Q for chemists about lanthanide-carboxyl structures and their formation

Factual Questions
1

In a lab I work at at school we are trying to create aggregates of the lanthanide gadalinium along with carboxyl groups which are attached to a long (20 or so) carbon chain. The problem is we are getting polymers of lanthanides with carboxyls instead of a large number of universally sized aggregates of a single gadolinium atom bonded to three carboxyl groups, which is what we want.

Does anyone have any advice, input or can you point me in the direction of any books or scientific articles on carboxyl/lanthanide structures and their formation? I would appreciate it as right now I’m trying to learn about the nature of whats going on and how to create gadolinium with 3 carboxyls on it instead of a massive polymer.

We are trying to create gadolinium 3+ atoms with metalloligand bonds to three carboxyl groups in the hope that this will fulfill the f-shell of the Gd.

My inorganic chemistry is a bit rusty though, and I learned more about S and P shells, not F shells. I know lanthanides have 7 electron shells (s=1, p=3, d=5, f=7) and I’m pretty sure that a gadolinium 3+ atom would have 6 electrons in its shell. From what I remember you want as many electrons to be unpaired as possible, but I thought that a 2+ charge was considered the de facto number of electrons which woudl mean that Gadolinium 3+ already has 6 unpaired electrons, and would prefer to only make one bond. Some Gd-ligand complexed like Gd-DTPA
or Gd-DOTA show 8 bonds. You’d assume that would happen as Gd in a 3+ state has room for 8 electrons.

However some pictures of Gd-DTPA or Gd-DOTA show the Gd bonded to a water molecule as well, making 9 ligands. Where do the extra water electrons go? Where do they fit?

In my view, maybe we are going about this all wrong for a few reasons.

  1. Gadolinium 3+ will only accept 1 electron to gain stability since it’ll have 7 unpaired electrons. it can hold up to 8. Trying to add 3 ligands (the professor thinks that Gd 0 is the most stable form) may be unwise as we’ll have 2 sets of paired electrons and 5 unpaired electrons. I don’t have any idea how to draw up the MO for an f shell bond though. I barely understand how to do it for S and P shells.
  2. Gadolinium can bond with up to 8 things, I don’t understand F block inorganic chemistry nearly well enough to know what happens to encourage metalloligand bonding, but I just don’t see how it is realistic to have only 3 bonds. Either we’d have 1 bond or 8 bonds.
    Would we be better off trying 1+ Neodymium, 2+Promethium, 3+ Samarium or 4+ Europium instead of 3+ Gadolinium? Are any of these molecules stable or affordable? Are things like 5+ Gadolinium stable or existant? We want something that’ll decay into smaller atoms when hit with neutrons. Maybe Gd is the best atom for this as it is (in the long term) designed for anti-cancer purposes.

Would doing the synthesis in an acidic environment help any? From what I’ve gathered by talking to the other professors there is resonance in the carboxyl oxygens, making 2 possible ligands on one molecule. If we left the hydrogen on and had a carboxylic acid would that still be able to metaolligand bond with the Gd?

Keep in mind that my chemistry isn’t great, so there can be problems with my view on this subject of f shells and how they relate to metalloligand bonding but any info is helpful.

2

I forgot to add this part

3

I’m guessing (I really don’t know for sure) that the goal is to have each carboxyl act as a bidentate ligand and to have 3 of these bidentate ligands forming a six sided ligand on the Gd. However we aren’t getting that, as I said. We are getting polymers of Gd bonded with carboxyl.

What is probably happening is each oxygen in the carboxyl is bonding to a different Gd atom, and each of those Gd atoms is bonded to one or more oxygens in another carboxyl, creating a gigantic chain of carboxyls & Gds. Does anyone know what we can do to prevent this from happening? I don’t understand MO theory well (especially as it relates to f electrons) so I don’t know how to encourage one kind of structure over another.

4

::BUMP::
First person: You’ve got Lanthanides in my fatty acids! :frowning:
Second Person: Your fatty acids are gumming up my Lanthanides! :mad:
First person has a taste: YUMM! :slight_smile:
Second person tastes: YUMM! :slight_smile:
Seriously, I admire your optimism in asking this, but it’s a little more esoteric than the usual humdrum question about singularities at the end of time.
Beyond the obvious of reducing the rate at which you add the gadolinium to the acid, I think you’re going to have to trudge off to the library.

5

OK I am an organic chemist, so lanthanides dont interest me much but…

Lanthanides are rather big and love oxygen. They can quite easily bond between 6-12 oxygen ligands depending on size. (see http://www.radiochemistry.org/periodictable/la_series/L7.html )

Whether something forms a polymer or simple coordination compound depends on the size of the ligand as much as anything. If you want the latter which you do need very bulky ligands which discourage sharing of carbonyl groups. As well carboxylic acids are only 2-coordinate, so the lanthanide being greedy (e.g wants a 10 or 12 coordination) wants to share with neighbouring carbonyls as well - and so makes a polymer.

best possibility - bulky carboxylic acid with an extra OH for coordination to lanthanide (or crystallise from an alcohol so the solvent can help coordination).

triphenylacetic acid could try due to its bulk -

6

First of all, ditto the optimism remark. My guess is you are doing research to find the anwer to the very question you are asking, which means nobody knows the answer or there would be no reason to do the research.

Second, I love the challenge, and whilie this isn’t my area of expertise in any way I’ll tell you what I know,

Lanthanides like to be +3. I don’t know why at the moment they just do.
Lanthanides like to be 12 coordinate. Once again, I don’t know why, they just do.

Ih the carbonyl groups on the structures you posted were donating into the metal as I suspect, these complexes would also be 12 coordinate.

Since this is research, the first thing I would do is question your results. Why do you think that you are getting a polymer instead of a single complex?

A fatty acid can be considered a bidentate ligand (2 coordinate). If you are adding 3 equivalents to the gadolinium, your gadolinium is quite naked (remember it likes to be 12 coordinate.) In order to compensate for this nakedness, it may be forming larger structures.

Are you using a coordinating solvent? A coordinating solvent may help fill some of the spaces and break up the polymer.

On reread I see that scm1001 has dealt with most of this. With the information we have, its about as good as you can expect.

7

Use a low concentration of the gadalinium.

8

I’m an organic chemist as well, so I won’t be able to answer very many of your questions with high precision. As a matter of fact, all I know about are transition metals, so I assume that what I know can be generalized.

(1) Transition metals tend to obey a rule of 18, not a rule of 8. I don’t know if the same holds for f-block metals, but that’s the rule I learned in inorganic chemistry.

(2) Frontier molecular orbital (FMO) diagrams are highly dependent on the coordination sphere – again, no specific knowledge of Lanthanides, but for transition metals, the d-orbitals get split depending on how many ligands there are.

(3) Low-valent transition metals are incredibly reactive. Several researchers have used sterically encumbering ligands to enforce extremely low valences on early transition metals – the result is that they’re so electrophilic they do some funky chemistry (such as bind dinitrogen!)

In terms of advice, one thing you may need to watch for is not polymerization, but micelle formation if you’ve got long, hydrophobic fatty acid chains mixed in with water. You may not be getting polymerization at all – just coordination to the Gd ion and then formation of a micelle.

As with the DTPA ligands, one of the best ways to fulfill a large coordination sphere is to increase the number of coordiating functionalities per ligand. This may require a little organic synthesis, but it should be fairly straightforward. I recall the Lanthanides preferring “hard” nucleophiles (such as phenols) over “soft” nucleophiles (such as carboxylic acids). I could be wrong, but if possible, you could try using multivalent catechol-based or salicylic acid-based ligands.

Prof. Ken Raymond at UC Berkeley does quite a bit of research on lathanides. If you have access to publications you might want to start a literature search there.

http://www.cchem.berkeley.edu/~knrgrp/lantha.html

9

That was my concern. I’m thinking maybe my professor was trying to form ionic bonds and not metallo ligand bonds between the carboxyl groups and the lanthanides. If so I have no idea how you pick and choose what kinds of bonds Gd ends up with. It comes in the form GdCl3 6H20, because CL is -1 and Gd is +3 maybe she feels that three -1 carboxylate groups will ionically bond with the Gd. But as I said my inorganic chem isn’t good enough to know how to choose what kind of bond (ionic vs metallo ligand) is formed with gadolinium. Can you pick a different solvent to change that or is it just not controllable?

Does anyone think it is possible to bridge the carboxylic acids together to form a crown ether structure? If so how would I go about doing that, and what research should I look at for more info?

Thanks.

10

Does anyone have advice on synthesizing crown ethers or if this is a good idea? I figure maybe something like this would work to build what we want. however I really am not sure how to synthesize a crown ether. All the crown ethers I see have an ethane molecule connecting the oxygens, couldn’t you use a 2nd group atom like magnesium to bridge the oxygens?

Then again we may get the same problem with endless polymer chains connected to each other, which isn’t what we want. What we wanted was 3 polymer chains around each gadolinium. Each polymer chain has a head with 8 carboxylates on it in a chain and a tail with about sixty 2-methyl-propanes.

11

When I need a crown ether I buy it. If you really want to make it yourself just look up a synthesis in scifinder or beilstein or something.

If I understand correctly, you’re asking about making a crown ether-like molecule with no carbons in it? Just a ring of Mg[sup]2+[/sup] and O[sup]2-[/sup] ions? I doubt very much that anyone will coax magnesium oxide into that sort of crystal structure.

Also, the crown ether is a nice covalent ligand, but you still need to counteract the 3+ charge on the Gd.

I have to run. More later…

12

I know very little about f-block chemistry, but seeing as you aren’t getting too many responses, I shall continue to throw my conjectures at you.

First, you may want to do some more reading. I don’t know the answers to most of your questions, but it seems like many of them could be found in review articles and textbooks. Get thee to a library. I’m sure you’re sick of literature, but it probably has more answers than we do.

I would think that Gd3+ is [Xe]4f[sup]8[/sup]. IIRC, f-electrons don’t get too involved in bonding. Assuming all the orbitals are accessable, I think you’ve got the option of a 32-electron complex (like an 18-electron d-block complex). If you just want to fill all d and f orbitals, you would have a total of 8 bonds (3 ionic, 5 covalent – the names are just conventions and should not be taken too seriously, btw).

You still have empty s and p orbitals for filling. It’s like with d-block complexes. For example, [Ir(COD)Cl][sub]2[/sub] (a dimer w/2 bridging chlorides, COD = cyclooctadiene) has Ir[sup]+1[/sup] (8 electrons). 2 more electrons would give a full 10-electron d-block, but the complex is actually 16-electron (for each iridium: 8(Ir) + 2 (chloride) + 4 (COD) = 16).

I have no idea what you’re talking about. There is no need to reduce the metal center to a rather unstable Gd2+. As for holding up to 8, see above.

When exactly is Gd(0) involved here? Neither Gd-DTPA nor Gd-DOTA is Gd(0); they are both Gd(III) by all conventions I know. I have no idea how the crystal/ligand -field splitting works for f-block complexes, so I can’t help you there.

I, too, doubt that you’ll only have three coordination sites occupied on a lanthanide center. They’re huge. Yes, you’re only going to have three ionic interactions (in this case RCOO[sup]-[/sup]), but any other e-donation that can occur will occur. Bonding in these cases generally occurs when a non-bonding orbital (like on a neutral oxygen) interacts w/an empty orbital on the metal, resulting in a stabalizing interaction.

Gd is one of the cheaper lanthanides. Just look up the prices of the starting materials to find out. They’re all pretty much going to be 3+ metal centers. There are a few exceptions. You’re still going to have 3 RCOO[sup]-[/sup]-type intereactions. Different metals will mostly just vary the number of covalent-type ligand interactions you can have.

A protonated carboxylic acid can still donate electrons to the metal center, but remember you still need something to counteract the 3+ charge. I don’t know if too low a pH may cause protonation of the metal center (in which case a hyride(s) will be the X ligand(s).

This post is too long.

13

[QUOTE=Wesley Clark]
maybe she feels that three -1 carboxylate groups will ionically bond with the Gd.

[QUOTE]
I think this is likely.

I’m not sure what exactly you mean by “metallo-ligand bond.”

I have no idea if this is possible. I’d check w/SciFinder or something. It wouldn’t, of course, be a crown ether, but I think I can envision what you may be thinking of. I’ve never seen such a molecule though.

Eh, save that for some better advice than mine.

I think I addressed your next post already, except for this:

I’m not sure exactly what you mean here. What exactly is your desired ligand:Gd ratio?

Ok now I actually have to go. I hope this all didn’t just confuse you more!

14

A lesson I (eventually) learnt as a PhD student was… “A year in the lab can save you literally hours in the library”
I’m not intending to sound patronising, as many of us are guilty of diving into research without going through the tedious literature reading, we wanna do experiemnts damnit!
Anyway, here’s how I would go about things

  1. Try to reduce the number of variables by defining precisely the type of complex you’d like to prepare (which metal, which oxidation state, which ligand, suspected stoichiometry etc)
  2. Do a thorough literature search (I think for this example Gmelin would be the way forward) to ascertain what has and has not been prepared. What solvents were used? Coordinating or non-coordianting solvents? Any X-ray structures reported?
  3. During the lit search, try to ascertain exactly how the complexes which have been prepared have been characterised.
  4. Start simple. What happens when you dissolve cerium chloride in methanol and add a long chain acid salt (1:3)? (I have no idea…will you see eg NaCl precipitate?) (Unless of course the lit knows the answer to this which I’m sure it does.)
  5. Don’t worry so much about trying to rationalise your hoped for results in terms of the orbitals of the metal. Whilst more knowledge is never a bad thing and this will certainly help in after the fact rationalisation of your results, I suspect your empirical results will dominate the work.
  6. If the lit’s overwhelming, email someone. Ask a couple of questions of authors you’ve found in the lit; I did this in my last post doc, Victor Snieckus helped me out, faxed me some experiemental, shed light on my problem and voila, a Nature paper;)
    (OK, it wasn’t quite that simple)

Anyway, good luck.

I never posted a chemistry question here, I thought it’d get zip replies. Now I know better…

15

So true, So true. :stuck_out_tongue:

16

This link might be of some use; if it isn’t, it does contain some pictures of the f orbitals.
I’d forgotten how beautiful they are.