P2 – Thomasclarkite-(Y)

Welcome back! Two down, 227 to go. Gulp.

The first two space groups were pretty simple prints (although my post-print work was a little shocking – if only someone would buy me a duel nozzle printer or an SLA printer… 😉 ) and I am really happy with how they turned out.

However, now things are getting serious.. We’ve moved out of the asymmetric territory of triclinic space groups and are now entering the heady heights of the monoclinic crystal system. Unlike triclinic where none of the sides or angles of the unit cell are equal, in the monoclinic system a, b and c are not the same distance, but the angles α and β are equal to 90 degrees.

P2 is the simplest of the space groups in the monoclinic crystal system and the fantastic people at: https://crystalsymmetry.wordpress.com/230-2/ selected ‘Thomasclarkite’ as their example for it.

Thomasclarkite is a newly (1998) discovered sodium containing rare-earth element carbonate, identified by J. D. Grice and R. A. Gault in Mont Saint-Hilaire, Quebec. Oh how I would love to go mineral hunting there one day! Sadly my one major mineral hunt ended with me falling down a cliff, but that is another story..

The .cif (from: http://rruff.geo.arizona.edu/AMS/minerals/Thomasclarkite-(Y) ) outputs the following in crystalmaker:

Thomasclarkite-(Y) with carbon, sodium, oxygen and yttrium in black, yellow, red and light blue respectively. Water positions are dark blue/grey.

Looks pretty but not printer friendly at all – firstly those floating water positions have to go and we need to increase both the bond sizes and unit cell to be able to print this..

Bumped up the unit cell to crystalmaker’s max (blue line, 10 pixels), changed the hydroxides to pink (wish I had neutron data to play with – then we’d see the hydrogen positions), removed the water and increased the carbon atomic radii to 0.75 angstroms. I’ve also increased the stick bonds to 50% the smallest atomic radii and we get the below:

Thomasclarkite-(Y) with water positions removed.

I think I’m going to take the executive decision not to print water positions (or any other absorbed molecules) unless they are part of the structure. I think it would be too tricky to keep them in for my Snapmaker – although I’m happy to upload .stl files if you want to give it a go – just let me know!

Importing the .stl into the snapmaker3D slicing software we get:

Scaled up to 500 % and rotated in y by 90 degrees

Only 7 hours! Although it is looking flimsy – I’ll give it a go but I’m not sure about this one at all.. Especially considering my record so far with the post-print work..

Check back later to see how it went!


P-1 – Success

After tweaking the deuterium atomic radii the print seemed to work well – although I am starting to discover the limitation of my poor Snapmaker is in the post-print. The scaffold is proving to be very difficult to remove without damaging the print.

A slightly battered print of the P-1 space group

Still – I’m happy with the .stl file so feel free to print out your own version! I’d recommend a soluble scaffold or a breakaway filament.

If you want to have a go yourself – check it out at:


Let me know how you get on if you give it a go!

P1 – Success (almost!)

We have success! The project begins properly now! After several failed attempts (check out the failures category for more details) we have our first space group print:

As you can see – although the unit cell printed, it was so thin that it snapped easily in post clean-up

Sadly my ham-fisted approach to cleaning up led to a couple of the unit cell axis being snapped. I don’t think you’d have that problem if you scaled it up even more than the 850 % that I used or used a breakaway filament such as PVA for the scaffold. Sadly I only have a single nozzle on my snapmaker so that is just a dream for me!

If you want to have a go yourself – check it out at:


I would love to see how yours comes out! Good luck and stay tuned for P-1 (it’s printing at home as I type.. Gulp!)


One down, only 229 to go. Gulp. Eventually, bar a minor filament failure and raft adhesion issues P1 seemed to work ok. So lets go through the same process with P-1. It’s important to remember that this is still a primitive triclinic system (a ≠ b ≠ c , α ≠ β ≠ γ or for the unit cell, the sides a,b and c are not equal, the angles α, β and γ are not equal) and the P-1 space group is related by inversion (hence the ‘-1’) to the P1 space group. But not just a simple inversion, check out; http://pd.chem.ucl.ac.uk/pdnn/symm3/sgtricl.htm for more information.

To the print! Again from the wonderful people at: https://crystalsymmetry.wordpress.com/230-2/

We the get the example:

Chalcanthite CuSO4 • 5(H2O)


CIF can be found here:

I’m using the one from neutron diffraction data by: ‘G. E. Bacon, D. H. Titterton, Zeitschrift fur Kristallographie, 1975, 141, 330-341′

With the same process of open the CIF in crystal maker and tweak it so we have polyhedra and a thick unit cell line (10 pixels) we get the following structure: (see
https://crystalprint.home.blog/2019/05/06/beginnings-p1/ for a more detailed run through of the process)

Chalcanthite viewed down c. Copper and sulphur polyhedra are blue and yellow respectively. Oxygen and deuterium, red and grey. Extra water molecules and stranded atoms hidden.

Looking pretty good, although I doubt those deuterium (grey, an isotope of hydrogen) positions will print and will cause havoc with my scaffolding. I’ve bumped the relative size of the bonds up to 80 % of the atomic radii of deuterium (remember this is about 3D printing, sometimes sacrifices will have to be made on the chemistry convention – a thicker bond is more representative of the probable electron position anyway..)

The .stl loaded into snapmaker, scaled up by 750 %

Ah. So even with those bonds bumped up to 80 %, and the model scaled up by 750 % we get floating hydrogen atoms. Crystalmaker will only let you increase bonds to 90 % of the atomic radii but even then, they are still not getting picked up by the snapmaker software. I then fiddled with the atomic radii of deuterium (usually 0.04 angstroms) but even having it the same as oxygen (1.2 angstroms), still no bonds!

Right, I’m going to have to slightly off-piste with this one – my number one objective is to print the unit cell of this space group, as faithfully as possible. So I’m going to make some changes to the atomic radii to do this. Firstly I changed the deuterium radii to 0.75 angstroms (it looked silly as 1.2 as well as being very, very wrong). Next in the ‘model>model inspector’ menu I changed entire ball-stick atomic radii to 65 % of actual. Don’t forget, those ball and stick models you see of crystals are not true representative models of the structure, actually there is very little space between the atoms and the spheres (in reality ellipsoids but that is another story), they are shrunk for convenience.

Chalcanthite viewed down c, this time with deuterium (grey) atomic radii set to 0.75 angstroms and then all ball radii scaled to 65 % of actual. Extra water molecules and stranded atoms hidden.

Boom. This looks much better. What about in snapmaker?

Scaled up to 750 %

Looking good, with a raft everywhere and standard snapmaker settings to normal, we’re looking at 30 hours and about 50 m of filament! Looking back at P1 we are printing out at just shy of 100 mm in the largest direction so I think P-1 can be shrunk a tiny bit.. At 550 % scale up we’re getting dimensions of 58.6 x 92.1 x 77.4 mm and a print time of 15 hours. Lets give this a shot, tune in later to find out what happens!

Beginnings part 2: P1

It’s done, the stl code has been sliced (put in a format the 3D printer understands) and all I need to do now is wait..

Of course there was the minor hurdle of the filament getting stuck on its feeder. The Snapmaker filament holder is just a little too short to handle Primia 1 kg filament but I think I caught it in time..

Whatever happens I’ll post about it and when I have a successful stl file it will be shared on the space group print project page!

Raft is down for the first P1 print, it’s not perfect due to the filament feeder getting stuck, but let’s try it out!

Beginnings: P1

tl:dr P1 triclinic is the simplest space group and it’s unit cell has no smaller asymmetric units. I’m going to try and print one and document the process below. As this is the first post, it’ll be longer than usual!

Scroll down to get to the actual printing, but first a bit of background to crystallography..

We have to start from somewhere so why not the simplest form of symmetry found in crystals? The P1 space group could be thought of as the Gaia or mother of all space groups. Every other space group fits into this one. It is also the crystallographers nightmare, when collecting the data atomic positions via X-ray crystallography as no two atoms in the unit cell are related by symmetry, you have to collect every. bit. of. data.

This can take a long time, depending on the complexity of the crystal and on the instrument being used. If, like me, you use single-crystal X-ray diffraction as your primary technique – you normally have a good idea of the quality of your crystal from your initial diffraction to work out the unit cell. Still, despite that I have put on many tiny hopeful specks of crystals, to come back 12+ hours later to a dataset that looks like a 2 year olds drawing of the night sky – why are there lines?! (You are looking for perfect spots – diffraction is essentially X-rays shining on atoms and bouncing off their clouds of electrons – how the atoms are arranged controls how the X-rays bounce off and where we detect them – we use those positions to start to build a model of the structure of the crystal.)

Remember a space group is just a shorthand way of describing how the atoms in a unit cell are related to each other via symmetry. (for an excellent introduction to unit cells, go to:
https://chemed.chem.purdue.edu/genchem/topicreview/bp/ch13/unitcell.php and for asymmetic units:

http://pd.chem.ucl.ac.uk/pdnn/symm3/asymunit.htm )

The fantastic people over at space group list project (
https://crystalsymmetry.wordpress.com/230-2/ ) have already identified a .cif and structure to work from:

Example – Gilmarite,  Cu3(AsO4)(OH)3

Gilmarite, Cu3(AsO4)(OH)3 a new mineral:, its description and crystal structure

Sarp H. , Cerny R.
European Journal of Mineralogy , 11 (1999) p.549-555


http://rruff.geo.arizona.edu/AMS/minerals/Gilmarite id=07530.cif&down=cif

Printing the P1 space group

So first stop is to open this .cif in your favourite modelling software, I use crystalmaker (
http://crystalmaker.com/ ) but other flavours are available.

From tiny acorns.. Don’t worry if this looks a bit alien – there are plenty of crystalmaker tutorials on their website

The above image is the CIF just loaded straight into crystalmaker. From past experience I know ball and stick models tend not to do well on a 3D printer so I am going to attempt a polyhedral model instead.

We have polyhedra!

Ok, looking a bit better, I’ve got rid of the unit cell (dotted blue line) turned the arsenic (purple) and copper (blue) atoms into polyhedra and kept the oxygen (red) as spheres – 3D printers are not good with corner joins! Those copper sites look a little odd however. To get them as proper polyhedra I had to tweak the copper-oxygen bonds to 2.6 angstroms (1 angstrom = 0.1 nanometre) which seems reasonable (if a bit more in the coordination complex area).

This time with copper-oxygen bond distances of up to 2.6 angstroms

Ok now we are in business, notice to get a full picture I have had to expand it to 27 unit cells – a single unit cell on it’s own doesn’t quite get the picture across – however is that really representative of the P1 space group? No, not in a print. Let’s try something different..

This time a single unit cell (and some of the neighbouring unit cell’s atoms to allow polyhedra to be drawn sensibly)

Much better – although I have no idea how well the unit cell line (blue) will print – if at all! I’ve set it to 10 pixels wide and ‘solid’ so hopefully the slicing software I use will pick it up.. Lets find out! Luckily crystalmaker has an ‘export to .stl’ function so that step is simple. I use a mix of Cura ( https://ultimaker.com/en/products/ultimaker-cura-software ) and the Snapmaker slicer to prepare my prints – we’ll try snapmaker first:

Well I can see a unit cell (solid lines) that is very promising!

So far so good. With a little bit of tweaking of scale (running at 750 %) and rotation in the z axis (I want the unit cell lines to be as easy as possible to print – the longer axis should be as close to vertical as possible and the shorter ones be an overhang angle which the printer can cope with) hopefully I’ll get something printable.

Ok – this is when I wish I had a dual nozzle printer..

So with the settings at: 750 % scale up, rotation in z – 71 degrees, raft everywhere and standard ‘normal’ settings for the snapmaker (see below image) we get a reasonable print time of just under 11 hours and 18.2 m of PLA.

Normal quality settings for Snapmaker – please change if you are using a different printer as needed

So far, pretty simple! Keep tuned to find out if the print actually worked..