How Paints Work - Part 2
Knowing what is supposed to happen can help if you are making your own.


The paint's medium can be almost anything that will:

  • hold the pigment or dye so that it can be seen;
  • allow light to pass through the pigment and then onto the viewer;
  • preferably not react chemically with the pigment;
  • be flexible enough for the support that it is on so that it doesn't crack over a short period of time; and,
  • not shrink or expand significantly over time.

You need to be able to apply the paint to the surface where some sort of physical transformation will occur and the paint will be in a state that will last. Either: you need to be able to reduce the viscosity of the paint by suspending particles in a relatively volatile liquid or by dissolving the paint in a solvent; or, you need to have a physical process that makes the medium set. You can have both if the medium reacts to solidify but is a bit too viscous without a solvent such as linseed oil. Examples of solvents are distilled turpentine and water.

So, what fits these requirements? Examples of some media are:

Medium Components Application
Mobility Improver Solidification Time Between Applications Time to Finish Flexibility Longevity Colour
Chemical Stability Storage Form Notes Comments   
Oil paints Linseed oil
Poppy oil
Safflower oil
Dilute with solvent Distilled Turpentine Solvent Evaporation then
oxidation/ polimerisation
Days Weeks Quite flexible although it hardens and becomes brittle with age Millennia Yellows a little but with a high enough PVC, not noticed Attacked by strong solvents Metal tube Is often supplied washed by acid (can attack some pigments) or alkali (causes soap formation and metal soaps leading to sweating) One of the earliest examples of painting with a medium is in some caves in Afghanistan where poppy seed oil was used as the medium.   
Encaustic Beeswax with Damar Resin Heat to melt None needed Cooling Down Minutes Hours a bit on the brittle side - only really suitable for stiff supports unless very thin Millennia Very Good Very Good Small Blocks Expensive to buy ready made paints but easy enough to make your own Encaustic dates back to Greco Egyptian times and before.   
Resins Damar Resin or Pine Resin Dilute with solvent Distilled Turpentine
or Alcohol
Solvent Evaporation Minutes Hours a bit on the brittle side - only really suitable for stiff supports Millennia Damar can yellow so high PVC needed Very Good but can be attacked by water or alcohol Made in situ from resin solution It is best to make this when you need it like some of the other media This dries very quickly if you use alcohol as the solvent. IPA is available as 99.9% pure IPA and is cheap. Use tends to be limited to applying coloured washes on things like frames although there is no reason why it shouldn't be used elsewhere.   
Water Colours
Acacia Gum
Dilute with solvent Water Solvent Evaporation Minutes Hours Quite flexible when thin Millennia Very Good Attacked by water Pans or Tubes This is used as water colours on paper but it can also be painted on other surfaces such as plaster and so on.
   When dry, the gum wets the surface of the pigment so the colours remain deep and saturated.
Early use dates back to Egyptian times. Easy to make your own. Honey there as a humectant but some commercial watercolour suppliers sometimes use glycerol which does not go solid.   
Egg Tempera
Egg Yolk Dilute with solvent Water Solvent Evaporation Minutes Hours Quite flexible when thin and young although hardens with time Millennia Very Good. Does not yellow so blues are as good after half a millennium as they were when they were painted. Attacked by water Made in situ from egg yolks and water This is made each time it is used. The medium goes off if it is kept wet and it cannot be re-wetted if it dries out.
   With this medium, it is the water that wets the pigment so when it dries out, the colours loose their depth - the minerals in the pigments are dry and are effectively just held in place by the medium.
This is one that dates back to works that are over a thousand years old. It has to be used when prepared as it cannot be stored for more than a couple of days. The paint cannot be mixed on the work, it has to happen on the palette. It has to be applied in thin layers because if it is applied thickly, it will crack.   
Egg White Dilute with solvent Water Solvent Evaporation Minutes Hours Quite flexible when thin Millennia Very Good. Attacked by water Made in situ from egg whites and water This is made each time it is used. The medium goes off and doesn't last more than a day when wet unless refrigerated.
   It is made by whisking egg whites until the make a froth then leaving that overnight. The liquid that is left when the remaining foam is removed is the glair. This is then mixed with the pigment and water.
This is one that dates back to works that are over a thousand years old.   
Casein from milk Dilute with solvent Water Solvent Evaporation Minutes Hours Finished paint is inflexible and brittle Month Very Good. Does not yellow so blues are as good after half a millennium as they were when they were painted. Very Good Made in situ from milk curds, lime and water This is made from milk by curdling it with vinegar then mixing the separated curds with lime then adding your pigment. It will keep if refrigerated for up to two weeks and when it is dry, it becomes waterproof. Casein paints can be bought from some sellers in tubes although like all commercial preparations, the version that has to sit for months if not years in an art shop and still work when you decide to buy it and use it, is probably going to have a lot of other things in there.   
Gelatin Dilute with solvent Water Solvent Evaporation Minutes Hours Quite flexible when thin Millennia Very Good. Does not yellow so blues are as good after half a millennium as they were when they were painted. Attacked by water Made in situ from gelatin and water This is made each time it is used. The medium goes off if it is kept wet although it can be re-wetted if it dries out and is thin - add some water to it and heat it up by putting the palette over some steam while stirring it.
   With this medium, like the egg tempera, it is the water that wets the pigment so when it dries out, the colours loose their depth - the minerals in the pigments are dry and are effectively just held in place by the medium.
This is one that dates back to works that are millennia old. It has to be used when prepared as it cannot be stored for more than a couple of days wet. The paint can be mixed to some extent on the work but is easier if mixed on the palette.   
Acrylic paints PVA Dilute with solvent Water Solvent Evaporation Minutes Hours Quite flexible Only been around for decades Very Good Attacked by water Tubes Many of the same colours used in oils can be used here although water is the solvent and not spirits so not all are available PVA (Polyvinyl Acetate) is a product of the 20th century chemical industry and therefore its longevity is completely unknown. Being an ester, it can be attacked by acids or alkalies. Many of the colours have titanium white added to them giving the paints an overall pale look.   
Fresco Plaster Dilute with solvent Water Solvent Evaporation then crystalisation Minutes Hours Brittle. Only suitable for rigid surfaces such as brickwork Millennia Very Good. Attacked by anything that attacks plaster such as mould when wet Made in situ from plaster and water This is made each time it is used. The medium goes off quickly and sometimes pieces need to be completed in small areas.
   The medium is alkaline so a lot of pigments are unsuitable.
   With this medium, like the egg tempera, it is the water that wets the pigment so when it dries out, the colours loose their depth - the minerals in the pigments are dry and are effectively just held in place by the medium.
This is one that can last millennia under the right conditions.   
Glaze Glass Suspend in Water Water After water evaporated, fire in a kiln to melt the medium Minutes Days Brittle. Only suitable for rigid surfaces such as metal and ceramics Millennia Very Good. Attacked by Alkali Powder from manufacturer Biscuit fired items are either dipped in glaze or have glazes painted on them. The articles are then fired at a higher temperature where the clay vitrifies and the glass powder melts, completing any chemical reactions to make the pigments. Often the colour of the pre-fired glaze can be quite different to that of the finished article. Colours such as Cobalt Blue (Silicate) start off as Cobalt Oxide which is grey. Also, you can mix a pigment such as cobalt oxide in with the clay which you then use decoratively on the article (such as dividing your stoneware clay into two, mixing pigment in one half, assembling a cylinder with one half with pigment then putting it on the wheel to make a spiraled pattern) and then glaze over it with a clear glaze to bring out the colour.   
Medium Components Application
Mobility Improver Solidification Time Between Applications Time to Finish Flexibility Longevity Colour
Chemical Stability Storage Form Notes Comments   

Pigment Volume Concentration (PVC)

Pigment Volume Concentration is a measure of the ratio of pigment to medium in terms of volume. It is volume that is important because some pigments are denser than others.

Some pigments are light and fluffy (French ultramarine and carbon black spring to mind) and other are quite dense (lead white and cinnabar, for example). The important thing about the ratio of pigment to the medium that it is in is the volume.

If you have a lot of medium, there might not be enough pigment in it to have the covering power that you need. However, if you had a lot of pigment in there, the paint would be more like putty or plaster and would not flow onto the canvas very well. If you diluted such a high PVC paint with thinner, when that evaporated off, there might not be enough medium there to hold the pigment in place.

In short, PVC is a fairly useful guide on how to mix paints for a consistent outcome.

With paints with media such as egg yolks, you should use a PVC of one to one so, a pile of pigment and add enough egg to be an equivalent volume. Then, you can add as much water as you need to but when the paint dries out, there will still be a PVC of 1:1

Solvents / Thinners

Solvents in paint - how it lowers the viscosity of paint by increasing the distance beytween the particles of pigments.
Copyright (c)2020 Paul Alan GrosseWe add solvents to our paints because without them, the paints would not be easy to apply to the artwork - they lower the viscosity of the paint, hence their other name - 'thinnners'.

A solvent effects a lowering of the paint's viscosity in two ways:

  • firstly, it makes the medium less viscous; and,
  • secondly, it increases the distance between the pigment particles.

In order to lower the viscosity of the medium, it needs to be miscible with it or dissolve it. In the case of water colours, the medium is usually acacia gum so the water that you add to it dissolves the gum, making it flow better than the solid lump of paint that you have in the pan - pigment held together that is applied dry is in effect clay pastel. In the case of oil paint, the solvent dissolves in the linseed oil, lowering its viscosity quite substantially. Additionally, it has to be chemically inert as far as the paint is concerned oterwise it would make changes that the artist would not be able to control.

The second effect - increasing the distance between pigment particles now takes effect because the volume of the paint is that much greater for a given amount of pigment so, with greater distances between pigment particles, they are flow over each other easier.

Once the paint has been applied, the solvent needs to evaporate, leaving no trace of it behind - anything that is left behind needs to be a solid or a compound that will turn into a solid - usually be oxidation in the case of oil paints. Anything left that does not become solid will destroy the integrity of the painting. As the solvent evaporates, the medium shrinks, bringing pigment particles closer together and increasing the paint's viscosity again.

So, to summarise, a solvent has to:

  • Dissolve in the medium;
  • Evaporate from the paint in a reasonably short period of time;
  • Leave no practical trace that it was there; and,
  • Not react chemically with the paint or surface.

Hydrophobic Media

Solvents for oil/wax/hydrophobic media tend to be some form of volatile spirit - turpentine, essential oils and so on.

Four solvents:- 
Gum Turpentine,
English Distilled Turpentine,
Steam Distilled Lavender,
Spike Lavender.
Copyright (c)2020 Paul Alan GrosseTurpentine is made from the destructive distillation of resin that has been collected from a plant - usually pine resin. The resin is heated up until the heat literally makes the resin fall apart - cracking it (in the chemical sense) - and the resulting volatiles are collected and called turpentine. From the fact that turpentine is obtained by heating up the source material to a temperature that will decompose it, the condensate is very likely to contain a number of chemicals that whilst volatile enough to turn into vapour at cracking temperatures, still have a high boiling point and therefore not be that suited to adding to paints to thin them as these lower-volatility compounds will be left in the paint.

It is the nature of distillation that if you simply flash-distill (that is, do it quickly) a mixture, you are going to get a product that has a range of molecular weights in it and therefore different volatilities. This 'gum spirit' is sold by Roberson. If you want to make sure that you only have the volatiles in the liquid, you need to re-distill it, taking care this time not to get what is called 'splash-over'. The light ends are your distillate and the less volatile compounds stay in the still's base. The light ends are sold by Winsor and Newton as 'English Distilled Turpentine'.

Lavender oil is steam-distilled from lavender plants and is used extensively in the perfume industry for things like washing powder, air care products and so on. It has the property that insects don't like it (many essential oils are pathological responses from plants to insect damage so they are there to cause damage to insects to reduce damage to the plant and the insects have evolved to not like the smell of lavender amongst other things). It is thought that Jan van Eyck used lavender oil as a thinner because it is a volatile oil so I obtained some for this test from 'naissance' (English Lavender) and 'biOrigins' (Spike Lavender).

By its very nature, the steam distillation process can pull over many compounds that have quite high boiling points so if a mixture is obtained by steam distillation, quite a high proportion of it can have boiling points that are undesirable for the artist who just wants a solvent for thinning paint or cleaning brushes. Fractional distillation can do that but steam distillation cannot guarantee it.

The predominant compounds in the lavender oils for the English lavender are: Linalool (around 30%); and, Linalyl Acetate (around 33%). For Spike Lavender, they are: Linalool (around 50%); Eucalyptol (around 25%); and, Camphor (around 13%). So, you can see that they will probably have different properties when it comes to rates of evaporation.

So, we have two variants each of two hydrophobic solvent products that can be obtained by artists for thinning down paint to use in oil paintings. How well do they perform?

Determination of Relative Evaporation Rates

Six samples:- The four above plus acetone and water. Copyright (c)2020 Paul Alan Grosse Determining absolute rates of evaporation is probably not that meaningful as there are going to be differences between each artist's studio and differences within each artist's studio for different times of day, time of the year and so on. More meaningful is the relative rates of evaporation so I decided on a method that would compare each so whilst the absolute results might vary according to air temperature, draught and so on, you can still tell just how much faster or slower one solvent is when compared to another.

I used aluminium make-up pots without the screw-on lids, 3.6cm in diameter and 2cm tall, numbered them and weighed them. I then accurately weighed in between 1 and 2g of each solvent. I also included two additional pots, one with acetone and one with rain water - the intention being that these would be familiar points of reference. I weighed them every quarter of an hour to start with and then every half hour on a balance that measured to 0.001g. The liquid in each produced a layer of liquid in the order of 1mm deep so to start with, the weight loss was a representation of how much mass can be lost through around 10 square centimetres of surface where migration of the liquid to the surface is unhindered. The experiment was performed in a domestic kitchen without any forced draughts. The room temperature was 22C - 23C throughout.

I then tested the four hydrophobic solvents for residues by weighing in approximately 1.5g of each solvent and letting them evaporate at room temperature, weighing them then putting them in the oven for 30m at 80C.


The turpentines evaporated steadily and relatively quickly, surprisingly at around the same rate as the water but the lavender oils were very slow and eventually had to be blown by a fan for around 6 hours before they had finished evaporating at room temperature.

I had included acetone and water firstly because most people are familiar with them but also, because I thought that they would bracket the other liquids - acetone being fastest and water being the slowest so it came as quite a surprise to find that even though the acetone behaved as suspected and the water performed as well as the turpentine, by far the worst was the lavender oil, both in terms of its rate of evaporation and the residue left behind.

Evaporation rates for Acetone (right scale) along with two types of turpentine and two types of lavender and water as a guide. Copyright (c)2020 Paul Alan Grosse

Solvent    English
    Acetone *    Rain
Water *
Brand    Winsor &
    C. Roberson
& Company
     naissance     biOrigins    Lucemill    Nature   
Initial Evaporation Rate 
/ mg cm-2 hour-1
    16    24    1    5    200    6   
Established Evaporation Rate
/ mg cm-2 hour-1
    8    9    1    2    180     Þ   
at Room Temperature
    <1%    5%    5%    3%    <1% §    <1%   
at 80C for 30m
    <1%    3%    2%    2%    <1%    <1%   
Residue - Physical    V. thin film,
    Thicker film,
    none    none   
Residue - Smell    None    Slight 'pine' smell    Slight, Sweet    Slight, Sweet    Olefinic,
'P'-Side (C3)
Acetone and water added to table to show rates for familiar liquids for comparison.
† Initial rate (approximate) reflects presence of light ends except in acetone.
‡ Acetone's evaporation rate was so high that its temperature was around 4C lower than the other samples.
Þ Water's latent heat of evaporation was so high it was around 1C lower than the other samples.
§ Some drops of water remained after the acetone had evaporated having condensed from moisture in the room - this soon evaporated.


Ideally, you need a solvent that will increase the volume of your paint sufficiently to allow you to apply it in the way you want, and then to evaporate, leaving nothing behind. Additionally, it should do this not so fast that it dries out whilst you are still painting with it but not so slowly that it is still there after your linseed oil has finished polymerising - reducing the volume of the paint layer after polymerisation would lead to a weakened paint layer.

Turpentine evaporates off at about the right rate for this process and the English Distilled Turpentine leaves no appreciable trace behind it. Whilst the Gum Turpentine evaporates at pretty much the same rate, it does leave between 3% and 5% residue in the form of a resin behind which is, in effect a low molecular weight version of the resin it came from - ie, colophony or pine resin. In order to reduce this to acceptable levels, all that needs to be done is to distill it so that these heavier ends don't end up in the final product and this is, what the English Distilled Turpentine is.

On the other hand, the lavender oils did not evaporate very quickly - at less than a fifth of the rate of the turpentine - and as they went on, the evaporation rate deteriorated considerably. Finally, both the lavender oil and the spike lavender oils both had around two per cent oil that was a heavy oil in one case and a grease in the other - neither of these being acceptable in oil paints.

So, if you want to add a solvent to your oil paint to make it paint better, go for the English Distilled Turpentine. If you can't get that then the Gum turpentine will do but whatever you do, don't waste your money or time on the lavender oil, wether it is ordinary or spike, it just takes too long to evaporate.

Hydrophylic Media

Hydrophylic media are basically all of those that will, as the name suggests, use water as the solvent. Water is water but it is not the water that is the problem, it is what is dissolved in it that is the problem.

Tap Water

Water from your tap has plenty of contaminants in it because the source of that water is often a reservoir or a river - note that reservoirs are filled by rivers. All sorts of contaminants get into rivers such as nitrates from fertiliser draining from fields, calcium from dissolving rock and so on. If you live in a hard water area, there will be more calcium in the water than in soft water areas.

The amounts of these contaminants in tap water are quite low - the calcium as calcium carbonate in my tap water is between 50 and 60ppm by weight, in other words, a litre has between 50 and 60mg of chalk dissolved in it.

So, whilst boiling gallons of water over years in a kettle will leave a scale on the inside, adding a few millilitres of tap water to your water colours isn't going to turn your artwork into a sea of scum. However some things in your water can react with pigments with undesirable effect or speed up the degradation of your artwork.

So, what alternatives are there to tap water and what was available centuries ago?

Deionised Water

Deionised water also called 'demineralised water' is tap water that has been through a couple of ion-exchange columns. One removes cations (things like ammonia, metal ions such as copper, calcium, magnesium and so on) and one removes anions (things sulphate, carbonate and so on). The resulting water is free from these inorganic salts but the columns don't remove organics (unless they react with the resin of the columns).

So, deionised water has the salts removed from it but not the organics.

Distilled Water

Distilled water has been boiled and the steam collected and condensed - usually in a piece of glass apparatus. This can have some salts in it if the boiler part of the apparatus has been running too long without the water being exchanged - some are designed to so replace this constantly so buildup and scaling doesn't occur as much - and droplets have been carried over with the vapour.

The head of the still is usually an anti-splash head, designed to reduce the level of misting. The best designs for anti-splash heads are based upon the cyclone so that liquids are separated out centrifugally (the largest one of these that I have seen was around 2m across as part of a petrochemical plant that I used to work in the lab for).

However, volatile organics and volatile salts can still make their way across into the distillate so again it is not perfect.

Tumble-Dryer Condenser Water

Like distilled water, this has evaporated by warm air from the damp clothes and the air/water vapour then cooled down by the dryer's heat pump to below its 'dew point' so that it condenses out. This is then pumped into a container so that you can tip it away when enough of it has collected.

Even though it has not been boiled as such, it has in effect gone through the distillation process and as such is very low in salts and so on.

However, even though it is distilled water, there are usually bits of lint in it and whilst you can filter these out and use the water as ironing water - it is not going to clog up your iron with salts - there are also other organic chemicals in there.

In chemistry, there is a process called steam distillation and it is used to distill from mixtures chemicals that would normally decompose if heated up to their boiling points. The ratio of water to the other chemical is a function of their molecular weights but you can get oils and very small quantities of light waxes to go over in the distillate - this is what is in fabric conditioner. Therefore, whilst these contain chemicals that are alredy part of the washing process and will not interfere with your steam iron, they are not part of the art process and therefore would be foreign to, say, water colours or other water-solvent-based arts

Therefore. appealing as this might sound as a potential source of water for art, it is not advised.

However, volatile organics and volatile salts can still make their way across into the distillate so again it is not perfect.

Rain Water

Rain water has got to be pure, hasn't it? Well, no. It is made by evaporating sea water, the vapour then travels upwards by usually a mile or so and then it condenses out and it falls as droplets through around a mile of air and then hits your collecting device.

The clouds are sufficiently far up for any problems of splashing not to exist although sea water can be whipped up into a spray that is carried upwards so that is no guaranteed. Additionally, there are tornadoes and water spouts (think of raining fish or frogs) which carry water and other contaminants upwards by large distances that can include as far as the clouds. And, then, there are sand storms - remember that Saharan sand and other material is carried across the Atlantic ocean to South America.

Next, it falls through the air and has the opportunity for around 5 minutes to pick up dust, particulates (algae, pollen, fungus spores and so on), along with soluble gasses such as oxides of nitrogen and carbon dioxide along with ammonia if it is in polluted air, before it is collected. If it has been raining for a short while, the contaminants that are not in the whole of the air such as industrial pollutants like ammonia and so on, will have been scrubbed out of the air by the rain that has already fallen - think of the air looking clearer after a storm.

Finally, it hits your collecting device and can pick up pollutants and dust that is on that although, again, if you let it rain for a bit first, a lot of that should be washed out of the way before you start collecting.

Once you have it in storage, if you let it stay in strong sunlight, there is bound to be at least one algae cell in there that will use the available nitrogen compounds and carbon dioxide to grow and you might well end up with a green tint to the water so keep it in the dark.

So, which is better? As an analytical chemist and an artist, I would go for the rain water stored properly and if that runs out, use tap water.

Grounds and Supports

In short, the ground is the surface that you draw or paint upon and the support is what holds it in place.

Stone / Brick Plaster The finished plaster surface is extremely absorbent. Usually, pigment is mixed in with the final plaster layer to produce the painting which is called 'Fresco.' In this case, the pigment environment is alkaline and only a subset of the normal set of pigments can be used. Artists' colourmen can advise on which pigments are and are not suited to this environment - one being L. Cornelissen.
Alternatively, the final plaster layer can be sized in order to seal it and then pretty much any type of paint can be used on the surface.
Porcelain - Usually, it is the surface of the porcelain that is painted on directly with paints that are in effect either powdered coloured glass or powdered minerals in powdered glass. Sometimes a slip made from clay with powdered minerals is used for painting pictured. This is all then fired at kiln temperatures and various chemical reactions take place, giving the final colours.
Alternatively, the surface an be treated as any other porous surface by treating it with size and then whatever colours in whatever medium you feel like if you are not going to fire it again.
Papyrus - Papyrus is already sized to a large extent by the glue-like chemicals in the plant's sap and there is no need to add anymore to it. Traditionally, it is painted on with inks made from soot and paints that are made from powdered rocks although Egyptian blue and vermilion (artificial cinnabar) were used by the Egyptians. On papyrus, the paint tends to stay where you put it.
Vellum / Parchment - Paint and ink is applied directly to the skin. It is not porous and the ink or paint just sits there until it dries. Any grease from your skin will stop the paint or ink from staying there and the surface generaly has to be prepared by going over it with pumice or some other fuller material.
Paper Size Paper for painting is usually made from waste cotton fibre with size in the water already so you don't have to size it yourself. You can paint on more size if you like, depending upon what you are doing.
Smooth - hot ironed - paper is better for drawing with inks as the loose fibres in the rough paper stock will catch on the nib/quill.
Wood Size / Gesso If you do not at least size the surface of wood, it bleeds anything that you put on the surface. With size, your work still sits where you put it but if you put gesso on there as well and then sand and scrape it, you have a nice, flat surface upon which to work. If you need to do punch work, then size the wood, stick on some cloth such as muslin or linen then apply gesso, smoothing with sand paper and then scraping as above. You can also use gesso to produce pastiglia which can be left or fashioned with various implements.
Canvas Size then Gesso Canvas will rot if you let linseed oil come into contact with it (so I am told) so first apply size and then coats of gesso and smooth as you want. Canvas is flexible so it is best not to make it too thick as the gesso will crack. Also, take into consideration the instability of canvas if you are considering using pastiglia in the painting by not making it too thick, otherwise it might simply crack and drop off.
Metal Gesso (optional) Metal is very stable and is used both for paintings that are going to be fired like ceramics and also for painting on with oils. Without a layer of ground to reflect the light, transparent paint virtually glows when compared to paint that has been applied to areas that have paint underneath - the light passes through the paint and instead of most of it being lost in the process of scattering, almost all of it is reflected back towards the viewer producing a substantially different effect. This method was used for a while to paint jewels from medieval times.
Chalk gesso ground on an oak support. Copyright (c)2020 Paul Alan Grosse
Gesso ground on solid oak support.
Silver Leaf under transparent oils making them appear to glow when comparing them to other areas of paint. Copyright (c)2020 Paul Alan Grosse
Transparent paint on silver leaf.

Mixing colours


We see colours in paints because the paint absorbs or otherwise prevents our eyes from seeing certain wavelengths. You can represent colours by using pure light or mixtures of light and a useful tool for predicting what will be seen when mixing paints can be made from this.

Move the mouse over the little CIE diagrams at the start of each section below to see that CIE diagram on the right.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Eyes. The iron-shaped CIE diagram on the right is a representation of hue and saturation that your eyes can see - the area for a given set of circumstances is called a 'gamut'. How dark or bright a colour is is not covered by the diagram because it is really a three-dimensional diagram of which this is a slice through the middle - white and black being at the hue centre, above and below the diagram.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Wavelengths. The wavelengths of the pure colours are around the curved edge because they represent a hue that can be reproduced by using light of a single wavelength.

Notice that there are no wavelength numbers along the straight edge at the bottom.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Absorption Wavelengths. Along the straight edge, between blue at 400nm and red at 700nm is a line of hues that need two colours of light to reproduce them or, from white light, can be represented by an absorption peak with a single wavelength. In the middle of the 'iron', there is a grey area and if you draw a line from the colour you see, through the spot at the centre, to the other side, you can see the wavelength of peak absorption that can be used to produce that colour.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse For example, an aqueous solution of potassium permanganate looks purple and if you look on the straight line where an absorption peak corresponding to 530nm (yellowy green) appears, you will see the purple reproduced. The actual permanaganate absorption is a number of peaks that mass together but their centre is around 530nm.

Yellows. Copyright (c)2020 Paul Alan Grosse Saturation. The further away from the edge that you get, the less saturated the colour. Taking yellow as an example, A laser would produce a single wavelength so it would be located at the edge of the diagram. Two LEDs can also convince you that you are seeing yellow light even though there is no yellow light there at all. If the wavelengths are close enough to each other, it can appear almost as saturated as the laser light. A yellow ink or dye only has to absorb blue light and that can appear quite saturated. The more light that you can see, the brighter it is and that is one of the things that is confusing to most people.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Screen. Your monitor screen does not use pure colours to produce the colours that you can see - these would not be bright enough and your eyes adjust to a lowering of saturation by automatically compensating for it so a compromise is made of a useful brightness against a useful purity.

One feature of the CIE diagram is that if you want to produce an intermediate hue by mixing two colours, the resulting hue and saturation will appear on a straight line between the two colours you started off with. By starting off with the optimised red, green and blue that your monitor can produce, you end up with the small RGB triangle on the right.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Printer. Following on from the RGB triangle, the CYMK triangle is based upon the even less saturated colours that your printer uses and as a result, has an even smaller gamut. The only reason that your printer seems to produce such saturated colours is that you are used to looking at images that are so dull and your eyes compensate for this to some extent.

Interestingly, some people claim that you can duplicate any colour there is just by using a palette of White, Magenta, Yellow, Cyan and Brown (black is not permitted although white is - note that you get a reasonably proficient black by mixing the brown with the cyan) and then they will match any colour from a colour print - of course, the gamut is so small that all they are doing is matching the printed colours and not the colour of the original subject. Single pigment paints give the best colours as you can see from the chart.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Lasers. Nominally, lasers will produce a single wavelength of light - in reality, it is a bit wider than a single wavelength but for most uses, this doesn't matter too much.

As the laser light is pure, it will sit at the edge of the chart. However, our eyes don't differentiate between pure colours and slightly broader ones and we will think that a light source that is centred around a particular wavelength is just as saturated up to a point, even though it is not. For this reason, we can look at the light from a red and a green source, mixed together and think that they are yellow. You can do this yourself with a red and green laser and even though there is no yellow light there, it looks yellow.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Early Colour Pigments. These are some of the colours that were available at the beginning of the 15th century - except for Egyptian Blue, that one is artificial and they had forgotten how to make it by then.

Terre Verte is an earth pigment - like the Siennas, Ochres and Umbers - but the main colouring agent is chromium oxide. For the reason that it is primarily an earth mixture that is dug out of the ground, there is a lot of variation in the colours you get for this pigment - saturation and brightness, but the hue is basically the same.

CIE Chromaticity Diagram. Copyright (c)2020 Paul Alan Grosse Common Mixtures. One thing you might notice from the above selection of colours is that there is little in the way of yellow-greens and purples. The latter was solved by mixing together rose madder and lapis lazuli (or, for a less pure version, rose madder and azurite) for the purple variants. This was easy because they don't react with each other.

The yellow-greens (or 'Vergaut') is a different story. Orpiment is Arsenic Sulphide and Verdigris and Malachite are copper compounds. Copper Sulphide is black so if you are using a medium that doesn't seal in the particles of pigment, hydrogen sulphide can be released from the Orpiment and will attack the Malachite or Verdigris so, if you are using Orpiment, the green that you mix it with needs to be Terre Verte.

However, Lead Tin Yellow II is all right with copper so either Verdigris or malachite is a mixture that will last. In oils, the medium seals the pigments in so it doesn't matter as much but it is still better to play safe - these mixtures are to be found in Jan Van Eyck paintings which date back to the early 15th Century so they do last but the passage of time is inevitable and chemicals do react in any medium so it is better to play safe.

Effect of Particle Size on Mixing

Diagram illustrating how particle size plays a crutial role in determining which colour dominates in a binary mixture. Copyright (c)2020 Paul Alan Grosse On the right are paint cross-section diagrams that represent what happens with paints with different sizes of opaque pigment particles imagine that the painting's viewer sees the painting from the top of each diagram.

We take two hypothetical colours in two particle sizes where each of the particles is spherical.

In diagrams one and two, the particle size is small whereas in three and four it is large.

Larger pigment particles shown by palette knife on grinding slab. Copyright (c)2020 Paul Alan GrosseIn diagrams one and three, the pigment is white whereas in two and four, it is black.

Paints one and three should look the same as each other as they have the same pigment to volume ratio and paints two and four should look the same as each other as they have the same pigment to volume ratio. For a given shape, it doesn't matter how big the particles are because they will always have the same particle to void ratio - in the case of paint, pigment to medium.

If you mix equal volumes of one and two together (5), you will get a mid grey because you can see equal areas of each pigment.

If you mix equal volumes of two and three together (6), paint two will tend to fill in the interstitial gaps in paint three and therefore dominate it. In short, the pigment with the smaller particles will tend to cover the pigment with the larger particles because the smaller particles fill in the gaps between the large particles more effectively.

This effect is more pronounced, the greater the difference between the particle sizes.

With transparent paints, this does not happen because this effect relies upon the light being reflected back to the viewer as soon as it has interacted with the pigment particles at the surface.

This sort of thing can be tolerated by the artist making their own paints (see image on the left of larger pigment particles shown by palette knife on the grinding slab) but if you are going to manufacture paints that are going to sell on, you need to make sure that your particle sizes are pretty much the same from paint to paint - something that would make paints from real pigments such as malachite, azurite, lapis lazuli, smalt, orpiment and realgar quite troublesome because the artist would have to have a better idea what was going on.

All images and original artwork Copyright ©2020 Paul Alan Grosse.