe when sulfuric acid ionized it reduced ferric iron back into ferrous iron, the solution standing with very excessive sulfuric acid.
The mentioned schirkorr reaction are for those who pretend to extract Fe and Nd out of solution as oxides. The schirkorr reaction resembles copper hidroxide reaction into CuO but I did not test it for neodimium: the reason for getting my hybrid hidroxide contamined results

Into the dissolution of the magnets:
77g magnets where mashed with a big rock (found the big rock more worthy tham a hammer :P) and thrown into a becker with all the Ni coating.
Stoichometrically it was put to react:
0,1424Nd + 0,9971Fe + 0,0712B + 2,4214HCl --> 0,9971FeCl2 + 0,1424NdCl3 + 0,0712B (s)
295g 30 - 33% threads hydrochloric acid were used. Water was adjusted to sustain just a good saturated solution of the chlorides. By the end of the reaction the acid was weak and could not dissolve the magnet completly: this was commonly mistaken as boron...
The nanometric particles of boron for sure reacted with water, acid and everything in its path! And could now be found in solution. As said, boron in this condition would react with everything on a little extent with each, but methods would further diminish its presence later, but not completly as expected stoichometrically, again. @_@
It was not put excess acid. Before calibrating the necessary water to suspend all the precipated from the acid bath, the so formed grey-purple (yea, strong purple) precipitat diluted to join the supernatant already dark purple solution. Foaming was noticed probably due to boron - that bastard ><
This foam, or even the dark purple turned yellow-green when exposed to atmospheric air. So the solution was keep closed at most steps.
Since hydrochloric acid is a monoprotic acid it was used for direct dissolution of thew magnets because its faster and forms less ionic agregates than sulphuric acid - except by the agency of that boron!!!
It was calculated that within that mass of 77g of magnets a volume of 0,34cm³ consisting of pure boron would remain in the vessel, thats pretty a lot and had to get off.
The hidroxides were then precipitated following the reaction:
0,9971FeCl2 + 0,1424NdCl3 + 2,4214NaOH --> 0,9971Fe(OH)2 + 0,1424Nd(OH)3 + 2,4214NaCl
A little more sodium hidroxide was added to neutralize any possibly formed boric acid into borax.
Water was adjusted to 3,2L to get all borax so prepared into solution, the hidroxides decanted, the supernatant thrown away.
Again, water was adjusted to 3,2L, the hidroxides decanted and the supernatant thrown away.
It would be expected that since the hidroxides are pure nothing else would remain in the vessel... I can tell thats not true - the hidroxides were not pure.
The hidroxides were then reacted with stoichometric ammounts of sulphuric acid: First considering only ferrous sulphate would form, but then it was decidedly added more acid as to consider formation of ferric sulphate.
So obscure was the iron already and so are the neodimium compounds!
0,9971Fe(OH)2 + 0,1424Nd(OH)3 + 1,7076H2SO4(aq) --> 0,4986Fe2(SO4)3 + 0,0706Nd2(SO4)3 + H2O
Heat evolved from dissociation of sulphuric acid and reaction with the hidroxides, raising the temperature to about 60ºC.
A greyish mass with a slight pink hue remained at the bottom after decanting. The supernatant was brown but translucid and did not resemble true ferric sulphate produced with hematite: I was obviously an hybrid.
Considering sulphuric acid dissociation constants:
H2SO4 - > H+ + HSO4- 2,4x10^6
HSO4- -> H+ + SO4- 1x10^(-2)
I does not dissociate completly!
But when reacting with strong bases:
HSO4- + OH- --> SO4(2-) + H20 the equilibrium is strongly changed, favouring formation of sulphate anion.
As iron hidroxide is not a strong base :/
hence the intriguing compounds described.
The greyish pink precipitate at the bottom was not, so, anything even close to pure neodimium sulphate, but something precipitated by it. At this point the solution had a volume of about 1 Liters: more than enough to dissolve both salts if they did it in time.
The solution was allowed to settle and the supernatant recovered at another becker. This collect supernatant was boiled until reaching a volume of about 700mL and then was re-added to the becker containg the grayish pink precipitate and mixed together au bain marie, at 100ºC for 40min.
The solution was stirred and then filtrated when still very hot. The filtrate was collected and washed with boiling hot water au bain marie then filtrated again. The filtrate was collected and washed with boiling hot water au bain marie then filtrated again (third time). All filtrated liquid were kept for analysis excpet for the last.
The resulting filtrate powder was set to dry in filter paper and its appearence is always yellowish no matter what light I expose it to: seems its very contamined with iron or Pr.
final product.JPG - 19kB
Last, the first filtrated liquid kept for further analysis was brought to a boil and reduced in volume from 700mL to about 300.
After a much longer period of time than expected this brownish solution gave green crystals!!

filtrate liquid crystalization.JPG - 27kB
and
filtrate liquid crys closeup.JPG - 25kB
Those green crystals when washed gives transparent crystals. Thats absolutly a sign of contamination.
Second filtrated liquid:
second wash lquid.JPG - 20kB


I came to a conclusion that its very important to add very excess acid when making such solutions to counteract hydrolisation speed of the new formed solved ions, when dealing with weak base metals.




~~~Poppy

[Edited on 11-10-2011 by Poppy]

blogfast25 - 10-11-2011 at 05:36

Wizzard:

Elemental boron does not react at all with HCl or H2SO4, AFAIK. ‘Magnet boron’ may be partly present as borides and those would blow off as boranes, explaining that typical meaty smell you get when reacting the magnets with acid.

Thanks for the tip on removing the coating: it will be used on my next attempt.

Poppy:

I suggest using H2O2 or even HNO3 to oxidise all Fe (II) to Fe (III), instead of relying on oxidation by air oxygen.

At a guess I’d say your top photo is of Nd(SO4)3 heavily contaminated with Fe2(SO4)/Fe2O3 as main contaminant. Does it look any different under TL lighting and incandescent lighting?

I suggest trying to purify it as follows: suspend it in a little water and add some strong alkali and simmer it: the alkali converts the Nd sulphate to Nd(OH)3.nH2O. Filter and wash cake properly. Dissolve the cake in the minimum quantity of HCl. There’s no need to filter. You now have a contaminated NdCl3 solution.

Separately prepare quite a quite large amount of fresh, thoroughly washed Fe2O3.nH2O and add some of it to the Nd bearing solution. Some of it will dissolve, then add some more. Allow standing overnight, making sure that some of the Fe2O3.nH2O doesn’t dissolve. You should thus obtain a NdCl3 containing suspension of Fe2O3.nH2O that is almost neutral (pH ≈ 7).

Filtering this solution you should find ALL the NdCl3 in the filtrate and ALL of the Fe (III) in the filter cake.

Poppy - 10-11-2011 at 06:44

I'll try purifying the yellow mass I got from picture above but I think the major other part is still inside the vessel with brown liquid and green crystals. The crystals have grown even more and I'll post a picture of them soon.

Blogfast,

you suggest some species might be hidrated iron oxide?
(Fe2O3 .nH2O or FeOOHHOH and all the heck of strange formulas iron can get in such conditions.) but I cant find a specific way to afford Fe2O3.nH2O

Between trying to react hematite with water and conc. alkali but nothing occured.

I will verify your method, thanks for advice

~~~Poppy

I need a credit card again so I can visualize internet books
After testing the method you said I guess I'll have to calcinate everything and start reaction all over, at least the componentes should be boron free by now.

[Edited on 11-10-2011 by Poppy]

blogfast25 - 10-11-2011 at 06:58

Quote: Originally posted by Poppy  
you suggest some species might be hidrated iron oxide?
(Fe2O3 .nH2O or FeOOHHOH and all the heck of strange formulas iron can get in such conditions.) but I cant find a specific way to afford Fe2O3.nH2O

Trying to react hematite with water and conc. alkali but nothing occured.

Iwill verify you rmethod, thanks for advice

~~~Poppy


Fe(OH)3 has a very small solubility product (Ks about 10<sup>-39</sup>;) and ferric ions hydrolyse easily (much more easily than Nd3+), so yes, hydolysed species are likely to contaminate your (top photo) sample.

Hematite (Fe3O4) does not react with alkali AT ALL. Iron oxides (II and III) are not oxoacid formers and cannot react with alkalis.

I think most of your Nd is in the stuff shown in the top photo: even the quantity (for 75 g magnets) seems to roughly fit the observation.

Mixing fresh ferric hydroxide with the dissolved contaminated sample, making SURE there's Fe(OH)3.nH2O/Fe2O3.nH2O left over should cause any Fe3+ in that solution to precipitate as Fe(OH)3.nH2O/Fe2O3.nH2O. That's what my preliminary test showed. The method is also used for cleaning Mn2+ solutions from Fe3+ contamination.

Poppy - 10-11-2011 at 07:07

Fe(OH)3.nH2O/Fe2O3.nH2O is common rust found from oxydised iron metal.
Preparation of this would require reacting pyrophoric iron with water. Otherwise large iron metal clumps would retain its metallic interior.
Not confident enough on my yellow mass content I will try this method soon.

[Edited on 11-10-2011 by Poppy]

blogfast25 - 10-11-2011 at 11:48

Quote: Originally posted by Poppy  
Fe(OH)3.nH2O/Fe2O3.nH2O is common rust found from oxydised iron metal.
[Edited on 11-10-2011 by Poppy]


No, no, no, I'm talking precipitated Fe(OH)3. Take any iron salt, dissolve it in water and (carefully!) add peroxide to it. Then cool, dilute somewhat and add alkai (like NaOH) to precipitate Fe(OH)3. Wash profusely to get rid of Na+, OH- and other anions.


MrHomeScientist - 10-11-2011 at 13:39

I'm really interested in this iron hydroxide method!

I also tried it last night on test tube scale, and I don't have my notes on me but here's a quick rundown.

I took 3mL of the magnet solution and added NaOH solution to it until no more precipitate formed. The precipitated hydroxides were initially green, and any exposed to air (e.g. on the filter paper) quickly turned brown. This is surely iron(II) hydroxide being oxidized to iron(III) hydroxide. (Remember that my magnet solution at this point contains both iron(II) and iron(III) sulfate, as well as Nd sulfate.) I filtered and washed the precipitate of hydroxides. While still wet, I scooped this into a test tube of 6mL of magnet solution and stirred it around a bit. I also added a few mL of water to rinse the goopy hydroxides off the side of the tube and down into the solution. After letting this sit and settle overnight, the color of the solution is a much lighter pink than it initially was. This of course might just be due to dilution from adding the extra water, though.

I then realized that I don't have any thiocyanate to use to see if anything actually happened! Are there any other methods to test for iron(III)? I'll try testing for iron(II) with ferricyanide like I did before, while I'm at it, and report back. Hopefully, if it turns out to be free of any iron, that would mean it's not necessary to convert everything to iron(III) first.

blogfast25 - 10-11-2011 at 14:51

Hi MrHS:

The iron HAS to be present as Fe (III) because FE9OH)2 isn't as insoluble.

To check for Fe w/o KSCN, add some adic and a piece of Zn or Al: both reduce Fe (III) to Fe (II), then test with ferricyanide.

I'm running a larger test with this ferri hydoxide method overnight: fingers crossed! ;)

[Edited on 10-11-2011 by blogfast25]

Poppy - 10-11-2011 at 17:51

Gentleman, lend me this opportunity to correct myself:

Fe(OH)3.n H2O / Fe2O3n.H2O IS NOT RUST

It is in fact a byproduct of the reaction of FeIII ions into FeIII hydroxide followed by the autoprotolysis of hydroxide molecules

(OH)(-) + (OH)(-) --> (O)(2-) + H2O

2Fe(3+) + 3O(2-) --> Fe2O3


blogfast25 - 11-11-2011 at 12:37

Quote: Originally posted by Poppy  
Gentleman, lend me this opportunity to correct myself:

Fe(OH)3.n H2O / Fe2O3n.H2O IS NOT RUST

It is in fact a byproduct of the reaction of FeIII ions into FeIII hydroxide followed by the autoprotolysis of hydroxide molecules

(OH)(-) + (OH)(-) --> (O)(2-) + H2O

2Fe(3+) + 3O(2-) --> Fe2O3



Poppy: you're looking too far afield: there are no O<sup>2-</sup> in solution.

Fe(OH)3 (Fe2O3.nH2O is just a different notation, really) forms most easily by adding alkali to a Fe3+ bearing solution:

Fe<sup>3+</sup>(aq) + 3 OH<sup>-</sup>(aq) + n H2O(l) === > Fe(OH)<sub>3</sub>.nH<sub>2</sub>(s)

£$$%&*())__%^&*(^%$£$

Well, this separation method based on adding ferric hydroxide to the Nd3+/Fe3+ solution works well. I’m inclined to say ‘it rocks!’, actually.

Two such separations were carried out on 2 lots of about 250 ml of ferric contaminated Nd3+ solutions, both with great success. In both cases the pH went up slowly to about 4 after addition of the ferric hydroxide (in both cases the same lot of ferric hydroxide was used). The supernatant solution then filtered to clear, testing negative for Fe3+ (KSCN) and very positive for Nd3+ (against oxalic acid).

This is roughly in line with Ksp = [Fe3+] x [OH-]<sup>3</sup> ≈ 10<sup>-39</sup>: at pH = 4; [OH-] = 10<sup>-10</sup> and thus [Fe3+] ≈ 10<sup>-9</sup>. Of course there are some subspecies at play like FeOH<sup>2+</sup> and maybe Fe(OH)<sub>2</sub><sup>+</sup>.

A few tips:

• Once the pH drop below 3, the [H2O+] concentration becomes low (< 0.001 M) and reaction slows down a lot. Help reaction along by heating the slurry and keeping it warm for a bit.
• Heating also helps the Fe(OH)3 particles to grow, allowing for faster filtering.
• Don’t try and squeeze the last bit of Nd3+ from the filter cake: Fe(OH)3 has a tendency to peptise at low ionic strengths: it then starts running through the filter as a colloid.

Then another batch of 4 magnets, 27.7 g by weight were dissolved in 122 ml of HCl 36 %, in a 1 L beaker on a low hot plate setting. After reaction had stopped, the solution was diluted to about 200 ml and filtered. The latter may not even be necessary here but because I didn’t get the cover off I didn’t want further contamination with Cu2+.

The solution was then iced and slowly oxidised on icebath with 100 ml of 9 % H2O2. It was diluted a bit more and then simmered to decay excess H2O2. At that point large amounts of ferric hydroxide started to precipitate: hydrolysis is pushed by heat. The pH was till low and I added the same ferric oxide slurry recovered from the filter of the last separation and kept the solution warm for some time. It’s now cooling and I will read pH tomorrow: at pH >= 4 the separation should be complete!


[Edited on 11-11-2011 by blogfast25]

 Pages:  1    3  ..  6