The kaolinitic sandstone (KS) samples in Wadi Siq-Rakiya area are from the white sandstone outcrops of the lower part of the Kurnub Sandstone Group of Lower Cretaceous. These outcrops are low relief and isolated hills of white sandstone crop out south of the Gharandal area (65 km north of Aqaba city) within and along the eastern mountainous ridges of Wadi Araba Graben.

Lithology of the kaolinitic sandstone outcrops are very fine, fine to medium-grained, angular, sub-angular to sub-rounded grains, friable, soft and partly reworked massive white sandstone containing scattered quartz pebbles and granules with white clay matrix. This lithology may result from crushing and subsequent weathering in a strongly deformed zone. The depositional environment of the lower part is interpreted as braided to meandering river channels (indicated by the presence of the channel-fill sandstone) with brief marine intrusions over the alluvial plain [1].

Three kaolinitic sandstone samples were taken from trenches excavated in the outcrops representing a thickness of 15 - 20 m. The sandstone samples are actually friable, poorly cemented sandstone with kaolinite clay mineral as matrix. Kaolinitic clays (thin seams) are interbedded within the sandstone beds (Table 1). A composite bulk sample for pilot study was collected from the same location. All samples were bagged, transported to Aqaba port, and shipped to the laboratory of the School at the University of Nottingham in UK.

The studied samples are friable sand with small sandstone lumps, which are easily breakable with little energy. There was no need for any crushing to reduce the size of the particles. Riffling, coning and quartering were carried out at the dry state. Different sizes of riffles were used to reduce the amount of the sample to different representative volume of sub-samples.

A detailed laboratory work was conducted to study the grain size analysis, the mineralogical variation and contents, the chemical components of the raw samples, and the dry and the wet sieved fractions (Figure 1). The results of this stage will be used as a foundation for the next evaluation and processing steps. Kaolinitic sandstone (KS3) sample was chosen to carry out the detailed characterisation study due to the fact that this sample represents the whole sequence of the deposit as well as the bulk sample.

Different methods of size analysis were employed in different stages of the study. Sieving technique was extensively used in this stage. In addition, other techniques, such as sedimentation and laser-sizer were also used in other stages.

Dry and wet sieve analyses were performed on the kaolinitic sandstone sample. The particle size distribution was determined using a set of sieves of aperture sizes of 1180, 850, 600, 425, 300, 212, 150, 106, 75, 53, and 38 μm. The weight retained on each sieve of dry and wet sieving was measured and the weight percentage was calculated [2]. The results of the weight percentage retained on each sieve are listed in Table 2.

The results of the kaolinitic sandstone sample are presented in cumulative and frequency distribution curves. For the purpose of plotting, the points on the cumulative and frequency curves are plotted in between two successive sieve sizes.

The cumulative undersize of the dry and the wet sieving of (KS3) sample shows that the median size (d₅₀) of the sample is about 190 μm, which indicates that the kaolinitic sandstone sample is mainly fine-grained size. The results indicate that the sand is poorly sorted (Figure 2).

The frequency distribution curves of the dry and the wet sieving of the KS3 sample show that the weight retained on sieve 212 μm has slightly increased from about (16.8%) in the dry sieving to about (18.6%) in the wet sieving. This increase could be explained due to the washing of the clay material from the surface of the grains that has dislodged the lumps and liberated the particles (Figure 3).

It is noticed from Table 2 of the KS3sample that the weight retained on the 38 μm sieve was decreased from 6.4% in the dry sieving to 3.1% in the wet sieving. This indicates that water has washed down the fine materials and the clay size particles to the −38 µm fraction.

The fine particles in the −38 μm wet sieve fraction constitute about 9.1% in weight of the whole rock. The fine material was analysed to determine the particle size distribution using “Mastersizer S” laser technique. The results have shown that 65% is less than 15 µm and 33% is less than 5 µm of this fraction, which stands for 6% and 3% of the whole sample respectively (Figure 4).

The chemical analysis for the major standard elements was carried out using X-Ray Fluorescence (XRF) spectrometry and the fusion bead moulding technique was employed for preparation test samples. The results of the chemical analysis of the whole rock samples are displayed in Table 3, while the wet and the dry sieved fractions of the (KS3) sample are listed in Table 4.

The results of mineralogical, chemical and grain size analysis of the −38 and −2 µm size fractions of the kaolinitic sandstone (KS3) confirmed that kaolinite was the only clay mineral found in the sample. The cumulative undersize curve showed that the fine fraction (−125 µm) contained about 34% by weight of the KS3 sample and the percentage of kaolinite increased towards the finer size fractions. Therefore, it was necessary to study this fraction in detail with a view to recovering kaolin.

The wet screened −125 µm fraction of the bulk sample was subjected to classification trials using hydrocyclone classifiers. The use of hydrocyclones was to produce fine size fraction in order to concentrate kaolin as a product. A sedimentation method was also applied to concentrate kaolin in the −5 µm size fraction. Figure 9 illustrates the procedures of kaolin concentration from the −125 µm fraction.

Hydrocyclone is a continuously operating classifying device, which utilises centrifugal force to accelerate the settling rate of the particles [3].

Two types of Mozley hydrocyclones were used. A 2-inch (50.8 mm) diameter hydrocyclone was employed to produce an overflow of “expected” −15 µm size fraction and another 10 mm diameter one used to produce an overflow of approximately −5 µm in size.

A substantial amount of the bulk sample was wet screened using a Russell Finex screen in order to produce a suitable quantity of −125 µm material as a feed for the hydrocyclone. The −125 µm fraction was agitated and diluted to obtain a concentration of about 5% - 6% solid by weight. The slurry, of pulp density 1039 kg/m³, was fed at 50 psi into the 2-inch hydrocyclone of 6.4 mm spigot and 14 mm vortex finder. This setting was expected to produce an overflow product of approximately −15 µm in size. A 20 second sample was collected from the overflow and the underflow. The wet and dried samples were weighed and the percentage of solids by weight, yield and the mass flowrate in kg/h were calculated (Table 5).

The results showed that the mass flowrate of dry solids in the 2-inch hydrocyclone was 205.2 kg/hr for the feed and 63 kg/hr for the overflow. The yield or recovery of solids to the overflow was 30.7% and in the underflow was 69.3%, while the weight of water recovered was about 78% and 22% for the overflow and underflow respectively.

The overflow product from the 2-inch unit was used as feed for the 10 mm hydrocyclone. The feed was adjusted to a solid concentration of 4% - 5% by weight and fed at a pressure of 100 psi using a 2 mm vortex finder and 1 mm spigot. It was stated that using the smallest vortex finder, highest inlet pressure and low feed pulp density produce the finest cut point (d₅₀) [4].

A 20-second sample was collected from the overflow and underflow streams. The wet and dried samples were weighed and the percentage solids by weight, yield and the mass flowrate in kg/h were calculated (Table 6).

The results obtained from operating the 10 mm hydrocyclone showed that the mass flowrate of the dry solids was 7.02 kg/hr for the feed, 1.98 kg/hr for the overflow and 5.04 kg/hr for the underflow. The yield or recovery of the solids to the overflow was 28.2% and 71.8% to the underflow, while the weight of water recovered was about 76% and 24% for the overflow and underflow respectively.

In another attempt to try to concentrate the kaolin in finer fraction, the overflow from the 2-inch hydrocyclone was used to produce a concentrate of −5 µm in size using the conventional sedimentation technique [5].

The results of the chemical analysis showed that the Al₂O₃ content of the −5 and 5 - 15 µm size fractions was 29.48% and 28.22% respectively (Table 7). The very close values would reflect poor refining as the quantity of kaolin in these fractions would be in direct proportion to the Al₂O₃ content. The results revealed that concentrating kaolin by refining method would not produce a high kaolin concentration in the fine fraction. This was probably due to the wide range of the particle size distribution present in the sample (from less than 1 µm

O/F: Overflow; U/F: Underflow.

O/F: Overflow; U/F: Underflow.

to around 15 µm), many kaolin particles occur as stacks, and the lack of good dispersion of the suspension. Therefore, there was a need for a shear force such as that encountered in the hydrocyclone to disperse the particles giving a greater chance for size separation.

The wet screened −125 µm size fraction and the overflow products from the 2-inch and 10 mm hydrocyclones were examined to determine the chemical, mineralogical and particle size analysis using XRF, XRD and Laser sizing techniques.

The particle size distribution of the wet screened −125 µm fraction and the overflow products of the 2-inch and 10mm hydrocyclones were analysed using Mastersizer-S Laser technique (Malvern). The results showed that classification had occurred using the hydrocyclones and the median (d₅₀) decreased from 28 µm for the −125 µm fraction to 6.4 µm and 3 µm for the overflows from the 2-inch and 10 mm hydrocyclones respectively (Figure 10).

The results of the chemical analysis of the hydrocyclones products and the wet screened −125 µm fractions are displayed in Table 8. The results indicate that the Al₂O₃ content increased from 8.43% in the wet screened to 27.89% and then to 32.58% in the overflows of the 2-inch and 10 mm hydrocyclones respectively.

Also the Al₂O₃ content (27.23%) of the underflow of the 10 mm hydrocyclone was still relatively high compared to the overflow (32.58%). This indicated that although separation had taken place based on the particle size, kaolin particles were still reporting with water to the underflow.

The same hydrocyclone products and the −125 µm fraction were analysed as well by private company [6] to confirm and double check by another part. The displayed results in Table 9 are in good agreement with those in Table 8.

NA: not available.

The results of the trace element analysis showed low values. Slightly high values of strontium (Sr) and sulphur (S) were noticed in the overflow products, which probably indicated the presence of Celestine (SrSO₄) heavy mineral as a source of strontium (Table 10).

The mineralogical investigation using XRD showed that quartz and kaolinite were the only minerals found in the wet screened fraction and the hydrocyclone products. The mineralogy of the overflow of the 10 mm hydrocyclone showed that although kaolinite was the major constituent, quartz still existed as a minor (Figure 11). It is indicated that very fine quartz particles (less than 5 µm) were recovered with kaolin in the overflow product.

As quartz was the only impurity provided that any heavy minerals present did not influence the Al₂O₃ content, then the only source of Al₂O₃ would be kaolinite. The kaolin content of the products was therefore measured based on the chemical assay of the Al₂O₃. Ideally, the weight percentages of the kaolinite constituents are; SiO₂: 46.55%, H₂O: 13.96%, and Al₂O₃: 39.49% [7]. A product of

100% kaolin content would contain 39.49% by weight of Al₂O₃ and therefore, the kaolin quantity in the products was calculated in proportional to the Al₂O₃ content as a ratio of (100:39.49).

Based on the results of the 20-second time samples from the 2-inch and 10 mm hydrocyclones (see Table 5 & Table 6), the kaolin grade and recovery results were calculated and are tabulated in Table 11. It was found that the 2-inch hydrocyclone recovered to the overflow 78% kaolin at a grade of 71% from the −125 µm fraction. The 2-inch overflow was fed to the 10 mm hydrocyclone and recovered 32% at a grade of 82.5% to the overflow stream.

It should be mentioned that there was still a considerable amount (68% recovery at a grade of 69%) of kaolin in the underflow product of the 10 mm hydrocyclone. Returning this underflow to the feed would probably increase the recovery of kaolin to the overflow. However, this was not part of the current work.

Based on the results of Table 11, the dry weight and recovery of the overflow products for the 2-inch and 10 mm hydrocyclones were calculated based on the dry weight of the wet screened −125 µm fraction. The results are displayed in Table 12.

The results showed that the kaolin grade (content) increased from 21% by weight in the −125 µm size fraction to 82% by weight in the 10 mm hydrocyclone overflow product while the recovery decreased from 43% to 14% (Figure 12).

The partition curve is the commonest method of representing hydrocyclone efficiency and is constructed by plotting the percentage of each particle size by weight in the feed, which reports to the underflow, against the particle size. Ref. [4] defined the cut point or d₅₀ size (separation size) as the point on the partition curve for which 50% of the particles in the feed of that size report to the underflow. Partition curves can be used to predict the products that would be obtained if the feed or separation size were changed. In all classifiers, it is assumed that solids of all sizes are entrained in the coarse product (underflow) liquid by short-circuiting in direct proportion to the fraction of feed water reporting to the underflow. Therefore, the partition curve can be corrected by utilising the following equation [8]:

y ' = y − R / 1 − R

where,

y' is the corrected mass fraction of a particular size reporting to underflow,

y is the actual mass fraction of that size,

R is the fraction of the feed liquid (water) recovered in the underflow.

The fraction (R) of the feed liquid recovered in the underflow was calculated in the separation processes of the two hydrocyclones (see Table 5 & Table 6). The fraction was 21.6% and 24.4% in the separation process of kaolin by the 2-inch hydrocyclone and the 10 mm hydrocyclone respectively. This meant that

21.6% of the feed material was short-circuited with water to the underflow without classification (i.e. not separated) in the 2-inch hydrocyclone, and 24.4% of the feed material in the 10 mm hydrocyclone. Therefore, the percentage of feed recovered to the underflow should be corrected in proportion to that short-circuiting fraction.

In the 2-inch hydrocyclone, the mass balance calculation of the products showed that the percentage of the feed (yield) which reported to the overflow was 31% and 69% to the underflow. The weight percentages of the particle size analysis of the products (i.e. overflow and underflow), the partition coefficient (% of feed to U/F), and the corrected partition coefficient are tabulated in Table 13. The results of the uncorrected and corrected of the percentage of feed to the underflow are plotted against the nominal size and illustrated in Figure 13. The results showed that the cut point (d₅₀) was 12 µm while the corrected cut point d_50(C) increased to 16 μm.

In the 10 mm hydrocyclone, the mass balance calculation of the products showed that the percentage of the feed (yield) which reported to the overflow and the underflow was 28.2% and 71.8% respectively. The weight percentage of the particle size analysis of the products (i.e. overflow and underflow), the partition coefficient (% of feed to U/F), and the corrected partition coefficient are presented in Table 14.

The results of the corrected and uncorrected of the percentage of feed to the underflow are plotted against the nominal size and illustrated in Figure 14. The results showed that the cut point (d₅₀) was 2.2 µm and the corrected d_50(C) was 3.9 µm.

*Arithmetic mean.

*Arithmetic mean.

The results showed that passing the overflow of the 2-inch hydrocyclone through 10 mm hydrocyclone reduced the d₅₀ of the products from 16 µm to 3.9 µm. The kaolin content increased from 70.6% to 82.5% in the overflow products of the 2-inch and 10 mm hydrocyclones respectively.

The overflow and the underflow products from the 2-inch and 10 mm hydrocyclones were tested by WBB Company [6] for number of physical properties. The results of the size distribution and quantity of minerals are presented in Table 15, while the brightness and colour measurements are shown in Table 16.

Brightness is defined as the ratio, expressed as percentage, of the radiation reflected by a body to the radiation reflected by a perfect reflecting standard (e.g. BaSO₄) measured at 457 nm wavelength. Whiteness is expressed as the difference between the percentage reflectance at 570 and 457 nm. L, a, b colour space is the CIE standard of colour measurement which determine the relative colour values of a sample by detecting reflections with a defined wavelength [9]. Values of L, a, and b are usually plotted on a chromaticity diagram.

The results showed that the ISO brightness of the starting material (71.5 to 73.4) were not startling while the fired brightness (87.3 to 90.2) were good being comparable with typical commercial products from English China Clay (ECC). The fired brightness of 90.2 for the 10 mm hydrocyclone overflow was probably partly due to a low degree of vitrification because of the low alkali (Na and K) content of the product. The good results of the fired brightness were encouraging for its potential use in the ceramic industry whereas the raw brightness was less important.

The silica alumina ratio was a little higher than desirable as the alumina content should ideally be more than 35% - 36%. The iron oxides content was considered low when compared with typical products. The results of the kaolin product are displayed against commercial kaolin products used as fillers and in the ceramic industry (Table 17).

*At 1200˚C; L: lightness; a & b: chromaticity coordinates.

Kaolin for paper manufacturing has to meet stringent specifications with regard to brightness, viscosity, particle size distribution, and abrasiveness. Due to the relatively high amount of quartz and low raw brightness, it is not expected that the concentrated kaolin would meet the paper manufacturing specification.

As filler, further classifying would be necessary to reduce the amount of fine quartz and increase the kaolinite content in order to upgrade it to the required specification. In the ceramic industry, despite the good fired brightness, low iron oxides, and accepted grain size distribution, the kaolinite content was still marginal for porcelain production. However, it could be used for other types of ceramic such as tableware, sanitaryware and earthenware when the results were compared with other kaolin produced by WBB.