Copyright © 1980 Ohio Acad. Sci. 0030-0950/80/0002-0052 $1.50/0

PRELIMINARY ARCHAEOLOGICAL EXAMINATION OF OHIO'SFIRST BLAST FURNACE: THE EATON (HOPEWELL)1

JOHN R. WHITE, Department of Sociology and Anthropology, Youngstown State University,Youngs town, OH 44555

Abstract. The Eaton (Hopewell) Furnace located near Struthers, Ohio was built in1802-1803. The first blast furnace west of the Alleghenies and the first industry ofany kind in the Western Reserve, it went out of blast circa 1808 due to a combinationof factors and fell into ruin. Historical sources on the Eaton are scarce and informa-tional sources are vague, but archaeological excavations carried out in 1975, 1976,and 1977 have led to some interesting findings concerning early blast furnace opera-tions. Subsequent chemical and metallurgical analyses of furnace artifacts and speci-mens provided insights into the level of efficiency of the operation and the quality ofthe raw materials, products, and byproducts. Foremost among these findings is thefact that the Eaton's use of bituminous coal in combination with charcoal was theearliest use substantiated in the New World.

OHIO J. SCI. 80(2): 52, 1980

The Eaton-Hopewell Furnace (33MH9)is located in Yellow Creek Gorge just200 m downstream from manmade LakeHamilton in Mahoning County, betweenthe cities of Struthers and Poland, twosuburbs of Youngstown, Ohio. It liesmidway up a steep slope with an inclinein excess of 45 degrees. The slope soil isclassified as Dekalb very stony loam, 25%to 50% slopes (DkF), characteristic ofvery steep valleys in Mahoning County(Lessig et al 1971, p. 78).

The furnace, built in 1802-1803, wasthe earliest blast furnace west of theAlleghenies and the earliest industry ofany kind in the Western Reserve. Itoperated with only one major interrup-tion until about 1808, when, due to acombination of factors including an in-efficient blast process, a shortage ofreadily available hardwood for charcoal,and an accidental blow-out, it went outof blast. Very little is known about theoperation years. What few accounts Ihave found written in local histories(Butler 1921, p. 658) are somewhat re-petitive (even in their errors), suggestingthat they were gleaned from the sameprimary and insubstantial source.

Prior to excavation, very little of thefurnace was observable and, like the re-

JManuscript received 30 January 1979 (#79-8).

mainder of the site, was either destroyedor buried under 175 years of erosionaloverburden. Only the tuyere arch anda small 1.75 m rim segment of the innerchimney of refractory sandstone werevisible. Evidence, including old photo-graphs, indicated that little more of thefurnace than this was exposed for at leastthe last 75 years. The cover vegetationwas so dense with elm, sycamore, wildgrape, sumac, and poison ivy that ittook 4 full days just to clear the area forgridding.

Excavations covering 3 seasons werecarried out by a crew consisting of 15Struther's High School seniors, 5 uni-versity archaeology students and recentgraduates, and a more or less steady sup-ply of university volunteers. The sitewas divided into 7 major excavationzones, 4 of which were of prime im-portance. This division served 2 pur-poses: it allowed for simultaneous sampl-ing and investigation in different areasof the site, and it facilitated deploymentof the field crew over the relatively re-stricted and precarious land area.

FURNACE ZONEThe Furnace zone consisted of the

furnace structure itself and the fill withinits perimeters. Overburden often reacheda depth of more than 2 m. Because of

52

Ohio J. Sci. OHIO'S FIRST BLAST FURNACE 53

the striking visual remains and great lo-cal interest, continuous efforts were madeto clear this zone. The end of the firstseason saw almost the entire furnace

structure uncovered to its full remainingsize. Excavation revealed the remainsof the entire bosh area and firepot orhearth, measuring 180 cm in height from

TIPPLE AREAi

SANDSTONEBLOCKS

JHARCOA

REFRACTORYSANDSTONE

OR BRICK

AND ISULATIONLAYER

DAM STONE IRONFIGURE 1. Schematic drawing of a typical early 19th century blast furnace.

54 JOHN R. WHITE Vol. 80

base to bosh and varying in width be-tween 1.7 m and 1.9 m. The bosh, whichis the widest point of the furnace and thestructural feature that allows for the sup-port of the charge, measured a maximumof 2.7 m in diameter. (See fig. 1 for aschematic of a typical early blast fur-nace.) Based on these specifications, anestimated production rate of 2 to 3 tonsof cast iron per day seems reasonable.

The inner chimney lining was madefrom shaped blocks of refractory sand-stone that were cemented with a hardmortar colored to a brick-red. Exami-nation of this mortar showed a con-stituency quite similar to that of thesamples of sand and soil analyzed fromthe site (table 1). In a laboratory ex-periment, red sand from the hearthopening (Sample 2) was mixed with waterand baked. This process resulted in thecreation of a friable concretion not ashard as but not very different from thecolor and texture of our mortar sample.This finding leads us to believe that thefurnace builders expeditiously weldedtheir sandstone blocks together withmortar made from local sand or mud andwater. The intense heat and pressureproduced by the furnace operation hard-ened and colored the mortar. Thechimney interface was patinated with athick incrustation of slag, and the taphole itself was clotted with the remains

of the furnace's last cast, a part of whichupon cessation of the blast had been al-lowed to cool within the firepot. Re-mains of the cast, which consisted of apudding-like conglomeration of charcoal,slag, ore, and iron, were discovered inthe form of a long runner extendingfrom the tap hole some 4 m out onto thecasting floor. The runner was coveredwith 100 cm of erosional soil. This find-ing suggests that the incompletely cookedcast erupted onto the casting floor as aresult of a furnace lining failure.

FURNACE WALLThe outer furnace wall was composed

of large hand-chiseled blocks of nativesandstone, some weighing several hundredpounds. The furnace was built into thegorge slope with the natural sandstonecliff constituting an integral part of itsconstruction. The insulating space be-tween the heavy outer wall and the re-fractory inner chimney was filled withsand that was subsequently oxidized to abright red-orange by the intense heatgenerated by the furnace. This sand,though dramatically different in color,proved to be quantitatively and quali-tatively identical to samples taken fromother areas of the site (table 1). Thesesands were local in origin and resultedfrom weathering and erosion of the con-tiguous sandstone cliff. When the fur-

TABLE 1

Dry sample analysis of Eaton-Hopewell sands and mortar {percent by weight).

CONSTITUENTS

Fe2OsSiO2

AhOsCaOM g OM n OTiO2

Na2OLoss on IgnitionMoisture

1.

yellow sandouterbrick

chimneyon south

arc offurnace

0 . 584.0

7 . 00 . 30 . 30 . 10 . 51.21.00 . 3

2.red sandhearth

opening

1.781.0

6 . 80 . 30 . 30 . 10 . 50 . 41.80 . 7

Sami

3.grey/black sand

hearthopening

below redsandlevel

1.184.0

5 .90 . 20 . 20 . 10 . 30 . 63 .21.2

>les and Sites

4.red sand

outerbrick

chimneyon south

arc offurnace

.0881.2

7 . 20 . 30 . 30 . 10 . 71.22 . 01.2

5.grey sand

overburdenfrom areaof hearthopening

.0881.2

7 . 60 . 20 . 20 . 10 . 61.01.79 . 6

6.grey sand

castingfloorarea

Zone Aunit

N2/W6

1.083.0

6 . 20 . 20 . 30 . 10 . 40 . 83 . 47 . 8

mortarbetween

refractorysandstoneblocks of

innerchimney

1.886.0

8 . 00 . 30 . 60 . 10 . 50 . 60 . 5

Ohio J. Sci. OHIO'S FIRST BLAST FURNACE

nace subsequently collapsed, the oxidizedsand spilled out over the immediate cast-ing floor area and served as a clear reddemarcation between the lower culturallevels on which it sits and the post-1808sandy loam overburden on top of it.

CASTING FLOOR (ZONE A)Zone A was the designation given to

the relatively flat and featureless areaimmediately adjacent to and south ofthe furnace. It was determined to be

the casting floor, the place where the rawmolten iron or loupe was set into variousmolds. This zone represented the areaof maximum furnace activity. The en-tire zone measured approximately 14 x 18m in area and was cleared to a depth ofbetween 10 and 100 cm revealing a flatsandy floor. The top several centi-meters contained a heavy concentrationof rock spalls and fragments. Withinthis spall level were found several arti-facts shaped like large, thick dog biscuits

d

cm

FIGURE 2. Hydraulic mortar bricquets recovered from the site indicate post-1907 activity.

56 JOHN R. WHITE Vol. 80

(fig. 2). These artifacts were determinedto be bricquets used for testing thetensility of hydraulic mortors (AmericanSociety for Testing and Materials 1963,p. 505). These devices were used byengineers during the construction ofHamilton Dam and hence were idealdating indicators for the spall level asmodern or post-1907. The spalls them-selves could well represent the result ofenergies put to the finish-shaping of theenormous sandstone blocks used in thedam. While no evidence was found ofa casting floor shed in this area, such astructure would have been a necessitybecause even the smallest amount ofprecipitation coming into contact withthe molten iron would cause a violentreaction.

Most of the artifacts that were foundcame from Zone A. These includedstove parts, fireback fragments, frag-ments of assorted heavy iron tools,spikes, heavy pins, staples, utensil piecessuch as Dutch ovens, trivets, and pans;and byproducts of the manufacturingprocess including sprues and scrap iron.Some artifacts were so encrusted withrust that identification was impossible.Several artifacts were made from wroughtiron and must have been brought to thesite from elsewhere because there is noevidence, historical or archaeological, ofthe Eaton's having a forge or facility fortheir production. One of these wroughtiron artifacts was a large staple that mayhave been part of the casting shed. Themore ubiquitous items included chunks ofslag and kidney ore. Two basic methodswere used to clean these artifacts: (1)more delicate pieces were cleaned byelectrolysis; (2) the hardier, bulkier arti-facts were muffled and sandblasted. Thisis a new, efficient and far speedier process(White 1976).

Less than 30 cm below the sand level,we found an extremely rocky talus levelcomposed of huge slabs and blocks ofsandstone that had, through the cen-turies, detached themselves from thecliff side. Apparently, the furnace build-ers created the necessary flat castingfloor by filling and covering the talus in-terstices with sand until a level workingsurface was achieved. Immediatelysoutheast of the flat casting floor was a

massive slag heap that sloped steeplyfrom the perimeter of the casting floor tothe creek level l l m below. A cut madeinto this slag heap at its top revealedevidence that the furnace had undergonesome major repair and relining during itsuse. Relining was done in a piecemealor patchwork manner rather than by acomplete overhauling. Fragments of dis-carded refractory sandstone were foundsandwiched between layers of slag andcinder. The slag heap probably servedas the general disposal area for all dis-carded material.

OTHER ZONESZone B was the designation given to a

small terrace roughly 6 x 6 m in arealocated about 8 m downslope east fromthe furnace mouth. It was here that theoriginal mechanism for supplying theblast was located. Excavations revealeda stone wall or footing and an almostsquare, flat sandstone slab floor measur-ing 160 x 158 cm along its sides. Thisstructure was all that remained of theblowing shed or wheelhouse. The over-shot wheel turned in the area between thewalls and raised and lowered the bellowsthat rested atop the square slab floor.

The least investigated of the majorexcavation areas was Zone C. This isthe designation given to the tipple areaapproximately 10 m up the cliff facefrom the bosh. Work was undertakenhere to determine what we could aboutthe charging process. We found morefragments of kidney ore, charcoal, andcoal in this area than we had anywhereelse on the site. This is as it should be,since it was from this spot that thecharges of fuel, iron ore, and flux (lime-stone) were supplied to the furnace.The abundance of high quality bitumi-nous coal in this zone led us to believethat the Eaton (Hopewell) used a com-bination of charcoal and raw coal asfuel. Chemical analysis of the cast ironsby scientists at Youngstown Sheet andTube company and other laboratoriessupported this conclusion because thefinished iron contained larger amounts ofsulfur than expected with simple char-coal reduction, i.e., between 0.060% and0.22% by weight (White 1978). Thisfact tends to reduce the effectiveness of

Ohio J. Sci. OHIO'S FIRST BLAST FURNACE 57

those arguments by historians and ar-chaeologists who proclaim that archaeol-ogy cannot add anything of a hardfactual nature to our knowledge of ourhistorical past. No extant record of theEaton (Hopewell) mentions this use ofcharcoal in combination with raw coal;in fact, coal used in this manner was theearliest reported for the New World(White 197S).

ANALYSESThe Eaton (Hopewell) coal was an-

alyzed and found to be of a relativelyefficient, high grade, and bituminoustype, averaging out as 4.06% ash, 38.04%volatile material, and 52% sulfur. Bycomparison, the Eaton (Hopewell) char-coal ranged between 1.94% and 7.30% inash content, between 40.02% and 44.58%in volatile material, and between 0.01%to 0.02% in sulfur content.

Eaton (Hopewell) iron ore is a varietyknown as kidney or reniform ore, gettingits name from its physical characteristics,i.e., it is generally kidney-shaped andreddish brown in color. Extremely dense,it occurs in both pockets or layers and asfloat material in Yellow Creek. Whilehigher quality iron ore exists (today themills use taconite pellets with 60-65%iron), the Eaton (Hopewell), with itsferrous oxide (Fe2O3) between 47.9% and58.6%, is considered good especially forits time.

The most ubiquitous byproduct of theironmaking process is slag, which is thelithic material created when the lime-stone flux mixes with impurities or non-ferrous materials in the iron ore. Thisslag, less dense than the molten iron,floats on the iron and is skimmed off anddiscarded. Old ironmasters had a maxim:"Take care of the slag and the steel willtake care of itself," which exemplified theimportance of this phase in the operation.Low viscosity and high sulfur-removingcapacity are the prime characteristics ofblast furnace slags (Muan and Osborn1964, p. 148). The Eaton (Hopewell)were relatively consistent in this desul-furizing ability, having between 0.30%and 0.50% sulfur. As is characteristicof blast furnace slags, the thermo-dynamically stable oxides of magnesium

(MgO), aluminum (A12O3), silica (SiO2),and calcium (CaO) were present. Slagcolors included blue, green, glassy black,and turquoise. The content figures wereroughly comparable for each of the dif-ferent colors. Analysis of Eaton (Hope-well) slags indicated that the furnace wasrun at a temperature between 2150 °Fand 2250 °F (1177-1232 °C) (White1977). As a point of comparison, mod-ern furnaces work at approximately2800 °F (1538 °C). Earlier blast fur-naces did not get respectively cooler asthere is a minimum effective temperaturefor ironmaking of 2200 °F (1204 °C).

We have learned a considerable amountabout early iron production in an areathat prides itself as one of the world'slargest steel producing centers, but thisis only a preliminary report and manyquestions still remain. Classification andanalysis of the artifacts and specimensfrom the site is almost completed, as isthe final site report containing morecomplete descriptions of site settings,geology and archaeology. Undoubtedly,publication of our final monograph willserve to raise even more questions, butarchaeology can answer some hard ques-tions about those periods and aspects ofour historical past traditionally neglectedby the record keeper and the historian.

Acknowledgments. I am indebted to Youngs-town Sheet and Tube company for its contribu-tions of time, equipment and expertise. Thehighest thanks must be reserved for FrankGalletta, research engineer, without whommuch of the analyses could not have been done.Finanacial support for the project came fromthe Gund Foundation, the Youngstown StateUniversity Graduate Research Council, andthe Struthers Board of Education.

LITERATURE CITEDAmerican Society for Testing and Materials

1963 Standard method of test for tensilestrength of hydraulic cement mortars.ASTM Standards in Building Codes, pp. 505-510. American Society for Testing and Ma-terials, Philadelphia, PA.

Butler, Joseph G. 1921 History of Youngs-town and the Mahoning Valley, Ohio. Ameri-can Historical Society, New York. 844 pp.

Lessig, Hever, W. F. Hale, P. W. Reese, andG. J. Post 1971 Soil Survey of MahoningCounty, Ohio. U.S.G.P.O. Washington D.C.

Muan, Arnulf and E. F. Osborn 1964 PhaseEquilibria Among Oxides in Steelmaking.Addison-Wesley, Reading, MA.

58 JOHN R. WHITE Vol. 80

Swank, James M. 1892 History of the Manu-facture of Iron in All Ages and Particularly inthe United States from Colonial Times to1891. American Iron and Steel Assoc,Philadelphia, PA.

White, Jr. R. 1976 Cleaning heavily encrustediron artifacts by calcination and inducedspalling. Curator 19: 311-315.

1977 X-Ray fluorescent analysis cf anEarly Ohio blast furnace slag. Ohio J. Sci.77: 186-188.

1978 Archaeological and chemicalevidence for the earliest American use of rawcoal as a fuel in ironmaking. J. Archaeolog.Sci. 5: 391-393.