Replication of Isaac Newton’s Regulus of Antimony

Regulus of Antimony made by Reduction with Metals                               

Because of its interesting semi-metallic properties, many early modern alchemists had a serious interest in antimony (see entry under “Antimony”).  However, antimony was typically available only as the mineral stibnite (Sb2S3) and thus had to be refined in order to arrive at a “regulus” or button of the metalloid.   This was typically performed in one of two ways: either by smelting the stibnite with a metal or with crude or partially calcined wine tartar, often in combination with saltpeter; when refined with tartar, the product was referred to as regulus per se. If a metal was chosen as the reducing agent, the choice was usually iron, but other metals were also employed for specific purposes.   Here we are concerned only with the reguli made by using one or another metal as a reducing agent.  Early modern chymists distinguished between “martial regulus,” produced by smelting the stibnite with iron, “jovial regulus” refined with tin, “venereal regulus” made with copper, “saturnine regulus” fabricated with lead, and “lunar regulus” made with silver.  A serious question emerges as to whether these different reguli actually contained significant quantities of the respective reducing agent after refining, or whether the initial metal remained only as a trace element.  In order to approach this question, successive attempts were made, based on the comprehensive experimental laboratory notebooks left by Isaac Newton (Portsmouth Additional MSS. 3973 and 3975, edited on the Chymistry of Isaac Newton site at  

All reguli were made using Newton’s recipe in Portsmouth Add. MS 3975, folios 42r and 42v. Unless noted, all samples were made using lab-grade chemicals. Newton suggests using small quantities of metal to produce the best regulus, so we used between 1 and 2 ounces of stibnite for each reaction. All weights are measured in grams. As one of the byproducts of the reaction, antimony trioxide [Sb2O3] is a toxic gas, all replications save the last were performed in a small electrical furnace that was placed inside of a fume hood. Newton lists four metals with which reguli may be produced: iron, copper, tin, and lead. In each of these cases, antimony has a higher electron affinity than the reducing metals, so a replacement reaction occurs: the antimony is reduced by the oxidizing metal of choice, giving elemental antimony and a metallic sulfur compound. As will be shown in the results, however, the process does not produce a pure elemental antimony, as large quantities of the reducing metal are still present in the regulus. Part of the antimony is also released as the gaseous antimony trioxide.

Regulus of Antimony made with Copper

Figure 1. Regulus of antimony made with copper. The regulus (silver-colored) is visible only after slag (dark grey) was split in two with a wrench.

Regulus of Copper: Sb2S3 + 3 Cu → 3CuS + 2Sb

Newton gives a ratio of ½ part copper for every 1 part stibnite to be fluxed with ¼ part potassium nitrate [KNO3]. Using a digital lab scale, we mixed 15.467 grams of copper with 31.040 grams of ground stibnite, then heated the mixture in an electric furnace at 800° C for 5 minutes, until the metals were molten. A flux of potassium nitrate—Newton’s saltpeter—weighing 7.840 grams was added to the crucible. The mixture was then returned to the furnace, and the temperature adjusted to 1,000° C. When the furnace reached 1,000° C (approximately 26 minutes), the crucible was removed and the liquid poured off into a greased crucible.

The receiving crucible split when the molten material was added. Upon breaking, we found a dark grey slag with no apparent regulus. The regulus became apparent only after splitting the product in two. Inside the slag was a bright silver-colored regulus with a crystalline structure (figure 1).


Figure 2. Regulus of antimony made with copper. Slag has been removed, and regulus split to show interior. The interior surface of the regulus has been polished slightly to facilitate examination with scanning electron microscopy.

Regulus of Copper: Sb2S3 + 3 Cu → 3CuS + 2Sb

We attempted to produce a second regulus with copper, this time doubling the weights and using 62.000 grams of stibnite, 30.940 grams of copper, and 15.677 grams of potassium nitrate. Once again, the stibnite and copper were mixed and placed in a furnace at 800° C for 5 minutes until molten, then fluxed with the potassium nitrate, and the temperature was raised to 1,000° C. As Newton does not specify how long the mixture should remain in the heat, we left it in for 15 additional minutes after the temperature reached 1,000° C, for a total of 38 minutes. The molten material was then poured into a greased crucible.

We used a different type of receiving crucible, which did not crack upon contact with the regulus and slag. When the product was cool, we removed it from the crucible and broke off the slag. The slag was dark grey, and inside was a well-formed regulus, approximately the size of a quarter, but slightly thicker (figure 2). Once the majority of the slag had been removed from the regulus, it was weighed, and found to be approximately 11 grams. Newton says that 12 ounces of antimony reduced with copper should produce a regulus of approximately 3⅓ ounces, or 27.5% of the starting weight. We did not fare as well: our regulus was only 17.75% of the starting weight of stibnite. The theoretical yield was 39.5 grams of antimony and the observed yield was 11.0 grams, or 28%.


Figure 3: Regulus of antimony made with tin, refined once.

Regulus of Tin: Sb2S3 + 3Sn → 3SnS + 2Sb

Newton uses 5⅓ ounces of tin and 3 ounces of potassium nitrate for every 12 ounces of stibnite. We combined 30.997 grams of stibnite with 13.820 grams of tin and put them in a furnace at 800° C for 5 minutes, at which point they were molten. We added the potassium nitrate, returned the crucible to the furnace, and raised the temperature to 1,000° C. As we had received decent results leaving the copper regulus in the furnace for an additional 15 minutes after reaching 1,000° C, we decided to leave the tin mixture in for 35 minutes, for a total of 1 hour 9 minutes. The molten mixture was poured into a greased crucible.

Only a small part of the mixture was still molten when the crucible was removed from the furnace. The regulus was the size of a pinhead, and was lost when the slag was broken. The slag itself was a dark grey color, the size of a quarter. We conclude that the majority of the antimony likely evaporated as antimony trioxide during its extended time in the furnace. In a second attempt to reduce stibnite with tin, we combined 62.178 grams of stibnite with 27.606 grams of tin and 15.588 grams of potassium nitrate in the same manner described above. This time, we let the mixture sit in the furnace for only an additional 15 minutes at 1,000° C before pouring off the molten material into a greased crucible.

This preparation of regulus fared much better than the first: upon breaking the dark grey slag, we retrieved a small button of regulus, which weighed 4.705 grams. Newton does not provide a weight for the regulus made with tin, but we suspect that it is greater than what we produced, given the results of the copper regulus. The theoretical yield of this reaction is 18.863 grams, and our observed yield was 4.705 grams, or 24.9%.

To further purify the regulus, Newton heats it until molten, then adds ¼ ounce of saltpeter for every 1 ounce of regulus present. We thus melted the regulus at 800° C for 5 minutes, then added 1.393 grams of potassium nitrate and raised the temperature to 1,000° C. As we had received decent results by letting the mixture sit in the furnace for 15 minutes at 1,000° C, we left the regulus in for a total of 33 minutes. When the refined regulus was poured into the crucible, it separated easily from the slag, which was a mix of a light grey solid and a chalky white powder. The regulus itself split into two small beads when poured out, and was a bright silver color (figure 3).


Figure 4: Slag in heating crucible produced from making regulus of antimony with golfer’s lead.

Regulus of Lead: Sb2S3 + 3Pb → 3PbS + 2Sb

When making this regulus, we did not use lab-grade lead, because we were worried about the condition of the powder we had. Instead, we used powdered lead purchased from a golf supply company. This lead was not pure, and contained traces of arsenic. Following Newton’s ratio of 8½ ounces of lead for every 12 ounces of stibnite, we mixed 21.829 grams of our powdered lead with 30.799 grams of the usual stibnite. We melted this in the furnace for 5 minutes, then fluxed with 7.718 g of potassium nitrate. We once again tried leaving the mixture in the furnace at 1,000° C for a longer period of time—this time an additional 30 minutes after the 33 minutes required to heat the furnace by another 200°. We poured out the molten product into a greased crucible.

The regulus separated from the slag with some difficulty (figure 4): we had to smash the slag once cooled, which broke the regulus into four pieces. We were only able to remove the remaining slag from the regulus by scraping it with a knife.

We then wanted to know if using golfing lead had significantly impacted the results, so we made a second attempt with lab-grade lead powder. We mixed 30.724 grams of stibnite with 21.750 grams of lead and 7.530 grams of potassium nitrate, and followed exactly the same procedure as used in making the first regulus of lead. The results suggest that there was no appreciable difference in using lab-grade lead vs. golfer’s lead.

Regulus of Antimony with Iron

Figure 5: Regulus of antimony made with iron, removed from slag. The surface was pocked and had a slight golden tint.

Regulus of Iron: Sb2S3 + 3Fe → 3FeS + 2Sb

Newton’s regulus of iron is prepared using 4½ ounces of iron and 3 ounces of potassium nitrate for every 12 ounces of stibnite. We mixed 31.019 grams of stibnite with 11.662 grams of iron and 7.817 grams of potassium nitrate in the same preparation used in the previous trials. After raising the furnace to 1,000° C, we let the mixture sit for an additional 30 minutes, then poured out the molten material into a greased crucible.

The slag was a dark grey and was covered with many bubbles. It broke easily from one side of the regulus, which was pocked and cratered (figure 5). It took some scraping with a knife to remove the slag from the other side of the regulus, which was much smoother. When the majority of the slag was removed, the regulus weighed 7.322 grams. Once again, our yield was less than Newton’s: he says that the regulus should be slightly more than ⅓ the weight of the original stibnite, while ours was just under ¼ the original weight. The theoretical yield of this reaction was 16.97 grams of antimony, while the observed yield was 7.322 grams, or 43.15%, thus making this the most successful of all the metals in producing a regulus. This is unsurprising: of all four reducing metals, iron is the least electronegative.

Star Regulus of Iron: Sb2S3 + 3Fe → 3FeS + 2Sb

The reaction that produces the stellate regulus of iron is the same as that which produces the non-starred regulus: the only difference is that instead of pouring out the molten material immediately after removing the crucible from the furnace, the regulus is left to cool slowly under the slag, enabling the characteristic star pattern to form in the crystal structure of the antimony. We used 31.025 grams of stibnite, 11.698 grams of iron, and 7.826 grams of potassium nitrate, prepared in the same manner as the previous regulus of iron. The crucible was left on a stand and cooled for 24 hours.

After 24 hours, we broke the crucible and removed the slag and regulus. It was difficult to remove the slag from the regulus, so we let it sit an additional 48 hours to soften. After 48 hours, the majority of the slag came off, but to our dismay, we did not see a stellate pattern. Such patterns are difficult to form, so the trial was not a failure.

Regulus of Iron: Newton’s furnace

For the final trial, we made another regulus of iron, but this time in an outdoor furnace that had been fashioned as a replication of Newton’s. Therefore, the procedure differed from the others: we combined approximately 30 grams of stibnite and 12 grams of iron in a crucible, which was then placed in a bed of coals in the furnace, which had reached approximately 800° C. After 5 minutes, we added the potassium nitrate and returned the crucible to the furnace. After about 30 minutes, we removed the crucible from the furnace and poured out the molten material into a greased crucible. The resulting regulus was about the size of a quarter, and had a golden sheen to it. It weighed about 7 grams.


Figure 6: Pyrite crystals formed on the surface of regulus of antimony made with iron (400x magnification with scanning electron microscopy).

As Newton believed that there were still appreciable quantities of the reducing metals in the regulus—hence their need to be refined several times—we used scanning electron microscopy (SEM) to determine the elemental composition of the reguli. Newton was correct: each of the reguli contained various amounts of their reducing metals, as well as other impurities. Some of the impurities came from the nature of the materials used – such as arsenic in the golfer’s lead – while others may have come from the sanding and cutting used to prepare the samples. We also discovered some interesting structures on the surface of the reguli: under 400x magnification, we were able to see a stellate pattern on the regulus produced with iron that had not sat under slag (figure 6). Though we at first assumed that this was the early stage of the formation of the stellate pattern in the crystal structure of the antimony, further testing through SEM showed that it was actually pyrite (FeS2) crystals. These crystals covered much of the surface of this regulus, and gave it a vaguely golden sheen in some lights.

Unfortunately, the results did not provide us with a complete picture of the elemental composition. In order to get a truly accurate result from SEM, a thin slice of the metallic sample must be smoothed to a mirror polish and embedded in an epoxy base. We thus plan to prepare our reguli in such a manner in the hopes of getting more accurate results from future SEM work.

Primary Sources

Beguin, Jean. Tyrocinium Chemicum. ed. Jeremia Barthio. Wittenberg. 1656.

Lemery, Nicolas. A Course of Chemistry Containing an Easy Method of Preparing Those Chymical Medicines. Translated from the 5th Edition in the French, by Walter Harris, M.D. London. 1686.

Macquer, Pierre-Joseph. Dictionnaire de Chymie. Paris. 1766.

Newton, Isaac. Portsmouth Collection Add. MS. 3975. The Chymistry of Isaac Newton. Ed. Newman, W.R. February 2006.

Newton, Isaac. Portsmouth Collection Add. MS. 3973. The Chymistry of Isaac Newton. Ed. Newman, W.R. February 2006.

Newton, Isaac. MS. Var. 259. The Chymistry of Isaac Newton. Ed. Newman, W.R. February 2006.

Stahl, George Ernhest. Opusculum Chemicum Physico-Medicum. Magdeburgicae. 1715.

Valentinus, Basilius. Basil Valentine with his Triumphant Chariot of Antimony with Annotations of Theodore Kirkringus, M.D. London. 1678.

Von Suchten, Alexander. Secrets of Antimony in Two Treatises: Translated out of High Dutch by a Dr. C, a Person of Great Skill in Chymistry. London. 1670. 


Critical Editions

Starkey, George. Alchemical Laboratory Notebooks and Correspondence. ed. William R. Newman and Lawrence M. Principe. Chicago: University of Chicago Press. 2004   

Copyright © 2018 by

Meagan S. Allen

Replication of Isaac Newton’s Regulus of Antimony