Science - Page 2
Kaolin
Kaolin, like just about all clay minerals, is made up by stacking layers tetrahedral silica and octahedral aluminium oxide (alumina). In kaolin, each stack consists of one of each type, but other types of clays have different structures, for example bentonite has two silica layers sandwiching a single alumina layer.
At the surface of the clay, the charges aren't balanced, and so the clay bonds with water. Water being a polar molecule, the hydrogen sticks to the oxygen on the clay, and the oxygen in the water is repelled. The strongest bound water is just 1 nanometer thick, but looser bound water extends up to 50nm from the clay surface. The lubricating power of the water, plus the very small platelets of clay, give it the plasticity we potters need.
During firing, it undergoes multiple transformations. Between around 550 and 900C the water is driven off as steam, and the kaolin is transformed to metakaolin - and if you put it into water, it won't go back to the original clay state. Then, between 925 and 950C, the metakaolin breaks down to form a cubic spinel crystal from the alumina and most of the silica, with the rest of the silica taking on an amorphous form.
From 1050C this breaks down into platelets of mullite (from alumina and silica) and high crystobalite (a crystalline form of silica). And finally from 1200C we start to see the long, needle like crystalline mullite forming. Note that some people give higher temperatures for the formation of mullite - this is true if dealing with pure materials, but the materials we potters use are far from pure, and the impurities act as a catalyst, bringing the temperature of formation down.
All of these are one way reactions, except that the crystobalite transforms to low crystobalite at 270C or below (to be covered more when we look at silica). However they don't occur that quickly, so each time we fire the piece we are transforming some, but not all, of the kaolin into mullite and crystobalite. The hotter and longer we fire, the greater the transformation. This gives clay much of its fired strength, but also makes it more prone to cracking in the kiln or in use, as we'll see when we look at silica.
Silica Changes
The diagram shows the transformations that may occur, the temperature and speed of the transformation, and the associated change of volume. High risk transformations are shown in red. And blue shows the sources of silica.
Silica can change between a number of phases during the firing - the chemical composition remains the same, but the atoms rearrange into different shaped crystals, and these changes also result in volume changes, which cause stresses to our pots in. The low/high crystobalite change may also occur with ovenware in use.
In some cases there is no breaking and remaking of chemical bonds, or repositioning of atoms, and these changes happen pretty well instantaneously, as with low/high quartz and crystobalite changes. These are potentially the most damaging, as our pots are seldom at an even temperature, so suddenly part of the pot changes in size whilst the rest is unchanged, and large stresses can build along the boundary. In other cases the changes are much more gradual, so the stresses are less. These slow changes are seldom completed over a single firing, and are mostly irreversible at the rate at which we heat and cool our kilns, so each time a piece is refired, more of the transformation will take place.
Note that the terms alpha and beta have been replaced with high and low, as physicists and geologists used alpha and beta in opposite senses which created confusion!
The silica doesn't just come from silica or flint in our glaze or clay body, both of which enter as low quartz. The kaolin (and other clay minerals) decomposes to high crystobalite and amorphous silica during the firing. And if we introduce plant ashes or glass into our glazes, this is also amorphous silica.
Note that impurities, such as fluxes and colourants, can have a significant effect on things. The diagram shows the difference in routes taken from quartz to crystobalite, but in many cases impurities also bring down the temperatures at which all these changes take place.
Glazes Melting
Whilst we talk of glazes, and of the clay body, melting, this is actually incorrect. Just look at the melting points of our materials, and you'll find that generally they are higher than when we know they become liquid, and often are higher than the top temperature of our kiln.
As I've said before, things kick off with materials like sodium migrating to silica (or already being there, for example in a frit), and the outcome of the resultant reaction is a liquid. Then the liquid spreads into contact with other materials, and they either dissolve into, or react with, the liquid to enter into the melt.
Exactly the same as when adding sugar to coffee.
Silica Glass
In this first of a series of posts that looks at what happens when the pot cools, and the glaze or the melted part of the clay sets, we'll start with the basics and look at silica. This is idealised, but we'll ad in the other components with time.
The pic shows a section of the silica glaze, with each blue silica atom surrounded by a tetrahedron (or 3 sided pyramid) of 4 oxygen atoms. You could draw lines from the centres of the oxygen atoms, and the silicon atom would sit in the middle.
Where oxygen atoms are drawn that are only attached on one side, you can either imagine that it has another silicon atom attached, and the glaze matrix continues on for ever, or you can wait until we start adding in other elements.
This network of silicon and oxygen is based on rings forming, but of variable size, and without any sense of order to them. This makes it a glass - if it were a crystal, the all the atoms would be in an ordered layout - silica has a number of configurations it can take up.
When liquid, the structure is much less ordered and more dynamic, with the strings and loops of silica continually breaking and reforming. As it cools down, the atoms become less energetic, and start organising themselves into a regular pattern - but as it cools, it becomes more viscous, and the silica sets solid before it can develop into a crystalline shape, like quartz. This high viscosity on setting is essentially what makes something into a glass rather than a crystal
Whilst silica is a great glass former, its melting point is too high for our kilns. So next week we'll look at network modifiers, or fluxes, that will bring the melting point down and also tidy up the loose ends on those oxygen atoms.
Silica and Alkali Metals
Here we have a glass/glaze of silica (light blue), oxygen (red), and various alkalis (lithium bright green, sodium dark green, potassium teal). The alkalis are connected to the oxygen atoms with relatively weak ionic bonds, replacing the hydrogen in last week's post. You can see how the tendency is to put the alkalis on the perimeter of the structure.
The alkali metals all have a charge of +1, so they can only attach to one of the oxygen bonds (with its charge of -2), and nothing else. This means that once an alkali has attached itself to an oxygen, the glass cannot grow any more in that direction. The oxygen bond has become a non-bridging oxygen (NBO), whereas if connected to silica it would be a bridging oxygen (BO). So the alkalis keep the glass structures small, which lowers both their melting point and their chemical stability.
When cooling from a liquid, everything is in a state of flux as things start forming structures and solidifying out, so alkali ions can easily just latch onto oxygen atoms that aren't yet covalently bonded to silica. But the alkalis can also break an existing bond between oxygen and silica. For example, a sodium ion would attach to the oxygen on one side of the broken bond, whereas on the other side an oxygen would attach with a covalent bond to the silicon, and then a sodium would form an ionic bond with the oxygen (bearing in mind that the sodium oxide has formula Na2O).
Also shown are phase diagrams for sodium and potassium against silica. Whilst phase diagrams are not always useful when looking at the melting of glazes, or the behaviour of clay bodies, they are good for looking at the cooling of a glaze, where everything is fully melted before the cooling begins.
Looking at the phase diagrams, as the glaze cools we first get pure silica solidifying, then the alkalis start to solidify as well, e.g. K2Si2O5 at 831C or Na2SiO3 at 1053, and as the temperature falls we get compositions of increasingly higher alkali and lower Si forming. Looking at the resultant structure of the solid, we get "flakelets" of high silica composition, with most of the alkalis sitting on the surface. These are separated by channels of high alkaline, low silica composition - all on a small scale, with the flakelets typically just a few microns across. This is worth remembering for when looking at things like phase separation, and glaze durability.
Silica, Alkali Metals and Alkali Earths
Adding alkali earths
Whilst the alkali metals have a charge (or valency) of +1, the alkaline earths all have 2 electrons in the outer shell, so have a valency of +2, as they are at their most stable when they have emptied the outer shell. This means that they can bond to two oxygen atoms in the silica matrix. Generally they bond to oxygens that are in separate silicon tetrahedra, as this is more stable than bonding to 2 oxygens in the same tetrahedron.
The great thing with this is that this means that an alkali earth effectively connects two tetrahedra - though the ionic bond is a lot weaker than the covalent bonds between silicon and oxygen. Thus they weaken the silicon matrix, rather than breaking it up, as the alkali metals do. This is the reason why alkali earths generally make more durable glazes than the alkali metals. But they cannot bring the melting point of the glaze down as far as the alkali metals, so we do need a proportion of alkali metals to get the glaze liquid, and then other materials can dissolve into the glaze.
Zinc is often bundled in with the alkali earths, as it too has a valency of +2. But, as we'll see in a later post, its behaviour is quite different in some ways.
The chemistry of how they breaks the bonds is very similar to last week's (q.v.), and again the alkali earths tend to form on the outside.
The image shows silicon (blue), oxygen (red), magnesium (pink), calcium (magenta) and strontium (purple), plus a couple of sodiums (green). Note how everything is bonded to everything else, apart from the two sodiums strung out on their own.
Periodic Table
Key to understanding the chemistry of ceramics and glazes are the elements and their arrangement in the periodic table.
With many people nit working today, and gathering their families round for a day of eating, drinking, receiving presents and arguing (and, for those so inclined, going to religious services being held by whichever cult you prefer}, here's a bit of a break from the joys of the festive season, an aide memoire to the elements by Tom Lehrer: https://bit.ly/3GiaW7z
For those of you into wrestling with wild clay, or crushing rocks for glazes, the Aristotelian version is probably more appropriate.
For those who haven't come across him before, Tom Lehrer was a mathematician and musician who write and performed music hall style songs with a particularly realistic view of life.
Getting on in years now, earlier this year he put all his recordings into the public domain, and they can be downloaded from there. So, if you like The Elements, have a browse and up your quota of Christmas Cheer!
Or, if his songs don't do it for you, the image shows the online periodic table from ptable.com, which is full of fascinating facts about everything to Flerovium and beyond.
0.3:0.7 - Help or Hindrance?
Some of you, no doubt, have heard of this ratio of alkalis to alkali earths in a glaze, purportedly guaranteed to create a stable glaze.
So where does it come from, and how reliable is it?
The oldest reference I found is in Parmalee's book Ceramic Glazes, where he attributes it to Hermann Seger, the scientist at the Royal Porcelain Factory in Berlin in the 1860s. A ratio of 0.3K2O:0.7CaO is described as the most satisfactory choice for porcelain (fired, at the time, to Cone 9 - 12). I just looked up Seger's Collected Writings, and this ratio was determined as the eutectic (or lowest melting point) for an Si/Al/K/Ca glaze - so nothing about glaze durability, and no generalisation to other alkalis, alkali earths or anything else. I also found that, in looking for for lead-free glazes, he said that the ratio can be anywhere between 06:0.4 and 0.2:0.8 to produce a high gloss, smooth surfaced, clear glaze (dishwashers had yet to be invented, and he is dismissive about leaching into food and drink, which wasn't a concern until the 1960s).
Parmalee quotes Seger in his chapter on lead-free glazes, and also gives eutectics found by others at different temperatures, e.g. Bourry (working at Sevres) found a ratio of 0.55:0.45 worked best at earthenware temperatures.
So the origins of the 0.3:0.7 ratio have absolutely nothing to do with glaze durability.
The other oft-quoted piece to this jigsaw are the talks by Matt Katz at NCECA: mid-Temperature Glaze Science (2012) and Glossed Over: Durable Glazes (2016). These make a number of assumptions, without evidence to back them up, and often incorrect, e.g.
- that caustic glaze attack in a dishwasher is the same chemistry as acid food attack - whereas in a caustic environment the whole glaze matrix dissolves, whereas in an acid environment it is more likely that the glaze is attacked by ion exchange
- that measuring the loss of gloss in a glaze corresponds to the amount it is attacked. This is true when you get pitting from glaze dissolution, and from the deposits that form in attacks from rainwater, but not necessarily so from ion exchange
etc etc.
So let's not get too hung up on following the 0.3:0.7 ratio, please!
Phase Separation
Before going on to look at what happens when we add additional materials to our glaze, particularly those that may be glass formers or that may result in crystal growth, we need to pause to look at phase separation, as thos often comes into the picture.
One source of phase separation is when we have two liquids that are immiscible, in at least some proportions - oil and water in mayonnaise is one example of this. If we have two substances A and B, and go from 100% A to 100% B, then as we increase B initially there is so little that it will dissolve into A, forming a homogenous mixture. But there comes a point (not necessarily close to 0) when B separates out into the isolated little globules we see at the bottom of the pic, suspended in a sea of A. Carrying on adding B, these globules expand and merge into one another, until we get the 3-d intermixed threads of A and B shown at the top of the pic. As the amount of B continues, it overwhelms A and we get the separate globules of A in a sea of B, before traces of A are fully dissolved in B, and we have a homogenous liquid.
The other route to phase separation is when crystals form in or, more frequently, on the surface of the glaze - potters call this crystalline glazes, or scientists call it glass ceramics. Initially our molten glaze is a homogenous liquid, with tiny crystals continuously forming, and then dissolving away as they haven't got enough critical mass to form on their own. Unlike the first situation, the liquid is perfectly happy in its homogenous state. But then something happens to create a larger starter crystal that is stable enough to start growing rather than dissolving - this seeding of the crystals may be due to something physical, such as a roughness on the clay surface or the glaze meeting the edge of the pot; or the glaze may be slightly cooler, so more viscous and less likely to dissolve the crystal; or it could be the presence of impurities that are stable and exist outside of the glaze matrix, with a structure similar to that of the crystal - titanium dioxide is frequently added to help promote crystal growth.