Received 24 May 2016; accepted 5 July 2016; published 8 July 2016
Geochemical studies of sediments are helpful in understanding the different sediment sources, element distribution pattern and evaluating the environmental conditions existing in an area. The mineralogical properties of sediments reflect the geological history of transport and sorting process. The dissolved chemical load and sediment flux of the Brahmaputra river has significantly higher rates of physical and chemical weathering than other large Himalayan catchments  -  . In total, the Brahmaputra carries over 73 million tons of dissolved material annually, which accounts for approximately 4% of the total dissolved flux to the oceans  . The focus on mineralogical, geochemical and geophysical studies and chemical composition of sediments of many Indian rivers were done by many authors  -  . As river sediments act as both source and sink for heavy metals therefore contaminants may eventually pass through the food chain and result in a wide range of adverse environmental effects. The estimation of silicate distribution in sediments is important because the total CO2 consumption by silicate weathering can be approximated by the total molar charge equivalents of all cations generated by silicate weathering. In many weathering environment, the chemical weathering of silicate minerals results in the formation of secondary clays. The heavy metal contaminations and silicate mineral distribution in sediments due to weathering of the Brahmaputra river tributaries have been discussed elsewhere by Saikia et al.   . This spectroscopic study is conducted to evaluate the concentration of clay and metals due to the natural and anthropogenic activities of the Brahmaputra river, which helps to assess the ecotoxic potential of the river sediments.
2. Experimental Methods
The surface sediment samples were collected from six locations viz. Sadiya (27˚49?33˝N, 95˚38?54˝E), Dibrugarh (27˚29?22˝N, 94˚54?58˝E), Dikhomukh (26˚59'20''N, 94˚24'42''N), Dhansirighat (26˚41'17''N, 93˚35'55''E), Kaziranga (26˚45?02˝N, 93˚26?49˝E) and Silghat (26˚36'50''N, 92˚55?58˝E) of the Brahmaputra river (Figure 1) and in each locations, five samples were collected at a depth 10 to 30 cm. The sample sites Sadiya, Dibrugarh, Dikhomukh Dhansirighat, Kaziranga and Silghat are denoted S1, S2, S3, S4 and S5 respectively. To eliminate the possibility of materials of the local origin, special care is taken on the sample collection by collecting about 50 - 100 meters away from the stream. Generally, in these sample collection locations, the sediments have been
Figure 1. The Brahmaputra River and its tributaries, the sample collection sites are confined between A (Sadiya) and B (Silghat), of length about 350 km of the upper part of the river.
deposited by the river during the summer season or in flood, therefore, all samples are collected in winter the season.
The samples are allowed to dry and the moisture contents are removed by heating the samples at temperature 110˚C for 10 min. The composition of the sediment samples were determined using a Philips MagiX PRO wavelength dispersive X-ray spectrometer with a rhodium anode X-ray tube was used, which may operated at up to 60 kV and current up to 125 mA, at a maximum power level of 4 kW. The precision and accuracy of the data is ±2%, and average values of three replicates were taken for each determination.
The sediment samples were crushed into fine powder for analysis. The powdered sample was homogenized in spectrophotometric grade KBr (1:20) in an agate mortar and was pressed with 3 mm pellets using a hand press. The infrared spectrum was acquired using Perkin-Elmer system 2000 FTIR spectrophotometer with helium? neon laser as the source reference, at a resolution of 4 cm?1. The spectra were taken in transmission mode in the region 400 - 4000 cm−1. The room temperature was 30˚C during the experiment. Raman spectra were collected using a Ar+ excitation source having wavelength 488 nm coupled with a Jobin-Yvon Horiba LabRam-HR Micro Raman spectrometer equipped with an Olympus microscope with 10×, 50× and 100× objectives and a motorized x − y stage and using 1800 gr./mm grating in the range from 100 to 4000 cm−1. Spectra were generally collected with counting times ranging between 10 and 60 s.
The enrichment factor (EF), contamination factor (CF), index of geo-accumulation (Igeo) and pollution load index (PLI) of the study sediments samples are ascertain by using the standard methods discussed elsewhere  .
3. Results and Discussions
The Raman spectra of the samples and spectral positions are tabulated in the Figure 2 and Table 1 respectively. The observed intense peaks between 100 to 200 cm−1 of the Raman spectra demonstrates the presence of clay minerals. The intense peaks at 144 and 123 cm−1 of the samples (S3 and S4) demonstrates the presence of kaoli-
Figure 2. Raman spectra of the sediment samples of the Brahmaputra River.
Table 1. Comparative peak positions (in cm−1) of Raman and infrared spectra with possible assignments and chemical phases.
nite, because in general, the kaolinite minerals are characterized by very intense bands around 143 cm−1  -  . The bands in this region are attributed to the symmetric bending modes of the O-Si-O and O-A1-O groups. The observed frequency at 144 cm−1 is attributed to the v2(E) mode of the AlO6 octahedron and the frequency at 123 cm−1 is attributed to the out of plane vibration of the Si2O5. The other bands in between 161 - 169 cm−1 are attributed to Raman active Eg(v2) vibration. The peaks around 161 - 169 cm−1 are also characteristic to anatase. The observed bands in between 212 - 220 cm−1 and 267 - 276 cm−1 are attributed to the vibrational modes B2(v3) and Al(vl) respectively. The bend around 212 - 220 cm−1 arises due to Fe-O (Hematite). Magnetite shows its main Raman peak near 667 cm−1, and is distinguishable from other Fe-oxides of structure, such as chromite, spinel, gahnite and franklinite. The peaks in the range 663 - 668 cm−1 (A1g) of all samples are attributed to the existence of magnetite. The other peaks at 301 - 306 cm−1 (Eg) of all spectra are indicative to magnetite in the samples. The peaks at 422 - 434 cm−1of all spectra are indicative to rutile in the samples. The peaks at 327 - 329 cm−1 and 388 - 391 cm−1 in the spectra of the samples S3, S4 and S5 correspond to the v2(E) mode of the SiO4 tetrahedron. The Si-O-Si stretching vibration is observed between 637-645 cm−1 in the samples S1 and S2. The band in the region 749 - 751 cm−1 is related to the stretching vibration of Si-O bonds. The spectral region 455 - 467 cm−1 and 790 - 792 cm−1 are observed in all samples and these bends are assigned to quartz. The Raman peaks due to feldspar is observed in between 504 - 512 cm−1. The peak at 511 cm−1 is indicative of albite. The relative intensities of the bends in the region 464 and 504 cm−1 is indicative to the presence of various amounts of moganite intergrowth with the dominant quartz in all the studied samples   . The peak observed in sample S3 at 388 cm−1 and S5 at 391 cm−1are very nearer to the main peak of goethite occurs at 386 cm−1 which suggestive to presence of goethite in the sample. The geikielite (MgTiO3) has a characteristic Raman peaks found at around 720 cm−1 and 490 cm−1. A weak band at 722 - 724 cm−1 and 486 - 488 cm−1 is observed in the spectra which indicate the presence of geikielite in the samples. Generally, the montmorillonite exhibits a peak near 705 cm−1 and it can be assigned to Si-O-Si vibration. All observed samples exhibit the peak 704 - 706 cm−1 in this region. The observed bends in between 910 - 920 cm−1 and 926 - 946 cm−1 reveals the bending vibrations of the inner hydroxyl and plain bending vibrations of the surface hydroxyls of kaolinite respectively  .
The infrared spectra of the studied samples represented in Figure 3 and the spectral positions are tabulated in Table 1. The infrared spectra have shown bands between 1200 - 450 cm?1 confirms the existence of quartz one
Figure 3. Infrared spectra of the sediment samples of the Brahmaputra River.
of the non clay mineral and invariably present in all samples. The presence of quartz in the samples can be explained by Si-O asymmetrical bending vibrations, Si-O symmetrical bending vibrations, Si-O symmetrical stretching vibrations at around 464 cm−1, 694 cm−1 and 778 cm−1 respectively. The observed doublet at 914 and 936 cm−l can also be recognized by kaolinite. The infrared peak corresponding to the range 536 - 541 cm−1 is arising due to Si-O asymmetrical bending vibrations and 641 - 649 cm−1 is arising due to Al-O-coordination vibrations and these peaks are indicative to the presence of orthoclase feldspar   . In the infrared spectra, the observed band at 777 - 780 cm?1 is arises due to Si?O symmetrical stretching vibration (v1), the band at 693 - 696 cm−1 is arise due to Si-O symmetrical bending vibration (v2), and the peaks around 468 cm?1 is arise due to Si?O asymmetrical bending vibration (v4) are indicative to quartz. The Si-O symmetrical bending vibrational peak at 695 cm?1 of the octahedral site symmetry is unique to the crystalline materials. All infrared spectra reveals peak at this range, therefore crystalline quartz particles present in the observed samples  -  . The infrared spectra reveals bands at 1014 - 1018 cm−1 are close to the SiO deformation band obtained for kaolinite. The absorption band at 1116 - 1120 cm−1 is identical to the Si-O normal to the plane stretching. The observed bands in the range 916 - 920 cm−1 are assigned to (Al-Al-OH) deformation respectively. The peaks around 920 cm−1 are attributed to presence of illite   . With the view of Keller and Pickett, 1949, the observed absorption peaks at 1615 - 1620 cm−1 in some sites indicate the presence of quartz in river sediments are weathered from metamorphic origin  . The infrared peak positions at 1614 and 1620cm−1 observed in the sample S3 and S1 respectively have good agreement with the observation on the quartz mineral obtained by Ramasamy et al., and Saikia et al.    .
The doubly degenerate symmetric stretch (ν3) at the region 1510 - 1521 cm−1of the infrared spectra are indicative to carbonates. All infrared spectra exhibits weak absorption bands at 2840 - 2842 cm−1 and 2924 - 2956 cm−1 arises due to symmetric and asymmetric stretching of CH group which suggest the presence of organic carbon in the studied samples   . The OH stretching modes of vibrations in between 3600 to 3800 cm−1 are observed in all samples. Generally four bands were found in this region at around 3620, 3649, 3664 and 3686 cm−1. These bands were arises due to the v4, v3, v2 and v1 stretching modes of vibrations. The comparative band positions of infrared and Raman are presented in the Table 1. The observed band positions in this region are similar to that of the band found for kaolinite. The variation or position shift of the OH stretching modes indicates the disorder nature of kaolinite in the samples. The frequency vibrations 3681 - 3699 cm−1 (v1), 3664 - 3667 cm−1 (v2) and 3650 - 3654 cm−1 (v3) are due to the three inner surface hydroxyls whereas the vibrations at 3620 - 3623 cm−1 (v1) is due to the inner hydroxyl  . The v1 band observed in infrared spectra around 3620 cm−1 has been assigned to the inner hydroxyl of kaolinite by many authors  -  . Generally the bands v1, v2 and v3 are arises due to the coupled antisymmetric vibrations, symmetric vibrations and due to symmetry reduction from an inner surface hydroxyl respectively    .
The oxide composition of the sediments in sample site S1 to S5 is estimated as: SiO2 (66.74 ± 2.07 wt%), Al2O3 (22.99 ± 2.14 wt%), Fe2O3 (2.04 ± 0.74 wt%), MgO (2.88 ± 1.25 wt%), MnO (0.09 ± 0.08 wt%), CaO (0.72 ± 0.17 wt%), Na2O (0.95 ± 0.20 wt%), K2O (1.07 ± 0.66 wt%) and TiO2 (0.94 ± 0.13 wt%). The metal concentrations in the sediment samples of Brahmaputra river are presented in the Table 2. The concentrations of the elements are compared with different reference data and results of the previous worker Subramanian et al.   . Average concentrations of Al, Fe, Ni, Pb, Ti, Zn, K, Ca, Co and Cr are found to be below of their respective reference values. Whereas the concentrations of Si, Mg, Mn and Cu has greater average values than the respective reference values. The concentration of K, Ca and Cr are slightly below the results of the previous worker Subramanian et al.   . The world surface rock represents the average lithology subjected to weathering in the hydrosphere. The world surface rock average prescribed by Martin and Meybeck is used as background value for investigation of enrichment factor (EF), contamination factor (CF), index of geo-accumulation (Igeo ) and pollution load index (PLI) of the sediments samples  . The average concentrations of all observed elements except Si and Mg have less than the world surface rock average as background level. The enrichment factor, contamination factor and geo-accumulation index of the study samples were depicted in Table 3.
Titanite is a common accessory mineral in sediments from the igneous and metamorphic origin and has affects low due to weathering. The strong positive correlation of Ti with Ca (0.96) suggests the presence of titanite minerals in the samples. The positive correlation of Ti with Mn (0.94) suggests the presence of pyrophanite (MnTiO3). The presence of MnTiO3 in the tributaries of Brahmaputra has been already reported by Saikia et al.   . The elements Pb and Fe expressed a strong positive correlation with Zn, Co and Mg, Mn, Ti, K, Ca respectively at 0.05 level. The other elements such as Al has strong positive correlation with Fe, Mg, Mn, Ti, K
Table 2. Comparative concentation of elements in Brahmaputra river sediments (in ppm).
*Brahmaputra basin sediment (BBS)  ; Indian river sediment average (IRSA)  ; Worlds river average (WRA); Worlds sur face rock average(WSRA)  ; Worlds soil average (WSA)  .
Table 3. Pearson’s correlation coefficient between metal elements of the Brahmaputra river sediments ( p < 0.05).
and Ca; Cu has strong positive correlation with Ni, Pb, Zn and Co; Ti has strong positive correlation with Ca; Zn has strong positive correlation with Co at this level of significance (Table 3). The strong correlation indicates that these elements have common sources. The strong positive correlation among Al, Fe, Mg and K suggests their association with clay.
The possible anthropogenic impact in the sediment is ascertain by enrichment factor (EF) based on the standardization of the analyzed element against a reference element. The element which has low occurrence variability is considered as a reference element. Generally geochemical normalization of the heavy metals data to a conservative element, such as Al, Si and Fe is employed. In this study Fe is considered as reference element of normalization because natural sources (1.5%) vastly dominate its input    . The calculated enrichment of different elements is presented in Table 4. According to Mmolawa et al., the categories of enrichment factor are deficiency to minimal enrichment (EF < 2); moderate enrichment (2 ≤ EF < 5); significant enrichment (5 ≤ EF < 20); very high enrichment (20 ≤ EF < 40) and extremely high enrichment (EF ≥ 40)  . Table 4 displays the enrichment factor of the all observed elements has a value in the range of minimal enrichment. The enrichment of Cu is relatively higher than other elements.
The index of geo-accumulation (Igeo) is characterized according to the Muller seven grades or classes profile of the geo-accumulation index i.e. the value of sediment quality is considered as unpolluted (Igeo is ≤0, class 0); from unpolluted to moderately polluted (Igeo is 0 - 1, class 1); moderately polluted (Igeo is 1 - 2, class 2); from moderately to strongly polluted (Igeo is 2 - 3, class 3); Strongly polluted (Igeo is 3 - 4, class 4); from strongly to extremely polluted (Igeo is 4 - 5, class 5) and Extremely polluted (Igeo is >6, class 6)  . The calculated Igeo values for all elements were negative (Table 4). Therefore, according to Muller’s classification, Brahmaputra river sediments were unpolluted (class 0). The total index of geo-accumulation (Itot) is defined as the sum of Igeo for all trace elements obtain from the site  . The total index of geo-accumulation for the Brahmaputra river sediment is −3.818 ± 0.593.
The metal contamination level of the sediment is ascertained by the level of contamination proposed by Hakanson  . According to Hakanson the classifications are: low contamination (CF < 1); moderate contamination (1 ≤ CF < 3); considerable contamination (3 ≤ CF < 6) and very high contamination (CF > 6). All elements except Si, Mg, Mn and Cu has low contamination value (Table 4). The sediment is moderately contaminated due to Si, Mg, Mn and Cu. The relative distributions of the contamination factor among the samples are: Cu > Si > Mn > Mg > Ni > Cr > Ti > Al > Co > Pb > K > Ca > Zn. The value of mean pollution load index of the sediments is estimated as 0.771 ± 0.046. The mean pollution load indexes of all sites suggest no overall pollution and are almost identical to the mean pollution load of the Subansiri river  .
Table 4. Enrichment factor, Contamination factor and Geo-accumulation index of the Brahmaputra river sediments.
Raman and infrared spectra indicate the most abundant constituents of the sediments are crystalline quartz with clay minerals which is identical to the compositional results. All infrared spectra of the studied samples exhibit peaks near 695 cm−1 which indicative to the presence of micro-crystalline quartz particles in the sediment samples. The identical clays are kaolinite, montmorillonite and illite, The other constituents present in the sediment are titanite, hematite, magnetite, pargasite, moganite, geikielite, feldspars (orthoclase, albite), carbonates and organic compounds. The presence of infrared absorption peaks in between 1614 - 1620 cm−1 is indicative to the weathered metamorphic origin of the silicate minerals. The observed positive correlation between Ti and Mn is indicative to the presence of pyrophanite (MnTiO3) mineral from the metamorphosed manganese deposition in the adjoin areas. The strong positive correlation among Al, Fe, Mg and K suggests their association with clay. The Raman peaks (at 722 - 724 cm−1 and 486 - 488 cm−1) and positive correlation of Ti and Mg (0.97) are indicative to presence of geikielite (MgTiO3) in the samples. The enrichment factor and contamination factor has a minimal value. The mean pollution load indexes of all sites suggest no overall pollution. The overall sediment is moderately contaminated due to Si, Mg, Mn and Cu. The relative distributions of the contamination factor among the samples are: Cu > Si > Mn > Mg > Ni > Cr > Ti > Al > Co > Pb > K > Ca > Zn. The negative value of geo-accumulation index indicates that the mean concentrations of metals Brahmaputra river sediments are lower than world surface rock average.
We thank Directors, National Geophysical Research Institute (NGRI-CSIR), Hyderabad, Indian Institute of Technology, Guwahati (IITG) and North East Institute of Science and Technology (NEIST-CSIR), Jorhat for their cooperation during this work. We also thank Dr. J.R. Chetia, Dibrugarh University, Dibrugarh, for his assistance in the FTIR analysis.
 Sarin, M.M., Krishnaswami, S., Dilli, K., Somayajulu, B.L.K. and Moore, W.S. (1989) Major Ion Chemistry of the Ganga-Brahmaputra River System: Weathering Processes and Fluxes to the Bay of Bengal. Geochimica et Cosmochimica Acta, 53, 997-1009.
 Harris, N., Bickle, M.J., Chapman, H., Fairchild, I. and Bunbury, J. (1998) The Significance of Himalayan Rivers for Silicate Weathering Rates: Evidence from the Bhote Kosi Tributary. Chemical Geology, 144, 205-220.
 Galy, A. and France-Lanord, C. (1999) Weathering Processes in the Ganges-Brahmaputra Basin and the Riverine Alkalinity Budget. Chemical Geology, 159, 31-60.
 Galy, A. and France-Lanord, C. (2001) Higher Erosion Rates in the Himalaya: Geochemical Constraints on Riverine Fluxes. Geology, 29, 23-26.
 Dalai, T.K., Krishnaswami, S. and Sarin, M.M. (2002) Major Ion Chemistry in the Headwaters of the Yamuna River System: Chemical Weathering, Its Temperature Dependence and CO2 Consumption in the Himalaya. Geochimica et Cosmochimica Acta, 66, 3397-3416.
 Singh, S.K. and France-Lanord, C. (2002) Tracing the Distribution of Erosion in the Brahmaputra Watershed from Isotopic Compositions of Stream Sediments. Earth and Planetary Science Letters, 202, 645-662.
 Singh, S., Sarin, M.M. and France-Lanord, C. (2005) Chemical Erosion in the Eastern Himalaya: Major Ion Composition of the Brahmaputra and d13C of Dissolved Inorganic Carbon. Geochimica et Cosmochimica Acta, 69, 3573-3588.
 Subramanian, V., Van’t Dack, L. and Grieken, V. (1985) Chemical Composition of River Sediments from the Indian Subcontinent. Chemical Geology, 48, 271-279.
 Subramanian, V., Grieken, R.V. and Dack, L.V. (1987) Heavy Metals Distribution in the Sediments of Ganges and Brahmaputra Rivers. Environmental Geology and Water Sciences, 9, 93-103.
 Ramesh, R., Subramanian, V. and Van Grieken, R. (1990) Heavy Metal Distribution in Sediments of Krishna River Basin, India. Environmental Geology and Water Sciences, 15, 303-324.
 Chakrapani, G.J. and Subramanian, V. (1990) Preliminary Studies on the Geochemistry of the Mahanadi River Basin, India. Chemical Geology, 70, 247-266.
 Singh, M., Ansari, A.A., Muller, G. and Singh, I.B. (1997) Heavy Metals in Freshly Deposited Sediments of the Gomti River (a Tributary of the Ganga River): Effects of Human Activities. Environmental Geology, 29, 246-252.
 Dekov, V.M., Araújo, F., Van Grieken, R. and Subramanian, V. (1998) Chemical Composition of Sediments and Suspended Matter from the Cauvery and Brahmaputra Rivers (India). Science of the Total Environment, 212, 89-105.
 Braun, J.J., Descloitres, M., Riotte, J., Fleury, S., Barbiero, L., Boeglin, J., Violette, A., Lacarce, E., Ruiz, L., Sekhar, M., Kumar, M.S.M., Subramanian, S. and Dupré, B. (2009) Regolith Mass Balance Inferred from Combined Mineralogical, Geochemical and Geophysical Studies: Mule Hole Gneissic Watershed, South India. Geochimica et Cosmochimica Acta, 73, 935-961.
 Saikia, B.J., Goswami, S.R. and Borah, R.R. (2014) Estimation of Heavy Metals Contamination and Silicate Mineral Distributions in Suspended Sediments of Subansiri River. International Journal of Physical Sciences, 9, 475-486.
 Saikia, B.J., Goswami, S.R., Borthakur, R., Roy, I.B. and Borah, R.R. (2015) Spectroscopic Characterization and Quantitative Estimation of Natural Weathering of Silicates in Sediments of Dikrong River, India. Journal of Modern Physics, 6, 1631-1641.
 Michaelian, K.H., Bukka, K. and Permann, D.N.S. (1987) Photoacoustic Infrared Spectra (250 - 10,000 cm-1) of Partially Deuterated Kaolinite. Canadian Journal of Chemistry, 65, 1420-1423.
 Gotze, J., Nasdala, L., Kleeberg, R. and Wenzel, M. (1998) Occurrence and Distribution of “Moganite” in Agate/ Chalcedony: A Combined Micro-Raman, and Cathodoluminescence Study. Contributions to Mineralogy and Petrology, 133, 96-105.
 Saikia, B.J., Parthasarathy, G. and Sarmah, N.C. (2008) Fourier Transform Infrared Spectroscopic Estimation of Crystallinity in SiO2 Based Rocks. Bulletin of Materials Science, 31, 775-779.
 Saikia, B.J. (2014) Spectroscopic Estimation of Geometrical Structure Elucidation in Natural SiO2 Crystal. Journal of Materials Physics and Chemistry, 2, 28-33.
 Saikia, B.J. and Parthasarathy, G. (2010) Fourier Transform Infrared Spectroscopic Characterization of Kaolinite from Assam and Meghalaya, Northeastern India. Journal of Modern Physics, 1, 206-210.
 Saikia, B.J., Parthasarathy, G. and Borah, R.R. (2015) Distribution of Microcrystalline Quartz in Glassy Fulgurites from Garuamukh and Kimin, India. Journal of Applied Mathematics and Physics, 3, 1343-1351.
 Clark, R.N., King, T.V.V., Kiejwa, M., Swayze, G.A. and Verge, N. (1990) High Spectral Resolution Reflectance Spectroscopy of Minerals. Journal of Geophysical Research, 95, 12653-12680.
 Saikia, B.J., Parthasarathy, G. and Sarmah, N.C. (2009) Fourier Transform Infrared Spectroscopic Characterization of Dergaon H5 Chondrite: Evidence of Aliphatic Organic Compound. Nature and Science, 7, 45-51.
 Saikia, B.J., Parthasarathy, G., Sarmah, N.C. and Baruah, G.D. (2008) Fourier-Transform Infrared Spectroscopic Characterization of Naturally Occurring Glassy Fulgurites. Bulletin of Materials Science, 31, 155-158.
 Johnston, C.T., Agnew, S.E. and Bish, D.L. (1990) Polarised Single Crystal Fourier-Transform Infrared Microscopy of Ouray Dickite and Keokuk Kaolinite. Clays and Clay Minerals, 38, 573-583.
 Brindley, G.W., Chih-Chun, K., Harrison, J.L., Lipsiscas, M. and Raythatha, R. (1986) Relation between the Structural Disorder and Other Characteristics of Kaolinites and Dickites. Clays and Clay Minerals, 34, 233-249.
 Tippie, V. (1984) An Environmental Characterization of Chesapeake Bay and a Framework for Action. In: Kennedy, V., Ed., The Estuary as a Filter, Academic Press, New York, 467-487.
 Mmolawa, K., Likuku, A. and Gaboutloeloe, G. (2011) Assessment of Heavy Metal Pollution in Soils along Roadside Areas in Botswana. African Journal of Environmental Science and Technology, 5, 186-196.
 Ya, Z.G., Zhou, L.F., Bao, Z.Y., Gao, P. and Sun, X.W. (2007) High Efficiency of Heavy Metal Removal in Mine Water by Limestone. Chinese Journal of Geochemistry, 28, 293-298.