Introduction

During the period of the Roman emperor’s Imperial Porphyry, a porphyritic igneous rock of trachyandesitic to dacitic composition with spectacular purple color (Fig. 1), was extensively quarried in the Mons Porphyrites area north of Jabal Dokhan in the Eastern Desert of Egypt (Fig. S1 in the electronic supplement). This was the only known source of the Imperial Porphyry in antiquity. Judging from local inscriptions dated at the reign of Emperor Tiberius (14–37 AD), the Mons Porphyrites site was discovered by the Roman explorer Caius Cominius Leugas who erected a black porphyry stele, which he dedicated to Pan-Min, the God of the Eastern Desert and the goddess Sarapis, indicating that he was the finder of the Imperial Porphyry and other rock types (van Rengen 1995, 2001). The text, written in Greek language, reads in English translation (V.v.S.): “Gaios Kominios Leugas, the finder of the quarries of the Porphyrites and Knekites and the black Porphyrites and various stones, has cut a prayer to Pan and Sarapis, the greatest gods, for the wealth of his children. Regnal year 4 of Tiberius, the Holy Kaisar, Epeiph 29”. This date corresponds to July 23, 18 AD. Thus, quarrying likely commenced before the year 18 AD, but came to a standstill in the middle of the fifth century AD (Klein 1988; Brown and Harrell 1995; Maxfield and Peacock 2001; Klemm and Klemm 2008). During this time c. 10,000 tons of dimension stone were extracted from numerous quarries (Klemm and Klemm 2001a). At Mons Claudianus, about 30 km south of Mons Porphyrites, light-colored quartz diorite was extensively quarried during the first to fourth century AD (Fig. 2 inset, S1) (Klemm and Klemm 2008; Harrell and Storemyr 2009).

Fig. 1
figure 1

a Roman column of brecciated Imperial Porphyry, mined in the Lykabettus quarries, 3 m in length, with a sculpture of two male figures, interpreted as Roman co-emperors of the tetrarchy period (293–313 AD; Delbrueck 1932), Vatican Library, Rome. b Close-up of the brecciated rock consisting of fragments with porphyritic texture, rich in white colored plagioclase phenocrysts, in a fine-grained matrix of dark purple color (from Lorenz 2012)

Full size image
Fig. 2
figure 2

Detailed geological map of the Wadi Um Sidri-Wadi Abu Ma’amel area (based on our own field work and Landsat image processing). Inset map: Geological map of the Eastern Desert of Egypt (modified after Eliwa et al. 2010; Ali et al. 2012)

Full size image

The Roman quarries are situated in high topographic positions between c. 800 and 1200 m on the hills forming the eastern and western slopes of Wadi Abu Ma’amel (Fig. 2) (e.g., Dardir and Abu Zeid 1972; Maxfield and Peacock 2001). In the nineteenth century, the quarries were detected again by the British explorers James Burton and John Gardner Wilkinson (1822/1823), the Italian explorer Giovanni Battista Brocchi (1823) and the French petrographer Achille Delesse (1852). In 1875, Georg Schweinfurth, a Baltic German explorer, investigated and named three of the different quarry areas (Schweinfurth 1887), i.e., the Rammius and Lykabettus quarries at the west side and the Lepsius quarries on the east side of the valley (Fig. 2). In the NW part of the area, Meredith and Tregenza (1950) detected the Northwest quarries and the Bigfoot quarries. Finally, Nick Bradford recovered the Northeast (Bradford) quarries (Maxfield and Peacock 2001) at the eastern side of Wadi Abu Ma’amel.

The Roman quarries were hardly visited since the end of the fifth century AD and, therefore, provide ample, undisturbed evidence for the production and dressing techniques of Roman dimension stone. Unfinished porphyry blocks show characteristic traces of different working stages (Figs. S2, S3). Downhill transport of the porphyry blocks was achieved by a system of stone chutes that connected the quarries with the main loading platform at the opening of the Wadi Umm Sidri towards the east. The quarry areas were accompanied by villages containing housing estates for workers, working areas for blacksmiths and stonemasons as well as relics of a large fort, a public bath and of two temples dedicated to Isis and Serapis (Fig. S2d) (Maxfield and Peacock 2001). The extremely remote, hyperarid mountainous mining area was connected to the ancient port of Καινή (now Qena) at the Nile River by a Roman trade route (Fig. S1), via porphyrites, which was first described by Strabo (63 BC–23 AD) and shown on a map by Claudius Ptolemy (c. 100–170 AD; Werner 1998). Porphyry is defined as “A general term for any igneous rock that contains phenocrysts in a finer-grained groundmass” (LeMaitre et al. 2002), i.e., displaying a porphyritic texture, as developed in many (sub-) volcanic rocks (see also Bates and Jackson 1997). To the best of our knowledge, the term πορφυρίτης (purple dye or purple stone) appeared first in the Greek inscription, on an Imperial Porphyry stele of Caius Cominius Leugas (van Rengen 1995, 2001).

A concise account on the geological setting, mineralogy, petrology and geochemistry of Imperial Porphyry is given by Williams-Thorpe et al. (2001). A much more extensive study was presented by Makovicky et al. (2016a, b), but is essentially focused on samples from the Lepsius quarries and some boulders from Wadi Abu Ma’amel. Our new investigation covers samples from Wadi Umm Sidri, Wadi Abu Ma’amel and all quarries. Therefore, new fieldwork and extensive sampling was carried out in the framework of an interdisciplinary Porphyry Project (Lorenz 2012), providing new insights into the mutual contact relationships between the important rock types and their petrographic characteristics. Moreover, results of geochronological and geochemical analyses of the samples collected will be discussed together with the crucial problem of the spectacular purple coloring of the Imperial Porphyry. Our results may contribute to a better understanding of the petrogenesis of this exceptional dimension stone and, at best, provide evidence for the exact provenance of individual works of art. In addition, an archeological sample WS-E9 from a medieval portable altar found in the castle “Schlössel” close to Klingenmünster in Germany was added to this study (Barz et al. 2012).

Cultural relevance of Imperial Porphyry

Imperial Porphyry was used for the production of architectural elements and works of art, exclusively reserved for the imperial court of the Roman Empire. High appreciation was mainly due to its characteristic purple color, traditionally regarded as an imperial status symbol (e.g., Klemm and Klemm 2001a). Careful polishing of the rock achieved a highly reflective finish and a sparkling appearance of the feldspar phenocrysts. Remarkably, brecciated varieties of Imperial Porphyry were regarded as the most attractive and thus were frequently used for fine works of art (Fig. 1 and Figs. S4, S5, S6). Under the Roman emperors Trajan (98–117 AD) and Hadrian (117–138 AD), the period of porphyries’ fashion reached a first maximum, while a second one occurred under the reign of Diocletian (284–305 AD), Constantine the Great (306–337 AD) and his followers on the throne of the Byzantine Empire.

The different steps of extraction and transport of Imperial Porphyry involved a wealth of masterly technical performances, all the more as a widespaced joint pattern allowed the production of large, cleavage- and fracture-free blocks, suitable for manufacturing of monolithic or sectioned columns, mantle pieces, pedestals, sarcophagi, sculptures, ceremonial bathtubs, magnificent giant bowls and other masterworks of Roman craftsmanship. For instance, individual segments of the Column of Constantine in Constantinople (330 AD) attain a weight of up to 45 tons (Fig. S6a). Moreover, it must be assumed that the Roman engineers had disposed of a fundamental knowledge on geological survey. Otherwise, they would have hardly been able to detect the occurrences of Imperial Porphyry, suitable for quarrying.

Works of art designed of Imperial Porphyry in Pharaonic Egypt are rare. Only small sculptures and plates have been shaped from loose boulders (Klein 1988). According to Del Bufalo (2013), the oldest example so far known is a man’s head from dynasty XXVI (664–525 BC). In contrast, many architectural elements and artworks from Roman antiquity are still preserved on-site in historic monument areas of the eastern and central Mediterranean, or are exhibited in museums.

Throughout the Middle Ages and, especially, in the Renaissance and Baroque eras, the demand for this rock type, called Porfido rosso antico by Italian sculptors, remained high and, after the closure of the ancient quarries, was only met by the use of antique spoils. These spoils, derived from destroyed Roman buildings, were either used as raw materials to design sculptures or floor and wall covers, or were incorporated, in their original form, in new churches or palaces. A compilation of the most important architectural elements and art objects made of Imperial Porphyry, which can be visited in Europe, is given in the electronic supplement.

General geological setting

The crystalline basement of the southern Sinai Peninsula and the Eastern Desert of Egypt was formed during the Pan-African Orogeny, due to collision of East and West Gondwana and the related closure of the Mozambique ocean in Neoproterozoic times (e.g., Shackleton 1986; Stern 1994). By this orogenic event, Neoproterozoic sedimentary sequences with intercalated (sub-) volcanic and plutonic rocks were deformed and metamorphosed under elevated pressures and temperatures, and subsequently intruded by Older I-type and Younger A-type granitoid magmas. (e.g., Abu El-Enen and Whitehouse 2013 and references therein; see Fig. 2 inset). The basement complex contains three different groups of Neoproterozoic volcanic rocks of mafic to intermediate, subordinately of felsic composition (Eliwa et al. 2006, 2014), i.e., the older metavolcanics, the younger metavolcanics (Stern 1981) and the Dokhan Volcanics. The later ones are represented by volcano-sedimentary successions containing varicolored lava flows, pyroclastic rocks and ignimbrites differing in thickness from a few tens of meters to about 1200 m (Wilde and Youssef 2000; Eliwa et al. 2006, 2010). Their composition ranges from basaltic andesite, via andesite, trachyandesite, and dacite to rhyolite. Moreover, subvolcanic rocks with porphyritic texture intruded into shallow crustal levels. Most of the Dokhan Volcanics are un-metamorphosed or, in some occurrences, underwent low-pressure metamorphism (Eliwa et al. 2006, 2014 and references therein).

The Dokhan Volcanics contain enclaves of the Older I-type granitoids, but on the other hand, are more or less of the same age as the late- to post-orogenic Younger A-type Granitoids (e.g., Stern and Gottfried 1986; Eliwa et al. 2006; see Fig. 2). Rb–Sr whole rock dating yielded wide age ranges of 610–560 Ma for the Dokhan Volcanics (Ressetar and Monard 1983; Ries et al. 1983) and of 610–550 Ma for the Younger Granites (e.g., Stern and Hedge 1985). A broadly conforming age range of 630–590 Ma was obtained by SHRIMP dating on single zircons from the Dokhan Volcanics (Wilde and Youssef 2000; Breitkreuz et al. 2010; and own data, see below). Simultaneously, the molasse-type Hammamat sediments, present outside the area of Fig. 2, were deposited forming interlayers with the volcanic rocks. The volcano-sedimentary sequence and the granites were intruded by contemporaneous, or slightly younger, swarms of NE–SW to E–W trending, bi-modal or composite dikes of andesite and rhyolite composition. These were emplaced in a highly extensional tectonic regime similar to the Basin and Range Province (Stern et al. 1988, and references therein).

Field relations

The ancient quarry sites at Mons Porphyrites are situated north of Jabal Dokhan in the central Eastern Desert about 50 km west of the seaside resort Hurghada at 27°14.2–16.1′N, 33°16.5–18.5′E (Fig. S1; compilation by Brown and Harrell 1995 and description by; Maxfield and Peacock 2001). Although evidence of subaerial volcanism exists in the Jabal Dokhan area, it became clear during our recent field campaigns that most of the “volcanic rocks” exposed in Wadi Um Sidri and Wadi Abu Ma’amel do not represent subaerial lava flows, but in fact are formed by shallow intrusion of magma. Presumably, this took place in at least two different batches, as indicated by structural evidence (see below). Consequently, we will avoid the term “lava” in the following description. Samples of different types of subvolcanic porphyries were collected from outcrops in Wadi Umm Sidri, from the northern entrance of Wadi Abu Ma’amel and from the Roman quarries. Sample localities are shown in Fig. 2.

Outcrop situation and state of rock preservation

Due to the present, hyperarid climatic conditions, blocks of dimension stone quarried by the Romans more than 1500 years ago are still absolutely fresh, as shown by traces of stone working (Figs. S2, S3). In contrast, older fractured rock surfaces are covered by a dark patina (Fig. 3d), the desert varnish. Accordingly, the structural features, characteristic of different rock types, are much better seen in objects of art, rather than in outcrop. Moreover, weathering under more humid conditions during the Pleistocene has produced a light to dark brown weathering crust, 1–5 mm thick, on surfaces of joints. Rock faces along the wadi slopes are covered by extensive drift sheets. These consist of sand, grit, weathered rock fragments or quarried freestone blocks up to 500 kg in weight, thus documenting the force of erosion produced by occasional rain storms, the last of which happened in 1999 and 2014.

Fig. 3
figure 3

Field relations exposed in Wadi Umm Sidri and Wadi Abu Ma’amel, Jabal Dokhan area: a Branching aplite dike in dark-colored aphyric rock; Wadi Umm Sidri; N27°18′15.6″ E33°19′11.7″. b Purple veinlets rich in Mn-epidote in deep reddish brown porphyritic Imperial Porphyry with white plagioclase phenocrysts. Northwest quarries (coin diameter = 2.5 cm). c Polished sample of subvolcanic rocks of Imperial Porphyry characterized by plagioclase phenocrysts altered to pink epidote/piemontite (Width of view is 8 cm). d Transitional contact between dark greenish gray Common Porphyry (left) and purple colored Imperial Porphyry (right) within a few centimeters. Lepsius quarries (coin diameter = 2.5 cm). e Greenish gray intrusive breccia of Common Porphyry, consisting of fragments up to c. 50 cm in diameter in a matrix of slightly darker color; Lepsius quarries; length of the hammer 40 cm. f Deep purple intrusive breccia of Imperial Porphyry. The fragments are purple and richer in phenocrysts of white feldspar than the matrix. Lykabettus quarries

Full size image

Rock types and field relationships

The great majority of the igneous rocks in the Wadi Umm Sidri and Wadi Abu Ma’amel area is represented by massive subvolcanic rocks, in part with brecciated structure (Figs. 1, 3e, f and Figs. S4b, d, S5c, e). The predominant Common Porphyry is a dark green colored porphyritic rock that contains phenocrysts of white, rarely pale pink, plagioclase and dark amphibole in a greenish gray to dark greenish gray matrix. Rarely, the larger crystals are aligned to form flow banding. Some rock portions contain amygdales, filled with secondary minerals, especially pistachio-green epidote and calcite. Rocks with aphyric texture are rare. They are exposed in northern Wadi Abu Ma’amel and Wadi Um Sidri, where they are intruded by a branching aplite dike (Fig. 3a).

The Roman quarries are the only ancient source of Imperial Porphyry. This conspicuous variety is distinguished, from the Common Porphyry, by its magnificent purple color of its matrix, whereas plagioclase phenocrysts are white or pale pink (Fig. 3b, c). Both rock types adjoin each other with irregular and diffuse contacts. The width of the gradational transition zone differs from a few centimeters (Fig. 3d) to several meters (Fig. S7a). Judging from evidence in outcrop and polished faces on works of art and architecture, most of the Imperial Porphyry displays a brecciated structure, showing different types of angular or rounded clasts in a darker and finer-grained matrix. Poorly defined angular to slightly rounded clasts, up to 50 cm in size, rich in feldspar phenocrysts, are set with relatively sharp contact in a matrix that is poorer in feldspar phenocrysts (Fig. 3e, f). Moreover, the Imperial Porphyry is intruded by hydrothermal veinlets of deep purple color, up to 1 cm wide (Fig. 3b). These facts suggest that the matrix of the Imperial Porphyry got its characteristic purple color by a later, pervasive overprint, performed by hydrothermal solutions (see also Makovicky et al. 2016a, b). At the climax of this process, even the plagioclase phenocrysts attain a pale pink color.

Although the Imperial Porphyry is by far best exposed in the Roman quarries in the upper parts of the subvolcanic intrusions, it also occurs down to the lower western faces of Wadi Abu Ma’amel, resting on top of a granitoid intrusion (Fig. S7b; see below). However, in these porphyry occurrences, a narrow-spaced joint system prevented quarrying of big blocks. In total, the subvolcanic sequence must have attained a minimum thickness of 600 m, presumably up to 900 m.

The Dokhan subvolcanics are intruded, with sharp contact, by coarse-grained, light colored quartz-bearing syenite, characterized by spectacular orthogonal jointing (Fig. S7c). Xenoliths of dark subvolcanic rock (Fig. S7d) clearly indicate that the syenite intrusion postdates the subvolcanic activity. In the southern part of the study area, the bulk syenite is coarse-grained, due to an intrusion depth of 600–900 m. A porphyritic structure has been attained only close to the contact with the subvolcanic country rock, where the syenite is more rapidly cooled. Moreover, a porphyritic structure is observed also in Wadi Um Sidri in the northern part of the study area, where the syenite has intruded a porphyry of common type, obviously at a shallower level. No additional alteration zones are visible in the immediately adjacent country rocks of Common or Imperial Porphyry type.

In the study area, the syenite and the Common Porphyry are dissected, with sharp contacts, by steeply inclined, approximately N–S trending mafic dikes, 0.7–1.5 m, rarely up to 5 m in thickness (Fig. 2). These fine-grained dikes, best exposed near the northeastern water well in Wadi Abu Ma’amel, are reddish brown and rich in needle-shaped amphibole. Swarms of NE- to ENE-trending aplitic dikes are more common in the areas of Imperial Porphyry exposures (Fig. 2). Whereas, due to the hydrothermal activity, the plutonic and subvolcanic rocks described have been subjected, along fissures, to alteration in deep reddish brown colors, the younger basic and acidic dike rocks were not affected by this process. A more detailed description of individual quarries is given in the electronic supplement.

Microscopic characteristics

The fine grain size of the matrix and the high degree of secondary alteration does not allow reliable modal analyses of the rock samples. Consequently, their classification is based on the bulk rock analyses. The mineral content for selected samples, based on microscopic observations and X-ray powder diffractometry, is listed in Table S1. Results of representative microprobe analyses are summarized in Table S2.

Purple Imperial Porphyry type

These rocks exhibit a well-developed porphyritic texture with phenocrysts of plagioclase occurring as single crystals or aggregates (Fig. 4a, c), but heavily altered into fine-grained aggregates of sericite + albite + quartz + hematite ± calcite ± pink Mn-bearing epidote ± pink piemontite. In addition, phenocrysts of amphibole occur in glomerophyric aggregates (Fig. 4b). They are largely altered to fine-grained, hematite-rich opacite (Fig. 4a), also containing pink Mn-bearing epidote, pink piemontite (Fig. 4c), dark mica, chlorite, tremolite and (?) talc. Sporadically, relics of the original brown amphibole are observed (Fig. 4c). The matrix of the Imperial Porphyry is extremely fine-grained and consists predominantly of albite + K-feldspar + quartz + “sericite” ± Mn-bearing pink epidote ± pink piemontite ± titanite ± apatite (Fig. 4a, b). Besides ilmenite, hematite is the clearly predominant ore mineral and occurs as tiny tablets visible in polished sections and also confirmed by X-ray powder diffraction.

Fig. 4
figure 4

Photomicrographs of Imperial and Common Porphyry: a Phenocrysts of plagioclase, highly altered to form pale pink Mn-epidote and albite, and amphibole partially to totally altered to opacite, set in a fine-grained matrix consisting of quartz, feldspars, platelets of hematite, and accessory tourmaline (upper right corner); sample D12-22, Northwest quarries; 1 polar. b Phenocrysts of twinned amphibole and sericitized plagioclase in a fine-grained matrix of quartz, feldspar and hematite; sample D12-43, Lykabettus quarries; crossed polars. c Phenocrysts of brown hornblende, partly altered to opacite and Mn-epidote in association with those of highly altered plagioclase set in fine-grained quartzo-feldspathic, hematite-bearing matrix; sample D12-42, Lykabettus quarries; 1 polar. d Newly grown, distinctly pleochroic crystals of piemontite in fine-grained matrix; sample D12-24, Northwest quarries; 1 polar. e Porphyritic subvolcanic rock of Common Porphyry showing phenocrysts of altered twinned plagioclase and vugs filled with epidote set in fine-grained matrix of plagioclase, quartz, biotite, epidote and Fe oxides; sample D12-12, crossed polars. f Common Porphyry with brecciated texture, containing porphyritic sub-rounded clasts in a matrix consisting of albite, epidote, sericite, biotite, chlorite, and opaques; sample D12-31, Northeast (Bradford) quarries; crossed polars

Full size image

Dark green colored common porphyry type

In the common porphyry, phenocrysts of plagioclase are white to pale green in color and are largely altered to aggregates of colorless, pale pink or yellowish green epidote + sericite ± calcite + opaques. Phenocrysts of amphibole occur in few samples only and are largely altered to chlorite + phlogopite ± talc ± bluish-green or colorless amphibole ± colorless or pale pink epidote; only few relics of initial brown amphibole are preserved. The matrix consists of very fine-grained aggregates of albite, quartz and epidote. The distinctly predominant ore mineral is ilmenite, commonly with minor hematite lamellae, whereas (titano-)magnetite is rare. At the contact zones between Imperial Porphyry and Common Porphyry, a transitional type with reddish brown matrix occurs, but with the same petrographic composition as the Common Porphyry.

Bulk rock composition

Geochemical analyses of major oxides were carried out by X-ray Fluorescence (XRF) on 35 representative samples, 16 of the Common Porphyry (including brecciated samples) and 19 of the Imperial Porphyry (including the archeological sample WS-E9). From these, 21 selected samples were analyzed for trace elements and REE by laser ablation inductively coupled plasma-mass spectrometry (LA-ICP-MS) (Table S3). For analytical methods, see electronic supplement.

Samples collected in Wadi Umm Sidri revealed significant variations in the concentrations of both major and trace elements, whereas in samples from the other localities, contents of trace elements are nearly identical, and only some of the major elements show certain variability (Table S3).

In the TAS diagram (Na2O + K2O) vs. SiO2 (Cox et al. 1979), all data points form a relatively wide cluster overlapping the fields of trachyandesite, andesite and dacite (Fig. 5a). In detail, most data points of the Common Porphyry plot into the dacite field, whereas most points of the Imperial Porphyry plot into the trachyandesite field or straddle the border to the dacite field. However, an andesite composition is indicated for nearly all samples using the Al–(Fe + Ti)–Mg cation plot of Jensen (1976) or discriminations based on relatively immobile trace elements such as the (Zr/TiO2) vs. (Nb/Y) diagram of Winchester and Floyd (1976), (Fig. 5b). All samples are characterized by Ni contents of 16–49 ppm and Mg# (100MgO/(MgO + FeOtot)) ratios of 28–37, distinctly lower than the composition of “primitive” rocks with Ni contents between 235 and 400 ppm (Sato 1977) and Mg# between 63 and 73 (Green 1971). This composition is far from that of primary mantle-derived magmas indicating substantial fractionation of olivine and/or pyroxenes from the initial magma composition.

Fig. 5
figure 5

Geochemical classification of subvolcanic rocks from Wadi Umm Sidri—Wadi Abu Ma’amel area, Eastern Desert of Egypt: a Total alkalis–silica diagram (Cox et al. 1979); discrimination between alkaline and subalkaline after Irvine and Baragar (1971); b selected enlarged part of Nb/Y–Zr/TiO2 diagram (Winchester and Floyd 1976)

Full size image

Within the TAS diagram (Fig. 5a), but also in diagrams using more immobile elements such as Zr, Ti, Nb, Y (+ Si) (Winchester and Floyd 1976, 1977), the samples display subalkaline affinity. In the diagram K2O vs. SiO2 of Peccerillo and Taylor (1976) (Fig. S8a), the data points scatter within the fields of the calc-alkaline series, clustering around the boundary between medium-K and high-K series (Rickwood 1989). Calc-alkaline affinity is clearly supported by further discriminations, e.g., the diagram FeOtot/MgO vs. SiO2 (Fig. S8b) after Miyashiro (1974), the Jensen cation plot (Jensen 1976), and the Th–Hf–Ta–Zr–Nb triangles of Wood (1980). An affiliation to low-Fe series is verified by the discrimination of Arculus (2003) in the FeOtot/MgO vs. SiO2 diagram (Fig. S8b).

The chondrite-normalized REE patterns show similar slopes that decrease from La to Ho and are almost flat from Er to Lu. There are no conspicuous differences between the Common Porphyry and the Imperial Porphyry (Fig. 6a, b). Although sample D12-05 shows the same trend, the enrichment factors are distinctly lower than for all other samples. In general, the REE patterns may indicate minor amphibole fractionation. In contrast to the majority of samples, which do not show a notable Eu anomaly, small negative anomalies with EuN/EuN* of 0.83 and 0.76 in samples D12-08 and D12-29 point to subordinate plagioclase fractionation (Fig. 6a).

Fig. 6
figure 6

Cl chondrite-normalized REE patterns for a the Common Porphyry type and b the Imperial Porphyry type using the normalization values of Sun and McDonough (1989). MORB-normalized trace element patterns for c the Common Porphyry and d the Imperial Porphyry using the normalization values of Sun and McDonough (1989) and Pearce and Parkinson (1993). The interpolated and extrapolated line of lowest enrichment factor (elements Nb, Zr, Ti, Y, Yb) is used to estimate the concentrations of other elements in the mantle wedge before modification during subduction (Pearce and Parkinson 1993)

Full size image

The MORB-normalized multi-element variation diagrams (Pearce and Parkinson 1993) are characterized by positive anomalies of the LIL elements Rb, Ba and K, the HFS elements Th, U and Pb as well as Al and Ga, whereas the HFS elements Ta, Nb, Ti, but also V display negative anomalies (Fig. 6c, d). Particularly the Ta–Nb anomaly underscores the calc-alkaline character of the samples.

U–Pb zircon dating

Two Imperial Porphyry samples (D12-24 and D12-36) and one Common Porphyry sample (D12-30) have been dated using SIMS (Secondary Ion Mass Spectrometry). Zircons were separated after crushing of the samples, using heavy liquids and a magnetic separator, then were handpicked under a binocular microscope to obtain the most transparent and inclusion-free grains. Microphotography shows euhedral crystals of zircon with sizes varying between 100 and 200 µm. Their internal zoning (Fig. 7) was detected by cathodoluminescence (CL) using a scanning electron microscope. Ion microprobe U–Th–Pb analyses were performed using a CAMECA IMS-1280 high resolution ion-microprobe at the Swedish Museum of Natural History in Stockholm (Nordsim facility). Details of instrument parameters and basic analytical techniques and data reduction have been given by Whitehouse et al. (1999) and Whitehouse and Kamber (2005). The zircon U–Pb isotope data are presented in Table S4. The analytical data are plotted as two-sigma error ellipses on Concordia diagrams (Tera and Wasserburg 1972), and the calculations were done using the Isoplot routines of Ludwig (2001).

Fig. 7
figure 7

Representative CL images with 206Pb/238U ages (in Ma) of zircon grains from the analyzed Imperial and Common Porphyry samples

Full size image

Sample D12-24 (27°14′58.68″N, 33°16′28.71″E) zircons are subhedral to euhedral and yellow to pale brown, and exhibit well-preserved oscillatory growth zoning (Fig. 7). A total of 15 measurements were carried out on 15 zircons (Table S4; Fig. 8a). The U content varies from 19 to 241 ppm, Th from 11 to 175 ppm, and the Th/U ratios are high (0.40–1.34), as expected for magmatic zircons (Corfu et al. 2003). In general, out of 15 zircon grains analyzed, 10 yielded a concordia age of 609 ± 4 Ma [2σ] (Fig. 8a) based on a group of 10 concordant and equivalent analyses (MSWD = 1.7). The rejected analyses were excluded from age calculation because three analyses are discordant (zr-4, zr-6 and zr-10), one analysis (zr-3) has a low U content (19 ppm) and shows a large uncertainty (± 55 Ma) [2σ] (Table S4). One spot (zr-14) is a concordant grain and yields an older 206Pb/238U age (713 ± 5 Ma) [2σ] than the other analyses, which is interpreted to represent an inherited zircon grain (Table S4).

Fig. 8
figure 8

Inverse (Tera–Wasserburg) Concordia diagrams for SIMS zircon U–Pb data from Imperial and Common Porphyry samples

Full size image

Sample D12-36 (27°14′51.81″N, 33°18′25.89″E) zircons are euhedral (100–200 µm) and yellow to pale brown. Cathodoluminescence (CL) images (Fig. 7) show well-developed zoning as expected for magmatic zircons. One measurement was carried out on each of the 15 grains (Table S4). Their U contents vary from 56 to 167 ppm, Th contents from 22 to 137 ppm and Th/U ratios from 0.16 to 0.97 (Table S4). 12 analyses cluster tightly, defining a concordia age of 600.4 ± 2.4 Ma (2σ; MSWD = 0.96; Fig. 8b).Two spots (zr-7 and zr-11) show reverse discordance and one spot (zr-1) is a concordant grain and yields an older 206Pb/238U age (623 ± 4 Ma), perhaps reflecting a xenocryst derived from older material (Fig. 8b).

Sample D12-30 (27°16′2.57″N, 33°18′5.04″E) zircons are euhedral (100–200 µm) and yellow to pale brown. One measurement was carried out on each of the seven grains (Table S4). U contents vary from 101 to 212 ppm, Th contents from 26 to 117 ppm and Th/U ratios from 0.24 to 0.74. Four analyses cluster tightly, defining a concordia age of 608 ± 6 Ma (95% confidence, MSWD = 1.9; Fig. 8c).One spot (zr-4) is discordant and distinctly younger than the other six analyses. Two spots (zr-1 and zr-2) are concordant grains and yield older 206Pb/238U ages of 786 ± 4 [2σ] and 828 ± 4 Ma [2σ], respectively, perhaps representing xenocrysts derived from older material (Fig. 8c).

Discussion

Geotectonic setting and petrogenesis

All samples show similar rare earth and trace element patterns (Fig. 6), thus indicating that they are derived from a similar source and/or have crystallized from a common parental magma. Based on their subduction-related trace element signature and the negative anomalies of certain HFS elements such as Nb, Ta and Ti, we conclude that the porphyritic andesite to dacite, both of Common and Imperial Porphyry types, were produced by partial melting of a prior subduction-modified upper mantle, either within an active subduction zone or a later transtensional tectonic regime (e.g., Johnson et al. 2011). Such transtensional tectonic regime is confirmed from the Th/Yb vs. Ta/Yb ratios plot (Fig. S9), where the Common and Imperial porphyries straddle the within plate and active continental margin geotectonic fields (Gorton and Schandl 2000). Lower NbN/LaN ratios (0.31–0.40, average 0.35) and Zr/Ti ratios of the subvolcanic samples (0.02–0.05, average 0.03) further confirm the arc settings (Pearce 1982; Pearce and Peate 1995; John et al. 2004; Pearce and Stern 2006). Higher Th/Yb ratios of the subvolcanic samples investigated (2.1–6.6, average 3.1) point to continental arc setting, where Condie and Kröner (2013) used a ratio of 0.5 for basalts to discriminate between oceanic and continental arcs. Positive Pb and Ba anomalies may be attributed to a Pb- and Ba-rich crustal component in the mantle source, such as subducted pelagic (meta-)sediments or interaction with granitic magmas. This conforms to the model, evaluated for the formation of the Dokhan magmas by Stern et al. (1988) and Moghazi (2003).

Field evidence, such as the regional distribution of the Hammamat sediments and the Dokhan volcanic suite in broad shallow basins (e.g., Eliwa et al. 2010; Bühler et al. 2014), as well as the ENE–WSW trending dike swarms, indicates that the magmas were emplaced into a N-to-NW-directed extensional crustal regime, at least between c. 600 and 575 Ma (Stern et al. 1988, and references therein). Such a geotectonic position testifies to a highly extended terrane (HET) as originally characterized by Olsen and Morgan (1995). This is underscored by the calc-alkaline, subduction-related geochemical signature of the igneous rocks in the Eastern Desert of Egypt, which is similar to the “type area” for HET, the Basin and Range Province (Zoback et al. 1981, and references therein, cf.; Wang and Shu 2012). Occasionally, HET syn-subduction or post-subduction extensional regime was misleadingly interpreted to be a rift or having a rifting phase (e.g., Burke 1978), a term also applied to the Dokhan volcanic suite by Moghazi (2003). However, the term rift is currently used synonymously to continental rift (Olsen and Morgan 1995), a tectonic process associated with magmatism of SiO2-undersaturated basalt to trachyte as well as nephelinite–carbonatite composition (e.g., Wilson 1989). The proposed geotectonic model is similar to the model assuming an extensional regime after crustal thickening by Stern et al. (1984, 1988), Stern and Gottfried (1986); Mohamed et al. (2000), and it is distinct from the compressional subduction regime model of Abdel-Rahman (1996) and the model of El Gaby et al. (1989) for the formation of the Dokhan magmas.

Intrusion mechanism

Based on their erosion forms with morphological steps as typically found in lava plateaus, the Common and Imperial Porphyry were interpreted as extrusive lava flows and pyroclastic deposits during former studies (Mohamed et al. 2000; Wilde and Youssef 2000; Moghazi 2003; Eliwa et al. 2006). However, Horseman et al. (2005, 2010), Gudmundsson (2012), and Menand (2011) present and illustrate a type of repeated magma emplacement in subvolcanic magma chambers, in which the individual magma pulses form sheets of nearly constant thickness within multiple sills. In our new interpretation, the occurrences of Common and Imperial Porphyry in the Wadi Umm-Sidri and Wadi Abu Ma’amel area most likely were formed by such a subvolcanic intrusion mechanism. The parental magma was derived from a mantle source and stored in a magma chamber in the lower crust, where it underwent minor lower crustal contamination and minor fractional crystallization of plagioclase and amphiboles. During an age interval between 609 and 600 Ma, this magma was then emplaced as subvolcanic magma pulses in a transtensional tectonic regime to form multiple sills beneath earlier erupted parts of the Dokhan Volcanic suite. The latter are dated to a volcanic period between 628 and 615 Ma (Breitkreuz et al. 2010) and occur just to the north of the study area.

Interpretation of the U–Pb zircon data

U–Pb SIMS dating on single zircons yielded ages of 600 ± 2 and 609 ± 4 Ma for two samples of Imperial Porphyry and 608 ± 6 Ma for a sample of Common Porphyry, all interpreted as crystallization ages. The age difference between the two Imperial Porphyry samples implies that these may have derived from two different sills of somewhat older and younger age, respectively.

Our age data are consistent with the SHRIMP U–Pb zircon data of 602 ± 9 and 593 ± 13 Ma of the Imperial Porphyry (Wilde and Youssef 2000). However, U–Pb SHRIMP ages obtained for zircons from other localities of the Dokhan volcanics cover a much wider range of 630–580 Ma (Wilde and Youssef 2000; Breitkreuz et al. 2010), broadly overlapping with the ranges of the Rb–Sr whole rock ages for volcanic rocks (610–560 Ma) and granites (610–550 Ma). There is no doubt that igneous activity took place, towards the end of the Pan-African Orogeny, in a dynamic setting around isolated volcanic centers and basin systems with different structural controls and different ages (Breitkreuz et al. 2010; Johnson et al. 2011). Contemporaneous to the intrusion of the post-collisional, calc-alkaline I- to A-type Younger Granites is the last phase of regional metamorphism of the crystalline basement in the Sinai Peninsula (Abu El-Enen and Whitehouse 2013) and the Eastern Desert (Abu El-Enen et al. 2016).

The coloring of the Imperial Porphyry

One of the most interesting features of the Imperial Porphyry is the presence of pink piemontite (Fig. 4d) and pale pink, Mn-bearing epidote. Both minerals either replace plagioclase phenocrysts as granular aggregates, giving the plagioclase a patchy color distribution from pure white to pale pink (Fig. 3b, c), or they form coarser grained aggregates, which occur as isolated patches, irregularly distributed in the fine-grained matrix (Fig. 4d). Some authors assume that the presence of piemontite is responsible for the impressive purple color of the Imperial Porphyry as a whole (Ghobrial and Lofti 1967; Basta et al. 1978; Klemm and Klemm 2001b). The main reason for the purple color of the Imperial Porphyry, however, is a dust of tiny hematite tablets, confirmed by X-ray diffraction that are widely and regularly distributed in the matrix of the rock. Conversely, the dark grayish-green Common Porphyry rather contains fine-grained magnetite/ilmenite, whereas hematite is subordinate. Such an interpretation was already presented by Dardir and Abu Zeid (1972) who have stressed that Mn-epidote is found in the common, green varieties of the Dokhan volcanics as well. On the other hand, Makovicky et al. (2016a) assigned the red coloration of the Imperial Porphyry to both the presence of piemontite and hematite dusting.

The transformation of the Common Porphyry to the Imperial Porphyry type can be explained by a pervasive hydrothermal-metasomatic overprint, leading to a widespread formation of hematite in the rock matrix. Such a process is also indicated by veins filled with purple-colored piemontite (width up to 1 cm, Fig. 3b), spreading out from fissures into the adjacent rock and thus forming a metasomatic front. The replacement of plagioclase crystals by Mn-epidote and piemontite may testify to an intensified metasomatic process.

Gradational contacts between the Common and Imperial Porphyries and the distribution of the Imperial Porphyry exposures along ENE- and NNW-trends (Fig. 2), parallel to the nearby Younger Granites indicate that the hydrothermal alteration around Mons Pophyrites (Fig. 2) can be related to the intrusion of the Younger Granites (e.g., Basta et al. 1978). These could have acted as a heat source for transforming meteoric water into hydrothermal fluid, which preferably moved along joints and faults. An interesting model has been proposed by Makovicky et al. (2016a,b) who relate the coloring to the increasing oxidation potential of the Earth’s atmosphere during the Second Great Oxidation event in late Neoproterozoic times (e.g., Frei et al. 2009; Campbell and Squire 2010). However, the gradual transition from the subvolcanic rocks of the predominant Common Porphyry type into the Imperial Porphyry type clearly testifies to a process of local rather than of global scale.

Conclusions

Based on field evidence, petrographic and geochemical characteristics, as well as on the U–Pb zircon SIMS ages, we propose the following model for the evolution of the Dokhan subvolcanic rocks of both, Common and Imperial Porphyry type:

  • Stage 1 An andesitic magma was formed by partial melting of a subduction-modified upper mantle.

  • Stage 2 The andesitic magma underwent differentiation towards dacite to trachyandesite composition by fractionation of minor amphibole and, if any, of minor plagioclase and, in addition, was modified by minor contamination with subducted sediments and/or admixture of granitic magmas.

  • Stage 3 At about 610–600 Ma, these compositionally modified magmas intruded into a highly extended terrane (HET) within a transtensional tectonic regime, thus forming multiple sills of nearly constant thickness. The presence of repeated magma pulses is indicated by the brecciated structure, visible in the field and on many objects of art and architecture.

  • Stage 4 Epidote, formed by propylitic hydrothermal alteration, produced the dark gray to greenish gray color of the matrix in subvolcanic rocks of Common Porphyry type.

  • Stage 5 Due to pervasive hydrothermal alteration under oxidizing conditions, fine-grained flakes of hematite were formed in the amphibole phenocrysts and the matrix of the Imperial Porphyry, thus producing its spectacular purple color. Between the two rock types, gradational transitions exist.

  • Stage 6 Ongoing pervasive hydrothermal alteration, combined with emplacement of hydrothermal veins, locally overprinted the Imperial Porphyry that was formed during stage 5. A more pronounced Mn-mobilization yielded Mn-epidote and piemontite that fill fissures and the cores of plagioclase (Fig. 4c).

The metasomatic alteration of stages 4–6 was caused by hydrothermal fluids, presumably formed by heating of meteoric water, whereby the contemporaneous Younger Granite sills may have acted as a heat source.