physical and mineralogical properties of experimentally heated chaff-tempered mud bricks:...

Journal of Archaeological Science: Reports 2 (2015) 80–93

Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports

j ourna l homepage: ht tp : / /ees.e lsev ie r .com/ jas rep

Physical and mineralogical properties of experimentally heatedchaff-tempered mud bricks: Implications for reconstruction ofenvironmental factors influencing the appearance of mud bricks inarchaeological conflagration events

Mathilde C.L. Forget a,⁎, Lior Regev b, David E. Friesem c, Ruth Shahack-Gross a,⁎a Kimmel Center for Archaeological Science, Weizmann Institute of Science, Rehovot 76100, Israelb Weizmann-Max Planck Center for Integrative Archaeology, Weizmann Institute of Science, Rehovot 76100, Israelc Institute for Archaeological Sciences, University of Tübingen, Rümelinstr. 23, 72070 Tübingen, Germany

⁎ Corresponding authors.E-mail addresses: [email protected] (M

[email protected] (R. Shahack-Gross).

http://dx.doi.org/10.1016/j.jasrep.2015.01.0082352-409X/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t
a r t i c l e i n f o
Article history:Received 10 October 2014Received in revised form 6 January 2015Accepted 9 January 2015Available online 26 January 2015

Keywords:Mud bricksHeatTemperatureCompositionDurationFiring atmosphereInfrared spectroscopy

Sun-dried mud bricks are used around the world and have been found in the Levant and Mesopotamia since theNeolithic period. Their form and composition lend important information pertaining to social and technologicalmeaning in human cultures. Fired mud bricks are well known in the southern Levant, often identified in Bronzeand Iron Age strata and used as a marker for destruction by fire (conflagration events). Only a few studies haveattempted to reconstruct conflagration conditions from fired mud bricks becausemany variables impact the for-mation of the final fired brick. These include brick composition, heat intensity (i.e., maximum temperature), heatduration and firing atmosphere (i.e., oxidizing vs. reducing). Themyriad combinations of these factorsmay resultin different appearance of fired bricks. Infrared spectroscopy is one method that has been exploited quite exten-sively in relation to fired clay-based materials: studies were conducted on powdered sediment samples for afixed duration and in oxidizing conditions, producing calibration curves that were then utilized for reconstruc-tion of past maximal heat. Here we report on an experimental study of the thermal behavior of mud bricksunder differing composition, heat intensity, heat duration and firing atmosphere. We carried out experimentsin a furnace oven using micro-thermocouples which allowed us to simultaneously measure heat across bricks,from edge to core. The resultingmud brickswere analyzed using Fourier Transform Infrared (FTIR) spectroscopy.We identify a previously unknown thermal effect that occurs in bricks temperedwith organicmaterialwhile theyarefired; namely a correlation between the amount of organic temper and elevation of temperatures up to 100 °Cabove the oven chamber temperature. We record the color patterns obtained at different temperatures and du-ration of heating, as well as the colors obtained from heating in different atmospheres. We report that the FTIRspectrum of bricks heated in oxidizing conditions differs from that of bricks heated in reducing conditions atthe same temperature. We note that the position of the main clay absorbance band cannot be used alone toinfer firing temperature as its shift is not systematic. We show that combining this parameter with the widthof the same band in the FTIR spectrum makes it possible to achieve better temperature reconstructions fromfired bricks. Lastly, we report a small scale case study inwhichwe tested the applicability of the experimental re-sults to the remains of amud brickwall unearthedwithin the largest known destruction event in the ancient cityof Megiddo, i.e., Tel Megiddo Stratum VIA. We show that this wall was burnt as one unit, having a reduced coreand oxidized outer part, where the core experienced temperatures in the range of 500–600 °C and the edge600–700 °C. The detailed analysis of brick compositions carried out in this study further allows us to reconstructthe ancient bricks' preparation recipe. The results of this study bear important implications for future studies ofarchaeological conflagration events, and the destruction phenomenon in general.

© 2015 Elsevier Ltd. All rights reserved.

.C.L. Forget),

1. Introduction

Mudbricks are commonbuildingmaterials used since theNeolithic inEurasia. These artificially shaped andmanipulated sediment blocks hold awealth of information that can be retrieved through archaeological re-search. First and foremost, identification of intact mud brick walls

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jasrep.2015.01.008&domain=pdf

http://dx.doi.org/10.1016/j.jasrep.2015.01.008

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.jasrep.2015.01.008

http://www.sciencedirect.com/science/journal/

http://ees.elsevier.com/jasrep

81M.C.L. Forget et al. / Journal of Archaeological Science: Reports 2 (2015) 80–93

makes it possible to reconstruct architectural units and thus site structure(or city/town plan; e.g., at Çatalhöyük, Mellart, 1975). Having a clear planof urban spacemay further be used to infer social, political, economic andgender-related issues. More information can be retrieved from the bricksthemselves, mostly through the study of their mineralogical and elemen-tal compositions. By so doing researchers have been able to identify thesources for brick material across the landscape in relation to thestudied sites, which further informs about human use of the land-scape, soil/sediment procurement and its influence on human–environment interactions (e.g., Goldberg, 1979; Nodarou et al.,2008). Identifying the source materials also enables discussion of orga-nization of labor and construction methods (e.g., Homsher, 2012;Emery & Morgenstein, 2007). Compositional similarities and differ-ences among bricks have also been used to argue that mud bricksare useful for the identification of production choices, which inturn reflect cultural and ideological choices, and even act as meansfor symbolic communication (Love, 2012, 2013). Recently it wasalso demonstrated that the study of the degradation products ofmud bricks provides important methodological insight which pertainsto the formation of stratigraphy and microstratigraphy, the identifica-tion of roofs and floors, and reconstruction of activity areas (Friesemet al., 2011; 2014a; 2014b).

Many of the studies related to mud bricks deal mostly with sun-dried bricks. Excavators over the years also noted the presence of firedmud bricks, yet relatively few studies explored these in detail. In mostcases the presence of fired bricks has been used to infer conflagrationevents. Differences in color of fired mud bricks have been noted bymany, however only a few researchers designed experiments aimingto understand the meaning of such color differences (e.g., Stevanović,1997). Experimental results indicated that the effect of fire on mudbrick color depends on many variables, including differences in brickcomposition, heat intensity, heat duration, and availability of oxygen(e.g., Maritan et al., 2006; Twiss et al., 2008; Love, 2012). It appearsthat this experimental complexity hampered continued researchinto burnt mud bricks. However, extracting information from burntmud bricks on heat intensity, duration and atmosphere is expectedto open up a window into very important social, cultural and historicalquestions. It may be useful for reconstruction of heat regime duringconflagration events which may allow for differentiation betweenintentional and accidental fire. It may be helpful in distinguishing be-tween types of destruction events, which in turn may affect historicalinterpretation. It is therefore important to study the effect of heat onmud bricks.

Clay minerals are a basic component in mud bricks. As in firing ofpottery, fired bricks also become consolidated and hardened, whichmakes fired bricks preserve better than unfired bricks. Clay mineralsare known from previous studies to be effective recorders of firingtemperature. For example, in heat-related mud-based industrial/domestic installations such as ovens, furnaces, andwheremetallurgicalactivities took place, researchers were able to reconstruct use tempera-tures through infrared spectroscopy (e.g., Eliyahu-Behar et al., 2012;Gur-Arieh et al., 2013). The same approachwas also deployed in potterystudies (e.g., Shoval et al., 2011). These studies showed that infraredspectroscopy enables to distinguish between heated and unheatedclay, and even determine rough estimates of temperature ranges(Berna et al., 2007; Eliyahu-Behar et al., 2012; Friesem et al., 2014a;Gur-Arieh et al., 2013). We note that thermoluminescence is also atechnique suitable for reconstruction of temperature from heatedclay-based materials, but relative to infrared spectroscopy it is moretime consuming and expensive. It is reasonable to assume that infraredspectroscopy will be useful in determining firing temperature of mudbricks, yet previous studies utilizing infrared spectroscopy did not testthe effects of heat intensity, duration and atmosphere on clay-basedmaterials.

The overall aim of the current study is to build a methodologicalframework that will allow reconstruction of the parameters that

result in what we identify in excavations as bricks of varying colors.We will test: (1) the thermal behavior of bricks as they are heatedunder controlled laboratory conditions, with varying compositions,heat intensity, heat duration and availability of oxygen, and (2) theeffect of the above changing factors on brick color and its clay infraredspectrum.

1.1. Background to the study of mud bricks in conflagration events in thesouthern Levant

Destructive conflagration events are documented in the archaeolog-ical record mainly since the Neolithic (Stevanović, 1997; Twiss et al.,2008). In the southern Levant, the number and magnitude of destruc-tions by fire are pronounced in the Bronze and Iron Ages, ca. 3rd to1st millennia BCE. Some of the massive destruction events have beenused as chronological anchors across the region and serve for interpre-tations of the history of the southern Levant, mostly within the realm ofbiblical archaeology. Despite the abundance of reported destructionlevels in many sites in the southern Levant, there is relatively littledeliberation on methodological aspects of this phenomenon (Dever,1992; Finkelstein, 2009).

Field identification of destruction events in Levantine Bronze andIron Age sites relies primarily on macroscopic criteria such as presenceof smashed vessels (rather than pottery sherds) on floors covered byash, sometimes including charred materials, and sometimes overlainby accumulation of stone and/or mud brick collapse (Finkelstein,2009). Differences in colors of mud bricks are often interpreted asreflecting exposure to different heat regimes, though exact tempera-tures cannot be determined based on color alone.

Utilization of microscopic, molecular and elemental methods to thestudy of heat intensity in Levantine Bronze and Iron Age destructioneventswas thus far conducted in two Iron Age contexts, employing Fou-rier Transform Infrared (FTIR) spectroscopy. The effect of heat on theFTIR spectrum of clay minerals has been extensively explored (Farmer,1974; Shoval et al., 1991, 2011; Berna et al., 2007; Savage et al., 2008;Clegg et al., 2012). It has been shown that reversible and irreversiblestructural changeswith heat occur in clayminerals, producing significantspectral changes that allow identification of heating temperature. Thesespectral changes are stable for long periods of time as shown by Shovalet al. (1991; 2011) who found that re-hydroxylation in ca. 3000 yearsold fired pottery is only partial (apparent by the presence of an OHband at 3620 cm−1, while the OH band at 3690 cm−1 is absent). Similarobservations were made by Berna et al. (2007), Eliyahu-Behar et al.(2012), Gur-Arieh et al. (2013, 2014) and Friesem et al. (2014a) whoworked in various locations in Israel, Uzbekistan and Greece, wheredifferent types of soils and sediments occur. They all point to the samespectral changes in heated clay minerals, and show that these spectralchanges are preserved for thousands of years. It is therefore concludedthat despite partial rehydroxylation, it is possible to use FTIR spectra ofclay minerals in order to deduce past firing temperatures.

In a studyof sediments associatedwith destruction, Berna et al. (2007)show a temperature gradient (ranging between 1000 and 500 °C)associated with a mass of collapsed burnt mud bricks in Tel Dor, dat-ing to the late Iron Age I. They studied this destruction event two-dimensionally, on an excavation section. Their study relied heavily onthe infrared analysis of clay minerals and included an experimental cal-ibration of the effect of heat on the infrared spectrum of clay minerals.They found that the highest recorded temperatures within the brickpile – in excess of 1000 °C – corresponded to yellowish-white sedimentsin the center of the pile, and that lower temperatures – ranging between500 and 1000 °C – corresponded to red, brown and gray colored sedi-ments (Fig. 8 in Berna et al., 2007). Namdar et al. (2011) studied anIron IIA destruction level at Tell es-Safi/Gath where they conducted de-tailed mapping of mineralogical and heat intensity changes acrossspace, in three dimensions in one excavation square (5 × 5 m). Theyidentified that the most intense heat was associated with the roof

82 M.C.L. Forget et al. / Journal of Archaeological Science: Reports 2 (2015) 80–93

of a collapsed structure, and that the beaten-earth floor was not ex-posed to high temperatures. In addition, they found that heat inten-sity across four large red colored mud bricks – with one of themhaving a black core – was homogenous (at ca. 500–600 °C), whichled them to suggest that mud bricks used for building were pre-fired in a kiln — a most unusual observation for this time period in theLevant.

In both studies cited above, the experimental calibration of heatintensity in relation to the corresponding infrared spectra of clay min-erals was done in oxidizing conditions. Here we will present a studythat experimentally calibrated the relationship between heat intensityand infrared spectra of clay minerals in both oxidizing and reducingconditions. This will allow us to match archaeological brick color withthe most appropriate firing atmosphere and thus most appropriate in-frared signal. In addition, in the studies reported above the calibrationswere produced by heating powdered sediment samples for the durationof 4 h. Herewe present experiments conducted on small butwholemudbricks. Our experiments take into consideration the effects of heatintensity, heat duration and amount of organic temper on the colorpattern of bricks, as well as the resultant infrared signals.

Wewill test our experimental results by presenting a small case studyof an intact mud brick wall segment unearthed in a major destructionevent believed to be associated with fierce fire, at Tel Megiddo (Israel).

Fig. 1. (a) Aerial photograph showing the location of Tel Megiddo in the southern Levant. (b) Scwall fragments and singular reddish-colored bricks unearthed in 2012 at the southern part of Awere obtained across this freshly broken brick profile, all utilized for FTIR spectroscopic analysisreferences to color in this figure legend, the reader is referred to the web version of this article

1.2. Background to the archaeological case study: An intact mud brick wallsegment in Stratum VIA at Tel Megiddo, Israel

Tel Megiddo, a mound site, is located in northern Israel (Fig. 1a).It has been excavated by Schumacher (1903–1905), The Universityof Chicago's Oriental Institute (Fisher, Guy and Loud 1925–1939),the Hebrew University of Jerusalem (Y. Yadin; 1960s and 1970s, in-termittently), and currently by an international expedition led bythe Institute of Archaeology in Tel-Aviv University (1992–present,by I. Finkelstein, D. Ussishkin and E. Cline) (Schumacher, 1908;Lamon & Shipton, 1939; Loud, 1948; Kempinski, 1989; Harrison,2004; Zarzecki-Peleg, 2005; Finkelstein et al., 2000; 2006; 2013). Itis a major biblical city that features multiple layers dating betweenthe 4th and 1stmillennia BCE. The early excavations atMegiddo focusedon large areas and large architectural structures. These vast excavationscontributed significantly to the establishment of the basic chronologicalsequence of the Bronze and Iron Ages in Israel and the Levant. The cityalso features several destruction events (Finkelstein & Piasetzky, 2009).

The most devastated level at the site is assigned to Stratum VIA,dated to the late Iron Age I (late 11th–early 10th centuries BCE;Boaretto, 2006). Evidence for destruction by firewas identified by all ex-cavators of the site, in almost all excavation areas (Finkelstein, 2009;Cline, 2011; and references therein). A famous marker of this stratum

hematic drawing showing TelMegiddo and highlighting excavation Area Q. (c) Collapse ofrea Q. (d) A longitudinally broken mud brick from Area Q and Level Q-7. About 10 samplesand served to determine preliminary past heating temperature. (For interpretation of the.)


is the red color of its bricks, interpreted to be a result of the fierce firethat annihilated the city and gave it its nickname “the red brick city”.The destruction includes public as well as domestic quarters of thecity and the debris is characterized, in all excavation areas, by thefollowing stratigraphic sequence: floor with smashed pottery, bones,metal objects and in some places human skeletons overlain by charredwooden posts, wood ash and thick (up to 1m) collapse of red andwhitebricks (withwhite bricks assumed to have been burnt at extremely hightemperatures). A short occupational gap seems to have followed thedestruction, based on the fact that the next phase of settlement at thesite (Stratum VB) is different in both the ceramic tradition and in thelayout of the city (Finkelstein, 2009; Harrison, 2004; and referencestherein). Past research of materials unearthed from this destructionlevel was macroscopic in nature. A pilot study conducted by us in2012–13 focused on mud bricks from excavation Area Q (Fig. 1b). Thisarea features primarily domestic archaeological remains dating, basedon pottery typology, to the Iron IIA (Levels Q-6–Q-4) and late Iron I(Level Q-7). Level Q-7, exposed at the end of the 2014 season in 7 exca-vation squares (a total of 175 m2), represents the Stratum VIA destruc-tion event based on its stratigraphic position and ceramic evidence, aswell as the presence of large amounts of red-colored mud bricks. Thelatter marks the top part of the Stratum VIA destruction layer (Fig. 1c).

At first sight all debris piles seem to be composed of reddish-coloredmud bricks. A closer look shows that color differences do occur. Westarted the investigation of the debris piles in 2012, utilizing FTIR spec-troscopy and conducting interpretations based on the method of cali-bration developed by Berna et al. (2007), i.e., powdered sedimentheated in an oven at oxidizing conditions to different temperatures forthe duration of 4 h. More than 130 samples have been studied fromN50 mud bricks from Level Q-7. The majority of bricks are orange toreddish-brown in color with no visual differences between the core andedge (Fig. 1d). Our initial results indicated a surprising homogeneity inthe infrared spectra, regardless of color differences among bricks. Theinfrared spectrum of the clay fraction indicates heating at 500–600 °C,especially based on the position of the main clay absorbance band atca. 1040–1045 cm−1 (Fig. 2). The infrared spectra of mud bricks fromLevel Q-7 also show presence of carbonated hydroxylapatite, a phos-phate mineral, based on the absorbance bands at ca. 565 and605 cm−1. In addition, the phytolith concentration in Level Q-7 mudbricks is rather unexpectedly high, being 23± 8million in 1 g sediment(cf., Friesemet al., 2011; 2014a showing less than 1million phytoliths inpre-modern mud bricks, as well as previous studies that used mudbricks as reference sediment in geoarchaeological studies of Bronzeand Iron Age sites in the Levant which also recorded very low concen-trations of phytoliths, e.g., Shahack-Gross et al., 2005; Albert et al.,2008; Shahack-Gross et al., 2009). The seemingly homogenous infraredsignal regardless of brick color may be misleading if interpretations are

Fig. 2.A representative FTIR spectrum from amud brick from Tel Megiddo, Level Q-7. Note the aabove ca. 500 °C. Calcite (CA) appears as a major component, while minor components include

based on calibration conducted under oxidizing conditions only and inconstant heat duration. Moreover, the presence of phosphate com-pounds and high concentrations of phytoliths indicate that the Q-7mud bricks included high amounts of vegetal temper (also reflected inhigh density of elongated voids after burnt vegetal matter). These pre-liminary observations thus call for a systematic study of the infraredchanges that occur in heated mud bricks, considering the presence ofvegetal temper in the brick composition, and the interplay betweenheat intensity, heat duration, and availability of oxygen. Below wedescribe the experimental study conducted, and its application on anintact mud brick wall segment that exhibits a variety of brick colors,unearthed in Square I/2 during the 2014 excavation season at TelMegiddo (Fig. 3).

2. Materials and methods

In order to study the thermal behavior of mud bricks under differentlevels of heat intensity, duration, material composition and availabilityof oxygen, several sets of experimental mud bricks were prepared.

2.1. Mud brick preparation

Experimental mud bricks were prepared using cultivated floodplainsediment collected from a roadcut profile at ca. 50 cm below surfaceabout 1 km from the site, in the Jezreel Valley north of Tel Megiddo.The sediment was manually broken into aggregates smaller than5 mm and stones larger than 3 mm were removed. Chaff was preparedfrom dry wheat stalks collected after harvest from a field near Megiddousing a kitchen blender (without lubrication). Bricks were preparedmanually by mixing weighed portions of disaggregated sediment, chaffand water (about 0.4 ml per gram of sediment) until a homogeneouspaste was obtained. The paste was poured into steel molds measuring5.5 × 7.5 × 4.0 cm (Fig. 4a) followed by drying in an oven at 50 °C for aminimum of 5 days. The resulting bricks were used for the experimentsdescribed in Sections 2.3 through 2.6. After initial drying (one day, whenthe bricks are firm but soft) a metal wire of 1 mm diameter was pressedinto the bricks in order to produce elongated voids into which thermo-couples of this same diameter would be inserted when heating experi-ments were conducted (Fig. 4b). The voids were made longitudinallyfrom the edge to the center of the brick, at 20 mm depth which is halfof the bricks' thickness. Drying then continued in the oven at 50 °C andafter ca. 3–4 days the bricks were taken out and left to stabilize forabout two weeks at room temperature. After drying the volume of thebrick was reduced by about 18% (Fig. 4b). This preparation process pro-duced sets of replicable, rather homogenous, experimental mud bricks.

bsence of structural water (in the 3600 cm−1 region) indicating exposure to temperaturescarbonated hydroxylapatite (CHAP), opal (OP) and quartz (Q).

Fig. 3. (a) Aerial photograph showing the southern part of Area Q (looking south) whereremains dating to Level Q-7 are exposed. The red arrow indicates the location of the intactwall segment studied here (excavation Square I/2). The smaller, light arrows indicate thelocation of three other intact brick wall segments identified in this archaeological layer.One excavation square is 5 × 5 m. (b) An intact wall segment in Square I/2 as it appearsin the square's north section. Note 8 visible bricks and their variability of colors. Thewall segment overlies a mass of yellowish-gray sediment that overlies a black horizontallayer which is interpreted as the Level Q-7 floor. The total thickness of destruction debrisin this locality is ca. 1.5 m. (For interpretation of the references to color in this figurelegend, the reader is referred to the web version of this article.)

Fig. 4. (a) Top viewof a freshly prepared experimentalmudbrick. (b) The samebrick as in (a), af600 °C for 2 h. Note an insignificant change in volume but a distinct color change. (For interpretaof this article.)


2.2. Firing of experimental bricks

The bricks were placed, one at a time, within a laboratory furnace(Adam Mandel, T21 type coupled with a temperature programmerEurotherm 3216). Seven mini-thermocouples were inserted into thepre-designed elongated voids and one thermocouple was placed nextto the brick to record the temperature in the furnace's chamber duringheating (Fig. 5). The heating rate was 10 °C/min in all experiments. Atthe end of the heating the oven was kept closed and cooling occurredovernight (ca. 14 h). During heating, the temperatures in all 8 thermo-couples were simultaneously recorded every 5 s using a MultiConCMC-141 data recorder (by Simex). The recorded data was processedusing the software DAQ Manager. Four major types of heating experi-mentswere conducted in a closed furnace oven, i.e., heatwas conductedfrom all directions (details in Sections 2.3–2.6 below).

2.3. Heat intensity experiment

In order to test the behavior and material properties of bricksheated to different temperatures, we used a set of bricks of the samecomposition – 6 g of chaff to 100 g of sediment – (ratio 6/100), same di-mensions, and same duration of heating (2 h). The only variablewas themaximum temperature to which the oven was set to — 400 °C, 500 °C,600 °C, 700 °C and 800 °C. Each of these experiments was conductedat least twice to control for reproducibility.

2.4. Composition experiment

In order to test the behavior andmaterial properties of bricks of differ-ent compositions, we maintained a constant brick size, maximum oventemperature (600 °C), and heat duration (2 h). In composition we referstrictly to the ratio betweenmineral sedimentmatter andorganic temper.We are aware that differentmineralogical compositions of sediments alsoplay a role, however there is a limit to the number of variables that can betaken into consideration within one experimental study. We may studysedimentmaterial of othermineralogical compositions in future research.In light of this practical limitation, four different brick compositions werestudied differing in the weight of chaff (in g) added to 100 g of sediment,thus producing the following ratios— 4/100, 6/100 and 10/100 (Table 1).As a control, one experimental mud brick was prepared without chaff(ratio 0/100). Two bricks of the same compositionwere heated in similarconditions in order to control for reproducibility.

ter drying. Note the change in volume (ca. 18%). (c) Thebrick as in (a) and (b) after firing attion of the references to color in thisfigure legend, the reader is referred to theweb version

Fig. 5. Experimental heating setup. Multilateral heatingwithin a furnace oven, showing anexperimental mud brick with inserted thermocouples ready for measurement, and onethermocouple in the air for recording the chamber's temperature.


2.5. Duration experiment

In order to test the behavior andmaterial properties of bricks heatedto different lengths of time, we maintained a constant brick size, com-position (ratio 6/100), and maximal oven temperature (600 °C). In allexperiments the oven reached the designed maximal temperature,and then stopped at 0, 15, 30, 60 and 120 min after maximal tempera-ture was achieved. These experiments were not repeated.

2.6. Oxygen availability experiment

Dried experimental mud brick of the 6/100 composition was brokeninto pieces ca. 1 cm3 each. Six pieces were heated together in the ovenchamber, three open to the chamber atmosphere and three wrappedin aluminum foil which reduced the oxygen availability to these brickpieces. This treatment was conducted separately at 500, 600, 700 and800 °C, each for the duration of 2 h.

2.7. Post-heating analyses

The thermal behavior of mud bricks in each type of experiment wasstudied byplotting the recorded temperatures using Excel software. Thematerial properties of the experimental bricks were studied as follows:each brick was broken in half. The color pattern was recorded byphotography and Munsell color chart in dry condition.

Table 1Experimental brick compositions prepared for the composition experiment. Bricks ofsimilar size and volume were prepared.

Ratio Weight (g) Volume (ml)

Straw Sediment Water

0/100 0 400 1104/100 20 500 2206/100 36 600 32010/100 50 500 300

Sediment samples of a few milligrams each were collected from thecore and at the edge of each brick. These were analyzed using a FourierTransform Infrared (FTIR) spectrometer (Nicolet 380, Thermo ElectronCorporation) using the KBr method and a reference library (Weiner,2010). The spectral changes to the clay fraction were evaluated withreference to our own experimental calibration (Section 2.6) as well asShoval et al. (1991, 2011) and Berna et al. (2007). Specific attentionwas given to the position of the main absorbance band of the claycomponent, and to its width. In order to quantify the width of this ab-sorbance band (whose maximum may be located between ca. 1030and 1100 cm−1 depending on the heat intensity the clay mineral hasbeen exposed to; e.g., Berna et al., 2007), we measured the width ofthis band at 66% of its height. We chose to conduct this measurementat 66% of the band's height rather than the conventional ‘full width athalf maximum’ because at the half maximum (50%) of this band werisked incorporating into themeasurement thewidth of the quartz aux-iliary band located at 1163 cm−1. Because the width of absorbancebands is affected by the degree of grinding during sample preparation(Regev et al., 2010a; Poduska et al., 2011), we tested the experimentallyfired brick pieces utilizing the grinding curve approach (Regev et al.,2010a). We found that the position of the main clay absorbance bandwas not affected by grinding, meaning that the relationship betweenband position and its width is independent of grinding (this was impor-tant to establish in order to evaluate the results presented below inSection 3.4.2 and Fig. 13).

2.8. Test case

Eight mud bricks, identified as one intact segment of a burnt wallwere sampled in Square I/2, Level Q-7, during the 2014 excavation seasonat Tel Megiddo. The bricks in this wall segment show a variety of colors,including brown, red, pink, yellow, white, black and gray (Fig. 3). Thesamples were analyzed using FTIR spectroscopy as described above andinterpreted based on our own calibration (Section 2.7) as well as thenew insights gained from the intensity, duration and oxygen availabilityexperiments conducted here. We developed an approach that first con-siders brick color as an indicator for burning atmosphere (i.e., oxidizing

Fig. 6. Time-temperature curves for bricks fired at oven temperature of 400 °C (orange), 500 °C(green), 600 °C (red), 700 °C (purple) and 800 °C (blue). The curves are adjusted on the X-axis for ease of comparison. For each time-temperature set the light colored curve representsthe brick's edge (4–5mm from the brick surface) and the dark colored curve represents thebrick's core (33–34 mm from the brick surface). The black curve reflects the chamber tem-perature. Note that in all experiments the chamber's temperature slightly overshootsabove the designated temperature but quickly stabilizes ca. 20 °C above the designated tem-perature. The temperatures measured within the bricks, whether edge or core, are alwayshigher than the designated experimental temperature. This phenomenon is termed in thetext as “elevated temperature effect”. (For interpretation of the references to color in thisfig-ure legend, the reader is referred to the web version of this article.)


or reducing conditions), and then used FTIR spectra for the estimation ofburning temperatures.

3. Results

3.1. Heat intensity experiment

Heating experimental bricks of similar dimensions, composition andfor the same duration of time resulted in one common thermal patternin which all bricks experience heating well above the temperature mea-sured in the oven chamber. This phenomenon is found in measurementsconducted both in the edge and core of the experimental bricks (Fig. 6).This ‘elevated temperature effect’ has a limited duration. At 400 °C, theeffect lasts over 3 h, becoming shorter as the temperature is higher, re-ducing to 40 min at 800 °C (note: the elevated temperature effect inthe bricks heated to 800 °C is the first peak noted while the secondpeak is related to the exothermic effect of calcite disintegration intoCaO and CO2). The maximum temperature reached is systematicallyabout 100 °C higher than the oven chamber temperature, except forabout 50 °C when heating to 800 °C (Fig. 6).

A closer observation into the thermal behavior within the experi-mentally fired bricks reveals differences in the behavior between edgeand core. In bricksfired at 400 °C the edge reaches the oven temperatureafter 40min and experiences an elevated temperature effect of ca. 50 °C(i.e. not exceeding 465 °C). The core reaches the oven temperature after50 min but experiences a more pronounced elevated temperature effect,and is in fact heated to ca. 500 °C. The elevated temperature effect on theedge of the bricks diminishes as the heating temperature is higher, and itis always lower than the elevated temperature in the core. At 800 °C thecore and edge reach the oven temperature at about the same time.

In order to understand the reason for the elevated temperature effectwe carried out experiments with different brick compositions and differ-ent heating duration.

3.2. Composition and duration experiments

Bricks with different ratios of chaff to sediment were fired at 600 °Cfor 2 h. A compositional ratio of 0/100 (no chaff) results in the core andedge temperatures of the bricks being similar to the temperature of theoven chamber, and no elevated temperature effect is identified (Fig. 7a,blue curves). With increasing ratios of chaff to sediment the elevatedtemperature effect in the bricks' core is higher. When a brick is re-fired at the same temperature, no elevated temperature effect is noticed(Fig. 7b). This experiment proves that the elevated temperature effect isdirectly correlated to the amount of chaff in the brick, i.e., to the

Fig. 7. (a) Time-temperature curves of brickswith different compositional ratio of chaff to sedimencurves are adjusted on the X-axis for ease of comparison. Dark-colored curves represent the corecurves) do not experience an elevated temperature effect, and that the elevated temperature effecsame brick fired twice to oven set-up of 600 °C. The curves are adjusted on the X-axis for ease of csecond firing of this brick did not produce such an effect (orange curves). (For interpretation ofarticle.)

exothermic combustion of organic matter (see more in the discussion,Section 4.2).

In order to understand whether the elevated temperature effect isrelated to heat duration we carried out experiments with similar brickcomposition and dimensions, heating to 600 °C, while the oven wasstopped at different duration times (Fig. 8). The temperature in thebrick decreases immediately upon oven shut-down. The elevated tem-perature effect continues only as long as the oven supplies energy forcombustion of the organic matter. When the heat source is turned offonce the chamber reached the designated 600 °C temperature, thebrick's edge experiences higher temperature than the core (Fig. 8a).Longer heat duration produces an elevated temperature effect withthe core temperature eventually exceeding that of the edge (Fig. 8bthrough d).

3.3. Macroscopic observations

3.3.1. At different heat intensityUnfired dry bricks are grayish brown in color (2.5Y 5/2; Fig. 4b).

After firing at 400 °C, the color of the surface and core of the bricks isnot very different from that of the unfired bricks except for possiblybeing somewhat darker in the core (dark grayish brown 2.5Y 4/2;Fig. 9a). After firing at 500 °C and 600 °C, brick colors appear yellowishbrown or light brown (10YR 5/4 or 7.5YR 6/4, respectively) with noclear visual differences between the surface and the core (Fig. 9b, c).Burned chaff fragments appear as white stringers (composed mainlyof opal phytoliths). After firing at 700 °C distinct visual differencesappear between the core and edge of the bricks, with a light browncolor (7.5YR 6/4) at the surface and penetrating to about 5 mmwithinthe bricks (except for the bottom which was in contact with the floor ofthe oven) and a dark gray (10YR 4/1) brick core (Fig. 9d). Such distinctcolor differences also occur in bricks fired at 800 °C with a yellowishred (5YR 5/6) surface about 3–4 mm thick and a light brown core(7.5YR 6/4; Fig. 9e). These observed color patternsmay reflect differencesin oxidation states.

3.3.2. At different conditions of oxygen availabilityAll brick pieces heated open in the oven chamber, i.e., experiencing

high availability of oxygen, are yellowish brown (10YR 5/4), lightbrown (7.5YR 6/4) and yellowish red (5YR 5/6) in color. All brick piecescovered by aluminum foil, i.e., experiencing low availability of oxygen, areblackened, grayish and brown at very high temperature (2.5Y 2.5/1, 2.5Y4/1 and 7.5YR 5/3, respectively; Fig. 10).

t, allfired to oven set-upof 600 °C. Blue: 0/100; purple: 4/100; red: 6/100; green: 10/100. Theand light-colored curves represent the edge. Note that bricks with no chaff (ratio 0/100; bluet increaseswith the increasing ratio of chaff to sediment. (b) Time-temperature curves of theomparison. The first firing resulted in an elevated temperature effect (red curves) while thethe references to color in this figure legend, the reader is referred to the web version of this

Fig. 8. Time-temperature curves of bricks fired at 600 °C for different duration times. The plots show oven shut-down (i.e., no energy supplied, represented by the arrow) at (a) 0 min,(b) 15 min, (c) 30 min and (d) 1 h. The dashed line represents the temperature of the oven, while the lighter solid line shows the temperature at the edge of the brick and the darkerline shows the temperature at the core. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


3.3.3. At different durationsFiring bricks of the same dimensions and composition at 600 °C for

different time durations result in visual color differences related to thesize of the blackened brick cores. These blackened cores presumably re-sult from partly combusted organic matter (char) which relates to lowavailability of oxygen as seen above (Section 3.3.2). The longer thefiringduration, the smaller the proportion of un-combusted organic matter(Fig. 11).

We note that brick color was not affected by differences in theamount of organic matter. Bricks which included a high ratio of organicmatter turned light brown as long as enough oxygen penetrated intothem, which is related to the duration of burning (i.e., the more organicmatter a brick contains, the longer it takes to oxidize it all).

Fig. 9. Photographs showing sections of bricks after firing at (a

3.4. Mineralogical changes with brick firing

3.4.1. Calibration of spectral changes in heated mud brick fragmentsThe experimental calibration of the effect of heat on the infrared

spectrum of experimentalmud bricks, which represent the compositionof alluvial sediment from Megiddo's vicinity, is presented. The infraredspectrum of unheated (control) brick sample shows that the sedimentis composed of clay, calcite and quartz (Fig. 12a), a composition typicalof Israeli soils and sediments (Singer, 2007). Specifically, we identify themain component of clay in the sediment used for preparation of exper-imental mud bricks to be montmorillonite (see in comparison to Fig. 6in Berna et al., 2007; see also Eliyahu-Behar et al., 2012; Shoval et al.,2011). Previous studies conducted calibration of spectral changes to

) 400 °C, (b) 500 °C, (c) 600 °C, (d) 700 °C and (e) 800 °C.

Fig. 10. Photographs showing the colors of brick parts heated under oxidizing conditions at (a) 500 °C, (b) 600 °C, (c) 700 °C and (d) 800 °C, and under reducing conditions at (e) 500 °C,(f) 600 °C, (g) 700 °C and (h) 800 °C. Approximate size of brick pieces is ca. 1 × 1 cm. (For interpretation of the references to color in this figure legend, the reader is referred to the webversion of this article.)


clay minerals under oxidizing conditions only (Berna et al., 2007;Eliyahu-Behar et al., 2012; Shoval et al., 2011; Gur-Arieh et al., 2013;Friesem et al., 2014a). They identified the spectral parameters thatallow approximation of heating temperature, to be the absence of ab-sorbance bands of structural water (especially the 3690 cm−1 band)when heating occurs above 500 °C, lowering of the absorbance bandsat 535 and 913 cm−1, and a gradual shift of the main absorbance bandfrom1035 to 1084 cm−1with increasing temperature.We note a signif-icant difference in the position of themain clay absorbance band obtain-ed at the same heat intensity, under oxidizing vs. reducing conditions attemperatures above 500 °C (Table 2). In general, under oxidizing condi-tions the position of the main clay band is lower than under reducingconditions. For example, at 600 °C in oxidizing conditions the main ab-sorbance band is located at 1041 cm−1 on average, while in reducingconditions it is located at 1046 cm−1 on average (Fig. 12b, c). This differ-ence is evenmore pronounced at 700 and 800 °C (Table 2).We also notedthat at high temperatures thewidth of themain clay absorbance band in-creases (Fig. 12d, e). At these temperatures there also appears a smallband at 3645 cm−1 which is characteristic of calcium hydroxide

Fig. 11. Photographs showing sections of bricks after firing at 600 °C for time durations(after oven chamber reaches the designated heating temperature) of (a) 0 min,(b) 15 min, (c) 30 min and (d) 1 h.

[Ca(OH)2], a product of calcite disintegration (equivalent to dry lime).The increase in band width seems to follow the heat intensity (Table 2).

Fig. 13a presents the relationship between the position of the mainclay absorbance band and its width, at different heat intensities and dif-ferent levels of oxygen availability (data from Table 2). The relationshipobtained shows: (a) that no significant difference is discerned betweenoxidizing and reducing condition firings at 500 °C; (b) that clear differ-ence in band position occurs for firing in reducing and oxidizing condi-tions at 600 °C and higher temperatures; and (c) that the width of themain clay absorbance band is significantly higher when heating occursabove 700 °C. These observations allow us to define zones withinFig. 13awhichwill correspond to rather safe temperature reconstruction(separated by dashed lines in Fig. 13a); a zone that corresponds to

Fig. 12. FTIR spectra showing the general appearance of (a) unheated brick, (b) brick heatedto 600 °Cwith oxygen and (c) 600 °Cwithout oxygen. Note the significant difference in theposition of the clay main absorbance band at the same temperature but under different at-mosphere. (d) 800 °C with oxygen and (e) 800 °C without oxygen. Note the significantlywider main clay absorbance band.

Table 2Position and width of themain clay absorbance band in relation to heat intensity (oven set-up temperature) and the firing atmosphere (oxidizing vs. reducing conditions). Data from thesmall experimental brick pieces (covered and un-covered by aluminum foil) which do not undergo a significant temperature elevation effect due to their small size. Thus oven set-uptemperature reflects actual firing temperature.

Oven set-up temperature (°C) Main clay absorbance band position (cm−1) Width at 66% of main clay absorbance band(wavenumbers)

Oxidizing conditions Reducing conditions Oxidizing conditions Reducing conditions

500 1043 ± 1 1044 ± 4 77 ± 5 86 ± 6600 1042 ± 3 1047 ± 1 92 ± 4 103 ± 5700 1039 ± 4 1052 ± 3 117 ± 8 209 ± 18800 1036 ± 4 1045 ± 5 180 ± 12 185 ± 14


unheated clay or clay heated at temperatures below 500 °C, a zone thatcorresponds to clay heated in the range between 500 and 700 °C, and azone that corresponds to clay heated in the range between 700 and900 °C.

3.4.2. Internal test: What do experimental bricks record?Table 3 summarizes the results of heat intensity and spectral param-

eters obtained from experimental mud bricks, showing the elevatedtemperature effect that occurs at the core of chaff-tempered heatedbricks. It also shows the colors obtained in the cores and edges of theexperimentally burnt mud bricks. In order to test the validity of ourscheme (Fig. 13a) for temperature reconstruction we plotted the aver-agedvalues of the experimental bricks (Table 3)whichunderwent elevat-ed temperature effect while fired, and present different color patternsthat seem to correspond to differences in oxygen availability during firing(Fig. 13b). The pattern obtained confirms the theoretical scheme. Thebrick fired at oven set-up of 400 °C was heated at its core almost to 500 °Cand shows FTIR parameters of heat-altered clay while its edge washeated to 465 °C and its spectral parameters indicate heating below500 °C. In all other brick parts studied, there is 1:1 correspondencebetween the actual measured temperatures and the theoretical schemewe proposed in Fig. 13a. This demonstrates that the scheme is valid andmay be used to reconstruct temperatures of burning of archaeologicalmud bricks. We note that despite apparent similarities in the colors ofcores and edges in the experiments carried out in oven set-up of 400,500 and 600°, the infrared spectral parameters of these brick portionsdo indicate some degree of reduction in the brick cores, judged fromthe position of the main clay absorbance band relative to that of theedges (Fig. 13b).

3.5. Archaeological test case: Spectral data from an intact burnt mud brickwall segment

Fig. 13c shows the main clay absorbance band parameters (positionvs. width) for samples collected from various bricks found in an archaeo-logical intact wall segment unearthed in Tel Megiddo, Area Q, Level Q-7(a famous destruction level also known as Megiddo's Stratum VIA). Weshow only samples with colors that could be unequivocally assigned tooxidizing or reducing conditions of formation, i.e., brown/yellowish/reddish vs. dark gray/black (respectively; n = 21). The results indicatethat two blackened brick parts were heated at temperatures below500 °C, while 4 other blackened brick parts have been heated at therange of 500 to 700 °C. Brown- and yellow-colored bricks fall withintwo heat regimes, 500–700 °C (n = 9) and 700–900 °C (n = 10). Themajority of brown and yellow brick samples converges around 680 °Cand 750 °Cwhen compared to Fig. 13b, and seem to correspond primarilyto oxidizing conditions when compared to Fig. 13a. In summary, itappears that the blackened brick portions were heated under reducingconditions at ca. 600 °C or below, while the yellow and brown brickportions were heated at ca. 700 °C. Note that the colors that correspondto reducing conditions are found in the central part of the wall segmentwhile the yellow and brown bricks surround them (see Fig. 3b).

4. Discussion

We studied here the interplay between heat intensity, heat duration,amount of organic temper and availability of oxygen and their effect onthe physical appearance (color pattern) and infrared signal of mudbricks. We introduce a new parameter in the infrared spectrum of clayminerals, namely the width of the main clay absorbance band, showingthat it grows significantly wider when smectite clay (montmorillonitein our case) is heated above 700 °C. We introduce a new scheme bywhich temperature reconstruction of heated clay is somewhat more ac-curate than the scheme presented by Berna et al. (2007). Specifically,we suggest working with the relationship between the main clay absor-bance band position and its width (Fig. 13). This allows determination ofthe following heating ranges: unheated or heated below 500 °C, heatedbetween 500 and 700 °C, and heated between 700 and 900 °C. Weshow that studying in tandem the color of heated mud bricks and theirclay spectral parameters as suggested above refines archaeological inter-pretation and demonstrated the utility of this new approach via a casestudy of a wall segment from Level Q-7 in Tel Megiddo (Section 3.5).

The infrared spectral results showed a systematic difference in theposition of the main clay absorbance band when firing temperature is600 °C and above, reaching higher wavenumbers in reducing than inoxidizing conditions. According toMaritan et al. (2006), iron transformsfrom Fe2+ to Fe3+ during firing of clay in oxidizing conditions and thistransformation is completed at about 450 °C. When heating continues,iron oxides such as hematite, gehlenite and hercynite form above 850 °C.On the other hand, when firing occurs in reducing conditions, Fe3+

is reduced to Fe2+ until 750 °C, followed by formation of Fe-spinel at850 °C, which is followed by formation of metallic iron above 1000 °C(Maritan et al., 2006). We suggest that the distinctive difference in theposition of the main clay absorbance between bricks fired in oxidizingand reducing conditions is due to the difference in the oxidation state ofiron. This suggestion requires further testing. Below we discuss thearchaeological and historical implications of our study.

The shift of the main clay absorption towards higher wavenumbershas been highlighted as an indicative feature for reconstruction of tem-peratures in previously published studies (Shoval et al., 1989; Bernaet al., 2007; and others). In this studywe noticed that the exact positioncannot be used to differentiate temperatures in the range of 500–700 °C(in oxidizing conditions) because the shift is not systematic (seeTable 3). Indeed, Berna et al. (2007)were cautious enough to differenti-ate temperatures below and above 1000 °C, implying that band positionis useful for coarse temperature determinations. We note that wherepure clay was experimentally studied (e.g., Shoval et al., 1989, 2011;and others), temperature reconstruction based on main clay band posi-tion could approach a ±100 °C accuracy. The difference between thosestudies and ours may lie in the fact that our experiments used bulk sed-iment rather than purified clay minerals. The interactions between clayand other sediment components, notably organic matter and calcite,may result in the a-systematic manner of main band position in the500 to 700 °C range. This highlights the limitations of working withnatural sediments, in the present and past alike.

Fig. 13. The relationship between the position of the main clay absorbance band and its width, at different temperatures and firing atmospheres. (a) Experimental, based on small brickpieceswhich donot undergo elevated temperature effect. Note the separation into clearly defined thermal zones—unheated (see in Fig. 13b), clay heated at 500–700 °C, and clay heated at700–900 °C. (b) Internal test of the utility of the scheme presented in Fig. 13a, using experimental mud bricks which underwent the elevated temperature effect while fired, and show arange of colors that may correspond to different conditions of firing atmosphere. Note the agreement between the theoretical scheme and the results from the experimental mud bricks.(c) Results from the archaeological mud brick test case. Note that blackened brick portions appear to have been exposed to either heating below 500 °C (n= 2) or heating in the range of500 to 700 °C (n=4), while brown and yellow colored brick portions have been burnt at 500 to 700 °C (n=9) aswell as at 700 to 900 °C (n=10). The blackened partswere found only inthewall's core, indicating that thewall burnt as one unitwith lower temperatures and oxygen availability in its core relative to its edges. (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of this article.)


From 700 °C and above, the main clay absorption becomes broader,first in the bricks' core, then also at the edges. This indicates that glassyphases have formed (Shoval et al., 1989). We note that although thisbroadening occurs at high temperatures in pure clay (Berna et al.,2007), it appears to be even broader when calcite is also present (c.f.,Regev et al., submitted for publication), presumably due to the formationof calcium silicates (see also Regev et al., 2010b).

4.1. Archaeological implications

4.1.1. Brick composition and manufactureMud bricks from Level Q-7 have been extensively studied. Apart

from including heat-altered clay minerals they include other materialsthat lend information about their composition and thus manufacture.A most interesting feature is that Q-7 mud bricks have very high

Table 3Summary of measured heat intensity, FTIR spectral parameters and color, obtained from the cores and edges of experimental mud bricks. The color is used to deduce the firing conditionsand the spectral parameters are used to reconstruct the temperature range to which the bricks have been exposed to. Note the agreement between actual measured temperatures andreconstructed temperature range.

Oven set-up temp. Max. actual temp. Clay position(cm−1)

Clay width(cm−1)

Color (Munsell values) Condition deduced from color Reconstructed firing temp.

EDGE 400 465 ± 1 1037 ± 0 68 ± 3 Dark grayish brown (2.5Y 4/2) Oxidation 500–700500 560 ± 2 1043 ± 1 74 ± 2 Yellowish brown (10YR 5/4) Oxidation 500–700600 678 ± 34 1041 ± 3 93 ± 1 Light brown (7.5YR 6/4) Oxidation 500–700700 750 ± 2 1040 ± 1 113 ± 3 Light brown (7.5YR 6/4) Oxidation 700–900800 832 ± 1 1039 ± 2 182 ± 11 Yellowish red (5YR 5/6) Oxidation 700–900

CORE 400 496 ± 8 1045 ± 1 79 ± 1 Dark grayish brown (2.5Y 4/2) Oxidation 500–700500 596 ± 7 1045 ± 1 94 ± 2 Yellowish brown (10YR 5/4) Oxidation 500–700600 682 ± 17 1046 ± 2 118 ± 5 Light brown (7.5YR 6/4) Oxid/reduct 500–700700 792 ± 15 1040 ± 2 180 ± 10 Dark gray (10YR 4/1) Reduction 700–900800 857 ± 3 1050 ± 3 172 ± 2 Light brown (7.5YR 6/4) Reduction 700–900


phytolith concentrations (23 ± 8 million in 1 g sediment). The archaeo-logical mud bricks include abundant elongated voids after grass temper,indicating that large amounts of chaffwere indeed used in theirmanufac-ture. The concentration of phytoliths in our 6/100 compositional ratioexperimental bricks is 15 ± 3 million phytoliths per 1 g of sediment(n = 6), which is similar (within the statistical error) to that in thearchaeological bricks.

Apart from the addition of large amounts of chaff during the manu-facture of bricks, it appears that other components were added. TheLevel Q-7 mud bricks include minor amounts of calcite whose infraredparameters (specifically of the ν2 and ν4 vibrational modes; Regevet al., 2010a) suggest that it is slightly disordered, as would be expectedfrom calcite forming after exposure to high temperatures. It is howeverunclear whether this heat-affected calcite originates from heated sedi-ment carbonates and/or from mixing vegetal ash products with sedi-ment during brick manufacture. Two observations support the secondoption, (a) the mud bricks contain phosphate minerals and (b) the highconcentrations of phytoliths, both materials are well documented inhearth and oven ashes from both ethnoarchaeological and archaeologicalcontexts, including at Tel Megiddo (Gur-Arieh et al., 2013; 2014). Wetherefore suggest that the recipe of preparation of the Level Q-7 mudbricks included a mixture of ash rake-out from ovens and hearths withnatural sediment.

The use of ash as an additive in brick manufacture in antiquity hasnot been highlighted enough in previous research.We show here initialevidence for this practice in the late Iron I city ofMegiddo. It will be inter-esting to studymoremud bricks across time and space in order to evalu-ate whether this practice was unique to Megiddo at this specific timeperiod, or was more wide spread. The potential practical contribution ofash in brickmanufacture is increasedmechanical strength (similar to ad-dition of cement; Binici et al., 2005, 2007), yet the thermal conductivitycoefficient in mud bricks with additional cement (lime) is higher thanthe bricks without lime cement (Binici et al., 2007).

In summary, the Q-7mud bricks have been prepared from amixtureof sediment, grass temper and ash. The grass temper in these bricksapproximated the 6/100 compositional ratio in our experiments.

4.1.2. Availability of oxygen during brick firingThe mud brick debris in layer Q-7 at Tel Megiddo varies in thickness

between 10 and 120 cm.Most of it is composed of separate bricks in ran-dom orientations, with few occasions of intact short walls. Across the en-tire excavated area, ca. 125m2, most of the bricks appear reddish-brown.Yellowish colors are also present, while gray or black colored bricks arerare (we do not provide Munsell values here because the variability istoo large). These observations suggest that brick heating occurredmostlyunder oxidizing conditions.

4.1.3. Heat intensity in fired bricksUsing the new approach presented in this study, that combines

macroscopic evaluation of firing atmosphere based on brick color, as

well as using the relationship between the position and width of themain clay infrared absorbance band (Fig. 13), we were able to showthat most mud bricks from an intact wall segment in Level Q-7 in TelMegiddo experienced heat of about 700 °C (on average) under oxidizingconditions, and that blackened brick portions from thewall's core expe-rienced burning under reducing conditions at 500–600 °C. In fact, thispattern resembles the pattern obtained during the experimental brickfiring conducted in the laboratory, indicating that the whole wallbehaved like a very large brick (with an outer oxidized edge and aninner reduced core). This finding seems to indicate that the wall wasbuilt from sun-driedmud bricks andwas later exposed tofire that heatedit intact.

Studying the heat intensity experienced by the bricks found in pilesacross the excavation area, most of which are reddish-brown in color,indicates that most bricks were exposed to temperatures of some600–700 °C. We note that all mud bricks studied thus far from LevelQ-7 at Tel Megiddo do not include structural water in association withclay minerals, indicating that re-hydroxylation is not an importantdiagenetic process and it does not hamper our reconstructions.

Based on our experimental results (Section 3.1), grass-temperedbricks experience an elevated temperature effect while they are heated.A similar phenomenon is known in northwestern Europe where indus-trial bricks such as the Fletton/London bricks ignite during kiln firingdue to the carbonaceous nature of the source clay. Yet, we could notfind published direct measurements of the effect of this ignition. Wehave shown that the elevated temperature effect is directly correlatedwith the ratio of organic temper to sediment. Based on the similarityin phytolith concentrations between the Q-7mud bricks and our exper-imental composition 6/100, we expect a similar elevated temperatureeffect to have occurred in the Q-7 mud bricks. In our experiments, theeffect within the bricks relative to the oven temperature was in excessof ca. 100 °C. Therefore, considering that based on FTIR spectroscopywe reconstruct exposure of clay minerals in Q-7 bricks to temperaturesof 600–700 °C, we infer that the environmental temperature aroundthose bricks was lower in ca. 100 °C, probably around 500–550 °C.

4.1.4. Brick firing durationOur experiments show that bricks heated for a short duration will

have black-colored cores (Fig. 11). This is known as black-coring(Gredmaier et al., 2011), due to a reductive atmosphere during thefiring. The size of the black core diminishes with heat duration. Mostof the Level Q-7 mud bricks do not show black cores. This indicatesthat the fire duration must have been long. Considering that in oursmall experimental bricks (volume of 125 cm3) absence of a black coreoccurred after 1 h of constant heat at 600 °C, and that the average volumeof the Q-7 mud bricks is about 100 times higher (ca. 17,000 cm3), it canbe deduced that the Q-7 bricks have been heated for a minimum of 1 h.At this stage of the research, the exact duration of firing cannot be esti-mated, and future experiments will test the effect of brick shape and


size/volume as well as heat conductance through bricks of variouscompositions.

4.2. Historical implications

Level Q-7 at Tel Megiddo represents the largest destruction the cityof Megiddo has known. It occurs across the tell in all excavation areas,denoted stratigraphically as Stratum VIA, and dates to the late Iron I(late 11th century BCE). Field observations described above, coupledwith our analytical results indicate that most mud bricks across a wideexcavation area have been heated in an environmental temperature of500–550 °C in what appears to be quite uniform conditions. Specifically,most bricks show color patterns indicative of firing in oxidizing condi-tions at about the same temperature. This finding raises important ques-tions regarding this destruction event, as it can indicate two entirelydifferent interpretations that may significantly change our understand-ing about the nature of this destruction event (and destructions in gener-al). One possible line of interpretation is that such homogenously firedbricks were produced in kilns prior to building the late Iron I city. Suchan interpretation implies that the “red brick city”may have been so be-fore its destruction, and that building with kiln-fired bricks occurred inthe southern Levant much earlier than currently suggested (supportingNamdar et al., 2011). Building awhole city fromkiln-fired bricks requireslarge amounts of fuel which may be obtained either through deforesta-tion and/or use of dung as fuel. No evidence currently exists for defores-tation, nor have brick-firing kilns been identified in Megiddo and itsvicinity. Taken together with the labor investment needed to producelarge amounts of kiln-fired bricks to build an entire Iron Age city, wefind this line of interpretation unlikely (but it cannot be ruled out). Theother possible line of interpretation is that the bricks were heatedhomogenously during the destruction event (in line with the ‘commonwisdom’). This line of interpretation is also not clear-cut: (a) Color pat-tern in Q-7 bricks is homogenous or symmetrical, as produced duringoven/kilnfiring (i.e., atmulti-directional constant heat). (b) It is expectedthat fire intensity during destruction will vary across space in relation tothe availability of combustiblematerials. So far we have not encounteredsignificant differences in fire intensity across the ca. 125 m2 of excava-tion. (c) A recent geo-ethnoarchaeological study reported that clay alter-ation by heat onmud brickwalls of an accidentally burnt barn and stablein a village in northern Greece was detected only on the outer fewmilli-meters of the bricks composing the burnt walls (Friesem et al., 2014a).

We are facing, therefore, an interpretational gridlock. Our small testcase of an intact brick wall segment may open the way for solving thisproblem. As stated above, the results indicate that the wall was burntas one unit with external oxidizing and internal reducing conditions,i.e., the fire occurred post-building with sun-dried mud bricks. Threeadditional lines of evidence should be used in future research. First,the excavation area should be enlarged and detailed mapping of evi-dence for heat intensity should be conducted (using FTIR spectroscopyand macroscopic observations) across space. This will determinewhether heat was indeed uniform across the destruction layer. Second,more experiments that target estimation of the heat duration needed totransform large sun-dried mud bricks (as well as thick wall segments)into almost uniformly heated fired bricks should be conducted. Lastly,but probably the most telling line of evidence, would be to study themagnetic directions of in-situ mud bricks. In a scenario that assumesthe use of kiln-fired bricks, one would expect that bricks found in-situwithin wall segments will show four magnetic directions at 90° to oneanother. In a scenario that assumes that whole walls have been heatedat once, all bricks should show one magnetic direction.

5. Conclusion

Our experimental studies into the thermal behavior of chaff-tempered mud bricks show that an elevated temperature effect occurswithin fired bricks due to an exothermic decomposition of the organic

temper. We found that the position of the FTIR main absorbance bandof clay is not shifted in a systematic manner in mud bricks which areproduced from poly-mineralic sediments, and also depends on theavailability of oxygen during firing. In the present study we exploredthe relationship between the position of the main clay absorbance andits width, and found an important behavior that allows approximationof heating temperatures in the following ranges: unheated to 480 °C;490–700 °C; above 700 °C. This refines previous observations by Bernaet al. (2007).

Mineralogical transformations deduced from FTIR spectra of heatedmud bricks, together with the macroscopic appearance of mud bricksmade it possible to reconstruct aspects in the manufacture and thermalhistory of mud bricks in Level Q-7 at Tel Megiddo. Application of theexperimental results in the study of Tel Megiddo mud bricks from theStratumVIA destruction layer in excavation area Q provides the followingarchaeological information:

(1) All mud bricks studied thus far have been heated.(2) The environmental heat was ca. 500–550 °C.(3) Most bricks have been uniformly heated (as may be expected in

kiln-firing).(4) An intact brick wall segment seems to have been burnt as one

unit after it has been built from sun-dried mud bricks.(5) Bricks were manufactured by mixing sediment, ash and chaff

vegetal temper.(6) No significant spatial variability in heat intensity (as it is recorded

in brick piles) has been identified across a 125m2 excavation area.

This archaeological information presents a complex scenario thatcannot be easily interpreted. The present study did not include all ofthe possible variables that should be considered in an archaeologicalcontext. While these variables, such as brick volume, rate of heatingand direction of heat are beyond the scope of the current study, we in-tend to explore them in future studies whichmay lead to a more defin-itive interpretation and possibly important historical implications.

Acknowledgments

This study was supported by a European Research Councilunder the European Community's Seventh Framework Programme(FP7/2007–2013)/ERC grant agreement no. 229418 to S. Weiner and I.Finkelstein where R. S-G acted as track leader, and the Kimmel Centerfor Archaeological Science, Weizmann Institute of Science. We aregrateful to the team at Tel Megiddo and the excavation directors I.Finkelstein and E. Cline as well as S. Weiner, director of the KimmelCenter.

References

Albert, R.M., Shahack-Gross, R., Cabanes, D., Gilboa, A., Lev-Yadun, S., Portillo, M., Sharon,I., Boaretto, E., Weiner, S., 2008. Phytolith-rich layers from the Late Bronze and IronAges at Tel Dor (Israel): mode of formation and archaeological significance.J. Archaeol. Sci. 35, 57–75.

Berna, F., Behar, A., Shahack-Gross, R., Berg, J., Boaretto, E., Gilboa, A., Sharon, I., Shalev, S.,Shilstein, S., Yahalom-Mack, N., Zorn, J.R., Weiner, S., 2007. Sediments exposed tohigh temperatures: reconstructing pyrotechnological processes in Late Bronze andIron Age strata at Tel Dor (Israel). J. Archaeol. Sci. 34, 358–373.

Binici, H., Aksogan, O., Shah, T., 2005. Investigation of fibre reinforced mud brick as abuilding material. Constr. Build. Mater. 19, 313–318.

Binici, H., Aksogan, O., Nuri, Bodur M., Akca, E., Kapur, S., 2007. Thermal isolation and me-chanical properties of fibre reinforced mud bricks as wall materials. Constr. Build.Mater. 21, 901–906.

Boaretto, E., 2006. Radiocarbon dates. In: Finkelstein, Israel, Ussishkin, David,Halpern, Baruch (Eds.), Megiddo IV. Tel-Aviv University, Tel-Aviv, pp. 550–557.

Clegg, F., Breen, C., Carter, M.A., Ince, C., Savage, S.D., Wilson, M.A., 2012. Dehydroxylationand rehydroxylation mechanisms in fired clay ceramics: a TG–MS and DRIFTS inves-tigation. J. Am. Ceram. Soc. 95, 416–422.

Cline, E.H., 2011. Whole lotta shakin' going on: the possible destruction by earthquake ofStratum VIA at Megiddo. In: Finkelstein, I., Na'aman, N. (Eds.), The Fire Signals ofLachish: Studies in the Archaeology and History of Israel in the Late Bronze Age,

http://refhub.elsevier.com/S2352-409X(15)00009-7/rf0005




















Iron Age, and Persian Period in Honor of David Ussishkin. Eisenbrauns, Winona LakeIndiana, pp. 55–70.

Dever, W.G., 1992. A case study in biblical archaeology: the earthquake of ca. 760 BCE.Eretz Israel 23, 27–35.

Eliyahu-Behar, A., Yahalom-Mack, N., Shilstein, S., Zukerman, A., Shafer-Elliott, C., Maeir,A.M., Boaretto, E., Finkelstein, I., Weiner, S., 2012. Iron and bronze production inIron Age IIA Philistia: new evidence from Tell es-Safi/Gath. J. Archaeol. Sci. 39,255–267.

Emery, V.L., Morgenstein, M., 2007. Portable EDXRF analysis of a mud brick necropolis en-closure: evidence of work organization, El Hibeh, Middle Egypt. J. Archaeol. Sci. 34,111–122.

Farmer, V.C., 1974. The infrared spectra of minerals. Mineralogical Society Monograph. 4.Batholomew Press, London.

Finkelstein, I., 2009. Destructions: Megiddo as a case study. In: Schloen, D.J. (Ed.), Exploringthe Longue Durée: Essays in Honor of Lawrence E. Stager, pp. 113–126 (Winnona Lake,Indiana).

Finkelstein, I., Piasetzky, E., 2009. Radiocarbon-dated destruction layers: a skeleton forIron Age chronology in the Levant. Oxf. J. Archaeol. 28, 255–274.

Finkelstein, I., Ussishkin, D., Halpern, B., 2000. Megiddo III. Tel-Aviv University, Tel-Aviv.Finkelstein, I., Ussishkin, D., Halpern, B., 2006. Megiddo IV: The 1998–2002 Seasons.

Monograph Series of the Institute of Archaeology. Tel-Aviv University, Tel-Aviv.Finkelstein, I., Ussishkin, D., Cline, E.H., 2013. Megiddo V: The 2004–2008 Seasons. Mono-

graph Series of the Institute of. Tel Aviv University, Archaeology.Friesem, D., Boaretto, E., Eliyahu-Behar, A., Shahack-Gross, R., 2011. Degradation of mud

brick houses in an arid environment: a geoarchaeological model. J. Archaeol. Sci. 38,1135–1147.

Friesem, D.E., Karkanas, P., Tsartsidou, G., Shahack-Gross, R., 2014a. Where are the roofs?A geo-ethnoarchaeological study towards the identification of roofs in the archaeo-logical record. Archaeol. Anthropol. Sci. 6, 73–92.

Friesem, D.E., Karkanas, P., Tsartsidou, G., Shahack-Gross, R., 2014b. Sedimentary processesinvolved in mud brick degradation in temperate environments: a micromorphologicalapproach in an ethnoarchaeological context in northern Greece. J. Archaeol. Sci. 41,556–567.

Goldberg, P., 1979. Geology of a Late Bronze Age mudbrick from Tel Lachish. Tel-Aviv 6,60–67.

Gredmaier, L., Banks, C.J., Pearce, R.B., 2011. Calcium and sulphur distribution in fired claybrick in the presence of a black reduction core usingmicro X-ray fluorescencemapping.Constr. Build. Mater. 25, 4477–4486.

Gur-Arieh, S., Mintz, E., Boaretto, E., Shahack-Gross, R., 2013. An ethnoarchaeologicalstudy of cooking installations in rural Uzbekistan: development of a new methodfor identification of fuel sources. J. Archaeol. Sci. 40, 4331–4347.

Gur-Arieh, S., Shahack-Gross, R., Maeir, A.M., Lehmann, G., Hitchcock, L.A., Boaretto, E., 2014.The taphonomy and preservation of wood and dung ashes found in archaeologicalcooking installations : case studies from Iron Age Israel. J. Archaeol. Sci. 46, 50–67.

Harrison, T.P., 2004. Megiddo 3. Final Report on the Stratum VI Excavations. 127. OrientalInstitute Publications, Chicago.

Homsher, R.S., 2012. Mud bricks and the process of construction in theMiddle Bronze Agesouthern Levant. Bull. Am. Sch. Orient. Res. 368, 1–27.

Kempinski, A., 1989. Megiddo. A City State and Royal Centre in North Israel, Munich.Lamon, R., Shipton, G., 1939. Megiddo I: Seasons of 1925–34. Strata I-V, Chicago.Loud, G., 1948. Megiddo II. Season of 1935–1939. 62. Oriental Institute Publications,

Chicago.Love, S., 2012. The geoarchaeology of mudbricks in architecture: a methodological study

from Çatalhöyük, Turkey. Geoarchaeology 27, 140–156.

Love, S., 2013. Architecture as material culture: building form and materiality in thePre-pottery Neolithic of Anatolia and Levant. J. Anthropol. Archaeol. 32, 746–758.

Maritan, L., Nodari, L., Mazzoli, C., Milano, A., Russo, U., 2006. Influence of firing conditionson ceramic products: experimental study on clay rich in organic matter. App. Clay Sci.31, 1–15.

Mellart, J., 1975. The Neolithic of the Near East. Scribner, New York.Namdar, D., Zukerman, A., Maeir, A.M., Katz, J.C., Cabanes, D., Trueman, C., Shahack-Gross,

R., Weiner, S., 2011. The 9th century BCE destruction layer at Tell es-Safi/Gath, Israel:integrating macro- and microarchaeology. J. Archaeol. Sci. 38, 3471–3482.

Nodarou, E., Frederick, C., Hein, A., 2008. Another (mud)brick in the wall: scientific anal-ysis of Bronze Age earthen construction materials from East Crete. J. Archaeol. Sci. 35,2997–3015.

Poduska, K.M., Regev, L., Boaretto, E., Addadi, L., Weiner, S., Kronik, L., Curtarolo, S., 2011.Decoupling local disorder and optical effects in infrared spectra: differentiatingbetween calcites with different origins. Adv. Mater. 23, 550–554.

Regev, L., Poduska, K.M., Addadi, L.,Weiner, S., Boaretto, E., 2010a. Distinguishing betweencalcites formed by different mechanisms using infrared spectrometry: archaeologicalapplications. J. Archaeol. Sci. 37, 3022–3029.

Regev, L., Zukerman, A., Hitchcock, L., Maeir, A.M., Weiner, S., Boaretto, E., 2010b. Iron Agehydraulic plaster from Tell es-Safi/Gath. Isr. J. Archaeol. Sci. 37, 3000–3009.

Regev, L., Cabanes, D., Homsher, R., Kleiman, A., Weiner, S., Finkelstein, I., Shahack-Gross,R., 2015n. Geoarchaeological Investigation in a Domestic Iron Age Quarter, Tel Megiddo,Israel. BASOR Submitted for publication.

Savage, S.D., Wilson, M.A., Carter, M.A., McKay, B., Hoff, W.D., Hall, C., 2008. Mass gain dueto the chemical recombination of water in fired clay brick. J. Am. Ceram. Soc. 91,3396–3398.

Schumacher, G., 1908. Tell el-Mutesellim I. Leipzig.Shahack-Gross, R., Albert, R.-M., Gilboa, A., Nagar-Hilman, O., Sharon, I., Weiner, S., 2005.

Geoarchaeology in an urban context : the uses of space in a Phoenician monumentalbuilding at Tel Dor (Israel). J. Archaeol. Sci. 32, 1417–1431.

Shahack-Gross, R., Gafri, M., Finkelstein, I., 2009. Identifying threshing floors in thearchaeological record: a test case at Iron Age Tel Megiddo. Isr. J. Field Archaeol. 34,171–184.

Shoval, S., Erez, Z., Kirsh, Y., Deutsch, Y., Kochavi, M., Yadin, E., 1989. Determination of theintensity of an Early Iron Age conflagration at Tel-Hadar. Isr. Thermochim. Acta 148,485–492.

Shoval, S., Beck, P., Kirsh, Y., Levy, D., Gaft, M., Yadin, E., 1991. Rehydroxylation of clayminerals and hydration in ancient pottery from the ‘Land of Geshur’. J. Therm. Anal.37, 1579–1592.

Shoval, S., Yadin, E., Panczer, G., 2011. Analysis of thermal phases in calcareous Iron Agepottery using FT-IR and Raman spectroscopy. J. Therm. Anal. Calorim. 104, 515–525.

Singer, A., 2007. The Soils of Israel. Springer, Verlag Berlin Heidelberg.Stevanović, M., 1997. The age of clay: the social dynamics of house destruction.

J. Anthropol. Archaeol. 16, 334–395.Twiss, K.C., Bogaard, A., Bogdan, D., Carter, T., Charles, M.P., Farid, S., Russell, N., Stevanović,

M., Yalman, N.E., Yeomans, L., 2008. Arson or accident? The burning of a neolithic houseat Çatalhöyük. Turkey. J. Field Archaeol. 33, 41–57.

Weiner, S., 2010. Microarchaeology: Beyond the Visible Archaeological Record. CambridgeUniversity Press, Cambridge.

Zarzecki-Peleg, A., 2005. Tel Megiddo During the Iron Age I and IIA–IIB: the Excavations ofthe Yadin Expedition to Megiddo and Their Contribution for Comprehending theHistory of This Site and Other Contemporary Sites in Northern Israel. The HebrewUniversity of Jerusalem, Jerusalem (Ph.D. dissertation).





































































































physical and mineralogical properties of experimentally heated chaff-tempered mud bricks:...

Documents