Pramāṇa · Search · About · API

Minerals and Metals in Ancient India v1.c26.s6.t1: Ancient Retorts and Furnaces at Zawar

Minerals and Metals in Ancient India v1.c26.s6.t1

Language
English

Ancient Retorts and Furnaces at Zawar

v1.c26.s6.t1.1

During the 1983 excavations two groups of furnaces were uncovered. A single bank of seven furnaces upon Zawar Mala contained small retorts 20 cm long and 10 cm in diameter. In old Zawar there was a more extensive arrangement of furnaces using larger retorts 30-35 cm long and 10-15 cm diameter. In both groups thirtysix retorts in a six by six arrangement were contained within the truncated pyramid of each furnace. The retorts were supported vertically on perforated bricks through which the condenser tubes passed into the cooler zinc collectors beneath (vide figs. 18.2-18.4).

v1.c26.s6.t1.2

The furnaces are in two parts consisting of a zinc vapour condensation chamber at the bottom and a furnace chamber at the top. The two chambers are separated by a perforated terracotta plate. The condensation chamber measures 65 cm by 65 cm and 20 cm in height. The perforated terracotta plate that separates the two chambers is a composite unit made up of four equal segments of 35 cm². It is 4 cm thick, well-baked, and sturdy. Its perforations include circular holes of two sizes: larger ones of 4 cm diameter, each of which are surrounded by a number of smaller holes of 2.5 cm diameter. Within the furnace the composite terracotta plate was found to be supported on a ledge in the furnace walls on all four sides and a single solid terracotta pillar placed below the junction of its four segments. (Hegde et al., 1986; Hegde, 1989)

v1.c26.s6.t1.3

Up above the perforated terracotta plate is the furnace chamber, in which 36 charged retorts (figs. 18.2-18.3) were arranged, inverted vertically; it may be presumed that 36 vessels were placed, one underneath each retort, to collect the condensed zinc vapour. This arrangement of downward distillation retort with the condensing unit underneath or distillation per descensum, is precisely what had been described in *Rasaratnasamuccaya* text (2.157-166; 9.48-50) to be discussed in chapter 9 of our Vol. II, and what was copied by William Champion in 1748 for his plant in Warmley, England. The brinjal-like retorts in Zawar are also similar to the *vṛntākamūṣā* described in *Rasaratnasamuccaya* (10.22-3).

v1.c26.s6.t1.4

Fig. 18.2 : Midpoint Vertical Section Through a Zinc Reduction — Distillation Furnace at Zawar Mala

v1.c26.s6.t1.5

Fig. 18.3 : Horizontal Section through the Furnace in Fig. 18.2 at Perforated Plate Level

v1.c26.s6.t1.6

Fig. 18.4 : Section Through Small Charged Zinc Reduction — Distillation Retort with Stick Still in Place

v1.c26.s6.t1.7

The furnace structures unearthed at Zawar did not include their roof part. However, curvature of their extant brick walls (*fig. 18.2*) suggests that the furnaces probably had domed roofs with a chimney-like opening at the top, similar to those of the vertical up-draught pottery-baking kilns of Mohenjo-Daro found by Mackay (1938). The opening would have served two purposes: (1) It would have acted as an outlet for the flow of hot gases generated within the furnace, and (2) it also facilitated continuous feeding of the furnace with fuel. The use of thin cow-dung cakes could have enabled quick attainment of a high temperature with a natural induced draught of free air coming through the front opening in the condensation chamber and the smaller holes in the perforated terracotta plate separating the condensation chamber from the furnace chamber. It is significant that the banks of furnaces at Zawar face south-west, probably to receive the natural draught during April to September. The dome-shaped superstructure of the Zawar furnace was possibly dismantled after the completion of each distillation process and rebuilt again for the next batch.

v1.c26.s6.t1.8

A very few undistorted retorts were found showing the original shape of the larger types to be cylindrical, ending in a cone. The retort walls are about 1 cm thick and of clay, containing large fragments of quartzite and phyllite, both local rocks. (Mookherjee 1964) The retorts were prepared in two parts: a condenser was luted onto the open main body of the retort after it had been filled with the charge. Of the condensers, only the heads survive, usually embedded in the retorts (*fig. 18.4*).

v1.c26.s6.t1.9

These heads appear to be of at least two types: a relatively large, thin-walled tulip-shaped variety and a thicker, more robust cone which was formed in a polygonal mould. Droplets of zinc-rich material are found just inside the broken, tapered ends of the heads. (Freestone *et al.*, 1985a : 231 and 243) The heads appear to have tapered into tubes which led the condensed zinc vapour to the collecting vessel underneath.

v1.c26.s6.t1.10

It appears that a cylindrical reed of 1.5 cm diameter was inserted into the retort after it was charged and the funnel part was luted on it (*fig. 18.4*); this is evident from the central hole which is preserved in many retort residues. (Freestone *et al.*, 1985b: 4) The reed helped to keep the charge within the retort when it was inverted and placed in the furnace. When ther furnace was fired, the reed burnt away leaving behind a cylindrical flow channel for the zinc vapour to flow freely out of the retort.

v1.c26.s6.t1.11

Petrography of the ceramic retort bodies and furnace walls shows that these were made from local (Tiri river bank) clays. (Freestone et al., 1985a) Analysis of the ceramic matrices, using Energy-Dispersive Analysis X-ray spectrometer (EDAX) attached to Cambridge S 600 Scanning Electron Microscope (SEM), revealed that these contain around fifteen weight per cent fluxes (FeO, MgO, CaO, Na₂O, K₂O). Thus the ceramics are closer to house bricks than to modern refractories in composition. (vide Table 18.1 for a complete analysis)

v1.c26.s6.t1.12

The temperature to which a ceramic has been subjected may be estimated by the examination of its micro-structure in the SEM. The methodology has been worked out by Tite et al. (1982). As we move from the inside of the retort wall to the outside in direct contact with the furnace flame, we find from the BSE micrographs of the polished thin sections, that the degree of inter-connection of the particles increases and the porosity decreases (indicating temperature 1000-50°C). Finally the pores are isolated and the trapped gases expand, causing the formation of rounded vesicles (T 1100-50°C) and at higher temperatures (1200-50°C) macroscopic bloating. (Freestone et al., 1987: 22)

v1.c26.s6.t1.13

Examination of the vitrification textures of the Zawar materials and comparison with refired examples shows that temperatures of C. 1200-50°C were attained during smelting. The furnace walls shows a gradient in vitrification structure from the high temperature, highly vitrified interior to the feebly vitrified exterior. The temperature gradient depends on the duration of the smelt as well as the temperature achieved in the furnace. Comparing the micromorphologies of the furnace wall at series of points with those of a series of refired examples, and using a simple heat-conduction equation, the time for which a furnace was at high temperature may be estimated. For the zinc smelting process at Zawar, duration of around 5 hr has been estimated. (Tite et al., 1982, 1985; Freestone et al., 1987)

v1.c26.s6.t1.14

In order to obtain some idea of the bulk composition of the retort, a thin slice throguh the body or the wall was crushed and analysed by atomic absorption spectrometry. (AAS: Hughes et al., 1976) The data (Table 18.1) were compared with the electron probe micro-analysis (EPMA; Freestone 1982) data for the clay matrix. The higher SiO₂ content in the retort body (62.49%) compared to that in the clay matrix (53.82%; vide Table 18.1) may be explained in terms of inclusion of large quartzite grains, as earlier observed by Mookherjee (1964), which is naturally reflected in the AAS analysis. Brooke reported in 1850 that the Rana's experiments in the previous year for smelting zinc had resulted in breaking of the crucibles because, according to an old man, the quartz gangue had not been removed from the ore prior to smelting. (Brooke, 1850: 214) We feel however, that the crucible was not properly made and strengthened by inclusion of quartzite and phyllite, the local rocks, as used to be done in the earlier centuries, and hence the failure of the Rana’s experiment.

v1.c26.s6.t1.15

Zawar Zinc Retort Wall: Analysis of Ceramic Materials

v1.c26.s6.t1.16

| | (1) | (2) | (3) | (4) |
| SiO₂ | 62.49 | 53.82 | 64.85 | 57.96 |
| TiO₂ | 0.71 | 0.27 | 0.44 | 0.59 |
| Al₂O₃ | 13.90 | 17.49 | 13.75 | 13.97 |
| FeO | 2.48 | 4.16 | 7.49 | 15.39 |
| MnO | 0.04 | b.d | b.d. | b.d. |
| MgO | 1.44 | 1.29 | 0.93 | 1.35 |
| CaO | 0.64 | 0.34 | 0.39 | 0.70 |
| Na₂O | 0.94 | 1.81 | 0.67 | 0.55 |
| K₂O | 3.08 | 3.11 | 9.90 | 7.54 |
| P₂O₅ | n.a. | 0.50 | 0.28 | b.d. |
| Cl | n.a. | 0.16 | b.d. | b.d. |
| S | n.a. | b.d. | b.d. | b.d. |
| ZnO | 2.18 | 3.22 | 0.61 | 0.61 |
| TOTAL | 87.90 | 86.17 | 99.11 | 98.66 |

v1.c26.s6.t1.17

(1) Bulk analysis of retort body (crushed slice) by atomic absorption spectrometry; (2) Clay analysis by EPMA or electron-probe micro-analysis; (3) and (4) EPMA analysis of partial melt (glass) from outer surface of retort.

v1.c26.s6.t1.18

n.a. = not analysed; b.d. = below detection limits

v1.c26.s6.t1.19

The glassy layer on the external surface of the used retort contained higher proportion of FeO and K₂O due to a partial melting of the ceramic which filled the pores and prevented the escape of vapours containing zinc. Some amount of zinc was retained in the retort body, a much smaller proportion appearing in the outer glassy layer (vide Table 18.1). The liquidus temperatures of the K₂O — FeO — Al₂O₃ — SiO₂ glasses on the exterior of the retort body are likely to be of the order of 1100-1125°C (Freestone and Powell, 1983) which are just sufficient for reduction of zinc oxide to the metal. (Smith, 1918; Freestone et al., 1985a : 238)

Cites (2)