Some aspects of Opal Synthesis

Abstract
Based on thirty years of active interest and experience in the heat treatment of corundum, the author provides an update on the latest Thai heat treatment process that results in spectacular changes in colour in ruby and sapphire - particularly with respect to the commercial production of various shades of orange sapphire.

AUSTRALIAN EXPERIENCE
It is strange how some things just keep turning up during one's lifetime. For me it's been orange sapphires; not just any shade of orange but an especially bright and pure orange. My first experience with these peculiar orange sapphires came in the mid 70's while heat-treating rough sapphire originating from Swanbrook Creek, Inverell, on the New England sapphire fields of New South Wales.

We were attempting to improve the colour of yellow, green and parti-coloured rough sapphire by heating them in air in an electric furnace to about 1550'C. On odd occasions, when opening the crucibles, our attention was immediately caught by one or two bright orange pieces of rough. After cutting, these occasional pieces produced stones of an exceptional orange colour (See figure 1).

Over the next twenty years, only very rarely did I encounter stones of similar colour, and then mostly these originated from Australian fields. I never once saw parcels or even small lots that displayed the same pure orange colour. That is until I was in Chantaburi last October, 2001.

THAI EXPERIENCE
You can imagine my surprise when one of my regular suppliers showed me parcels of orange sapphire of a very similar shade to those I had observed so many years ago in the heat-treated sapphires from Inverell (See figure 2). I was told the material was Tanzanian, from Songea, that had been subjected to a 'new' form of heat treatment. Further explanation revealed that Songea ruby, when subjected to this new process, lost it's brownish shade and turned a more pure red reminiscent of the colour of Burmese red spinel (See figure 3). More importantly, less popular Songea material in the brownish yellows, reddish browns, palish yellows etc. also changed colour to various shades of bright orange grading through to reddish orange (See figure 4).

I was told the process was a recently discovered secret, and that my supplier had spent over US$1 million in buying back Songea sapphire he had previously sold into the market. This was later confirmed by a third party I spoke to in Chantaburi who had sold back Songea stone for double what he had paid one month previously. Here was a news event that was obviously going to 'rock' the gem industry.

INVESTIGATIONS BEGIN
Upon my return to Australia I sent my old friend R.W. (Dick) Hughes an email about the 'new' treatment, attaching a photo or two. Dick had also heard of the new treatment (R.W. Hughes, 30th October 2001, personal communication) and was in the process of seeking further information.

On the 28th December, 2001, gemmologist Ken Scarratt mentioned to Dick Hughes that, under immersion, he had observed orange coloured rims around pink cores in the newly treated orange-pink sapphires¹. This news, and further detail on this 'new' treatment, were announced² on The American Gem Trade Association's website on January 8th, 2002.

By January 2002 the whole gemmological world was abuzz with news about the 'new' treatment process - especially when this was applied to Malagasy pink sapphires^3,4,5 _ making these treated sapphires look so much like naturally coloured 'padparadscha' sapphires that some internationally recognised gemmological laboratories had already issued certificates stating just that (R.W Hughes, 2002, personal communication).

Interest in this 'new' treatment process continued, but concentrated on the new 'padparadscha' coloured sapphires. After all, what had been a very rare colour suddenly was available in large, and often calibrated, parcels.

By the end of January, just in time for it to be the 'talk' of the Tucson Gem Show, the GIA reported the presence of beryllium in very small quantities in the orange rim of these sapphires, increasing in concentration towards the surface of the stone. It appeared that this 'new' treatment process was, in fact, a new form of treatment by diffusion⁶.

February 25th 2002 saw the issue of a 'Unified ACTA, CIA, Gilbelin, SSEF disclosure policy' for this 'new' corundum treatment coming out of Thailand⁷. The essence of this policy was:

The name 'padparadscha' would not be applied to this material, and that the sapphires were to be described by Species - natural corundum, Variety Treated (orange) sapphire, Treatments - Indication of heating, and Comment - The orange colouration of this stone is confined to a surface-related layer.

As this policy made no mention of diffusion, the question - What if the orange colour went right to the core of the stone? - needed to be answered.

Many senior gemmologists and corundum experts believed there were still many problems to be resolved, and questions to be answered, concerning this new process⁸.

What surprised me most was that all the interest concentrated on the pink Malagasy material that had orange rims (See figure 5). Nobody seemed at all interested in what I believed was a much more significant aspect of the 'new' treatment. That was the conversion of the colour of Songea ruby from a 'garnet' red to a rather nice and much more acceptable 'spinel' red colour (See figure 3).

I re-examined the stones I had purchased in October, and could find no evidence of orange rims. The only unique feature I observed was the presence of blue 'haloes' of colour around undefined inclusions (See figure 6). The presence of these was later mentioned in the Spring 2002 issue of Gems & Gemmology.

Early March once again saw me in Chantaburi, and once again all the talk was about this 'new' treatment. Topics of discussion ranged from 'How dare those foreigners tell us Thais what we can and cannot do,' to, 'Oh, all the Thai merchants are trying to cheat us using secret treatment methods.' There was a lot of name calling on both sides, but the real truth of the matter was that nobody, other than those actually involved in applying the 'new' treatment process, knew what it involved. And, if they did, they were not about to reveal their secrets¹⁰.

My supplier would not buy back the orange goods I had purchased in October, using as an excuse the claim: If I do that for everybody, I will go bankrupt. Things were very quiet on the orange sapphire front, with everybody adopting a wait-and-see approach. But, I had another surprise in store.

Before I left Chantaburi, another supplier remarked: Terry, do you know the yellow sapphires you are buying are also treated by the same method as the orange sapphire you bought last October? Then, he went on to say: Colourless corundum from Sri Lanka has been turned to strong yellow and gold by this method for over 18 months now.

I was shocked, for I hadn't heard of this treatment before. Was this statement true, or just another rumour? The answer came in Bangkok a week later.

For over thirty years a good Thai friend has always given me the truth when it comes to heat treatments. In October 2001, when I had shown him the orange stones I had bought in Chantaburi, he hadn't even heard of the 'new' treatment process. However, during the ensuing six months he'd done a lot of investigating, and was very quick to show me samples of the range of applications the 'new' heat treatment process had been put to. These conversions included:

Malagasy pinks to 'padparadscha' colour
For the first time I was shown treated rough material, and the resulting preforms prior to their faceting (See figure 7). A slice was taken from one piece, and its pink core with an orange rim was clearly evident (See figure 5).

Improved colour in Thai ruby
For a long time, 'old' off-coloured ruby from the Thai-Kampuchea fields has been sought out to be treated and/or retreated¹¹. A limit had been reached on how much this brownish or purplish volcanic type ruby could be improved. That limit disappeared once this 'new' treatment appeared. In figure 8 can be seen what were originally 'off colour' old Thai rubies. These now do not appear to display a rim of colour that is different from that in their cores.

Sri Lankan colourless, pale blue, and yellow
sapphire converted to yellow and gold sapphire
These proved to be samples of the treated yellows my friends in Chantaburi told me about (See figure 9).

When cut in half, it was fairly easy to see their colourless cores¹², but in faceted stones the colourless cores varied from very difficult to impossible to see (See figure 15).


Fig. 1. Heat treated orange sapphire, Inverell NSW. (Octagonal-cut weighs 1.32 ct)		Fig. 5. Section of Malagasy pink sapphire subjected to the 'new' treatment. (8 mm)

Fig. 2. Songea sapphire after being subjected to the 'new' treatment prior to repolishing. Note etching of polished surfaces.		Fig. 6. A Songea ruby subjected to the 'new' treatment displaying a blue 'halo' around an unidentified inclusion. (x60)

Fig. 3. Songea ruby after being subjected to the 'new' treatment (2.06 ct)		Fig. 7. Malagasy pink rough sapphire after being subjected to the 'new' treatment, with resultant coarse and fine preforms.

Fig. 4. Songea sapphire after being subjected to the 'new' treatment. (4.25mm diameter)		Fig. 8. Thai ruby after being subjected to the 'new' treatment.

Green and green-blue Kanachanaburi sapphire to yellow and gold sapphire
Perhaps the most spectacular results I was shown that day were green sapphires from Kanchanaburi that had yellow rims, leaving no doubt that they had been subjected to some gemmological 'tomfoolery' (See figures 10 & 11). I found this particular application of the 'new' treatment especially fascinating, as the demarcation between the 'new' and 'old' colour was very distinct. Problems with this application of the new' treatment arose when it was explained to me that the effects of the 'new' process can be made to penetrate right to the core of the stone, effectively converting a green sapphire into a yellow sapphire (See figure 12).

Some time later, when I had a chance to examine a section of this material under magnification (See figure 13), some interesting observations could be made. For example, it could be clearly seen that although the boundary between the yellow and the green colours was in general terms parallel to the surface of the stone, it wasn't an exact match. The interface was 'stepped' when passing at an angle to the colour banding, and was 'flat' when parallel with the colour banding.

After examining a photograph of this effect, Dr John Emmett of the USA based consultancy Crystal Chemistry explained (J.L. Emmett, June 6th 2002, personal communication) this phenomena in the following words:

It appears to me that, in the language of the band theory of solids, that beryllium is a very shallow electron acceptor. For a stone to appear blue-green, both titanium and iron must be present. The titanium is a donor and donates an electron to trivalent iron making it divalent (thus the Fe²+ - Ti⁴pair is formed). Iron is an acceptor also, but not so shallow as beryllium. Thus as beryllium diffuses into a stone containing both Fe and Ti it will oxidize the iron by grabbing that electron. Thus the Fe²+ - Ti⁴ pair is neutralized, and the blue-green color is replaced by yellow which is probably a combination of Fe³+ and the trapped hole color center.

Since beryllium is taking part in a solid state chemical reaction, it no longer diffuses freely. This is what is called trap diffusion, or diffusion in the presence of a chemical reaction. The rate, or penetration, will be modified by the concentration of the natural impurity it reacts with in the stone. Thus it will not remain perfectly conformal with the surface.

Everything above is my current theory of what is happening. I have not proved it all, but that is what I am currently trying to do.

Despite all these wonderful samples, my Thai friend still had no idea how this 'new' process was actually carried out. He did, however, state the process was 'time dependant', and could be reversed.

Upon my return to Australia, and on the advice of Dick Hughes, I sent samples of these treated sapphires to Dr. John Emmett of Crystal Chemistry in the USA for analytical purposes. Shortly after, my attention was caught by a photograph on the GRS Cernresearch Swisslab LTD's website of an unknown material appearing as a white opaque mineral after the heat treatment of corundum¹³. This appearance reminded me of my early heat treatment of the Inverell yellows and greens, because not only did we get the orange stones but also odd opaque white particles of a then unidentified mineral. All we knew was the mineral prior to the treatment process must have looked very much like green or yellow corundum. After considerable effort we finally identified this white mineral as chrysoberyl. So could this be the source of the beryllium the CIA had discovered? Could I have been inadvertently carrying out this 'new' treatment process some thirty years ago?

On April 10th I emailed off this idea to Dick Hughes and John Emmett. While both agreed this might be a source, it was John's reply (J.L. Emmett, June 6th 2002, personal communication) that proved most interesting. He suggested that chrysoberyl could well be the source of beryllium for the new process, perhaps first by accidental mixing with sapphire when one considers the amount of chrysoberyl coming out of Madagascar and Tanzania, and second by deliberately grinding up of chrysoberyl to make a flux containing beryllium.

Outstanding inventors are not usually university trained; they just happen to be people who try a lot of things and are very careful observers.

Early in April, John Emmett was still not sure if beryllium was the only factor involved in this 'new' process, so we decided to wait until he had finished his study before announcing our thoughts about chrysoberyl. But we were pre-empted by an article published in late April on the GIA website, and subsequently in Spring 2002 edition Gems & Gemology, that reported¹⁴ chrysoberyl as a possible source for the beryllium used in the 'new' treatment process.


Fig. 9. Amixture of colourless, pale blue and pale yellow Sri Lankan sapphire subjected to the 'new' treatment.

Fig. 10. A piece of rough Kanchanaburi green-blue material after being subjected to the 'new' treatment.		Fig. 13. Polished section of Kanchanaburi sapphire subjected to the 'new' treatment (x12)

Fig. 11. A large cut Kanchanaburi green sapphire after being subjected to the new treatment		Fig. 14. Note the variation in the thickness of the pink 'rim' due to the amount removed during performing after the 'new' treatment.

Fig. 12. Kanchanaburi green and blue-green sapphires that have been subjected to the 'new' treatment. Note the depth of the effect varies from stone to stone.		Fig. 15. It is hard to imagine how the presence of the colourless core could be detected in these Sri Lankan sapphires that have been subjected to the 'new' treatment. (1.75mm diameter)

May 4th saw what was perhaps the first definitive publication relating to the methodology involved in the 'new' treatment process. John Emmett and Troy Douthit made a presentation to a meeting of GIA, AGTA-CTC, GIT and SSEF at Carlsbad, California. They described¹⁵ how they had duplicated the results of the 'new' treatment by a series of experiments using beryllium-doped dry powders and flux under high temperature conditions. To quote from their submission:

The experiments reproduced the complete range of colours and diffusion phenomenology that are observed in gemstones that are in the marketplace, plus a few new colours.

June Ist saw the GIA Gem Trade Laboratory and the AGTA Gem Testing Center make a joint announcement that they were "changing their reporting policies for the new corundum treatments^"16, and reported that:

The new wording reflects two significant determinations with regard to these treatments that:

Further these gemtesting laboratories reviewed their use of the term 'diffusion', as previously used in a gemmological sense, with the words:

As a result of these discussions and a review of the technical literature, it is clear that 'bulk diffusion' is scientifically correct and should replace the terms 'surface diffusion' and 'diffusion treated' as they were previously used gemmologically.

THE FUTURE
So just what's the 'score' at the end of June 2002? How does the average gemmologist cope with stones subjected to this 'new' treatment. More importantly, how can they identify them with assurance? It would be wise to be suspicious of any corundum displaying a shade of orange; but this is hardly a sufficient action for any qualified gemmologist.

Most of the illustrations published, to date, show a nice colour rim with boundaries parallel to the cut stone's surface. This may well occur when cut stones have been subjected to the 'new' treatment process and then repolished in the same shape. But, what about when rough is treated and then cut. The areas of colour would not appear as rims, would not necessarily parallel any surface of the cut stone, and in fact could very easily be mistaken for natural colour banding.

In figure 14 it can be noticed that the thickness of the pink rim left remaining after preforming the treated rough varies considerably from one part of the stone to another, and from stone to stone. Often the rim of induced colour has been almost completely removed.

The biggest problem in identifying corundum that has undergone this 'new' treatment will be when no outer rim is visible; either because the induced colour continues through to the core of the stone, or the colour and size of the stone make visual detection of a differently coloured rim difficult (See figure 15).

Sensitive analytical equipment may indicate increased levels of, say beryllium, towards the outside of a stone; but how can normal gemmological equipment be used to detect this?

Ladies & gentlemen, fellow front line gemmologists, Have we finally met our match?

Acknowledgements:
The author wishes to thank the following colleagues for their advice and assistance: Dr John Emmett, Richard Hughes, Arnun Jarugosol, Nuttachai Ratanasheworn, Graham Henry, and Susie Robson.