SOX lamps: History of low-pressure sodium lighting

Early history to 1920s - Difficulties to be overcome - 1930s to 1960 - 1960s to present

SOX light in South Wales, U.K., with old concrete lamp-post.Left: Tonna, South Wales, U.K., May 2007. The lamp-post, made of concrete, is probably 50 years old or more. The lantern is newer.

The term "SOX" has only been around since the 1960s. Sodium lighting was introduced in the 1930s, but the history of electric lighting goes back well into the 19th century. As you will see, a great amount of technological research and development has been required to make the sodium lamp…

Early lighting history to 1920s

Electric lamps can be divided into three types: carbon arc, incandescent, and discharge lamps. The first electric lamps of any sort were carbon arc lamps which first appeared in the 1830s. These lamps worked on the principle of discharging air (i.e. neutralising the electrical charges that exist in the air) between two carbon rods using an electric current; this luminous discharge is called an "arc". The first example of streetlighting was using these carbon arc lamps, in Godalming, Surrey, in 1881, provided by Sir William Siemens; the main roads had arc lamps while the side roads were lit by incandescent lamps, which are similar to tungsten-filament light bulbs as described below:

An old-fashioned domestic light bulb (prior to the energy-saving type) was an example of an incandescent lamp. Incandescent lamps contained a filament (which was made of tungsten) which glow white-hot when an electric current passes through it. This glow is called incandescence. This filament lamp was developed in the late 19th century. This type of light was of course most commonly seen in the home: it is not usually seen in streetlighting, although some installations have existed.

1930s streetlightingRight: 1930s streetlighting. This scene was of course still a futuristic concept in the very early days of electric lighting. (Colin Grimes)

Sodium lamps belong to the third category of lamp — the discharge lamp, which are classed as a type of fluorescent light source. We will see in more detail later how these work, but in short, a discharge lamp contains vapour of a certain substance (e.g. sodium) which gives off light when an electric current passes through it. The key difference between incandescent and discharge light sources is that incandescent light sources are much hotter than discharge lamps. For example, the filament in an old-fashioned domestic light bulb operated at around 2500°C, and the light that was produced was purely because the bulb was so hot. On the other hand, a sodium lamp operates at around 260°C. Another difference is that whereas incandescent sources (i.e. the domestic light bulb) give off a continuous spectrum of light (i.e. with all the colours of the rainbow), discharge lamps only give light of certain colours.

The first discharge lamp was made, using mercury, in 1900 by Cooper-Hewitt. The lamp produced a greenish-white light.

However, sodium can be used to make a more efficient discharge lamp, and the patent for such a lamp was taken out in 1919 by Arthur H. Compton (1892-1962) of Westinghouse Electric in the USA. Compton made a great contribution to the study of light and its behaviour, and received the Nobel Prize in 1927. The reason for choosing sodium is that it has a high vapour pressure, which means that it is relatively easy to vaporise as required for use in a discharge lamp.

Because the low-pressure sodium lamp is very efficient, it was very suitable for lighting the streets in large numbers; hence its yellow/orange light became a familiar sight. In scientific terms, the term ‘efficacy’ is the ratio of the total light output (lumens) to the input power (watts), and the efficacy of a low-pressure sodium lamp is around 180 lumens per watt (depending on wattage), compared with around 13 lm/W for a domestic light bulb with a tungsten filament. Efficacy is a useful method of comparing one light source with another, being a measure of the lamp’s effectiveness; it is slightly different to efficiency in that efficiency is the ratio of output watts (of useful light) to input watts (in other words, how much use the lamp makes of the power it is given — the higher the efficiency, the less power that is wasted.). The colour of light given off by a low-pressure sodium lamp (yellow or orange — a wavelength of 589 nanometres, or 589 millionths of a millimetre) is also close to the peak sensitivity of the human eye (555 nanometres), which is the reason for the lamp’s effectiveness. Many animals, though, do not see sodium light as brightly as humans.

The term “low-pressure” is used because the sodium vapour is used at normal atmospheric pressure (or generally below .05 bar above atmospheric pressure); high-pressure sodium lamps, or SON lamps, also exist. They give off a light that is closer to white, with either a salmon-pink or a lemon-yellow tinge. (Ref: James Hooker, Simon Cornwell)

1930s streetlightingLeft: 1930s streetlighting with the lanterns centred over the road (Colin Grimes)

Difficulties to be overcome

Before the sodium lamp could be introduced commercially, a few difficulties had to be overcome. Because sodium is a material that causes chemical reactions very easily, for example against most types of glass, it was necessary to develop special sodium-resistant glass. In 1920, such a glass was developed by the aforementioned Arthur H. Compton. Compton's glass was 60% borate (i.e. an oxide of the chemical element boron) with oxides of aluminium, sodium and calcium.  

The second difficulty arises in the fact that a sodium lamp needs a ‘kick’ in order to get it going, i.e. a surge of initial voltage. In order to be able to start the lamp with a voltage that was not too excessive, it was necessary to add a mixture of gases to the lamp (of which more later): the design of the sodium lamp was aided by the invention of the neon tube in 1922. Neon is now used to enable the sodium lamp to warm up from cold. Neon is well-known for its red glow, commonly used in advertising signs, and its presence in the sodium lamp is the reason why the lamp glows red when it is switched on. In 1927 F. M. Penning developed certain mixtures of gases, including neon, which are now part of the design of the sodium lamp, and which we will explore later.

A third problem was in maintaining the correct operating temperature of the lamp at 260°C, which required good heat insulation. Developmental work to cover this problem was published by M. Pirani in 1930. Pirani recognised that in order to achieve good heat insulation, an outer glass ‘jacket’ would be useful. As well as this, Pirani did some experiments with the sodium lamp inside an oven (at 350°C) in order to mimic ideal heat insulation, and to understand the importance of good heat insulation. The efficacy of his lamp was far higher than what was achievable at that time without using an oven. Pirani’s work helped to bring about the practical use of sodium lamps (although, thankfully, we do not need to use sodium lamps inside ovens!).

As we know, these problems were overcome, but even so the efficacy of a sodium lamp was only 50 lumens per watt in the 1930s, as opposed to around 200 lumens per watt as is possible today (depending on the wattage of the lamp). Also, the total light output was only a fraction of modern values — 7000 lumens in the 1930s, which had become 33,000 lumens by the early 1980s. Even so, the sodium lamp was still the most efficient light source in its early days — a paper detailing just how much light output could be produced was published by Compton and van Voorhis in 1923. (ref: J. W. Denneman)

1930s to 1960: sodium makes an appearance

The first UK low-pressure sodium installation at Croydon, 1932.Right: The first installation of low-pressure sodium lamps in the UK in Purley Way, Croydon, 1932. (Bob Cookson)

And so, in the early 1930s the low-pressure sodium lamp duly appeared as a street light. The streets outside the Osram company in Berlin were lit with these lamps in August 1931, and low-pressure sodium lighting was also installed by Philips in Holland in June 1932, on the road between Beek and Geleen in the Netherlands. These first lamps were contained in a detachable 'Dewar jacket' (named after Sir James Dewar) in order to maintain heat insulation.

The first public installation of low-pressure sodium light in the UK was on Purley Way in Croydon in December 1932. This location was chosen simply because it was just outside the Philips factory where lamps were manufactured.

After three years with successful results, the initial 100W lamps (and their lamp-posts) were replaced with 150W lamps made by a manufacturer called Wardle, and some parts of the installation existed until the 1970s. The Wardle lamps were suspended over the centre of the road on wires. When Purley Way had its streetlighting upgraded in 1979, Philips again used SOX lamps — their more efficient 135-watt MA50s.

Right: A surviving example of the first Wardle sodium streetlights to be installed in the 1930s in Liverpool. This lantern is now fitted with a modern 90W SOX tube. (Colin Grimes)

Liverpool followed Purley Way in 1935 with the first installation being on Queens Drive, Mossley Hill, using 100W (later 85W) and 150W lamps (later 140W — the wattages of sodium lamps were re-rated in 1938). They survived until 2000 using 90W SOX lamps, when they were replaced with 10-metre steel columns with a mixture of SOX and SON (high-pressure sodium) lamps. (Ref: J. W. Denneman, Colin Grimes)

The wattages of sodium lamps have been reclassified more than once in their history. The first reclassification came in 1938, as mentioned above:

Before 1938 After 1938
50W 45W
70W 60W
100W 85W
150W 140W
  Left: Reclassification of SO sodium lamps in 1938. SO lamps are the predecessors of SOX, SOI and SLI lamps. (ref: Simon Cornwell)

This reclassification was done by a group called ELMA (Electronic Lights Manufactures' Association of Great Britain). This cartel was formed in 1919, formed from the ‘lamp rings’ — groups of manufacturers who fixed wattages, lamp life and prices (being basically GEC, BTH and Siemens).

After the government stepped in, ELMA eventually became a trade union, although even today there are still agreements made between the lamp manufacturers. (ref: Simon Cornwell)

Developments in sodium lamp design in the 1950s.Right: Changes in the design of low-pressure sodium lamp tubes, 1932-1960 (J. W. Denneman)

However, these sodium lamps were not used everywhere. From the 1930s to the 1960s they were really seen as major highway and motorway lighting — not suitable for residential schemes or high streets, where incandescent, mercury and fluorescent lighting ruled.

The design of the lamp (the tube that actually produces the light) did not change significantly between the early 1930s and 1955, although much research was carried out after the Second World War in its technical development (of course, there was little streetlighting during the blackout period of the war). Four wattages were available: 45, 60, 85 and 140-watt lamps. The outer glass jacket did not become an integral part of the design of the lamp until the late 1950s; until this time, as mentioned, it was detachable. However, further technical difficulties involving the glass were marring the sodium lamp achieving optimum performance.

The glass of the inner tube was not perfectly designed: during the life of the lamp it would turn brown due to attack by the sodium, and this caused the light output to fall over time (see the table below). When an improvement in the glass was introduced in 1955, the light output later on in the life of the lamp was improved. However, this new glass had less friction towards the sodium inside it when it was molten during operation; the liquid sodium would roll along the tube and cause an uneven distribution when it solidified after the lamp was switched off. This uneven distribution of sodium would cause an uneven production of light, with some parts of the tube not producing enough. To solve this problem, grooves were added to the tube to prevent movement of the sodium (see "b" in the diagram above). But there was another problem: the glass also absorbed argon more easily, which is used in the lamp to allow for a lower ignition voltage to be used. This meant that a different mixture of gases had to be used that did not involve argon, and this caused the efficacy to fall slightly as shown the table below(the gas mixture that is also present with the sodium in the lamp is discussed in the next chapter, along with the reasons why argon is useful and why it has since been re-introduced as part of the gas mixture).

But even this solution was not good enough. The grooves that now existed in the lamp were at a hotter temperature during lamp operation than the rest of the tube where the glass surface was flat. This meant that during lamp operation, sodium vapour would condense back into liquid on the cooler flat areas of glass and cause a series of "sodium mirrors" along the length of the tube, which reduced the light that then could be emitted from the lamp.

Hence the lamp was re-designed (again) in 1958. By placing small dimples along the length of the tube to collect the sodium condensation, there was no need for grooves, and the sodium mirrors no longer formed. These dimples are shown in "c" in the diagram above as little ‘bumps’ along the top and bottom halves of the tube, although it is a feature of Philips lamps more than any other manufacturer. The sodium will condense in the dimples first, as these are the coolest areas of the tube. The dimples also ensure that the distribution of sodium is more uniform along the length of the tube. However, because these dimples are vulnerable, it was now necessary to mount the whole lamp tube in a special sealed-on outer jacket instead of the previous detachable version. It was also now possible to use argon once again in the lamp, provided that the glass was specially treated with argon. This change back to argon caused the efficacy to rise slightly, and the light output level was now maintained for longer during the whole life-span for this 1958 lamp (see the table below) because of the absence of sodium mirrors being formed. By 1958, the average life of a low-pressure sodium lamp was 4000 hours, compared with 2500 hours in 1938.

Change of efficacy in lumens per watt of early sodium lamps as they were developed (J. W. Denneman):

  Operating time
Year 0 hours 100 hours 4000 hours
1932-1954 85 72 44
1955 84 79 52
1958 87 82 68
1960 104 100 88

Another improvement in the glass was made in 1960, which resulted in less argon being absorbed and the resulting improvement in efficiency raised the efficacy once again as shown above. (ref: J. W. Denneman)

1960s to present

The type of low-pressure sodium lamp referred to as “SOX” did not appear until the early 1960s. The term “SOX” indicates a low-pressure sodium lamp with the lamp tube bent into a U-shape, and with a metal-oxide glass coating (which we will come to in a moment). Before this time SO/H and SOI lamps were used. An explanation of where the terms SOX, SO/H etc. originate from is given on the page “How they work” — “The parts of a lamp”.

Some major technological advances regarding low-pressure sodium lamps were made in the 1960s. As we will see in later chapters, it is very important to be able to keep as much heat inside the lamp as possible, as they are best run at a temperature of 260°C inside the U-tube. In 1964, with this in mind, thin heat-reflecting films of exotic materials were introduced. A microscopically thin coating of these materials on the inside of the outer jacket successfully dispensed with the cumbersome heat-reflecting glass sleeves of the earlier SOI design. The composition and thickness of the film is designed to reflect the infrared radiation given off (i.e. the heat) back into the lamp while still allowing visible light to pass through. To begin with, a film of pure gold of just 0.00005mm thick was employed — this is still the best heat reflector, but it also absorbs a lot of light. Bismuth was tried in some German lamps, but the biggest improvements came with the use of metal-oxide semiconductor films and the name ‘SOX’ was introduced with these lamps. These materials can be altered so that they reflect a high proportion of infra-red radiation, while still being transparent to visible light.

Between 1964 and 1966 a range of SOX lamps having stannic oxide (i.e. tin oxide) coatings was marketed. As a result of these heat-reflecting coatings on the glass, there was a great increase in the effectiveness of the lamps, meaning that the lamp wattages had to be reduced; this happened around 1968/69. The reduced wattage made the lamps very marketable as the cost of electricity was rising. The tin-oxide material was superseded in 1966 with a 0.00032-mm coating of tin-doped indium oxide film. The corresponding increase in efficacy saw the wattages of SOX lamps reduced once again with the lamps now called “SOX Plus”. SOX Plus have a longer life, running for 16,000 hours compared with 12,000 hours for the standard lamp. Most often, however, these lamps are still referred to as “SOX” rather than “SOX Plus”.

SO/H, SOI/H, SOX and SOX-E lampsRight: SO/H, SOI/H, SOX and SOX-E lamps (Colin Grimes)

A further improvement was made in 1983 with SOX-E lamps. These have been designed for lower heat losses than for ordinary SOX lamps, and as a result the required wattage of the lamp has been reduced, e.g. from 35W to 26W, or from 180W to 131W. (ref: James Hooker; Philips information leaflet)

As a result of all these developments, the wattage of lamp which a lantern requires has fallen drastically over the years. A 1930s 150W SO/H lamp became a 140W SO/H lamp from 1938. This became the 140W SOI/H lamp in the 1950s, and then the 100W SOX in the 1960s, which was subsequently re-rated at 90W, and nowadays the equivalent 66W SOX-E lamp is available. So the wattage of this size of lamp has fallen to less than half from 150W to 66W.(ref: Colin Grimes)

Above is a photo of four types of low-pressure sodium lamp. On the left (the one with the white cap) is a 60-watt SO/H which had a removable arc tube. This was more or less the first type of sodium lamp used in streetlighting in the 1930s-50s. The second lamp (the one with the grey and gold cap) is a 60-watt SOI/H lamp; this was the second design from the 1950s/60s and was superseded by the modern 35-watt SOX. The next lamp along is an early design of a 35W SOX from the 1970s. However, in the 1960s the lamps were actually rated at 40-watts; they were only reduced by 5 watts after the infra-red coating was improved. The fourth tube shows a 26-watt SOX-E lamp. This is a more modern design with improved technology and infra-red coating which reduced the required lamp wattage down to 26 watts — another example of how the modern rating is less than half the rating of a SO/H and SOI/H lamp. When holding a lamp tube up to the light, it is possible to tell the type of coating from the colour of the surface reflections in the film. Stannic oxide lamps produce a yellow/orange colour, whereas indium oxide takes on a greenish hue. SOX-E coatings are less straightforward as they can either impart a reddish or bluish colour to reflected light: in fact, it is possible for a SOX-E coating to look red to the eye, but blue when photographed.(ref: Colin Grimes, James Hooker)

SOX tube just switched on (red)SOX tube fully lit (yellow)
Above: A 35W SOX-plus lamp (left) just switched on and (right) fully lit. (Colin Grimes)

Taking up the thread once again from the 1960s, in the later part of this decade mercury and sodium lamps were generally being used on main roads, with incandescent, mercury and fluorescent lamps on the smaller streets (although there was a lot of variation from one area to the next).

The lighting situation changed with the 1970s energy crisis, where the efficacy of the low-pressure sodium lamp again worked to its advantage. Because of the shortage of energy sources, sodium lamps took over from other lamps in great numbers, with this trend accelerating in the 1980s and into the 1990s. (ref: Simon Cornwell)

In more recent years, the major supplier of SOX tubes was Philips Lighting at Hamilton in Scotland, with complete SOX streetlights also being made on the same site. Until June 2000, Osram also manufactured this light source near Manchester at its Shaw factory, an old multi-storey cotton mill building with a different type of lamp manufactured on each floor. There was also a GE base at Leicester. Sadly, as of 2021 the worldwide production of SOX lamp tubes has ceased altogether, the Philips Hamilton factory having closed in November 2019. (ref: James Hooker, Allan Court, Mike Barford)

© Matthew Eagles 2005-2018. Last updated 29th May 2021.