Mineralogy and geochemistry of pozzolans from the Tombel Plain, Bamileke Plateau, and Noun Plain monogenetic volcanoes in the central part of the Cameroon Volcanic Line

Mbowou Ngantche Igor Fulbert · Owona Se´bastien· Chako Tchamabe Boris ·Lissom Justin · Lanson Bruno · Ekodeck Georges Emmanuel

Abstract Pozzolans from the Tombel Plain, Bamileke Plateau, and Noun Plain, 3 monogenetic volcanic fields in the central part of the Cameroon Volcanic Line (the Tombel Plain, Bamileke Plateau, and Noun Plain), were explored in order to constrain their petrology and make some predictions on their pozzolanicity. The rocks in this study include alkaline and subalkaline basalts, trachybasalts, and basanites. Most of these rocks present an overall composition that overlaps with primitive mantle,suggesting rapid ascent of magmas, limited crustal contamination and crystal fractionation of olivine, clinopyroxene, and feldspar. The pozzolans present enrichment of LREE relative to HREE and high chondrite normalized ratios of La/Yb and Tb/Yb, ranging between 7 and 20 and ＞1.9 respectively, similar to those of Ocean Island Basalts. Like other nearby volcanoes, partial melting in a dominantly garnet-bearing mantle zone can be assumed.Quantitative mineralogy by X-ray diffraction revealed various mineral phases with dominantly plagioclase,augite, olivine, and Fe–Ti oxides. The samples contains important amorphous phase up to 23, 51, and 69 wt% in the Tombel Plain, Noun Plain, and Bamileke Plateau,respectively. This elevated amount of amorphous phases together with the sum of SiO2, Al2O3, and total Fe2O3(SAI = 68.50–83.50 ＞70 wt%)according to ASTM C 618 standard and the sum of CaO, FeO, and MgO (CIM =14.5–30.52 wt% and 23.58–31.08 wt%) suggest interesting pozzolanicity character for the studied pozzolans.

Keywords Pozzolans · Mineralogy · Geochemistry ·Pozzolanicity · Monogenetic volcanic fields · Cameroon Volcanic Line

1 Introduction

Pozzolans are unconsolidated pyroclastic rocks or tephra,resulting from ‘‘dry’’ explosive magmatism and consist mostly of juvenile components such as scoria, free crystal,and volcanic glass that vary in size, color, shape, and texture (Morrissey et al. 2000). Pozzolans have been identified in various locations along the Cameroon Volcanic Line (CVL) notably at volcanic cones in monogenetic fields in the Tombel Plain (Nkouathio 2006;Nkouathio et al.2008;Tchamdjou et al.2017),Baleng lake within Bamileke plateau and in Noun Plain (Wotchoko et al. 2005).

Fig. 1 a Location of Cameroon in Africa.b The Cameroon Volcanic Line(CVL)and c sampling map of pyroclastic rocks(pozzolans)along the central part of CVL

Monogenetic volcanoes of the CVL display incomplete magmatic suites limited to basanites,basalts,and accessory hawaiites while polygenetic volcanoes are characterized by a complete magmatic series (Sato et al. 1990; De´ruelle et al. 2007; Chako Tchamabe 2014; Tiabou et al. 2019).Monogenetic volcanoes in the world are known to be geochemically and volcanologically simpler (Brenna et al.2010)and smaller in terms of volume of magma erupted as compared to polygenetic volcanoes. Heterogeneity of the mantle under some volcanoes has been pointed out as the source of slight differences, notably on MgO content,observed on the composition of the lavas belonging to monogenetic volcanic cones or field. This has been reported for example at the Crater Hill volcano in New Zealand (Smith et al. 2008), the Udo tuff in South Korea(Brenna et al. 2010), the Parı`cutin in the Michoaca´n-Guanajuato volcanic field in Mexico (Erlund et al. 2010), the Auckland volcanic field in New Zealand (McGee et al.2012, 2013), the Mt Gambier Complex in Australia (Van Otterloo et al. 2014), the Barombi Mbo Maar complex(Chako Tchamabe 2014), the Debunscha Maar complex(Ngwa et al. 2017), and the Baossi—Warack volcanoes(Tiabou et al.2019)in Cameroon(Central Africa).Some of these compositional differences result from successive partial melting of distinct but contiguous source with varying composition(McGee et al.2012,2013;McGee and Smith 2016),from polycyclic activity of the volcanoes that will enable deposition, through the same vent, of lavas originated from distinct magmas derived from mantle zones with slightly different percentage of garnet (Chako Tchamabe 2014) or, most of time, lavas resulting from mixing of two or many batches of magma of different composition(Erlund et al.2010,Otterloo et al.2014;Ngwa et al. 2017; Tiabou et al. 2019).

Numerous studies done on monogenetic volcanoes have shown that products from their eruption are lava flows and pyroclastic materials from a strombolian eruptive style as well as phreatomagmatic deposits (Nemeth et al. 2003;Martin and Nemeth 2006;Valentine et al.2007;Kereszturi et al. 2011; Nemeth and Kereszturi 2015).

Most of the petrogenetic studies done, on monogenetic volcanoes along the CVL have focussed mostly on lava flow and cinder cones(Tiabou et al.2015,2019;Sato et al.1990) and maar diatreme volcanoes such as Nyos (Aka et al. 2004, 2008; Hasegawa et al. 2019), Barombi Kotto lake (Tamen et al. 2007; Asaah et al. 2020), Lakes Elum,Wum and Oku (Asaah et al. 2015), and Debunscha (Ngwa et al. 2017). Little attention has been paid to the geochemistry of these pyroclastic materials, at least in some volcanic Plateau (Wandji et al. 2000, Wotchoko et al.2005). However, nothing has been done on the pozzolanic monogenetic cones identified so far along the line.Furthermore, no study has been carried out to point out possible petrological links amongst these pozzolanic cones with nearby volcanoes.

Fig. 2 Photographs showing some outcrops of the studied pozzolans.a Reddish to medium brown stratified deposits in the Tombel Plain; b Reddish to medium brown and Black pozzolan deposit surrounding Baleng Crater Lake in the Bamileke Plateau, c deposits showing an alternating beds of fine-grained ‘‘reddish’’ (dark brown) and coarse-grained black pozzolan at Foumbot quarry (Noun Plain); d Uniform black pozzolan in front of Petpounoun lake (Noun Plain);e Dark yellow—colored pozzolans materials that constitute a soil of Monoun village (Noun Plain) and follows topography

Most of these cones, especially those in the Tombel graben, like the Djoungo volcanic cone, are currently exploited,and their pozzolans used as an additive to clinker in the manufacturing of pozzolanic cement in local cement factories(Billong et al.2013;Tchamdjou et al.2017).The pozzolanicity or pozzolanic activity of natural pozzolan refers to both its capacity of binding lime and the rate at which the binding reaction takes place(Pichon 1994;Segui 2011;Walker and Pavı´a 2011).This characteristic depends on the chemical and mineralogical composition of the pozzolan, the type and proportion of its reactive silica content, and the particle’s specific surface area (Pichon 1994;Rodriguez-Camacho and Uribe-Afif 2002;Massazza 2007; Segui 2011).

The aim of this study is to perform a geochemical study while bringing new mineralogical data on pozzolanic outcrops located in the Tombel Plain, Bamileke Plateau, and Noun Plain, three volcanic provinces of the central part of the CVL. X-Ray Diffraction (XRD) analysis allowed making a better comparison and differentiation of the samples by evaluating their mineral (crystal) and nonmineral (amorphous) components. The chemistry of primitive melts from pozzolan clasts helped to understand their petrogenesis as well as making predictions on their pozzolanicity based on ASTM C 618 standard. The comparison of the chemistry of pozzolans with those of nearby volcanoes will help to complete data record on composition of mantle source of magma within the central parts of CVL.

Fig. 3 Pozzolans in hand specimen showing variability in grain sizes and color.a Vesiculated medium brown lapilli-ash(Lim52);b vesiculated medium brown lapilli-block-ash (TMC17); c vesiculated dark green–brown lapilli-ash (BL32); d highly—vesiculated black lapilli-block-ash(Lim53); e dark green-grey ash-lapilli (Bki48); f highly—vesiculated dark grey lapilli-block (TMC18); g highly—vesiculated medium olive lapilli-ash pozzolan (MV38);h medium olive ash-lapilli (ML44); i highly—vesiculated dark grey lapilli-ash (MB46). Colors are in accordance with color chart exhibit in Geo—explore store (www.geoexplorestore.co.za)

Fig. 4 Scanning Electron Microprobe(SEM)images of the prototype studied pozzolans.a Sample Kpe´04 from Tombel plain;b Sample BL31 from Bamileke Plateau and c sample MB46 from Noun Plain. Note overall presence in samples from the three areas of vesicles that evidence degassing while cooling

2 Geological setting

The CVL previously called Cameroon Line (Cantagrel et al. 1978) and renamed as ‘‘Cameroon Hot Line’’ (CHL;De´ruelle et al.2007)defines a SW–NE alignment(trending 30°; Fig. 1) of volcanic massifs and extruded plutons extending on about 1600 km from Pagalu` Island in the Gulf of Guinea to Chad Lake (De´ruelle et al. 2007). It exhibits the unique worldwide example of an active intraplate volcanic alignment build onto both the oceanic and the continental plates (Fitton and Dunlop 1985; De´ruelle et al.2007).The oceanic part lies within the Gulf of Guinea and comprises the islands of Pagalu`, Sao Tome´, Principe,and Bioko(De´ruelle et al.2007).The continental sector is a succession of large polygenetic volcanoes separated by monogenetic volcanic fields, where Hawaiian, strombolian and moderate vulcanian to explosive phreatomagmatic eruptions constructed several short lavas flows,pyroclastic cones and maar volcanoes (Sato et al. 1990; Nkouathio et al.2008;Chako Tchamabe 2014;Ngwa et al.2017).The Tombel Plain, Bamileke Plateau, and Noun Plain in the southernmost continental part of the CVL host most of these pyroclastic cones(Figs. 1c,2).The Tombel plain,for example, displays series of Pleistocene (0.6–0.05 Ma)explosive events (Lee et al. 1994) with about 115 pyroclastics cones of various size (Nkouathio et al. 2006;Tamen et al.2007)among which mostly consist of black to reddish basaltic pozzolan cones. In the Noun plain, for example, about nine strombolian cones have been recognized as such (Wotchoko et al. 2005; Moundi et al. 2009).The tephra ring of Lake Baleng in the Bamileke Plateau was also evidenced as pozzolans (Wandji et al. 2000).Around the localities of Limbe and Batoke at the foot of Mt Cameroon, some volcanic cones are also made of pozzolans (Bardintzeff and McBirney 2000).

Volcanic rocks from CVL seem to be almost geochemically identical in both ocean and continental sectors.They are dominantly alkali-rich OIB,having similar trends in trace elements (Halliday et al. 1988, 1990; Lee et al.1994, De´ruelle et al. 2007; Asaah et al. 2015). Ballentine et al. (1997) defined a HIMU mantle as the origin of the entire CVL. Fitton and Dunlop (1985) added a metasomatized mantle as part of the source for CVL’s lavas. Asaah et al. (2015) provide in a recent review, complementary information on the petrogenetic evolution of the CVL highlighting mantle heterogeneity and distinct geochemical signatures even for individual volcanoes.These lavas form a magmatic series from picrites to rhyolites (Nkouathio et al.2008), with picrites and alkaline basalts linked to the Noun plain areas (Wotchoko et al. 2005). Lavas from the Tombel Plain, Bamileke Plateau, and Noun Plain provide an opportunity to discuss the petrogenetic evolution of the southernmost continental part of the CVL.

3 Materials and methods

3.1 Fieldwork and sampling

Brief descriptions of the deposits were made onsite and hand specimen collected for further laboratory analyses.Sampling was limited to outcrops, where pozzolans are currently being collected for industrials purposes. Figure 1c and Table 1 present the list of 37 bulk samples collected in the central part of the CVL for mineralogical and geochemical analyses. This includes 22 samples derived from 11 cones of Tombel Plain, 12 samples from five cones in the Noun Plain and three samples from a single cone studied on Bamileke Plateau.

Fig. 5 Quantitative mineralogy of Pozzolans. a Average distribution of mineral assemblage and amorphous of the studied pozzolans within the three volcanic fields and Prototypes of XRD patterns of samples b Kpe07 from Tombel plain; c BL32 from Bamileke plateau and d MV38 from Noun plain. See supplementary materials for others analyzed XRD patterns

3.2 Mineralogical analysis

Quantitative mineralogical analyses were performed at the X-Ray Diffractometry Laboratory of the Institut des Sciences de la Terre (ISTerre) in Grenoble, France, from X-ray diffraction patterns collected with a Bruker D5000 X-ray diffractometer equipped with a Sol-XE Si(Li)detector from Baltic Scientific Instruments and operated at 40 kV and 40 mA. Juvenile scoria fragments from natural pozzolans field samples carefully selected based on their abundance of vesicles in hand specimen and hand lens observation were first ground in an agate mortar to a powder of 0.315 mm diameter. One gram of this powder was mixed with 20–30 ml of ethanol for 10 min wet grinding in a McCrone micronizing mill containing 48 small cylindrical agate pestles (Hillier 2003). Ethanol was also used to clean the pestle and the mortar and to recover the slurry from the mill. That slurry was dried in an oven for 8 h and smoothly disaggregated using agate mortar.The dry powder was then prepared as a randomly oriented mount and data was recorded from 8 to 90° using 0.026°steps and 16 s counting time per step. XRD data interpretation was performed using the open-access software Profex (3.7.0), which is based on the Rietveld method(Doebelin and Kleeberg 2015), and the American Mineralogist Crystal Structure Database that contains XRD data for known minerals (www.rruff.geo.arizona.edu). Each sample was run twice using 400–800 mg of the powdered rock.The first pattern was used for phase identification and quantitative phase analysis of crystalline phases, whereas the second pattern,collected after addition of 20 wt%of an Al2O3(corundum) internal standard, was used to quantify the amorphous component of samples (Jenkins and Snyder 1996; Madsen et al. 2011; Raven and Self 2017). The rare occurrence and limited variety of clay minerals allowed their identification and quantification, using XRD patterns obtained from randomly oriented powders.

3.3 Major and trace elements analysis

Representative fresh coarse lapilli to bomb sizes pozzolans collected from the deposits were prepared and analyzed at the GEOLAB (Ontario, Canada) for their major and trace element abundances following the laboratory techniques and methods(Geoscience Laboratories 2015).Special care was taken to select only the freshest core/parts of the samples for the analyses.

Loss on ignition (LOI) was performed by calcinating about 20 g of the sample at 105 °C under a nitrogen atmosphere then, at 500 and 1000 °C under an oxygen atmosphere. The results were represented in percentage,and the reduction of weight noticed after calcination compared to the weight of the fresh sample.

Major elements were determined by X-Ray Fluorescence (XRF) by fusing 10 grams of the burned samples with a borate flux to produce a glass bead for the analysis.Trace elements were determined by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Sample preparation involved Agate Mill Preparation (SAM-AGM) techniques to reduce the size of rock particles to less than 90 μm(170 Mesh). The use of this method reduces the amount of contamination by Cr, Al, and Fe. Then, the Closed Vessel Multi-Acid Digest (SOL-CAIO) method (Geoscience Laboratories 2015) was used for the complete dissolution of the silicate rock sample.Finally,prepared solutions were analyzed by ICP-MS and ICP-AES to determine their composition in trace elements. Ferrous irons were quantified by potentiometric titration with a standardized permanganate solution to bring more accurate data in iron determination. For that purpose, samples were first dissolved in an aggressive non-oxidizing acid mix. On the other hand,measurements of geochemical standards(DUP:Laboratory Duplicate) were carried out to guarantee the quality and accuracy of the analyses.The CIPW norm was determined using the GCD kit (Janousek et al. 2006).

4 Results

4.1 Petrography

The pozzolan deposits consist of pyroclastic materials that slightly differed on outcrops by some physical characteristics such as grain size, color, and vesicle abundance(Table 1, Figs. 2, 3, 4). In the Tombel Plain and Bamileke Plateau, pozzolans deposits are exposed in the form of large cones, very often well-stratified and graded (Fig. 2a,b), while those in the Noun Plain appear either as small cones(Fig. 2c,d)or large sub-horizontal expanses without any apparent grading (Fig. 2e). In hand-samples, the pozzolans consist of a mixture of grains with sizes varying from blocks (φ ＞64 mm) to ash (φ ＜2 mm). Lapilli is the dominant component. The color of pozzolans also varies from brownish to darkish(Table 1,Fig. 3),and most samples show vesicles (Figs. 3a–c, 4).

?

Fig. 6 Total alkali versus silica (TAS) classification diagram (Le Bas et al. 1986) for the analyzed pozzolans compared to lavas of the nearby large volcanoes

4.2 Mineralogy

Table 2 and Fig. 5 present the percentages of minerals and amorphous phases of the analyzed samples. For commodity, the names of minerals have been abbreviated in the text, according to Kretz (1983) to show the mineral assemblage in samples. In almost all samples, plagioclase and augite are the most dominant mineral phases,followed by olivine and Fe–Ti oxides. K-feldspar and nepheline are scarce. Samples from Barombi Kotto Island, Mt Koupe,Penja, and Manjo in the Tombel Plain contain primarily labradorite (26–56 wt%), augite (10–54 wt%), olivine(5–15 wt%), ilmenite (0.5–8 wt%), magnesite (0.7 wt% in only one sample), smectite (42.3 wt% in one sample),franklinite 7.1 wt% (in only one sample), Ti-magnetite,muscovite, hematite, nepheline, and calcite. The mineralogical assemblage comprises Lab + Aug + Ol + Ti-Mag + Mag ± Qtz ± Musc ± Ne ± K-Fpr ± Hem ±Cal ± Mags ± Ilm ± Gib ± Gt ± Fkl. In the Bamileke Plateau, the mineralogical assemblage is Aug + Lab +Ol ± Hem ± Mag ± Musc ± Ne ± Ti-Mag ± Gt, with augite ranging between 38 and 40 wt%, labradorite between 20 and 25 wt%and olivine between 5 and 7 wt%.Ti-magnetite, nepheline, hematite, and goethite are accessories. In the Noun Plain, the mineralogical assemblage is composed of Aug + Lab + Ol + Mag ± Qtz ± Musc ±Hem ± Cal ± Ne ± Ti-Mag ± Ilm ± Fkl ± Fl-Ap ±Hem. Samples exhibit augite (12–46%), labradorite(9–21%), olivine (5–27%) as main phases, while Ti-magnetite (1–3 wt%), quartz (＜1%) nepheline, hematite,franklinite, fluorapatite, hematite, and ilmenite as scare phases.The amorphous phase contents vary between 0 and 50.8 wt% in the Tombel Plain (average of 15.4 wt%),between 5.2 and 81.2 wt% in the Bamileke Plateau (average of 36.6 wt%) and, between 1.0 and 69.1 wt% in the Noun Plain (average of 31.7 wt%).

4.3 Major elements geochemistry

Major element compositions and calculated CIPW norm of the studied pozzolans are presented in Table 3. The SiO2content ranges from 42.20 to 45.10 wt% in the Bamileke Plateau,between 44.71–55.02 and 43.21–57.38 wt%in the Tombel and the Noun Plains,respectively.The Al2O3content is quite constant (14.45–15.09 wt%) in the Bamileke Plateau in respect to the Tombel and Noun Plains where it varies between 12.08–19.44 and 11.99–15.83 wt%,respectively. The T-Fe2O3content is also constant(13.39–14.66 wt%)in the Bamileke Plateau compared to the Tombel and Noun Plains where it ranges between 10.85–14.25 and 9.56–14.00 wt%, respectively. The CaO content varies between 9.59 and 11.03 wt%in the BamilekePlateau, between 6.16–13.06 and 5.56–10.87 wt% in the Tombel and Noun Plains, respectively. Lavas from Noun Plain are the most magnesium-rich types with MgO varying between 5.40 and 13.77 wt%compared to those in Bamileke Plateau and Tombel Plain between that display 8.29–8.96 wt% 3.61–10.38 wt% respectively. The total alkali(Na2O + K2O)varies between 3.55 and 4.72 wt%in the Bamileke Plateau,from 3.06 to 6.03 wt%in the Tombel and 3.81–6.12 wt%in the Noun Plains.Most of the samples present acceptable LOI values ≤4 wt% (Table 1). However,a few of them(BKI 48,TL09,TPja13,BL32,and N34)present values between 4 and 8.51 wt%. These might be considered as weathered (Crummy et al.2014) and will be used with caution in further petrological interpretations.

Fig. 7 The Tombel Plain,Bamileke Plateau and Noun Plain bivariate diagrams of major elements versus MgO for analyzed pozzolans compared to the CVL nearby volcanoes.Markers are as in Fig. 5

The CIPW norm reveals normative nepheline and olivine on half of the studied sample. The Tombel Plain pozzolans display lowest values of normative nepheline(0–5.8 wt%) and olivine (0–10.40 wt%) while those from Bamileke Plateau and Noun Plain exhibit high- to highestvalue of normative nepheline (4.30–24.92 wt%) and olivine (10.53–21.39 wt%) apart the tephritic sample BL31 with Ol = 8.26 wt%.

Figure 6 highlights the variability of rock types.Amongst the 24 samples studied in the Tombel Plain,there are 13 alkaline basalts, five sub-alkaline basalts, five trachybasalts, and one basaltic andesite. The Noun Plain includes seven tephrites and basanites, two trachybasalts,one alkaline basalt, one basaltic trachyandesite, and one trachy-andesite. The Bamileke Plateau samples are two basanites and one alkaline basalt.

Figure 7 shows bivariate plots of selected major elements (SiO2, TiO2, CaO, K2O, Al2O3, FeO, and Fe2O3)versus MgO. The analyzed lavas along with lavas from neighborhood Mt Cameroon, Mt Manengouba, and Mt Bambouto volcanoes define continuous trends in most of these variation diagrams with a decrease in SiO2, Al2O3,and Fe2O3as MgO increases. Contrarily, the FeO, K2O,Na2O,CaO,and TiO2generally increase as MgO increases.

Fig. 8 The Tombel Plain,Bamileke Plateau,and Noun Plain trace elements variations versus MgO for analyzed pozzolans compared to the CVL nearby volcanoes. Markers are as in Fig. 5

4.4 Trace elements composition

Table 3 reports the trace elements (ppm) compositions of the studied samples and Fig. 8 shows bivariate plots of selected high field strength elements (HFSE), large-ions lithophile elements (LILE) and transition elements against MgO. HFSE’s display a slight negative correlation with MgO in the Tombel Plain. Apart from Hf and Y (not shown) that follow the same trend as in Tombel Plain,HFSE’s remain constant in the Bamileke plateau and Noun Plain (Fig. 8). LILE such as Rb, Cs, and Sr, are also constant with increasing MgO in the Noun plain. Samples of the Tombel Plain can be discriminated between enrichedand poor-LILE; the enriched-LILE lava patterns decrease as MgO increases until ～8 wt% while the poor-LILE lavas patterns increase as MgO rises until ～8 wt%(Fig. 8). The transition elements (Ni, Cr, Co) exhibit a positive correlation with increasing MgO for all samples irrespective of the volcanic field. Scandium and V (plots not shown) follow the same trend for all the samples from Tombel plain but remain constant for samples from Bamileke plateau and Noun plain.

Figure 8 shows the studied lavas normalized to primitive mantle multi-elements patterns compared with nearby large stratovolcano samples.The Tombel Plain samples are compared with those of Mt Cameroon and Mt. Manengouba (Fig. 9a), while Noun Plain and Bamileke Plateau samples are compared with those of Mt Bambouto(Fig. 9b). All samples follow the same pattern with their nearby volcanoes and show a general OIB-like pattern(Fig. 9a, b). An enrichment of LREE relative to HREE is evident with negative anomalies in some compatible elements such as Cs, Rb, U, Th, and Pb, and a positive anomaly in one incompatible element (Nb).

5 Discussions

5.1 Crystal fractionation

The overall mafic composition (SiO2＜52 wt%) of the studied samples (Fig. 5) with high MgO values (up to 10.38 wt% in Tombel plain, ～8 wt% in Bamileke Plateau and between 8.04 and 13.77 wt% in Noun Plain) and Cr, Ni, and Co contents, which overlap with the expected composition for melts equivalents to primitive magma(Cr;300–500 ppm, Ni; 300–400 ppm, Co; 50–70 ppm Beier et al. 2006, 2008; Madureira et al. 2011) suggest rapidmagma ascend. Consequently, a significant effect of the crystallization process during their ascension to the surface cannot be expected. This could be supported by the significant contents in amorphous phase in most samples(Table 1) even though the positive trends of transitions elements such as Co, Cr,Ni (Fig. 8)against MgO (Brenna et al. 2010) and the negative trends of Al2O3and Na2O(Fig. 7)against MgO highlight fractionation of both olivine and clinopyroxene minerals. The mafic character and overall nearly primitive melt composition indicate that these samples can be used to discuss the nature and source composition of their magmas though it is primarily essential to check if those magmas did not suffer crustal assimilation during their passing through the Earth crust.

Fig. 9 Primitive-mantlenormalized (Sun and McDonough 1989) multielement diagrams for a the Tombel Plain, b Bamileke Plateau and Noun Plain analyzed pozzolans from compared to the CVL nearby volcanoes

5.2 Crustal contamination

Crustal contamination has already been evidenced in many CVL igneous rocks (e.g., Marzoli et al. 1999; Rankenburg et al. 2005; Kamgang et al. 2008, 2013; Tchuimegnie Ngongang et al.2015).The assimilation of crustal melts by a rising magma can increase its contents in some incompatible elements. Hart et al. (1989) indicated for example,that rising magmas that have faced crustal assimilation on passing through the thick continental crust could have ratios of La/Nb ＞1.5 and La/Ta ＞22. Such magmas also are expected to have a composition (Nb/Yb and Th/Yb ratios) that deviates from that of mid-ocean ridge basalts(MORB) or OIBs (Pearce 2008; Pearce et al. 1995). Our studied samples present La/Nb ratios between 0.6 and 0.9 and La/Ta between 10.4 and 16.04. They also plot within the MORB-OIB array in the Nb versus Nb/U and Nb/Yb versus Th/Yb (Fig. 10) that suggests a negligible contribution of crustal materials. This aligns with their high-MgO values, the lack of a negative Nb–Ta spike on the primitive mantle-normalized multi-elements and the ratios of some other incompatible elements such as Zr/Nb (3–5),and Ba/Nb (6.1–8.5) that equal those of OIBs (Weaver 1991; Sun and McDonough 1989).

Fig. 10 Diagrams of Th/Yb versus Nb/Yb and Nb/U versus Nb showing the OIB signature of the pozzolans (after Pearce and Peate 1995; Pearce 2008). N-MORB, E-MORB, and OIB data from (Sun and McDonough 1989). Markers are as in Fig. 5

5.3 Mantle sources

Fig. 11 Mantle sources and CVL relationships of the Tombel Plain,Bamileke Plateau,and Noun Plain pozzolans.a The plot of chondritenormalized ratios of Tb/Yb and La/Yb to distinguish between spinel and garnet mantles. Chondrite values from Sun and McDonough(1989). Horizontal line, according to Wang et al. (2002). b Fields defined for samples from Mt Cameroon, Mt Bambouto and Mt Manengouba. Note the distribution of our samples in the different fields especially samples from the Tombel Plain with similarities with those from Mt Cameroon, Mt Bambouto and Mt Manengouba.Markers are as in Fig. 5

The studied samples, as demonstrated above, exhibit an OIB affinity like most volcanic lavas of the CVL (Asaah et al. 2015; Kamgang et al. 2013; Ngwa et al. 2017; Suh et al. 2009; Yamgouot et al. 2016). They also show a composition that overlaps those of nearby volcanoes, suggesting that their magmas were certainly generated in similar sources in the mantle. Most of the CVL magmas derive from the melting of mantle rocks,dominantly within the garnet lherzolite stability field (Asaah et al. 2015;Atouba et al.2016)although cases of melting in the mantle spinel lherzolite field (Lee et al. 1994; Nkouandou and Temdjim 2011) and the participation of pyroxenites have been reported (Kamgang et al. 2008; Tchuimegnie Ngongang et al.2015).Using the ratios of La and Yb normalized to chondrite values, (La/Yb)N, Farmer (2003) suggested that lavas with a (La/Yb)N≤5 and those with a (La/Yb)N≥5 derived from melting in spinel and a garnet lherzolite stability zone respectively. Accordingly, the highest (La/Yb)Ni.e. 7–20 of our lavas suggests that their parental magmas originated from partial melting in a dominantly garnet-bearing mantle zone (Fig. 11a), i.e., at depths greater than 80 km. This can be supported by their(Tb/Yb)N＞1.9 (Wang et al. 2002). Still, the variations in trace elements compositions (La/Yb, Ba) of the studied samples might indicate either heterogeneity in the mantle source or variations in the degree of partial melting as shown by samples from the Tombel Plain that share features of both Mt Cameroon, Mt Manengouba, and Mt Bambouto magmas (Fig. 11b).

5.4 Pozzolanicity hints: influence of the mineralogy to the suitability of studied pozzolans in cement manufacturing

Portland cement remains the best cement used in the industry nowadays. However, because of its high cost and that of its main constituent, the clinker, the use of pozzolans compliant with standards such as NF EN 197-1 and ASTM C-618 remains a good alternative. The abundance and accessibility of pozzolans in the southern and central continental portions of the CVL might thus represent a prime material for the local industries. However, a pozzolan should present a certain quality, especially an excellent activity to be used as an additive in cement manufacturing and depends directly on its chemical and mineralogical compositions (Walker and Pavı´a 2011). The mineralogy and reactivity relationships of aggregates mainly obey to amorphous silica or cryptocrystalline content(Lombardi et al.1997;Prezzi et al.1997,1998).Also,the abundance of amorphous phases in a pozzolan is demonstrated to increase its reactivity or pozzolanicity(Pichon 1994; Segui 2011) improving, therefore, the mechanical properties of the pozzolan by promoting the cohesion and densification by the elimination of porosity(Weshondo 2012).

Though determining the pozzolanicity of our samples was out of the scope of the present paper, the extensive usage of these deposits in the cement factory might also be related to their mineralogical composition. For instance,most of our studied samples display a significant(＞10 %)amorphous phases content (Table 2, Fig. 5a), up to 50 %for Kpe´04,Kpe´07,and TL08 in the Tombel plain(Fig. 5b);to 81.2 %for BL32 in the Bamileke Plateau(Fig. 5c);and to 69.1 for MV38 in the Noun Plain (Fig. 5d); suggesting that they could present an interesting pozzolanicity, and therefore usable as additive in cement factory. The pozzolanicity of a natural pozzolan can also be appreciated based on the sum of SiO2, Al2O3, and Fe2O3T (SAI) or their sum of CaO,FeO and MgO(CIM)since those oxides allow the development of vitreous phase whereas other oxides such as Na2O are against amorphous phase formation (Weshondo 2012). According to the ASTM C 618 standard, a good pozzolan must have an SAI greater than 70 wt%.The SAI of our studied pozzolans vary from 68.50 to 83.50 wt% indicating that they almost all would fulfill the requirement of the standard. However, their CIM vary widely, between 14.5 and 30.52 wt% in the Tombel plain,between 20.67 to 22.03 wt% in the Bamileke plateau, and between 23.58 and 31.08 wt%in the Noun Plain.Two sets of samples can thus be distinguished; the first one with CIM far less than 29% and the second one made up of samples having a sum CIM ＞29 %. The latter set of pozzolans with CIM ＞29 wt% could be considered the more prone to develop an amorphous phase that governs the reactivity of pozzolans. From that point of view, only two pozzolans from Tombel plain(Lim53 and Lim54)and four from Noun Plain (ML37, MV38, FM43, and MB46)that present a relatively high amount of CIM (＞29 wt%)seem to be suitable as an additive in cement manufacture.A detailed study is being carried out in order to verify this hypothesis.

6 Conclusion

The mineralogical and geochemical studies of pozzolan deposits from Tombel Plain, Bamileke Plateau, and Noun Plain in the continental portion of the CVL have been realized to assess their petrology and their possible pozzolanicity or ability to be used as additives for cement manufacture. In summary, the studied samples comprise alkaline and subalkaline basalts, trachy-basalts, and basanites. They exhibit an OIB affinity, and alike most volcanic lavas of the CVL, their composition suggests that their magmas were generated in a dominantly garnetbearing mantle zone. Though the insignificant effect of crustal contamination and crystal fractionation were noticed, the variations in trace element compositions (La/Yb, Ba) suggest either heterogeneity in the mantle source or variations in the degree of partial melting. The main mineralogical assemblages including augite, plagioclase,followed by Fe–Ti oxides and olivine,with few K-feldspar,nepheline, and quartz as accessory phases, as well as high glass content of most samples, suggest their occurrence through the rapid ascent of the magma. According to their Activity Index(AI),the studied pozzolans,especially those in the Tombel Graben, with important amorphous phases,seem to be suitable as additives for cement manufacture.Additional tests providing information on Compressive Strength,SAI and durability that form part of ASTM C 618 requirements are being carried out to confirm the real importance of those materials in cement manufacture.

AcknowledgementsAuthors are thankful to the ISTerre of Grenoble(France) and Nathaniel Findling for hosting and training the first author on the XRD technique and Rietveld method interpretation as part of his Ph.D. research work. The authors are also thankful to the Ministry of Higher Education of Cameroon for providing Research Modernization Allowance (RMA) that has been very helpful to support part of the field trip and Geochemical analysis expenses.