Sciencemadness Discussion Board - The trouble with neodymium...

Chelating agents like EDTA are quite non-discriminating: they form strong complexes with many (most, in fact) non-alkali metals, including Fe and Nd.

U and Nd being in the same column is really a meaningless coincidence. Nd belongs to the First f-block (Lanthanides), which show very strong inter-group chemical similarities. U belongs to the Second f-block (Actinides), which show also strong inter-group chemical similarities (but less so than Lanthanides).

Due to their electronic configuration structures, both groups also show similarities between each other.

Similarity between U and Nd however is really quite weak. Yes, both form EDTA complexes but then so do... erm... Al and Pb! U forms oxidation states from +2 to +6 (but +3 isn't very stable), Nd only does +3. U forms very stable complex anions and cations, Nd only forms complexed Nd(+3) cations.

By contrast, the Fe/Nd separation methods presented and tested to destruction in this thread are simple to use and require mostly only OTC chemicals. That's why we chose them...

[Edited on 11-6-2015 by blogfast25]

would the other method work? Using CaF instead of NH₃HF?

blogfast25 - 11-6-2015 at 06:29

Quote: Originally posted by DalisAndy

would the other method work? Using CaF instead of NH₃HF?

Huh? CaF2 (NOT CaF), aka Fluorite, is water insoluble.

MrHomeScientist - 12-6-2015 at 05:33

From Andy's first link:

"Neodymium is found most often in nature in Misch metal, monazite sand, and the mineral bastnasite. Due to the fact that these minerals and compounds also contain lanthanides and other rare earth elements, it is difficult to isolate neodymium. The process of neodymium isolation and extraction is highly complex; due to the complexity of the process, neodymium is never isolated on a small scale laboratory basis."

We'll see about that...

I'll show them! I'll show them all!!

blogfast25 - 12-6-2015 at 06:21

From Andy's first link:

"Neodymium is found most often in nature in Misch metal, [...]

HAHAHAHHAHHHAHAHAHHHAHHHAAHHHAHAHAH!

MrHomeScientist - 12-6-2015 at 09:20

Totally missed that one! I was suspicious of the 'science fair' website though.

blogfast25 - 12-6-2015 at 10:14

Totally missed that one! I was suspicious of the 'science fair' website though.

"And here, next to my native gold and silver, is some of that native Misch metal y'all wanted to see! I bought it off an anonymous Craiglist seller, years ago."

The Volatile Chemist - 13-6-2015 at 12:47

Quantification of legitimacy involves an index of illegitimate conjunctions and its <->it's errors within a given paper.

DalisAndy - 21-8-2015 at 07:54

How would one isolate the fluoride salt from the bisulfate?

MrHomeScientist - 21-8-2015 at 09:53

What fluoride from what bisulfate? I don't think anyone ever used bisulfate in this thread. Clarify what you mean.

While I'm here, I'll mention that I'm gearing up to get back on track with this experiment. I sincerely hope to have something to report by the end of the weekend.

DalisAndy - 25-8-2015 at 07:06

How did you guys separate the neodymium fluoride and ammonium sulfate?

gdflp - 25-8-2015 at 08:09

NdF₃ is insoluble in water while (NH₄)SO₄ is soluble in water as well as NH₄F and Nd₂(SO₄)₃.

MrHomeScientist - 25-8-2015 at 10:12

A fact you would have known had you (gasp!) READ THE THREAD.

[Edited on 8-25-2015 by MrHomeScientist]

aga - 25-8-2015 at 12:31

Reading this thread takes more than the standard 8 second attention span of modern people.

Please condense it to a photo, a few smileys and an instant answer (plus a 'like' button) to make it suitable for SmartPhone users.

Something like this :-

Magnet -> Pure Neodymium

The Volatile Chemist - 30-8-2015 at 13:44

Haha, indeed. I'm sure the information will be condensed into a more readable format. Until then, it is the responsibility of the reader not to be lazy.

CrossxD - 12-2-2016 at 13:46

maybe you can percipitate iron using aluminum you will end up with Nd III+ and Al III+ then you add NaOH aluminum will stay in solution as NaAl(OH)4 and neodymium will percipitate as Nd(OH)3......teoreticaly

[Edited on 12-2-2016 by CrossxD]

IrC - 12-2-2016 at 14:18

From Andy's first link:

"Neodymium is found most often in nature in Misch metal, [...]

HAHAHAHHAHHHAHAHAHHHAHHHAAHHHAHAHAH!

I had to laugh when I read that as well. I have a few 1 lb ingots of it, they are amazingly clean considering they were dug up in mines. My biggest dilemma is figuring out what process underground shapes them into such nice ingot appearing forms. Anyone know the answer to this?

blogfast25 - 12-2-2016 at 15:19

Quote: Originally posted by IrC

My biggest dilemma is figuring out what process underground shapes them into such nice ingot appearing forms. Anyone know the answer to this?

You're wrong. They grow on trees, actually.

Atrum - 11-7-2016 at 08:17

I recently joined the fun in extracting Neodymium compounds from rare earth magnets. I have had decent success in isolating Nd sulfate, and oxalate (the easy one).

My initial process was the sulfuric acid separation. The magnet was a small 1" cube that had no nickel plating on it. It was old and the plating had a chip in it, and so over the last 5 years had completely come free from the plating.

Like a noob I forgot to weigh the magnet before proceeding.

1. Dissolved in 30% H2SO4. No heating was involved in an attempt to limit any possible reaction with boron. This took about a week to fully dissolve.

2. I oxidized the resulting mixture with 30% H2O2 until no further reaction could be seen.

3. I added a 2 molar solution of NaOH to the magnet soup and precipitated the insoluble hydroxides.

4. Washed the hydroxides with distilled water.

5. After filtering the hydroxides off, I reacted the brown sludge with 50% Sulfuric acid. It dissolved but it left a very fine tan precipitated of something. I was not sure whether this was an Nd compound or an iron compound. It was completely insoluble in water. So I used some conc. HCl and obtained a bright yellow solution. This suggests iron chloride I believe.

Based on that my guess as to what the insoluble compound was is Iron hydroxide sulfate (Fe(OH)(SO4)). I could be completely wrong about that though.

6. I tested the supernatant for iron using salicylic acid (I did not have any Potassium Ferricyanide). It was positive. I then tested for Nd using oxalic acid. Also positive.

7. I boiled the supernatant which darkened in color upon heating which is characteristic of iron compounds being thermochromic. After removing nearly 500mL of water, I started to see some pink crystals. I filtered these off and dried them. The crystals were not quite the right color (under any lighting conditions) and some recrystallizing is needed.

The Nd oxalate 1 was first oxidized to remove any iron(II). The Nd oxalate 2 was not oxidized and resulted in a very yellow powder of Nd oxalate.

[Edited on 7-11-2016 by Atrum]

blogfast25 - 11-7-2016 at 09:41

@Atrum:

Welcome to the club.

Atrum - 11-7-2016 at 14:21

@Atrum:

Welcome to the club.

Thanks.

I was wondering if you could provide any insight to the precipitate I mentioned in my post?

Did you come across anything like that during your processes?

blogfast25 - 11-7-2016 at 16:47

I was wondering if you could provide any insight to the precipitate I mentioned in my post?

Did you come across anything like that during your processes?

Quote:

5. After filtering the hydroxides off, I reacted the brown sludge with 50% Sulfuric acid. It dissolved but it left a very fine tan precipitated of something. I was not sure whether this was an Nd compound or an iron compound. It was completely insoluble in water. So I used some conc. HCl and obtained a bright yellow solution. This suggests iron chloride I believe.

Most likely a neodymium sulphate. Many of us have noticed there are several forms of Nd sulphate hydrate.

To check, isolate precipitate and treat it with strong NH₃ and some time. Wash carefully and dissolve precipitate in HCl (NOT sulphuric), then test solution with oxalic acid for Nd. I'd put good money on a positive Nd test, executed that way.

[Edited on 12-7-2016 by blogfast25]

Atrum - 12-7-2016 at 15:02

@blogfast25

Thanks, I will give that a go.

Here is a photo of the precipitate that I centrifuged to wash it. Filtering was impossible. Very Very fine precipitate. It goes through filter paper like water through a sieve.

blogfast25 - 12-7-2016 at 15:17

@blogfast25

Thanks, I will give that a go.

Here is a photo of the precipitate that I centrifuged to wash it. Filtering was impossible. Very Very fine precipitate. It goes through filter paper like water through a sieve.

Well, you you've got enough for that test.

Atrum - 12-7-2016 at 15:21

Well, you you've got enough for that test.

Yeah, that is just one tube out of 4 that have the same amount. Hopefully I will be able to extract the rest of the nd from it. I am hoping for some large crystals of the sulfate eventually.

elementcollector1 - 13-7-2016 at 00:13

This was probably obvious to everyone else, but I felt it was worth noting:

-I've left a solution of oxalic acid and precipitate of iron/neodymium oxalate to chelate the iron out for over a year now. Not much happened, so I thought to just let time do the work.

-Put in some concentrated hydrogen peroxide yesterday, after seeing other users' experience with oxalates. Within 24 hours, most of the iron had leached out of the precipitate, leaving the supernatant solution a vivid green and the precipitate an increasingly lighter shade of whitish-pink.

Given that I started with about 8 hard drive magnets, this leaves me with quite a bit of oxalate to work with. Hopefully I can calcine it at a low enough temperature that it won't become inert like the previous stuff...

Atrum - 13-7-2016 at 10:06

Quote: Originally posted by elementcollector1

This was probably obvious to everyone else, but I felt it was worth noting:

-I've left a solution of oxalic acid and precipitate of iron/neodymium oxalate to chelate the iron out for over a year now. Not much happened, so I thought to just let time do the work.

-Put in some concentrated hydrogen peroxide yesterday, after seeing other users' experience with oxalates. Within 24 hours, most of the iron had leached out of the precipitate, leaving the supernatant solution a vivid green and the precipitate an increasingly lighter shade of whitish-pink.

Given that I started with about 8 hard drive magnets, this leaves me with quite a bit of oxalate to work with. Hopefully I can calcine it at a low enough temperature that it won't become inert like the previous stuff...

So, I am curious. How many different solvents have you tried when trying to dissolve Nd oxalate?

elementcollector1 - 13-7-2016 at 10:35

So, I am curious. How many different solvents have you tried when trying to dissolve Nd oxalate?

None but water, actually. I didn't figure oxalate would be soluble in much. At any rate, my pink neodymium oxalate is currently filtering out, and looking very nice while doing so.

I do still have the results of my attempt to dissolve the previous, calcined stuff - the precipitate's turned a sandy pink, and the solution itself is quite clear. Could this be neodymium sulfate, or just 'wet', unreacted neodymium oxalate? Hard to tell.

blogfast25 - 13-7-2016 at 11:12

So, I am curious. How many different solvents have you tried when trying to dissolve Nd oxalate?

It's not really a question of 'How many different solvents have you tried'.

1. Nd oxalate is supremely insoluble, that makes it hard to dissolve in anything including strong acids.

2. Oxalic acid is not volatile: no matter what acid you throw at it, the oxalate ions are always there.

One's 'best bet' is to transform the oxalate, e.g. by oxidising it. Permanganate oxidises H₂Ox to CO₂ (leaving you also with Mn(II), of course).

Hypochlorite (bleach) might also do it but would probably slow because thin bleach contains only a few w% of ClO^-.

[Edited on 13-7-2016 by blogfast25]

Atrum - 13-7-2016 at 16:46

So, I am curious. How many different solvents have you tried when trying to dissolve Nd oxalate?

That sounds like it could be interesting to try out. I appreciate the information blogfast.

blogfast25 - 13-7-2016 at 17:30

This source:

http://www.nrcresearchpress.com/doi/pdf/10.1139/v61-211

... indicates oxidation of oxalate with HClO (acidic conditions) is possible, with high activation energy (heating required, basically):

Quote:

There is some indication that, if the reaction is relatively exothermic, the activation energy is lower. This is perhaps to be expected, since if the new bond to oxygen is strong, this would tend to lower the energy of the activated complex as it is formed. However, this trend is not universal; for instance, the very exothermic oxidation of oxalate [by HClO] has quite a high activation energy. This particular reaction may be a little different from the others, since it is apparently not an oxygen transfer but an electron transfer; although it could be formulated either way.

A combination of bleach and a weak acid (like vinegar) + some heat and time, might just slowly dissolve an insoluble oxalate to chloride/acetate.

[Edited on 14-7-2016 by blogfast25]

elementcollector1 - 19-7-2016 at 09:39

A picture of the neodymium oxalate:

Still has oxalic acid crystals in it, but I don't particularly mind - those should simply decompose upon heating, right?

I think I'll take this to the university this fall and calcine it there, with their temperature-controlled furnace - what's the lowest temperature I can do this at? The furnaces typically run for 24 to 48 hours, so time is not imperative.

The Volatile Chemist - 22-7-2016 at 11:55

That's quite a lot...is the neodymium oxalate itself very crystalline, or just the oxalic acid mixed in with it - it's hard to tell in the picture.

Atrum - 22-7-2016 at 12:18

I am slowly building up a good amount of Nd Oxalate. It seems to me that it is the best method to get it free of iron contamination. I am still playing around though.

On the side I am trying to get a crop of nice green Iron(III)Oxalate crystals. I find the thermochromism of iron oxalate to be interesting. Bright Green when cold, brownish red when hot.

elementcollector1 - 28-7-2016 at 21:32

I'm considering using sodium metal to reduce my neodymium fluoride. This would presumably be rammed into a graphite crucible, followed by the neodymium fluoride, followed by external heating from a propane blowtorch. Does anyone see any problems with this method?

MrHomeScientist - 1-8-2016 at 06:16

I did some calculations a while back that suggested sodium won't work as well. It's probably back a few pages in the thread somewhere, so here they are again:

∆H values:
NdF3 = -1657 kJ/mol
MgF2 = -1124..2 kJ/mol
NaF = -576.6 kJ/mol
LiF = -616.0 kJ/mol

For Mg
2NdF3 + 3Mg = 3MgF2 + 2Nd
∆Hrxn = 3(-1124.2) – 2(-1657) = -58.6 kJ/mol

For Na
NdF3 + 3Na = 3NaF + Nd
∆Hrxn = 3(-576.6) – (-1657) = -72.8 kJ/mol

For Li
NdF3 + 3Li = 3LiF + Nd
∆Hrxn = 3(-616) – (-1657) = -191 kJ/mol

So lithium produces quite a bit more heat than the others, though I’m not sure how that number corresponds with actual reaction temperature. It at least suggests that Li should favor the easiest reaction conditions.

Also I built my setup to deliver inert atmosphere to the crucible, mostly because I'll be using lithium but also to protect the newly formed Nd from the air. It's possible it will sink to the bottom of the melt and be protected anyway, but it's something to consider.

God I need to finish this project. I finally have all the pieces in place and can proceed at any time. I just need to get off my ass and do it.

blogfast25 - 1-8-2016 at 09:11

So lithium produces quite a bit more heat than the others, though I’m not sure how that number corresponds with actual reaction temperature. It at least suggests that Li should favor the easiest reaction conditions.

Also I built my setup to deliver inert atmosphere to the crucible, mostly because I'll be using lithium but also to protect the newly formed Nd from the air. It's possible it will sink to the bottom of the melt and be protected anyway, but it's something to consider.

God I need to finish this project. I finally have all the pieces in place and can proceed at any time. I just need to get off my ass and do it.

Li is the go to reducing agent in this case.

Na might work but will need stronger heating.

Both will need strong heat + inert blanket (Ar), to obtain the Nd as a regulus.

Powdered Nd in a alkali metal fluoride matrix will be inseparable, unless you heat to to well above 1000 C: the liquid Nd will then separate out.

No matter how you play it, you'll need to heat to close to 1000 C, with adequate Ar blanket.

It might be tempting to remove heat after reaction has taken place but it's probably not a good idea, if you want adequate metal/slag separation.

[Edited on 1-8-2016 by blogfast25]

MrHomeScientist - 1-8-2016 at 09:36

The last problem I had to solve was connecting the Ar tank to "The Gavel" (my inert atmosphere crucible). For some reason the fittings for argon tanks don't connect with regular pipe, so I had to order a special adapter.

The next problem I foresee is being able to heat to the proper temperature. The Gavel won't fit in my mini propane furnace, so I bought some kaowool to wrap it in. If that won't hold in heat well enough, I'll have to spring for a larger, perhaps custom-made, furnace setup.

Dan Vizine - 2-8-2016 at 12:27

I've been constructing furnaces that operate to 1200 C with ceramic fiber board, Alumel wire, some high-temp. ceramic adhesives or (even furnace cement) and ceramic coil forms for many years. I'd estimate you could buy everything needed to bring a 200 to 300 cubic inch space to 1200 C for $150. Control is a whole different issue. I bought a $25 PID controller on eBay and a $10 high current solid state relay for $10. I had a K thermocouple just laying around. This will give you a nice furnace with years of service. Of course, gas-heated rigs are cheaper but control is limited and of course, there are exhaust issues...

If you seriously consider the electric option and want some pointers, I've made all the mistakes that you can make, I've got it down to a reliable formula.

MrHomeScientist - 10-8-2016 at 09:57

I'd love to learn from your experiences. An electric furnace would be a very valuable thing to have, not just for this particular experiment.

Brain&Force - 20-8-2016 at 21:31

I think I may have accidentally come across a much better separation method for neodymium from magnets.

I've been experimenting with gadolinium, terbium, and holmium acetates, which are conveniently non-hygroscopic but very soluble in water. They are also quite poorly soluble in ethanol, though crystals of the acetates will fall apart in it.

Contrast this behavior with boric acid, iron(III) acetate and nickel acetate, which are highly soluble in ethanol.

I recommend you test this method (I'm about to test this myself). Dissolve a neodymium magnet in acetic acid and hydrogen peroxide. Remove the supernatant and evaporate it to dryness. Redissolve the material in ethanol. The neodymium acetate should remain as a powder. Filter with ethanol rinses and you should get quite pure neodymium acetate (of course there could be other lanthanides).

I haven't tried this myself, but theoretically it should get rid of nickel, iron, and boron in the magnets.

j_sum1 - 20-8-2016 at 23:03

That's some good thinking B&F. I like where that is going.
I have been pondering recently how easy it is too overlook acetic acid for reacting with metals. Sure, it tends to be slow but it has a lot of plus sides.

Brain&Force - 23-8-2016 at 09:50

Yeah, the non-hygroscopic salts but very water-soluble salts are the biggest advantage IMO. Just a quick note - keep the concentration of acetic acid low. I've found that highly concentrated acetic acid will not solubilize the acetates, and that glacial acetic acid does not appreciably attack holmium at room temperature for thirty minutes. (I suggested hydrogen peroxide to keep iron as iron(III) acetate which retains extremely high solubility in ethanol). I think iron(II) acetate will also be soluble in ethanol, so that should be another bonus.

Also, I think you can form a soluble neodymium acetate ethanol complex with strong heating, so I'd avoid heating it while it's in ethanol. That is definitely the case in methanol for dysprosium acetate.

elementcollector1 - 26-11-2016 at 15:27

Okay, time to try this again. Me and Brain&Force have been discussing possible alternative amateur routes to pure neodymium salts for a while now, and while I ran an earlier attempt this year, it culminated just a week ago in utter failure. So, this time, I'm going to post my method(s) beforehand, in the hopes that somebody will be able to spot problems before they occur.

Also, this might turn into my senior thesis.

The goal of this series of tests is to identify which chemical separation route provides the most promising results for the amateur when trying to isolate neodymium compounds from magnets. Several separation methods will be evaluated, consisting of the potassium oxalate route, the ethanol route, and the selective precipitation route. From there, the oxalate will be further tested for its solubility in ionic neodymium(III) salts, as according to this source.

The reaction idea is as follows:

First, a 2"-by-2"-by-1" neodymium magnet (approximate formula Nd₂Fe₁₄B) is heated to 500 degrees Fahrenheit (approx. 260 degrees Celsius) in a standard convection oven for 1 hour to at least partially demagnetize it (for ease of processing). Once done, the nickel casing is stripped away as best as possible, leaving only the sintered material underneath. The approximate weight of this material, assuming none has been lost so far, should be 478.5 to 491.6 grams, assuming the reported density of 7.3-7.5 grams/centimeter cubed is correct. This results in an estimated neodymium content of 127.7 to 131.2 grams.

This material is then dissolved in an excess of 25% acetic acid (diluted from glacial acetic acid) to form green iron(II) acetate tetrahydrate, Fe(C₂H₃O₂)₂(H₂O)₄, pale mauve neodymium(III) acetate tetrahydrate, Nd(C₂H₃O₂)₃(H₂O)_x, and an unknown boron product (possibly elemental boron, 'boron acetate,' or boric acid, or a mixture). Assuming the boron product is mostly insoluble, it is filtered off, along with any plastic or organic binding material that may have been used in the manufacture of the magnet.

Nd₂Fe₁₄B (s) + HC₂H₃O₂ (aq) -> Fe(C₂H₃O₂)₂(H₂O)₄ (aq) + Nd(C₂H₃O₂)₃(H₂O)₄ (aq) + (unknown boron compound, likely boric acid?)

Finally, this mixture is treated with an excess of approx. 11-14% hydrogen peroxide solution, which has been freeze-distilled from topical grade 3%. This oxidizes the iron(II) acetate to iron(III) acetate, also known as basic iron acetate, Fe₃O(C₂H₃O₂)₆(H2O)₃]C₂H₃O₃. The hydrogen peroxide does not affect the neodymium(III), because it is a more stable ion.

Fe(C₂H₃O₂)₂(H₂O)₄ (aq) + Nd(C₂H₃O₂)₃(H₂O)₄ (aq) + H₂O₂ (aq) -> Fe₃O(C₂H₃O₂)₆(H2O)₃]C₂H₃O₃ (aq) + Nd(C₂H₃O₂)₃(H₂O)₄ (aq)

The solution thus far is split into three equal parts.

The first third of the solution is mixed with an excess of aqueous potassium oxalate (K₂C₂O₄) to precipitate neodymium oxalate (Nd₂(C₂O₄)₃), chelating the iron(III) into soluble, vivid green potassium ferrioxalate (K₃(Fe(C₂O₄)₃)₃). The pale pink neodymium oxalate is filtered off, washed, and dried to remove any excess water.

3 Fe(C₂H₃O₂)₂(H₂O)₄ (aq) + 2 Nd(C₂H₃O₂)₃(H₂O)₄ (aq) + 6 K₂C₂O₄ (aq)-> Nd₂(C₂O₄)₃ (s) + K₃(Fe(C₂O₄)₃)₃ (aq) + 3 KC₂H₃O₂ (aq)

(This reaction is not properly balanced.)

This oxalate is then split once again into three equal parts by mass. The first part is calcined with a torch, forming pale blue neodymium oxide (Nd₂O₃) and carbon dioxide according to the following reaction:

2 Nd₂(C₂O₄)₃ + 3 O₂ -> 12 CO₂ + Nd₂O₃

This is weighed to determine the yield of Nd, assuming a perfect 1/6 of the Nd from the magnet was in the starting oxalate. Theoretical perfect yield, therefore, is approximately 24.85 to 25.5 grams of neodymium oxide, representing 21.3 to 21.9 grams of neodymium.

The second portion of the oxalate is weighed and placed in a pre-prepared solution of aqueous neodymium chloride, and left to dissolve. Because this test is more qualitative than quantitative, results will depend on how much, if any, of the oxalate remains.

The third portion, as with the second, is weighed and placed in a pre-prepared aqueous solution of pure neodymium acetate, to see if a similar complexing reaction occurs.

Returning to the second portion of the mixed iron and neodymium acetates, it is boiled to dryness and mixed with ethanol. According to Brain&Force's suggestion, this should result in the iron(III) acetate dissolving in the ethanol, leaving behind insoluble neodymium acetate tetrahydrate. This acetate is then dried, weighed and compared to the theoretical yield of 128.4 to 132.0 grams. This is then dissolved in water, and tested with thiocyanate for iron content.

The third and final portion of the initial mix of iron and neodymium acetates is subjected to NurdRage's method of selective precipitation, to see if it works between iron and neodymium. This would be done by taking about half of the solution and precipitating out the hydroxides with NaOH, followed by adding the washed hydroxide mixture back into the other half of the solution. The reactions are as follows:

Fe₃O(C₂H₃O₂)₆(H2O)₃]C₂H₃O₃ (aq) + Nd(C₂H₃O₂)₃(H₂O)₄ (aq) + NaOH (aq) -> Nd(OH)₃ (s) + Fe(OH)₃ (s)+ NaC₂H₃O₂ (aq)

Nd(OH)3 (s) + Fe₃O(C₂H₃O₂)₆(H2O)₃]C₂H₃O₃ (aq) -> Fe(OH)3 (s) + Nd(C₂H₃O₂)₃(H₂O)₄ (aq)

While this has never been tested to my knowledge, the difference in electropositivity between iron and manganese (+0.28) is less than the difference in electropositivity between iron and neodymium (+0.69), which from my admittedly ignorant interpretation means that this reaction will go forward with greater ease than its manganese counterpart. If the reaction does indeed work, then this would be an easy alternative to both the oxalate and the ethanol route for separating iron from neodymium. The resultant solution would be filtered, and tested with thiocyanate for presence of iron.

The eventual goal, once all has been said and done, is to convert any soluble neodymium salts produced into NdF₃, which will be then converted into Nd metal.

Questions so far:

-Why does basic iron acetate have such a ridiculously complex formula?
-Is there anything in this post that strikes you as partially or completely incorrect?

[Edited on 11-26-2016 by elementcollector1]

MrHomeScientist - 28-11-2016 at 12:08

Some comments:

To demagnetize the magnets, all that's needed is to blast them with a propane torch for a minute or two until they separate from the metal tongs you're holding them with. But perhaps you wanted a more repeatable process for your procedure.

I don't believe 25% acetic acid would do anything to elemental boron, so your first step is probably just B as a product.

Once you've made Nd2O3, it will likely be difficult to transform it into anything else. I seem to remember Woelen saying his commercial sample was impervious to just about everything.

This is a great project for a thesis. So many things to experiment with and test. The ultimate extraction method would be from the paper I've attached: "Direct Extraction and Recovery of Neodymium Metal from Magnet Scrap". The apparatus is a little tough to build though!

Attachment: Direct Extraction and Recovery of Neodymium Metal from Magnet Scrap.pdf (257kB)
This file has been downloaded 814 times

Every time I see a post in this thread I get a little kick to finish my part of the experiment. I need to beat everyone else to Nd!

elementcollector1 - 28-11-2016 at 16:49

Sadly, I lack a propane torch. One of the many drawbacks to trying to do chemistry while at university...

In that case, the boron can be easily filtered off and saved for later.

I actually have a commercial sample on the way as we speak, so that I can give comparisons between what I isolate from the magnets and the 'pure' substances. From what I've studied, the trend seems to be that the reactivity largely depends on the thermal history - the lower the decomposition temperature of the oxalate, the better.

That paper's fascinating, but you're right, the setup's nearly impossible! It requires molten magnesium, for one thing, which I'm almost certain is pyrophoric, and for another, it requires essentially that the magnesium/neodymium mix be distilled in a stainless steel container. I bet the apparatus could be simplified, and if I ever get my furnace working it's definitely something to try, but I doubt this is the most practical route for an amateur. On a related note, though, how soluble are the other lanthanides in molten magnesium?

Tomo - 28-11-2016 at 23:41

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

Same thing here, I think I need some stronger beakers. Did you give it another try?

[Edited on 29-11-2016 by Tomo]

elementcollector1 - 2-12-2016 at 20:28

Also, quick question about reducing NdF₃: Wasn't calcium metal suggested at one point as a suitable alternative?

HoF (CaF2) = -1228 kJ/mol (Wolfram Alpha)

2 NdF3 + 3 Ca -> 2 Nd + 3 CaF2: -370 kJ/mol, more than the equivalent with lithium. Is there some reason this is not preferred that I missed? I read over the entire thread, but couldn't seem to find it. Is it due to calcium fluoride's high melting point?

elementcollector1 - 6-12-2016 at 12:40

Update: Just got my sample of pure neodymium oxide from Snaucke Elements on EBay. It arrived as a white/gray powder, with distinct pink tinges under incandescent light.

Most of the sample was then mixed with diluted acetic acid and boiled, in the hopes that this would dissolve the oxide. To my surprise, this worked!

Once the bubblegum-pink neodymium acetate had crashed out, I then decided to test if it dissolved in ethanol. The results gave some important clues:

-The filtrate upon first coming out was the same dark pink as the concentrated aqueous solution. This is presumably due to residual water in the filtered acetate.

-As time went on, however, the filtered liquid became more dilute in coloration, and the drops coming out of the bottom of the filter paper were clear in coloration. The solid acetate on top retained mostly the same volume, though it had largely disintegrated into a paste. This indicates that the solid neodymium acetate is not soluble in ethanol, making separation of iron by selective dissolution in ethanol a viable procedure, provided the iron/neodymium acetates are boiled to dryness.

EDIT: Well, what do you know! The ethanol and water ended up separating just a few minutes after I made this post, and something cloudy has formed at the interface.

[Edited on 12-6-2016 by elementcollector1]

Velzee - 6-12-2016 at 12:49

Yay!

Dan Vizine - 23-5-2017 at 10:42

Quote: Originally posted by elementcollector1

Also, quick question about reducing NdF₃: Wasn't calcium metal suggested at one point as a suitable alternative?

HoF (CaF2) = -1228 kJ/mol (Wolfram Alpha)

2 NdF3 + 3 Ca -> 2 Nd + 3 CaF2: -370 kJ/mol, more than the equivalent with lithium. Is there some reason this is not preferred that I missed? I read over the entire thread, but couldn't seem to find it. Is it due to calcium fluoride's high melting point?

I'm confident that you're absolutely right and that it's superior to commercial lithium.

MrHomeScientist - 4-10-2017 at 10:44

The reason lithium is preferred is its fluoride's melting point. Ideally the reaction is run similar to a thermite, where all products are in the liquid phase. This affords good metal-slag separation.
Melting points:
Nd - 1021 C
LiF - 845 C
CaF2 - 1418 C

The reaction's target minimum running temperature is one where everything is molten, thus using LiF reduces this requirement to 1021C.

The other consideration is reaction enthalpy. Careful choice of reducing agent yields extra heat as part of the reaction, helping keep your external heating requirements low. Coincidentally I just yesterday found my calculations for reaction enthalpies for various reducing agents (the more negative the value, the greater the heat released):

Zn - deltaH = 1020.8
Ga - deltaH = 494
Mg - deltaH = -58.6
Na - deltaH = -72.8
Li - deltaH = -193.79
Ca - deltaH = -363.73

These were calculated by using the reaction equation for each:
A NdF3 + B M == C Nd + D MFx
where ABCD are the coefficients and M is the chosen metal. deltaH for elements is 0, so the calculation boils down to
dH = D*deltaH(MFx) - A*deltaH(NdF3

In retrospect, I used the values for the solids but the products will be liquid. This changes the numbers a little, but preserves their order in the list.

Calcium produces the most heat but requires the highest temperature to melt, and I'm not sure if the tradeoff is favorable or not. If you have calcium metal, it might be worth a shot!

I've been thinking about this experiment again recently. God, I need to just try the last step already.

unionised - 4-10-2017 at 11:35

Interesting.
"A binary eutectic composition of Formula and Formula was observed to melt at 769°C."
From
http://jes.ecsdl.org/content/104/11/661

MrHomeScientist - 5-10-2017 at 07:32

Formula and Formula you say?

From the link (in case it breaks): "A binary eutectic composition of 80.5 mole%LiF and 19.5 mole%CaF2 was observed to melt at 769°C."

Somewhat less than LiF alone, but the reaction still needs to achieve 1021C to melt the Nd.

Lab Rat - 10-4-2018 at 15:50

This thread's been quiet for some time now... I hope my intrusion isn't too unwelcome, but I've been reading and re-reading the earlier bits of this thread a lot in the past few weeks, and now I'm finally throwing my hat in the ring here with some questions!
I'm a simple man: all I really want is to grow some pretty pink crystals of Nd2(SO4)3*8H2O (yeah, there are other, more accessible compounds that fill this role, but honestly I also just want the coolness of them being a Nd compound

). Based on earlier discussions in this thread, I processed a half-inch cube Nd magnet using the oxalate separation method: demagnetize, remove plating, dissolve in sulfuric acid, filter, oxidize all Fe(II) to Fe(III), precipitate Nd2(Ox)3 with oxalic acid. That worked fairly well, I think, as it has for other people in this thread. Later, I was planning on converting the oxalate to the oxide, then redissolving in H2SO4 for the final sulfate.
...Buuut, then I read Blogfast's posts on page 4, noting his discovery that Nd2(Ox)3 from this method contains significant Fe(III) contamination.

...BUUUT, then I read Blogfast's OTHER posts on page 4 about separating Nd+3 from Fe+3 using fresh iron hydroxide/oxide-hydroxide/whatever you want to call it! It seems perfect: FeO(OH)*nH2O is easy to prepare, and apparently, when added in excess to Nd/Fe sulfate solution, produces almost perfect precipitation of iron as... something, the old posts are actually very unclear about what the iron becomes, but it appears that an oxide of iron is involved.

And that's basically my question. I've literally been thinking about this iron separation problem all day, and I can't work it out. How does adding solid Fe(OH)3 (or FeO(OH), or whatever) to a solution of a ferric salt cause all the iron cations to precipitate out? What do they turn into? What does the Fe(OH)3 turn into? What's the reaction/mechanism? And why does it leave the Nd virtually untouched?

[Edited on 4/10/2018 by Lab Rat]

MrHomeScientist - 16-4-2018 at 06:02

Doesn't look like anyone knows! Unfortunately blogfast isn't around currently to answer. Perhaps some sort of common ion effect?

I had great results with separation via the potassium sulfate double salt route. If your goal is to make crystals, though, iron removal may be less important. The very act of growing crystals will purify your desired product! You just might have to brush the iron crud off the surface.

Lab Rat - 16-4-2018 at 12:44

Hey, thanks for replying! Eh, that's ok, I'll try the iron thing out on a small batch anyway just in case. And yeah, even just separating Nd salt by crystallization is probably sufficient for me, but... I have just as much doing the chemistry as I do looking at the final product!

As for the double salt method, GAAH, it seems like everyone has had luck with that except for me! Recently, I dissolved a 45.3 gram "de-shelled" magnet in 80 mL of 9M H2SO4, diluted it to 600 mL (to let all salts stay dissolved), filtered it, and added 7.3 g dissolved K2SO4 (which I believe is stoichiometric)... And nothing happened. I added a bunch more, and I DID finally get precipitate, but the reaction was very slow. So okay, whatever, chemistry isn't as clean as I dream of it being, I need to add more than one equivalent.

But the other problem is that I can tell that the reaction was incomplete, because the supernatant solution was still different colors under fluorescent and incandescent light, indicating remaining Nd.
I figured, hey, maybe I just needed to add even more K2SO4. To test that theory before doing it, I instead made a small amount of saturated K2SO4 solution and added a few drops of my magnet solution into that. Weirdly, though, no additional precipitate!

Is it possible that my magnet solution was too acidic, causing the double salt precipitation to become unfavorable or something? Or maybe the reaction DID go to completion, and I'm wrong about the different colors thing. When I get the chance to work on it again, I'll try oxidizing to Fe(III) with H2O2 and testing for Nd with oxalic acid.

[Edited on 4/16/2018 by Lab Rat]

MrHomeScientist - 17-4-2018 at 05:35

I ran into a similar problem. We found that just adding the solid K₂SO₄ directly to the magnet solution is far better than adding a solution. I guess the extra water interferes with precipitation for some reason (doesn't make a lot of sense to me, but it does work. It could be a pH thing like you mentioned). Just add the solid salt, stir vigorously for something like 30 minutes, allow to settle, then check the supernatant for color. Changing solution colors definitely indicates Nd still in solution.

Lab Rat - 17-4-2018 at 15:11

Ooooooh, VERY good info, thanks so much! And now that I think about it, I'd be too afraid to try and modify the pH with base, because introducing alkali or ammonium ions in the presence of sulfate might trigger early precipitation of the double salt before I want it.

Lab Rat - 3-5-2018 at 10:42

Ok, here's a little update from me, if anyone's interested. In my second post from a couple weeks ago, I mentioned that my double salt Nd precipitation didn't appear to go to completion, as evidenced by different solution color under different lighting. When I got to work on it again, I took a small amount of my solution (containing mostly Fe(ii) with suspected leftover Nd(iii) sulfates, as well as a lot of excess potassium sulfate), filtered it, oxidized the iron, and added plenty of oxalic acid to test for neodymium. To my pleasant surprise, the test failed! No precipitate.

Either I was completely wrong before and the reaction DID go to completion and I literally just observed the color wrong, or leaving it alone for two weeks allowed it to move towards completion. Honestly, I'm thinking it was probably the first thing lol

Filtered and washed the double salt, then reacted it with excess KOH, and filtered and washed the Nd hydroxide (pale green under fluorescent, pink under incandescent). I'm gonna wait to continue until I finish my vacuum filter setup--all I need is the right water pump to connect to my aspirator, and the tubing--so that the filtering and washing steps are faster.

Bezaleel - 7-5-2018 at 08:33

Thanks for posting all this, Lab Rat! Great work!

I've been working on Nd as well. Separating it with oxalic acid worked well for me, and rinsing with ~5% H2O2 removed most of my Fe contamination. I then thermally decomposed the Nd-oxalate, to be left with a brown oxide, the brow colour being due to contamination with Pr. A few percent of Pr is enough to make it dark brown. When dissolving this "didymium"oxide in H2SO4 soln., the sulphate crystals were a bit more reddish than those of completely pure Nd-sulphate.

If you decide to thermally decompose some of your Nd-oxalate, please do not strongly calcine it, or your oxide may become hard to dissolve in sulphuric or any other acid.

Besides, I've been working on the dissolution of industrially calcined Nd2O3 (which is pure with respect to Pr-content), which is a very difficult task. So far I failed in coming up with a good method to do this.

Lab Rat - 10-5-2018 at 12:21

Hey man, sounds good! Yeah, to be honest, there's no way I can expect my sample to be free of praseodymium, but the ionic radii of Pr and Nd are so similar that I would expect Pr contamination to simply fill the same holes as Nd in the crystal structure.

As for the calcination, I've seen that issue arise once or twice in the depths of this thread, but I don't really understand it... If you heat Nd2(Ox)3, you get Nd2O3. And if you heat it REALLY hot, you still get Nd2O3, right? How can it be that strongly calcining it makes it harder to dissolve? Does it transform into an incredibly stable solid state structure at high temperatures or something?

As for dissolving Nd2O3, I wonder if you've seen this paper, it might be userful: https://www.degruyter.com/downloadpdf/j/ncrs.2002.217.issue-...

It's actually about dissolving praseodymium oxide to get the sulfate, but I'm pretty confident the method would work with Nd, too. Basically, they wet the oxide powder with water first, then added conc. H2SO4 dropwise to dissolve the oxide and form the sulfate. I guess then for additional purification/clean crystal growth, they dissolved it in 1:1 methanol:water (v/v), and added a tiny bit of 2,2'-bipyridine (presumably to form some kind of bipy coordination complex). Evaporation yielded green needle crystals.

woelen - 14-5-2018 at 03:47

Strong heating of oxides makes the oxide more compact and more crystalline. The reactivity of the sample then strongly decreases and this can go so far that it becomes nearly impossible to dissolve the oxide in anything else than molten glass or molten alkalies.

I have encountered many oxidis which have become like this:
- Nd2O3
- Er2O3
- Pr2O3
- SnO2
- TiO2
- Co3O4
- Cr2O3
- Al2O3
- Fe2O3
- Sb2O5
- Nb2O5

Once you have an oxide in this inert state it has become nearly inaccessible for aqueous chemistry. In this way I lost well over 100 euros worth of money by buying oxides of metals and not being capable of using them in any meaningful way.

Bezaleel - 8-6-2018 at 06:07

Quote: Originally posted by Lab Rat

(...)

As for dissolving Nd2O3, I wonder if you've seen this paper, it might be userful: https://www.degruyter.com/downloadpdf/j/ncrs.2002.217.issue-...

It's actually about dissolving praseodymium oxide to get the sulfate, but I'm pretty confident the method would work with Nd, too. Basically, they wet the oxide powder with water first, then added conc. H2SO4 dropwise to dissolve the oxide and form the sulfate. I guess then for additional purification/clean crystal growth, they dissolved it in 1:1 methanol:water (v/v), and added a tiny bit of 2,2'-bipyridine (presumably to form some kind of bipy coordination complex). Evaporation yielded green needle crystals.

Well, it is evident from their description, that they use a weakly calcined Pr-oxide. Otherwise the dropwise addition of sulphuric acid would not make any sense. Wetting a hard-calcined oxide does nothing at all to such oxide. (I had a bit of success by partially dissolving hard-calcined Nd2O3 in sulphuric acid by boiling it for over an hour in a distilation set-up.)

Adding the sulphuric acid dropwise seems to me like a way of constraining the excess of sulphuric acid to a minimum. I assume that an excess of the acid is needed in order to solve the oxide (even if it is only weakly calcined). They are still left with at least a little excess acid of though, because they heat it to only near-dryness, instead of evaporating the excess H2SO4.

It seems to me that the addition of the CH3OH/H2O mixture is their way to neutralise the excess acid. And indeed a precipitate forms, presumably of Pr(OH)3.nH2O. It could be that here the role of the pyridine comes in, as an agent to dissolve the hydroxide again, so that the hydroxide will not serve as a nucleation agent for the Pr-sulphate they are after. I only suppose this; if anyone knows a better explanation, please post!

So I think that the mixture of methanol, water and the excess H2SO4 yields them a more or less neutral solution, which would prevent H2SO4 from entering the crystal lattice. (When you crystallise Pr-sulphate from a strongly acid solution, H2SO4 *will* enter into the lattice - I know from experience. Successive steps of calcining such results and recrystallising will reduce the amount of acid in the Pr-sulphate by a factor of ~100 in each step.)

For me the big question about the article you mention is the role of the 2,2'-bipyridine. It is used in a large amount. The calculated molar ratio is

Pr2O3 : bipy = 1 : 1.9

or

Pr : bipy = 2 : 1.9 (almost one bipy ligand for each Pr atom)

That their solution turns orange shows that there must be a strong interaction between the bipy ligands and the f-orbitals, since the green colour of Pr is due to transitions involving the f-electrons.

Lab Rat - 19-6-2018 at 08:10

Sorry, not related to what we were just talking about, but two quick things about Nd magnet separation:

1. In the sulfate double salt precipitation method, don't heat the mixture above 50°C. Honestly there's no reason to heat it at all (I don't really know why I did it anyway), but one batch I warmed and stirred just to quicken the precipitation (which, in my experience, is a very slow reaction, taking upwards of 24 hours to complete), and another batch I warmed upwards of 70°C and stirred, and that precipitate appears noticeably orange (probs iron 3) under fluorescent light. I'll probably be able to separate it out somehow later.

2. I don't recommend washing/drying the sulfate double salt with acetone. I didn't have a vacuum filter setup until just now, so to dry my product, I'd often suspend it in acetone and re-gravity-filter, because the acetone evaporates quicker than water. But in my experience, the double salt appears slightly soluble or at least colloidal in acetone. Either way, I don't really recommend it.

Honestly, neither of these things have been suggested anywhere above, but like... just so ya know, I guess!

[Edited on 6/19/2018 by Lab Rat]

Dan Vizine - 19-8-2018 at 14:55

Quote: Originally posted by Lab Rat

And that's basically my question. I've literally been thinking about this iron separation problem all day, and I can't work it out. How does adding solid Fe(OH)3 (or FeO(OH), or whatever) to a solution of a ferric salt cause all the iron cations to precipitate out? What do they turn into? What does the Fe(OH)3 turn into? What's the reaction/mechanism? And why does it leave the Nd virtually untouched?

[Edited on 4/10/2018 by Lab Rat]

This isn't a strictly chemical process. It has a lot to do with adsorption. Ferric oxyhydroxide (from ferric chloride plus a hydroxide) has long been used to wash heavy metal ions out of ground-water.

It's unlikely that this method would necessarily leave Nd totally untouched, but this physisorption process is very pH sensitive and operating near the notch of the typical V-shaped solubility curve for the targeted metal hydroxide is the aim. Normally, for ferric iron that is about pH 7.8. This is close to the pH 7 that blogfast25 mentioned on p. 4. Presumably, the solubility minima for Nd lies at a somewhat higher pH.

fusso - 20-8-2018 at 02:44

Is there any solvents that can dissolve amorphous B? CS2?

Poppy - 21-8-2018 at 17:57

Similar dissolves simular. Most like the way silicon oil cleans dust better than soap does clean it. I remind carbon is more soluble in water than alcohol, and > than organic solvents. Thats because pi bounds and some e pairs make their fashion to surpass the dynamics of the solvents requirements to carry the C particles along. Also, not less important you have to first of all undo the clumps of amorphic B then they will inevitably solubilize. Then you go 2 steps: create a silly putty of B + surfactant (anything that matches the maximum dipole moment achievable for B bonds), heat and stir, then dilute in suitable solvent for both, same way you would first rub grease with soap and they dilute both in water. For the surfactant I would suggest sulphonic acid.

[Edited on 8-22-2018 by Poppy]

fusso - 8-9-2018 at 13:50

when I use oxalic acid to precipitate Nd3+ from the magnet chloride solution, should I add excess oxalic to complex all Fe3+ or slight excess relative to Nd3+ is enough?

Lab Rat - 26-9-2018 at 09:04

Quote: Originally posted by Dan Vizine

Ahh, very interesting, thanks a ton! I've ultimately decided that I don't trust my abilities/equipment enough to rely on this adsorbtion process, and will instead probably just recrystallize my Nd2(SO4)3 at the end to purify. Also, I'm highly suspicious that the double salt precipitation step is pH sensitive, and its efficiency is low in the presence of acid. I tried to use Fe(OH)3 to neutralize excess acid, but making Fe(OH)3 is a pain, and so is filtering the excess back out (small particles). I'll try using bicarbonate in the future instead.

Wizzard1 - 26-2-2019 at 13:19

To the folks getting their Nd from R.E. Magnets and HCl -

Anybody have an explanation for the perceived color differences between batches / ages of magnets? I'm assuming it's other rare-earth impurities, mostly Pr of course.

Older magnets yield a Nd sulfate with almost a fluorescent rose color, very pink, while newer ones have a much more neutral rose color, but some batches come out a a shade more towards lavender.

Just wondering.

[Edited on 26-2-2019 by Wizzard1]

fusso - 23-4-2019 at 23:22

I've dissolved some Nd magnets into HCl, I want to ppt Nd as oxalate. Is stoich amount of oxalic acid (just enough to ppt all Nd2ox3) enough or do I need huge excess (much more than 3eq relative to all metals) to keep all Fe3+ as the soluble ferrioxalate ion?

SantinoCase - 7-9-2019 at 04:26