Reconstructing Indian alchemy: Making silver, second method
Out of all the recipes we chose, this one probably had us scratching our heads the most: The Rasaprakāśasudhākara’s second recipe for making silver. It features uncertain, and exotic materials, and an unusual method.
खण्डं कर्षप्रमाणं हि सुमलक्षारकस्य हि
वेष्टितं नरकेशेन द्रुते नागे निमज्जितम् //७४//
निर्वापितं निम्बुजले चैकविंशतिवारकम्
द्रुते शुल्वस्य गद्याणे रक्तिकापञ्चमात्रकम् //७५//
कल्कं दद्यात्प्रयत्नेने तारवर्णं प्रजायते
गयाणे चतुरो वल्लान् रूप्यं दत्त्वा प्रगालयेत्
जायते रुचिरं तारं सत्यमेतदुदीरितम् //७६//
One should envelop a piece of white arsenic weighing 1 karṣa (= ca 10 grams) in human hair, dip it into melted lead and quench it in lemon water twenty-one times.
One should place five raktika ( = ca 0.62 grams) of the paste into a gadyāna (= ca 4 grams) of melted copper with care. A silver colour will emerge.
Add four vallā (1.5 grams) of silver to the gadyāna and melt this.
Shining silver will emerge. This is said to be true.
Let’s begin with problem ingredient one: white arsenic. The text reads “sumalakṣāraka” – a term we have not been able to find elsewhere, apart from in one other verse from the Rasaprakāśasudhākara’s eleventh chapter. We did, however, find reference to “somalakhāra”, a Marathi word, meaning “white arsenic”. The Sanskrit term “somala”, without the “kṣāra”, is also understood to be white arsenic. So our first decision was to disregard the spelling in the recipe, and interpret sumalakṣāra as somala, calcium arsenate, or white alkali of arsenic: white arsenic for short.
White arsenic comes under a number of names in rasashastra texts: somala, gaurīpāṣāṇa, malla(ka), śaṅkhaviṣa, pīta, vikaṭa, haṭacūrṇaka, mūṣaka. Unlike the other forms of arsenic, such as orpiment (haritāla, tāla) or realgar (manaḥśilā), white arsenic doesn’t seem to have been used widely in early Indian medicine or alchemy. The Rasaratnasamuccaya (chapter three, verse 126) notes that its extract is useful for the binding of mercury and for augmenting its potency, that it is unctuous (snigdha), and removes the humours (doṣaghna), but gives no further information on its medicinal or alchemical applications. The 20th-century Rasataraṅginī (chapter 11, verses 138-144) has a bit more to add to white arsenic’s therapeutic applications:
“It improves a person’s virility. It cures sandhigata vāta (arthritis), phiraṅga roga (syphilis), agnimāndya (loss of appetite) and viṣama jwara (intermittent fevers). It improves the body radiance, cures chronic anaemia (pāṇḍuroga) and chronic pratamaka śwāsa (dyspnoea). If rājayakṣma roga is detected in initial stages and treated with this medicine, the disease will be cured. Hṛdaya śūla and associated jwara and hṛdaya dourbalya will be cured with this medicine. It is also beneficial in atisāra roga. If used for external application, this medicine acts like ‘kṣāra’. Its external application alleviates inflammations, burning sensation and local temperature.” (translation by Ravindra Angadi, Rasataraṅginī, page 179)
[Commercially available orpiment and realgar]
You cannot buy white arsenic, so we decided to prepare ash (bhasma) of arsenic sulfide (= orpiment) for the purpose of this experiment. The Materia Medica of the Hindus (Udoy Chand Dutt 1877:39) states that “White arsenic was obtained artificially in ancient days by roasting orpiment.” We looked for recipes for orpiment ash and found several – none, however, from the Rasaprakāśasudhākara.
We did find two recipes for it in Materia Medica of the Hindus (Dutt 1877:42):
“Bhavaprakasa recommends that orpiment should be powdered and made into a ball with the juice of Boerhavia diffusa (punarnava) and placed in the centre of a pot full of the ashes of that plant. The pot should now be covered with a dish, luted with clay, and heated over fire for twenty hours. When cool, the ball of roasted orpiment is taken out from the pot and reduced to powder”
Another process is as follows. Take of purified orpiment and yavakshara, equal parts, rub them together with the juice of Vitex Negundo (nirgundi), and roast the mixture in a closed crucible. The resulting compound for both these processes is described as a white camphor-like substance.”
Our first step was to purify the arsenic sulfide (haritāla). Andrew used his notes from Sri Lanka for a preliminary simple version, which advised soaking yellow, red, or white arsenic in plantain juice for three days. In a further purification method, he boiled it in ash pumpkin juice. Once purified, it was ground and triturated with a decoction of punarnavā. The first recipe of the Materia Medica of the Hindus (MMH) recommends using the juice of this plant but we did not have fresh plant material and so had to substitute with the decoction. The mixture was then rolled into a small ball and dried in the sun.
[punarnavā- Boerhavia diffusa]
[Ball of arsenic (orpiment), mixed with punarnavā decoction, placed on a bowl of punarnavā ash, with more plant ash beneath]
[Topped with punarnavā ash lid]
[Sealed crucible in sand bath]
Next, Andrew mixed powdered punarnavā into a ball with punarnavā decoction and made a surround for the arsenic ball. This was placed below and above the ball, its two parts cupping the ball. This cup was then dried. In the MMH, the arsenic ball is surrounded by a pot full of punarnavā ash. We did not have quite enough punarnavā to make the required amount of ash, so instead, Andrew placed the arsenic ball with its punarnavā surround in a pot filled with plant ash. This was then sealed and dried. It was then roasted on an open fire for 12 hours. The MMH advised 20 hours of roasting, but as we were using gas, we thought 12 hours would be sufficient. Upon opening the crucible, Andrew noted that a bit of sulphur had sublimated and that some red arsenic crystals had also formed. Some parts of the arsenic ball were white but the majority was either silver or brown. The silver colour means sublimation occurred, so the temperature must have gotten higher than anticipated. The material was recovered, ground and stored for analysis. We considered this a fail, and decided that we would not use it for the recipe.
We prepared a new batch of the arsenic sulfide ground with punarnavā juice and made another ball. This was dried and again placed in the pot, this time, we decided to lower the cooking temperature and expose the ball. It was placed directly into punarnavā powder and capped with the same lid of herbs we’d previously used, the idea being that the heat would have less material to pass through before reaching the ball, but overall, the heat would be much lower this time as the flame would be set to a single ring. On the first attempt both rings had been used. The pot was sealed and cooked for 6 hours at low temperature. Upon opening, the interior was identical. Same red and yellow crystals, and the arsenic was silvery and white in places, but brownish when ground. We know
sublimated arsenic is silvery, so again the temperature in the pot must have been too high.
You can view the film of this method below.
Andrew decided to try the second method given in the MMH, according to which the arsenic is ground with chaste tree (nirguṇḍī) liquid and then mixed with an equal quantity of yavakṣāra ash (ash of barley). This was dried into four flattened balls and placed in a bed of barley ash in a pot that was then sealed and dried. The whole thing was then roasted, this time at a low temperature for 12 hours. The result was better than the two previous attempts, yielding a greyish white ash.
However, over a week, it oxidised and looked darker. Andrew was nevertheless satisfied that most of the original sulfur content was sublimated and had left a pale metallic oxide, though not quite the camphor-like powder described in the MMH.
[Ground orpiment, before chaste tree liquid is added]
[Barley, roasted into an ash, soaked in water - after boiling the water off, the salt remains, and is added to the orpiment]
[Orpiment with barley ash]
[Orpiment formed into cakes, placed on sand, to be roasted in an enclosed vessel]
This film shows the second method.
Though the result was not bad, Andrew decided he wanted to try one further way of obtaining orpiment ash, this time using a recipe from the Rasataraṅginī. Here, the orpiment was ground with Aloe vera and then formed into cakes. Layered with ground cuttlefish bone above and below, each cake is placed in between two cow dung cakes that are then fastened together. All of this is then placed into a roasting pit, where it is surrounded by further cow dung cakes as fuel. After roasting, the pieces of oxidised orpiment are removed and ground. The pieces were light-coloured, though not quite white.
[Orpiment with ground cuttlefish bone on cow dung cakes]
[The cow dung cakes that contain the orpiment are roasted in a pit, with further cow dung used as fuel]
[The resulting orpiment ash]
This film shows the third method:
Making alchemical silver
We now started on the actual recipe, using the batch of orpiment ash from the second method. Its first step involves enveloping ten grams of white arsenic in human hair, and then dipping it into melted lead and quenching that in lemon water. This is repeated twenty-one times.
Have you ever tried enveloping powder in hair? Spoiler alert: It's difficult.
Actually, there is something to say about the use of human hair in the first place. It is an unusual ingredient, not often found in Sanskrit alchemical literature. It is found in four main contexts: for constructing a particular kind of crucible (see Rasaratnākara Rasakhaṇḍa 2.44) and also for making a mud seal for a crucible (see, for example, Rasaratnākara Vādakhaṇḍa 17.27); for mixing diamonds and gold (see Rasārṇava 14.159-160) and finally, for making hair oil, i.e., oil made from hair, which is then applied in a procedure dedicated to liquifying mica essence (Rasaratnākara Vādakhaṇḍa 17.33). We could not find any parallel to our silver-making recipe. Was this a misreading in the edition?
Since we currently have no access to manuscripts of the treatise, we decided to take it at face value. Which left us with the practical problem of how to envelop powder in hair. Andrew had two solutions for this: 1) mixing the arsenic ash with aloe to make a clay-like nugget, and 2) using gum arabic to make the hair stick after enveloping the arsenic nugget. Neither solution has a textual basis, and it’s uncertain whether this would impact the outcome. In the absence of a better solution, we went with it. We should note, however, that while the first batch Andrew made was made using the acacia gum, he was able to make a second batch without gluing the hair with it.
[Not too pretty: orpiment ash mixed with aloe, enveloped in hair, fixed with Acacia gum]
Next step: Dipping it in lead. For this, Andrew melted lead using a sand bath for even heat distribution. Lead has a high density, so it was a bit of a challenge to submerge the hair nuggets in it. However, we found that you can roll them in the lead and they pick up a surface coating of lead that way. It was somewhat tricky to then quench the lead-coated hair nuggets in lemon juice without inadvertently fusing the tongs together in the process. Andrew managed, somehow, by dipping only parts of the nuggets into the lead at one time, keeping the tongs free of lead. The dipping and quenching was supposed to be done twenty-one times: because Andrew only did parts of each nugget at one time, it was more like fifty to sixty dippings and quenchings.
[Lead in sand bath]
[Hair-orpiment ash nuggets ]
[Lemon juice for quenching]
Next, the arsenic-hair-lead nuggets would be added to melted copper. Here, the text mysteriously speaks of adding “the paste” (kalka) to the copper. What paste? We discussed grinding the balls, but Andrew found that the lead didn’t really yield to grinding. Instead, he cut one of the nuggets in half, exposing the arsenic middle, and added that to the copper. Just as the hair-arsenic-nuggets were not easily immersed in the lead, the lead-coated nuggets also didn’t sink into the copper, but floated on top, simply burning off there. Andrew tried to stir them in, but didn’t manage to get them fully mixed with the copper: they kept on coming back up to the surface. As in the previous experiments, Andrew also added gunja seeds to hinder oxidation. These do not mix with the alloy, but just float on top, forming a barrier to the air. The resulting alloy was in fact quite even, despite the difficulties in mixing the lead, arsenic, and copper.
Note that in the recipe, while one first starts with 10 grams of arsenic and an unspecified amount of hair and lead, only 600 milligrams of this, in the form of a paste, are added to the copper. Given that there are no instructions on how the paste is made, the question arises whether there is another step in between that is omitted in the text, or that is simply missing in the edition. Something through which there is perhaps a loss of the arsenic-hair-lead mixture, therefore accounting for the mismatch in amounts between the first and second steps?
Adding liquid mercury (Hg) would be a reasonable contender for a missing step, since the mercury would break the lead and arsenic down and absorb them, and the lead-arsenic-mercury mixture would also have the consistency of a paste. Heating this mixture would also result in a reduction of material, which may account for the small amount of the paste added to the melted copper. However, this is just speculation, and since we neither have access to manuscripts of the text, nor found parallel recipes in other alchemical works, we could not ascertain whether such an additional step was indeed required. We therefore just used the named ingredients of the recipe.
[Copper is melted in the furnace]
Here are the ingots of the copper-lead-hair-arsenic mixture (pure copper on the right for comparison):
There was still one final step: mixing the prepared ingots with silver. The recipe asks for quite small quantities, which are difficult to process and pour, so Andrew applied the recommended proportions rather than the actual recommended amounts. In the final melting and mixing of the silver and copper-lead alloy, he again used gunja seeds to hinder oxidation.
Despite all our doubts about ingredients and proper procedure, after polishing, the final product did look like silver, albeit with a slightly rosy hue.
[Alchemical silver, before and after filing]
This recipe seemed unnecessarily complicated. Why not just add lead and arsenic to copper, rather than making the hair nuggets first? It might have made sense if the copper and then the silver were only used to coat the nuggets further, but given that the wording of the text suggests that the hair-arsenic-lead nuggets were made into a paste before being added to the copper, this does not seem to have been intended.
It was generally rather difficult to know what was intended, given the irregularities in the use of terms for substances (sumalakṣāra rather than somala) and indeed the choice of substances (hair), and the unexplained mention of a paste (kalka) to be added to the copper. And, not only were we unsure whether white arsenic was the correct substance to begin with, we also hadn’t quite produced a completely convincing version of it. All in all, almost every step of this recipe was fraught with difficulty and doubt.
On the bright side, the ratio of the silver to the other ingredients was quite low: 1.5 grams of silver to 4 grams of the copper-lead-arsenic alloy. So at least it was not quite as risky an investment as our other silver-making recipe, in which silver formed half of the weight of ingredients. And the final ingots do, more or less, look like silver. So perhaps we can call this a success of some sort?
In the end, it seems to us that while this was a valiant effort, we did not quite replicate what the authors had in mind, mostly because we suspect some crucial information was missing. Perhaps we will find the full recipe at some point and will revisit this formula (though Andrew says he never wants to do the hair nugget thing again!).
Here is the film: