Science
Whilst some ceramics instructors may talk about the chemistry or the science behind their courses, few actually go beyond concepts like the UMF and Stull, both of which are now over a hundred years old, and much is based on ad hoc experimental results. And, whilst there are arguments about maintaining the history of our craft, to my mind we should take advantage of the advances in science that have been made over the last century, whether to improve our making or just to have a better understanding of what we are doing.
So welcome to Science Sundays! Every week I'll do a post about some of the science behind ceramics - whether it is something basic such as understanding chemical notation, or something deeper such as understanding what happens to our materials as they melt, or how colours are developed in glazes.
To begin at the beginning...
The basic component of chemistry is the atom. This is made up of a nucleus containing positively charged protons (charge +1), and bulked out with neutral neutrons that have about the same weight and size, but no charge. Spinning round the nucleus are the electrons, which have a negative charge (-1), next to no weight, and are immeasurably small. The basic state of an atom is that the number of electrons matches the number of protons, so they have no charge as the + and - charges cancel out.
You'll find quite a lot of abouts in my descriptions - I'm keeping things as simple as they can be for us potters, but dropping in words like about so those with curious minds can go and forage further into the matter.
Each element is identified by its having a unique number of protons - the atomic number - and an atomic weight, which represents the relative weight of the atom. For our purposes, protons and neutrons have a weight of 1, and electrons can be ignored, so the atomic weight is simply the total number of protons and neutrons, beginning with Hydrogen at 1 and steadily increasing (I'll come to them not being exact whole numbers in anther post). If we counted out 6x10^23 atoms, their weight in grams would be about the same as their atomic weight.
Coming back to the electrons, there are various models of how they are organised. As a starting point, we'll go with the idea of electron shells. Here there are a number of layers of electrons going round the nucleus. The innermost one holds just 2 electrons, then 8, 18, 32... as we work outwards. All electrons in a given shell have about the same energy level, which increases as we work outwards. Hydrogen has its single electron in the first orbit, and as the atomic number of elements increases, the electron shells are increasingly filled up.
You may think why do I need to know this? Well, atoms go one better than Lady Bracknell and her handbag - they can both gain and lose electrons, the charge of the atom changing as this happens. This is key to how atoms can join together as molecules, to be covered in another post.
Electrons
Let's look a bit closer at the electrons orbiting the nucleus in their shells, which are made up of subshells, but that needn't concern us here.
The shells and subshells are not wholly filled in most elements (the exception being the noble gases like neon and xenon). When all are fully filled, the element is very stable, but as they become increasingly part filled they also become less stable, and are hungry to become more stable by either gaining electrons to fill up the gaps, or by losing electrons to vacate a subshell. This is called the oxidation state, whole numbers representing the number of electrons gained (negative) or lost (positive). Some combinations of electrons in their subshells are more stable than others, so different elements take on different oxidation states, and some (like iron) can have more than one oxidation state.
Some compounds, such as magnetite (or Fe3O4 - the 3 and 4 would be subscripts if I could do that here!) may appear to have an oxidation number that is not a whole number - oxygen has a coordination number of -2, so this would give iron an oxidation number of 8/3. But closer examination shows that magnetite is actually a mixture of FeO and Fe2O3, with the iron having oxidation states of 2 and 3.
The image shows the major oxidation states of the elements widely used by potters. Other oxidation states can exist, but are less stable and so are found less often.
When referring to chemicals, the oxidation state can be shown using roman numerals. This is normally shown in brackets, e.g. Fe(II), or iron (II) oxide for FeO, though sometimes the roman numerals may appear as a superscript. For chemical mixtures, such as magnetite, they may be denoted as Fe(II, III)3O4 in situations where it isn't necessary to give the breakdown of the mixture, FeO.Fe2O3.
For those who studied chemistry a number of years ago, you probably learned about valencies. Oxidation states can be seen as a further refinement of valencies.
So what happens to these charged atoms, and where do the electrons come from or go to? That's the basis of chemical bonds, in another post.
Atomic bonds
As we saw last week, many elements do not have a very stable outer shell of electrons, and would be more stable if they could gain or lose some. And this is made possible by atomic bonds, which can be seen as atoms combining to have a more stable configuration of electrons. There are 3 primary types of atomic bonds:
Metallic bonds are where pure metals or alloys lose excess electron, which encompass the atoms in a soup of electrons. We won't deal with this further, as potters don't really deal with metals (as opposed to compounds of metals plus something else, e.g. oxides and carbonates).
Ionic bonds are where excess electrons in one atom are transferred to another atom with a shortage of electrons, leaving the donor atom with a positive charge, and the recipient with negative charge. This change in charges results in the atoms being attracted to each other, but there is no physical connection between the two atoms, so they can easily switch which atom they are bonded to. If dissolved or melted into a liquid, and a pair of electrodes are put into the liquid, a current will flow as the atoms are attracted to the oppositely charged electrodes. Examples of ionic bonds include salt (NaCl), and fluxes bonded into the glaze matrix.
Covalent bonds are where electrons are shared between atoms - you can envisage this as the outermost electron shells overlapping, and the shared electrons orbiting round both nucleii. Thus the bond is between a specific pair of atoms, unlike ionic bonds, and this results in very strong bonds between the atoms. In some cases, such as diamond and glazes, the bonds form a continuous structure, which has a high melting point and is very strong but brittle, such as diamonds and the silicone, aluminium and oxygen of the glaze matrix. In other cases, such as carbon dioxide, there is no extended structure, and the solid and liquid depend on weak inter-molecular forces, and so have little strength and low melting and boiling points.
So far this is a very idealised view of bonds - we'll look at some of their complexities next week, as well as how it is decided which type of bond will form between a given pair of atoms.
More on Bonds
We looked at the two types of bonds between atoms found in ceramics: ionic bonds where electrons leave one atom to join another, and covalent bonds where electrons are shared, spending their time spinning round two atoms.
But this is a bit of a simplification and, as with so many things nowadays, bonds lie on a spectrum of which ionic and covalent bonds are the extremes. All the inbetweenies are called polar covalent bonds - here the electrons spin round both atoms, but spend more of their time round one than the other. If you average out the time spent on each atom, you'll see that those where the electrons spend most of their time effectively have a negative charge, with the others having a positive charge, and the total charge being zero.
So what determines which type of bond is formed? This brings us to electronegativity, which ranks the tendency of an atom to attract electrons. If two atoms have the same electronegativity then shared electrons will spend an equal time between both, forming a perfect covalent bond. As the difference in electronegativity increases, the electrons will spend more time at the atom with higher electronegativity. A difference of less than 0.5 is generally seen as purely covalent; over around 1.7 to 2.0 is ionic; and in between is polar covalent.
The image shows electronegativity on the periodic table, highlighting potter's materials.
So what does this mean? Well in ceramics we're concerned with oxides, and oxygen has an electronegativity of 3.44, second highest to Fluorine at 3.98. None of the materials potters regularly use will form a purely covalent bond with oxygen, but Fe, Co, Ni, Cu, B, Si, Sn, Pb, P and Sb all form polar covalent bonds. Meanwhile, all the fluxes plus Zr are ionic bonds, and Ti, V, Cr, Mn, Zn, Cd and Al are in the grey area between the two.
And in practice we'll see the effect of this in several ways: the polarity of water molecules affecting how clay wets out; the polarity of tin oxide being key to the formation of copper red glazes; and which elements go where in the UMF (which Matt and Rose Katz have also been working towards from an experimental approach).
Naming of Parts
Just as in Henry Reed's poem from the 2nd World War about conscripts learning the names of the parts of the Enfield rifle, we need to learn what to call the various chemical substances that we deal with, both in plain language and as chemical formulae.
The formulae are easiest. First, each element is given a 1 or 2 letter symbol, e.g. Fe for iron. Then, for a molecule, we give the most metallic elements first, and a subscript suffix giving their number (you'll have to imaging the subscripts and superscripts here!), for example H2O for water.
In written form, for simple ionic compounds the non-metal is given the suffix -ide, e.g. sodium chloride for NaCl. If the metal can have multiple charges, then we put the charge in Roman numerals in brackets, e.g. FeO is iron (II) oxide and Fe2O3 is iron (III) oxide.
If it is an ion on its own, as opposed to an ionically bonded compound, it has a superscript suffix giving the charge, e.g. NaCl is solid salt, but dissolve it in water and it becomes Na+ and Cl- ions.
Where ions are made of multiple atoms (such as the carbonate ion, CO3 2+), things are a little less clear cut, with suffixes -ate and -ite depending on the oxidation state - but luckily this is not relevant to the materials potters generally use.
For covalent (or molecular) compounds, proportions of elements are not restricted to the charges on the atoms, so we need to be more explicit as there are many more possibilities. So we prefix the name of the non-metal with the number of atoms, e.g. carbon monoxide or carbon dioxide - of course carbon isn't a metal, but it is more metallic than oxygen, which we'll touch on next week. Sometimes the mono is left out for brevity. Where necessary a prefix is used for the metal too, e.g. dihydrogen monoxide is the formal name for water. But older common names persist, such as nitric and nitrous oxide, and water!
Idiosyncrasies exist, harking back to poorly understood chemistry, e.g. bicarbonate refers to a carbonate ion with a hydrogen atom, and the OH- ion referred to as hydroxide
Older writings may use other naming systems. These include older names for elements, such as ferrous for iron and stannic for tin; and using the suffix -ous for the lower charge and -ic for the higher, e.g. ferrous oxide instead of iron (II) oxide (and prefixes hypr-, per- and hyper- available to extend the range). And often shorter names are used for brevity, where the meaning is clear, e.g. iron oxide instead of iron (III) oxide.
Potters, of course, have their own misnomers, e.g. many use chrome instead of chromium, confusing the shiny metallic coating with the element, or using the colour to differentiate compounds, such as red or black iron oxide.
If this naming of parts excites you, or you just want to make sense of a name, you can download the full IUPAC Nomenclature or Inorganic Chemistry here: https://bit.ly/3RZOZ0M
The Periodic Table
Most of the elements were discovered between 1750 and 1900, and several chemists proposed ways to organise them, so that those with similar properties are grouped together. In 1869 the Russian chemist Dmitri Mendeleev created a pattern that is basically the same as the current Periodic Table, and then the last major change was in 1913, when the Dutch physicist Antonius van den Broek proposed that the atomic number was used (i.e. the number of protons), as opposed to the atomic weight that Mendeleev used.
Hydrogen, with an atomic weight of 1, is at the top left. the atomic number increases from left to right, then top to bottom. The gaps occur because the elements are organised so that those in a column have similar properties - although it wasn't known initially, this is due to the configuration of electrons in each element.
The rows are called periods, with 1 at the top, and the columns are groups, with 1 at the left - though many of the groups also have common names, such as the alkali earths for Group 2, and Halogens for Group 17. The two rows normally drawn below the main body of the table should be drawn in the bottom two periods, but are moved below to save space - and they are not included in the groups. The table can also be subdivided into blocks, named s, p, d and f based on their electron orbitals. There is a bit of debate over how some elements should be grouped, but generally that doesn't affect the materials we deal with.
The image shows the periodic table, with the names of the various groupings shown. In future posts we'll look at these groupings in more detail, to understand why they behave as they do in ceramics. The one name you won't see is "heavy metal" - this is a meaningless populist phrase, with almost as many definitions as there are elements.
For those wanting to investigate the table and grouping of atoms, I suggest the online periodic table at https://bit.ly/3oGnPyv . And for an alternative ordering, look up the song "The Elements" by Harvard mathematics lecturer Tom Lehrer.
Oxygen
Which element is most important in ceramics?
I bet most of you are thinking that it must be silicon, as the basis of both silica and clay.
But I'd suggest that it could be oxygen.
First, let's look at the finished ceramics. If we put aside materials that didn't melt in the firing, such as grog or coarse silica in clays, the UMF (or Seger) view of the finished product is seen as a mix of glass formers (like silicone), stabilisers (like aluminium), modifiers (e.g. sodium) and colourants (e.g. copper) - but this forgets the oxygen to which all of these are bonded, either with covalent bonds for the glass formers and stabilisers, or ionic bonds for the modifiers. Without the oxygen we'd have a metal alloy, not a ceramic. And whilst silica is the most common glass former, we also often use boron and phosphorus, and there are many others such as vanadium.
Secondly, without oxygen we'd have no water, without which the forming of clays and application of glazes would be even more troublesome than most potters already find their craft. And we need water because of its specific properties that allow it to stick to and lubricate the clay particles, and keep our glazes in suspension - neither of these would work with most other liquids.
Behind all of this, there is the chemical nature of oxygen. It's outer shell of electrons has two vacancies, giving it a charge of -2. This enables it to act as a link between other atoms, so it can link to 2 positively charged atoms - the basis of large scale crystals such as those of silica or alumina. It is also highly reactive (only beaten by fluorine out of the materials potters use), and reacts with all the other elements that we use. And, of course, a large proportion of the air we breathe consists of Dioxygen (O2), where two oxygen atoms are covalently bonded by each sharing 2 electrons with the other. (As a side note, you can also get peroxides, where adjacent oxygen molecules only form a weaker single bond, such as hydrogen peroxide H2O2 or barium peroxide BaO2. These are called peroxides as shorthand for hyperoxides, as they contain more oxygen than the normal oxides, H2O and BaO).
Perhaps it is ignored so much because we don't get recipes that include oxygen as an additive - it has already reacted with everything else we use. But my vote goes to oxygen.
All in a Name
What do we mean by terms like ceramic or glaze, and do scientists mean the same thing? If the words have different meanings in these two groups, we need to do a mental translation when reading any scientific literature in the field.
For us potters, the word ceramic comes from the Greek keramikos, meaning to do with clay or tiles - this word was long extinct, and only came into use around 1850 when Hellenomania meant Western Europe decided that its roots lay in (an idealised view of) ancient Greece (and Rome). This meant that using Greek and Latin words became hip, your gap year was spent looking at Greek and Roman ruins, and having a "good" education meant learning Latin and classical Greek - plus it allowed the groundwork to be laid for racial purity and a master race. It's really quite impressive how Western Europe reverse-adopted Greek civilisation as its parent, despite its tenuous links. And when I was in Athens, my ability to ask for a chariot to take me to Sparta was of limited use, and an ability to blurt out classical aphorisms didn't help poor old Boris!
So to us ceramics is the posh boy of pottery (which came from 13th Century old English and French, possibly with Germanic roots). Potters produce earthenware tableware, ceramicists create porcelain sculptures! But both are based on clay, which is made of kaolinitie, containing, silicon, aluminium, oxygen and water, though other things like feldspar and grog may also be present in the clay.
But to the materials scientist, ceramics encapsulates a much wider range of materials. They see ceramics as any non-metallic, inorganic solid - of which glasses and crystals are subsets. As alternatives to the potter's alumina-silica oxides they may use aluminium or zircon or other metals formed into oxides, nitrides, carbides and silicides (yes silica, but this time with a negative charge). So, as well as the potter's products, they include glass, semiconductors, concrete, optical fibres, hip joints, ceramic knives, tank armour and more.
Glazes and glasses and crystals will be a future post, as my interest in etymology and culture took over too much of my interest in science today!
PS Spot the image links!
Glass
After looking at what we mean by ceramics in general, the next thing we need to look at is glass - this is what all fully melted, non-crystalline glazes are made of, and it is also what is formed when part of the clay body melts, to glue together all the unmelted components.
So what is a glass? Until 100 years ago, it was thought to be silica based, but now we have glasses from a number of inorganic materials, as well as metallic and organic glasses - so a glass does not need to be a ceramic material. But two properties are found in all glasses.
First, there is no long range atomic arrangement - glasses are not crystals, even if over a short range some areas have crystal-like structures.
Secondly, it exhibits glass transformation behaviour over a temperature range called the glass transformation region. And so what is this behaviour? Whilst a non-glass material will form a regular crystal when cooled below its melting point, for a glass it does not solidify at the melting point, but becomes a super-cooled liquid. Then, at the glass transition temperature, Tg, the start of the glass transformation range, the atoms start moving towards a crystalline structure, but because the liquid is supercooled it is highly viscous, and this happens slowly. Eventually it becomes so viscous that the atoms cannot move any more, and this setting point is the bottom end of the glass transformation range. This is shown in the image, where the enthalpy is effectively the amount of energy of the atoms. By extending the two straight lines above and below the transition range to where they meet, we get to the fictive temperature, Tf, which can be used to put the glass transformation characteristics into a single number.
The other thing to note it is the effect of the rate of cooling. The more slowly it cools down, the more order develops, so ultimately the melt forms a crystalline structure, not a glass. Also, slower cooling gives a smaller glass transformation range that occurs at a lower temperature.
Now we know the essence of what a glass is, we can look at the various approaches to its formation and structure in a future post.
Glass formation
There is no single comprehensive model for predicting what will form a glass, but there are basically 2 approaches: the structural one, looking at the final structure; and the kinetic one, looking at the behaviour when cooling. Here we'll look at the structural approach, leaving the kinetic one to next time.
Zachariasen developed the first model, where instead of the silica tetrahedra being joined in a regular fashion they are joined in a random manner. The rules he came up with are:
1 each O is bonded to no more than two glass forming ions
2 the coordination number of the glass formers is 3 or 4
3 the polyhedra only share corners, not edges or faces
4 at least 3 corners of each polyhedron are shared, so creating a random continuous network
He was the first to differentiate between the glass formers and modifiers (what we call fluxes). Also, his work suggested how the modifiers are part of the structure (of which more later)
K.H.Sun, and then Rawson, looked at bond strengths to categorise the elements, and introduced intermediates (which may be a former or modifier). This produced the following (strongest glass formers and weakest modifiers first):
Formers: Al, Si, B, V, P, Zr, Sb
Intermediates: Ti, Pb, Zn, Zr, Al, Cd
Modifiers: Sn, Pb, Mg, Li, Zn, Ba, Ca, Sr, Cd, Na, K, Cs
You'll note that some elements appear more than once - this depends on the coordination number they take up.
or by Rawson:
Formers: Pb, B, Sb, Si, Sn, V, P, W
Intermediates: Ti, Al, Zr, Zn, Cu, Fe, Ni, Cr, Mn, Co
Modifiers: Ca, Li, Sr, Ba, Cd, Mg, Na, K
This showed how Boron can be such a good glass former, and the importance of eutectics in glass formation.
Dietzel looked at the field strength (the ratio of valency:inter-atomic distance), giving a slightly different grouping. It also predicts the likelihood of phase separation when there is more than one glass former: the bigger the difference, the more likely there will be phase separation, or one of the glass formers will crystallise out.
Formers: P, Si, B
Intermediates: Ti, Al, Fe, Be, Zr, Mg
Modifiers: Mn, Fe, Ca, Sr, Pb, Ba, Li, Na, K
None of these is perfect, and other approaches have been tried too.
Glass - Kinetic Aspects
We've looked at which materials could form a glass/glaze based on their chemical properties alone. This week we'll look at kinetic aspects of glass formation, i.e. how atoms behave during the formation process, and how the dynamics affect whether or not a glaze is formed.
Let's start at looking at what happens when a material goes from liquid to solid normally. First of all a number of nucleii form. As the liquid cools, the tendency to form nucleii, and their capability to grow large enough to be stable, increases. But against this the liquid is becoming more viscous, so from this point of view the ability to form a nucleus is decreasing. This means that there is quite a narrow temperature range within which stable nucleii will form.
Similarly for the growth of solidified material around the nucleus - crystallisation. This occurs at a lower temperature than nucleus formation, but there is a balance between a lower temperature resulting in faster crystallisation, and a lower temperature reducing the rate of crystallisation due to increased viscosity. So again we get a temperature window within which crystalline solidification will occur. Outside of this area, a glass will form.
Putting this together, we get the graph shown, with blue indicating a glass, and white a non-glass solid. The minimum cooling rate to form a glass is given by the line from the melting point Tm that is tangential to the curve. If we cool slower than this, then the atoms have time to move into their crystalline positions, and no glass is formed. If we cool faster, we get a glass.
This curve tells us several things. First, a high viscosity at Tn, the temperature that is worst for glass formation, will help. Our silica and boron glazes have this all the way from the melting point down.
We also want to minimise nucleii. Having a complex glaze recipe with lots of constituents helps, as the atoms find it harder to congregate into nuclei; also adding strong fluxes such as lead and the alkali metals.
So, whilst measuring the curve for a given mix is difficult, this approach gives us a complementary view of glass formation, particularly showing the importance of viscosity.
Melting
Understanding how our clays and glazes melt is key to being able to work with them successfully, as shown in this piece by David Louveau .
So I'll do a series of posts explaining exactly what happens to clays and glazes in the kiln.
But first, a top level view of what happens in the melting process.
As the material heats up, various things may happen whilst it is still very much a solid. The molecules may become unstable and break down into smaller molecules via thermal decomposition, often giving off things like water, carbon dioxide or oxygen as gases. In the case of organic matter, this may break down to the extent that it all escapes as gases, often grabbing oxygen from the atmosphere or other molecules around it to help with this. And, if molecules get hot enough, they may start to melt (e.g. sodium, potassium or lead).
But before they melt, there are likely to be solid state reactions where particles touch - sintering. These reactions are slow because solids are chemically more stable than liquids, and the contact area is smaller, but the temperature at which these start is often surprisingly low, starting below 500C. The product may be solid or liquid, but let's stick with solids for now.
As these reactions proceed, our lump of matter shrinks a bit, and the edges and corners are rounded off, as the reaction generally results in voids being filled, but the overall shape is unchanged. If there were no sintering, and no materials decomposing, we would expect the size to increase with temperature. From here on I'm going to use a commercial high temperature glaze as an example, where the process has been studied in depth. For this glaze, sintering occurred between 1125 and 1175C - though the kiln temperature is rising much faster than the rate of sintering reactions, so at 1175 sintering would not be complete without a long hold to give time for both the heat to penetrate our sample, and for the sintering reactions to complete. In this study, the sample was a cube, and the stages were determined by the size and profile of the sample, rather than by trying to determine the state throughout the sample.
Next, melting starts - either from the raw materials reaching their melting point, or due to the product of the solid state reaction being a liquid, or material dissolving into the liquid. Here we see continued shrinkage, and a change in shape, as the faces of the cube start to bulge out, and the edges pull in - effectively the cube is trying to become a sphere as it becomes increasingly liquid, due to surface tension. At 1300C, this sample hits its maximum bulge, and minimum contact area with the surface it is resting on. Beyond this temperature, decreasing viscosity means that the sample will first tend to become a hemisphere (reached at 1352C), and then flatten out further as it becomes even less viscous (the point where the sample height was 45% of the original was taken for this, reached at 1360C for this glaze).
In case you're wondering, the tests followed a fast firing for a commercial porcelain, going from cold to 1390C and down to cold again over just 5 hours.
During the cooling stage, two more temperatures are important. The softening temperature, Ts at 790C is when the sample is stiff enough not to deform under its own weight, and Tg at 773C can be seen as the annealing temperature of the sample, below which internal stresses will not be relieved by heating.
Solid State Reactions
We talk about things melting in the kiln, but the process starts with solids reacting in various ways.
One of these is in solid state reactions, where compounds in touch with each other react by ions diffusing from one compound to another. This can start at surprisingly low temperatures, often comfortably below 1000C. The end product may have a high melting point, as with many of the the aluminates or silicates, so may not physically melt at kiln temperatures; or may have a low melting point, like the lead silicates and products of reactions between fluxes and/or colourants.
For these reactions to happen, as well as the two substances being physically connected, there also needs to be enough energy for the reaction. This may come from heat, giving the atoms more energy, or there may be weaknesses in the structures of the substances - these may be points or edges, or cracks, or voids or intrusions of other substances into the structure. The reactions are slow compared to those in liquids, as they rely on physical contact between the reagants, and then the ions diffusing from one to the other.
Real world examples of these effects are finely ground substances entering the melt at a lower temperature, as they have a larger surface area to volume ration, so proportionately more material is exposed to surface reaction, plus in milling to a smaller size more cracks and other defects will have been created. Another example is in flint (the rock) having a lower melting point than silica from sand, as in production it is calcined to make it easier to grind, and the calcining will introduce more defects.
As solid state reactions proceed, these tend to fill up voids in the structure, and there will be some shrinkage. This brings the materials into closer contact with each other, thereby speeding up any possible reactions.
In case you're wondering, the image is of Ben Hope, in the north of Scotland. I couldn't think of a decent image to illustrate the topic directly, but scree slopes down mountains are a good example of solids mixing, and I used to go up this one regularly when working as an outdoor activities instructor up there in my youth
Water
Today I'll talk about water, and in particular how many of the materials we use have water attached to them. These compounds are referred to as hydrates, or the hydrated form, as opposed to the anhydrous form with no water attached. And, unsurprisingly, the water that comes off is called the water of hydration.
For inorganic compounds, the water is generally only loosely bonded to the rest of the molecule. Water, being a polarised molecule, attaches on via hydrogen bonds (a type of van de Waal's force) - the positively charged hydrogen in water attracts a pair of electrons from the oxygen atom.
An example of this is magnesium carbonate, which can form di, tri and pentahydrates, bonding with 2, 3 or 5 water atoms. The trihydrate is the most common, MgCO3.3H2O. Potters way also come across references to light and heavy magnesium carbonate - light is hydromagnesite, Mg5(CO3)4(OH2).4H2O, and heavy is dypingite, Mg5(CO3)4(OH)2ยท5H2O. All of these first lose their water on heating (at 200 - 450C), and then the carbonate, becoming MgO. The only question is to know how much water and carbonate there is, to weight out the materials correctly to get the right amount of MgO (the rest being lost as part of the Loss on Ignition, LOI). Other materials with this loosely bound water (with decomposition temperatures in brackets) include aluminium oxide (220C), cobalt carbonate (140C), colemanite (105 - 420C), gerstley borate (45 - 420C) and nickel carbonate (90 - 200C).
For some materials, the water is bound in much more strongly, with ionic or covalent bonds. Clay is an example of this (covered elsewhere), as is talc (hydrated magnesium silicate, Mg3Si4O10(OH)2). On heating, the water comes off from the (OH)2, with the remaining oxygen forming MgO. But the structure determines the high temperature at which this happens. Talk is made of sheets of 3 layers: a middle one of Brucite - Mg(OH)2 - sandwiched between two layers of silica tetrahedra. So, to release the water, the whole structure of talc must be destroyed, which happens at over 900C.
The other example is bone ash. Here the water is ionically bound to calcium, and is given off at 800C and above.
Thermal decomposition
Many materials we use break down during the firing of their own accord, not through interacting with other materials. The prime examples are the carbonates breaking down to the oxides, but also some oxides break down to simpler forms. So what happens?
It all comes down to the strength of the bonds, and chemists have a number of ways of looking at this which, for our purposes, are near enough the same. One measure is the reactivity series, and another is the electropositivity of the element. The less reactive, or electropositive, an element is, the weaker the bonds, and so thermal decomposition is more likely and will occur at a lower temperature. Putting this onto the periodic table, as we move from the bottom left towards the top right, the greater the likelihood that thermal decomposition will take place. Or, of the materials we use, the list below starts with those most likely to decompose, and ends with those least likely:
Cu (200-290), Sb, Pb, Sn, Ni (270-420), Co (335), Cd, Fe, Cr, Zn, Mn (200-400), Ti, Al, Mg (510-550), Ca, Sr (667-1497), Ba, Li, Na, K
The numbers in brackets are the decomposition temperatures (in C) of the carbonates. These should be seen as guidelines - decomposition starts very slowly at lower temperatures, and the rate increases with temperature, so is quite hard to measure in absolute terms, and the data has come from various sources.
If you look at the image, the strong positive charge on the calcium ion means that the electrons on the carbonate ion are pulled towards the calcium (darker green), and spend less time at the carbonate end (paler green). This means that there is a strong bond between the oxygen nearest the calcium, and weaker bonds between the carbon and other two oxygens. So, as things heat up and the atoms are vibrating around with more energy, eventually the carbon dioxide comes off as a gas, leaving calcium oxide behind. Very similar things happen when an oxide or hydroxide decomposes. Sometimes it is a multi-stage reaction - for manganese and cobalt, there are several stages in going from the carbonate to the oxide that is most stable at kiln temperatures.
In praise of complex things
Some potters like to mix up their materials from the simplest, purest materials possible - going to pure quartz, kaolin, and the oxides and carbonates that we use. This does have the advantage that they know exactly what is going into their materials (assuming that what comes out of the bag corresponds to the analysis that they have, but that's a different story). But it also has a significant downside. Mixing up these pure materials creates just that, a mixture - there is no chemical interaction between the components until they are fired.
If we look at materials that consist of multiple components - ball clays, feldspars, slips, frits and stains - the components are generally much more tightly bound - it could just be from being very well mixed and compressed, or it could be that they have been melted together. This has a number of potential advantages. The better mixing and greater surface contact means that they will generally melt earlier, and will form a truly homogenous mix more easily. If they have been subjected to heat, or enough pressure, some of the reactions that we want to occur in the kiln may have already started, so they will fire more easily, and may have given off gases that could cause problems. The downside is that for the naturally occurring materials we are unlikely to know their precise composition - but often that doesn't matter, so long as we have a consistent supply. And often another benefit of using less processed materials is that they tend to be cheaper (though not for frits and stains).
So whilst there are benefit in using pure materials when developing a new glaze, once developed it is worth seeing if you can reformulate it to use more complex materials. Swap the kaolin out for ball clay (with suitable adjustments). But beware that if the original raw material is no longer available the suppliers may just offer a blend of materials, with the same chemistry but different behaviours. Secondly, some raw materials present problems that may outweigh any advantages - the one that comes most readily to mind is when the raw material is soluble in water.
Something for you to experiment with!
Melting (again)
When looking at the melting of a clay or glaze, many people will look at the overall chemical composition of the mixture, then turn to phase diagrams and pick out the lowest temperature eutectic as the starting point of the melt, then the next one up, and so on, hopping from eutectic to eutectic until everything is melted.
But this assumes that we have a well mixed, fine grained homogenous material, so everything is in contact with everything else. In the real world, this is not the case. Furthermore, the melting of feldspar, one of our basic materials, is complex.
Most feldspars are a mixture of sodium and potassium minerals, rather than a single mineral containing both sodium and potassium. When heated, they undergo incongrous melting - this means that whilst part of the feldspar melts (generally the sodium containing part), part will stay solid until a significantly higher temperature. Plus, in a clay or glaze some of the melt behaviour is reliant on diffusion (as opposed to all particles melting and mixing together), which is sluggish. This is why many people say feldspar has a wide melting range, rather than a single melting temperature - it actually has two melting points.
The feldspar will begin to melt on the outside surface of grains, and also along cracks and fault lines extending into the grains. Sodium and potassium leaves the feldspar and reacts with silica (both from decomposed kaolin and from silica in the body) to form a melt, whilst calcium, magnesium and iron diffuse into the feldspar. If the body also contains a calcium feldspar, then sodium, calcium and aluminium join the silica melt, whilst potassium, iron, magnesium and silica migrate into the feldspar
In time, the feldspars melt fully, but do not form a homogenous mixture with the rest of the clay or glaze. Instead, the materials pretty much stay in place, so although the clay or glaze may melt fully, the composition will very across the piece. Also, a melt halo forms round the outside of each particle, where the composition s that of the surrounding silica, modified by the materials that have melted in and out of the feldspar.