Introduction

The use of rocks in construction (both as foundation, crushed stone, and aggregates) has gone on since time immemorial. However, the demand for and utilization by construction projects has significantly increased in recent years as has the sophistication in building technology, civil engineering, and infrastructural development (Lester 1981; Smith and Collins 2001). Research shows that the world demand for construction aggregates in 2008 was about 24.9 billion metric tons (Freedonia 2009). World demand for construction aggregates is forecast, according to Freedonia (2009), to expand 2.9 % annually through year 2013 to 28.7 billion metric tons. Nelson and Bolen (2008) state that the total U.S. aggregate demand by final market sector for the year 2008 was 30–35 % for non-residential buildings (offices, hotels, stores, manufacturing plants, government and institutional buildings, and others), 25 % for highways, and 25 % for housing.

In the Abakaliki area of southeastern Nigeria (Fig. 1), pyroclastic rocks are found associated with the thick Albian Asu River Group sediments. The occurrence and petrology of these rocks (commonly referred to as the ‘Abakaliki pyroclastics’), have been extensively studied by earlier researchers such as Okezie (1957, 1965), Farrington (1952), Uzuakpunwa (1974), and Olade (1979). About 10,000 metric tonnes of crushed rock mainly pyroclastic rocks are produced on daily basis in this area. These products are being used in most construction projects (especially as road stones) in southeastern Nigeria.

Fig. 1
figure 1

Geological map of the Abakaliki metropolis showing studied pyroclastic suites

Full size image

Despite common structural failures, including a few instances of building collapse, these aggregates are widely used (Aghamelu and Okogbue 2011; Aghamelu et al. 2011). Notwithstanding their widespread use, sparse information is presently available on the impact of the geotechnical and geochemical properties of these pyroclastics when used as aggregates for construction projects. In this research, an attempt has been made to employ basic but standard laboratory tests to assess the geotechnical and geochemical properties of the Abakaliki pyroclastics, in terms of their suitability as road aggregates. Based on the results obtained, the causal contribution of the pyroclastic material to known failures of road projects in the area is evaluated.

Previous researches on rocks as road aggregates

Today, almost all types of rocks, ranging from igneous, sedimentary, to metamorphic, are used in engineering constructions. This is mainly due to the increase in demand for construction aggregates as well as advancing sophistication in the technology that enables handling and processing of a range of construction materials. Despite this improving technical sophistication, some rock types naturally perform better than others. For instance, ‘trap rocks’, which comprise rocks of basaltic composition, perform generally better as road stone than ‘hydrophilic rocks’, those of granitic composition. Krynine and Judd (1957) observed that trap rocks have less affinity for water, (i.e., they are ‘hydrophobic’) whilst having a high affinity for bitume. Conversely, hydrophilic rocks, as the name implies, have high affinity for water but less less affinity for bitumen. If hydrophilic rocks are used as a road stone in this region, the likely long term result is the breakup or ‘stripping’ of the road surface. However, trap rocks do not often serve well as a surface finishing material (Krynine and Judd 1957), as they present a hard, dense, interlocked surface, difficult to polish (Bimel 1988). Therefore, determining the suitability of a rock type for road construction material should be made prior to construction.

Previous research has shown that aggregates, to a large extent, derive their behavior from the geological characteristics of their parent rock material (Goswami 1984; Waltham 1994); thus, the geological classification itself may provide the potential user an indication of the expected field performance of aggregates under consideration, prior to utilization in construction projects. According to Goswami (1984), Ballivy and Dayre (1984), and Brattli (1992), the engineering or mechanical behavior of most common rock types (and their derived aggregates) is largely controlled by geology. Hence, it is important that petrographic (modal) analysis be carried out to assess the geological attributes of the aggregate to be used in a construction project. Whilst geological factors govern the suitability of a particular rock as an aggregate, Hartley (1974), however, concluded that petrographic fabric of the rock has a significant bearing on the mechanical properties of any rock type. However, given the complex nature of the constituent minerals and their bonding, petrography cannot be used as the only means of evaluating these properties as potential aggregates.

Owing to the complexity of rock properties, some laboratory tests have been designed to assess the ability of aggregates to withstand crushing, impact, and abrasive stresses in service (Waltham 1994), especially as road aggregates. Rocks and aggregates used in highway pavement, especially the surface course, are subjected to the abrasion caused by horizontal forces from wheels of automobiles. The resistance of these rocks to predetermined forces applied in the laboratory is measured against industry standard calibrations namely, aggregate impact value (AIV), Los Angeles abrasion value (LAAV), ten percent fines value (TFV), and flakiness index (FI) tests (Waltham, 1994). It is also possible nowadays to measure the durability or rate of deterioration of aggregates over `engineering time' by means of an aggregate water absorption test (ASTM C127 1990). Density or specific gravity of rocks has been considered an excellent indictor of its durability (ASTM C128 1990). For example, Reidenouer (1970) observed that aggregates with specific gravity greater than 2.65 would perform better than those with specific gravity lower than 2.65. Neville (1973) and Eze (1997) have associated durable rocks with high unconfined compressive strength (UCS) values (usually above 200 MPa).

Description of the study area

Geographic setting

This study covers the Abakaliki metropolis and its suburbs (see Fig. 1). Abakaliki is the capital city of Ebonyi State, southeastern Nigeria. It is encompassed within latitudes 6°15′N and 6°20′N and longitude 8°05′N and 8°10′N and covers an area approximately 320 km2. The state has a good network of roads; however, despite the national investment of aggregate sealing major infrastructure routes, many of these paved roads experience failure over time.

The Abakaliki area experiences two distinct climates, referred to as the dry and wet (rainy) seasons. The wet season begins in March and ends in October and is characterized by frequent, high volume rainfall, relatively low temperature, and high relative humidity. The dry season begins in October and ends in February and is characterized by infrequent rainfall, high temperature, and low relative humidity. Temperature in the dry season ranges from 20 to 38 °C and during the rainy season, 16 to 28 °C. The average monthly rainfall ranges from 3.1 mm in January to 270 mm in July. Average annual rainfall varies from 1,500 to 1,650 mm.

Physiography

The relief is generally undulating, lying at around 100 to 200 m above-sea-level (ASL), and is predominantly underlain by shale. Localised high ground is present as a belt of south west to north east trending, conical hills of intrusive pyroclastic material. These residual hills are generally above 200 m ASL and locally exceed 400 m ASL (such as Juju Hill at 492 m ASL). The predominance of shale as the underlying countyrock, has resulted in a landscape mostly devoid of deeply incised valleys and erosion channels. Drainage is dominated by the Ebonyi River and its tributaries: Ebia, Iyi-Udene, and Iyiokwu Rivers. The area is generally waterlogged during the rainy season, especially at low elevations, including those traversed by the paved roads.

Geology

The Abakaliki pyroclastic rocks consist of a sequence of tuff, mafic lavas, pyroclastic flows, agglomerates, and amygdaloidal lavas of basaltic composition, alkaline in nature (Ofoegbu and Amajor 1987; Obiora and Umeji 1995); these are intruded into and extrude through the Albian Asu River Group Shale (Reyment 1965). In this region, lightcolored, grayish tuffs and unstratified lavas, form the dominant lithology of the Abakaliki pyroclastics (Olade 1978, 1979). Previous researchers have variously described and dated the Abakaliki pyroclastics as follows: interstratified with Albian shales (McConnel 1949); post-Albian (Tattam 1960; Farrington 1952; Hoque 1984); pre-Albian or Aptian (Uzuakpunwa 1974); and pre-Albian and overlying the basement complex (Olade 1978). Different varieties of pyroclastics exist in some other parts of southeastern Nigeria, for example, at Ezillo (Ofoegbu and Amajor 1987), Uturu (Ukaegbu 2008), and Lokpan-Ukwu (John-Onwualu and Ukaegbu 2009).

Methodology

Sampling and testing

A total of fifteen fresh representative samples from different pyroclastic rock bodies within the Abakaliki area were collected as lumps with the aid of sledge hammer and stored in polyethylene sacks. The sampled pyroclastic bodies are herein designated as A-PR01 to A-PR15 and are presented in Fig. 1. To ensure that the moisture content change and other alteration processes did not affect the test results, polyethylene sacks used were sun-dried 48 h before use and firmly sealed the moment they contained samples. Laboratory testings commenced in all cases within 48 h of sampling.

The various laboratory tests and testing followed standards specified by British Standard Institution (BS 812 Parts 1 and 2 1975a, b; 812 Parts 100 and 110 1990a, b), International Society of Rock Mechanics (ISRM 1979, 1981), and American Society of Testing and Materials (ASTM C127 1990; C128 1990; C535 1988; C88 1990). The tests include petrographic modal analysis, water absorption, specific gravity, Los Angeles abrasion, aggregate crushing, Ten percent fines, flakiness index and Schmidt hammer rebound. Petrographic analysis was carried out in accordance with standard procedure for petrographic examination of rocks given by ISRM (1978). Schmidt hammer rebound and unconfined compressive strength were performed on bulk samples, while the rest were on aggregates. In most cases, the recorded values of tested parameters are the average of at least three or more sets of tests. Descriptive, correlation, and regression analyses were carried out using proprietary computer software, GenStat Discovery (Edition 3).

Results and discussion

Modal analysis

The results of the petrographic modal analyses are summarized in Table 1. A photomicrograph of one of the studied thin sections is presented in Fig. 2. Analysis shows that the Abakaliki pyroclastic rocks consist predominantly of phenocrysts of plagioclases (ranging between 21 and 60 %) and pyroxenes (ranging between 6 and 12 %). These two predominant phenocrysts are set in a compact, chaotic mixture of unsorted angular to subangular lithic fragments (shale and mudstone), which range from 11 to 32 %, and a highly altered basaltic groundmass (consisting of volcanic glass, plagioclase, chlorite, and secondary carbonate), which ranges from 11 to 36 %. Amygdaloidal calcite and quartz are also present and range from 0 to 10 % and 0 to 4 %, respectively. Opaque minerals constitute between 3 and 8 % of the pyroclastics. The grain sizes of the major constituents are as follows, <1–4 mm for plagioclase and pyroxene phenocrysts, <1–2 mm for lithic fragment, amygdaloidal calcite and quartz, and the glassy groundmass, significantly, <1 mm. The textures of the samples range between fine-grained and porphyritic (see Table 1).

Table 1 Results of modal analyses on the samples of the Abakaliki pyroclastics
Full size table
Fig. 2
figure 2

Photomicrographs of two samples, representing porphyritic (A-PR02) and fine-grained (A-PR09) textures, of the Abakaliki pyroclastic rocks (pl—plagioclase, px—pyroxene, ca—calcite, qz—quartz, op—opaque mineral, gm—groundmass, and Lf—lithic fragment)

Full size image

As can be seen from Table 1, sample A-PR9 has plagioclase (21 %) significantly lower than the others, suggesting that mineralogical composition is different. Earlier researchers (Farrington 1952; Hoque 1984; Obiora and Umeji 1995) have noted that considerable variations exist in the mode of formation, texture, and mineralogy of the Abakaliki pyroclastics. That sample A-PR9 is mineralogically different may mean that it is a distinctive variety of the pyroclastic rocks studied here. This suggestion is also supported as sample A-PR9 also recorded the highest amount of glassy groundmass (36 %). The pyroxene in the studied rocks is suspected to be augite. Ofoegbu and Amajor (1987) had suggested occurrence of augite in the pyroclastics, which has mostly been replaced by chlorite and calcic-plagioclase. They equally noted that some of the pyroxenes have recrystallized and altered to albite, carbonate, and epidote. Iron, apatite, and titanium oxides may likely constitute the opaque and accessory minerals (Ofoegbu and Amajor 1987).

The presence of particular deleterious minerals can render an aggregate unsuitable for construction use. According to Krynine and Judd (1957), high content of quartz (above 60 %, Bangar 2005), such as in granite, renders aggregates ‘hydrophilic’, which has higher affinity for water than bitumen. Use of hydrophilic aggregates results in ‘stripping’ (Krynine and Judd 1957) in road pavements. Low quartz (0–4 %) in the Abakaliki pyroclastics may mean that they are ‘hydrophobic’ (i.e., higher affinity for bitumen than water). While chlorite is considered a weak mineral in rocks and aggregates, plagioclases and volcanic glass minerals readily convert to clays under favorable climatic condition, especially high humidity (Krynine and Judd 1957; Jaeger and Cook 1969) and fluctuating temperature, as is the case in the tropics.

Although the modal analysis was carried out on fresh samples, alterations are, however, inevitable overtime. Obiora and Umeji (1995) have noted that volcanic glass in the Abakaliki pyroclastics alters to form expansive clays (montmorillonite or smectites) within few days of exposure. Hence, samples, such as A-PR01, A-PR02, A-PR07, A-PR09, and A-PR13, with very significant contents of lithic fragment and glassy groundmass will be susceptible to faster alteration and material deterioration.

Water absorption (W a)

The results of W a analysis are presented in Table 2. The results indicate that W a ranges from 0.7 to 1.5 %, with mean value of 1.04 %. Krynine and Judd (1957) note that the amount of water an aggregate can absorb tends to be an excellent indicator as to the strength of the aggregate, in other words, its weakness. They observed that strong aggregates will have a very low W a value, usually below 1.0 %. Eze (1997) had observed that above 4.0 % W a, the acceptability of an aggregate as a construction material would require further geotechnical analysis to ascertain. According to ASTM C127 (1990), W a of construction aggregate should be <2.5 %. The fact that most (8 out of 15) of the samples from the Abakaliki pyroclastics recorded W a values above 1.0 % makes it mandatory that further engineering tests must be carried out to establish their suitability as a construction aggregate.

Table 2 Results of water absorption and specific gravity tests on the Abakaliki pyroclastics
Full size table

The general low W a (all below 2.5 %) may also buttress the freshness or very low degree of weathering of the plagioclase or feldspathic contents of the tested samples. Feldspathic minerals in rocks on weathering commonly yield clay, which in turn increases the W a capacity of such rocks. Cobanoglu et al. (2003) had reported high W a values (13.40–25.90 %) in some tuffs from Turkey and associated it partly with high percentage of clay inside the tuffs. It is probable that the plagioclase and/or other feldspathic minerals in their tested tuff samples had undergone appreciable degree of weathering, resulting in existence of clay within the tuff.

Specific gravity (SG)

As shown in Table 2, the SG of the pyroclastic aggregates ranges between 2.52 and 2.78. Krynine and Judd (1957) note that the SG or unit weight of a rock depends on the density of its constituents and the amount of water in the pores. For instance, if the rock is dominated by plagioclase, calcite, or chlorite, its SG should range between 2.62 and 2.90 (see Lambe and Whitman 1969). It has also been reported (Jaeger and Cook 1969) that SG is a reflection of the amount of heavy elements (especially Fe and Mg) present in a rock. In most cases, SG can serve as an indirect means of determining the amount of stable and potentially durable minerals in aggregates.

Aggregates with high SG values (above 2.65; Reidenouer 1970) are generally suitable in construction. The Abakaliki pyroclastics have mean SG of 2.65 which may be an indication of good suitability as a construction material. The appreciable SG of the Abakaliki pyroclastics may have resulted from significant amount of pyroxene and iron oxides. SG, however, is likely to drop, in service, as appreciable weak minerals (plagioclase and chlorite) are recorded and are prone to deterioration when exposed to weathering conditions (Bell 1993).

Los Angeles abrasion value (LAAV)

A summary of the results of LAAV tests on the studied samples is presented in Table 3. The table indicates that LAAV of the tested samples ranges from 11 to 25 %, with a mean value of 19 %. Waltham (1994) has noted that the smaller the LAAV, the less the aggregate abrades. According to Lefond (1975), exclusion of aggregate with low LAAVs is relevant only to coated 20 mm aggregate (i.e., pre-coats) applied to hot rolled asphalt wearing course layers. Generally, fine-grained rocks usually yield crushed aggregates with low LAAV (below 15 %), while coarse-grained rocks yield higher values (above 15 %, Waltham 1994). This may explain the reason that trap rocks (i.e., basalt and its likes) have their LAAVs fall within acceptable limits, generally less than 10 % (Waltham 1994).

Table 3 Results of aggregate tests on the samples of the Abakaliki pyroclastics
Full size table

The tested aggregates from the Abakaliki pyroclastic rock have mean LAAV of 19 %, suggesting that the aggregates may not serve well as a surface course in road projects despite their fine-grained nature. This seeming anomaly can be explained by the presence of weak minerals (plagioclase and chlorite) in the studied pyroclastics. Indeed, Van Rooy and Nixon (1990) had documented that texture and mineralogy are factors that control the toughness and durability of rocks and aggregate. And since LAAV is a measure of toughness of aggregates (Waltham 1994), there is little wonder therefore that the LAAV of the Abakaliki pyroclastics is higher than expected.

Extreme climatic conditions (e.g., high rainfall combined with high, fluctuating temperatures) as is the case in Nigeria and other countries in the tropics, would provide favourable conditions for feldspathic minerals and glassy material to decay (or weather) to clay. Such deterioration would profoundly affect the performance of the Abakaliki pyroclastics as a construction aggregate, in service. Based on this study, the worst construction material performance is to be expected on samples A-PR01, A-PR02, A-PR07, A-PR09, and A-PR13, which have a very significant contents of lithic fragment and glassy groundmass. Therefore, the alteration and decay of the rock mass of these samples of pyroclastic material is likely to be higher than others sampled.

Aggregate crushing value (ACV)

As shown in Table 3, the aggregate crushing values (ACVs) of the tested samples range between 8 and 15 %, with a mean value of 12 %. Waltham (1994) had recorded ACV of 14 % on basalt and that compares well with the ACV values of the Abakaliki pyroclastics. Generally, fine-grained rocks yield aggregates with higher crushing resistance than coarse-grained rocks that yield aggregates with lower crushing resistance; thus, the lower the ACV, the stronger the aggregate, in terms of its ability to resist crushing. Waltham (1994) had also noted that an aggregate with ACV below 5 % would serve well as a construction material, while that with ACV above 35 % would most likely perform poorly as a construction material. Lefond (1975) considered 30 % as the upper ACV limit for reliable aggregate.

Based on these recommendations, the Abakaliki pyroclastics would yield aggregates that would likely serve well when used in construction projects where crushing stresses are dominant. This favorable ACV of the aggregates from the pyroclastics resulted most probably from their fine-grained to porphyritic texture, as it has been widely noted by several researchers (Lefond 1975; Waltham 1994; Eze 1997) that coarse-grained rocks record relatively high percentage loss in values in the ACV test. Modal analysis of the rock samples has shown that the studied samples have fine-grained to porphyritic texture, with the particle sizes of the groundmass significantly <1 mm and that for the phenocrysts generally <4 mm.

Ten percent fines value (TFV)

Table 3 shows that the TFV of the tested aggregate samples ranges between 102 and 202 kN, with mean value of 138 kN. The TFV is considered one of the important parameters used to assess aggregate prior to selection and designing in construction projects where impact and crushing stresses are expected. It usually supplements ACV determination on aggregates. Bell (1993) notes that coarse-grained igneous rocks such as granite are not generally as suitable as the fine-grained types, as they crush more easily. On the other hand, the very fine-grained and glassy volcanics are often unsuitable; since when crushed, they produce aggregates with sharp edges (Bell 1993).

Waltham (1994) has noted that an aggregate which will serve well generally as a construction material would have its TFV above 400 kN, while those with TFV below 20 kN would likely be unreliable and non-durable, especially when used as a road stone. The tested aggregate samples from the Abakaliki pyroclastics all have their TFV between 102 and 202 kN, suggesting fair suitability as construction material, especially in road projects. Despite the view of an earlier researcher (Bell 1993) that texture and origin control the TFV results of some rocks, especially granite, it is evident from this study that the Abakaliki pyroclastic rocks are texturally and mineralogically different from the granitic rocks. They, therefore, fall outside the class of those rocks where texture and origin could control their TFV results. Attempts to correlate the TFV values of the pyroclastics with texture, mineralogy, or groundmass yielded no correlation result, suggesting that a more complex combination of factors may be at play in the case of Abakaliki pyroclastics.

Flakiness Index (FI)

As shown in Table 3, the FI of the tested aggregates ranges between 18 and 32 %, with mean value of 26 %. A flaky aggregate has low density under compaction and less strength than a cubical or an angular aggregate. It also provides less texture when used in surface dressing (Krynine and Judd 1957). Waltham (1994) has pointed out that aggregates with FI value of 20 % and below will serve well as a construction aggregate, while those with values above 70 % are unlikely to be acceptable. According to Waltham (1994), very good aggregate for use as a road stone should have FI of 3 % or below. This indicates that aggregates from the Abakaliki pyroclastics, which have FI between 18 and 32 %, would provide marginal performance as construction aggregate and would be considered poor as a road stone.

The rating of the Abakaliki pyroclastics as marginal for use as construction aggregate and totally unacceptable as road stone (as indicated by the high flakiness index), is possibly the rocks are volcanic in origin and generally fine-grained. Bell (1993) had noted that fine-grained volcanic rocks produce aggregates with sharp edges during crushing. Table 3 shows that the highest FI value of 32 % was recorded on a fine-grained sample (that is, A-PR10), with a correspondingly highest total content of plagioclase and glassy groundmass. The lowest FI value of 18 % was, as expected, recorded on sample A-PR13, which had a porphyritic texture and lower total content of plagioclase and glassy groundmass. These results may suggest that total content of plagioclase and glassy groundmass combined with the crystal fabric characteristics of the rock, has a bearing on the FI results.

Schmidt Hammer Rebound Number (R)

The results of the Schmidt Hammer Rebound tests on the tested samples are presented in Table 4. The table indicates that the rebound number (R) varies from 32 to 48 %, with mean value of 38 %. Roberts (1977) and Franklin and Dusseault (1989) had pointed out that soft rocks, such as shales, would generally have R that ranges from 0 to 30 %, while hard rocks, including igneous rocks, would generally record high R that may be up to 60 %. This suggests that R, to a large extent, is controlled by rock texture, mineralogy, and origin. As can be seen in Tables 1 and 4, samples A-PR09 and A-PR13, despite noticeable difference in their mineralogical composition, both have nearly equal R values. It is very likely that the near equality in their R values was brought about by similarity in texture; both samples are fine-grained. Krynine and Judd (1957), Jaeger and Cook (1969), Bell (1993), and Waltham (1994) had opined that hard rocks would usually be indentified with good strength and mechanical properties and generally good performance as construction materials. The Abakaliki pyroclastic rocks, with R ranging from 32 to 48 %, are definitely not in the soft rock category, but on the lower rating of hard rock.

Table 4 Results of strength tests on the samples of the Abakaliki pyroclastics
Full size table

Uniaxial compressive strength (UCS)

As shown in the Table 4, the UCS values of the tested samples range from 40 to 122 MPa, with mean value of 69 MPa. Eze (1997) had noted that a sound rock (devoid of fissures and defects) would have UCS close to or above 200 MPa for it to be considered a good aggregate source for most civil engineering works. Neville (1973), also considers 200 MPa as the threshold below which a material is considered to be a poor aggregate material. On this basis, the Abakaliki pyroclastics would be said to yield unsatisfactory aggregates for civil engineering works.

A rock classification scheme based on UCS is presented in Table 5. Of the fifteen Abakaliki pyroclastic rock samples tested, sic classify as moderately strong, five as strong, and the other four as very strong rocks. Despite these relatively favourable ratings, the Abakaliki pyroclastics would not qualify as good aggregates following Neville’s (1973) and Eze’s (1997) observations. This poor rating as an aggregate source, when judged by the UCS analysis, may be accounted for by the appreciable amounts of glassy groundmass, plagioclase, and shaley to muddy lithic fragmentc, as shown in Fig. 2.

Table 5 Strength classification of rock based on UCS
Full size table

Remarkably, samples A-PR09 and A-PR10, with the lowest and highest contents of plagioclase, respectively, have both classified as very strong (see Table 5). This may be due to similarity in texture; both are fine-grained. Sample A-PR02 with lowest total plagioclase and lithic fragment content classifies as moderately strong alongside other samples with significantly higher total plagioclase and lithic fragment contents. A possible explanation to this result, could be that although sample A-PR02 has less content of phenocrysts, the phenocrystals in this sample are considerably larger in grain size when compared to phenocryst grain sizes in the other samples (see Fig. 2). This difference may be the cause of the lower UCS strength rating of sample A-PR02.

Statistics

Statistical analyses on the values of the tested parameters indicate strong degree of positive relationships, expressed in terms of correlation coefficient (r 2): between sum of Fe and Mg oxides versus SG, r 2 = 1 (Fig. 3); R versus UCS, r 2 = 0.9525 (Fig. 4); UCS versus SG, r 2 = 0.9364 (Fig. 5); UCS versus TFV, r 2 = 0.9343 (Fig. 6); and TFV versus SG, r 2 = 0.9087 (Fig. 7). The trend suggests that presence of heavy minerals and oxides had yielded appreciable specific gravities which have direct influence on both the physical and mechanical properties of the samples. Strong degree of negative relationships, in terms of r 2, was found to exist between SG versus W a, r 2 = −0.9417 (Fig. 8); UCS versus W a, r 2 = −0.8988 (Fig. 9); TFV versus W a, r 2 = −0.8876 (Fig. 10); and R versus W a, r 2 = −0.8527 (Fig. 11). The trend, statistically, identifies moisture increase as detrimental to the physical and mechanical properties of the tested samples.

Fig. 3
figure 3

Cross plot of sum of iron and magnesium versus specific gravity

Full size image
Fig. 4
figure 4

Cross plot of Schmidt hammer rebound number versus unconfined compressive strength

Full size image
Fig. 5
figure 5

Cross plot of unconfined compressive strength versus specific gravity

Full size image
Fig. 6
figure 6

Cross plot of unconfined compressive strength versus ten percent fines value

Full size image
Fig. 7
figure 7

Cross plot of ten percent fines value versus specific gravity

Full size image
Fig. 8
figure 8

Cross plot of specific gravity versus water absorption

Full size image
Fig. 9
figure 9

Cross plot of unconfined compressive strength versus water absorption

Full size image
Fig. 10
figure 10

Cross plot of ten percent fines value versus water absorption

Full size image
Fig. 11
figure 11

Cross plot of Schmidt hammer rebound number versus water absorption

Full size image

No relationship was found to exist when UCS was plotted against ACV for both linear (r 2 = −0.0337, Fig. 12) and logarithmic (r 2 = −0.0365) relationships, despite the fact that UCS and ACV represent strength factor in material. Irfan (1994) had noted that aggregate parameters must not always be related statistically. The reason for such non-relationship may include nature of tested samples and quantity of plotted data (Al-Harthi 2001). In this study, UCS tests were conducted on bulk rock samples, while ACV tests were carried out on aggregate samples.

Fig. 12
figure 12

Cross plot of aggregate crushing value versus unconfined compressive strength

Full size image

Suitability of the Abakaliki pyroclastics for road construction projects

The geotechnical assessment provided by this study, indicates that the Abakaliki pyroclastics, as a road aggregate source, would yield aggregates that are marginally satisfactory for use as road stones. Although the tested samples meet a number of requirements (TFV and Wa are within acceptable limits), they fail to meet most of the requirements as a road stone; their AIV, LAAV and FI values are above the recommended limits (see Table 6). The failure of the pyroclastic aggregates to meet all the industry standards for use as road stone may account for their poor field performance. Aghamelu and Okogbue (2011) and Aghamelu et al. (2011) have noted the common road failures in the Abakaliki metropolis (Fig. 13); possibly these failures are not be entirely due to subgrade problem (as suspected by these authors) but may be partly associated with the widespread use of the pyroclastics as a road construction aggregate in the area.

Table 6 Comparison of the Abakaliki pyroclastics with general specification as road stone
Full size table
Fig. 13
figure 13

Some failed paved road (FPR) portions within Abakaliki metropolis (Locations FPR 1-Police Junction, Abakaliki-Mile 50-Ishieke Road, FPR 2- Gilbert Street Junction, Ogbaga-Izzi-Nodo Road, and FPR 3- Near Mile 4 Hospital, along Abakaliki, Mile 50-Ishieke Road. Photographs taken on 26/04/09)

Full size image

A comparison of the road aggregate properties of the Abakaliki pyroclastics with some other rock types commonly used as road aggregates is presented in Table 7. As revealed in the table, the SG of the Abakaliki pyroclastics is significantly lower than those of basalt, dolerite, and hornfels, while their AIV is appreciably higher than that of dolerite. Also, their ACV is generally lower than those of granite and quartzite, and their LAAV generally higher than those of dolerite, hornfels, and graywacke. The TFV of the pyroclastics is significantly lower than that of dolerite.

Table 7 Some representative values of road stone properties of some common aggregates
Full size table

As previously noted, the pyroclastic rocks are classified geologically, as fine-grained to porphyritic, basic igneous rocks. Basalt and dolerite (also referred to as ‘trap rocks’) are similarly classed as basic igneous rocks. However, unlike the pyroclastic suite, basalt and dolerite yield good road aggregates. The differing geotechnical properties between these rock types, suggests that mode of formation, mineralogy, and texture are important factors that must be considered in determining the likely when predicting geotechnical behavior of the particular rock types when subjected to ‘field’ conditions. This primary assessment, would provide an indication as to whether the rock type would likely provide a good road aggregate.

Conclusions

Petrographic modal analysis and geotechnical investigations carried out on samples from the Abakaliki pyroclastics have enabled the following conclusions to be drawn on the suitability of this rock type for use as a road aggregate;

  1. 1.

    The Abakaliki pyroclastic rock yields aggregates with marginal performance in road projects. Although the aggregates meet the TFV and W a specifications, they fail to meet a number of other requirements as a road stone, such as AIV, LAAV, and FI. In particular, statistical analysis identifies their susceptibility to moisture increase as detrimental to the physical and mechanical properties of the tested samples.

  2. 2.

    Petrographic modal analyses indicate significant difference in the mineralogy and texture of the Abakaliki pyroclastic rocks. This is likely due to the variation in the magmatic and depositional processes that occurred during the emplacement of the volcanic suites. The presence of shaley to muddy lithic fragment and glassy groundmass is recorded in all the tested samples. In most cases, samples with fine-grained texture and low total content of plagioclase and glassy groundmass registered more favourable engineering properties thus indicating these were better suited as road aggregates.

  3. 3.

    Based on determinations of the mineralogical composition of the Abakaliki pyroclastic rock suite (i.e., significant amounts of glassy groundmass, shaley to muddy lithic fragments and plagioclase) and rock nature, it is judged that this material may experience deterioration in service and pavement constructed with it may fail. This prognosis is especially likely to prove to be the case, when the material is exposed to extreme climatic conditions (high volume rainfall and high and fluctuating temperatures) as is the case in Nigeria and the tropics in general.

  4. 4.

    The aggregate samples, however, recorded low quartz content (≤4 %), which may mean that they would have higher affinity for bitumen than water. Hence, stripping would likely not result in pavement constructed with the pyroclastic aggregate.

  5. 5.

    The abundance and cheap availability, as well as high cost and scarcity of alternative replacement as aggregate source in the Abakaliki area, will continue to be a prominent factor in the utilization of the pyroclastics as road aggregates in southeastern Nigeria.

  6. 6.

    This study has shown that although rocks may classify geologically alike, their mode of formation, mineralogy, and texture are factors that must be considered when predicting their geotechnical behavior both as a construction and as a road stone aggregate.