Arsenic removal using biosorption with Chitosan

(1)

Arsenic removal using biosorption with Chitosan

Evaluating the extraction and adsorption performance of Chitosan from shrimp

shell waste

A Minor Field Study

MSc. Thesis

Robin Westergren

(2)

Arsenic removal using biosorption with Chitosan

Evaluating the extraction and sorption performance of Chitosan from shrimp

shell waste

Robin Westergren

2006

Supervisors

Lic. Martha Benavente Assoc. Prof. Olle Wahlberg

Facultad de Ingeniería Quimica Department of Inorganic Chemistry Universidad Nacional de Ingeniería (UNI) Royal Institute of Technology (KTH) Managua, Nicaragua Stockholm, Sweden

Assoc. Prof. Joquín Martínez

Department of Chemical Engineering Royal Institute of Technology (KTH)

Stockholm, Sweden

Examiner

Prof. Lars Kloo

Department of Inorganic Chemistry Royal Institute of Technology (KTH)

(3)

Preface

This study has been carried out within the framework of the Minor Field Studies Scholarship

Programme, MFS, which is funded by the Swedish International Development Cooperation

Agency, Sida /Asdi.

The MFS Scholarship Programme offers Swedish university students an opportunity to carry

out two months’ field work, usually the student’s final degree project, in a Third World country.

The results of the work are presented in an MFS report which is also the student’s Master of

Science Thesis. Minor Field Studies are primarily conducted within subject areas of importance

from a development perspective and in a country where Swedish international cooperation is

ongoing.

The main purpose of the MFS Programme is to enhance Swedish university students’

knowledge and understanding of these countries and their problems and opportunities. MFS

should provide the student with initial experience of conditions in such a country. The overall

goals are to widen the Swedish human resources cadre for engagement in international

development cooperation as well as to promote scientific exchange between unversities,

research institutes and similar authorities as well as NGOs in developing countries and in

Sweden.

The International Office at KTH, the Royal Institute of Technology, Stockholm, administers

the MFS Programme for the faculties of engineering and natural sciences in Sweden.

Sigrun Santesson

Programme Officer

MFS Programme

International Office, MFS

(4)

Summary

Nicaragua is a country in which the toxic metal contamination of freshwater resources has become an increasingly important problem in certain regions posing a threat to the environment as well as to human health. Among the metals found in the waters of Nicaragua, arsenic is one of the most problematic since its long time consumption is connected to serious health problems such as cancer and neurological disorders. The arsenic contamination of water recourses in Nicaragua is mostly attributable natural factors, even though anthropogenic activities including gold mining may be a contributing factor.

In this work the biopolymer Chitosan was studied as a potential adsorption material for the removal of arsenic from aqueous solutions for water treatment design purposes.

The Chitosan used in this study was extracted from shrimp shells with an overall yield of 40% and a deacetylation grade of 59%. The maximum adsorption capacity was determined to 20.9 mg As/g at a controlled pH of 5.5 using the Langmuir isotherm. The adsorption was found to be strongly pH dependant with a fourfold increase in adsorption capacity when pH was well under the pKa of Chitosan. The pH dependence indicates that ionic exchange was the most important mechanism. No difference in adsorption capacity with respect to the initial pH of the solution was detected in the pH range 3-7. This was attributed to the ability of Chitosan to act as a weak base in water solutions.

The arsenic was desorbed from Chitosan using NaOH, (NH4) 2SO 4 and NaCl, with a 1M NaOH solution being the most efficient displaying a concentration ratio of 1.08. The NaOH and (NH4) 2SO 4 solutions displayed a steep desorption curvature with a large fraction of the arsenic being easily desorbed. The arsenic was, however, not completely desorbed from the Chitosan implying that the adsorption capacity would decrease for the coming cycles. Being a biopolymer the Chitosan is quite easily degraded in acid and alkali solutions, which might be a limiting step for the process applicability.

(5)

Resumen

En Nicaragua, la contaminación de los recursos hídricos con metales tóxicos se ha convertido en un importante problema, el cual se ha ido incrementando en ciertas regiones planteando una amenaza al medio ambiente y a la salud humana. Entre los metales encontrados en las aguas de Nicaragua, el arsénico es uno de los más perjudiciales ya que su consumo por un tiempo relativamente largo esta relacionado a serios problemas de salud tales como cáncer y desordenes neurológicos. La contaminación de arsénico, en los recursos acuáticos en Nicaragua, se atribuye mayormente a factores naturales, aunque las actividades antropogénicas, incluyendo la minería para la extracción de oro, puede ser un factor que contribuya a incrementar los niveles de contaminación.

En este trabajo, se estudió el biopolímero quitosana como un potencial material adsorbente para la remoción de arsénico de soluciones acuosas para propósitos de diseño en el tratamiento de agua.

La quitosana usada en este estudio fue extraída del caparazón de camarón con un rendimiento global del 40% y un grado de desacetilación del 59%. Utilizando la isoterma de Langmuir, se determinó que la máxima capacidad de adsorción fue de 20.9 mg As/g a un pH controlado de 5.5. Se encontró que la adsorción depende fuertemente del pH con un aumento cuatro veces mayor en la capacidad de adsorción cuando el pH esta por debajo del pKa de la quitosana. La dependencia del pH indica que el intercambio iónico es el mecanismo más importante. Además, se detectó que no hay diferencia en la capacidad de adsorción con respecto al pH inicial de la solución en un rango de pH de 3-7. Esto fue atribuido a la habilidad de la quitosana de actuar como una base débil en soluciones acuosas.

El arsénico fue des-adsorbido de la quitosana usando NaOH 1M, NaOH 0.1M, (NH4)2SO4 1M y NaCl 1M. Los resultados mostraron que con la solución de NaOH 1M, la des-adsorción fue más eficiente, con una relación de concentración de 1.08. Así mismo, con las soluciones de NaOH y (NH4) 2SO4 se obtuvieron curvas de desorción más inclinadas, con respecto a la curva obtenida con la solución de NaCl, indicando que el arsénico es más fácilmente des-adsorbido. Sin embargo, el arsénico no fue completamente des-adsorbido de la quitosana implicando que la capacidad de adsorción decrecerá en los siguientes ciclos.

Ya que la quitosana es un biopolímero, el cual es fácilmente degradado en soluciones ácidas y alcalinas, puede ser un paso limitante para su aplicabilidad en los procesos.

(6)

Sammanfattning

Nicaragua är ett land där förekomsten av toxiska metaller i yt- och grundvatten har blivit ett problem i vissa regioner, då de utgör ett alvarligt miljöhot såväl som ett hot mot människors hälsa. Bland de metaller som påträffats i Nicaraguas vattenresurser så utgör arsenik en av de mest problematiska, då dess konsumtion är förknippad med livshotande sjukdomar som cancer och störningar på centrala nervsystemet. Arsenikföroreningen i Nicaragua är främst av naturligt ursprung även om mänskliga aktiviteter såsom gruvdrift tros bidra till de förhöjda halterna.

I detta arbete studerades biopolymeren Chitosan som ett potentiellt adsorptionsmaterial för rening av arsenikförorenat vatten.

Det Chitosan som användes i denna studie utvanns ur räkskal med ett utbyte på 40% och en deacetyleringsgrad på 59%. Den maximala adsorptionskapaciteten för Chitosan bestämdes till 20.9 mg As/g genom approximering till Langmuir isotermen vid ett kontrollerat pH värde av 5.5. Adsorptionskapaciteten fanns vara beroende av pH med en fyrfaldig ökning då pH var under pKa värdet för Chitosan. pH beroendet indikerar att adsorptionen sker företrädesvis med en jonbytes mekanism. Inga skillnader i adsorptionskapacitet kunde påvisas till följd av skillnader i vattenlösningars initial pH i området 3-7. Detta tillskrivs Chitosanets förmåga att agera som en svag bas i en vattenlösning.

Arseniken kunde desorberas från Chitosan med hjälp av NaOH, (NH4) 2SO 4 och NaCl lösningar. 1M NaOH fanns vara den mest effektiva med ett koncentrations förhållande (CR) på 1.08. NaOH och (NH4) 2SO 4 lösningarna uppvisade branta desorptionskurvor, vilket innebär att en stor del av arseniken lätt kan desorberas. Arseniken kunde dock inte i något försök desorberas fullständigt vilket skulle ge upphov till en försämrad adsorptions kapacitet för de kommande cyklerna. Då Chitosan är en biopolymer bryts den lätt ner under sura och basiska förhållanden, vilket skulle kunna vara en begränsning för processens tillämplighet.

(7)

1 Introduction

1.1 Background

Despite being a country with ample supply of fresh water resources, shortage of safe drinking water has become an increasingly important problem in certain regions of Nicaragua. The increasing contamination of toxic metals have caused a serious degradation of rivers, lakes and ground water reserves and has been recognized as a threat to the environment as well as a threat to the population of affected regions (Benavente et al 2005) . Among the metals found in the natural waters of Nicaragua arsenic is one of the most problematic due to its documented toxic and carcinogenic effects at low concentrations (Espinoza 2005).

A study monitoring the waters of the Zapoyol community in central Nicaragua revealed that people for some time had been consuming ground water contaminated with arsenic (Espinoza 2005). The concentrations in some cases exceeded 100 µg/l, which is ten times the permissible limit stated by the World Health Organization guidelines for drinking water quality. The occurrence of arsenic in Nicaraguan waters can be attributed to both natural and anthropogenic factors and the sources differ between different regions. Groundwater interactions with arsenic rich rocks and geothermal activities often combine with arsenic rich discharges from gold mining activities to complicate remediation strategies (Espinoza 2005).

There exists a number of treatment processes to remove arsenic from water effluents including sulphide precipitation, co-precipitation with iron and metal hydroxides and coagulation processes. The standard methods for removal of arsenic from industrial effluents are, however, often expensive or fail to concentrate arsenic in small waste volumes (Dambies et al. 2002). Thus, there is an urgent need to develop a cost efficient treatment technology capable of separating arsenic from both drinking water and industrial effluents. One method that recently has gained some of attention is biosorption, in which dead biomass is used to concentrate the metals. Chitin and Chitosan are two biopolymers that can be derived from shrimp shells that previously have displayed a high capacity to fix a great variety of metals (Guibal 1999). Since Nicaragua produces about 5.5 thousand tons of shrimps per year, of which the residual shells constitute about 20% of the production volume, there is a plentiful supply of raw material for the production of these polymers (Benavente 2001). This study was therefore initiated to evaluate weather biosorption with locally produced Chitosan could be used to treat arsenic contaminated water in Nicaragua.

1.2 Objectives

(9)

The specific objectives were to:

Produce the Chitosan that was to be used in adsorption experiments from dried shrimp shells at laboratory scale, determine the overall yield and the grade of deacetylation.

Determine the isotherms for arsenic adsorption on Chitosan with special reference to the influence of solution pH.

Screen for effective desorption agents by determining the eluation curves for the different solutions and comparing their overall efficiency.

Use adsorption and desorption data to establish the chemical mechanisms for adsorption.

(10)

2 Arsenic water contamination – A global problem

The occurrence of Arsenic in surface and ground waters can be attributed to a number of different sources of both natural and anthropogenic origin. Due to the chronic toxicological effects of arsenic, increased levels of arsenic have become a problem in several countries among which India, Bangladesh are the most critical with over 40 million persons affected (Smedley and Kinnburgh 2002). It is, thus, an environmental problem of global concern and indeed a very active field of research. Other countries which suffer from elevated levels of arsenic include Argentine, Chile, Mexico, China, Hungary and as will be discussed more thoroughly Nicaragua.

2.1 Arsenic in natural waters

The range of As concentrations found in natural waters is large, ranging from less than 0.5 µg/l to more than 5000 µg/l. Typical concentrations in freshwater are less than 10 µg and often lower than 1µg/l (van Loon and Duffy 2005). Higher concentrations are rarely found in ground waters. In certain high-As regions, however, up to 90 % of the drinking water wells show concentrations exceeding 50µg/l. These large-scale As ground water problems are often found in two types of environment. Firstly, inland or closed basins in arid or semi-arid areas, secondly, in strongly reducing aquifers. Both environments tend to contain geologically young sediments and to be in flat low-lying areas where ground water flow is slow and sluggish (Smedly et al. 2002). Arsenic rich ground waters can also found in areas of geothermal activity and on a more localised scale, in areas of mining activity and in areas where oxidation of sulphide minerals have occurred. Common for most arsenic rich environments is that there is great variability in the As concentrations found within an affected area.

2.2 Aqueous speciation of arsenic

(11)

2 4 6 8 10 12 - 10 0 10 p e pH H₃A sO₃ AsH3 AsO43− H2AsO3− H2AsO4− H3AsO4 HAsO42− As ( c ) [H₃AsO₃]_TOT= 10.00 mM t= 25°C

Figure 2.2.1. pe-pH diagram for aqueous As species in the system As- O2- H2O at 25°C and 1bar total pressure (Smedly et al. 2002).

The redox potential and pH are the variables that determine the speciation of arsenic in a water body. Figure 2.2.1. illustrates that under oxidising conditions H2AsO4- is the dominant species at pH values under 6.9, whilst at higher pH, H2AsO32- becomes dominant. H3AsO4 and AsO43- may be present in extremely acidic and alkaline conditions respectively. Under reducing conditions and pH less than 9.2 the uncharged arsenite species H3AsO30 predominate. 2 4 6 8 10 12 0. 0 0. 2 0. 4 0. 6 0. 8 1. 0 F r a c t i o n pH H₃AsO₃ H₂AsO₃− H4AsO3+ HAsO₃2− [H3AsO3]TOT = 1.00 µM

(12)

2 4 6 8 10 12 0. 0 0. 2 0. 4 0. 6 0. 8 1. 0 F r a c t i o n pH AsO₄3− H2AsO4− H3AsO4 HAsO₄2− [AsO43−]TOT = 1.00 µM

Figure 2.2.3. Arsenate speciation as a function of pH at the redox conditions, where the chosen species dominates.

In Figure 2.2.2. and 2.2.3.the single variable diagrams for Arsenite and Arsenate respectively can be seen. In natural waters organic As can be formed in areas of high biological activity and may be quantitatively important in areas of industrial pollution (Smedley et al. 2002). In the presence of high concentrations of reduced S, dissolved As-sulphide species can be significant.

2.3 Arsenic mobilisation

Compared to other heavy metalloids and oxyanion forming elements such as Se, Sb, Mo, Cr, U and Re, arsenic displays a rather exceptional sensitivity to mobilisation at the pH values found in ground waters (pH 6.5- 8.5) under both reducing and oxidizing conditions (Smedley et al. 2002).

(13)

oxyanion species while its reduced Cr(III) form behaves like other cations at near neutral conditions. Other oxyanions may form insoluble suphides in S-rich reducing conditions. Arsenic is in this way distinctive in being mobile under both reducing and oxidising conditions (Smedley et al. 2002).

Two different triggers have been identified to be responsible for the release of arsenic in groundwaters. The first trigger is the development of a pH>8.5, which is often an effect of high evaporation and weathering rates in arid or semi-arid environments. This elevation of pH enhances the desorption of As from mineral oxides, especially Fe oxides, or prevents As adsorption. The second trigger is the development of strongly reducing conditions at near neutral pH conditions, leading to the desorption of As from mineral oxides and to the reductive dissolution of Fe and Mn oxides, also leading to As release. Thus, in high As areas there is often a strong correlation between high concentrations of As(III) and Fe(II) and typically low sulphate concentrations (van Loon and Duffy 2005). Large concentrations of phosphate, bicarbonate silicate and and possibly organic material can enhance the desorption of arsenic due to the competition for adsorption sites.

2.4 Arsenic toxicity

The valency state of arsenic has an important role for its behaviour in the environment with respect to transport and accumulation. The chemical form in which arsenic is present, however, also is a key factor in assessing its toxicity. Changes in the degree of oxidation are recognized to have an important effect on the degree of bioavailability and its magnification of arsenic in food chains (Jain and Ali 2000).

When considering the toxicity of metals in general, simple hydrated metal ions posses the highest toxicity while strong complexes and species associated with colloidal particles are considered less toxic. Trivalent arsenic species are considered to be more toxic than the pentavalent form. Studies of the toxicity to humans are rather limited but Arsenite is about 60 times more toxic than the oxidised arsenate (Jain and Ali 2000). Organometallic compounds of tin, mercury and lead are often more toxic than the corresponding inorganic species. This is particularly true for simple methylated species. Organoarsenic compounds, however, pose an exception to this rule since they reduce their toxicity when methylated (Jain and Ali 2000). The organoarsenic compounds are about 100 times less toxic than inorganic arsenic compounds. Methylation of inorganic arsenic has actually been described as the most important detoxification process in the human body since it reduces the affinity of the compound for tissue (Vahter and Marafante 1988).

(14)

2.5 Arsenic water contamination in Nicaragua

The number of studies previously conducted to determine the occurrence of arsenic in Nicaraguan ground and surface waters are quite few there is little information in literature covering the topic. The knowledge about sources of arsenic as well as natural release mechanisms is therefore quite sparse. A geo physical survey performed by CIRA in the (Centro para la Investigación en Recursos Acuáticos de Nicaragua), however, established that Sebaco-Matagalpa region display increased arsenic groundwater concentrations with severe effects for both the environment and its population (Espinoza 2005).

Of the 57 water samples included in the survey, 21 showed total arsenic concentrations between 10 and 122 µg/l, which is above the guidelines established for drinking water. The highest concentrations of arsenic where found in the community of El Zapote. The total arsenic concentration in rocks and soil were 14.98 µg g-1_and 57.19 µg g-1_{respectively, water concentrations were found to be 122.15 µg l}-1_{. In this community cases of} hydroarsenicism were reported in the year 1996 from people drinking water from a well for a period of 6 months with an arsenic concentration of 1,320 µg l-1. People from this community were affected by irreversible health problems, which forced the populations to change its water supply. The dug well used by the community today displays an arsenic concentration of 122.15 µg l-1_{, which is still above the guideline for drinking water} (Espinoza 2005). The elevated arsenic concentrations of the Sebaco-Matagalpa region associates with the hydrothermal activity originated from tertiary volcanism and are thought to spread in a NW-SE path (Espinoza 2005). The area affected by arsenic contamination might therefore be larger than the area covered by the study. Except from the natural release of arsenic, there is also an anthropogenic input of arsenic from the gold mining industry in certain regions. Since the gold ore findings occur together with arsenopyrite large quantities of arsenic is displaced in the ore tailing dumps (Smedley et. al 2002). These tailing dumps are thereafter subject to subsequent leakage of various metals including arsenic (Bennavente 2005). As higher grade ores are depleted more complex sulphide ores and concentrates are being processed. The processing of more complex materials results in higher levels of impurities such as arsenic in the process streams, which will make this an increasingly important issue in the future (Lorenzen et al. 1995).

2.6 Arsenic removal methods

As discussed earlier the chemistry of arsenic is quite different from other toxic elements which are usually found as cations in natural water. Thus normal precipitation is insufficient for the removal of arsenic and other methods have been developed for arsenic. Sulphide and co-precipitation with iron hydroxides are often used for as an alternative waste water treatment (Lorenzen et al. 1995).

(15)

3 The Adsorption Process

In terms of process engineering, adsorption is a chemical separation technique involving a fluid flowing over a solid. The strength of adsorption is its capacity to remove mere traces of solutes, making this method especially useful for pollution control of dilute solutions. Compared to other chemical separation methods the adsorption process depends to great extent on experimental data of the solution-solid interaction for design purposes.

3.1 Adsorption mechanisms and materials

A common characteristic of adsorbent materials is a high porosity and highly irregular geometries which result in a large specific area. Since molecules can adsorb to some extent on all surfaces the amount adsorbed shows proportionality to the surface area of the sorbent. There are, however, several types of adsorption phenomena which all require individual explanation.

Electrostatic attraction to a charged surface is perhaps the most common and occurs on a large variety of materials including clays metal oxides as well as organic materials. The surface charge is often dependent on the properties of the surrounding solution. Iron and aluminium oxides are examples of inorganic sorbent materials where the surface charge depends on whether they occur in their protonated (positive), neutral or deprotonated (negative) state. Organic material, analogously, often display a variable surface charge with the deprotonation of carboxyl groups or protonation of amino groups. The electrostatic attraction is therefore strongly dependant of the solution pH.

The electrostatic interactions between sorbent and sorbate are more correctly described in terms of an ion exchange process where co-ions may replace each other. The equilibrium position of the reaction depends on the nature of the sorbent and the nature and concentration of the dissolved species adjacent to the sorbent (van Loon and Duffy 2005).

In addition to electrostatic interactions molecules may also be adsorbed by covalent binding. Such chemical processes are often referred to as specific adsorption and may be irreversible. Here the adsorption depends on a chemical, as opposed to, electrostatic affinity between the sorbent and the species in solute. In adsorption processes involving covalent bounding there is a weaker dependence of surface charge and thus also of the solution pH. A combination of electrostatic and covalent bounds often occurs to complicate the picture (van Loon and Duffy 2005).

3.2 Diffusion and mass transfer

(16)

controlling steps involved in the adsorption process are therefore mainly intraparticle diffusion and the actual sorption through electrostatic interactions or covalent binding.

Since adsorption by electrostatic interactions is more or less an instantaneous process, while complexation and chelation often occur with a lower rate some conclusions about the sorption mechanism can be drawn from the kinetics of the reaction (Guibal 1995).

3.3 Adsorption equilibrium and isotherms

In order to design an adsorption process the isotherms for each solute-sorbent system need to be determined experimentally. The isotherms describe the chemical equilibrium for the specific conditions of temperature, pH, co-ion existence etc. and thus represent the maximum achievable adsorption capacity for a given system. The plotted isotherm shows the amount adsorbed as a function of the solution equilibrium concentration. A downward curvature is often referred to as a favourable sorption characteristic since the sorption capacity increases rapidly for low equilibrium concentrations. This terminology implies that adsorption is commonly used to capture small amounts of solutes from dilute solutions.

To quantitatively describe the adsorption, mathematical descriptions are often used to withdraw information from the adsorption data. Three commonly cited isotherm models are the linear, Freundlich and Langmuir which implicitly hold information about the sorption mechanism. The Linear isotherm, which is rarely occurs but is sometimes assumed for its simplicity, is given as.

eq

KC

q

=

(Eq. 3.3.1)

Where q is the concentration in the adsorbent and Ceq is the concentration in solution. The linear isotherm simply states that the amount adsorbed is proportional to the concentration of solute. A fact that is often true in a certain range of the isotherm but usually does not fit adsorption data over an entire isotherm.

Compared to the linear isotherm the Langmuir model uses some theoretical basis to better describe the adsorption process. By assuming that there are a limited number of sites on the adsorbent and a monolayer adsorption a mass balance can be stated.













+













=













sites

empty

sites

filled

sites

total

(Eq. 3.3.2)

The sites are further subject to a chemical equilibrium













+













site

empty

solute

bulk

_











site

filled

(Eq. 3.3.3) In quantitative terms introducing the equilibrium constant

(17)

Combining Equation 3.3.2 and 3.3.4 gives the expression













+

























=













solute

bulk

K

solute

bulk

sites

total

K

site

filled

1

(Eq. 3.3.5) Or eq eq m

KC

q

+

=

1

(Eq. 3.3.6)

Were qm is the monolayer saturation capacity (mg/g sorbent) and K the equilibrium constant. By testing this plot with adsorption data we can get some information about the sorption mechanism.

The third of the common isotherms called the Freundlich isotherm is given by

n

Ky

q

=

(Eq. 3.3.7)

Were both K and n are Freundlich constants. The Freundlich relation is differs from that of Langmuir in that it does not consider all the sites on the adsorbent to be equal, but rather that adsorption becomes increasingly difficult as the sorbent becomes more saturated. Furthermore, the Freundlich assumes that there can be interactions between the adsorbed molecules and that multilayer adsorption is possible. Thus the equation does not propose any particular mechanism of adsorption but assumes several different interactions between the sorbent and the adsorbate. The Freundlich model is strictly empirical, but has been found to fit adsorption data of small molecules in small concentrations (van Loon and Duffy 2005).

3.4 Adsorption in a continuous systems

The isotherm gives, as stated above, the maximum achievable adsorption capacity for a given solute sorbent system. In a system with a continuous flow of solute equilibrium is rarely the case and a lower adsorption capacity is therefore to be expected. When using a packed bed adsorption system for water treatment purposes there are, however, some advantages compared to a stirred tank.

The compact packed bed permits a faster mass transfer from bulk solution to the sorbent than any stirred tank analogue (Cussler 1997). Although, the main advantage of using a packed bed is the effect of a counter current flow with a more effective treatment as result. In the case of the tank reactor the equilibrium is reached with the average depleted solution, while the packed bed approaches equilibrium with the concentrated feed solution (Cussler 1997).

3.5 Regeneration of the sorbent

(18)

for the sorption-desorption process can be defined as a concentration ratio between regeneration fluid and the incoming water for a packed bed system (Volesky 1999).

]

[

]

[

Feed

Eluate

CR

=

(Eq. 3.5.1)

Another important parameter that often determines the efficiency of the regeneration that can be used in batch experiments is the solid to liquid ratio (S/L ratio) (Jeon and Ha Park 2005). Combining the S/L ratio with the desorption ratio gives an expression of the concentration ratio of a batch experiment similar to that stated above. Generally as the S/L ratio increases as the desorption ratio decreases, however complete desorption is not always the desirable case, but rather to find an optimum high the concentration ratio while maintaining an acceptable desorption ratio (Jeon and Ha Park 2005).

Some qualitative information about the desorption process is given by the adsorption isotherm. An isotherm which is strongly favourable for adsorption will be unfavourably when it comes to regenerating the sorbent. Unfavourable desorption reflects that it will be difficult to regenerate the sorbent completely which leads to a decreased adsorption capacity for the following cycles (Cussler 1997).

4 Chitosan biosorption

Biosorption is defined as an adsorption process in which biomass is used to concentrate solutes. In the search for cheap and effective adsorbents containing natural polymers Chitin and its derivate Chitosan are very interesting since they hold many of the desired properties for sorbent materials such as being biodegradable and cost efficient (Guibal et al. 1999). One advantage of using biosorption is simply that the sorbent is a waste product that can be recycled.

4.1 Chitin and Chitosan production

(19)

NH O O O OH O H CH₃ C H3 CH₃ O

Picture 4.1.1 Chitin structure showing the hydroxyl and acetamido groups on the glucose unit.

However, Chitin in biomass is closely associated with proteins, lipids, minerals that have to be removed quantitatively in order to obtain a high quality sorbent. The major inorganic components are magnesium and carbonates (Percot et al. 2002).

4.2 The structure and chemical properties of Chitosan

Chitosan is a linear polymer, chemically described as a poly(N-glucosamine) with hydroxyl and amine groups present at the 2,3- and 5- position in the glucose unit respectively. The novel adsorption capacity of Chitosan can be attributed to its functional groups. The hydroxyl groups increases the hydrophilicity of the polymer, which enables diffusion into to the polymer network allows adsorption from aquatic solutions. The hydroxyl- and amino groups also have a high reactivity and can react with solutes in a number of different ways (Crini 2005). Since the amino group is perhaps the most important when it comes to adsorption capacity, the degree of deacetylation is an important parameter to asses the quality of the Chitosan.

NH2 O O O O H OH CH₃ C H₃

(20)

In spite of these favourable properties some problems can occur in chemical process applications. For example Chitosan is soluble in acidic media and, therefore, Can not be used as an insoluble sorbent under such conditions (Varma et al. 2004). The stability, acid and alkali resistance of Chitosan may be enhanced by crosslinking reactions although this procedure leads to a loss in adsorption capacity. Another problem occurs when Chitosan in the form of flakes or powder is used in adsorption columns. Due to the low porosity around 0.85 of pure Chitosan this procedure causes a significant pressure drop when used in a sorption column. To avoid this problem Chitosan beds have been developed (Crini 2005).

4.3 Biosorption mechanisms on Chitosan

The high Nitrogen content of Chitosan makes up a large number of active sites that are subject to different chemical interactions in water solutions. The amine groups of the Chitosan polymer are weak bases that will deprotonate water according to equation 6.1.1.

− +  ←

 →



−

+



+

−

NH

H

O

Chitosan

NH

OH

Chitosan

₂ ₂ ₃ (Eq. 4.3.1.)

According to Elson et al. the pKa for Chitosan is 6.3 Thus when Chitosan is slurried in water it will slightly increase pH of the solution (Elson et al. 1980). The direct consequence of this acid-base reaction is that the adsorption on Chitosan will be dependant on pH. The amine sites in their deprotonated form may bind metals through chelation mechanisms (Crini 2005). In its protonated form, on the other hand, the Chitosan possesses electrostatic properties. Thus, it is possible to adsorb metals through anion exchange mechanisms according to equation 5.3.2. − − +  ← − − +

₊

_

_{ →}

_

₋

₊

−

NH

X

Y

Chitosan

NH

Y

X

Chitosan

₃ ₃ (Eq. 4.3.2)

5 Materials and methods

5.1 Material

The Chitosan used in the adsorption experiments was obtained at laboratory scale from shrimp shells provided by CAMANICA S.A. Deproteinization of the shrimp shells was made using a 0.5% NaOH solution under boiling temperature for 30 minutes with constant agitation. The liquid was thereafter separated by filtration and shrimp shells were transferred to a new beaker. The residual shrimp shells were boiled for 10 minutes in a 3% NaOH solution. This procedure was repeated for 3 times. The shells were thereafter agitated with a NaClO solution for 30 minutes in order to remove all pigments.

The separated shells were then demineralised with a 1.25M HCl in room temperature for 30 minutes. The solid phase was separated and dried at 60°C for at least 24 hours. The product was thereafter weighed to determine the chitin content of the shrimp shells.

(21)

The Chitin was deacetylated using a 50% NaOH solution. The reaction rate was increased using a boiling water bath for 30 minutes. The Chitosan product was obtained by floatation in the beaker. The Chitosan was washed in distilled water until pH reached 7. The product was thereafter weighed and the Chitosan yield was determined.

100 )

(

)

(

Pr

%

=

×

g

Chitin

g

oduct

Chitosan

(Eq 5.1.2)

The dried Chitosan was grinded and passed through a size excluding mesh to separate particles with a diameter above 0.05 mm.

The Chitosan quality was analysed using IR spectrometry according to the following procedure: 0,1g of Chitosan was grinded, mixed with 0.2 g KBr and left to dry for 3 hours at 100°C. The sample was pressed into a pastille and an IR spectra was produced using a Magna 550 IR spectrometer. The deacetylation percentage was calculated as the relationship between the characteristic Chitin and Chitosan peaks.

5.2 Experimental procedure of arsenic adsorption

A certified arsenic standard solution was used to prepare solutions with initial concentrations raging between 5- 500 mg/l. To achieve the desired pH for the sorption experiments, micro volumes of nitric acid and Sodium hydroxide solutions were added to the solutions. Measurements of pH were performed with a model 410A ORION pH meter. Adsorption experiments where carried out by putting 0.5 g of Chitosan (dry sorbent mass) in contact of 20 ml of arsenic solution. The samples were collected and separated after 1 hour of agitation at 200 rpm. The final concentration of Arsenic in the solution was determined by Atomic Absorption spectrometry. The Arsenic content, q (mg As/g Chitosan) was determined by a mass balance between the solid and the liquid phase. Kinetic experiments were performed as described above while collecting and separating samples over time.

5.3 Experimental design

The experiments to determine the isotherms were designed to study the influence of solution pH on the sorption capacity of Chitosan. All experiments were carried out in duplicate accepting a 5 % deviation between the replicates for calculations of a mean value.

5.4 Desorption experiments

(22)

5.5 Arsenic analysis and equipment

(23)

6 Results and discussion

6.1 Chitosan yield and quality

Chitin was extracted from the dried shrimp shells with an average yield of 25%. The Chitosan was obtained with a yield of 40% after deacetylation and washing of the Chitin. The actual washing is thought to represent a significant loss of polymer, which could easily be diminished with a more sophisticated washing technique and a finer mesh filter.

In accordance with the method used by Moore and Domzy, the degree of deacetylation was calculated as the fraction between the characteristic peaks of the carboxyl oxygen of the amid, and the nitrogen- hydrogen stretch of the deacetylated Chitosan (Moore and Dumzy 2000).

Figure 6.1.1. IR spectra for the produced Chitosan.

In figure 6.1.1. we ca see the IR spectra for the produced Chitosan. The peak found at 3409 cm-1is typically belongs to that of N-H stretching. The broadness of the peak may indicate tat there is both symmetric and asymmetric stretching belonging to both amine and amid groups. At 1654 cm-1 there is a sharp peak, which is typical for the carbonyl stretching of an amide group.

(24)

And the deacetylation percentage was thereafter determined as

Acetyl

N

ion

Deacetylat

=

100 −

%

−

%

Where the absorbance is equal to the minus logarithm of the transmittance on the IR spectra

T

A

=

−

log

Thus

%

03 .

59

97 .

40

100 %

%

97 ,

40

33 .

1

100

556 .

0

303 .

0 %

=

−

=

























=

−

ion

Deacetylat

acetyl

N

Compared to commercially produced Chitosan a 59% grade of deacetylation is quite a poor result. Commercially produced Chitosan normally holds a degree of deacetylation above 85% (Elson et al. 1980), which implies that the produced Chitosan in this study would have a lower adsorption capacity compared to a commercial analogue.

6.2 Adsorption kinetics

In order to establish the proper conditions for the equilibrium experiments, the study of the isotherms were preceded by a brief study of the sorption kinetics. At the point where no further adsorption occurs the Chitosan has reached equilibrium with the solute.

Adsorption kinetics 0 0,2 0,4 0,6 0,8 1 1,2 0 1000 2000 3000 4000 5000 Time (min) C (t )/ C o _{150 mg/l} 500 mg/l

Figure 6.2.1. describes the sorption ratio as a function of time, with samples taken at 30 min, 60 min, 12 h, 24h, 48 h and 72 h.

(25)

Chitosan flakes used (d<0,05mm), which reduced time required for intra particulate diffusion of Arsenic to the adsorption site.

Figure 6.3.1. also displays that the kinetics were not affected by the initial concentration, since equilibrium was reached within 60 minutes in both cases. This was confirmed by all concentrations included in the isotherms, which were established to be at equilibrium after 60 minutes.

6.3 Adsorption isotherms

Figure 6.3.1, 6.3.2 and 6.3.3 shows experimental adsorption data for different values of initial pH in the solution along with the approximated Langmuir and Freundlich models.

Isotherm pH 3

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 0 100 200 300 400 500 Ceq (mg/l) q ( m g A s /g C h it o s a n ) Experimental data Langmuir Freundlich

(26)

Isotherm pH 5 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 0 200 400 600 Ceq (mg/l) q ( m g A s /g C h it o s a n ) Experimental data Langmuir Freundlich

Figure 6.3.2. Arsenic adsorption isotherm at pH 5. Experimental data and modelled curves illustrating adsorption capacity as a function of equilibrium concentration.

Isoterm pH 7

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 0 200 400 600 Ceq (mg/l) q ( m g A s /g C h it o s a n ) Experimental data Langmuir Freundlich

Figure 6.3.3. Arsenic adsorption isotherm at pH 7. Experimental data and modelled curves illustrating adsorption capacity as a function of equilibrium concentration.

(27)

Initial

pH Langmuir model Freundlich model qm (mg/g) b (l/mg) K (l/g) n

pH 3 5,1 0,0075 0,35 0,4

pH 5 5,05 0,0065 0,26 0,44

pH 7 5,05 0,006 0,3 0,42

Table 6.3.1. Approximated Langmuir and Freundlich constants for the experimental data.

The results from these experiments concluded in table 6.4.1. does not, however, support the hypothesis that a decrease in solution pH would increase the adsorption capacity. The maximum adsorption capacities were estimated to 5.1, 5.05 and 5.05 mg As/g Chitosan respectively using the Langmuir model. There was thus no significant difference in adsorption capacity with respect to initial pH of the solution.

Isotherms

0 0,5 1 1,5 2 2,5 3 3,5 4 0 100 200 300 400 500 Ceq (mg/l) q ( m g A s /g C h it o s a n ) pH 3 pH 5 pH 7

Figure 6.3.4. Adsorption isotherms illustrating the effect of initial pH on.

The Similarity between the experiments can also be illustrated graphically, as in figure 6.4.4. The results can, however, be explained by the ability of Chitosan to act as a weak base in aquatic solution. When the final pH of the solutions was measured they were all found to be slightly above neutral pH regardless of the initial pH which means that the Chitosan was only partly protonated and cationic.

(28)

Adsorption isotherm

0 2 4 6 8 10 12 14 16 18 20 0 200 400 600 Ce q (mg/l) q ( m g A s /g C h it o s a n ) pH>pKa pH<pKa Langmuir Langmuir

Figure 6.3.5. Arsenic adsorption isotherms for controlled pH<pKa and pH>pKa respectively.

Isotherm pH<pKa

0 2 4 6 8 10 12 14 16 18 0 50 100 150 200 250 Ceq (mg/l) q ( m g A s /g C h it o s a n ) Experimental data Langmuir Freundlich

Figure 6.3.5.Arsenic adsorption isotherm for a controlled pH<pKa.

Langmuir model Freundlich model

qm (mg/g) b (l/mg) K (l/g) n pH<pKa 20,9 0,012 0,9 0,55 pH>pKa 5,1 0,0075 0,35 0,4

Table 6.3.2. Adsorption coefficients for the modelled Langmuir and Freundlich curves.

(29)

significantly increased the adsorption capacity of Chitosan. This implies that ionic exchange was quantitatively the most important adsorption mechanism.

The adsorption capacity calculated with the Langmuir model (Table 6.4.2.) also differs from those previously cited in literature. Guibal et al. estimated the maximum sorption capacity of Chitosan to be 7 mg As/g (Guibal 1995). Elson et al. calculated the column capacity of Chitosan to be 224.76 mg As/g for acidic conditions (Elson et al. 1980) and Benavente et al. calculated the column capacity to 38.7 mg As/g (Benavente et al. 2006). These big differences in sorption capacity can not be explained by different grade of deacetylation, but that some other phenomenon is present. Studies of the uranium sorption on Chitosan showed that the sorption capacity is a function of the particle size (Guibal et al. 1995). A plausible explanation could be that the adsorption capacity is dependant of the diffusion into the polymer network. Intraparticular diffusion has already been established to affect the kinetics of the adsorption may also be the case of the arsenic adsorption and could explain the discrepancies in adsorption capacities in previous studies.

One other factor that contributed to the increase in adsorption with decreasing pH was the form in which the arsenic was present. The speciation of arsenic suggests that arsenates were present as H2AsO4- and HAsO4-2 (pK2 6.8), which both could be adsorbed through ion exchange to the cationic Chitosan. As pH decreases below 3 the speciation favours the uncharged H3AsO4 specie. A preliminary study confirmed that the adsorption capacity of uncharged arsenite species was significantly lower than that of arsenates. The adsorption capacity in Figure 6.4.5 was therefore close to an optimum favoured by the protonation of the amine groups as well as arsenic speciation. This conclusion is also supported by Dambies et al. (Dambies et al. 2002).

A problem of decreasing the final pH of the Chitosan-solute mixture was that the Chitosan started to dissolve. Although the dissolution of Chitosan did not adversely affect the adsorption performance, it would be a serious disadvantage in a continuous column system. The dissolution of sorbent would not only lead to a loss of sorbent, but would also create a large pressure drop over the column by reducing the particle size.

6.4 Desorption and sorbent recycling

The desorption of Chitosan is a topic that has been sparsely examined compared to the adsorption of various metals. An initial screening for an efficient desorbing solution was therefore conducted.

(30)

NaOH desorption S/L=42

0 5 10 15 20 25 30 35 0 10 20 30 40 Eluation volume (ml) A s c o n c e n tr a ti o n ( p p m ) NaOH 1M

Figure 6.6.1. Desorption of arsenic containing Chitosan with a 1M sodium hydroxide solution.

The first desorption step resulted in a 50.2% recovery of the adsorbed arsenic and a concentration ratio of 0.82. This implies that even though the NaOH enables the desorption of arsenic, the eluated liquid held a lower concentration than the incoming feed solution and that all adsorbed arsenic was not stripped.

NaOH desorption S/L=100g/l 0 5 10 15 20 25 30 35 40 45 0 20 40 60 80 100 Eluated volume (ml) A rs e n ic c o n c e n tr a ti o n ( p p m ) NaOH 1M NaOH 0,1M

Figure 6.6.3 Desorption of arsenic containing Chitosan with NaOH solutions of 1M and 0.1M.

(31)

Studying the curvature of figure 6.6.2. and 6.6.3. one can see that the desorption was increasingly difficult for each step. Since the adsorption of arsenic was a favourable process it was expected that desorption would be unfavourable, since small amounts of arsenic were strongly adsorbed to the Chitosan. This means that the performance in terms of concentration ratio was lowered for each desorption step. The steepness of the curve, however, indicates that most arsenic is desorbed rapidly even though it is not desorbed completely.

In the search for other desorption agents, NaCl was chosen to test weather the presence of competing anion could strip the arsenic by reversing the ion- exchange equilibrium. Due to its low cost NaCl was an especially interesting desorption agent to study. Figure 6.6.3. shows the desorption curve using a 1M NaCl solution. Even though chloride has been showed to have a low affinity for binding to Chitosan (Vold et al. 2003) the concentrated NaCl solution was able to desorb the arsenic. The concentration ratio for the first desorption step is, still, only 0.38, which is less than half compared to desorption with NaOH.

NaCl desorption S/L=50g/l

0 2 4 6 8 10 12 14 16 0 50 100 150 200 250 300 Eluated volume (ml) A rs e n ic c o n c e tr a ti o n ( p p m ) NaCl 1M

Figure 6.6.3. Desorption of arsenic containing Chitosan with a 1M sodium chloride solution.

(32)

Amoniumsulphate desorption S/L=50

0 5 10 15 20 25 30 0 50 100 150 200 250 Volume of eluent (ml) A rs e n ic c o n c e n tr a ti o n ( p p m ) Amoniumsulphate 1M

Figure 6.6.4. Desorption of arsenic containing Chitosan with a 1M ammonium sulphate solution.

The poor concentration ratios attained in this study are probably due to the concentrated feed solutions that were used in the adsorption studies (up to 500ppm). The concentration ratios displayed is, therefore, an indicator of the desorption solution’s relative efficiency. When the adsorption process is used for more dilute solutions the efficiency of the desorption is expected to increase in terms of concentration ratio.

It has previously been demonstrated by Guibal et al. that Chitosan can be successfully desorbed and reused 10 times with unaffected sorption capacity for the oxyanions Molybdate and Vanadate using a 1M NaOH solution (Guibal 2004). This is a strong indication that arsenic could be desorbed with good results in a column system since the oxyanions share a similar sorption mechanism.

6.5 Process applicability

When considering the biosorption process with Chitosan from a broader perspective it is important to identify the limiting parameters of its applicability.

Being a biopolymer, Chitosan is rather easily degraded. The poor stability of the Chitosan polymer is a property that has serious drawbacks for its use in adsorption columns. As stated above it was noticed that the Chitosan started to dissolve during pH controlled experiments when acid was added. This was also observed during desorption experiments using strong alkali solutions. The degradation of the polymer seems to be limiting for the reuse and, therefore, the overall the cost-effectiveness of the treatment method.

(33)

(34)

7 Conclusions

The degree of deacetylation in the produced Chitosan was significantly lower than that of commercial equivalents and contained impurities. The fact that adsorption capacity should be proportional to the grade of deacetylation was, however, not reflected in the calculated adsorption capacity.

The maximum theoretical capacity of Chitosan to adsorb arsenic was calculated to 21 mg As/ g Chitosan at final a pH of 5.5. This is comparable to the adsorption capacity of arsenic on activated carbon from coconut shells (Lorenzen et al. 1995) and about 10 times smaller than that of commercial ion-exchangers (Elson et al. 19809. Compared to other adsorption studies carried out on Chitosan, Arsenic is one of the metals that display the lowest adsorption capacity.

The capacity of Chitosan to adsorb Arsenic was strongly pH dependant with a fourfold increase in adsorption capacity when the pH was well under the pKa of Chitosan. The pH dependence supports the hypothesis that ion-exchange was the most important mechanism. Though, it can not be excluded that surface sorption was present. Desorption studies also suggests that some arsenic was bounded covalently to the Chitosan.

The pH of the initial solution had no significant effect on the adsorption capacity in the pH range 3-7. When optimizing the adsorption process with respect to pH it was important measure the final pH of the solution, since the Chitosan increased the pH when it was slurried in water.

Arsenic desorption was achieved using NaOH, NaCl or (NH4)2SO4, with NaOH of 1M being the most effective. The desorption had an unfavourable curvature that made it difficult to desorb the Chitosan, while maintaining a high concentration ratio.

The possibility to regenerate the sorbent and the fact that the sorption proceeds more effectively under acidic conditions direct the biosorption process to treatment of industrial raw water and mining effluents.

(35)

8 Acknowledgments

This study would not have been realised without the support and assistance of a number of persons to whom I wish to express my gratitude.

First and foremost, I would like to thank Martha Benavente for leading the experimental part of this work and for taking such good care of me and Anna in Nicaragua.

Thank you, Olle Wahlberg, for being a mentor in the field of environmental chemistry as well as being the principal supervisor and examiner of this thesis.

I also want to thank Joaquín Martínez for invaluable support, for putting this project together and for soccer games in Managua.

(36)

9 References

Álvarez E., Benavente M., 2005 Determinación de isotermas de Adsorción de Fe, Cu y Zn en Quitosana. Tesis para optar al Título de Ingeniero Quimica. Universidad Nacional de Ingeniería, Nicaragua.

Artola A., Balanguer M. D., Rigola M., 1999, Competitive biosorption of copper, cadmium, Nickel and Zink from metal ion mixtures using anaerobically activated sludge, In proceeding of International biohydrometallurgy Symposium (IBS’99), San Lorenzo de El Escorial, Madrid, Spain, June 175-185, Ed Amils, R.,Ballester., Elsevier

Benavente M. 2005, Washing and leaching of gold ore tailings containing cyanide compounds and removal of heavy metals from mining effluents. Unpublished specification of doctoral thesis.

CAPRE Guidelines, 2000, Regional committee for drinking water institution and sanitation for Central America, Panama and Dominican Republic.

Cussler E.L., 1997, Diffusion, mass transfer in fluid systems 2nd_{edition. Cambridge University press.}

Dambies L., Roze A., Roussy J., Guibal E., As(V) removal from dilute solutions using Molybdate impregnated chitosan beds, In proceeding of International biohydrometallurgy Symposium (IBS’99), San Lorenzo de El Escorial, Madrid, Spain, June 277-288, Ed Amils, R.,Ballester., Elsevier

Dambies L., Vincent T., Guibal E., 2002, Treatment of arsenic-containing solutions using chitosan derivatives: uptake mechanism and sorption performances, Water Research Vol. 36, pp. 3699-3710.

Espinoza M. A., 2005, Distribution of arsenic pollution in the groundwater of the South West basin of Sebaco Valley - Matagalpa, Nicaragua Centro para la Investigación en Recursos Acuáticos de Nicaragua - CIRA/UNAN

Elson C. M., Davies D. H., Hayes E. R., 1980, Removal of arsenic from contaminated drinking water by a chitosan/chitin mixture, Water Research Vol. 14, pp. 1307-1311.

(37)

Guibal E., Interactions of metals with Chitosan-based sorbents: A review, 2004, Separation and purification technology Vol 38, pp. 42-74.

Jain C. K. and Ali I., 2000, Arsenic occurrence, toxicity and speciation techniques, Water Research Vol. 34, pp. 4304-4312.

Lorenzen L., van Deventer S. J., Landi W. M. 1995. Factors affecting the mechanism of adsorption of arsesnic species on activated carbon. Miner Eng. 1995 Vol. 8, pp557-569.

Percot A., Viton C., Domard A., 2003, Optimization of Chitin from shrimp shells, Biomacromolecules 4, 12-18. Persson, P.O. 2005. Miljöskyddsteknik – strategier och teknik för en hållbar utveckling. Industriell ekologi, KTH

Sjörén A., 2006, Removal of Mercury using biosorption with Chitosan. Diploma work. At the department of inorganic chemistry, Royal Institue of Technology (KTH), Stockholm, Sweden.

Smedley P.L. and Kinniburgh D. G., 2002, A review of the source, behaviour and distribution of arsenic in natural waters, Applied geochemistry Vol. 17, pp. 517-568.

Thornton I., 1995, Sources and pathways of arsenic in south-west England: In Arsenic Exposure and Health, Science and Technology Letters, Northwood, England.

Tsezos M., 1999 Biosorption of metals. The experience and outlook for technology development, In proceeding of International biohydrometallurgy Symposium (IBS’99), San Lorenzo de El Escorial, Madrid, Spain, June 171-174, Ed Amils, R.,Ballester., Elsevier

Vahter M. and Marafante E., 1988, In vivo methylation and detoxification processes of arsenic, In The Biological Alkylation of Heavy Elements, pp. 105-109. Royal Society of Chemistry, London.

van Loon and Duffy, 2005, Environmental chemistry- A global perspective,

Vold I.M.N., Vårum K.M., Guibal E, Smidsrod O, 2003, Binding of ions to Chitosan- selectivity studies. Carbohydrate polymers Vol. 54, pp. 471-477.

(38)

Arsenic removal using biosorption with Chitosan