Arsenate Phosphors

Around the same time that Luke Thorington had perfected the magnesium fluoro-germanate phosphor with Westinghouse in the USA, a remarkably similar material was brought to the attention of the Philips Laboratories in Eindhoven by a Mr. Travnicek of Graz, Austria.  The material was magnesium arsenate activated by tetravalent manganese, 6MgO.As2O5:Mn⁴⁺, and during the previous few years Philips had built up substantial expertise in the preparation of Mn⁴⁺phosphors.  Previously the company had pinned its hopes on a phosphor based on Magnesium titanate with the same activator, that material having a very desirable red emission but unfortunately its quantum efficiency could not be raised to levels making it suitable for use as a mercury lamp phosphor.

The Eindhoven phosphor team was quickly able to adapt its titanate processes to investigate the arsenate instead, and it was this material which ultimately led to Philips' decision to bring the HPL fluorescent mercury lamp into mass production. (Incidentally in Dutch light source nomenclature, the MB lamps are known as the HP type, while their MBF fluorescent counterparts are classified as the HPL type.  H = Hg or Mercury, P = high Pressure, L = Luminescent).

The properties of this magnesium arsenate phosphor turned out to be highly favourable for its application in the company's HPL / MBF products.  Its fluorescence is bright red peaking sharply at 656nm (Figure XX), it absorbs the whole UV emission from the mercury arc tube, its quantum efficiency is 75% for both the long and shortwave UV stimulation, and its efficiency is stable right up to 300C (Figure XX).  Its first apparent disadvantage is that the very broad absorption of UV radiation extends slightly into the visible spectrum, and consequently some blue light is absorbed and the phopshor has a faint yellowish colouration.  Nevertheless this is considerably less than is the case for earlier phosphors and the colour output of arsenate lamps is in no way rendered greenish.

Figure XX - Arsenate Emission                   Figure XX - Temperature Stability

Previously Philips had made HPL lamps in only very small quantities indeed, based on the early developments in sulphide phosphors.  But from 1950 onwards the company marketed its mercury fluorescent lamps with coatings of the arsenate phosphor which proved to be a very valuable asset.  Other lampmakers around the world had been forced to adopt Westinghouse's magnesium fluoro-germanate material and were tied to that source, while Philips had the freedom to make lamps based on its own developments.  The red ratio of Philips' new lamps was 7.5-9%, somewhat higher than the 7% that had been attainable with magnesium fluoro-germanate, and with germanium being a very expensive raw material the price of the Philips phosphor was also some four times cheaper.  The thickness of the phosphor coating was chosen so as to deliver the same luminous flux from coated lamps as the clear versions.  This could have been made thicker and allowed the phosphor to absorb all of the UV generation of the arc tube and emit more red light, but at these kinds of thicknesses blue absorption became a problem, thus shifting the colour point towards the green.

Chemical behaviour of Magnesium Arsenate

In the beginning however, it was unfortunately learned that magnesium arsenate suffered a sharp decline in its efficiency during the lifetime of the HPL lamps to which it was applied, and the red ratio fell very rapidly.  While other phosphors had to be abandoned because of their lack of durability, some ingenious thinking at Philips enabled the engineers to overcome this problem through a very simple approach.

It was determined that part of the minutely small concentration of Mn⁴⁺activator had been reduced to Mn²⁺, this occurring because of the slightly reducing atmosphere present inside a nitrogen-filled mercury lamp outer bulb.  Experiments were therefore conducted with new lamps containing trace amounts of oxygen in the nitrogen filling of the outer bulb, the new oxidising atmosphere theoretically being sufficient to prevent the reduction of Mn⁴⁺.  Indeed it was subsequently proven that with 0.3% to 0.5% of oxygen in the nitrogen filling, there was no decline in the red radiation of the phosphor throughout lamp life.

However, there is a reason why the outer bulbs of mercury lamps were nitrogen filled in the first place - it is because the smallest traces of oxygen are sufficient to cause oxidation of the molybdenum foil seals which carry the current through the quartz of the arc tube.  In the oxygen doped lamps the seals would oxidise, expand, and split open the arc tubes within a matter of only some tens of hours of operation.

This problem can be overcome by reducing the temperature at the outward extremities of the molybdenum foil pinch-seals simply by increasing the length of the seal.  In fact good lamps were made this way in which the seal temperature was reduced from 500°C into a safe zone below 250°C, but such a long seal could not easily be made on automatic production equipment.  The long-seal lamps were therefore not a viable solution to the problem.

The reason for having used 0.3-0.5% of oxygen in the lamps was because during life, the various metal parts in the lamp also oxidise and bind this oxygen through getter action, and therefore enough oxygen had to be provided at the outset to last for the entire lamp life.  It is unfortunate that the amount required was so high that it caused the seals to oxidise.  It was as a result of this recognition that Philips devised an ingenious solution - to fill the bulb with a gas which very slowly broke down under the action of the internal UV radiation, continuously providing just enough oxygen to protect the phosphor but not so much that the seals would oxidise.  The gas chosen was carbon dioxide.

It was found that 0.5 torr filling of CO2 provided an adequate source of oxygen to protect the phosphor.  Such a pressure is rather low though and causes the arc tubes to run somewhat hotter than usual - previously a nitrogen fill pressure of 500 torr had been used, this offering sufficiently high thermal conductivity and convection losses to maintain the arc tube at its correct temperature.  Very fortunately it was found that increasing the CO2 pressure to similar levels had no adverse effect, and the partial pressure levels of oxygen do not rise with increasing CO2 pressure.  The reason is because the volume CO + O2 evolved as a result of the dissociation of CO2 is roughly 50% greater than the volume of CO2 itself, hence increasing the CO2 pressure in the bulb actually results in a lower rate of dissociation and reduced oxygen partial pressure.

In later work it was discovered that the reduction of the Mn4+ activator could be inhibited by replacing part of the MgO raw material with Li2O.  This increased the UV and high temperature stability of the material and allowed the critically balanced carbon dioxide filling system to be abandoned in favour of the original and cheaper nitrogen.  An interesting spinoff was that the incorporation of lithium also slightly changed the heights of the various emission peaks of the material.

Manufacture of Magnesium Arsenate Phosphor

It is found that the fluorescent intensity is very strongly depending on MgO/As2O5 ratio, so much so that the composition corresponding to the orthophosphate (3MgO.As2O5:Mn⁴⁺) is practically non-fluorescing.  It is a strong UV-absorber but simply does not fluoresce.  Increasing the MgO concentration results in an increase in fluorescent intensity up to the point where a maximum is obtained at 6MgO.As2O5:Mn⁴⁺, the composition which was employed in all lamps.

The tolerances to which this phosphor must be made are quite critical.  Any under-dosing of MgO results in the formation of the non-fluorescent UV absorbing orthophosphate phase mentioned above, markedly decreasing the fluorescence.  Adding an excess of MgO simply results in an excess of that compound in the finished product.  It is not a UV absorber and therefore the rate of fall off in fluorescence is not so severe as for under-dosing, but it does still diminish due to extra scattering.

During production, part of the Mg is added in the form of MgF2 which accelerates the rate of the reaction, just as is the case for magnesium fluoro-germanate.  With that phosphor it has the effect of doubling the fluorescent intensity, but with magnesium arsenate it only has the effect of increasing the durability of the material to ensure that the phosphor does not lose its fluorescent intensity during lamp life.