Zinc and mercury retorts, direct-fusion brass

Covers period texts read as archive.org OCR text. Page numbers are printed folios read from the OCR running heads; OCR is rough.

  • J. Percy, Metallurgy: Fuel, Fire-Clays, Copper, Zinc, Brass, etc. (London, 1861), archive.org metallurgyartofe00percrich. Zinc pp. 520-596, brass pp. 608-622; works seen 1848-1859.
  • J. A. Phillips, A Manual of Metallurgy, 3rd ed. (London and Glasgow, 1859; first ed. April 1852), archive.org manualofmetallur00philrich. Mercury pp. 452-463.
  • A. Ure, Dictionary of Arts, Manufactures, and Mines: vol. 1 (7th ed., 1875, dictionaryofarts01ureauoft), art. Brass pp. 464-466; vol. 3 (1861, b21780559_0003), art. Mercury, about pp. 58-63 (heads garbled). Ure's text dates from 1839, so it describes practice before 1840.
  • W. R. Ingalls, The Metallurgy of Zinc and Cadmium (New York, 1903), archive.org metallurgyzinca00ingagoog. LATER than the era: used for dates, the early Liege furnace and blue powder. A "Hofman (1903)" is Ingalls; Hofman's book is 1922 and was not found as text.
  • T. Egleston, The Metallurgy of Silver, Gold, and Mercury in the United States vol. 2 (1890), metallurgysilve02eglegoog, pp. 808-813. LATER (California, 1870s), small iron retorts.
  • Not searched: Makins, Grace's Guide. Purpose period facts on zinc and mercury distillation in retorts and on direct-fusion brass. INFERRED marks the writer's own reading.

A. Zinc distillation, 1830-1870

  • Belgian (Liege) process: "first practised in the vicinity of Liege in the early part of the present century, but in what year I am unable to state"; the furnace and its apparatus "have remained substantially the same to the present day" (Percy, p. 578). Dony: "In 1805 the Abbe Daniel Dony, a chemist of Liege, Belgium, discovered independently the method of zinc smelting and from his experiments the Belgian zinc industry developed" (Ingalls, p. 2, 1903). Ingalls: Belgian furnaces "experienced no radical changes from Dony's original design until the introduction of gas firing" (p. 428). The original ore was calamine only (Percy, p. 578).
  • English (Bristol) process: zinc in large closed pots, vapour led down through a hole in the pot bottom into a chamber below, "per descensum". Percy saw it near Swansea 1848 and Neath 1859 and doubts it is English in origin (p. 550). Ingalls: Champion's Bristol works "in 1740, or about that time"; "continued in use until about 1860" (pp. 1-2). Percy: Watson places Champion's Bristol works at about 1743; Champion's patent of 1739 (OCR) "had no reference to zinc", that of 1758 covered spelter and brass from calcined blende and charcoal (pp. 520-521).
  • Silesian process: per ascensum, vapour led out of the retort mouth downward through condensing tubes (Percy, p. 558); built at Wessola in 1799 or 1800 from the pots of a glass furnace (Ingalls, p. 2). Percy describes Dillwyn's works at Llansamlet, Swansea (pp. 558-576).
  • Carinthian: vertical clay pipes 40 in long, 4.5 in bore, in a wood-fired arched chamber, "formerly practised" (Percy, p. 585); it "did not long survive" (Ingalls, p. 2).
  • At one British works in 1861 both ran, about 100 t of zinc a week, 17 pounds a ton (Silesian) against 18 (Belgian) with ore (Percy, p. 584).

Belgian furnace (Percy, pp. 578-585, from Thum 1859-60 and Vieille Montagne drawings)

  • Retorts: cylindrical, closed at the back, clay made and moulded as the Silesian (p. 579). Size: 8 in outside diameter, 3 ft 6 in long in the engraved furnace (the OCR reads ";> feet G inches"; 5 ft could not lie in its chamber, 4 ft from front to back inside) or about 3 ft 3 in at Moresnet, whose chamber is 6 ft 6 in from front to back inside (p. 581); Ingalls has the Moresnet retort 1.1 m by 0.15 m, inclined 0.078 to 0.158 m (p. 429).
  • Setting: rows one above another in a narrow arched chamber, fire-place along the bottom; retorts rest on ledges in the back wall and incline toward the front so corrosive slag flows to the cooler end; two retorts per niche; a row of thick EMPTY retorts over the fire shields the first row; furnaces built in fours round a stack (pp. 580-581).
  • Ironwork: all inner parts of good fire-brick; "The furnace is securely clamped and braced by iron standards and tie-rods" (p. 581, figs. 146-148). Ingalls (LATER, pp. 374-375): cast-iron buckstaves are used on distillation furnaces; a standard stands each side of every working door; in a distillation furnace rods cannot pass through the walls, so the standards' feet are held by flagstones or a beam bedded in the ground and only upper tie-rods are used, 1 1/4 in round, looped over the heads of an opposite pair and drawn up by a turnbuckle or by a U-yoke and a pair of nuts; rods are kept out of brickwork near the fire.
  • Retorts per furnace: 78 (large Vieille Montagne), 61 (Moresnet) (p. 581); the first Liege furnaces had 40 small ones in eight rows (Ingalls, p. 429).
  • Condenser: a clay nozzle luted into the retort mouth, a sheet-iron vessel (receiver) on its outer end with a small hole in the closed end; the nozzle is smeared with milk of lime so crusts part easily (pp. 579, 584). The sheet-iron tubes are the English process's (p. 551); the Belgian nozzle is clay with an iron receiver.
  • Charge: finely ground CALCINED calamine of about 50 per cent zinc, moistened and mixed with carbonaceous matter; "most non-caking coal-slack", which "may advantageously be mixed with half its weight of cinders or coke-dust"; low in sulphur. Charged "twice daily - once in the morning and once at night" (p. 583).
  • Per retort: lower half of the furnace 12 to 12.5 kg (26.5 to 27.5 lb), upper half 6 to 7 kg (13 to 15 lb); the hotter rows are loaded more; poorer and iron-rich ore goes to the upper retorts (p. 583).
  • Proportion, Moresnet day (Thum, p. 584): 1035 kg calamine, 549 kg reducing matter (coal), 371 kg zinc plus 37 kg zinc fume = 39.0 per cent yield of the ore, 1479 kg fuel coal, 3.8 retorts and 11.5 nozzles replaced. INFERRED: reducing matter is about half the ore by weight.
  • Firing: coal on a grate; no degrees given, only that lower rows run hotter than upper (p. 583).
  • Drawing the zinc: the receivers are placed when vapour shows; "after the lapse of 6 hours most of the nozzles will be filled with melted zinc, which is withdrawn by means of a little scraper", repeated before charging afresh (p. 583). Ingalls (1903, Practice in Distillation chapter, folio not captured): metal tapped three times in 24 h into a kettle, coke dust thrown on, oxide and blue powder skimmed for the next charge, clean metal poured into slabs of about 50 lb.
  • Yield and loss: a Moresnet furnace treats 1200 kg ore in 24 h and gives 450 to 470 kg zinc, "inclusive of that in the fume"; the actual yield of ore of 50 per cent is 38 to 39 per cent. Loss 11 to 12 per cent of the zinc, half recovered from residues, half lost "partly by volatilization from the nozzles, partly by the escape of zinc from bad retorts, and partly by imperfect reduction"; volatilization alone 1 to 2.5 per cent; coal per ton of zinc 5.46 t with the reducing coal, 3.98 t without (pp. 584-585). Blue powder (zinc fume with oxide film): "about 5 to 10% of the zinc distilled", reworked in the next charge (Ingalls, p. 626; LATER).
  • Retort life and replacement: "Cracked or perforated retorts must be replaced by new ones, which have been previously gradually heated to redness in an adjoining kiln. This operation generally occurs every morning" (p. 583). INFERRED life: 61 retorts at 3 or 4 a day, about 15 to 20 days. A retort cost 1 fr 60 to 2 fr 10 (1857); one man made 18 to 20 a day by hand (p. 583).

Silesian furnace (Percy, pp. 558-576, Dillwyn's practice by W. Penrose)

  • Retort clay: about equal weights of best Stourbridge clay and powdered old glass-pot sherds, tempered for hours, moulded in sections, dried three months, 1.75 cwt dry; annealed red hot before setting (pp. 559-560).
  • 24 retorts per furnace, two to an arched recess; 24-hour cycle: emptying from 6 a.m., 3 to 4 hours to rake out and recharge, heat raised to near white by 4 p.m., zinc condenses from about 5 hours after charging, brisk from 6 p.m. to 3 or 4 a.m., over by 6 a.m. (pp. 569-570).
  • Charge for 24 retorts: 1568 lb calcined blende plus 150 lb sweeps, 5 cwt fine bituminous coal, 2 cwt coke or ashes, mixed; about 71 lb per retort, more in the hot ones (p. 569). "A Silesian furnace lasts on an average 2 years, with a monthly consumption of 8 retorts" (p. 574).
  • Leaks are mopped with fire-clay on a rod through roof holes; opening them chills the retorts and "checks distillation sometimes to the extent of 15 per cent of the metal" (p. 570). Rough zinc is melted in an 8 cwt iron pot, skimmed (the "sweeps" go back in the charge) and laded into iron moulds (pp. 558, 570).

English pot furnace (Percy, pp. 550-558)

  • Six pots in an octagonal house with a cave below; each pot has a hole in its bottom over a sheet-iron tube (short conical upper, long lower, 1/16 in iron, clayed); pots of Stourbridge clay and old pot, 25 shillings each; a cracked pot is replaced red hot with wheeled tongs (pp. 551, 556-557).
  • Six pots take 20 cwt calcined blende in alternate layers with coke (not mixed); 67 h a charge; 6 to 8 cwt spelter; 22 to 27 t coal and 3 t raw blende per ton of zinc (pp. 557-558).

Which a small works would build: not stated. INFERRED: the Belgian row of small clay retorts on a coal grate can be built and mended one retort at a time (under 2 francs each); the English pot needs a house with a cave.

B. The roast

  • Blende is roasted to oxide, calamine calcined; the roast is a separate step before the retort. Percy: blende "is one of the most troublesome to roast sweet", but does not clot by heat, so "may from the first be exposed to a much higher temperature than most other sulphides" (p. 541). Roasted at a gentle red heat it gives sulphurous acid, oxide and sulphate; the sulphate needs a strong red heat to decompose (p. 541).
  • Furnace for blende: a reverberatory flat-bedded calciner, "in the construction of which there is nothing peculiar" (English works): charge 16 cwt, roasted in 24 h with about 2 t coal, stirred at intervals, 20 per cent loss of weight (p. 550). A double-bedded one on the reduction furnace's waste heat: 10 x 6 ft beds, 5 cwt, 12 h at a red heat on the upper bed then 12 h hotter on the lower (p. 551). Silesian: one ton per charge on an upper floor 15 ft x 9 ft, stirred every ten minutes, drawn to the lower floor at the eleventh hour, 22 h in all, "dull to bright redness", until sulphurous vapours cease; zinc lost in roasting about 2 per cent (pp. 567-568). "It is the custom in most zinc-works" to roast in a two-bed calciner on the waste heat of the distilling furnace, 12 h per bed (p. 569).
  • Calamine: at Moresnet it is calcined in circular KILNS, 7 ft 2 in across the mouth, 9 ft 8 in at the widest, 17 ft 6 in high, in alternate layers of ore (lumps of about 6 in) and inferior coal, 3 to 4 per cent coal, drawn 4 to 6 times in 24 h, 25 t a day; fine ore goes to double-bedded reverberatories, 6 h a bed (Percy, p. 582). So calamine WAS calcined in coal-layered shaft kilns, as iron ore was.
  • Blende in a shaft kiln: not the practice in the period. Percy describes only reverberatory calciners for blende. Ingalls (1903): "Zinc blende alone cannot be successfully burned in stalls" (heaps); shaft kilns "are used at several works in Europe for the desulphurization of blende without fuel, but their action is incomplete and the roasting has to be finished in reverberatory furnaces"; kilns take lump ore only (pp. 67-68; LATER, naming Freiberg and Upper Silesia). Short answer: blende was roasted in reverberatory calciners; a kiln only part-roasted lump ore.

C. Mercury from cinnabar

Small works (Phillips 1859; Ure, pre-1840 practice)

  • Before 1635 mercury was distilled "per descensum" in two stacked earthen pots; the Palatinate then moved to "galleries", first earthenware then iron retorts (Ure vol. 3).
  • The Palatinate gallery (Deux-Ponts): 30 to 52 earthen retorts (cucurbits) in one furnace, each with a stoneware receiver part filled with cold water; "Into each of the retorts are introduced from 56 to 70 lb of cinnabar, and from 15 to 18 lb of quick-lime, a mixture which should fill about two-thirds of its capacity". Fuel pit coal on a grate; dome pierced for draught (Phillips, p. 461). Ure's text says ores are put "into retorts of cast iron, sheet iron, or even stoneware, in mixture with an equal weight of quicklime"; the two disagree on lime (equal weight against about a quarter; OCR); Ure's is the older text. Lime takes the sulphur as sulphide and sulphate of calcium and the mercury is volatilised; "in a few places, ... the iron cinder driven off from the blooms by the tilting hammer" replaces lime (Ure). Iron filings are not named in either; Phillips names a bar of iron only for precipitating assays (p. 456).
  • Landsberg, Bavarian Palatinate, 1847: gas-style retorts in masonry, an iron lid and a cast iron eduction tube, a cast-iron condenser 18 in across full of water to a little above the tube ends, with a water valve against suck-back, in a trough of running cold water, inclined to an iron chest; a charge of 5 cwt per retort, "almost totally expelled in the course of three hours" (Phillips, pp. 462-463). Ure's design for the same: retorts at least 1 ft square and 7 ft long holding 6 or 7 cwt of ore and lime, 4 in tubes dipping under the water of an 18 in condenser pipe about 20 ft long, benches of three retorts (Ure vol. 3).
  • Egleston (1890; LATER; California 1870s): iron retorts set in brick, D-shaped, 18 in wide, 12 in high, 9 ft long; charges of 200 to 350 lb every 12 h; at another works 10 per cent quicklime on a 150 lb charge, 4 h firing, 500 lb ore in 12 h; a 6 in pipe 30 ft long into iron boxes with a water seal; mercury ran 2.5 h after charging; men wore wet sponges and were often salivated. He calls the method "the oldest, the simplest, and the worst" and says it was dropped at New Almaden about 1850 (pp. 808-813).

Large furnaces

  • Idria: mines found 1497; an aludel apparatus on the Almaden model set up 1750; the great furnace of about 180 ft x 30 ft erected 1794 (Ure; Phillips, p. 458: 1000 to 1200 quintals of ore per charge to 80 or 90 of mercury, firing 10 to 12 h at cherry red, a week per cycle); 1812 yield about 8.5 per cent of the ore.
  • Almaden: aludel furnaces ("buytrones"), ore piled over a perforated arch, earthen aludels 12 rows of 25 luted with loam, brushwood fire 12 to 13 h, cooling 3 or 4 days; ore averages 10 per cent (Phillips, pp. 453, 459-460). Bustamante is not named in Phillips, Ure or Egleston. A web search gives aludel furnaces at Almaden from 1646 and a Bustamante furnace "designed in 1720" in a UNESCO file; modern, conflicting, so no date is taken.

Zinc and mercury in one furnace or retort setting: not documented; no source read describes it. INFERRED from the colours given: mercury ores fire at cherry red (Phillips, p. 459), zinc near white heat (Percy, p. 569).

D. Direct-fusion brass

  • Emerson: "James Emerson obtained a patent, in 1781, for making brass by the direct fusion of its two metallic elements, and it is now usually manufactured in this way" (Ure vol. 1, p. 464).
  • Before it: cementation, "granulated copper, called bean-shot, ... with calcined calamine ... and charcoal in a crucible", 3 parts copper to 3 of calamine and 2 of charcoal (Ure, p. 464). Percy: "Until a comparatively recent period all brass was made by the old process of cementation, which has been almost entirely superseded"; it was patented by Champion in 1779 (p. 612). Calamine brass was still made at Llanelly in 1859 (p. 612).
  • Crucible and furnace: Birmingham furnaces "about 10 inches square and 2 feet deep"; London ones round; good coke fuel (Percy, p. 618). Ure: 10 to 12 in square, about 2 ft deep, Stourbridge fire-clay or rarely plumbago crucibles on iron bars, packed round with coke (p. 465). Calamine-brass pots for comparison: 8.5 in wide, 12 in deep; king pot holding 120 lb, small pot 84 lb (Percy, pp. 614, 616). Hegermuhl: 120 lb in four crucibles, about 30 lb each (Ure, p. 466).
  • Order: copper first, heated red in ingot, then melted; "The zinc is gradually and cautiously added to the copper when the latter has just melted" (Percy, p. 618). Ure: "The ingot copper is first introduced into the crucible, and when this is melted the proper proportion of zinc is cautiously added, the mixture being stirred with an iron poker". Ure's older route: slips of copper plunged into melted zinc, the rest of the copper later.
  • Cover: no charcoal cover on the melt is stated in these texts; charcoal appears only dusted in oiled ingot moulds and as the reducing charge of cementation. The coke round the pot is fuel. Not documented.
  • Proportions: "about 2 parts by weight of copper and 1 of zinc" for brass unqualified (Percy, p. 608; Ure, p. 464: "very nearly 2 parts of copper to 1 of zinc"). Muntz metal 60 copper to 40 zinc, patent of 1832, sheathing zinc not over 38 per cent (Percy, pp. 619-620). Hegermuhl charge: 41 lb old brass, 55 lb copper, 24 lb zinc (Ure, p. 466).
  • Loss to fume: "there is always an inevitable loss from the volatilization of zinc, for which a due allowance is made to the founders" (Percy, p. 618); Ure says the metals' affinity prevents loss "in a great measure" and gives 2.5 to 4 lb waste per 120 lb charge, i.e. about 2 to 3 per cent (p. 466). In a reverberatory the fume cannot be predicted, so zinc is added and the fracture tested (Percy, p. 620).
  • Pour: skimmed, poured into sand moulds, or into closed iron ingot moulds, warmed, lightly oiled and dusted with charcoal inside (Percy, p. 618; Ure, p. 465). Old moulds were granite slabs.
  • Time: calamine brass 10 h a charge (Percy, p. 615); Ure's two-melt route 8 to 9 h per melt (p. 465); a direct-fusion crucible's time is not stated.

Gaps

  • Furnace temperatures in degrees (only colour terms); retort wall thickness and English pot size (in Percy's figures, not read); the English charge "box" is unsized.
  • Zinc and mercury in one setting; a sourced Bustamante date; charcoal cover on direct-fusion melts.
  • Makins, Hofman (1922) and Grace's Guide not reached.