Top Banner
My Profile Log HOME ABOUT US CONTACT US HELP Home / Chemistry / Industrial Chemistry Ullmann's Encyclopedia of Industrial Chemistry Water Standard Article Hermann Weingärtner 1 , Ernst Ulrich Franck 2 , Gabriele Wiegand 3 , Nicolaus Dahmen 4 , Georg Schwedt 5 , Fritz H. Frimmel 6 , Birgit C. Gordalla 7 , Klaus Johannsen 8 , R. Scott Summers 9 , Wolfgang Höll 10 , Martin Jekel 11 , Rolf Gimbel 12 , Robert Rautenbach 13 , William H. Glaze 14 1 Ruhr-Universität Bochum, Physikalische Chemie II, Bochum, Federal Republic of Germany 2 Institut für Heiße Chemie, Forschungszentrum Karlsruhe, Karlsruhe, Federal Republic of Germany 3 Institut für Heiße Chemie, Forschungszentrum Karlsruhe, Karlsruhe, Federal Republic of Germany 4 Institut für Heiße Chemie, Forschungszentrum Karlsruhe, Karlsruhe, Federal Republic of Germany 5 Institut für Anorganische und Analytische Chemie, Technische Universität Clausthal, Clausthal-Zellerfeld, Federal Republic of Germany 6 Engler-Bunte-Institut der Universität Karlsruhe, Bereich Wasserchemie, Karlsruhe, Federal Republic of Germany 7 Engler-Bunte-Institut der Universität Karlsruhe, Bereich Wasserchemie, Karlsruhe, Federal Republic of Germany 8 Technische Universität Hamburg-Harburg, Arbeitsbereich Wasserwirtschaft und Wasserversorgung, Hamburg, Federal Republic of Germany 9 University of Cincinnati, Department of Civil Engineering, Cincinnati, Ohio, United States 10 Forschungszentrum Karlsruhe, Institut für Technische Chemie, Bereich Wasser- und Gastechnologie, Karlsruhe, Federal Republic of Germany 11 Technische Universität Berlin, Fachgebiet Wasserreinhaltung, Berlin, Federal Republic of Germany 12 Gerhard-Mercator-Universität Gesamthochschule Duisburg und Rheinisch-Westfälisches Institut für Wasserchemie und Wassertechnologie, Mülheim, Federal Republic of Germany 13 Rheinisch-Westfälische Technische Hochschule, Institut für Verfahrenstechnik, Aachen, Federal Republic of Germany 14 Department of Environmental Sciences and Engineering, School of Public Health, University of North Carolina, Chapel Hill, NC, United States Copyright © 2002 by Wiley-VCH Verlag GmbH & Co. KGaA. All rights reserved. DOI : 10.1002/14356007.a28_001 Article Online Posting Date: June 15, 2000 Recommend to Your Librarian Save title to My Profile Email this page Print this page Abstract | Full Text: HTML BROWSE THIS TITLE Article Titles A Z Topics SEARCH THIS TITLE Advanced Product Search Search All Content Acronym Finder Abstract The article contains sections titled: 1. Water as a Solvent 1.1. Properties of Pure Water 1.1.1. Molecular Properties 1.1.2. Hydrogen Bonding and Water Structure 1.1.3. Bulk Properties of Liquid Water 1.2. Aqueous Solutions 1.2.1. General Solvent Properties 1.2.2. Solutions of Simple Nonpolar Gases 1.2.3. Solutes with Hydrophilic Groups 1.2.4. Electrolyte Solutions 2. Water at High Pressure and Temperature 2.1. Properties of Ice 2.1.1. Thermodynamic Properties 2.1.2. Transport Properties Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience page 1 of 82
82
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Water Ullmans

My Profile Log

HOME ABOUT US CONTACT US HELP

Home / Chemistry / Industrial Chemistry

Ullmann's Encyclopedia of Industrial Chemistry WaterStandard Article

Hermann Weingärtner1, Ernst Ulrich Franck2, Gabriele Wiegand3, Nicolaus Dahmen4, Georg Schwedt5, Fritz H. Frimmel6, Birgit C. Gordalla7, Klaus Johannsen8, R. Scott Summers9, Wolfgang Höll10, Martin Jekel11, Rolf Gimbel12, Robert Rautenbach13, William H. Glaze14

1Ruhr-Universität Bochum, Physikalische Chemie II, Bochum, Federal Republic of Germany

2Institut für Heiße Chemie, Forschungszentrum Karlsruhe, Karlsruhe, Federal Republic of Germany

3Institut für Heiße Chemie, Forschungszentrum Karlsruhe, Karlsruhe, Federal Republic of Germany

4Institut für Heiße Chemie, Forschungszentrum Karlsruhe, Karlsruhe, Federal Republic of Germany

5Institut für Anorganische und Analytische Chemie, Technische Universität Clausthal, Clausthal-Zellerfeld, Federal Republic of Germany

6Engler-Bunte-Institut der Universität Karlsruhe, Bereich Wasserchemie, Karlsruhe, Federal Republic of Germany

7Engler-Bunte-Institut der Universität Karlsruhe, Bereich Wasserchemie, Karlsruhe, Federal Republic of Germany

8Technische Universität Hamburg-Harburg, Arbeitsbereich Wasserwirtschaft und Wasserversorgung, Hamburg, Federal Republic of Germany

9University of Cincinnati, Department of Civil Engineering, Cincinnati, Ohio, United States

10Forschungszentrum Karlsruhe, Institut für Technische Chemie, Bereich Wasser- und Gastechnologie, Karlsruhe, Federal Republic of Germany

11Technische Universität Berlin, Fachgebiet Wasserreinhaltung, Berlin, Federal Republic of Germany

12Gerhard-Mercator-Universität Gesamthochschule Duisburg und Rheinisch-Westfälisches Institut für Wasserchemie und Wassertechnologie, Mülheim, Federal Republic of Germany

13Rheinisch-Westfälische Technische Hochschule, Institut für Verfahrenstechnik, Aachen, Federal Republic of Germany

14Department of Environmental Sciences and Engineering, School of Public Health, University of North Carolina, Chapel Hill, NC, United States

Copyright © 2002 by Wiley-VCH Verlag GmbH & Co. KGaA. All rights reserved. DOI: 10.1002/14356007.a28_001 Article Online Posting Date: June 15, 2000

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Abstract | Full Text: HTML

BROWSE THIS TITLE Article Titles A–Z Topics

SEARCH THIS TITLE

Advanced Product Search Search All Content Acronym Finder

Abstract The article contains sections titled:

1. Water as a Solvent1.1. Properties of Pure Water1.1.1. Molecular Properties1.1.2. Hydrogen Bonding and Water Structure1.1.3. Bulk Properties of Liquid Water1.2. Aqueous Solutions1.2.1. General Solvent Properties1.2.2. Solutions of Simple Nonpolar Gases1.2.3. Solutes with Hydrophilic Groups1.2.4. Electrolyte Solutions2. Water at High Pressure and Temperature2.1. Properties of Ice2.1.1. Thermodynamic Properties2.1.2. Transport Properties

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 1 of 82

Page 2: Water Ullmans

2.1.3. Electrolytic Properties2.1.4. Mechanical Properties2.2. Properties of Fluid Water2.2.1. Thermodynamic Properties2.2.2. Transport Properties2.2.3. Electrolytic Properties2.2.4. Other Physical Properties2.3. Properties of Water Vapor2.4. Water in the Supercritical State3. Water Analysis3.1. Sampling and Sample Preservation3.2. Physicochemical and Sum Parameters3.3. Inorganic Analysis3.3.1. Determination of Cations3.3.2. Anion Analysis3.3.3. Determination of Dissolved Gases3.3.4. Quick Test Processes3.4. Organic Analysis3.4.1. Spectrometric Methods3.4.2. Gas and Liquid Chromatography3.5. Biochemical Methods4. Hydrological Cycle and Water Use4.1. World Water Balance4.2. Hydrological Cycle4.3. Demand for Water4.4. Source of Water Used4.5. Water Treatment5. Adsorption Processes in Water Treatment5.1. Introduction5.2. Properties of Activated Carbon5.3. Adsorption Theory5.4. Adsorption Equilibrium5.5. Adsorption Kinetics5.6. Biological Processes5.7. Design of GAC Systems5.8. Design of PAC Systems5.9. Reactivation of GAC6. Ion Exchange6.1. Fundamentals6.1.1. Characterization of the Chemical State of Water6.1.2. Representation of the Chemical State of Water6.1.3. Ion Exchangers6.2. Cation-Exchange Processes6.2.1. Strongly Acidic Resins in the Na+ Form6.2.2. Strongly Acidic Resins in the H+ Form6.2.3. Weakly Acidic Resins in the Hydrogen Form6.3. Anion-Exchange Processes6.3.1. Strongly Basic Resins in the Cl– Form6.3.2. Strongly Basic Resins in the OH– Form6.3.3. Strongly Basic Resins in the Form6.3.4. Weakly Basic Resins6.4. Combined Processes6.4.1. Demineralization6.4.2. Partial Demineralization7. Flocculation7.1. Introduction7.2. Definitions7.3. Chemicals Used in Coagulation and Flocculation

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 2 of 82

Page 3: Water Ullmans

[Top of Page]

1. Water as a Solvent Hermann Weingärtner

1.1. Properties of Pure Water

7.3.1. Inorganic Chemicals7.3.2. Organic Chemicals7.4. Kinetics of Aggregation7.5. Process Technology for Coagulation and Flocculation7.5.1. Dosing7.5.2. Rapid Mixing7.5.3. Reactors for Floc Formation7.5.4. Integrated Flocculation Systems7.5.5. Operational Aspects8. Filtration8.1. Introduction8.2. Slow Sand Filters8.3. Polishing Filters8.4. Rapid Filters8.4.1. Design8.4.2. Areas of Application and Modes of Operation8.4.3. Filter Performance9. Membrane Separation Processes in Water Treatment9.1. Principles9.1.1. Flow, Selectivity, Driving Forces9.1.2. Mass Transport Resistance in Front of the Membrane9.2. Membranes9.2.1. Organic Membranes9.2.2. Inorganic Membranes9.3. Modules9.3.1. Modules with Tubular Membranes9.3.2. Modules with a Flat-Sheet Membrane9.4. Fouling and Scaling9.4.1. Membrane Blockage due to Crystallization (Scaling)9.4.2. Membrane Blockage due to Contaminants (Fouling)9.5. Module Arrangements (Plant Design)9.6. Use of Membrane Processes in Water and Wastewater Technology9.6.1. Desalination of Seawater and Brackish Water9.6.2. Recovery of -Caprolactam9.6.3. Leachates from Dump Sites9.6.4. Wastewater from the Dye Industry10. Oxidation Processes in Water Treatment10.1. Physical and Chemical Properties of Chemical Oxidants10.2. Production of Chemical Oxidants10.3. Uses of Chemical Oxidants10.3.1. Drinking Water Treatment10.3.1.1. Disinfection10.3.1.2. Chemical Oxidation10.3.2. Wastewater Treatment10.3.3. Oxidation in Pulp and Paper Treatment10.3.4. Advanced Oxidation Processes10.4. Toxicology and Environmental Health11. Acknowledgement

1.1.1. Molecular Properties In comparison to other solvents, water and aqueous solutions in many respects show unique properties that are highly sensitive to temperature and pressure [1-3]. These properties result from the structure and the charge distribution of the water molecule. In a simple four-point-charge model, this charge distribution is depicted as a quadrupole with two protons and two lone electron pairs at the corners of a tetrahedron with oxygen in its center (Fig. 1). In the isolated molecule the O–H

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 3 of 82

Page 4: Water Ullmans

internuclear distance is 0.096 nm, with an H–O–H bond angle of 104.3°, which is close to the tetrahedral angle [2], [3]. This nonspherical charge distribution gives rise to electric moments. The dipole moment of the isolated water molecule is 1.85 D (6.14×10–30 C · m), the dipole vector acting along the bisector of the H–O–H angle with the negative end directed toward the oxygen. Thus, water is a highly polar solvent.

When considering intermolecular interactions, the electrostatic forces have to be supplemented by terms accounting for the core repulsion of the molecules and the attractive dispersion (van der Waals) forces. For water, these terms are often approximated by the corresponding figures for the isoelectronic neon, which has a core radius of 0.14 nm. In computer simulations, models of this type are able to reproduce the major features of the properties of liquid water [4].

Figure 1. Four-point-charge model of water after [4]. The partial charges + refer to the protons, the charges – to the lone electron pairs. The experimental dipole moment is reproduced with + = – = 0.236 e, where e is the elementary charge.

1.1.2. Hydrogen Bonding and Water Structure The molecular geometry and charge distribution of water favor the formation of a hydrogen-bonded network with four O–H–O bonds per molecule. Hydrogen bonding can be proved by infrared and Raman spectroscopy. Three normal vibrations of the isolated H2O molecule occur in the vapor phase: the symmetric stretching mode at a wave number = 3675 cm–1, the

asymmetric stretching mode at 3756 cm–1, and the bending mode at 1595 cm–1 (Fig. 2). Hydrogen bonding leads to shifts in the wave numbers of these modes, a change in their band contours, and the appearance of new bands associated with intermolecular vibrations [3]. These spectral features are often used to monitor changes in hydrogen bonding as a function of temperature and pressure or induced by added solutes. In addition, the chemical shift in the 1H NMR spectrum of water provides a sensitive tool for measurement of hydrogen bonding.

The structure of hexagonal ice Ih (Fig. 3), which is the stable polymorph of ice at atmospheric pressure, is an ideal representation of this hydrogen-bonded network [3]: Each oxygen is surrounded tetrahedrally by four others at a distance of 0.276 nm, the structure being isomorphous with the tridymite structure of silica. The protons lie approximately along the O–O axes. Because of hydrogen bonding, the intramolecular OH distance (0.101 nm) is greater than in the isolated molecule. If compared with closely packed structures with up to 12 nearest neighbors, the ice Ih structure is very open. The large molar volume of ice Ih is, for example, responsible for the volume contraction of ice upon melting.

In the liquid state the long-range order is lost, but short-range positional and orientational correlations are retained, which can be studied by X-ray and neutron scattering [3]. At room temperature a structure is perceptible in the local environment, which extends to about 0.8 nm. On average, one molecule has 4.4 nearest neighbors at a distance of 0.284 nm. These figures and the location of second-nearest neighbors correspond closely to the requirements imposed by the ice Ih structure. However, intermolecular O–O and O–H distances and H–O–H angles show a broad distribution around the average value, and due to thermal reorientation, the lifetime of a given configuration is only of the order of some picoseconds. With increasing temperature, the three-dimensional network gradually breaks down. Thus, with rising temperature, water loses its unique properties, although Raman spectroscopy shows that even at the critical point at 374 °C, some degree of hydrogen bonding is still retained [5].

Any model of liquid water must accommodate these features, but none of the existing models seems to be immune to criticism. For a long time, so-called mixture models have been used to explain the properties of water and aqueous solutions. These models assume the existence of two (or more) physically distinguishable water species, corresponding to molecules with “broken” and “unbroken” hydrogen bonds [2], [3], according to equilibria of the general form H2O (icelike) H2O (gaslike), which depend on temperature and pressure and are changed by solutes. From the present point of view, the need for such models is overcome largely by the possibility of performing more rigorous model-free analyses based on statistical mechanical methods and computer simulations, so that many of these models are only of historical interest.

Figure 2. Normal vibrations of the water molecule: 1 is the symmetric stretching mode, 3 the antisymmetric stretching mode, and 2 the bending mode

Figure 3. Structure of hexagonal ice Ih. The spheres give the positions of the oxygen atoms.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 4 of 82

Page 5: Water Ullmans

1.1.3. Bulk Properties of Liquid Water A knowledge of the bulk properties of pure water and their anomalous behavior is a prerequisite for understanding its solvent properties, because many methods and rules for evaluating and predicting data in nonaqueous systems are of little value for aqueous solutions. Most of these anomalies occur near room temperature or up to 100 °C. The significance of these anomalies for processes in nature and in sustaining life has been outlined [3]. Doubtless, these anomalies result from the strong hydrogen bonding that occurs, however, in many other substances as well. Hence, the existence of strong hydrogen bonds cannot fully explain the uniqueness of some properties of water. Rather, it is the existence of the three-dimensional network that distinguishes water from other hydrogen-bonded liquids. Despite great progress in liquid-state theory, the treatment of such hydrogen-bonded systems remains difficult.

Isotopic Composition. Natural (“ordinary”) water is a mixture of isotopic species made up by the hydrogen isotopes and (deuterium) and the oxygen isotopes , , and . Their masses and natural abundances are compiled in Table 1.

Proton exchange equilibria of the form yield further mixed species. The and are usually assumed to be distributed randomly on the oxygens. Then, the four major species of ordinary water occur in the proportions [6]

Table 1. Isotopic masses and natural abundances of hydrogen and oxygen isotopes [11]

The molar mass of ordinary water is 18.012 g/mol. The variation of these proportions in water from different natural sources is usually smaller than variations as a result of water purification, and except in work of the highest accuracy, the properties of ordinary water can be set equal to the properties of . Practically pure (heavy water) and highly enriched - and -labeled water are available commercially, so that some data also exist for water of other isotopic compositions [7]. In the following, the term “water” is used for ordinary water with natural isotopic composition.

p – V – T Data and Invariant Points. Water is the only inorganic substance that, under ambient conditions found in nature, occurs in the solid, liquid, and gaseous states. Its state of aggregation is characterized by the pressure p, the temperature T, and volume V (or density ). Only the p – V – T properties of the liquid phase are discussed here. In most applications, the properties of the liquid along the saturation curve can be set equal to the properties at atmospheric pressure.

The liquid range is limited by the triple point at which solid ice Ih, liquid water, and steam are in equilibrium and by the critical point. The triple point of ordinary water is bound by the definition of the temperature scale to 273.16 K = 0.01 °C. The normal melting and boiling points at 1.01325 bar are 0 and 100 °C, respectively. The triple-point pressure is pt = 6.113×102 Pa. Recommended values for the critical temperature, critical pressure, and critical density are [8] tc = 373.98 °C, pc =

22.05 MPa, and c = 322 kg/m3. Table 2 lists volumetric properties of the liquid from 0 to 100 °C [8], [9]. Here, is the mass density; p = (1/V )(∂V/∂T )p, the thermal expansion coefficient; and T = (1/V )(∂V/∂p)T, the isothermal compressibility. The thermal pressure coefficient v = (∂p/∂T )V can be calculated by v = p/ T.

Table 2. Density , thermal expansion coefficient p, and isothermal compressibility T of ordinary water at atmospheric pressure [11]

Atomic mass Natural abundance

Mol % wt %

H (natural) 1.0080 1.00783 99.985 99.970 2.01410 0.015 0.030

O (natural) 15.9994 15.9995 99.759 99.73216.9991 0.037 0.03917.9992 0.204 0.229

t, °C , kg/m3p×106, K–1

T, bar–1

0 999.84 –68.14 50.88

10 999.70 87.90 47.82

20 998.20 206.6 45.90

25 997.05 257.1 45.25

30 995.65 303.1 44.77

992.23 385.4 44.24

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 5 of 82

Page 6: Water Ullmans

At atmospheric pressure, ice Ih contracts by ca. 8.2 % on melting — from a density of 916.8 kg/m3 to 999.87 kg/m3 in the liquid. The slope of the melting pressure curve depends on the volume change on melting. The anomalous volume contraction of water leads to a depression of its melting point with rising pressure. This depression continues up to 207 MPa and –22 °C, where the ice polymorph changes to ice II. Close to 0 °C, a pressure of 133 bar is required to lower the melting point by 1 °C.

The volume contraction on melting continues in the liquid phase up to 3.98 °C, where the density reaches a maximum value and the thermal expansion coefficient changes sign from negative to positive. Over the entire temperature range considered, the thermal expansion coefficient of water is one order of magnitude smaller than that of other liquids. The isothermal compressibility is unusually large and possesses a minimum near 46.5 °C. Both extremes represent anomalous properties of pure water that tend to disappear at high pressure.

Actually, liquid water may be supercooled below the equilibrium melting point [10]. Experiments can be performed fairly easily down to –20 °C, which is roughly the limit of heterogeneous nucleation. With special techniques, temperatures as low as –40 °C have been achieved with liquid water. At 1 bar, homogeneous nucleation sets a limit at –43 °C. Many of the anomalous properties of liquid water become even more unusual in the supercooled, metastable regime [10].

Specific Heat. At the normal melting point, the isobaric specific heat cp increases from 2.072 J K–1 g–1 in the solid to

4.228 J K–1 g–1 in the liquid state, where it remains almost unchanged up to 100 °C [8], [9] (a very shallow minimum is found near 35 °C) (Table 3). The specific heat is particularly amenable to molecular interpretation and has often been used to estimate the extent of hydrogen bonding in water [2], [3].

Table 3. Isobaric specific heat cp, vapor pressure p, latent heat of vaporization h, and surface tension of water at saturation pressure

Liquid – Vapor Equilibria. The vapor pressure at saturation is given in Table 3 [9]. The slope of the vapor pressure curve yields the latent heat of vaporization h [8], [9], which is ca. 2500 kJ/kg at the normal melting point. Up to 100 °C, a slight decrease in h is observed (Table 3). Compared with other liquids, the latent heat of vaporization is anomalously high, and its values have often been used to estimate the energy of hydrogen bonds in water. Based on comparison with data for similar molecules, the heat of vaporization of non-hydrogen-bonded H2O should be ca. 550 kJ/kg, about one-fourth of the actual values between 0 and 100 °C. The difference is ascribed to the energy of the hydrogen bonds.

The vapor pressure is well investigated for other isotopic compositions — above all heavy water. Vapor pressure data for other isotopic forms usually are represented by the partition coefficient , where H2O represents light water and x the heavier isotopic form. Figure 4 shows the partition coefficient of heavy water as a function of temperature [7]. Over a

40

50 988.03 457.8 44.18

60 983.19 523.4 44.51

70 977.76 584.0 45.17

80 971.79 641.3 46.15

90 965.31 696.3 47.43

100 958.36 750.0 49.02

t, °C cp, J g–1 K–1 p, bar h, kJ/kg ×103, N/m

0 4.217 0.00611 2500.5 75.64

10 4.191 0.01228 2467.9 74.23

20 4.182 0.02338 2453.4 72.75

30 4.178 0.04245 2429.6 71.20

40 4.178 0.07382 2405.9 69.60

50 4.180 0.12346 2309.9 67.94

60 4.184 0.19936 2357.7 66.24

70 4.189 0.31181 2333.1 64.47

80 4.196 0.47379 2308.2 62.67

90 4.205 0.70123 2282.7 60.82

100 4.216 1.01325 2256.6 58.91

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 6 of 82

Page 7: Water Ullmans

large range of temperature, heavy water is less volatile than light water, so its melting and boiling points (3.82 and 101.42 °C) are higher than those of . The inverse is true above 200 °C, and the critical point of is about 3 °C lower.

Surface Tension. Below 100 °C, the surface tension [8], [9] (Table 3) refers to the interface between water saturated with air and air saturated with water. It is distinctly higher than that of other liquids, including most hydrogen-bonded ones. On the basis of simple theories, the quantity / 2/3 should be a function of the temperature distance T – Tc from the critical point.

This is observed approximately, but for water the temperature function of / 2/3 does not exhibit the predicted universal slope found with most other liquids.

Transport Properties. The dynamic viscosity of liquid water (Table 4) is known with great accuracy [6], [8], [9], and the data are often used to calibrate viscometers. The dynamic viscosity is higher than that of comparable substances without hydrogen bonds. The decrease of dynamic viscosity with increasing temperature cannot be described with sufficient accuracy by the conventional Arrhenius-type expression with two adjustable constants 0 and E

unless the energy of the viscous flow E is assumed to be temperature dependent. In other words, the plot of the logarithm of viscosity versus the inverse temperature does not yield a straight line, as found for many other liquids. More complex expressions are therefore in use [10]. The temperature dependence of viscosity is of some interest in interpreting processes in supercooled water [10]. The pressure dependence of the dynamic viscosity of water is anomalous as well: Whereas the viscosities of liquids usually increase with increasing pressure, the viscosity of water initially decreases up to a minimum at ca. 60 MPa. This decrease has clear structural implications: the hydrogen-bonded network becomes deformed with rising pressure, and this tends to facilitate molecular motion. An unusually large increase of viscosity (ca. 23 % at 25 °C) is found when ordinary water is replaced by heavy water. At higher temperature, this isotope effect diminishes. This unusually large effect is attributed to a higher structuredness of heavy water.

Table 4. Viscosity and thermal conductivity of liquid water at atmospheric pressure

In gases, the thermal conductivity is proportional to viscosity, but this correlation is lost in the liquid phase, where for most liquids a linear decrease is found with increasing temperature. In contrast, water shows an anomalous increase in up to ca. 130 °C [6], [8], [9] (Table 4). Despite its considerable technical importance, understanding of the thermal conductance of liquids on a molecular basis is poorly developed.

The self-diffusion coefficient D characterizes the mean-square displacement of water molecules due to thermal agitation. In ordinary water at 25 °C, D = 2.30×109 m2/s [11], corresponding to a displacement of the order of 10–4 m during 1 s. The temperature and pressure dependence of D and the isotope effects are correlated strongly with those of the reciprocal viscosity.

Autoprotolysis. Self-dissociation of water is characterized by the ion product or its negative decadic

Figure 4. Partition coefficients of light (a) and heavy water (b) as a function of temperature

t, °C ×106, Pa · s , mW K–1 m–1

0 1792 565

10 1307 584

20 1002 602

25 890.3

30 797.5 617

40 653.2 631

50 547.1 642

60 466.6 652

70 403.9 660

80 353.8 669

90 312.8 675

100 278.3 679

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 7 of 82

Page 8: Water Ullmans

logarithm. Here and are the activities of the H+ and OH– ions. At 25 °C, Kw = 1.008×10–14. Accordingly, the pH value of pure water is . The ion product increases with increasing temperature and increasing pressure [12] (seeSection Electrolytic Properties). Autoprotolysis is of paramount importance for the Brønsted acid – base concept. According to this scheme, water is classified as an amphiprotic solvent that, depending on the solute, exhibits acidic or basic properties.

Electric Properties. Water exhibits a nonvanishing electrical conductance. Values of the specific conductance as low as 3× 10–8 Ω–1 cm–1 have been achieved, but in practice, water with ≈ 1×10–6 Ω–1 cm–1 is used even in accurate experiments. The discrepancy results from impurities not removable by simple distillation or ion exchange and from dissolved carbon dioxide.

The dielectric constant (relative permittivity) of water ( = 87.81 at 0 °C) is very high and decreases substantially with increasing temperature to 55.44 at 100 °C [13]. The value at 25 °C is 78.46. These high values are attributed to a cooperativealignment of molecules in the hydrogen-bonded network. The dielectric constant is a key quantity for interpreting solute – solvent interactions, above all in electrolyte solutions. Its change with temperature and pressure gives rise to drastic changes in the solvent properties of water.

Other Properties. Refractive Index. At 25 °C, the refractive index of water is 1.3340. It decreases slightly with increasing temperature.

Standard Enthalpy and Entropy of Formation. The standard enthalpy of formation of water at 25 °C and 1 bar is –286.2 kJ/mol, and the corresponding entropy is 69.98 J mol–1 K–1 [14].

Cryoscopic and Ebullioscopic Constants. The cryoscopic constant Kf = 1.853 g K–1 mol–1 characterizes the freezing-point

depression caused by 1 mol dissolved species in 1000 g H2O. The ebullioscopic constant Kb = 0.515 g K–1 mol–1 characterizes the boiling-point elevation caused by 1 mol dissolved species in 1000 g H2O.

1.2. Aqueous Solutions 1.2.1. General Solvent Properties At ambient conditions, water is a good solvent for many ionic and polar inorganic and organic substances. To some degree, it can even accommodate nonpolar particles and larger nonpolar molecular groups in the interstitial voids of its hydrogen-bonded network. At high temperature the solvent behavior of water can change drastically. Many comparatively inert substances that mix only poorly with water at room temperature become highly soluble or even completely miscible. In contrast, due to the decreasing dielectric constant, water loses its favorable characteristics as a solvent for ionic compounds.

A useful classification distinguishes among three types of hydration, i.e., ion hydration; hydration of organics with polar functional groups such as OH, C=O, and NH; and “hydrophobic” hydration of inert particles and substances with nonpolar molecular groups. Of these, ion hydration has traditionally received the most interest and is comparatively well characterized. The hydration of substances containing polar functional groups is determined by the extent of solute – solvent hydrogen bonding. In such cases, a highly complex and specific behavior is found. Hydrophobic hydration is a nonspecific “nonbonded”interaction that depends on whether water can build up a stable structure around the inert particle, thereby retaining or strengthening its hydrogen bonds. Hydrophobic effects are therefore determined by the nature of water – water interactions rather than by water – solute interactions, and are driven by a favorable entropy change rather than by energetic contributions. Such processes are very sensitive to temperature, and this strong temperature dependence is one of the essential characteristics of hydrophobic hydration.

1.2.2. Solutions of Simple Nonpolar Gases The solubility of simple low-polarity gases (noble gases, aliphatic hydrocarbons such as methane or ethane, or inert inorganic compounds such as SF6) in water at 25 °C is lower than in most organic liquids [15] (Fig. 5). Moreover, in contrast to the normal behavior found with other solvents, the solubility decreases as a function of temperature. Only at high temperature (e.g., >100 °C) does the solubility pass through a minimum and then have a positive temperature coefficient like that of normal solvents [16]. Extensive tabulations of gas solubilities in water are available [17].

The thermodynamics of dissolution of inert gases in water and the structural properties of such systems have received wide attention [18], [19], because they are often used to model the hydration of nonpolar groups in more complex molecules. The same interactions play an important role in determining the properties of colloidal and macromolecular solutions, and in the manner in which proteins acquire their native structure [19], [20]. The various properties may be investigated by considering the transfer of 1 mol inert solute from a nonaqueous to an aqueous environment. The latter solutions are nonideal in the sense that the changes in volume and enthalpy on dissolution are negative. However, this is counteracted by a very large negative change of the excess entropy over the entropy of ideal mixing. This dominance of the entropy over the enthalpy term leads to a strongly positive free energy of mixing and, hence, to the low solubility of inert gases in water.

Figure 5. Solubility of simple gases in water and cyclohexane; x2 is the mole fraction of the gas (reproduced from [21] by permission)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 8 of 82

Page 9: Water Ullmans

FRANK and EVANS [21] have explained this process by an ordering of water molecules and an increase in the degree of hydrogen bonding around the inert solute. This assumption is supported mainly by the existence of solid hydrates of gases and of other small molecules with definite cage structures (clathrates) [22]. In liquid water, the environment of inert gases and inert molecular groups such as the methyl group are now believed to exhibit some residues of the clathrate structure.

At higher solute concentrations, a second important effect occurs: so-called hydrophobic interactions or hydrophobic bonding[18-20]. This term denotes the formation of close-encounter configurations between two inert solutes or molecular groups that are stabilized by a favorable common water structure around the pair (Fig. 6). With larger solutes the configuration shown in Figure 6 may, for example, correspond to the encounter of two methyl groups or larger side chains accompanied by a decrease in the number of nearest water molecules. Hydrophobic bonding may include not only configurations with direct solute contact but also configurations in which one layer of water molecules is located between the hydrophobic pair. Since this association process annihilates solute – water contacts, the association is analogous to that expected if an inert solute is removed from an aqueous environment. Thus, the thermodynamic parameters characterizing hydrophobic bonding have signs opposite to those discussed for hydrophobic hydration.

Figure 6. Schematic representation of a hydrophobic association process A) Two methane molecules form a dimer; B) Stabilization of the conformation of a biopolymer by hydrophobic bonding between two methyl groups or side chains The spheres represent methyl groups or side chains. The shaded area indicates that the processes occur only in an aqueous environment.

1.2.3. Solutes with Hydrophilic Groups The hydration of polar groups is highly specific, and with molecules carrying both polar groups and inert residues, distinguishing between the various effects is usually difficult. Some insight is often obtained by considering the variation of solution properties within homologous series. Typical examples of such systems include aqueous solutions of alcohols, amines, ketones, ethers, fatty acids, nitriles, or polar gases.

A useful classification that may establish some order in this confusing field is based on the strength and number of proton-accepting groups of the nonaqueous interaction partner [23]. For example, nitriles and ketones are weak proton acceptors, and mixing with water is endothermic. In combination with the entropic term, such solutions are slightly stable thermodynamically and liquid – liquid immiscibilities with upper consolute points are usually found, except for the lowest members of the homologous series. If the solute is capable of accepting stronger hydrogen bonds, the mixing process may become exothermic over some temperature ranges, and a tendency exists for a lower consolute point or a closed miscibility gap. Such behavior is often found with alcohols, ethers, or amines. Finally, in some mixtures, hydrogen bonding is very strong. These are thermodynamically very stable and never separate into two liquid phases. Examples are mixtures with polyhydric alcohols such as glycerol and amides.

Even though knowledge of the hydration of comparatively simple systems is imperfect, the results are of major importance forinterpreting the role of water in more complex systems of biological interest. There, hydrophobic hydration and hydrophobic interactions have gained wide attention because these nonbonded interactions play a major role in determining the stability of conformations of biomolecules (e.g., proteins) in solution [19], [20].

1.2.4. Electrolyte Solutions Traditionally, electrolyte solutions have received far more attention than nonelectrolyte systems [24]. The dissolution of a salt in a solvent may be visualized by two steps consisting of (1) dissociation of the salt into gaseous ions and (2) solvation of these ions in the solvent. Consequently, to achieve dissolution, the free energy of solvation must overcome the lattice energy. The polarity and high dielectric constant of water near room temperature favor this process. This, at least qualitatively, explains why the solubility of many salts is higher in water than in nonaqueous solvents. The resulting heats of solution are low, compared to the lattice energies, and in total, the solution process can be exothermic or endothermic.

The dissolution of ions changes the water structure, because the electric field associated with the ionic charge favors definite orientation of the water molecules near the ions. However, other effects also must be accounted for, such as ion – water hydrogen bonds and conditions imposed by the need to form hydrogen bonds with water molecules beyond the first sphere. For anions, most experiments, including neutron-scattering studies [25], give evidence that a hydrogen-bonded configuration in which one of the water protons is directed toward the anion (Fig. 7, right) is preferred to that favored by electrostatic arguments (Fig. 7, left).

With larger univalent ions such as K+ or I–, even the water configuration in the first hydration sphere may be quite isotropic. This leads to the structure-breaking effect: On dissolution of the salt, water behaves as if its temperature were increased, corresponding to a decrease in the solvent structure. This picture is derived from thermodynamic data such as hydration entropies [21], transport coefficients such as viscosity [24], and spectroscopic data [26], [27]. An illustrative example of the quite unusual effects obtained by the addition of structure-breaking salts is the concentration dependence of viscosity at room temperature. For salts such as potassium iodide, the viscosity decreases with increasing salt concentration, in contrast

Figure 7. Possible configurations of an anion – water pair

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 9 of 82

Page 10: Water Ullmans

to the usual expectations. Salts that follow this pattern are termed structure breakers as opposed to structure formers. This structure-breaking effect exhibited by alkali-metal ions (except Li+) and halide anions (except F–) constitutes an anomalous property of aqueous solutions. Multivalent cations and most multivalent anions are structure formers.

Because of the high dielectric constant of water near room temperature, most salts of univalent ions are dissociated completely even at high concentration. At high dilution of the salt, the Debye – Hückel theory of electrolyte thermodynamics and its extensions to transport processes provide a correct description of such systems [24]. In the strict sense, these are, however, limiting laws for vanishing salt concentration, as exemplified by the famous Debye – Hückel limiting law for the mean ionic activity coefficient, the logarithm of which exhibits a square-root dependence on salt concentration [24].

Progress in statistical mechanics has also led to an understanding of concentrated electrolyte solutions. Although such theories are notoriously complex, PITZER and coworkers have derived rather simple semiempirical equations for electrolyte thermodynamics that fit, within experimental accuracy, a variety of data on pure and mixed electrolytes in aqueous solutions over wide ranges of salt concentration and temperature [28].

At elevated temperature, where the dielectric constant of water is low, extensive ion pairing may occur [16]. The same is true for electrolytes of higher valence such as sulfates of alkali, alkaline-earth, and transition-metal ions, because the Coulombic interactions between the ions are stronger than in the case of unielectrolytes [24]. In addition, very specific association can be observed, for example, with zinc, cadmium, or mercury halides, where stable complexes are formed [24]. To a certain degree, such ion pairing can be incorporated in PITZER's approach. A more common way, however, is a description in terms of equilibria between distinct ionic species, which are then treated by a simple law of mass action and the corresponding thermodynamic relations [24].

[Top of Page]

2. Water at High Pressure and Temperature Ernst Ulrich Franck, Gabriele Wiegand, Nicolaus Dahmen

Water plays a part in many technical processes and is used in a wide range of temperatures and pressures. The critical point above which liquid phase and vapor phase no longer coexist is determined by the critical data tc = 373.98 °C, p

c = 22.05 MPa, and c = 322 kg/m3 (see Section Bulk Properties of Liquid Water).

2.1. Properties of Ice 2.1.1. Thermodynamic Properties Twelve modifications of ice are known [29-31]. The phase diagram of ice in the p – T projection in the temperature ranges –100 to +600 °C and pressures up to >104 MPa is shown in Figure 8 [32], [33]. The densities at 500 °C have also been entered for comparison. Natural ice exists as ice Ih and crystallizes hexagonally. In general, only this modification is of interest for technical processes. A metastable cubic variant of this modification also exists. In Figure 8, ice IV (a metastable modification, which occurs in the phase region of ice V) and ice VIII (with a triple point at ca. 5 °C and 2100 MPa, which is formed from ice VI or ice VII at higher pressure and lower temperature) are not shown. Ice IX and ice XII have also been excluded from Figure 8 because they lie outside the temperature and pressure region depicted in this figure. All known triple points are listed in Table 5 [31], [34].

Table 5 shows that modification II, for example, is not formed directly from liquid water on cooling or compression but can originate only from modifications Ih, III, V, or VI. An anomaly of water is that the melting pressure curve for ice Ih exhibits a negative slope up to the triple point Ih – III – liquid at 207 MPa and –22 °C. The values of the melting pressure curve ice Ih – liquid water are summarized below in tabular form down to this triple point [35]. e

The following equation describes the melting pressure curve of ice Ih in equilibrium with water [36]:

The sublimation pressure curve of ice Ih – water vapor is known down to –110 °C:

t, °C p, MPa 0.0098 6.1156×10–4

0.0 0.10132

–5.0 59.8

–10.0 110.8–15.0 155.9–20.0 193.2–22.0 207.4

t, °C p×103, hPa 0 6108

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 10 of 82

Page 11: Water Ullmans

Table 5. Triple points of ice and water modifications

The heat of fusion of ice Ih – water is 333.69 kJ/kg at 0 °C, and the heat of sublimation of ice Ih – water vapor is 2838 kJ/kg at 0 °C [37], [38].

The molar heat capacity Cp of ice was investigated for ice Ih. Down to –140 °C, the following equation is valid [39]:

The parameters of this equation were determined by fitting the data of GIAUQUE and STOUT [35].

–10 2597

–20 1032

–30 380.1

–50 39.37

–70 2.59

–90 0.093

–110 0.0013

Coexistent phases Pressure, MPa Temperature, °C

Ih – liquid – gaseous 6.1×10–4 0.01 *

Ih – liquid – III 207 –22 *

Ih – II – III 200 –30.15 **

Ih – II – VI 1050 –120.15 **

Ih – VI – XII 1550 –168.15 **

II – III – V 350 –21.15 **

II – V – VI 970 –100.15 **

III – liquid – V 346 –17.0 *

V – liquid – VI 626 0.16 *

VI – liquid – VII 2200 81.6 *

VI – VII – VIII 2100 ≈5 *

* from [34]. ** from [31].

Figure 8. Pressure – temperature phase diagram of water in the solid and liquid states Densities at 500 °C have been included for comparison.

T, K Cp, J mol–1 K–1

10 0.276

20 2.050

50 7.931

100 15.879150 22.025200 28.211250 34.829270 37.481

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 11 of 82

Page 12: Water Ullmans

2.1.2. Transport Properties For the thermal conductivity of ice, the following equation is valid in the temperature range 0 to –170 °C [37], [38]:

The self-diffusion coefficient D is [34]

2.1.3. Electrolytic Properties Relative Permittivity. The relative permittivity was determined on polycrystalline ice at a frequency of = 0 Hz:

The relative permittivity increases with decreasing temperature. The phase change from one ice modification to the next-denser one is always accompanied by a jump in to considerably higher values along the phase equilibriumcurves with liquid water [32]. Within one modification region, the relative permittivity increases only slightly with pressure. The more densely the water molecules are packed, the higher is the value of .

The relative permittivity depends on the frequency of an applied electrical field. This frequency dependence provides information on the rotation of dipole molecules in the lattice. The dielectric relaxation time is a measure of how fast the polarization caused by an electrical field decays after the field is turned off. Reorientation times of the water molecules can also be determined by NMR studies [34].

The electrical conductivity of ice was determined, e.g., at –10 °C [34], [40]:

t, °C –0.1 –20.9 –44.7 –65.8 91.5 97.4 104 133

2.1.4. Mechanical Properties The mechanical properties of ice provide information on its stability and load capacity. The mechanical properties of ice Ih are listed below [37], [38]:

2.2. Properties of Fluid Water

Mohs scratch hardness at freezing point ≤2

at –78.5 °C ≈6

Tensile strength (–10 °C) 1 – 2 MPaCompressive strength (–10 °C) 3 – 6 MPaYoung's modulus (–16 °C) 9.42 GPaShear strength (–16 °C) 3.55 GPaPoisson ratio (–16 °C) 0.33

(average)

2.2.1. Thermodynamic Properties For water in the fluid state, equations of state are developed on the basis of experimental p – V – T data and fixed points (e.g., critical data). Exact and sometimes very comprehensive equations of state exist, especially for technical applications [41]. They must sometimes by subdivided for different phase regions. Equations of state that are restricted to parameters interpretable by molecular physics must take into account the high dipole moment of the water molecules and hydrogen bonds between them, especially when the equations apply to moderate temperatures and very high densities [42], [43].

The three-dimensional phase diagram of water is depicted in Figure 9.

Tabular compilations list p – V – T data up to a maximum pressure of 3000 MPa and maximum temperature of 2000 °C [44-

Figure 9. Three-dimensional p – V – T phase diagram of water; the coexistence liquid – gaseous region lies in the darkly marked area [44] C.P. = Critical point

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 12 of 82

Page 13: Water Ullmans

46]. They are based on International Association for the Properties of Steam (IAPS) recommendations and on the International Formulation Committee (IFC) Formulation for Industrial Use. Selected data for temperature and pressure under saturation conditions are presented in Table 6. Corresponding volumes of the gas and liquid phase along the vapor-pressure curve up to the critical point are included [45]. A comparison between the data listed in the SI steam tables [45] and those compiled in the NBS – NRC steam tables [46] shows that for <300 °C, the values deviate from each other by only ca. 0.1 %. The deviations increase with increasing temperature and can reach 1 %. The enthalpy of vaporization h is calculated from the saturation data; it is zero at the critical point. Table 7 shows specific volumes for selected pressures and temperatures in technically relevant regions and also in the supercritical state.

Table 6. Dependence of pressure p, specific volume v of water (l) and water vapor (g), and enthalpy of vaporization h on the temperature in the saturation state at temperatures <100 °C [45]

Table 7. Pressure and temperature dependence of specific volume v (10–3 m3/kg) of water and superheated steam up to 800 °C and 100 MPa [45]*

t, °C p, MPa v1×103, m3/kg vg, m3/kg h, kJ/kg

110 0.14327 1.0519 1.210 2230.0120 0.19854 1.0606 0.8915 2202.3130 0.27013 1.0700 0.6681 2173.6140 0.3614 1.0801 0.5085 2144.6150 0.4760 1.0908 0.3924 2113.3160 0.6181 1.1022 0.3068 2081.2170 0.7920 1.1145 0.2426 2048.0180 1.0027 1.1275 0.1938 2013.2190 1.2551 1.1415 0.1563 1976.8200 1.5549 1.1565 0.1272 1938.5210 1.9077 1.1726 0.1042 1898.4220 2.3198 1.1900 0.08604 1856.2230 2.7976 1.2087 0.07145 1811.8240 3.3478 1.2291 0.05965 1764.6250 3.9776 1.2513 0.05004 1714.5260 4.6943 1.2756 0.04213 1661.5270 5.5058 1.3025 0.03559 1604.7280 6.4202 1.3324 0.03013 1543.6290 7.4461 1.3659 0.02554 1477.6300 8.5927 1.4041 0.02165 1406.0310 9.8700 1.4480 0.01833 1327.6320 11.289 1.4995 0.01548 1241.1330 12.863 1.5615 0.01299 1143.7340 14.605 1.6387 0.01078 1030.7350 16.535 1.7411 0.008799 895.8360 18.675 1.8959 0.006940 721.2370 21.054 2.2136 0.004973 452.6371 21.306 2.2778 0.004723 407.4372 21.562 2.3636 0.004439 351.3373 21.820 2.4963 0.004084 273.5374 22.081 2.8407 0.003458 108.7374.15 22.120 3.1700 0.003170 0.0

p, MPa t, °C

0 100 150 200 250 300 350 400 500 600 700 800

0.1 1.0002 1696 1936 2172 2406 2639 2871 3102 3565 4028 4490 4952

1.0 0.9997 1.0432 1.0904 205.9 232.7 258.0 282.4 306.5 354.0 401.0 447.7 494.3

5.0 0.9977 1.0412 1.0877 1.1530 1.2494 45.30 51.94 57.79 68.49 78.62 88.45 98.09

10 0.9953 1.0386 1.0843 1.148 1.2406 1.3979 22.42 26.41 32.76 38.32 43.55 48.58

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 13 of 82

Page 14: Water Ullmans

Figure 10 shows the specific heat capacity at constant pressure cp as a function of pressure and temperature. The specific heat capacity at the critical point approaches infinity. In the critical region, the isobars show a pronounced maximum as a function of temperature. For increasing pressures, the maxima become smaller but disappear only above 300 MPa.

Values for the isobaric expansion coefficient for temperatures in the range 200 – 1000 °C and pressures between 1 and 100 MPa are listed in Table 8. They pass through a maximum between 300 and 600 °C. The isothermal compressibility T is defined as

Table 9 shows the dependence of T on temperature and pressure in the pressure range 1 – 100 MPa. The maximum error

is ±0.016×10–5 MPa–1. The isothermal compressibility and the isobaric expansion coefficient are quantities derived from the p – V – T data.

Table 8. Isobaric expansion coefficient p in 10–3 K–1 of water and water vapor as a function of pressure and temperature [44] *

15 0.9928 1.0361 1.0811 1.1433 1.2324 1.3779 11.46 15.66 20.80 24.88 28.59 32.09

20 0.9904 1.0337 1.0779 1.1387 1.2247 1.3606 1.6662 9.947 14.77 18.16 21.11 23.85

25 0.9881 1.0313 1.0748 1.1343 1.2175 1.3453 1.6000 6.014 11.13 14.13 16.63 18.91

30 0.9857 1.0289 1.0718 1.1301 1.2107 1.3316 1.5540 2.831 8.681 11.44 13.65 15.62

40 0.9811 1.0244 1.0660 1.1220 1.1981 1.3077 1.4896 1.909 5.616 8.088 9.930 11.52

50 0.9767 1.0200 1.0605 1.1144 1.1866 1.2874 1.4438 1.729 3.882 6.111 7.720 9.076

60 0.9723 1.0157 1.0552 1.1073 1.1761 1.2698 1.4083 1.632 2.952 4.835 6.269 7.460

70 0.9682 1.0116 1.0501 1.1005 1.1663 1.2541 1.3793 1.567 2.467 3.972 5.257 6.321

80 0.9641 1.0076 1.0452 1.0941 1.1573 1.2401 1.3547 1.518 2.188 3.379 4.519 5.481

90 0.9602 1.0037 1.0405 1.0880 1.1488 1.2273 1.3335 1.479 2.013 2.967 3.964 4.841

100 0.9565 0.9999 1.0359 1.0821 1.1408 1.2155 1.3148 1.446 1.893 2.668 3.536 4.341

* The horizontal lines denote the transition between liquid and gaseous state.

Figure 10. Specific heat capacity of water up to 800 °C and 100 MPa; saturated vapor: x = 1, saturated liquid: x = 0 [45]

p, MPa t, °C

200 300 400 600 800 1000

1.0 2.714 1.926 1.563 1.168 0.941 0.789

10 1.317 3.167 2.711 1.398 1.023 0.825

20 1.255 2.63 7.030 1.720 1.121 0.866

30 1.201 2.295 36.75 2.105 1.223 0.906

40 1.153 2.060 7.972 2.538 1.325 0.944

50 1.111 1.884 4.890 2.976 1.423 0.982

60 1.073 1.745 3.711 3.343 1.512 1.016

70 1.038 1.631 3.068 3.563 1.589 1.047

80 1.007 1.537 2.655 3.602 1.650 1.073

90 0.979 1.456 2.363 3.491 1.694 1.095

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 14 of 82

Page 15: Water Ullmans

Table 9. Pressure and temperature dependence of the isothermal compressibility T of water (in 10–5 MPa–1) [47]

In the area of technical thermodynamics, the enthalpy – entropy diagram (h – s) presented in Figure 11 is of interest. It is especially important in the design and operation of power plants.

The velocity of sound w in water also depends on pressure and temperature. In the liquid state and at atmospheric pressure (0.1 MPa), it increases with increasing temperature. In the gaseous region and at constant higher pressure (isobaric), the sound velocity passes through a maximum at ca. 100 °C. At constant temperature, the velocity of sound increases with increasing pressure. Values for the sound velocity up to 600 °C and 100 MPa are listed in Table 10 [46].

Table 10. Sound velocity w in m/s in water and water vapor up to 600 °C and 100 MPa [46]

100 0.952 1.386 2.144 3.288 1.721 1.112

* The horizontal line denotes the transition between liquid and gaseous state.

p, MPa t, °C

0.0 10 20 30 40 60 80 100

10 49.461 46.564 44.725 43.633 43.100 43.287 44.778 47.426

20 48.093 45.358 43.604 42.546 42.015 42.135 43.492 45.942

30 46.772 44.196 42.527 41.507 40.982 41.048 42.285 44.555

40 45.498 43.077 41.492 40.514 39.998 40.020 41.151 43.257

60 43.076 40.955 39.542 38.652 38.167 38.129 39.080 40.902

80 40.815 38.979 37.737 36.943 36.500 36.434 37.243 38.830

100 38.699 37.135 36.063 35.371 34.980 34.911 35.612 37.002

Figure 11. Enthalpy – entropy diagram of water [44]. x = 1: saturated vapor (x = 0 would correspond to the saturated liquid)

p, MPa t, °C

0.0 20.0 100 200 300 400 600

0.1 1401.0 1483.2 472.8 533.7 585.7 632.2 713.9

1.0 1402.6 1484.8 1543.8 517.9 578.4 628.2 712.5

10 1418.1 1500.6 1564.3 1361.3 919.7 581.0 698.9

20 1434.9 1517.6 1586.2 1396.4 1004.4 505.8 684.8

30 1451.3 1534.2 1607.1 1429.1 1073.0 419.0 672.7

40 1467.9 1550.6 1627.3 1459.7 1131.3 627.3 664.1

60 1501.9 1583.0 1665.7 1515.8 1228.1 850.7 666.1

80 1538.1 1615.5 1701.9 1566.4 1307.6 990.8 702.9

100 1577.1 1648.6 1736.2 1612.7 1375.8 1096.7 767.3

* The horizontal lines denote the transition between liquid and gaseous state.

2.2.2. Transport Properties Thermal Conductivity. The thermal conductivity of water as a function of pressure and temperature in the range 200 – 800 °C and 10 – 100 MPa is shown in Table 11. Experimental data for higher temperatures and pressures exist: for 250 °C up to 300 MPa [48] and for 250 to 510 °C up to 95 MPa [49]. The values in Figure 12 have been taken from international

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 15 of 82

Page 16: Water Ullmans

tabular compilations [45]. The tolerances are around 1.5 % for 0.1 MPa and 0 °C, but they reach 8 % for 100 MPa and 600 °C. A characteristic feature is that the thermal conductivity of water shows a very pronounced peak at the critical point. This indicates that the change in structure of water is especially great at the critical point. For the supercritical region, the isobars pass through a minimum up to about 50 MPa. This minimum is more pronounced, the lower the pressure. Above ca. 50 MPa, the isobars decrease monotonously without passing through a minimum.

Table 11. Thermal conductivity of water and water vapor (10–3 Wm–1 K–1) as a function of pressure and temperature in the range 200 – 800 °C and 1 – 100 MPa [45] *

Dynamic Viscosity. The dynamic viscosity of water has been studied experimentally in the high-pressure region up to 300 MPa and 500 °C [50]. Figure 13 shows the temperature dependence of dynamic viscosity for various pressures in the sub- and supercritical regions. These data have been taken from SI steam tables [45]. Like other liquids, liquid water has the highest viscosity at low temperature and high pressure (i.e., the region of high density). The viscosity of water decreases rapidly with increasing temperature, even at high pressure. In water vapor, the viscosity increases at low pressure and density with increasing temperature. Between these two limits with, on the one hand, a high and, on the other hand, a low density, a large region exists in which the viscosity of water barely changes (see Section Bulk Properties of Liquid Water).

Table 12 shows the viscosity, thermal conductivity, and isothermal compressibility of water [45] along the saturation curve of liquid and vapor in the temperature range 150 – 374 °C.

Table 12. Dynamic viscosity , thermal conductivity , and isothermal compressibility T of water (′) and steam (″) in the saturation state [45]

p, MPa t, °C

200 300 400 500 600 800

1.0 35.9 44.3 56.0 68.0 81.0 108.6

10 675 557 66.9 75.6 89.4 118.1

20 684 576 104.9 91.6 98.6 122.7

30 692 593 337 114.1 112.3 130.2

40 700 608 399 152.9 129.2 139.3

60 715 634 476 277 176 161

80 729 653 521 346 235 185

100 742 672 561 412 288 215

* The horizontal line denotes the transition between liquid and gaseous state.

Figure 12. Self-diffusion coefficient of water up to 700 °C and 150 MPa [51] a) 25 °C; b) 200 °C; c) 400 °C; d) 500 °C; e) 600 °C; f ) 700 °C

Figure 13. Dynamic viscosity of water up to 800 °C and 100 MPa [45]

t, °C ′×106, Pa · s

″×106, Pa · s

′×103, W K–1 m–1

″×103, W K–1 m–1

T′×103, MPa–1

T″×103, MPa–1

0.0 1793 9.216 561.0 17.07 0.5101 1 636 700

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 16 of 82

Page 17: Water Ullmans

At the critical point, the thermal conductivity and specific heat capacity extend to infinity.

With increasing temperature, the compressibility increases in liquids but decreases in the gas phase. At the critical point, it approaches infinity.

Self-Diffusion. Very few investigations have been conducted on the diffusion coefficient of water in technically relevant regions of high pressure and temperature [51]. The self-diffusion coefficient as a function of pressure up to 150 MPa and temperature from 25 to 700 °C is shown in Figure 12. In this region, a pressure increase always results in a decrease in the self-diffusion coefficient.

25 890.5 9.866 607.1 18.55 0.4523 316 020

50 547.1 10.62 643.5 20.36 0.4421 92 124

100 281.9 12.27 679.1 25.09 0.4909 10 029150 182.5 13.99 682.1 31.59 0.6215 2 205200 134.4 15.71 663.4 40.10 0.8856 722.4250 106.2 17.49 621.4 51.23 1.470 322.1300 85.96 19.65 547.7 69.49 3.211 195.0350 65.88 23.81 447.6 134.6 16.67 215.2374 43.2 ∞ ∞

2.2.3. Electrolytic Properties The ion product (see Section Bulk Properties of Liquid Water) has been the subject of diverse studies and discussions. The dependence of the ion product on temperature and pressure is presented in Table 13 [52]. It shows that the logarithmic ion product (e.g., at 400 °C and 100 MPa) assumes a value of approximately 10.8 compared with a value of about 14.0 under normal conditions. The autoionization of water is favored by increasing temperature and increasing density. Thus, an increase in pressure at constant temperature always results in an increase in ion product. Shock-tube experiments have shown that water is ionized completely above 1000 °C and 20 GPa.

Table 13. Ion product –log Kw/(mol kg–1)2 of water up to 1000°C and 1000 MPa [52]

For a liquid, water has a very high relative permittivity of ca. 80 under normal conditions because of its unusual structure as a result of the formation of hydrogen bridges (see Section Bulk Properties of Liquid Water).

The relative permittivity of water and steam up to 500 °C and 500 MPa [53] is given in Table 14. The relative permittivity decreases with increasing temperature (isobaric). If the pressure is increased isothermally (i.e., if density is increased), increases. In the critical region, the relative permittivity is ca. 6. Pressure and temperature influence not only the value of relative permittivity, but also the solution behavior of water. High relative permittivities favor solubility and ionization of electrolytes.

Table 14. Temperature dependence of the relative permittivity of water and steam for selected pressures [53]

p, MPa t, °C

0.0 25.0 100 200 300 400 600 800 1000

pSat14.938 13.995 12.265 11.289 11.406

25 14.83 13.90 12.18 11.16 11.14 19.43 23.27 24.23 24.93

50 14.72 13.82 12.10 11.05 10.86 11.88 18.30 19.92 20.80

75 14.62 13.73 12.03 10.95 10.66 11.17 15.25 17.35 18.39

100 14.53 13.66 11.96 10.86 10.50 10.77 13.40 15.58 16.72

250 14.08 13.28 11.61 10.45 9.91 9.74 10.18 11.11 12.02

500 13.60 12.83 11.22 10.00 9.34 8.99 8.85 9.11 9.42

1000 12.91 12.21 10.68 9.45 8.71 8.25 7.78 7.68 7.85

p, MPa t, °C

0.0 25.0 100 200 300 400 500

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 17 of 82

Page 18: Water Ullmans

0.1 87.81 78.46 1.00 1.00 1.00 1.00 1.00

1.0 87.86 78.49 55.44 1.02 1.02 1.01 1.01

10 88.28 78.85 55.76 35.11 20.39 1.17 1.11

20 88.75 79.24 56.11 35.52 21.24 1.64 1.32

40 89.64 80.00 56.77 36.28 22.56 10.46 2.34

60 90.49 80.72 57.39 36.96 23.58 13.28 4.90

80 91.29 81.42 57.98 37.59 24.43 14.88 7.50

100 92.04 82.08 58.55 38.17 25.17 16.05 9.29200 95.20 84.94 61.00 40.66 28.07 19.69 13.86300 97.69 87.34 63.10 42.65 30.15 21.94 16.25400 99.72 89.39 64.94 44.33 31.81 23.62 17.89500 101.42 91.16 66.57 45.82 33.21 24.96 19.14

** The horizontal lines denote the transition between liquid and gaseous state.

2.2.4. Other Physical Properties The interfacial tension of water as a function of temperature at the saturation vapor pressure in the temperature range 150 – 374 °C is listed in Table 15. At 20 °C and 0.1 MPa, of water reaches a value of 72.75 mN/m, which is very high for liquids. With increasing temperature, the interfacial tension decreases almost linearly until it reaches zero at the critical point [45]. In principle, the isothermal pressure dependence of the interfacial tension of pure water cannot be measured. To generate pressure, the interfacial tension of water must always be measured in relation to a second partner. If gases are used for this purpose (e.g., nitrogen, methane, carbon dioxide, neon, or argon), a decrease in interfacial tension with increasing pressure is observed because the gases dissolve in water and also are adsorbed at the interface. Only at >150 MPa does the interfacial tension increase slightly [54].

Table 15. Surface tension of water at 0.1 MPa [45]

The refractive index n of water increases with increasing pressure from 1.334 at 0.1 MPa to 1.347 at 100 MPa.

2.3. Properties of Water Vapor Most of the physical properties of water vapor have been described in Section Properties of Fluid Water. Its standard enthalpy of formation from the elements at 25 °C and 0.1013 MPa, H °298.15 K is 241.99 kJ mol–1 K–1; its standard entropy

of formation S°298.15 K is 188.85 J mol–1 K–1.

For estimation of p – V – T data at moderate pressure, a virial equation is often used:

where v is the specific volume in m3/kg.

The serial development is often stopped after the second virial coefficient B. A compilation of values for the second virial coefficient as a function of temperature is given in Table 16 [55].

Table 16. Second virial coefficient of water vapor [55]

t, °C , N/m

150 48.74200 37.69250 26.06300 14.30350 3.65374 0

t, °C B, 106 m–3 mol–1 t, °C B, 106 m–3 mol–1 t, °C B, 106 m–3 mol–1

20 –1251.5 200 –206.0 575 –37.5

50 –812.2 275 –132.4 675 –26.2

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 18 of 82

Page 19: Water Ullmans

2.4. Water in the Supercritical State [34], [56], [57] Above the critical temperature of 374 °C, water possesses many unusual properties at high density. In contrast to the liquid state, for example, mixing with nonpolar gases and many organic compounds with large, nonpolar molecules is generally possible. In addition, supercritical water can dissolve electrolytes with the formation of ions if its density and relative permittivity are sufficiently high. Thermophysical properties such as viscosity and relative permittivity of supercritical water can be changed continuously in wide regions by pressure variation. Thus, water under supercritical conditions can be used as an extraction agent as well as a versatile reaction medium (e.g., for hydrogenation and oxidation). Combinations of pyrolysis, hydrolysis, and oxidation have already been proposed and developed for the emission-free annihilation of pollutants.

[Top of Page]

3. Water Analysis Georg Schwedt

3.1. Sampling and Sample Preservation (see also Sampling; Sample Preparation for Trace Analysis). Sampling [58-60] is frequently the crucial step in the flawless analysis and evaluation of water. Depending on the type of problem, a distinction must be made between single or random samples and mixed, composite, or average samples (

Sampling – Probability Sampling). In view of representative sampling, the parameters to be considered are very complex. From a homogeneous body of water, a single sample is sufficient to achieve representative sampling. In contrast, in the case of surface waters and stagnant piping systems, several individual samples must be taken. In flowing water, both nonconstant currents and nonhomogeneous water bodies are present. In general, composite samples provide information on certain periods of time, but top loads cannot be discerned. For this reason, the type of sample collection to be used is often stipulated by law to maintain limiting and guideline values. The following DIN processes (no. 38 402) exist for the sampling of unpolluted and polluted waters. Corresponding U.S. processes are found in [60].

Parameter-specific sampling is required, especially in chemical wastewater analysis. Here, too, a legal basis has been established [61], [62]. Further information is contained in ISO standards 5667–1 and -2 (corresponding to EN 25 667–1 and -2) “Water quality sampling: Guidance on the design of sampling programme” and “Guidance on sampling techniques.”

The type and requirements of sample preservation [58-60] depend on the constituents to be determined, their concentrations, their possible mutual interactions, and especially the activity of microorganisms contained in the sample. Keeping the sample in darkness and cooling it represent the most general and the most effective processes. For transport to the laboratory, a temperature of 2 – 5 °C is usually sufficient, but for longer storage times, deep freezing at –20 °C is required. Chemical preservation processes depend on the parameters or substances to be determined. A compilation of the preservation processes and techniques for 75 chemical parameters, as well as microbiological examinations, is given in EN ISO 5667–3 (1994, German version 1995).

3.2. Physicochemical and Sum Parameters In the broadest sense, the term in situ analysis refers to analytical methods that can be applied directly to the sample to be investigated for the measurement of physicochemical and sum parameters, or individual components. In this chapter, in situ analyses refer only to processes that must be conducted directly at the sampling site because changes can occur on the way to the laboratory.

In situ analytical processes for water are based mostly on electrometric measuring principles. They include the measurement of temperature, pH, oxygen and redox potential, and conductivity. Moreover, determination of light absorption in the visible and UV regions belongs to the general physicochemical analysis of water samples. A survey is given in Table 17.

Table 17. Physical and physicochemical parameters in water analysis

100 –459.8 375 –82.0 775 –18.1150 –295.4 475 –54.5 975 –7.6

Sampling of wastewater A 11Stagnant water A 12Groundwater aquifers A 13Raw water and drinking water A 14Flowing water A 15Seawater A 16Falling, wet precipitation in the liquid state of aggregation A 17Water from mineral and medicinal springs A 18Water from swimming and bathing pools A 19 Tidal waters A 20Cooling water for industrial use A 22

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 19 of 82

Page 20: Water Ullmans

As a result of the large number of substances present in water of various types, cummulative determinations have been developed. These provide preliminary information on the degree of pollution or the presence of certain groups of substances. In the narrower sense, sum parameters more or less completely identify organic substances after their oxidation to carbon

Parameter Method Remarks

Temperature, K ( °C)

electronic measurement with temperature probe (DIN 38 404-C 4)

for temperature-dependent quantities: solubility of gases, pH value, electrical conductivity, density

Density, g/cm3 pycnometer, precision aerometer, bending vibration method (DIN DEV-C 9)

in stronger mineralized waters

Electrical conductivity , Ω–1 cm–1

conductivity cell with cell constant K measure electrolyte concentration

Conductance G, Ω–1

(EN 27888-C 8, ISO 7888) check K, determine temperature coefficient ,25

pH value a (H3O+) electrometric (potentiometric) measurement with glass electrode and saturated calomel or silver – silver chloride reference electrode (DIN 38 404-C 5)

measure at sampling site because of rapid changes due to physical, chemical, and biological processes: asymmetry balance, correction of electrode slope, consideration of temperature compensation, measurement in low-electrolyte waters with long adjustment times

Redox potential, mV

electrometric (potentiometric) measurement with Pt electrode against reference electrode (DIN 38 404-C 6)

interference by electrode coating and in low-oxygen waters by small changes in concentration of dissolved oxygen

Absorption in visible (436 nm) or UV range (254 nm)

measurement of spectral decadic absorbance A ( ), determination of the spectral absorption coefficient (SAC) ( ) or the specific spectral absorption coefficient (SAC/DOC *) corresponding to ( ) DOC (436 nm: EN ISO 7887-C 1; 254 nm: DIN 38 404-C 3)

especially a measure of the concentration of aromatics (254 nm); determination of color (436 nm)

Turbidity by nephelometry: measurement of attenuation of transmitted (860 nm) monochromatic IR light or of scattered radiation (860 nm, scattering angle 90°) (EN 27 027-C 2, ISO 7027)

detection of sand, loam, and clay particles, of bacteria and other microorganisms

* DOC = dissolved organic carbon; see Table 18.

dioxide (Table 18). The total organic carbon (TOC) value gives the total content of organic substances, the permanganate index detects only easily oxidizable substances, and the amount of biochemically degradable substances is given by the biological oxygen demand (BOD). Determination of parameter ratios provides useful information for water monitoring self-purification processes [58], [63]. For instance, the BOD5 : COD ratio represents a characteristic value that is specific for wastewater and characterizes its biodegradability. Cumulative determinations such as electrolytic conductivity, residue on evaporation, cation exchange, and pH (Tables 17 and 18) can also be used for plausibility testing (e.g., via ion balances) in conjunction with the results of individual determinations. Table 19 shows the differentiation of the sum parameters for haloorganic substances, which also contribute to the characterization of a group of substances. The determination procedure is based on the combustion of a sample at ca. 1000 °C in a stream of oxygen. The combustion products enter an electrolysis cell, and chloride is titrated coulometrically (the sum of all halides is given as chloride value) [64].

Table 18. Chemical and biochemical sum parameters

Parameter Definition: determination method

Total dry residue GT, mg/L

volume based, mass of dissolved and undissolved (nonvolatile) water constituents (based on unfiltered sample) left after a prescribed drying process (DIN 38 409-H 1)

Filtrate dry residue FT, mg/L

residue on evaporation; volume-based mass of dissolved (nonvolatile) water constituents left after a prescribed filtration and drying process, according to EC drinking water guideline at 180 °C (DIN 38 409-H 1: 105 ± 2 °C)

Substances removable by

according to DIN 38 409-H 2

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 20 of 82

Page 21: Water Ullmans

Table 19. Differentiation of sum parameters for haloorganic compounds (parameters in µg/L, calculated as Cl)

filtration A, mg/L

Residue on ignition GA, mg/L

ignition (550 °C) of the filtrate dry residue (DIN 38 409-H 2)

Cumulative determination by cation exchange, mmol/L

exchange of cations by oxonium ions

Total hardness, degrees of hardness [e.g., d GH b or as d (Ca2+ + Mg2+)], mmol/L

EDTA-Na2 a and Eriochrome Black T (pH 10) (DIN 38 406-E 3–3) or by

AAS c: DIN 38 406-E 3–1 or ISO 7980

Acid and base capacity (KA,4.3, KA,8.2 and KB,4.3, KB,8.2), mmol/L

neutralization titration with HCl or NaOH against methyl orange or phenolphthalein or with electrometric end point indication (pH 4.3 or 8.2)

Calcuable therefrom: m and p alkalinity, mmol/L d

acid capacity: volume-based quantity of hydronium ions that a sample can absorb until it reaches a certain pH; base capacity: volume-based quantity of hydroxyl ions that a sample can absorb until it reaches a certain pH. An aqueous CO2 solution with a ion concentration (mmol/L) of 1 % of the concentration of dissolved CO2, a temperature of 25 °C, and an ion strength of 10 mmol/L has a pH of 4.3. An aqueous solution with >1 mol/L

ions, a temperature of 25 °C and an ionic strength of 10 mmol/L has a pH of 8.2. Details in DIN 38 409-H 7. Acid and base capacity are determined for the calculation of c(CO2), , and . Influenced by the presence of other weak acids

TIC, mg/L (calculated as C)

total inorganic carbon to be derived from the carbon sum , which is determinable for Q

C > 0.5 mmol/L, according to TIC = m alkalinity – p alkalinity (DEV D 8); or

determination in the course of TOC determination d

TOC, mg/L (calculated as C)

total organic carbon determination by oxidation of all C compounds to CO2 (UV irradiation in the presence of an oxidant such as peroxodisulfate at ca. 60 °C (measurement of CO2 by IR spectrometry, coulometry, or

conductometry, or after reduction to methane by GC/TCD e or FID f ) (DIN 38 409-H 3)

DOC, mg/L (calculated as C)

dissolved organic carbon (mostly after membrane filtration – 0.45 µm pore size)

POC, mg/L (calculated as C)

particulate (undissolved) organic carbon

VOC, mg/L (calculated as C)

volatile organic carbon

Permanganate index, mg/L

mass per unit volume of oxygen that is equivalent to the mass per unit volume of consumed permanganate ions (permanganometric, detects easily oxidizable substances) (DIN 38 409-H 5, ISO 8467)

COD, mg/L determination of oxidizability with potassium dichromate, volumetrically or photometrically (at 340, 445, or 585 nm) (DIN 38 409-H, ISO 6060)

BOD, mg/L determination of the volume-based mass of oxygen that is consumed for oxidation of the biochemically oxidizable constituents contained in a 1-L water sample in n days (generally n = 5, BOD5) with the metabolic activity of a corresponding microbiocenosis at 20 °C (DIN 4045; DIN 38 409-H 51, ISO 5815); oxygen determinations volumetrically, manometrically, or electrometrically with oxygen electrode

a Ethylenediaminetetraacetic acid disodium salt. b GH = Gesamthärte (total hardness). c Atomic absorption spectrometry. d For definition of m and p alkalinity see Section Characterization of the Chemical State of Water. e TCD = thermal conductivity detector. f FID = flame ionization detector.

Parameter Detectable substances

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 21 of 82

Page 22: Water Ullmans

Most of the processes mentioned also exist in a variant for on-line analysis. Equipment for on-line analysis allows the course of processes in the waste treatment plant to be controlled with robust (i.e., often also no- or low-maintenance) measuring devices.

3.3. Inorganic Analysis The analytical procedures for inorganic cations and anions range from simple analytical colorimetric testing methods in the field, spectrophotometry, electrochemical methods (especially polarography and inverse voltammetry), and ion chromatography (IC) to atomic absorption spectrometry (AAS), optical atomic emission spectrometry with inductively coupled plasma (ICP – OES), as well as inductively coupled plasma coupled with high-resolution mass spectrometry (ICP – MS) [65], [66].

TOX total halogens in organic compounds (total organic halogens)DOX dissolved organic halogensAOX adsorbable (mostly on activated carbon) organic halogensPOX purgeable organic halogens (DEV H 25)VOX volatile organic halogensEOX extractable organic halogensTOCI total of organic chlorineDOCI dissolved organic chlorine

3.3.1. Determination of Cations Because of the efficiency of instrumental methods, cation analysis is carried out mostly in water without any special sample preparation. However, trace enrichment processes [67], [68] or digestion (usually UV digestion for the destruction of organic substances) must be applied in the case of less selective methods, such as spectrophotometry, or because of interference by organic substances. [Trace enrichment is conducted with simultaneous matrix separation on selective ion exchangers and, after chelation with complexing agents such as dithiocarbamates, on chemically modified silica gel or polymer adsorbents or on activated carbon, which can also be used as flow injection analysis (FIA) technique.] A survey of the generally accepted analytical methods and of the cations that can be analyzed is presented in Table 20. The analytical methods standardized for water analysis are collected in the “German Standard Methods for the Examination of Water, Wastewater, and Sludge” (Deutsche Einheitsverfahren zur Wasser-, Abwasser- und Schlammuntersuchung, DEV).

Table 20. Methods of cation analysis

Apart from classical spectrophotometry (and fluorescence spectrometry for a few elements), high-performance methods of atomic spectrometry are available today. These include atomic absorption spectrometry with flame AAS ( Atomic Spectroscopy – Flame Atomic Absorption), graphite furnace technique (flameless AAS; Atomic Spectroscopy – Electrothermal Atomic Absorption), the hydride technique (for elements that form volatile hydrides), the cold vapor technique (for Hg, Atomic Spectroscopy – Hydride and Cold Vapor Techniques, and as a multielement determination method,

Method Determinable cations

Spectrophotometry (with selective reagents, partly as FIA-technique), colorimetrically as quick test

Al, Ag, As. Cd, Co, Cr(III, VI), Cu, Fe, Hg, K, Mn, Ni, Pb, Zn

Fluorescence spectrometry (with selective reagents)

Al, Be, U

Atomic absorption spectrometry (AAS), flame or graphite tube technique

Digestion/extraction with dithiocarbamate (DIN 38 406-E 21)

Ag, Bi, Cd, Co, Cu, Ni, Pb, Tl, Zn

Hydride AAS-technique As, Bi, Ge, Pb, Sb, Se, Sn, Te

Cold vapor technique (DEV-E 12) Hg

Atomic emission spectrometry with inductively coupled plasma (ICP – OES) (DIN 38 406-E 22)

Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Ti, V, W, Zn, Zr

ICP mass spectrometry (ICP – MS) Ag, Al, As, B, Ba, Bi, Ca, Cd, Co, Cr, Cs, Fe, Hg, K, Li, Mg, Mn, Mo, Na, P, Pb, Rb, S, Sb, Se, Sn, Sr, Th, Ti, Tl, U, V, W, Zn

Polarography and voltammetry (DIN 38 406-E 16)

Cd, Co, Cu, Ni, Pb, Tl, Zn

Potentiometric stripping analysis Ag, As, Bi, Cd, Co, Cu, Fe, Hg, In, Mn, Ni, Pb, Sb, Se, Sn, Tl, U, Zn

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 22 of 82

Page 23: Water Ullmans

optical emission spectrometry with inductively coupled plasma as the excitation source ( Atomic Spectroscopy – Inductively Coupled Plasmas). The latter can be coupled with high-resolution mass spectrometry (ICP – MS technique) to achieve still lower detection limits and larger linearity ranges and to measure the abundance ratios of isotopes [69]. Despite the advances in and efficiencies of these methods, matrix effects and interference still must be taken into account. They must often be eliminated by special sample preparation, such as selective extraction or concentration. In addition, electrochemical methods such as polarography and voltammetry [70], especially the stripping technique — after concentration on the electrode — ( Analytical Voltammetry and Polarography – Stripping Voltammetry.) and potentiometric stripping analysis (PSA, Analytical Voltammetry and Polarography – Stripping Chronopotentiometry (SCP).), are important in water analysis. In PSA, metal ions capable of amalgam formation are first deposited on a working (usually glassy carbon) electrodeby potentiostatic electrolysis. The metals are then oxidized and again dissolved (stripped), the time required being proportional to their concentration in solution. The potential – time curves are evaluated.

In cation analysis, besides determination of total concentrations, element species analysis (speciation), i.e., differential analysis of one element according to different chemical bonding and physical form of state, is becoming more and more important [71]. Examples are the determination of chromium as Cr(III) and Cr(VI) [72], arsenic as As(III) and As(V) [73], and inorganic and organic tin compounds [74]. For speciation, coupling processes with a chromatographic separation system (gas and liquid chromatography) and atomic spectrometric detection have gained importance [75].

3.3.2. Anion Analysis Spectrophotometry is the classical analytical method in anion analyis, as well. For some elements present in the anionic state in water (e.g., B, V, Mo), methods of cation analysis are also used (Table 20). In addition, for some special anions such as fluoride, nitrate, and cyanide or sulfide, ion-selective electrodes [76] have proved useful in direct potentiometric determination (with use of the standard addition process and integrated in FIA systems for calibration). Most of the processes can be automated; i.e., they can also be employed as FIA systems (Table 21). Furthermore, special problems such as determination of the total nitrogen or total phosphorus concentration in wastewater samples can also be solved by using the FIA technique [77], [78]. In these processes, digestion of all nitrogen compounds (oxidation under UV radiation with peroxidisulfate) to nitrate or of organic phosphorus compounds to orthophosphate (by acid and microwave radiation) is integrated into the flow system.

Table 21. Methods of anion analysis

Since the early 1980s, ion chromatography has been of utmost importance in anion analysis [79]. Being a type of high-performance liquid chromatography, it uses separation columns with anion exchangers on a polymer or silica gel basis, as well as ion-pair separation systems in conjunction with different detection methods such as conductivity, UV, and amperometric detection [59]. If solutions with carbonate – hydrogencarbonate as eluting ions are employed, the suppressor technique for neutralization of the mobile phase (by means of an ion-exchange membrane in the H+ form) is used in conjunction with conductivity detection. Ions that absorb UV, such as nitrate, nitrite, bromide, iodide, or thiocyanate, can be detected directly (by using UV transparent eluents); ions that do not absorb UV can be detected by indirect UV detection (UV-adsorbing eluent such as phthalate buffer or complexons [80] ); and electrochemically active (oxidizable) anions, such as iodide and nitrite, can also be detected amperometrically [81]. This wide range of separation and detection methods is a characteristic feature of ion chromatography. In anion analysis, the latest method is capillary electrophoresis in the form of zone electrophoresis [82].

Method Determinable anions

Spectrophotometry (after chemical conversions with selective reagents)

borate, chromate, cyanide, fluoride, iodide, nitrate, nitrite, phosphate, silicate, sulfate, sulfide, sulfite, thiocyanate (others such as borate/B, molybdate/Mo, vanadate/V; see also cation analysis in Table 20)

Potentiometry with ion-selective electrodes

fluoride, cyanide, nitrate

Ion chromatography

(DIN 38 405-D 19) in slightly polluted water

fluoride, chloride, nitrite, phosphate, bromide, nitrate, sulfate

(DIN 38 405-D 20) in wastewater

chloride, nitrite, phosphate, bromide, nitrate, sulfate

(DIN 38 405-D 22) chromate, iodide, sulfite, thiocyanate, thiosulfate

3.3.3. Determination of Dissolved Gases The most frequently determined dissolved gases in water are oxygen, ozone, carbon dioxide, and chlorine.

Oxygen. The classical Winkler process for the measurement of oxygen is based on the conversion of manganese(II) to manganese(IV) hydroxide under alkaline conditions (“oxygen fixation”), with subsequent iodimetric determination (EN 25 813,ISO 5813). Measurements with an oxygen electrode (as oxygen sensor) according to Clark, are based on the following principle. In an electrolyte-filled reaction chamber, a gold cathode and a silver electrode, which is involved in the redox processes and acts as the anode, are polarized with a direct voltage of 790 ± 10 mV. The reaction chamber is sealed with a gas-permeable membrane. Oxygen from a water sample reaches the cathode by diffusion [e.g., via a teflon (polytetrafluoroethylene) film] and is reduced. Silver dissolves correspondingly at the anode. The current flowing in the cell is

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 23 of 82

Page 24: Water Ullmans

proportional to the partial pressure of oxygen [58].

Carbon Dioxide (see Section Characterization of the Chemical State of Water). The carbon dioxide content (“free carbonic acid”) can be calculated from the total inorganic carbon (TIC = m alkalinity — p alkalinity), with the m and p alkalinities (see Section Physicochemical and Sum Parameters) determined by titration taking the pH and conductivity into account (in the pH range 6.0 – 8.4) [58]. Carbon dioxide can also be measured with the aid of a gas-sensitive electrode. In the case of a CO2 electrode, as with the oxygen sensor, gas molecules diffuse through a semipermeable membrane into an electrolyte solution (hydrogencarbonate). On entry of CO2, the pH of this hydrogencarbonate solution is shifted, corresponding to the equilibrium between carbon dioxide and hydrogencarbonate. The pH shift is measured. The standard processes DEV-D8 [calculation of dissolved carbon dioxde (free carbonic acid), carbonate, and hydrogencarbonate ions] and DIN 38 404-C 10 (calcium carbonate saturation of a water sample) are available for investigations of the carbonate balance including the calcium carbonate system that are important in practice.

Chlorine. For the determination of free chlorine and total chlorine (e.g., in swimming pool water), both chemical and physicochemical processes (using an electrode) are available. The photometric or colorimetric (also possible volumetrical) process (DIN 38 408-G 4) is based on the reaction of elemental chlorine (as well as hypochlorous acid and hypochlorite ions, where the sum = free chlorine) with N,N-diethyl-p-phenylenediamine to give a red oxidation product. Chloramines can be detected only after conversion with iodide (bound chlorine). The content of chlorine can also be determined with a gas-sensitive electrode (measurement of redox potential).

Ozone. Like free chlorine, ozone is determined photometrically.

3.3.4. Quick Test Processes For most of the parameters described in Sections Anion Analysis and Determination of Dissolved Gases, chemical quick tests — also called alternative tests compared to laboratory processes — are available on the market. They include test papers (for “semiquantitative” analysis of various metal ions), test strips, colorimetric tests, and gas testing tubes for substances that can be purged out of water [83].

As an analysis involving carrier-bound reagents, test strips are square reaction zones made of paper applied to plastic strips. With the help of a pocket instrument, test strips can also be evaluated by reflectometry.

The application of gas testing tubes has been extended from workplace monitoring to include water analysis. In water analysis, volatile substances are purged out of water with the help of air and a gas-washing bottle. A certain volume of air is sucked through the testing tube (located between gas-washing bottle and pump) by using a bellows pump. The length of the colored zone is a measure of a definite concentraton of the gas to be determined.

Colorimetric quick tests are available in the form of cuvette tests with color comparison, as a rotary-disk comparator, or a color-scale sliding comparator. If two cuvettes are used — one cuvette for the untreated sample — the self-color of a water sample can be compensated for. Visual colorimetry according to the Lambert – Beer law based on the principle of remission measurement (in a direct-light process) with greater layer depths (compared to the rotary-disk comparator) and higher sensitivity finds application in color chart tests. The color of a sample of water after the addition of reagents is compared with the colors on a plastic color chart valid for one concentration range in each case (lower and upper limits). The reagent sets used in these processes can also be applied in photometric determinations for on-the-spot analysis with the help of battery-driven pocket photometers.

3.4. Organic Analysis The analysis of organic substances in water is characterized on the one hand by cumulative detection on the basis of sum parameters (see Section Physicochemical and Sum Parameters) and by biochemical and biological test methods (see Section Biochemical Methods) for total effect-oriented analysis. On the other hand, analysis of individual substances is carried out by using efficient chromatographic separation methods [84]. The groups of substances most often analyzed are pesticides, polycyclic aromatic hydrocarbons, haloorganic hydrocarbons, phenols, hydrocarbons from mineral oils and mineral oil products, and surfactants.

3.4.1. Spectrometric Methods Unlike the analysis of inorganic water constituents, UV - visible (UV – VIS) spectrometry plays a much smaller role for organic substances. Substances determined by UV – VIS spectrometry include, e.g., uric acid and pyridine as single substances; phenols (phenol index DIN 38 409 H 16), organic complexing agents such as ethylenediaminetetraacetic acid (EDTA), nitrilotriacetic acid (NTA), and phosphonic acids (as bismuth complexation index IBIK; humic acids; lignosulfonic acids [85]; and surfactants (DIN 38 409 H 20 and 23) as groups of substances.

For turbid water, derivative spectrometry is the method of choice not only to eliminate the interference of colloids and particles, but also to obtain characteristic spectra (“shoulders” of the original spectrum give minima in the first derivation), e.g., in the determination of phenol [59].

Fluorescence spectrometry ( Ultraviolet and Visible Spectroscopy – Fluorimetry is based on the property of substances absorbing energy (excitation) in the form of light (UV – VIS range) and emitting it as radiation with lower energy, i.e., longer wavelength (fluorescence). Fluorimetry is most important as a detection method in conjunction with separation methods. It is also used to determine the concentration of chlorophyll A (DIN 38 412 L 33) in algae in the examination of the trophic state of waters; in the qualitative analysis of polycyclic aromatic hydrocarbons (PAHs); in surfactant analysis; and in the characterization of humic substances. Humic substances are characterized spectroscopically [86] by spectral absorption coefficients at 245 and 436 nm and by fluorescene with an excitation wavelength of 330 (or 341) nm and fluorescence radiation of 410 (or 460) nm. Complexable paramagnetic metal ions quench this fluorescence.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 24 of 82

Page 25: Water Ullmans

Infrared spectroscopy is based on the absorption of electromagnetic radiation for the initiation of mechanical vibrations of atoms or functional groups in a molecule. This method is used primarily in the characterization and quantification of hydrocarbons. Sample preparation involves extraction with CCl4 or 1,1,2-trichloro-trifluoroethane (DIN 38 409 H 18). Nonhydrocarbons (such as surfactants) that are also extracted in this process are separated on a polar adsorbent. The characteristic absorption bands used are 3.38 µm ( = 2958 cm–1) for the CH3 group, 3.42 µm ( = 2924 cm–1) for the CH2

group, and 3.30 µm ( = 3030 cm–1) for the aromatic CH group ( Infrared and Raman Spectroscopy – The Concept of Group Frequencies, Infrared and Raman Spectroscopy – Methyl and Methylene Groups, Infrared and Raman Spectroscopy – Alkene Groups, Infrared and Raman Spectroscopy – Aromatic Rings). Humic substances and surfactants [87] can also be characterized by IR spectroscopy.

3.4.2. Gas and Liquid Chromatography As physicochemical separation methods, chromatographic methods [84] based on the distribution of substances between a stationary and a mobile phase in conjunction with various detectors play an important role in qualitative and quantitative analysis, especially of organic substances (see also ion chromatography, Section Anion Analysis). Gas chromatography with a gas as the mobile phase is especially suitable for volatile substances (i.e., substances that evaporate without decomposition). The stationary phase consists of a solid (adsorbent) or fluid. The highest separation efficiencies are attained in capillary columns. Liquid chromatography with a liquid mobile phase is conducted as thin layer (planar) chromatography (TLC) or as high-performance liquid chromatography (HPLC) in columns (with small particles of 3 – 10 µm diameter as separating materials (for details, see [84]).

Gas Chromatography. In general, sample preparations are required for gas chromatography. These preparations involve direct transfer of the substances to be analyzed from the water sample to the separation column or first to an organic (easily evaporable) solvent. Headspace analysis ( Gas Chromatography – Static and Dynamic Head Space Analysis) is used for volatile organic water constituents such as halogenated hydrocarbons [88], [89]. The sample of water is heated to a given temperature in a headspace sample vial provided with a silicone membrane and rubber seal. A controllable equilibrium is established between the liquid and the gaseous phase. With a gastight syringe, a defined volume can be withdrawn from the headspace via the silicone membrane and rubber seal of the sample vial and used for GC analysis without the interfering matrix water. In the purge-and-trap process, a highly pure stream of helium gas is first passed through the water sample. Thehighly volatile organic compounds enter the helium gas phase (purge process) and are then retained and concentrated in an adsorption tube (with an adsorbent such as tenax, i.e., a linear polymer of para-2,6-diphenylphenylene oxide; trap process). The substances are desorbed by rapid heating (thermal desorption) and then transferred directly to the sample inlet of a gas chromatograph ( Gas Chromatography – Thermal Desorption Units). Other sample preparation processes for GC analysis are liquid – liquid (e.g., with n-pentane) and solid – liquid extraction (concentration and separation from water on adsorbents made of chemically modified silica gels with functional groups such as n-alkyl, usually C18, phenyl, aminopropyl etc.). Microextraction is also possible [90], [91]. The efficiency of a GC analysis system is determined not only by the separation column but also by the type of detector used. For relevant water constituents, a compilation of GC processes with the detector used in each case is presented in Table 22 (examples: pesticides [92], polychlorinated biphenyls [93], coupling with a mass selective detector [94], and mass spectrometry [88]).

Table 22. Applications of GC analysis

Thin layer chromatography is widely applied, especially in the analysis of polycyclic aromatic hydrocarbons. For this purpose, one-dimensional or two-dimensional separation is conducted after extraction with chloroform in accordance with

Group of substances GC process or detector (DIN)

Halogenated hydrocarbons with low volatility electron capture detector (ECD) (DIN 38 407-F 2)

Highly volatile halogenated hydrocarbons

After extraction with, e.g., n-pentane or n-hexane

ECD (DIN 38 407-F 4)

Headspace analysis ECD (DIN 38 407-F 5)

Phenoxyalkane carboxylic acids with liquid – liquid extraction and derivatization

mass spectrometric detector (MSD) (DIN 38 407-F 14)

Phenols (selected) after derivatization ECD (DIN 38 407-F 15)Aliphatic hydrocarbons flame ionization detector (FID)Organic solvents FIDSelected organic nitrogen and phosphorus compounds (pesticides) after concentration by solid-phase extraction

phosphorus – nitrogen detector (PND) (DIN 38 407-F 6) combination of GC and mass spectrometer

Aromatic hydrocarbons photoionization detector (PID)Hydrocarbons combination of GC and FTIR (Fourier

transform infrared spectrometry)Benzene, toluenes, xylenes (BTX) FID (DIN 38 407-F 9)Vinyl chloride ECD (DIN 38 413-P 2)Polychlorinated biphenyls (PCBs) ECD

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 25 of 82

Page 26: Water Ullmans

DIN 38 409-H 13. This analysis is limited to six representative PAHs. Quantitative evaluations can be achieved by means of a thin layer scanner via UV absorption or fluorescence measurement.

Automated multiple development (AMD) is used as a special development technique in the separation of selected organic pesticides (DIN 38 407-F 11). This is a modern technique for stepwise development, which — unlike the classical process — starts with the solvent having the highest polarity and ends with the solvent having the lowest elution strength. Development is carried out in a vertical chamber using high-performance (HP) TLC plates. It is started with a run of 8 mm and is continued up to 25 times with a 3-mm run each time. After each step, the solvent is removed from the chamber by suction, and each step is preceded by layer conditioning. The AMD technique with the important effect of spot focusing is one of the most efficient developments in TLC.

High-Performance Liquid Chromatography. In comparison with GC and TLC, HPLC has the advantage that aqueous phases can be used on reversed-phase materials (silica gels with n-alkyl groups such as C18, RP-18, Liquid Chromatography). Thus, extraction of the substances to be analyzed from water samples and transfer to an organic phase are usually not required. Concentration can also be conducted in a precolumn directly before the separation column. Another advantage (especially over TLC) is the large selection of different types of detectors (UV – VIS detectors with fixed or variable wavelengths or diode array systems for the detection of spectra; and fluorescence, amperometric, and conductivity detectors) and the possibility of coupling with mass spectrometry [95]. The level of instrument technology makes a high degree of automation possible — starting with sample application, precolumn concentrations, gradient elution (e.g., increasing methanol concentration in a water — methanol eluent mixture), and recording of spectra, and ending with evaluation. In water analysis, this gradient elution is applied mainly to separation of pesticides (DIN 38 407-F 12) [96] and polycyclic aromatic hydrocarbons (ISO 7981–1, DEV process F 8) [97].

3.5. Biochemical Methods Biochemical and biological testing methods [98] are used to test the effects of chemical and physical parameters, i.e., especially of substances dissolved (present) in water, on enzymes (biochemical) or on certain test (water) organisms. The term “ effects” refers not only to the acute and chronic toxicity of individual substances, but also to detrimental biological effects in general caused by all the noxious agents together (i.e., total noxiousness). Synergistic and antagonistic effects are also discerned. Although the qualitative and quantitative identification of single substances is possible in chemical and physicochemical analyses (Sections Sampling and Sample Preservation , Physicochemical and Sum Parameters, Inorganic Analysis, Organic Analysis), no information can be provided on the joint action of several potentially toxic substances. If a damaging effect is determined in the proceses described below, the substances responsible must be identified by chemical analysis. Thus, biochemical and biological test methods are of significance as a precursor to instrumental analysis. They represent an important preliminary test and supplement the analytical determination of substance concentrations.

The enzymatic and biochemical processes employed include enzyme inhibition tests and enzyme-coupled immunochemical tests, i.e., enzyme-linked immunosorbent assays (ELISA; ). ELISA tests use the ability of organisms to form antibodies against foreign higher-molecular mass substances and to bind the foreign substances with these antibodies. By coupling environmentally relevant low-molecular mass substances to larger proteins, antibodies are produced that respond specifically to these low molecular substances (haptens) [99]. The antibodies obtained are immobilized on the inner wall of plastic tubes (or in the cavities of microtiter plates) for an antigen – antibody reaction. The water sample and an enzyme conjugate, which consists of the analyte coupled with an enzyme, are then placed in a tube of this type. The free analyte molecules in the sample and the analyte molecules bound to the enzyme compete for the free binding sites on the immobilized antibodies. After a rinsing step, a solution containing the enzyme substrate is added. A dye is produced as a result of the enzymatic reaction. The higher the intensity of the color, which can be measured photometrically, the lower is the concentration of analyte molecules in the sample. The ELISA tests are available for a series of pollutants, especially for herbicides such as alachlor, cyanazine, atrazine, 2,4-dichlorophenoxyacetic acid, and metolachlor; for insecticides such as aldicarb, carbofuran, and chlorodienes, such as chlorodane, aldrin, dieldrin, heptachlor, endrin, and endosulfan; for benomyl, metalaxyl, and the fungicide captan; and also for PAH and polychlorinated biphenyls (PCBs).

Enzyme Inhibition Tests. Of the enzyme inhibition tests, the acetylcholinesterase inhibition test for phosphate esters, phosphate thioesters, and methyl carbamate is of special interest Enzyme and Immunoassays – Enzyme Assays: Determination of Inhibitors) [100]. In this test, enzyme activity is inhibited by substances that act as neurotoxins. The inhibitory effect can be determined photometrically on the basis of the substrate reaction (e.g., with thiocholine). A biosensor is produced by combining an enzyme immobilized on a membrane with a pH (flat-bed) electrode [101]. The substrate reaction can then be determined from potential measurements. For the analysis of a drinking water sample, the activity of the enzyme membrane of the electrode is determined by dipping the membrane into a substrate-containing (acetylcholine) solution. As a result of enzymatic hydrolysis, acetic acid is released in the enzyme membrane. The resulting shift in pH is measured by a pH electrode and is a measure of the enzyme activity. After incubation in an inhibitor-containing sample of drinking water, the active center of the enzyme is blocked irreversibly, e.g., by phosphate ester molecules. Consequently, the activity of the enzyme is reduced. Subsequent determination of activity with the same substrate solution as before shows a lower activity. In this manner, detection limits of 0.1 µg/L can be achieved under optimum incubation conditions without previous analyte enrichment.

Biological Tests. Of the biological tests, the luminescent bacteria test DIN 38 412-L 34 and L 341) [102] is widely applied because it is easy to handle. A NaCl solution is added to the deep frozen test charges to give a temperature of 15 °C, and the bioluminescence (as a part of the bacterial metabolism) is measured. After addition of a wastewater sample and an incubation time of 30 min, the luminescence is measured again. The decrease in luminescence is a measure of bacterial toxicity. For wastewater analysis, dilution series are made to determine the dilution for an inhibition of 20 % by means of interpolation.

The reviews “Water Analysis” in the journal Analytical Chemistry survey the developments in methods and analyses over a period of two years [103].

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 26 of 82

Page 27: Water Ullmans

[Top of Page]

4. Hydrological Cycle and Water Use Fritz H. Frimmel and Birgit C. Gordalla

4.1. World Water Balance The appearance of the earth is characterized by the natural occurrence and physical properties of water. As the basis of all life, water is an essential component of nature. Although the total global water volume of 1.38×109 km3 accounts for only about 0.23 ppm of the mass of the earth, it covers 71 % of the earth's surface. The distribution of water among different regions of occurrence is shown in Table 23. Fresh water, including ice deposits, accounts for only 2.6 % (36×106 km3) of the total volume. With reasonable effort, only about 10 % of the fresh water available is technically accessible to man at present (Table 24). Thus, the water supply of the earth's population will be based essentially on this volume of about 3.7×106 km3.

Table 23. Assignment of water according to regions of occurrence according to [104]

Table 24. Breakdown of proportions of total fresh water usable for water supply according to [104]

4.2. Hydrological Cycle Many of the physicochemical properties of water, which lead to its normal appearance in nature, are due to the strong dipole character of the H2O molecule (dipole moment: 6.13 ×10–30 Cm, 1.85 D, see also Chap. Water as a Solvent).

The fact that water occurs in nature in all three states of aggregation is of special importance. The flow of energy that reaches the earth's surface with incident solar radiation results in an average surface temperature of 15 °C [105], with considerable variation as a function of time and location. This energy input causes the evaporation of water from oceans, other surface waters, and vegetation. In the atmosphere, water vapor condenses and can be transported as visible clouds by wind. Water returns to the surface of the earth as rain, snow, and ice. It then either superficially flows away, seeps away, or contributes directly to surface water. Thus, it is again available for evaporation, partly via vegetation (Fig. 14). On an annual average, the earth's 149×106 km2 of dry land receives 745 mm of precipitation (111 000 km3), of which 477 mm (71 000 km3) evaporates from the land and 40 000 km3 flows into the sea [106].

From a global point of view, this cycle is closed (i.e., no water is lost). In individual regions, however, considerable deviations from the theoretical mean values for the mass fluxes occur due to long-range transport and climatic differences. Even in a relatively small country such as Germany, which belongs to the temperate climate zone, substantial differences are observed. Thus, the average temperature in July is between +17 °C and +18 °C in the North German lowlands, but about +20 °C in the Upper Rhine rift. The amounts of annual precipitation are between 500 and 700 mm in the North German lowlands, between 700 and 1500 mm in the low mountain ranges, and up to more than 2000 mm in the Alps [107]. The long-term average annual temperature in Germany is +9 °C. Since 1990, increases of up to 2 °C have been observed in the annual temperature and deviations of up to 5 °C in the monthly average values compared with the long-term values, have

Region Volume, 106 km3 Percentage

Oceans 1348 97.39Polar icecaps, icebergs, and glaciers 27.82 2.01Groundwater and soil moisture 8.06 0.58Lakes and rivers 0.225 0.02Atmosphere 0.013 0.001Total 1384 100

Region Percentage Volume, 106 km3

Underground (up to a depth of 800 m) 9.86 3.55Lakes 0.35 0.13Rivers 0.003 0.001Atmosphere 0.04 0.014Total 10.25 3.7

Figure 14. Schematical representation of the hydrological cycle

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 27 of 82

Page 28: Water Ullmans

been recorded. Despite these temperature variations with time, the precipitation balance remained almost unchanged. The long-term average value for precipitation over the area covered by the western German states (248×103 km2) is 837 mm/a, while the value for the eastern states (former German Democratic Republic, 109×103 km2) is 612 mm/a. Taking the small distance (a few hundred kilometers) between these two regions — that are in the same climatic zone — into consideration, this variation is remarkable. For the entire Federal Republic of Germany, this gives an average precipitation level of 768 mm/a, which corresponds to a precipitation of 274×109 m3/a [105]. The water supply for Germany ensues from the difference between the long-term average values for precipitation and evaporation plus the long-term average value for inflow from other countries (Table 25) [107].

Table 25. Annual supply of water in Germany (based on 30 years of balance data, reference period 1931 – 1960)

4.3. Demand for Water About 70 wt % of the human body consists of water. The WHO works from the fact that an adult weighing 60 kg must consume 2 L water per day in the form of drinking water, beverages, or food to avoid damage to his health. The quantities that apply to small children (body weight 10 kg) and infants (body weight 5 kg) are 1 and 0.75 L, respectively. A much higher water consumption per person is observed, depending on the standard of living and the economic structure. The FAO speaksof a shortage of water in a certain country or region when the annual water supply from domestic sources is <1000 m3 per resident. Thus, a world population of 6×109 people will require at least 6×1012 m3 water per year. Although this amount is only about 0.16 % of the total water available for drinking purposes, clearly discernible bottlenecks occur in some countries. They result from an inadequate regional supply of water. Examples of countries having a shortage of water (FAO, as of 1990) are given below, together with the amounts of water per citizen per year:

Germany has an annual water supply of about 2000 m3 per person. According to data released by the Statistical Federal Office (Statistisches Bundesamt), the total water requirement in 1991 was 47.9×109 m3 and was distributed as shown in the following:

Precipitation levels and amounts of water

Relative values, %

Average supply of water 163.5× 1 09 m3/a 100

Precipitation minus evaporation 95×109 m3/a 58

Average inflow from other countries

68.5×109 m3/a 42

Average precipitation 768 mm/a 100(274×109 m3/a)

Average evaporation 501 mm/a 65(179×109 m3/a)

Rwanda 885 m3

Botswana 797 m3

Algeria 754 m3

Burundi 663 m3

Syria 615 m3

Kenya 613 m3

Tunisia 465 m3

Israel 372 m3

Jordan 222 m3

Yemen 222 m3

Mauretania 203 m3

United Arab Emirates 188 m3

Libya 154 m3

Saudi Arabia 148 m3

Egypt 35 m3

Cooling water, thermal power stations 28.8×109 m3

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 28 of 82

Page 29: Water Ullmans

A breakdown of the utilization of water provided by the public water supply is given below:

In households and small businesses, a daily water consumption of about 145 L per person was estimated over the past ten years. With 140 L per person (as of 1994), a downward trend has been observed in the past two years [108]. A breakdown ofwater consumption in households is given in the following:

The length of the mains employed by the public water supply to distribute water covers about 386 000 km (estimation for 1994) [108]. The materials used for pipes (as of 1992) are listed below [109]:

In Europe, the statistical data for water consumption per person are in the range of 116 to 260 L/d (see below):

4.4. Source of Water Used Sources used by the public water supply in Germany as of 1994 include [108]

Industry 11×109 m3

Public water supply 6.5×109 m3

Agriculture, irrigation 1.6×109 m3

Households and small businesses 68 % Industry 13 %Loss in the mains and consumption by waterworks 13 %Public facilities 6 %

Baths, showers, hygiene 40 %Toilets 31 %Laundry 13 %Dish washing 7 %Cleaning 5 %Garden 2 %Cooking, drinking 2 %

Cast iron 49 %Plastic 26.5 %Cement 13 %Steel 8.7 %Other 2.8 %

Austria 215 LBelgium 116 LDenmark 175 LFinland 150 LFrance 161 LGermany 145 LHungary 153 LItaly 214 LLuxembourg 183 LThe Netherlands 173 LNorway 167 LSpain 131 LSweden 195 LSwitzerland 260 LUnited Kingdom 161 L

Groundwater 63.6 %Springs 7.8 %

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 29 of 82

Page 30: Water Ullmans

Groundwater is the most important natural source of drinking water in Germany [110].

Water resources are used regionally in different ways, depending on geographic location and on geological, pedological, and hydrological conditions (Fig. 15).

As a result of their earlier history, different waters differ in their chemical composition and biological nature. Water pollution protection was the first issue of European Community environmental policy [111] and has been most extensively dealt with. The quality requirements demanded for the production of drinking water are stipulated in a European guideline [112] and, in Germany, in technical regulations [113], [114]. The requirements to be fulfilled by drinking water are also formulated in an ECguideline [115] which has to be transferred into national law of the member states. In the German decree on drinking water [116], [117], these requirements are translated into national law. The guiding principles for technical realization of the production and distribution of drinking water contained in DIN 2000 [118] have also been included partly in the decree on drinking water.

In the United States, the EPA establishes maximum contaminant level goals (MCLG) and maximum contaminant levels (MLC) as well as best available technologies for regulated water contaminants that are stipulated in the Safe Drinking Water Act (SDWA) with current amendments. A review of promulgated, proposed, and anticipated EPA drinking water regulations, which also concern disinfectants/ disinfection byproducts, analytical methods, information collection, rankings of regulatory priorities, and public notification requirements, is given in [119].

If the requirements stipulated for drinking water are not met by raw water, it must be treated before it is used as drinking water.

4.5. Water Treatment A wide range of processes is available for the treatment of water (e.g., [120], [121]). The selection of individual treatment stages and the entire concept of treatment depend primarily on the water quality of the source used, its productivity, and consumer requirements. This means that optimum water treatment depends in each case on a particular situation, which is usually unique. For this reason, many different concepts are encountered in practice. In principle, the individual units of many treatment systems are largely similar. The water treatment concept of the Wiesbaden – Schierstein waterworks is presented in Figure 16. The large number of technical stages used permits the conversion of Rhine water to drinking water, which not only is qualitatively perfect but corresponds quantitatively to the demands of the population.

The versatile use of the Rhine and — despite relevant decrees and laws — the poor raw water quality found, as well as the effects of disasters and disturbances on water quality, are a real challenge to a water treatment strategy. The individual treatment stages shown in Figure 16 comply with the state of the art. In the case of better raw water conditions, some of these stages can be omitted. In the total concept of the Wiesbaden – Schierstein waterworks, the key function of the subsoil as a natural or, at least, close-to-natural treatment stage is discernible (see also [122], [123]). The bank filtrate of the river water supplements the dominant role played by groundwater in the public water supply of Central Europe. Through utilization of subsoil, nature takes on the function of technical measures in wide areas by means of its cleansing power. The reliable use of this natural self-cleansing power requires a knowledge of its capacity and responsible, future-oriented management. This includes a knowledge of the hydrological and geological situation, the allotment of protective zones with restricted use of the land, and extensive prevention of groundwater pollution [124]. The entire water utilization cycle with water above and below ground must also be regarded and assessed as a closely linked system. The German water management law (Wasserhaushaltsgesetz) [125] provides the legal framework for this purpose.

Typical concentrations of water constituents as present in some types of raw and finished water (drinking water) are given in Table 26. The limiting values found in the German decree on drinking water are given for comparison. Larger German water-supplying companies have been selected as examples. Note that the drinking water supply of the city of Munich is based on

Riverbank filtration and replenishment of groundwater 15.3 % Surface water 13.3 %

Figure 15. Regional differences in the use of water resources in Germany (as of 1994) [108]. By courtesy of the Bundesverband der deutschen Gas- und Wasserwirtschaft e. V. (Federal Association of German Gas and Water Suppliers)

Figure 16. Principle of water treatment in the Wiesbaden – Schierstein waterworks

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 30 of 82

Page 31: Water Ullmans

the direct utilization of groundwater without technical treatment: i.e., the drinking water quality is the same as the raw water quality. The Lake Constance water supply is cited as an example of the utilization of standing surface waters (lakes, reservoirs), and the water supply of Düsseldorf reflects the use of the Rhine riverbed filtrate (riverbank filtrate) as a typical example of utilization. Accordingly, characteristic compositions are obtained for the raw waters. Groundwater has a relatively high water hardness (calcium and magnesium ions), especially when it comes from limy subsoil, has a slightly elevated nitrate content compared with the geogenic background, and otherwise shows the absence of larger influences of civilization, which is guaranteed by the protective function of the ground. These influences would be discernible from increased values for boron and phosphate. The water from Lake Constance (raw water) is correspondingly softer and exhibits the characteristic chemical features of surface water, e.g., a noticeable concentration of organic water constituents (DOC). In addition, slight anthropogenic influences are discernible, which can be seen from the fact that phosphate and the widely used EDTA are detected. The raw water quality of the Rhine bed filtrate is characteristic of the load of a variedly and greatly used flowing body of water, a load that is reduced by the natural cleansing power of the ground.

Table 26. Typical concentrations of constituents and properties of raw and drinking water compared to selected limiting values in the decree on drinking water

Parameter Calculated as

Limit Groundwater Munich a

Surface water Lake Constance b

Riverbank filtrate (Rhine near Düsseldorf) c

Raw/drinking water

Raw water

Drinking water

Raw water

Drinking water d

Arsenic, mg/L As 0.01 <0.001 0.0014 0.0014 0.002 0.002Cadmium, mg/L Cd 0.005 <0.001 <0.00002 <0.00002 <0.0001 <0.0001Chromium, mg/L Cr 0.05 <0.003 0.00020 0.00018 <0.0005 <0.0005Cyanide, mg/L CN 0.05 <0.0005 <0.0005 <0.0005 <0.01 <0.01Fluoride, mg/L F 1.5 0.13 0.083 0.08 0.09 0.09Lead, mg/L Pb 0.04 <0.003 <0.0002 <0.0002 <0.001 <0.001Mercury, mg/L Hg 0.001 <0.0005 <0.000002 <0.0000013 <0.0003 <0.0003Nickel, mg/L Ni 0.05 <0.003 0.00065 0.00067 <0.0008 <0.0008Nitrate, mg/L NO3 50 8.5 4.55 4.65 16.3 16.6

Nitrite, mg/L NO2 0.1 <0.005 <0.005 <0.005 <0.01 <0.01

Polycyclic aromatic hydrocarbons (sum), mg/L

C 0.0002 <0.00003 <0.00005 <0.00005 <0.00005 <0.00005

Organochlorine compounds (sum), mg/L

0.01 <0.0005 <0.0001 <0.001 <0.005 <0.005

Tetrachloromethane,

mg/L CCl4 0.003 <0.0001 <0.00002 <0.00002 <0.0001 <0.0001

Herbicides and pesticides, mg/L

single substance 0.0001

single <0.00003

single <0.00005

single <0.00005

single <0.0001

single <0.0001

total 0.0005 total <0.00016

total <0.0001

Polychlorinated, polybrominated biphenyls and terphenyls, mg/L

n. deter. single <0.000005

single <0.000005

single <0.00001

single <0.00001

Antimony, mg/L Sb 0.01 <0.001 0.00013 0.00013 <0.0018 <0.0018Selenium, mg/L Se 0.01 <0.001 0.00014 0.00013 <0.0005 <0.0005Color, m–1 0.5 <0.1 n. detect. n. detect. 0.08 <0.01Turbidity, TE/F 1.5 0.05 0.20 0.07 0.2 0.1Threshold odor number 2 at 12 °C n. detect. n. detect. 3 at 20 °

C1 at 20 °C

3 at 25 °C 1 at 25 °CTemperature, °C 25 10.4 5.0 5.2 13.0 13.0pH value at T 6.5<pH<9.5 7.62 8.05 (5.0 °

C)7.93 (5.2 °C)

7.24 (13.0 °C)

7.26 (13.0 °C)

pH after calcite saturation at T

pHC 7.91 7.89 n. deter. n. deter.

Electric conductivity (at 25 °C), µS/cm

2000 537 322 320 702 698

Oxidizability, mg/L O2 5 0.5 3.7 2.7 3.11 0.60

Aluminum, mg/L Al 0.2 <0.02 0.0073 0.0035 0.003 0.009

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 31 of 82

Page 32: Water Ullmans

Ammonium, mg/L NH4 0.5 <0.05 <0.01 <0.01 <0.01 <0.01

Barium, mg/L Ba 1 <0.005 0.018 0.018 0.09 0.08Boron, mg/L B 1 <0.05 0.025 0.025 0.09 0.09Calcium, mg/L Ca 400 82.4 48.9 48.6 78 76Chloride, mg/L Cl 250 5.5 5.3 5.7 66 65Iron, mg/L Fe 0.2 <0.01 0.0055 0.0019 <0.06 <0.06Kjeldahl nitrogen, mg/L N 1 n. deter. <1.0 <0.3 n. deter. n. deter.Magnesium, mg/L Mg 50 19.6 8.0 8.0 10.8 10.7Manganese, mg/L Mn 0.05 <0.001 0.0008 0.00014 0.01 <0.01Phenols, mg/L C6H5OH 0.0005 n. deter. n. deter. n. deter. <0.02 <0.02

Phosphorus, mg/L PO4 6.7 <0.01 0.074 0.077 0.32 1.22

Potassium, mg/L K 12 1.0 1.3 1.3 4.8 4.8Silver, mg/L Ag 0.01 <0.001 <0.00005 <0.00005 <0.0002 <0.0002Sodium, mg/L Na 150 3.6 4.5 4.5 41 41Sulfate, mg/L SO4 240 25.3 33.4 33.3 65 66

Hydrocarbons, mineral oils, mg/L

0.01 n. deter. n. deter. n. deter. <0.01 <0.01

Surface active substances

Anionic substances, mg/L

MBAS e 0.2 n. deter. <0.1 <0.1 <0.01 <0.01

Nonionic substances, mg/L

BiAS f n. deter. n. deter. n. deter. n. deter. n. deter.

Copper, mg/L Cu 3 <0.001 0.00075 0.0007 0.025 0.001Zinc, mg/L Zn 5 <0.01 0.0006 0.0012 0.008 <0.004Number of colonies at 20 °C

100 0 – 2 21 0 0 – 1 0 – 1

at 36 °C 100 0 – 2 9 0 3 0 – 1

Escherichia coli/100 mL at 36 °C

0 n. detect. 10 n. detect. 0 – 1 0

Coliforms/100 mL at 36 °C 0 n. detect. 10 n. detect. 26 0Residue on evaporation, mg/L

328.5 194 n. deter. n. deter.

Residue on ignition, mg/L n. deter. 158 n. deter. n. deter.Base capacity up to pH 8.2, mmol/L

0.26 (10.4 °C) 0.04 (9 °C)

0.06 (9 °C) 0.51 (16 °C)

0.51 (16 °C)

Acid capacity down to pH 4.3, mmol/L

2.45 (20 °C)

2.44 (20 °C)

3.26 (18 °C)

3.21(18 °C)

Carbonate hardness, °dH 14.2 °KH 6.86 6.83 9.0 8.8Sum alkaline earth (total hardness), mmol/L

2.869 1.58 1.58 2.39 2.34

°dH 16.1 8.85 8.85 13.4 13.1

Silicon dissolved, mg/L Si 1.9 1.34 4.2 4.7Strontium, mg/L Sr n. deter. 0.44 n. deter. n. deter.Oxygen, mg/L O2 9.2 10.0 26.6 2.8 4.7

DOC, mg/L C <0.2 1.2 1.1 1.17 <0.55Spectral absorption coefficient (254 nm), m–1

n. deter. 3.06 1.32 3.48 0.76

AOX, mg/L Cl n. deter. 0.005 0.026 <0.010 <0.010Nitrilotriacetic acid (NTA), mg/L

n. deter. <0.0005 <0.0005 <0.0005 <0.0005

Ethylenediaminetetraacetic acid (EDTA), mg/L

n. deter. 0.0022 0.0015 0.0036 0.0005

Total activity, Bq/L 137Cs + 134Cs <5 <0.1 <0.1 n. deter. n. deter.

a By courtesy of Stadtwerke München. b By courtesy of Bodensee Wasserversorgung. c By courtesy of Stadtwerke Düsseldorf. d n. deter. = not determined; n. detect. = not detectable. e MBAS = Methylene blue-active substance.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 32 of 82

Page 33: Water Ullmans

For the two treated waters, comparison of raw water and drinking water data shows clearly that many parameters are not significantly changed after treatment. The effects of oxidative and adsorptive water treatment steps are observable in the improvement of both the organoleptic values, such as color, turbidity, and odor, and the microbiological parameters (decrease in coliform organisms and number of colonies). These quality criteria also represent the primary target values of drinking water treatment. Furthermore, a decrease is observed in the values of those parameters that are partially (spectral absorption coefficient, oxidizability) or exclusively (DOC) a measure of the concentration of organic water constituents [126]. The elimination of organic water constituents is important to reduce the danger of future regrowth of bacteria in the mains during water distribution and also to restrict the formation of trihalomethanes during future chlorination. Halogen-containing disinfection byproducts can also increase the AOX value produced by industrial wastewater. The increase of phosphorus in the finished water from the riverbank filtrate is due to the addition of a phosphate – silicate mixture as a corrosion inhibitor.

Despite the quality improvement of surface water (e.g., [127] discernible since the 1980s, the need for efficient water treatment continues to exist. This results from the obligation of the public water supply system to provide sufficient amounts of qualitatively perfect drinking water at all times. In addition, precautions against accidents and disasters that can impair the quality of surface waters and are not excludable require appropriate protective measures.

Water for technical use is also of particular significance. Great demands are made on boiler feed water, process water, water for pharmaceutical and food technology uses, and water for swimming pools. These demands can be met only by purpose-oriented water treatment processes (see e.g., [128], [129] and Chaps. Ion Exchange and Membrane Separation Processes in Water Treatment). As a rule, a combination of various process steps is required.

[Top of Page]

5. Adsorption Processes in Water Treatment Klaus Johannsen and R. Scott Summers

5.1. Introduction Adsorption processes are used for the removal of dissolved organic substances for production of drinking water from both surface and groundwaters. The most important adsorbent for water treatment is activated carbon, and it was initially applied to improve the taste and odor of water. Activated carbon is also used to remove micropollutants, such as pesticides and halogenated hydrocarbons, and to limit the formation of disinfection byproducts (DBPs), such as trihalomethanes, by removal of natural organic matter (NOM), which can form disinfection byproducts upon chlorination. Activated carbon is applied in two forms: powdered activated carbon (PAC) and granular activated carbon (GAC). Powdered activated carbon is added as a suspension to the water to be purified and removed by sedimentation or filtration. In GAC treatment, fixed beds are used that contain the activated carbon. Monographs give an overview of the fundamentals and application of activated carbon in drinking water treatment [130-134].

5.2. Properties of Activated Carbon (see also Carbon, Carbon – Mechanical Properties, Carbon – Adsorption Properties) The most common raw materials used in the production of activated carbons are bituminous coal, peat, wood, lignite, coke, and coconut shells. Typical production processes include charring by pyrolytic carbonization of the raw material at <700 °C, followed by thermal activation with oxidizing gases such as steam, CO2, and air at 800 – 1000 °C. Development of the pore

structure in this manner yields internal surface areas of 500 – 1100 m2/g. The ability of activated carbon to adsorb large quantities of material is related directly to its porous nature. Pore volume and surface area distributions are determined by mercury porosimetry and nitrogen and benzene absorption. The intitial portions of the nitrogen adsorption isotherms are oftenutilized to calculate the total pore surface area of an adsorbent. Standardized aqueous test methods, such as the molasses test, methylene blue number, phenol test, and iodine number, have been developed to evaluate the equilibrium adsorption capacity of PAC or GAC. In particular, the iodine number gives a good indication of the microporosity and is often correlated with the BET surface area [130], [131].

To minimize operational problems when using GAC in packed beds, the particle size and particle-size distribution must vary only within certain limits. Typical average particle diameters of granular activated carbon are in the 0.9 – 1.6-mm range. Smaller particles yield faster overall adsorption kinetics; however, too many particles smaller than 0.5 mm may cause high pressure losses in the fixed bed. Typical average particle diameters of powdered activated carbon are between 0.01 and 0.04 mm. Smaller particle sizes impede handling and separation of PAC. Values for the material density of GAC are in the 1800 – 2000 kg/m3 range. The dry particle density of GAC (i.e., dry mass per particle volume) is between 500 and 850 kg/m3, while its wet particle density is between 1300 and 1500 kg/m3. The apparent (bed, bulk, or filter) density (i.e., dry mass per vessel volume) of GAC is in the 300 – 500-kg/m3 range; that for PAC is slightly higher (400 – 700 kg/m3) [130], [131], [133], [134].

5.3. Adsorption Theory ( Adsorption) Adsorption is the accumulation of a substance, the adsorbate, onto a solid, the adsorbent; it is driven by physical or chemical forces.

Physical adsorption is based on nonspecific electrostatic forces, such as dipole – dipole interactions, hydrogen bonding, and van der Waals forces. In general, nonpolar compounds dissolved in a polar solvent (water) are more strongly adsorbed by a nonpolar adsorbent (activated carbon). Compared with chemical adsorption (chemisorption), physical adsorption is less specific, operates over longer distances, has weaker forces and energies of bonding, and is reversible to a greater degree.

Chemisorption forces approach that of a covalent or electrostatic chemical bond, and because of the specificity of the bond

f BiAS = bismuth-active substance.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 33 of 82

Page 34: Water Ullmans

between adsorbate and surface, the adsorbate cannot accumulate at more than one molecular layer; i.e., a monolayer of adsorbate is formed. In water treatment, physical adsorption dominates.

In general, adsorbability on activated carbon is inversely related to the water solubility of the adsorbate and the pH. The adsorption of a substance increases with increasing number of double bonds ( -bonds) in the molecule, increasing molecular size, and increasing ionic strength of the solution. However, for high molar mass compounds such as humic substances, size exclusion can be observed [131]. A good relationship between the surface area or volume of micropores of activated carbons and their adsorption capacity for several compounds has been found [131], [134].

5.4. Adsorption Equilibrium ( Adsorption – Thermodynamics, Equilibria, and Heat of Adsorption) Adsorption equilibrium studies — termed isotherms, because they are conducted at constant temperature — establish the relationship between the concentration of adsorbate in the liquid phase c and its concentration in the solid phase q. For determination of equilibria, the bottle point technique [131], [134] is normally used. Here, defined quantities m of PAC or ground GAC are added to a series of bottles containing a volume of solution L with an initial adsorbate concentration c0. The samples are mixed until equilibrium is reached: i.e., no measurable change in the liquid-phase concentration occurs. Activated carbon is separated from the liquid phase, and c is determined. Based on a mass balance for fresh carbon, the solid-phase concentration q is calculated according to

The equilibration time depends on particle size. For carbon with particle sizes <0.1 mm, an equilibration time of 2 – 5 d is adequate.

The two most commonly used models to describe the equilibrium relationship between c and q are the Langmuir [135] and the Freundlich equations [136]. Both are two-parameter equations. The Langmuir equation has a maximum solid-phase concentration value. In the low-concentration range, a linear relationship exists between the concentration of adsorbate in the solid and the liquid phase. The Freundlich equation is a empirical power function of

where KF is the Freundlich constant and n is the Freundlich exponent. These two parameters can be determined from the measured data by nonlinear regression of c and q or by linear regression of log c and log q. The slope of log c versus log q is equal to the exponent n, and the value of q at c = 1 equals KF. The Freundlich equation is most often used to describe experimental data, but it is valid only in the concentration range of these data. Experimental adsorption data for a variety of compounds and carbons are summarized in [131], [137], [138].

When more than one adsorbable component is present in a solution, the compounds compete with each other for available adsorption sites on the carbon surface. If the complete reversibility of adsorption is assumed, the ideal absorbed solution (IAS) theory accounts for these interactions by using data only from the single-solute isotherms [139-143]. The ideal absorbed solution theory has the flexibility to incorporate many of the single-solute isotherm model equations; thus, for multicomponent systems it can be used with the model that best describes the single-solute data. The use of the IAS theory coupled with the Freundlich equation was applied successfully to multicomponent systems with solutes of similar characteristics [131]. However, in other situations, the IAS theory was not as successful and modifications were made to account for unequal competition for active sites and irreversible adsorption [144]; changes in the Freundlich parameters at low concentration [141], [145]; and differences in accessible surface area [146].

Most (80 – 95 %) of the natural organic matter in water is adsorbable. However, the composition of natural organic matter in water is complex; only a small percentage of the components can be identified specifically, and quantification of the few identifiable compounds is problematic because of their low concentration. Typically, NOM is quantified by the amount of dissolved organic carbon or the UV absorbance at 254 nm. This complexity of natural organic matter presents a problem in describing adsorption equilibria. The adsorption analysis approach, developed to describe the adsorption of an unknown mixture, divides NOM into fictive components of different adsorbabilities and uses the IAS theory to quantitate their competition [131]. The adsorption analysis approach has been applied successfully to several waters to demonstrate the influence of initial NOM concentration, pH, treatment processes, carbon type, and presence of tracer compounds. One drawback is that it does not account for the effects of size exclusion, which is likely important for the molecular size range of NOM [131].

In most polluted raw water, the NOM concentration is much higher than that of specific micropollutants to be removed. The adsorption capacity of carbon for the specific micropollutants can be lowered significantly in the presence of adsorbing NOM. Figure 17 illustrates the influence of background NOM on the adsorption isotherms of the pesticide metazachlor [N-(2,6-dimethylphenyl)-N-(1-pyrazolylmethyl)-chloroacetamide] [147]. At both pesticide initial concentrations c0, the presence of 10 mg/L DOC causes a reduction of ca. 60 % in the adsorption capacity. Furthermore, Figure 17 shows the impact of the initial concentration of pesticide on the isotherm. At c0 = 2 µg/L, the adsorption capacity is much lower than at c0 = 200 µ/L for both DOC values. Thus, even the presence of 0.3 mg/L DOC causes a capacity reduction at the low initial concentration of the micropollutant.

(5.1)

(5.2)

Figure 17. Influence of initial concentration and background DOC concentration on metazachlor adsorption isotherms [147]

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 34 of 82

Page 35: Water Ullmans

These effects can be explained by the occurrence of competitive adsorption between metazachlor and background NOM. Models have been developed that calculate the competitive adsorption between target compound and background NOM using the IAS theory [147-150]. These competitive effects have been found for most pesticides, but not for some chlorinated hydrocarbons [131].

5.5. Adsorption Kinetics ( Adsorption – Kinetics) Although adsorption at an active site of the carbon surface is a fast reaction, the resistance to mass transfer of the adsorbate from the bulk liquid to the adsorption sites is high enough that the overall kinetics of the adsorption process must be considered in the design of both PAC and GAC systems. The driving force for mass transfer is the concentration difference of the adsorbate in the bulk solution and in the carbon particle. This exists as long as the solid-phase concentration on carbon is not in equilibrium with the liquid-phase concentration.

For the fixed-bed application of GAC, the breakthrough curve for the ideal plug-flow reactor (PFR) with no mass-transfer resistance is shown in Figure 18 (curve a). This ideal breakthrough curve may be used to define the time at which the amount of adsorbate applied to the GAC bed equals the amount that can be adsorbed at equilibrium. This time is sometimes referred to as the stoichiometric breakthrough time tstoich and can be calculated from the following mass balance equation and isotherm data:

where c0 is the influent concentration, q0 is the solid-phase concentration in equilibrium with c0, is the volumetric flow rate, and m is the adsorbent mass. Mass-transfer resistances, axial dispersion, and unfavorable equilibrium conditions can cause the breakthrough curve to deviate from ideal PFR behavior as shown for a single solute in Figure 18 (curve b). The total amount of solute adsorbed should be the same for the ideal PFR and the case with mass-transfer resistance and is the difference between the quantity applied and that in the effluent. This can be calculated by the following mass balance:

where ce is the effluent concentration. The value of the integral corresponds to the area bounded by the c = c0 line and the breakthrough curve. It represents the total capacity of the GAC in the reactor. Since the total amount adsorbed satisfies Equation (5.4) regardless of the shape of the breakthrough curve, the shaded areas F1 and F2 in Figure 18 must be equal. If they are not equal, the calculated capacity from the isotherm q0 does not agree with the column capacity.

In practice, solutes to be removed from water are present in mixtures and, as discussed in Section Adsorption Equilibrium , the presence of other adsorbates — both more weakly and more strongly adsorbing — will decrease adsorption of the target compound. This is also illustrated in Figure 18 (curve c). The concentration and adsorbability of the different components of the mixture determine the shape and time of the breakthrough curve of the target compound. If the mixture contains high concentrations of solutes that are adsorbed more strongly, ce of the target compound may become greater than c0 (i.e., “overshoot” may occur). If the mixture contains mostly specific compounds, breakthrough of the target compound can be calculated by using competitive adsorption models. If significant concentrations of NOM are present, competitive adsorption models are not viable, when a component of NOM is adsorbed irreversibilly. Compared to the target compound, this NOM component progresses deeper into the GAC bed and is adsorbed first. Because of its irreversible adsorption, this NOM component “blocks” the GAC surface (fouling), thus reducing the adsorption capacity for the target compound. Currently, only site-specific calibrated models can simulate this adsorption behavior [131], [134].

The breakthrough behavior of NOM is important in controlling the formation of DBPs and is illustrated in Figure 18 (curve d). Typically, an immediate breakthrough of 5 – 20 % is displayed, and if this level is maintained for any length of time, this fraction is considered nonadsorbable as shown in Figure 18. If ce begins to rise immediately, then the immediate breakthrough may also include a fraction due to external mass-transfer resistance. The NOM breakthrough curve normally begins to plateau at a value 10 – 30 % lower than c0, which can be attributed to very slow adsorption or biological degradation [131], [134].

A wide range of mathematical models have been developed to account for equilibria and kinetic components such as axial dispersion, external (film) mass transfer, and intra particle pore and surface diffusion [131]. The most commonly used model for micropollutants is the plug-flow homogeneous surface diffusion model, which includes external mass transfer and intra particle surface diffusion, and can utilize the IAS theory and the Freundlich equilibria model [131]. A simplified model considers the internal mass transfer as a linear driving force [151], [152]. For NOM adsorption the pore diffusion component seems to be important.

5.6. Biological Processes Most drinking water sources contain a biodegradable NOM component in a concentration of 5 – 20 %, which is increased an additional 5 – 20 % by ozonation. Since this fraction is not removed in most treatment plants, biological degradation occurs to

(5.3)

(5.4)

Figure 18. Comparison of breakthrough curves a) Ideal (negligible mass-transfer resistance); b) Single solute; c) Specific solute in a mixture; d) NOM

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 35 of 82

Page 36: Water Ullmans

some degree in all GAC contactors. In addition to removing some of the biodegradable NOM fraction and associated DBP precursors, biological degradation activity on GAC can extend its adsorption capacity by reducing the overall organic load to the GAC. The negative aspects of biological activity on GAC include growth of zooplankton in the GAC bed, increased microorganism concentrations in the effluent, release of colonized GAC fines that are difficult to disinfect, and under anaerobic conditions, the formation of taste- and odor-causing compounds. Depending on temperature and biodegradability, component bioacclimation times are in the range of 5 – 30 d. Those GAC columns that are run for extended periods of time and reach a steady-state removal level, as shown in Figure 18 for NOM, are often termed biologically activated carbon (BAC) systems. Biological activated carbon systems rely on biodegradation for removal of pollutants, but the GAC adsorption capacity attenuates influent pulses, which allows the biomass to degrade desorbing substances. Compared to other support media such as sand, GAC also has the advantage of providing larger accessible surface area for attachment of biomass. When preozonation of water is used, the interaction with GAC must be considered, because ozonation increases biodegradability but decreases the adsorbability of NOM on GAC by formation of polar substances such as aldehydes, ketones, and ketoacids. An ozone dose in the range of 1.0 – 0.5 g O3 per gram DOC is sufficient for disinfection and NOM oxidation without significantly decreasing adsorbability [131], [134].

5.7. Design of GAC Systems The important parameters and typical values used to characterize the design and operation of GAC adsorbers are listed below [131], [134]:

Of particular importance are the filter velocity vF (also termed the superficial linear velocity or surface loading rate), which is defined as

and the empty-bed contact time EBCT, which is defined as

For a given flow rate, vF determines the filter area, while EBCT determines the length and volume of the carbon bed, both of which influence the capital costs of a GAC system. The EBCT also has an impact on operation and maintenance costs, because shorter EBCTs lead to faster breakthrough of pollutants and subsequently more frequent replacement of the GAC bed is required. The throughput parameter in bed volume BV is a normalized expression of the volume of water treated VL or the filter operation time tF and is defined as

The carbon use rate CUR is the mass of GAC required to treat a volume of water to achieve a given treatment objective and is defined by

where F is the density of the filter layer. For a given flow rate, CUR dictates the GAC demand from which the cost of GAC replacement or reactivation can be calculated. Background NOM decreases the GAC adsorption capacity for specific organic compounds and thus increases CUR. Values of CUR predicted from distilled water isotherm data were compared with the observed CUR of pilot and full-scale column runs for which breakthrough data were available [153]. Figure 19 shows the relationship between the natural water correction factor (observed CUR: distilled water CUR) and the distilled water CUR. For compounds that are strongly absorbed by the carbon bed (i.e., low CUR), the natural water correction factors are higher (10 – 100) than for substances that are adsorbed only weakly. This is due to the fact that in treating strongly adsorbed compounds, the GAC is exposed to NOM for a longer time, thus, more fouling occurs.

Bed volume VF 10 – 50 m3

Cross-sectional area AF 5 – 30 m2

Length l 1.8 – 4 mFilter velocity vF 5 – 15 m/h

Empty bed contact time EBCT 5 – 30 minOperation time tF 100 – 600 d

Throughput BV 4000 – 30 000 m3/m3

Carbon use rate CUR 5 – 50 g/m3

(5.5)

(5.6)

(5.7)

(5.8)

Figure 19. Relationship between natural water correction factor and distilled water carbon usage rate [153]

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 36 of 82

Page 37: Water Ullmans

In groundwater treatment plants, the GAC is normally placed at the end of the treatment train but before final disinfection. In surface water treatment plants, GAC can be used as a filter adsorber or placed as a stand-alone unit after filtration, termed the postfilter adsorber. In both cases, predisinfection of water with chlorine-based disinfectants should be avoided because the disinfectants can react with the GAC and thus reduce its adsorption capacity for organic compounds and form undesired chlorinated compounds [134]. Old water filters that have been retrofitted as filter adsorbers tend to have low EBCTs, 3 – 9 min, which limits their application in most cases to taste and odor improvement and chemical spill control. Postfilter adsorbers generally have longer EBCTs, 10 – 30 min. They can be used effectively to control a wider range of compounds, including micropollutants, NOM, and DBP precursors.

Prediction and Simulation of GAC Breakthrough Behavior. For site-specific design purposes, a variety of approaches have been applied to predict or simulate GAC breakthrough behavior. Parallel pilot columns that utilize the same GAC and influent water as the full-scale system are accurate and reliable predictors of breakthrough behavior [154], [155]. They also have the advantage of not requiring the use of numerical models in interpretation of the data, but they have the same operation time as the full-scale system. Another approach is to collect isotherm and adsorption kinetic data and use the data as input to fixed-bed numerical models for predicting breakthrough [131], [151], [152], [156], [157]. Some extensions of these models have been made to take into account the impact of NOM on the breakthrough behavior of micropollutants [131], [147].

Minicolumns utilize small particle and column sizes to decrease the operating time of a column to a small percentage of that required in the full-scale column. Small-scale columns are minicolumns in which similitude to full-scale GAC systems has been maintained. The relationship among particle size, column length or EBCT, and operation time of small-scale and full-scale columns is determined by dimensional analysis. Successful application of small-scale columns yields breakthrough curves that are equivalent to those of a full-scale adsorber. The method was pioneered by FRICK [131] and was developed further and applied extensively by CRITTENDEN and coworkers [158-160], leading to the rapid small-scale column test (RSSCT). The major problem is that the impact of background NOM on the adsorption capacity and diffusion kinetics of micropollutants is not always reflected accurately in small-scale columns. Rapid small-scale column tests have been used successfully to predict the breakthrough of NOM and DBP precursors [160], [161].

5.8. Design of PAC Systems Powdered activated carbon can be applied effectively to control periodic taste and odor problems and spills of synthetic organic compounds in raw source water. The PAC dosage (i.e., the mass of PAC per volume of water treated) typically ranges from 5 to 50 mg/m3 for the improvement of taste and odor. Long-term removal of specific compounds, such as pesticides and volatile organic carbons, NOM, and DBP precursors is possible but usually only at uneconomically high activated carbon dosages. If long-term high dosages of PAC are required, GAC systems are usually more economical. The advantages of PAC compared to GAC are (1) low capital costs, because PAC is added to existing contactors; (2) rapid utilization of the adsorption capacity, because PAC is 25 times smaller than GAC; and (3) flexibility in operation, because the dosage, and to some degree the point of application of PAC, can be controlled easily. Its disadvantages are that PAC cannot be reactivated and that it can be carried over into the distribution system [131], [134].

Powdered activated carbon can be applied dry or as a slurry at several points in a surface water treatment plant, as shown inFigure 20 (points A – E). The most common points of addition are at the coagulant rapid mixing tank (B) and the flocculation tank (C). The PAC is removed from the process stream during sedimentation and filtration, and effective contact times of 30 – 60 min can be achieved. Addition of PAC at the raw water intake (A) results in longer contact times, especially if presedimentation is employed. Each PAC addition point has advantages and disadvantages with respect to contact time, kinetics, and competing organic compounds. The PAC added at point A is not incorporated initially within the floc particles, as it is when added at points B and C; i.e., addition at point A allows for faster adsorption kinetics. When PAC is added at point A, the disadvantage is increased competition for active sites from organic compounds that are lated removed by coagulation. The PAC added at point D will have the shortest contact time, but incorporation of floc and competing organic compounds is decreased. When PAC is added at point E, incorporation of competing organic compounds and floc is at a minimum and the PAC is retained within the filter for longer times. This all results in an increased adsorption capacity, thus improving effectiveness. However, filter run times are shortened, and PAC breakthrough into the distribution system must be avoided because the presence of PAC in the distribution system can lead to bacterial regrowth problems. The retention of PAC particles within filters can be aided with the addition of polyelectrolytes [131], [134].

Figure 20. Flow sheet for PAC application in a conventional surface water treatment plant

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 37 of 82

Page 38: Water Ullmans

5.9. Reactivation of GAC Activated carbon has a finite adsorption capacity, and after exhaustion of this capacity, the used GAC must be replaced with fresh or reactivated GAC. Reactivation conditions are similar to those employed in manufacturing activated carbon by thermal activation, which means that part of the carbon surface may be included in the reaction. Typically, thermal reactivation of GAC occurs in four stages: (1) thermal desorption of the absorbed compounds; (2) thermal decomposition of the adsorbed compounds followed by desorption of the decomposition products; (3) carbonization of the nondesorbed products formed during thermal decomposition or chemisorbed during the adsorption step; and (4) surface reactions between the carbonaceous residuals and water vapor or oxidizing gases to form gaseous products. Each of these stages is associated with a particular temperature range and consists of several simultaneous steps. If properly designed and operated, reactivation can yield 100 % recovery of the adsorption capacity. However a 5 – 15 % loss of GAC occurs during a reactivation cycle, which includes backwashing, transportation, and reactivation; this necessitates the use of fresh makup GAC [131], [134].

[Top of Page]

6. Ion Exchange Wolfgang Höll

(see also Ion Exchangers)

In water technology, ion-exchange resins are used to eliminate cations or anions and replace them with other cations or anions. In this way the composition of the components of the water is changed, and as a consequence, a change in water properties occurs. Among the components of natural water carbonic acid and its dissociation products play a particular role because these species form the dominant buffer system of water. This buffer system is affected by almost every ion-exchange process. Beyond the mere exchange of ionic species the properties of water are therefore influenced in a way that is characteristic for each type of exchange. The fundamentals of ion exchange and aspects of its application have been presented in several books and innumerable publications [162-167]. This chapter illustrates the influence of ion exchange on the properties of the water treated, i.e., its chemical state, which may be characterized by pH value and alkalinity.

6.1. Fundamentals 6.1.1. Characterization of the Chemical State of Water Inorganic Components of Natural Water. The main inorganic components of natural water include the following:

In addition, natural water usually contains nonionic impurities (dispersed substances that cause turbidity, colorants, colloids, or organic matter), which are not discussed in this chapter.

Dissociation of Carbonic Acid and Water and the Solubility of Calcium Carbonate. The above list shows that carbonic acid is present in water in the form of four different species. If nondissociated H2CO3 is ignored because of its low concentration, the sum of the concentrations of carbonic acid species — usually designated the total inorganic carbon (TIC) — is

The concentrations of CO2, , and are linked by the dissociation of carbonic acid [dissociation constants K1 (T, I ) and K2 (T, I ), where I denotes ionic strength and T is temperature].

The ionic strength I is defined by

where z (i) is the valence of species i. Further dependencies are normally negligible. By using the ionic strength, activity coefficients of monovalent species can be calculated according to

For I < 0.1 mol/L the second term is negligible. Activity coefficients of bivalent species result from f (z = 2) = f (z = 1)2.

The dissociation constants K1 and K2 (for = 25 °C, I = 0 mol/L) can be calculated according to [168], [169]

Cations of strong electrolytes Ca2+, Mg2+, Na+, K+

Anions of strong electrolytes , , Cl–

Species of carbonic acid CO2, H2CO3, ,

Water and its dissociation products H2O, H+, OH–

Silica SiO2

(6.1)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 38 of 82

Page 39: Water Ullmans

The relative amount of negative charges of carbonic acid species is expressed by the so-called equivalent factor :

which lies in the range 0 ≤ ≤ 2 [170].

The dissociation of water molecules is characterized by the ionic product of water:

If the liquid phase is in equilibrium with solid CaCO3, the concentrations of the respective ions are linked by the solubility product:

The solubility product of CaCO3 (for = 25 °C, I = 0 mol/L) can be calculated from [171]

with T0 being 298.15 K.

m and p Alkalinities. One of the fundamental properties of each electrolyte solution is its electroneutrality. Given the species on the above list, this condition can be expressed as

Rearrangement of this equation with respect to the origin of the species from either strong or weak electrolyes yields

The algebraic sum of equivalents on the right-hand side of this equation can be determined approximately by titration with methyl orange (color change at pH 3.1 – 4.4) as indicator. Both sides of Equation (6.6), therefore, define the so-called methyl orange alkalinity (m alkalinity, total alkalinity, TAC) [170]. By means of the abbreviation

the right-hand side of Equation (6.6>) can be written as

Carbonic acid-bearing solutions are usually characterized also by a second quantity which is derived from titration with phenolphthalein as the indicator (color change at pH 8.2–9.8), the so-called phenolphthalein alkalinity (p alkalinity):

This relationship results from the definition of m alkalinity and from the condition

For natural water with pH values between 6 and 8.5 the m alkalinity approximately equals the concentration of ions and the value of the p alkalinity that of carbon dioxide. The advantage of using m and p alkalinities is that they allow characterization of the water without tedious analyses of all inorganic components [170].

Interdependence Between Carbonate Balance and Solubility of Calcium Carbonate. In natural water the solubility of calcium carbonate is almost never adjusted correctly. The water is either oversaturated or undersaturated with CaCO3. The state of saturation can be characterized by the saturation index S I:

(6.2)

(6.3)

(6.4)

(6.5)

(6.6)

(6.7)

(6.8)

(6.9)

(6.10)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 39 of 82

Page 40: Water Ullmans

For positive values of S I, CaCO3 can be precipitated from water; for negative values, it can be dissolved.

The content of the components unaffected by either the precipitation or the dissolution of CaCO3 is characterized by the

parameter m – 2 c (Ca2+), which remains constant in both cases. Subtracting 2 c (Ca2+) from both sides of Equation (6.8) and expressing c (Ca2+) on the right-hand side by Equation (6.11) finally yield the relationship

which gives the interdependence between the carbonate balance and the solubility of calcium carbonate for an arbitrary value of the parameter m – 2 c (Ca2+) [170], [172].

(6.11)

(6.12)

6.1.2. Representation of the Chemical State of Water One of the classical graphic representations of the carbonate balance and thus of the chemical state of water was introduced by TILLMANNS, who plotted the concentration of “free” carbonic acid [c (CO2)] as a function of [173]. With respect to this familiar plot, a generalized modification giving the (negative) p alkalinity as a function of the m alkalinity was chosen to illustrate the changes of the state of water in ion-exchange processes. The generalization is obtained by combining Equations (6.8) and (6.10) to yield the following relationship:

Since both and are functions of pH, Equation (6.13) represents the equation of a straight line at constant pH [170]. The set of lines is plotted in Figure 21. The state of equilibrium with solid CaCO3 corresponding to Equation (6.12) is given by the

set of curves for which m – 2 c (Ca2+) has a given constant value. Since TIC cannot be negative, all possible states of solutions of salts of carbonic acid must lie above the line with p = m. The plot allows a simple representation of most of the changes of the chemical state of water.

Any mathematical calculation of carbonate balance requires that four quantities be known (i.e., temperature, ionic strength, and m and p alkalinities). By this means, an arbitrary fifth quantity can be calculated. Graphical plots allow representation of the interdependence of only three quantities; two more must be assumed to be constant. To fulfill this condition, all calculations were made for 25 °C and a constant ionic strength of 0 mol/L. The latter assumption is a crude simplification in view of the changes in composition of water. However, ionic strength affects only the calculated pH values (through the activity coefficients), whereas the m and p alkalinities remain unaffected.

Further graphical representations of the chemical state have been developed by HALLOPEAU and by CALDWELL and LAWRENCE [174], [175]. However, these plots are not discussed here.

(6.13)

Figure 21. Interdependence of m alkalinity, p alkalinity, and pH (temperature 25 °C, ionic strength I = 0 mol/L)

6.1.3. Ion Exchangers ( Ion Exchangers, Ion Exchangers – Properties) [165]

Resin Structure and Type. An ion-exchange resin consists of a polymer matrix containing functional groups. The matrix consists of polystyrene or of acrylic polymers, both cross-linked by means of divinylbenzene (DVB). The DVB concentration normally varies between 4 and 12 %. However, for particular purposes, resins with either less than 4 % or up to 25 % DVB are available.

The ability to exchange ionic species is due to the functional groups present in the resin. Either these must be fixed within the matrix structure by a suitable functionalization step during synthesis or they are already part of the organic monomers. According to the types of functional groups and their degree of either dissociation or protonation, strongly and weakly acidic or basic ion-exchange resins can be distinguished. Functional groups of common exchange resins are summarized below:

Strongly acidic – H+

Weakly acidic –COOHStrongly basic

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 40 of 82

Page 41: Water Ullmans

Dissociation of Functional Groups. The dissociation of functional groups of cation and anion exchangers can be written as

Overbars refer to the resin phase. By applying the law of mass action, the following constants are obtained:

where the indices c and a denote cationic and anionic resins.

The pK values can be derived from these constants:

The pK values of cation- and anion-exchange resins are summarized in the following:

The range of operation of different resin types can be derived by means of pK values. If pH p , the percentage of dissociation of charge-bearing functional groups exceeds 50 % in cation exchangers for pH > pK and in anion exchangers forpH < pK. As a crude rule, the pK therefore represents the limit of applicability of an exchange resin.

6.2. Cation-Exchange Processes

–N+(CH3)3OH– (Type I)

–N+(CH3)2C2H4OH–(Type II)

Weakly basic –NR′2 · H2O

–NRH′ · H2O

–NH2H2O

(6.14a)

(6.14b)

(6.15a)

(6.15b)

(6.16)

Strongly acidic <1Weakly acidic 4 – 6Strongly basic >13Weakly basic 6 – 9

6.2.1. Strongly Acidic Resins in the Na+ Form

Application of strongly acidic resins in the sodium ion form is the classical ion-exchange process for water softening. Softening makes possible the elimination of alkaline-earth ions from water. These species may cause troublesome CaCO3, Mg(OH)2, and CaSO4 scales on heat-transfer surfaces. Furthermore, increased concentrations of alkaline-earth ions require increased consumption of detergents in washings. Softening processes have been introduced in industrial and drinking water treatment, both in municipal water treatment systems and in home softeners [176-178]. An equilibrium diagram for water softening is shown in Figure 22.

In terms of a formal chemical reaction, the process reads as follows:

Symbols in parentheses indicate the stoichiometry of the exchange. Because of the greater affinity of the resin for bivalent species, calcium and magnesium ions are removed very efficiently. Sodium chloride solutions (brine) are used to regenerate exhausted resin. In practice, the resin is converted only partly to the sodium form; the relative amount of loading is about

Figure 22. Equilibrium diagram for water softening. The following raw water composition is assumed: c (Ca2+) = 3 mmol/L, c (Mg2+) = 0.5 mmol/L, c (Na+) = 1 mmol/L, c (Cl–) = 2 mmol/L, TIC = 5.5 mmol/L R = Raw water; P = Product water

(6.17)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 41 of 82

Page 42: Water Ullmans

50 % of the exchange capacity [176].

The pure neutral exchange of alkaline-earth species for Na+ ions does not change the concentrations of the species involved in the dissociation of carbonic acid. However, the saturation index is strongly affected. The decreased calcium concentration results in a decrease in the numerical value of m – 2 c (Ca2+). Since the CaCO3 – H2CO3 equilibrium is now described by the respective curve, the product water is far away from saturation and becomes lime agressive.

6.2.2. Strongly Acidic Resins in the H+ Form

Strongly acidic resins in the H+ form are normally applied in demineralization processes prior to the elimination of anions by means of anion exchangers. Because of the exchange of practically all cations for H+ ions, salts are converted to the corresponding mineral acids:

Regeneration is carried out by using HCl or H2SO4. However, with respect to the precipitation of CaSO4, the concentration of sulfuric acid used for regeneration must be relatively low. As a consequence, the regenerant volume is greater. The amount of acid that must be applied depends on the tolerable leakage, mainly of sodium, during the service cycle.

As long as no carbon dioxide is degassed from the water, TIC remains constant. Therefore, the development of the state of water follows a straight line with a gradient equal to –1. The final state depends on the concentration of strong acids (Figure 23).

(6.18)

Figure 23. Change of chemical state of water during cation exchange with strongly or weakly acidic resins in the hydrogen form. The raw water composition in Figure 22 was assumed. R = Raw water; Pw = Product water after weakly acidic resin; Ps = Product water after strongly acidic resin

6.2.3. Weakly Acidic Resins in the Hydrogen Form If weakly acidic resins in the hydrogen form are allowed to come in contact with natural water, a so-called reaction-coupled exchange occurs: protons from carboxylic groups of the resin react with anions of weak acids, normally and . Thus, nondissociated carbonic acid is formed. To maintain electroneutrality, the resin adsorbs cations. Corresponding to their stronger affinity, alkaline-earth ions are mainly eliminated; Na+ and K+ are exchanged only if the raw alkalinity exceeds the sum of the equivalents of calcium and magnesium ions. With respect to the simultaneous elimination of alkaline-earth ions and of alkalinity, the process — called softening with dealkalization — represents a partial demineralization [176].

The process can be written as

Softening with dealkalization can be applied as the first step during demineralization to eliminate a substantial part of the totalsalinity. Furthermore, it can be combined with a subsequent softening using a strongly acidic resin in the Na+ form. In this way, cooling water or feed water for low-pressure boilers can be produced. Finally, softening with dealkalization also has been proposed to soften drinking water.

Strong acids must be used for complete conversion of the resin to the hydrogen form during regeneration. Since the functional groups show a very strong preference for protons, only slightly more than the theoretical amount of acid has to be applied. Furthermore, the acid need not be concentrated or very pure. Weakly acidic resins can also be regenerated by using weak acids (i.e., citric or carbonic acid). In the latter case, however, only partial conversion to the hydrogen form is achieved [179-182].

During the combined softening – dealkalization with weakly acidic resins, the chemical state of water changes along a line with a gradient of –1 until m = 0 and pH 4, provided no CO2 degasses. In practice, however, acrylic resins split part of the neutral salts. Therefore, the final m value is slightly negative at a pH of ca. 3.5. Figure 23 demonstrates the change of state of water for the application of strongly acidic and weakly acidic resins in the hydrogen form.

6.3. Anion-Exchange Processes

(6.19)

6.3.1. Strongly Basic Resins in the Cl– Form

Strongly basic resins in the Cl– form allow the elimination of troublesome anionic impurities such as and . They have also been proposed and applied to the elimination of chromate, arsenate, and selenate [183-185]. In contact with

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 42 of 82

Page 43: Water Ullmans

natural water, and are exchanged easily. However, an uptake of nonpreferred species such as by the resin also results:

Corresponding to the sequence of selectivity, subsequent zones with different resin loading are formed in filters. The boundaries between these regions pass the filter as traveling waves. Each nonpreferred species is replaced by stronger preferred ones and dumped out of the filter. As a consequence, the effluent undergoes variations mainly in concentration. Initially, part of the bicarbonate is eliminated from water, accompanied by a decrease of pH. However, during a later part of the filter operation, bound bicarbonate species are replaced in the resin, which leads to higher concentration and pH values in the effluent than in the feed. These variations depend strongly on the individual properties of the resin and the composition of raw water. Therefore, this cannot be generalized. Consequently, the exchange of anions for Cl– ions cannot be demonstrated graphically.

Strongly basic resins in the chloride form are frequently used to eliminate nitrate ions from drinking water. Unfortunately, standard strongly basic resins exhibit a marked preference for sulfate species. To decrease the predominant uptake of nonhazardous sulfate ions, so-called nitrate-selective resins have been developed. These are strongly basic resins of type I with ethyl or butyl instead of methyl groups [186], [187].

Apart from acting as ion exchangers the resins can also be used as adsorbents for organic matter such as humic or fulvic acids. For normal anion-exchange processes, regeneration is carried out by means of 4 – 8 % brine solution. To force the uptake of nonpreferred Cl– ions, substantial amounts of brine in excess of the stoichiometric quantity are required [176]. These scavenger resins are regenerated by using alkaline brine solution.

(6.20)

6.3.2. Strongly Basic Resins in the OH– Form

Strongly basic resins in the OH– form replace all anions with OH– ions. Normally, OH– loaded resins are applied almost exclusively during demineralization in combination with a strongly acidic cation exchanger, either in subsequent filters or in mixed beds. If the feed water for the basic resin is the effluent of a strongly acidic resin, the anions of strong acids are exchanged for OH– and water molecules are generated:

Because of the high degree of protonation of nitrogen atoms in quaternary ammonium groups even at relatively high pH, these resins are also capable of removing nondissociated weak acids such as carbonic acid or silica according to

Regeneration is carried out by means of 4 – 8 % aqueous NaOH solutions. Because of the weak affinity of the resins for OH–, the regenerant solution must be of high purity.

In a graphical representation the uptake of anions of strong acids and of H2CO3 can be subdivided into two parts: (1) During

the elimination of , , and Cl–, no change in TIC occurs. Therefore, the state changes along a line with TIC = constant until m = 0. (2) During the sorption of carbonic acid, the state changes along the ordinate toward p = 0. The real development runs as a superposition of both parts on an arbitrary line between the states of the feed and of the product water (Fig. 24, curve a). The elimination of silica cannot be represented in such a diagram.

(6.21)

(6.22)

Figure 24. Change of chemical state of water during exchange of anions for OH– (curve a) and for ions (curve b). The raw water composition in Figure 22 was assumed. R = Raw water; P = Product water; P′ = (Hypothetical) state of product water without precipitation of CaCO3

6.3.3. Strongly Basic Resins in the Form Strongly basic resins of the form exchange all anions for ions:

According to Equation (6.23), anion exchangers in the bicarbonate form may be used to eliminate nitrate from drinking water [188], [189]. However, this kind of exchange can be applied only if the concentration of calcium ions is low. Otherwise, the

(6.23)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 43 of 82

Page 44: Water Ullmans

solubility product of calcium carbonate is exceeded, yielding uncontrollable precipitation of solid CaCO3. Apart from their application as single exchangers, anion exchangers in the bicarbonate form are used in the Desal and Carix processes [190-192].

Regeneration can be carried out either in a single step by using NaHCO3 or H2CO3 (+ MgO) or in two subsequent steps by means of NaOH or aqueous NH3 and CO2 [189], [193].

During the exchange of anions for species, an increase of m at constant p is encountered. Thus, the state moves on a horizontal line. If the theoretical end point lay beyond the saturation curve, CaCO3 would be precipitated, which limits the further increase of m. The state of the product water is then represented by the intersection of a line p = constant with the saturation curve (see Fig. 24, curve b).

6.3.4. Weakly Basic Resins In neutral or alkaline media, weakly basic resins in the free-base form are only weakly protonated. Because of the corresponding lack of positive charges they cannot act as anion exchangers. As a consequnce, they can adsorb an-ions only from acidic media with pH <6. These resins are therefore able to eliminate strong acids from the effluent of a strongly acidic cation exchanger in the H+ form. The uptake of carbonic acid by the resin is, however, negligible, with the exception of acrylic weakly basic resins. The exchange, which can also be classified as an adsorption of acids, can be represented as

Regeneration requires alkaline solutions such as NaOH or aqueous NH3. As with weakly acidic resins, the regenerant solution need not be concentrated or extremely pure. Because of the pH-dependent protonation of the amino groups, regeneration requires only slightly more than the theoretical amount. Like strongly basic exchangers, weakly basic resins can be used to remove humic substances.

The development of the chemical state of water treated with weakly basic resins closely resembles that of water treated with strongly basic resins. The negative m value of the feed water is increased until m = 0. Thus, the development follows a straight line with TIC = constant. The uptake of small amounts of carbonic acid by the resin leads to a decrease in TIC according to individual resin properties.

Figure 25 demonstrates the advantage of degassing CO2 only after using the weakly basic resin. Because of the leakage of sodium in cation exchange, the m alkalinity of the effluent of a weakly basic resin becomes slightly positive. As follows from Figure 25, for this case the pH of the product water is high for low CO2 concentrations and lower at high concentrations. Because of the pH-dependent protonation of the functional group the exchange capacity of the resin is necessarily greater at a high TIC.

6.4. Combined Processes

(6.24)

Figure 25. Change of chemical state of water during treatment of an acidic solution with a weakly basic resin (no uptake of CO2). The raw water composition in Figure 22 was assumed. R = Raw water; P = Product water

6.4.1. Demineralization The simplest form of demineralization of groundwater (e.g., for boiler feed water) consists of the subsequent application of a strongly acidic resin in the H+ form and a strongly or weakly basic resin in the OH– form or free-base form, respectively. Depending on the composition and quality of the fresh water to be treated and the desired product water quality, a weakly acidic resin can also be applied in the first step followed by a strongly or weakly basic anion-exchange resins [176]. This process design also saves regeneration costs. The development of the state of the water can be compiled easily from the previous sections.

Figure 26 demonstrates changes of the state of water. Furthermore, it clearly shows the advantages of using weak electrolyte resins for the removal of alkalinity and strong acids, and of efficient degassing prior to the use of a strongly basic resin.

Figure 26. Change of chemical state of water during demineralization by combined application of weakly and strongly acidic resins, a weakly basic resin, a degassing step, and a strongly basic resin. The raw water composition in Figure 22 was assumed R = Raw water; P = Product water; I1 – I4 = Intermediate states; – – – Equilibrium with atmospheric CO2

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 44 of 82

Page 45: Water Ullmans

For the demineralization of brackish water, other concepts have been proposed, for example, the Desal process [189], [190] ( Ion Exchangers). The process is unique because of its exclusive use of weak electrolyte resins. Its key is the application

of an acrylic weakly basic resin, which in contrast to styrene-based resins can adsorb ions. In the first step, an anion exchanger converts all the salts to salts of carbonic acids, which can be split by a weakly acidic resin applied in the second step. Carbon dioxide evolved in this step is used for conversion of a second weakly basic resin from the free base to the

form. Despite its advantages, the Desal process has never been used on a pilot or technical scale.

6.4.2. Partial Demineralization Partial demineralization is achieved either by the application of a sequence of (completely regenerated) exchange resins not allowing full demineralization or by the use of incompletely regenerated resins. The simplest partial demineralization using completely regenerated exchangers involves softening with dealkalization by using a weakly acidic resin in the hydrogen form (see Section Weakly Acidic Resins in the Hydrogen Form). Another example is the combination of a weakly acidic resin in the hydrogen form and a strongly acidic resin in the Na+ form. The change of the state of the water can be derived easily from the corresponding plots.

Among the processes using incompletely regenerated resins, the Sirotherm and Carix processes ( Ion Exchangers) are the most important.

The Sirotherm process, also developed for the treatment of brackish water, uses a resin with a mixed weakly acidic and weakly basic functionality [194], [195]:

The service cycle runs at ambient temperature; regeneration, however, is carried out by means of hot water, thus making use of the temperature-dependent dissociation of water. Since the treatment of NaCl solutions does not influence carbonate balance, no change in chemical state occurs as defined in Section Characterization of the Chemical State of Water.

The Carix process consists of the combined application of a weakly acidic resin in the free-acid form and of a strongly basic resin in the form. Both resins are used as mixed beds. In the service cycle, the resin mixture replaces neutral salts with carbonic acid. The resins are regenerated as a mixture by using carbonic acid. The process is applied for partial demineralization of drinking water. The simplified formal representation of the process is [192]:

In the regeneration step the exchange capacity generated on the cation exchanger is greater than that on the anion exchanger. As a consequence, the elimination of neutral salts in the service cycle is accompanied by a softening or dealkalization by the cation exchanger.

By using ion-exchange resins completely converted to the hydrogen and forms, process development can be divided into three steps. First, the cation exchanger reduces the m alkalinity to about zero, accompanied by a corresponding increase in concentration of dissolved carbon dioxide; second, the anion exchanger replaces anions of strong acids with species to yield an increase of m alkalinity. Thus, the cation exchanger can again replace cations by protons (Fig. 27). Because of the limited efficiency of carbonic acid, only partial regeneration of both resins is achieved. As a consequence, the exchange on both resins remains incomplete, yielding considerable leakage of both anions and cations. The possible development of the chemical state of the water is plotted in Figure 27.

[Top of Page]

7. Flocculation Martin Jekel

7.1. Introduction Most surface waters contain finely dispersed substances that cannot be separated completely by simple sedimentation or rapid filtration. These dispersed particles are responsible for the turbidity of water and may consist of minerals such as silica and clays, algae cells, bacteria, viruses, organic detritus, or insoluble pollutants [196]. In groundwater, particles can be formed by a preceding precipitation (e.g., lime precipitation for hardness removal) or by oxidation of iron(II) and manganese ions. The finely dispersed particles can be aggregated to form removable flocs by flocculation techniques using suitable

(6.24)

Figure 27. Development of the chemical state of water in the Carix process. The raw water composition in Figure 22 was assumed. R = Raw water; P = Product water; P′ = Hypothetical product water; – – – = Hypothetical development with completely regenerated resins; ——— = Development under realistic conditions

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 45 of 82

Page 46: Water Ullmans

chemicals known as coagulants and flocculants.

The aggregates formed can be removed to a wide degree by subsequent separation processes such as sedimentation, flotation, or filtration. In the production of high-quality water such as drinking water, sedimentation and flotation usually are followed by rapid filtration, while direct filtration of flocs (see Chap. Filtration) is limited to raw water with low suspended solids and coagulant demand.

The removal of particulate substances is usually the primary objective of flocculation, but important cases exist in which dissolved inorganic and organic substances must be removed effectively by coagulants and flocculants via precipitation, coprecipitation, or adsorption onto the flocs. Examples are phosphates, some heavy metals, or dissolved organic substances of natural or anthropogenic origin. In view of the new standards for lower concentrations of organic substances in drinking water, the removal of DOC has become a major task in water treatment, especially in raw water colored by natural humic substances or having a high DOC [197].

7.2. Definitions Scientific and technical terms that are used frequently in flocculation are listed and explained below [198-201]:

Colloids are particles with diameters in the range of 0.001 – 1 µm. They do not settle, and they cause turbidity. Their high specific surface areas induce important interfacial forces and phenomena between the colloid particles.

Suspended particles have diameters >1 µm and are detected under a light microscope. They may be removed by settling if their density is greater than that of water.

Aggregation denotes the general process of forming larger flocs from primary particles.

Stabilization. Colloidal and suspended particles can be stabilized against aggregation. Stabilization is caused by van der Waals attractive forces, by the electrostatic surface charge of the particles, by an adsorbed layer of polymers, or by forces of hydration. Most important in aqueous systems are the negative surface charges formed by the dissociation of protons from hydroxyl functions belonging to silica groups or from acidic carboxylic functions [196], [198], [199].

Destabilization. Stable particles are rendered unstable through neutralization of their negative surface charges by adsorption of positively charged coagulants.

Coagulation refers to (1) the reduction or elimination of electrostatic repulsion forces between particles via addition of certain coagulants, and (2) in technical terms, the first phase of floc formation after chemical mixing and destabilization, but before dosing of flocculants (see below) [199].

Flocculation. (1) Flocculation is the mechanism of floc aggregation by high molar mass polymers that adsorb on the particles and form bridges between them. (2) In water treatment, flocculation is the phase of formation of large flocs by addition of a secondary chemical (after addition of the coagulant) called a flocculant or flocculant aid [199]. (3) The term can be used for the entire water treatment process (coagulation plus flocculation), as in this chapter. In English-speaking countries, the terminology “coagulation and flocculation” is also used frequently to describe this process.

Coagulants are chemicals used for destabilization in the first phase of floc formation. Most common are the salts of Al3+ and Fe3+ ions; sometimes, organic cationic polymers are used.

Flocculants designate chemicals used to improve the formation of larger flocs. Flocculants are high molar mass water-soluble polymers. Flocculant aid or coagulant aid is used as a synonym.

Precipitation is the formation of insoluble substances from dissolved matter and the chemicals added. The important coagulants Al3+ and Fe3+ salts can precipitate as hydroxides in the neutral pH region and are thus removed from water.

Coprecipitation is the inclusion of dissolved or particulate substances into a precipitate (e.g., of aluminum or iron(III) hydroxide). Particles are then aggregated by the important sweep coagulation mechanism, with the amorphous, gelatinous hydroxide flocs acting as particle collectors.

7.3. Chemicals Used in Coagulation and Flocculation 7.3.1. Inorganic Chemicals ( Flocculants – Inorganic Flocculants) The most important coagulants in water treatment are the salts of iron(III) and aluminum ions, available in various commercial forms and purities. Iron(III) chloride or aluminum sulfate (alum) is used frequently because of low cost. Mixed coagulants are also available, especially for applications in wastewater treatment. These coagulants dissolve readily in water, and the metal ions form hexaquo complexes, , which are acidic species and give up protons [202].

The aquo complexes are stable below pH 3 – 4 (Al3+) and 1 – 2 (Fe3+). After stepwise formation of hydroxo – aquo complexes with increasing pH, insoluble hydroxides are formed at higher pH (Al3+, pH > 5.5 – 6.0; Fe3+, pH > 4.5 – 5) as voluminous, amorphous flocs. The residual concentrations of dissolved metal ions are well below the drinking water standards of 0.2 mg/L for Fe and Al. The Al(OH)3 dissolves as aluminate at pH > 7.5 – 8.0, which must be avoided in water treatment [203]; Fe(OH)3 redissolves at pH > 10 – 11 to form .

The cationic partially hydrolyzed species present in a slightly acidic medium can condense to form polymeric products with a high positive charge. Various oligomeric and polymeric products are known [e.g., and ]. For more

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 46 of 82

Page 47: Water Ullmans

details on hydrolysis and condensation reactions, see [202]. Cationic aluminum polymers are available as commercial products or can be produced by partial hydrolysis and polymerization. These coagulants are usually called polyaluminum salts, but their composition and degree of hydrolysis vary widely.

Iron(III) and aluminum salts exhibit two distinctly different mechanisms in particle aggregation [204]:

1. Cationic polymeric species adsorb on negatively charged particles and neutralize them at an optimum dosage (charge and neutralization). Overdosing leads to restabilization because of charge reversal.

2. Insoluble hydroxides precipitate and enmesh the particles by the sweep coagulation mechanism. Optimum coagulation is achieved under conditions of maximum precipitation of the hydroxides. The nature and concentration of particles have no effect.

The sweep coagulation process is the dominant mechanism in most cases, whereas charge neutralization may be important in the coagulation of algae, humic colloids, or more concentrated suspensions and at a pH of <6.0 – 6.5.

The optimum dosages of Al3+ or Fe3+ may vary with water source and, because of frequent changes in raw water quality, even in the same plant within a few hours. The typical dosages given below are a preliminary guide for frequent cases:

1. Treatment plants for river water with settling: 0.05 – 0.2 mol/m3 Me3+

2. Treatment plants for lake and impounded water with direct or in-line filtration: 0.01 – 0.1 mol/m3 Me3+ 3. Treatment plants for color and DOC removal: 0.01 – 0.05 mmol Me3+ per milligram DOC; optimum pH values for Al3+:

5.5 – 6.5; for Fe3+: 4.5 – 6 [205], [206]

Laboratory tests are usually performed to find the most suitable coagulation conditions before adjusting dosages in the plant.

Other inorganic chemicals are much less common, such as “activated silica,” a polymeric flocculant, or Mg(OH)2, formed at pH >10.5 [200].

7.3.2. Organic Chemicals ( Flocculants – Natural Organic Flocculants, Flocculants – Chemical Structure) Organic substances added in flocculation are linear, water-soluble polymeres (synthetic, natural, or modified natural polymers) with mean molar masses in the range of 104 – 107 g/mol. The following monomers are included [200]:

Acrylamide (nonionic), acrylic acid (anionic), N,N-(dimethylaminopropyl)methacrylate (cationic), and ethyleneamine (cationic).

The resulting polymers can be nonionic, cationic, or anionic; the charge density can vary from low to very high, depending oncopolymer composition. The molecular diameter of the polymers in solution is 10 – 500 nm.

Most commercial polymers are copolymers, containing at least two monomer types. The monomer content of these products is limited to 0.05 – 0.1 % if they are used in drinking water treatment [207].

Highly charged cationic polymers can be applied as coagulants, instead of Al3+ and Fe3+ salts because of their effective charge neutralization mechanism at optimum dosage [208]. The dosages required can be rather high (1 – 10 g/m3).

High molar mass polymers with low to medium charge densities are flocculants, acting by the adsorption-bridging mechanism [198]. Typical dosages are in the range of 0.05 – 1.0 mg/L, and the polymer should be added at least 30 s after addition of the metal salts. Flocs formed with the aid of these polymers are larger, are more resistant to erosive and destructive forces, and settle faster.

In addition to synthetic polymers, natural or modified natural polymers such as starch, modified starch, or alginates can be used. Since natural polymers tend to be less effective, are unstable, and may support bacterial growth, they must be applied with more care.

7.4. Kinetics of Aggregation The particle transport processes necessary for particle collisions are usually slower than mixing, chemical reactions (e.g., hydrolysis), adsorption, and destabilization. The kinetic rate of aggregation is thus determined by the particle collision rate, which itself depends on three basic transport processes [198], [209].

Particle diffusion is based on Brownian movement and dominates in concentrated colloidal systems if the particle diameter is <0.5 – 1 µm. This perikinetic coagulation is probably not important in water treatment or is limited to the first short phase until microflocs appear with diameters >1 µm.

Hydrodynamic shear gradients imposed by stirring or turbulent flow are responsible for different particle velocities. This important collision mechanism (orthokinetic coagulation) can be described for the initial period by

where N and N0 are the total particle concentrations at time t or t = 0; is the collision efficiency factor; is the mean shear

gradient (s–1); is the volume concentration of solids; and t is time.

The collision efficiency factor depends on the extent of destabilization of the particles (chemical factor), while represents

(7.1)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 47 of 82

Page 48: Water Ullmans

shear conditions. The mean value is used for turbulent conditions and can be calculated [198] according to the following: For back-mix reactors

where P is the power input of stirrer; V the reactor volume; and the dynamic viscosity.

For pipes and channels

where p is the pressure loss; v the flow rate; l the length of the pipe; and the dynamic viscosity.

The orthokinetic equation is not longer valid if larger flocs are formed and destroyed again. In this case, the formula must be extended by a floc destruction term [210], describing the fact that larger flocs are formed at low values, while high values induce smaller flocs. In designing flocculation reactors, the use of two or more reactors can reduce overall reactor volume. The first reactor is operated at high values for high collision rates but small flocs, whereas the second reactor is needed to produce large flocs at low values (see Section Process Technology for Coagulation and Flocculation).

The final flocs are separated by either settling, flotation, or filtration, whereby the different floc velocities induce further collisions and floc growth. This third transport process is significant in settling and flotation tanks.

7.5. Process Technology for Coagulation and Flocculation The major process steps in coagulation and flocculation are listed in Table 27. Steps 1 and 2 are physically different processes but are quite rapid and usually occur in the same reactor, the rapid mix system. Steps 3 and 4 are performed in one, two, or more reactors, depending on plant design.

Table 27. Process steps for coagulation and flocculation systems

(7.2)

(7.3)

Step Task

Dosing of coagulants and rapid mixing

homogeneous distribution of chemicals

Destabilization adsorption of coagulants on colloids and suspended particles, precipitation of dissolved substances

Aggregation to microflocs (coagulation)

rapid formation of small flocs from destabilized particles in high shear gradients, without flocculants (polymers)

Aggregation to macroflocs (flocculation)

formation of removable large flocs, with or without a flocculant (polymer)

7.5.1. Dosing Coagulants and flocculants are generally prepared as a stock solution or delivered as a concentrated solution. Some restrictions exist on the concentration of stock solutions and their predilution before dosing. Iron(III) and aluminum salts may start to precipitate or hydrolyze at concentrations <1 % (Fe, Al) and pH values >1 – 2. Polymeric aluminum salts should be diluted only to the ratios given by the producer. All metal salts are acidic or alkaline solutions and require safety precautions. Iron(III) salts are especially corrosive.

Organic polymer solutions exhibit high viscosities at concentrations >0.2 – 0.5 %, depending on their molar mass. Dissolution sometimes requires several hours, and the stock solution may not be highly turbulent because of loss of efficiency and should be stored for only a limited time.

National and upcoming international standards exist regarding the purity of all applied chemicals to avoid secondary pollution in the finished drinking water. Cationic polymers, especially, are not approved for use in several countries because their risks are unknown.

7.5.2. Rapid Mixing The small flow of a dissolved coagulant and flocculant must be distributed rapidly in the large water flow by suitable reactors. Efficient, rapid mixing is very important for destabilization of colloids. The present state of knowledge and technology is summarized in [211]. Rapid mixing reactors can be designed as plug-flow or back-mix systems. The latter apparently have some deficiencies in terms of coagulation, because the coagulant species may not be transported evenly to all particles in a short time. Mixing studies indicate that the metal salts should be mixed completely within 0.1 – 1 s, which may be difficult to achieve in large plants [211].

Possible rapid mixers are listed in Table 28. If polymeric flocculants are dosed after the primary coagulants (usually metal salts), the retention time between dosing points should be at least 30 s or more to avoid a higher polymer demand. Polymers should not be mixed with excessive intensity to limit microfloc destruction and polymer breakage.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 48 of 82

Page 49: Water Ullmans

Table 28. Rapid mixer and destabilization units

Basic type Examples, characteristics

Pipes and channels (plug flow)

without internals, with orifices or in-line mixers

with pipe restrictions and expansionswith static mixerswith high-speed injection

Stirred vessels (back mix) detention time 10 – 120 sshear gradient >1000 s–1

special design to control back mixing and retention time distributionmixing energy can be varied

Turbulent zones before, at, or after pumpshydraulic jumpsventuri tubesweirs with overfall

7.5.3. Reactors for Floc Formation A variety of reactors for aggregation exist in water treatment, depending on historical developments, national regulations and experience, and the technical solutions of plant designers. The basic principles are as follows [199]:

1. Stirred tanks (back-mix reactors)

2. Pipes (plug-flow reactors) 3. Static systems (channels with baffles, plug flow, and some back mixing) 4. Sludge contact systems (sludge recycling or retention in flocculators) 5. One or more of these reactors in series

Stirred tanks are used widely and allow adjustment of the mean shear gradient independently of plant flow. The tanks are designed in rectangular form or as round vessels with baffles. Retention times at nominal flow are generally between 10 and 30 min (total); longer times are needed if only one reactor is used. The required retention time is lower (2 – 10 min) in direct filtration plants. The mean shear gradients adjusted by the stirrer speed are in the range of 10 – 100 s–1. In two or more sequential reactors, the value declines typically from ca. 100 s–1 to 10 – 30 s–1. If polymeric flocculants are added, the final shear gradient can be higher, ca. 25 – 50 s–1, because of more rigid flocs. Inlets and outlets must be arranged properly to avoid short-circuits. The stirrers used are mostly of the blade or grid type, and stirrer speed should be variable to optimize floc formation under various conditions of raw water and flow.

Flocculation in pipes [212] offers the possibility of using existing piping and saving space, but the efficiency of flocculation depends on flow rate. At a low flow rate, the kinetics of aggregation are slower and solids may settle, while at high flow the flocs tend to be destroyed. Polymeric flocculants are required, and the pipe design should avoid narrow bends and other high turbulence. Pipe lengths required are 20 – 50 m, corresponding to ca. 0.5 – 2-min detention time.

Static Channel Flocculators. Similar reactor characteristics are found in static channel flocculators equipped with baffles to induce back mixing and establish shear gradients by the head loss. These systems are in wide use in developing countries because of their simple design and construction [213], [214]. The channels are usually arranged one by one in flat, rectangular tanks, requiring more surface area compared to other flocculators. Easy cleaning of the channel bottom is essential. The width of the channels may be constant or increasing in the direction of flow to lower the value. Total detention times are generally about 20 – 40 min. Design parameters and examples are given in [214].

Sludge contact systems use the positive kinetic effect of high volume concentrations of suspended solids (see in Eq. 7.1) by retaining the flocs in reactors by various methods. Besides more rapid floc formation, denser and larger flocs are formed, which settle faster. Sludge contact is established only in combination with sedimentation. Basically, two techniques are applied:

1. Recycling of settled sludge to the flocculators (stirred tanks)

2. Sludge blanket systems in upflow tanks with vertical or inclined walls

A sludge blanket is established when the flocs have a settling rate equal to the upflow rate. Raw water enters the bottom after rapid mixing via a pipe system, and destabilized particles are captured by the large flocs of the blanket. Excess sludge is removed via sludge hoppers. Sludge blanket reactors need an extended period of startup and are somewhat sensitive to rapid flow increase. They have advantages for low-turbidity and algae-laden raw water. Their upflow rate is in the range of 2 – 4 m/h. A commercial system uses a pulsating raw water flow [215]. Equipped with parallel plates, a more recent version can operate at an 8 – 12-m/h upflow rate. New flocculators use a stirred upflow tank to form pellet-like flocs of high density [216].

7.5.4. Integrated Flocculation Systems

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 49 of 82

Page 50: Water Ullmans

Because of the great variety of reactors and floc separators employed only a few examples are shown here. Classic process schemes including settling are depicted in Figures 28 and 29. The total detention time is 2 – 4 h, and the systems are most suitable for treatment of river water with a variable content of mineral solids. The settler itself can be designed as a rectangular (Fig. 28) or a circular (Fig. 29) tank with a circular flocculator in the inner section [200]. Settled sludge may be recycled to the flocculators by a pump.

The introduction of parallel plates or tubes into sedimentation units can reduce retention times. Such a system is shown in Figure 30, including coagulant mixing (a), sludge recycle, and polymer mixing (b); two stirred flocculators (c); parallel plate settlers (d); and a sludge thickener (e) below the plates. The total detention time is only 30 min at full capacity [200], [217].

Direct filtration of flocculated water can be employed for low-turbidity raw water from lakes and dams. The flocs can be quite small (microflocs) to be filtered off in dual- or triple-media filters, at rates of 10 – 15 m/h. An example is shown in Figure 31 [218].

Flotation is a feasible floc removal process for low-density flocs, formed in colored, humic acid water or in algae-laden raw water. The version used mostly is dissolved air flotation (DAF, Wastewater), recycling 5 – 10 % of the clarified water, which is saturated with air under a pressure of 5 – 10 bar. The flocculated water and recycled pressurized water are mixed; gas bubbles are formed, attach to the flocs, and cause them to float [219]. This process typically requires less space than classical flocculation processes, no polymers, somewhat shorter flocculation times, and no additives for flotation.

Since the 1970s, the introduction of preozonation before coagulant addition has proved successful to decrease turbidity and increase the removal of algae, to lower coagulant demand, or to extend filter run times in direct filtration [220]. Ozone is applied typically in dosages of 0.5 – 2 mg/L, depending on raw water conditions. Other oxidants such as chlorine or chlorine dioxide are less efficient or are avoided because of the formation of harmful byproducts.

Figure 28. Rectangular settlers a) Rapid mixing; b) Flocculator; c) Sedimentation

Figure 29. Circular system

Figure 30. Process scheme of treatment plant with multiple stirring, parallel plate settler, and thickener a) Coagulant mixing; b) Polymer mixing; c) Stirred flocculator; d) Parallel plate settler; e) Sludge thickener

Figure 31. Direct filtration plant with three-layer filter a) Rapid mixing; b) Flocculator; c) Three-layer filter

7.5.5. Operational Aspects The selection of suitable coagulants and flocculants and the continuous task of adjusting their dosages to optimum values are based mainly on experience, rule of thumb, and extensive laboratory and pilot-plant testing. The great number of parameters that influence flocculation and the lack of quantitative models require adapted testing procedures. The jar-test is the most widely used laboratory test, suitable for conventional plants and chemical selection [198], [199], [221]. Other tests include filtration or batch flotation. Flow-through pilot plants are usually designed according to full-scale systems and allow more experiments, especially regarding hydrodynamic parameters. However, problems occur in scaling-up pilot systems [222] because of different flow patterns. Large-scale plants are therefore often equipped with variable-speed stirrers or changeable reactor geometry to adjust the reactors to flow variation. On-line control and dosage adjustment have been implemented sucessfully in some plants in the 1980s, based on novel techniques and sensor systems [223], but their application require careful evaluation.

[Top of Page]

8. Filtration

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 50 of 82

Page 51: Water Ullmans

Rolf Gimbel

8.1. Introduction In this chapter, only the elimination of turbidity-causing solids (colloidal or finely suspended particles) by filtration is described. Consequently, this is clearly distinguished from filtration through layers of activated carbon (see Chap. Adsorption Processes in Water Treatment) or ion exchangers (see Chap. Ion Exchange), which is used primarily for the removal of fully dissolved water constituents.

Surface Filtration. The removal of turbidity from water can be carried out by using surface filtration (either cake filtration or cross-flow filtration; see Fig. 32 A and B). In surface filtration, the dimensions of the particulates are greater than the pore size of the filter medium (or of the filter cake possibly already present on it). Therefore, particles are retained because of sieve effects, while water passes through the filter medium under a pressure gradient. Surface filtration is used in many different areas of industry for solid – liquid separation ( Filtration – Washing of Filter Cakes, Filtration – Optimal Cycle Time). The liquids treated range from extremely dilute solutions to highly concentrated suspensions or sludge (dewatering). Therefore, a large number of different processes and equipment are available, which operate according to the basic principle of surface filtration. These also include membrane filtration processes (see Chap. Membrane Separation Processes in Water Treatment).

Deep-Bed Filtration. In the treatment of wastewater and drinking water and the production of cooling and process water, by far the predominant amount of water is subjected to deep-bed or depth filtration to eliminate turbidity (see Fig. 32 C) (

Filtration – Deliquoring of Filter Cakes). In contrast to surface filtration, particles are retained in the interior (i.e., trapped in the depths) of a porous filter layer, which usually consists of a bed of granular material. To maintain the depth effect of filters of this type, the particulates to be separated must be considerably smaller than the pore size of the filter layer. Otherwise, the top part of this layer changes to cake filtration after a very short filtration time, resulting in an excessive pressure drop in the uppermost filter layer.

Deep-bed filtration is used especially for the purification of liquids having very low solid concentrations or when the filtrate is required in larger quantities and with a high purity (i.e., practically free of turbidity). For this reason, deep-bed filtration is applied mainly in the aforementioned areas of water technology. A volume concentration of solids of <0.05 % can serve as a rough upper limit for the economically meaningful use of depth filters.

To achieve separation, particles must be transported to the surface of the filter material by various transport mechanisms in the interior of the depth filter. At the surface, contact between filter material and particles occurs. In a second step, sufficiently stable adhesion of particles to the filter material must take place. The entire separation process is very complex and can varygreatly depending on application. This variability is due to the complicated flow conditions in the filter layer and to the large range of properties of normally used filter materials. Moreover, the turbidity-causing solids to be separated are diverse and can range from solid mineral particles to particulate biomasses, particle agglomerates, and hydroxide flocs (see Section Inorganic Chemicals ) of very low shear strength. Finally, dissolved inorganic and organic water constituents (including polymeric filter aids) greatly influence particle separation. In addition, biological processes can occur.

The design and mode of operation of depth filters are as diverse as the factors influencing deep-bed filtration. Filters can be divided into different groups according to the typical filtration rate in each case (water volume per unit time and total filter cross-sectional area). These groups are discussed in detail below. Today, rapid filtration plays a dominant role in water filtration. With rapid filtration as an example, the separation mechanisms of depth filters, which can also be influenced by appropriate pretreatment processes (flocculation, see Chap. Flocculation), and the possibilities of describing total filter behavior (as a basis for optimal filter design) are discussed.

8.2. Slow Sand Filters Slow sand filters are operated at relatively low filtration rates of about 0.1 m/h. This type of filtration played a dominant role in the treatment of surface waters for the production of drinking water in the beginning of the 20th century and is of great importance until now. Here, purification processes are very similar to those in natural ground filtration. Apart from mechanical sieve effects and physicochemical deep-bed filtration mechanisms (transport and adhesion processes), microbiological processes are especially important [224-226].

Slow sand filters are usually employed in the form of open basins, which can have a filtration area up to 10 000 m2. The filters are provided with a suitable drainage system or infiltrate the subsoil directly to artificially recharge groundwater (see Fig. 33). The active filter layer consists of a bed of sand with a depth of 0.5 – 1.5 m and a grain size that can vary from 0.2 to 1 mm. The head of water can reach ca. 1.5 m. A new filter develops its full efficiency only after a so-called “Schmutzdecke” (dirt blanket) has been formed in the uppermost filter layer. This is formed from the retained water constituents by a very complicated combination of diverse physical separation mechanisms and biochemical processes. The “Schmutzdecke” can reach a height of a few centimeters. The purification processes involved in slow sand filtration proceed mainly in this layer.

Figure 32. Schematic representation of various types of turbidity filtration A) Cake filtration; B) Cross-flow filtration; C) Deep-bed filtration a) Suspension flux; b) Filter medium or filter layer; c) Filtrate flux; d) Filter cake

Figure 33. Schematic representation of slow sand filtration a) Biological filter layer (Schmutzdecke); b) Fine sand, height 0.8 – 1.5 m; c) Supporting layers, height 0.2 – 0.4 m; d)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 51 of 82

Page 52: Water Ullmans

Excellent water purification can be achieved with slow sand filters. Not only the finest inorganic particles but also particulate organic substances, especially protozoa, bacteria, and viruses, are retained. In addition, dissolved degradable organic substances can be mineralized and nitrogen compounds oxidized by aerobic biological processes. Moreover, smaller amounts of dissolved interfering substances (e.g., heavy metals) can be retained in the “Schmutzdecke”.

A ripening period of several days is required to form the Schmutzdecke. The length of a filter run (the time between two cleaning steps) is usually some months. At the end of a filter run, the pressure drop in the topmost filter layer is increased to such an extent that the required water throughput can no longer be maintained. In general, the necessary filter regeneration is then carried out by peeling off the uppermost 3 – 5 cm of the filter layer.

The disadvantages of the slow sand filter are a large area requirement and laborious filter regeneration. For this reason, new slow sand filtration plants are rarely built in industrialized countries. In many existing plants, attempts are being made to lengthen the filter run by using appropriate pretreatment processes for turbidity removal (flocculation, sedimentation, rapid filtration).

8.3. Polishing Filters Polishing filters, which include precoat, disk, and cartridge filters, are characterized by the relatively small radii of the filter pores (micrometer range) and the shallow depth of the active filter layer (millimeter to centimeter range). In filters of this type, apart from physicochemical transport and adhesion processes, sieve effects on the surface of the filter layer can also be significant, depending on the ratio of the size of suspended particles to pore size. Thus, these filters cannot be regarded as pure depth filters in many cases. In contrast to slow sand filtration, however, biological processes are negligible in polishing filtration. With pressure differences up to several bar, polishing filters can be operated at average filtration rates of about 1 m/h. Depending on the specific application, larger deviations from this average can occur [227].

Precoat Filters. In precoat filters, turbidity removal occurs in a layer of a fine granular or fibrous filter aid having a characteristic size of a few micrometers. This layer has previously been filtered onto a support. The aids employed are mainly aluminum silicate (e.g., perlite) and diatomaceous earths. Other filter aids are also used in some cases. For example, powdered activated carbon or a finely dispersed ion-exchange resin gives not only highly efficient separation of colloidal sustances, but also retention of dissolved organic material or ions [228].

In precoat filtration, depending on the conditions, sieve effects on the surface of the precoat layer can dominate after a relatively short filter run, resulting in clogging of the surface. To prevent this and guarantee sufficient permeability of the resulting filter cake, filter aids are often added continuously to the suspension to be treated. In addition, the filter aids of the precoat layer can be conditioned with suitable polymers in such a way that they exhibit especially favorable properties for turbidity removal. The same polymers used in flocculation and rapid filtration are applied.

Precoat filters are usually available as closed apparatus in which the supports have a space-saving flat or tubular arrangement as far as possible. After a filter run that lasts from hours to days depending on the application, the precoat layer is generally discharged by means of a pressure surge. After renewed precoating, which requires a few minutes, the filter is ready for use again.

Disk and Cartridge Filters. In contrast to precoat filtration, the active filter layer in leaf and cartridge filters need not be built up by a filter aid but is available in a preprepared form. Apart from this characteristic difference, the modes of action and operation of different types of polishing filters are similar. The preprepared filter layers can be made of fabric, fleece, felt, or wound string especially in the case of cartridge filters. Finely porous sintered materials such as fibrous or powdery metals, ceramic materials, and plastics are also used. In many cases, disk and cartridge filters require not only mechanical but also chemical cleaning. Alternatively, filter materials are used only once [229], [230].

Filter Performance. In general, excellent filtrate quality can be achieved with polishing filters. They are used especially in thetreatment of boiler feed water and ultrapure water, such as that required in the electronic, chemical, and pharmaceutical industry. In addition, these filters are often employed in smaller drinking water treatment plants, in emergency water supply facilities, and in the treatment of water for swimming pools. Polishing filters are used frequently in the food industry for clarification of beverages. In special cases, filters of this type are also used in wastewater treatment and in the purification of various solutions (e.g., in surface technology).

8.4. Rapid Filters The term rapid filter is derived from the relatively high filtration rate through the filter, which is on average ca. 10 m/h. Filters of this type consist of a relatively coarse-grained filter material (millimeter range) having a filter bed height of 1 – 2 m. The deposition of particles within the filter layer — which generally acts as a real depth filter — is based primarily on physicochemical transport and adhesion mechanisms, which can also be influenced by biological processes.

In industrialized countries, rapid filtration has the greatest practical importance of all filtration processes used in the treatment of drinking water and industrial water or in advanced wastewater treatment. This importance is attributed not only to the low space requirement as a result of high filter rate, but also to the simple design and mode of operation of rapid filters and to the largely automated filter regeneration in which filter material is backwashed within the filter tank.

Drainage system; e) Outflow controller

8.4.1. Design The active layer of rapid filters consists of a bed of granular materials having an average diameter between 0.5 and 4 mm and a bed height of 0.5 – 2.5 m. If a fine-grained filter material is used, low bed heights are sufficient for good removal

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 52 of 82

Page 53: Water Ullmans

efficiency. However, the disadvantage is that the capacity for the particulate load on the filter is exhausted quickly, which results in a short filter running time. Quartz sand, anthracite, filter coke, activated carbon, or pumice serve as filter materials [228].

The single-medium design of a classic rapid filter is represented schematically in Figure 34. It consists of a distributing device for raw water (a), an active filter bed (b), which is to be regarded as homogeneous; and a filter bottom (d), which supports the filter material and allows the evenly distributed passage of filtered water into the clean-water chamber (e). The filter bottom made of steel or reinforced concrete either is equipped with filter nozzles provided with fine slits or is made of special porous components. To guarantee perfect functioning of the elements of the filter bottom, especially if fine filter materials are used, one or several supporting layers (c) are included between the active filter bed and the filter bottom. Each supporting layer has a grain size ca. two to four times coarser than the material above it.

As a rule, the direction of flow in rapid filters is from the top downward. The filter material should have a particle-size distribution as uniform as possible. Otherwise, fine material becomes concentrated in the top filter layer during filter backwashing. This can result in a change to surface filtration, which leads to an excessive pressure drop. To prevent this effect and distribute particle deposition over the greatest possible bed depth, dual-media filters are used frequently (see Fig. 35). In this type, the upper filter layer consists of a relatively coarse material of low density (filter coke in Fig. 35), while the lower layer has a fine grain size and high density. If the grain sizes and densities of the filter materials in the two layers are well suited to each other, this structure remains even after filter backwashing. Because of its large pore size, the upper layer has a high load capacity for particulates, whereas particle retention is not necessarily complete. In comparison, the behavior of the lower layer is the opposite. As a result, especially in the case of high-turbidity loads, longer filter running times can be attained with dual-media filters than with single-medium filters. Three-media filters have also been used successfully. Typical examples of the design of multimedia filters are given in Table 29.

Table 29. Exemplary design of multimedia filters [228]

Figure 35 shows the design of closed rapid filters ( pressure filters). Filters of this type are usually made in the form of cylindrical steel tanks. Filter surfaces up to 50 m2 are achieved in vertical tanks. Horizontal tanks provide much larger filter surfaces but usually have relatively low bed heights. The usual pressure drops vary between 0.2 and 2 bar at filtration rates of about 5 – 30 m/h.

Open rapid filters (gravity filters) are often used in large plants. The pressure drop occurring in the filter bed must be compensated by a head of water above the bed. The height varies between 0.3 and 3 m of water column (corresponding to a pressure drop of 0.03 – 0.3 bar). Usual filter flow rates are in the range of 3 – 15 m/h. Open filters are usually made in the form of rectangular reinforced concrete tanks. Filter surfaces up to 150 m2 have been achieved. Apart from the classic type of rapid filter shown here, several special designs exist with upward or horizontal flow through the filter layer. Continuously operated rapid filters are also being used increasingly. The loaded filter material is withdrawn continuously from the bottom and returned at the top after a washing stage [231], [232].

Figure 34. Basic design of a rapid filter a) Raw water distribution; b) Active filter layer (filter bed); c) Supporting layer; d) Filter bottom; e) Clean water chamber

Material Bed depth, m Bed weight, kg/m3 Grain size, mm

Dual-media filters Filter coke, pumice, etc. 0.5 – 1.2 500 – 700 1.7 – 2.5Sand 0.6 – 1.5 1000 0.8 – 1.2Three-media filters Activated carbon 0.3 – 0.6 250 – 350 3.0 – 5.0Filter coke, etc. 0.6 – 1.2 500 – 750 1.5 – 2.5Sand 0.5 – 0.8 1000 0.6 – 0.8

Figure 35. Exemplary design of a closed dual-media filter a) Inlet; b) Outlet; c) Filter bottom; d) Sand; e) Filter coke; f ) Water distribution

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 53 of 82

Page 54: Water Ullmans

8.4.2. Areas of Application and Modes of Operation In water and wastewater treatment, the different applications of rapid filters can be divided into the following main groups [228].

1. Separation of Single Particles. Applications of this type can be encountered, e.g., in the removal of turbidity from surface water that is to be used as drinking water and industrial water. The particles are inorganic substances (e.g., clay minerals) and organic particles such as plant cells, algae, or bacteria. Further applications are the separation of CaCO3 particles after a water-softening step and the treatment of boiler feed water and swimming pool water. In all these cases, biological processes are usually of secondary importance.

2. Separation of Particle Agglomerates and Hydroxide Flocs. If turbidity-causing solids are difficult to filter out as single particles, agglomeration is usually carried out before rapid filtration, or the particles are embedded in iron or aluminum hydroxide flocs by means of a flocculation process. Dissolved substances can also be bound in the flocs (see Section Introduction). Applications of this type are encountered frequently in the production of drinking water and industrial water from surface waters and in advanced wastewater treatment. In the latter, the hydroxide flocs to be separated contain not only turbidity but also colloidal and dissolved phosphorus compounds. Biological processes in the filter can greatly influence turbidity removal, especially in advanced wastewater treatment.

3. Removal of Iron, Manganese, and Ammonium. In the treatment of groundwater to produce drinking water and industrial water, the rapid filter usually has the task of eliminating divalent iron and manganese ions. These ions must be converted to a separable form by chemical or biochemical oxidation before or during actual filtration. In this widely encountered application of rapid filters, chemical – catalytic as well as biological processes are important. Ammonium can also be eliminated by biochemical oxidation.

In many cases, single-medium filters are sufficient for relatively low levels of turbidity in raw water. For higher-turbidity loads, multimedia filters can be of advantage. The importance of multimedia filters has increased since the 1970s because the direct filtration process is often used for economic reasons. In direct filtration, water (without preliminary separation of flocs by sedimentation or flotation) is passed directly through rapid filters after a flocculation stage.

Especially in direct filtration, pretreatment of water is very essential. Apart from the type and amount of flocculant used, the conditions of addition (residence time up to the filter, energy input in the flocculation stage, etc.) are also important. Moreover, the flocculation or filter aids added, which are mostly synthetic, water-soluble polymers, can clearly improve the properties of flocs with regard to filtration. Oxidative pretreatment of raw water (e.g., with ozone) can have a similar effect.

In normal depth filtration, the space between the grains of filter material is filled completely with particulates and water. In dry filtration, however, air is passed through the filter bed together with water. This special process can be used in water treatment when the aerobic degradation or oxidation processes proceeding in the filter bed are so intense that a continued supply of oxygen via air is required. This applies, for example, to the treatment of water containing high concentrations of ammonium or iron. Dry filtration is usually operated as a preliminary stage in normal rapid filtration because turbidity retention in the dry filter tends to be unstable [228].

Operation. In the operation of rapid filters, it is frequently preferred to keep the filtration rate constant during a run. This requires suitable control because the resistance to flow in the filter bed increases with increased filter running time due to turbid deposits. A constant filtration rate can be achieved by throttling the filter outflow at the beginning of the run. Throttling is then reduced, corresponding to the increased pressure drop due to loading of the filter bed. In open filters, especially; the same result can be achieved by an appropriately increasing head of water. In some cases, operation with variable filtration rates is preferred, the rate decreasing to about half of the starting value during a filter run. Depending on the raw water properties, the average filtrate quality can be somewhat better than that obtained with a constant filtration rate in the same filter unit. This results from the fact that the lower the flow rates, the lower is the probability of turbidity deposits being detached and washed out due to the shear forces in the filter.

The running time of rapid filters is normally between 10 and 150 h. In this time, both filtrate quality (expressed as volume concentration of turbidity c) and pressure drop over the filter bed ( pV) change, as shown qualitatively in Figure 36. In the continuous monitoring of filtration plants, a simple turbidimetry is generally used in practice to determine the turbidity level. As shown in Figure 36 three characteristic phases of a filter run can be distinguished. The level of turbidity in the filter outflow decreases up to time t1. This is the period of ripening, which can be very pronounced, especially when chemical – catalytic or biological processes are important for filtering action. The level of turbidity in the filtrate is approximately constant between t1 and t2. The breakthrough phase of the filter starts at t2. The filter run can continue until the maximal permissible concentration of turbidity-causing particles is reached (e.g., at t3); then the filter must be backwashed. Furthermore, Figure 36 shows that the pressure drop over the filter bed increases with filter running time and reaches its maximum permissible value at t4. One of the important optimization tasks in rapid filtration consists of making t3 and t4, the two times that limit the filter run, coincide as much as possible. For safety reasons (i.e., to make sure that the turbidity in the filtrate never becomes too high), the t3 value chosen in practice is often ca. 10 – 20 % greater than t4, and attainment of the maximum permissible pressure drop is used as the primary criterion for terminating a filter run.

Figure 36. Change in pressure drop and in turbidity with time in the filtrate of a rapid filter at a constant filtration rate

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 54 of 82

Page 55: Water Ullmans

Regeneration. At the end of its running time, a classic rapid filter must be regenerated. This is usually carried out in a backwashing process in which the upward flow through the filter layer is so strong that the turbid deposits on the filter grains are detached and discharged. In single-medium filters, backwashing is often carried out with water only. The flow rate must be so high that the filter bed is converted from a fixed bed to a fluidized bed. To achieve more intensive purification of the filter grains, air can be used in addition as backwash medium. In the first phase, the filter layer with a slight water head is whirled up only with air, causing a reduction of possibly hardened filter grains. Subsequently, backwashing is often conducted with air and water, followed by water alone.

Multimedia layers are treated analogously. In most cases, however, simultaneous air – water backwashing is omitted because the danger of losing filter material with the wash water is too great. In multimedia filtration, the last phase involves a reclassifying process, which causes the different filter media to separate so that the original layer setup of the filter is reproduced. In this process relatively high wash water rates are applied; expansion of the filter bed can reach ca. 40 % in the process.

A filter backwash usually requires 10 – 20 min. Depending on grain size and density, the wash air and wash water rates vary from ca. 30 to 100 m/h with a filter bed expansion of 10 – 50 %. In a well-designed filter wash program, the water consumption based on the filter area should be not greater than 2 – 5 m3/m2 per wash. If the removal of turbidity in the filter depends greatly on biological or chemical – catalytic processes, excessive filter backwashing can be a disadvantage because a relatively long period of filter ripening is required.

8.4.3. Filter Performance To design and operate a rapid filter optimally, both the volume concentration c of a certain type of particle and the filter load

, which are functions of filter layer depth z and filter running time t, should be predeterminable under given feed conditions. The load determines the pressure distribution p in the filter layer and, finally, the total pressure drop. To describe c, , and p as a function of z and t, a mass balance, a kinetic equation for the mass transfer from the flowing phase to the stationary filter material, and a relation to describe the pressure distribution in the filter are required for a differential filter element [228], [233], [234].

The simplified mass balance is as follows:

where vf is the filtration rate; c the volume concentration of particles; the filter load (volume of particles removed, based on filter layer volume); z a space coordinate (filter layer depth); and t the filter running time.

For the kinetic equation, the relation proposed by IWASAKI [235] is generally used. It corresponds formally to first-order reaction kinetics:

The filtration coefficient is a measure of filter efficiency. It corresponds to the probability of a certain particle being retained in the filter layer. The filtration coefficient depends on the properties of the filter material (e.g., grain size) and of the particles (e.g., particle size, density, and surface charge) and on filter operating conditions (e.g., filtration rate). However, also changes with increasing filter running time because of the increase in filter load with time. For this reason, is usually formulated as follows [236]:

0 represents the filtration coefficient in the initial phase of filtration (i.e., as long as the influence of the filter load is still

negligible); f is a correction function, which essentially describes the effect of the filter load on particle deposition. In the idealized model case, f decreases linearly with filter load, i.e., f = 1 – / s ( s: saturation or maximum load). Then, the following analytical solution for the distribution of the particle concentration in the filter layer can be obtained from the two partial differential equations for the mass balance and for the kinetic equation

where c0 is the volume concentration of particles in the filter inflow. In case of a more complicated dependence f = f ( ), numerical solutions for describing the distribution of the concentration and load must be applied. The same holds true in describing the pressure distribution in the filter layer, which is usually based on the Carman – Kozeny equation [228], [233], [234] ( Filtration – Filtration Models).

For the initial phase of filtration (t 0), the above equation results in an exponential decrease in the volume concentration

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 55 of 82

Page 56: Water Ullmans

of particulates with bed depth, according to the following:

where 0 is determined by various forces that act within the filter layer on a particle to be deposited. These forces must not only bring a particle into contact with the surface of the filter material (transport mechanisms), but also make possible sufficiently stable particle adhesion (adhesion mechanisms) on contact. Particle separation is thus influenced by many parameters, including the filtration rate and the size, shape, and surface charge of the filter material. The properties of water such as density and viscosity, the pH value, and the other dissolved water constituents are also important. In addition, the properties of the turbidity-causing solids, such as density, particle size and shape, surface charge, and shear strength, sometimes play a dominant role. The important classic separation mechanisms in deep-bed filtration of liquids are shown in Figure 37. In the case of interception (Fig. 37 A), a particle is assumed to move through the filter layer with its center along a stream line. It can come in contact with the surface of a grain as a result of the constriction of stream lines due to neighboring filter grains. In the case of sedimentation (Fig. 37 B), the particle can leave its original stream line and come in contact with the filter material under the effect of gravity. If the turbidity-causing particle has a size <1 µm, a particle – grain contact can also occur due to particle diffusion (Fig. 37 C).

Despite the extremely diverse influencing factors, the success of particle deposition in rapid filters usually increases with larger particle size, higher particle density, lower filtration rate, and smaller filter grain size. However, a decreasing filter grain size results in an increasing pressure drop. In addition, excessively high electrostatic respulsion forces as a result of surface charges of the same sign on turbidity-causing particles and filter materials must not occur. These repulsion forces can prevent particle deposition. In this respect, appropriate water pretreatment can be of considerable significance.

Under normal conditions of water treatment, a range of ca. 1 – 10 m–1 can be given as a rough guide for 0. The maximum load s attainable is usually < 2 % [228], [236].

[Top of Page]

9. Membrane Separation Processes in Water Treatment Robert Rautenbach

(see also, Membranes and Membrane Separation Processes)

Since the 1980s, membrane processes have been used increasingly in separation technology. The main emphasis was put on the treatment of aqueous solutions. In many cases, such as the desalination of brackish water, membrane processes have set the standard against which all other alternatives must be compared with regard to operating and investment costs.

In general, membrane processes are characterized by the following properties:

1. They can operate at ambient temperature.

2. They separate on a purely physical basis so that the components are not altered and can, in principle, be reused. 3. They have a modular setup so that they can be adapted to any separation capacity relatively easily and later

expansion is possible (in contrast to biological processes).

With regard to the production and treatment of drinking water and wastewater treatment, the microfiltration – ultrafiltration and nanofiltration – reverse osmosis are of special significance ( Membranes and Membrane Separation Processes – Ultrafiltration, Microfiltration, and Reverse Osmosis). This chapter is restricted to these processes.

In these processes, the component that passes preferentially through the membrane is the solvent (i.e., water in the case of aqueous solutions). In all the aforementioned processes, the driving force for this transport results from a transmembrane pressure difference. However, the magnitude of pressure difference and, especially, the type of membrane used differ in these processes. Whereas pore membranes are used in microfiltration and ultrafiltration (UF), nonporous solution-diffusion membranes are employed in nanofiltration (NF) and reverse osmosis (RO).

9.1. Principles

Figure 37. Schematic representation of classic particle deposition mechanisms in deep-bed filtration of liquids A) Interception; B) Sedimentation; C) Particle diffusiond = diameter of filter grain; x = diameter of turbidity-causing particle

9.1.1. Flow, Selectivity, Driving Forces Microfiltration and Ultrafiltration. In pore membranes, selectivity results basically from the pore-size distribution of the membrane. Characterization of the membrane is carried out by permeation experiments with aqueous solutions of macromolecular substances of varying molar masses (Fig. 38).

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 56 of 82

Page 57: Water Ullmans

Like the flow through a particle bed, the flow of solvent through pore membranes (i.e., the performance of the membrane) is a viscous flow proportional to the transmembrane pressure difference.

where ″ is the mass flow rate per unit area (flux), w is the density of water, and p is the pressure difference across the membrane. The constant A is membrane and substance specific (influence of solvent viscosity) and is determined experimentally. In most applications of micro- and ultrafiltration, a surface layer forms on the membrane that consists of retained components. This formation can rarely be prevented and is often even desirable. This layer is more important than the real membrane with regard to the separation characteristic and performance (permeate flux). Therefore, in the case of micro- and ultrafiltration, experiments under operating conditions are mandatory — processes cannot be designed based on performance data of the membrane alone.

Since the thickness and structure of the surface layer are influenced by conditions of flow of the feed solution along the membrane, the hydrodynamics in the membrane module is of great importance. In many cases (e.g., the separation of protein from whey), the permeate flux can in practice be influenced only by hydrodynamics and is independent of the transmembrane pressure difference (Fig. 39, see also Section Mass Transport Resistance in Front of the Membrane).

The measured values shown in Figure 39 are steady-state values, i.e., values obtained after a certain operating time, corresponding to Figure 40, curve a.

When a reversible surface layer is formed, an equilibrium exists between deposition of the retained components on the upper side of the surface layer as a result of permeate flux through the membrane and removal of surface layer components due to shear stresses caused by tangential flow. The use of ultrafiltration instead of sedimentation for the separation of biomass after a biological treatment stage represents an example of this behavior. In this case, a constant permeate flux of ca. 100 L m–2 h–1 has been observed over several weeks with tubular modules at tube flow rates of ca. 4 m/s [237]. Formation of a reversible surface layer is often encountered in ultrafiltration (and microfiltration), but the behavior with time shown in curve b (formation of an irreversible surface layer) is encountered just as often. When an irreversible surface layer is formed, the operating time after which the unit should be cleaned is then determined by the economics of the process.

Reverse Osmosis. In nonporous membranes for reverse osmosis and nanofiltration, transport generally is determined by the mass transport resistance of the membrane and not by a surface layer. However, modeling mass transport in the membrane is more difficult than in ultrafiltration, especially in the case of organic – aqueous solvent mixtures and, in nanofiltration, in the case of a multicomponent salt solution with monovalent and multivalent anions.

If the membrane is regarded as a continuum in which the permeating components are first sorbed, then transported by diffusion, and finally desorbed, equations can be derived for any permeating component “i ”, e.g., i = W (water), i = S (salt or dissolved organic substance) by the integration of the generalized Fick law over the membrane thickness.

where is the molar flux, Dio the thermodynamic diffusion coefficient, the chemical potential, and z the thickness of the membrane. Depending on the degree of simplifying assumptions, these equations vary with regard to complexity and

Figure 38. Separation characteristics of ultrafiltration membranes PEG = Poly(ethylene glycol)

(9.1)

Figure 39. Influences on the performance of ultrafiltration A) Influence of pressure difference (operating conditions: wp = 0.8 %, = 25 °C, n = 500 min–1; membranes: = UF-00

Kalle; = UF-0 Kalle; = CA-19000 Kalle); B) Influence of flow conditions (operating conditions: = 25 °C, p = 4 bar a) n = 550 min–1; b) n = 300 min–1) n = impeller speed

Figure 40. Behavior with time of surface layer-controlled flux a) Reversible formation of surface layer; b) Irreversible formation of surface layer

(9.2)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 57 of 82

Page 58: Water Ullmans

descriptive accuracy. Table 30 shows three such equation systems for reverse osmosis, as well as cases in which they describe flux and selectivity with sufficient accuracy.

Table 30. Models for describing mass transport in reverse osmosis membranes

The following assumptions are common to all three relations:

1. Equilibrium of the chemical potentials i at the phase boundaries on both sides of the membrane: i.e., , (where the index F denotes feed, M denotes membrane, P permeate, FM feedside of the membrane, and

PM permeate side of the membrane). 2. No coupling of the permeate flows of different components: i.e., only the gradient of the chemical potential of the

component i is regarded as the driving force for transport of component i of a mixture through the membrane

The question which equation can or must be used to describe local mass transport (i.e., the mass transport at a certain point of the membrane) for a certain application can only be solved experimentally. For instance, in the reverse osmosis of dilute sodium chloride solutions, a constant rejection R can be assumed — i.e., a constant ratio of the (local) salt concentration in the permeate to the local feedside concentration. In contrast, this ratio (or the rejection) is highly concentration dependent in the case of an aqueous butanol solution.

The problem can be explained by the example shown in Figure 41. In the case of a phenol – water mixture, significant differences between experiments and calculations are observed even at very low phenol concentrations when the calculation is based on model 2. The three-parameter model 3 must be used here.

Nanofiltration. In principle, nanofiltration membranes are nonporous solution – diffusion membranes like reverse osmosis membranes [239]. However, they have a low rejection for low molar mass substances. Nanofiltration is characterized by two extremely interesting properties:

1. The ability to fractionate organic components of varying molar mass and/or structure in aqueous solutions (e.g., mixtures of mono- and polyhydric alcohols)

2. The possibility of utilizing the Donnan effect ( Membranes and Membrane Separation Processes; Ion Exchangers – Principles ) whereby through the addition of one ionic species (e.g., ions), the rejection capacity for another ionic species (e.g., ) is drastically decreased.

This selectivity of nanofiltration for organic components in the molar mass range 100 g/mol < M < 300 g/mol is shown in Figure 42. For example, the rejection capacity of the membrane “Desal 5” is only ca. 30 % for ethylene glycol, but more than 80 % for glucose. These results were obtained from experiments with binary solutions of the same osmolarity (i.e., the same driving force) [238].

Model Transport relation Application

1 production of ultrapure water

from drinking water

2 production of drinking water

from brackish water

3 concentration of aqueous – organic solutions

wastewater treatment

Figure 41. Decrease in flux in reverse osmosis caused by traces of organic compounds (phenol) [238] (operating conditions: phenol – water mixture, p = 30 bar; = 25 °C) a) Measurements performed with membrane SU 700; b) Measurements performed with membrane FT 30 ——— = calculated according to model 2; – – – = measured

Figure 42. Separation characteristics of nanofiltration in organic aqueous systems (operating conditions: p = 10 bar; = 25 °C; osmotic concentration xF = 200 mOsmol per kilogram H2O)

1 = Methanol; 2 = Ethanol; 3 = n-Butanol; 4 = Ethylene glycol; 5 = Triethylene glycol; 6 = Glucose; 7 = Sucrose; 8 = Lactose

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 58 of 82

Page 59: Water Ullmans

The Donnan effect is illustrated in Figure 43. The addition of ions, in the form of Na2SO4, reduces the rejection capacity

for Cl– ions significantly, even to negative values, while the rejection capacity for Na+ ions increases slightly.

Figure 43. Influence of the addition of in the form of Na2SO4 on Cl– rejection (operating conditions: membrane XP 45; c NaCl = 0.05 mol/L) [239] a) Na+ rejection; b) Cl– rejection

9.1.2. Mass Transport Resistance in Front of the Membrane In reverse osmosis and nanofiltration, where mass transport is membrane controlled in normal operation, the separation performance is usually overestimated if only the resistance of the membrane is considered. A concentration boundary layer on the side of the raw mixture (in front of the membrane) can, in unfavorable cases, clearly reduce the performance of the membrane and the quality of separation. The steady-state local conditions are shown qualitatively in Figure 44. For the purpose of this discussion, a binary solution is used.

The concentration profiles indicated schematically in Figure 44 are formed in front of the membrane due to the selectivity of the membrane. The concentration w of the preferentially permeating component W (water) decreases toward the membrane, and the concentration of the preferentially rejected component S, corresponding to wS = 1 – wW, increases. This phenomenon is called concentration polarization ( Membranes and Membrane Separation Processes – Concentration Polarization in Filtration Processes). On condition that concentration gradients in the direction of flow are negligibly small compared with the orthogonal gradients, the concentration profile is described by

with k ≡ D/ as mass-transfer coefficient for negligible wall flux (m″P 0) and with

the subscripts S1 and S2 denote the salt at point 1 and point 2, wSP is the salt concentration of the locally produced permeate. The mass-transfer coefficient can be calculated with sufficient accuracy by means of equations valid for heat transfer, using the Sherwood number instead of the Nusselt number and the Schmidt number instead of the Prandtl number. The correlations valid for tubes and rectangular channels are listed below.

where

Laminar flow fully developed hydrodynamic but developing concentration boundary layer (Sieder and Tate [240])

Turbulent flow (Linton and Sherwood [241])

The degree of concentration polarization is definitely determined by the hydrodynamics in front of the membrane. Good

Figure 44. Concentration polarization on the feed side of a reverse osmosis or nanofiltration membrane

″S, diff = mass flux of salt per unit area by diffusion; ″S, conv = mass flux of salt per unit area by convection; ″

W, diff = mass flux of water per unit area by diffusion; ″W, conv = mass flux of water per unit area by convection;

F = thickness of the concentration layer

(9.3)

(9.4)

Characteristic numbers Reynolds number = Re = dh v/

Schmidt number = Sc = /Dij

Sherwood number = Sh = k dh/Dij

dh = hydraulic diameter

dh = d (tube with diameter d )

dh = 2 h (channel with height h and width b, b h)

v = flow rate = kinematic viscosity

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 59 of 82

Page 60: Water Ullmans

hydrodynamics is of great importance for the development of modules (technical membrane configuration).

Estimation of the effect of concentration polarization in a technical membrane module is shown in Figure 45. The ratio of real flux to maximal possible flux (pure membrane resistance) is plotted against the ratio of the mass-transfer coefficient k to the pure water flux A · p (see Fig. 45). Clearly, the more efficient the membrane, the more important are proper hydrodynamics. (In order to control concentration polarization, k must increase with increasing values of A · p).

A special case of Equation (09.3) describes ultrafiltration in the surface layer-controlled region. In this region the stationary final value of the transmembrane flow (see Fig. 40, curve a) becomes fully independent of the transmembrane pressure difference (Fig. 39 A).

If the concentration in the boundary layer reaches saturation concentration, the system compensates for any further increase in driving force by an increase in surface layer. For a binary system consisting of water and dissolved macromolecules (e.g., proteins), which are quantitatively retained by the UF membrane, Equation (9.3) becomes

In this case, the permeate flux is determined only by the saturation concentration wP,max, the concentration of the feed, and the hydrodynamics.

9.2. Membranes ( Membranes and Membrane Separation Processes – Membrane Preparation and Membrane Module Constructions)

For the projecting and process development of any membrane separation process, a large range of selective and stable membrane materials is available. These materials have been optimized for many different applications. Because of the great variety of materials, the nature and efficiency of membranes must be described separately for each process.

Nevertheless, to better understand the processes occurring in a membrane, a short general introduction to the materials and structures of commercially available membranes is given. This discussion is limited to synthetic solid membranes, which are made of both organic and inorganic materials.

Figure 45. Influence of mass-transfer coefficient on the local performance of a membrane module [operating conditions: system water – NaCl – FT 30 SW; A = 4×10–7 m s–1 bar–1; B = 9.33×10–5 kg m–2 s–1; b = 8 bar per wt % NaCl;

F = 1000 kg/m3; wS* = 0.035 %; pF = 60 bar; pp = 1 bar; R = 1; ″p = A i ( p – b wS);

m″p (wS2)/m″p (wS1) = permeate flux with concentration polarization/permeate flux without concentration polarization a) bwS1/ p = 0.1; bwS1/ p = 0.3; c) bwS1/ p = 0.5; d) bwS1/ p = 0.8

(9.5)

9.2.1. Organic Membranes In organic membranes [242-244], the choice of membrane-forming polymer for a concrete separation problem is not arbitrary but is oriented to the following properties of the polymer:

1. Molar mass

2. Chemical structure and spatial arrangement of the macromolecules 3. Interactions between different macromolecules

The structural properties determine both the macroscopic properties of the membrane, such as thermal, chemical, and mechanical stability, and the microscopic “inner” properties, such as the permeability of the polymer to a certain component.

The demand for the highest possible permeate flow requires that the actual selective layer of a membrane be as thin as possible because the flow of a component is inversely proportional to membrane thickness.

As a consequence, membranes usually have an asymmetric structure. They consist of a very thin skin (active layer) and a porous support layer. As the actual selective barrier for mass transport, the active layer essentially determines the separation efficiency of the membrane.

In the case of asymmetric polymer membranes, integral-asymmetric and composite-asymmetric structures can be distinguished. Integral-asymmetric membranes are made by phase inversion (i.e., precipitation of the polymer from a homogeneous solution). The active layer and the substructure consist of the same material. The most successful materials used for integral-asymmetric reverse osmosis membranes are, e.g., cellulose acetate and polyamide.

In the case of composite-asymmetric membranes, a homogeneous polymer layer, which is as thin as possible, is applied to a microporous or porous structure. These membranes, called composite membranes, allow individual optimization of the selective layer and the porous support layer with regard to separation efficiency and mechanical, thermal, and chemical stability. In reverse osmosis, the support layer of many membranes is made of polysulfone, and the selective layer is made of, e.g., polyethyleneimine.

9.2.2. Inorganic Membranes

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 60 of 82

Page 61: Water Ullmans

Apart from organic membrane materials, inorganic membranes have become increasingly important since the 1980s [245]. Their advantages are:

1. High temperature stability

2. Chemical stability 3. No aging; long service lives 4. Backwashing possible 5. Controllable separation limit and separation efficiency

Their disadvantages are

1. Sensitivity to tension and bending

2. High investment costs 3. Use of temperature-sensitive sealing materials, which limits full utilization of the temperature stability of the

membranes

Inorganic membranes are not nearly as important as organic membranes. Materials used are carbon and zirconium oxide – aluminum oxide.

9.3. Modules ( Membranes and Membrane Separation Processes – Membrane Preparation and Membrane Module Constructions)

As mentioned above, membrane systems are modular (i.e., made up of individual modules). In module development, the following requirements, which are partly contradictory, are essential:

1. Good hydrodynamics (e.g., no dead zones)

2. Mechanical, chemical, and thermal stability of the module 3. High packing density of the module 4. Inexpensive production 5. Good cleaning possibilities 6. Inexpensive replacement of a membrane 7. Low friction losses

Module design must be oriented to these requirements. Since the main emphasis is on one or another requirement, depending on application, many completely differently designed module types are available on the market. Disregarding design details, the modules can be divided into two classes and six types:

Tubular Membranes Flat-sheet MembranesTubular module Plate and frame moduleCapillary module Spiral-wound moduleHollow-fiber module Disk module

9.3.1. Modules with Tubular Membranes Tubular Module. In this type of module, the membrane lies in the form of a tube on the inner side of pressure-stable pipes having a diameter of 6 – 24 mm. The mixture to be separated is pumped through the inside of the tube; the permeate is collected outside. If the supporting pipe is made of a material that is impermeable to the filtrate, a thin porous tubular fleece (e.g., of porous polyethylene) is placed between the supporting pipe and the membrane. This fleece does not impede the transport of the filtrate to the bores arranged at short intervals in the supporting pipe and provides the necessary support for the membrane, especially in the region of these bores. Membranes are partly replaceable and partly attached firmly to the supporting material. To increase the relatively low packing density (<80 m2 of membrane area per cubic meter of space enclosed), many producers arrange several modules in one jacket pipe (Fig. 46).

Capillary Module. Unlike tubular modules and modules with flat-sheet membranes, in which the membrane film is mechanically supported by a porous supporting structure, capillary and hollow-fiber membranes are pressure-stable.

The capillary module consists of larger (di = 0.5 – 6 mm) membrane capillaries, which have an asymmetric structure with the selective layer on the inside. The capillary module can be compared with a shell-and-tube heat exchanger. The membrane capillaries are arranged in parallel bundles potted in epoxy resin, thus forming a head plate at both ends. Capillary modules have a higher packing density than tubular modules. However, mass transfer conditions are inferior because of laminar flow (see Section Mass Transport Resistance in Front of the Membrane).

Figure 46. Tubular modules in jacket pipe

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 61 of 82

Page 62: Water Ullmans

Hollow-Fiber Module. Hollow-fiber modules have much smaller diameters (outer diameters 85 – 200 µm) than capillary modules. Hollow-fiber modules are employed in reverse osmosis.

In reverse osmosis, the mixture to be separated is always on the outside of the fiber (i.e., the selective layer is also on the outside), while the permeate flows in the lumen. In contrast to capillaries, hollow fibers are exposed to external pressure. These are favorable conditions, and accordingly hollow-fiber modules can resist pressure differences up to 100 bar. The individual fibers of a hollow-fiber module are assembled to a bundle and installed in a pressure vessel.

9.3.2. Modules with a Flat-Sheet Membrane Plate and Frame Modules. Plate and frame modules are used in reverse osmosis, microfiltration, and ultrafiltration. All designs have the following essential elements: the flat membrane; the support plate for the membrane; and plates for guiding the flow on the feed side. The elements are assembled in form of stacks.

An advantage of plate modules is that the permeate from each membrane pair is collected separately. Thus, membrane defects can be recognized and repaired by disconnecting the defect membrane pair without shutting down the entire module. In all plate modules, a disadvantage is the high number of individual seals and the relatively high friction losses caused by 180° changes in flow direction.

Spiral-Wound Modules. The spiral-wound module is the most widely used module type. It was first developed for reverse osmosis and is also used in nanofiltration and ultrafiltration. The design of a spiral-wound module is shown in Figure 47. In this module, one or several membrane pockets, each with a netlike spacer made of plastic, are wound spirally around the permeate-collecting tube. The membrane pocket consist of two membranes with a porous plastic fleece between them (permeate spacer). The membrane pocket is sealed on three sides and is connected to the perforated permeate-collecting tube on the fourth open side. Feed solution enters at the front surface and flows in axial direction through the module, while the permeate flows spirally within the porous support layer toward the collecting tube.

Spiral-wound modules have gained acceptance to a great extent because they combine a simple and inexpensive design with relative insensitivity to fouling and a relatively high packing density (up to 1000 m2/m3).

Disk Module. The concept of the disk module is comparable with that of the spiral-wound module. Here, too, two individual membranes are combined with a fabric fleece between them. However, the entire periphery of the disk is sealed. The permeate is collected centrally through an opening provided with an O-ring. A clamping bolt is also located in the central hole. It compresses the membrane disks and spacers in such a way that the permeate flow is sealed off from the raw mixture. The advantages of the disk module are, above all, the low pressure drops on the permeate side and a low susceptibility to plugging on the feed side (this is due to the more “open” spacers of the disk module, compared to the feedspacer of the spiral-wound module). The disadvantage is a relatively low packing density (<400 m2/m3) (Fig. 48).

Disk modules are used today up to pressure differences of p = 200 bar. This is of interest when, in the solutions to be treated, the osmotic pressure determines the limit of concentration. Spiral-wound modules are operated today at maximum pressure differences of 120 bar.

9.4. Fouling and Scaling ( Membranes and Membrane Separation Processes – Concentration Polarization and Membrane Fouling)

The efficiency of reverse osmosis and nanofiltration is often limited not by the osmotic pressure or the viscosity of the feed solution, but by two other phenomena:

1. Scaling: Membrane blockage due to increase in concentration of the dissolved constituents in the feed until the solubility limit is exceeded and the constituents precipitate on the membrane surface

2. Fouling: Membrane blockage because of suspended or colloidally dissolved substances carried in or because of growth of microorganisms

Figure 47. Schematic of a spiral-wound module a) Membrane; b) Permeate-release layer; c) Spacer; d) Cover

Figure 48. Disk module

9.4.1. Membrane Blockage due to Crystallization (Scaling) If the solubility limits of certain substances are exceeded in an aqueous system during the concentration of constituents, the substance precipitates and can form a layer on the membrane surface. In many cases, these layers can be removed from the

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 62 of 82

Page 63: Water Ullmans

membrane by washing with acid (e.g., citric acid). Especially in the case of spiral-wound and hollow-fiber modules, however, flushing the crystal sludge out of the module is difficult or impossible. For this reason, scaling must be avoided by, e.g.,

1. Chemical conversion, elimination, or stabilization of the substances

2. Restriction of concentration increase in such a way that the saturation concentration of none of the dissolved components is reached even at the critical spot of highest concentration (i.e., the membrane surface at unit exit)

To estimate this concentration limit or to estimate the required raw water pretreatment, the following properties in water analysis must be taken into account:

1. Temperature, pH, conductivity, total dissolved solids (TDS)

2. Acid – base capacity (m alkalinity, p alkalinity) 3. Concentration of Na+, Ca2+, Ba2+, Sr2+, Mn2+, Fe2+ 4. Concentration of Cl–, , , F–, , (from m and p alkalinities) 5. Concentration of SiO2, free Cl2, O2, and CO2 (from m and p alkalinities)

For the most important dissolved solids, calculations of the solubility limit or maximum possible concentration in the module are available in the literature [246-252].

Important water pretreatment measures (e.g., against CaCO3-scaling) are acid dosing ; precipitation; and

softening (substitution of Na+ for Ca2+).

Measures against CaSO4 scaling are limiting or even lowering of the water recovery rate so that CaSO4 precipitation is

avoided in all events; softening (substitution of Na+ for Ca2+); precipitation (Ca2+ CaCO3 ); and stabilization by polyphosphate or organic antiscalants.

9.4.2. Membrane Blockage due to Contaminants (Fouling) In practice, the formation of a surface layer due to suspended or colloidally dissolved substances carried in (and often not eliminated despite very careful pretreatment) or to the growth of microorganisms is the weak point of reverse osmosis. Although the thickness of the surface layer formed during operation is barely measurable, it represents a considerable additional resistance for the permeating components. Therefore, for safe operation of reverse osmosis plants, the fouling behavior must be studied in practice-related experiments with the real system “prefiltered raw water – original membrane.” Pretreatment measures for the prevention of fouling in reverse osmosis are [253-257] ( Flocculants)

1. Filtration through a sand or multimedia deep bed, usually in combination with flocculation

2. Polishing filtration using cartridge filters 3. Microfiltration or ultrafiltration 4. Flotation

9.5. Module Arrangements (Plant Design) Especially in reverse osmosis, nanofiltration, and ultrafiltration, proper hydrodynamics on the feed side are of central importance. In a continuously operating plant, the mass flow rate (decreasing) and the concentration (increasing) on the feed side change because essentially pure water is transferred from the feed to the permeate at every point of the membrane. In this connection, two module arrangements that take this problem into account are especially important (Figs. 49 and 50): (1) the tapered design and (2) the feed and bleed design, a plant structure made of several module blocks with circulation pumps.

Figure 49. Flow diagram of tapered design

Figure 50. Flow diagram of feed and bleed design (module blocks with circulation pumps)

The tapered design is normally used in very large plants for desalination of sea- and brackish water. Plants with feed and bleed design are used especially in wastewater treatment where smaller mass flow rates with a greater contamination potential must be treated.

9.6. Use of Membrane Processes in Water and Wastewater Technology 9.6.1. Desalination of Seawater and Brackish Water Fresh water can be separated from sea- or brackish water by (1) evaporation, if the phase transition of water during the separation process is accepted, and (2) membrane processes, where no phase transition occurs [258]. At present and in the near future, seawater will be desalinated mainly by evaporation processes. However, reverse osmosis is an attractive alternative to thermal processes.

The membranes used in the desalination of seawater and brackish water are either composite membranes (e.g., Dow-

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 63 of 82

Page 64: Water Ullmans

Filmtec's FT 30, the membrane Desal 3 of Desalination Industries) or asymmetric phase inversion membranes made of polyamide (e.g., by DuPont). The modules used are spiral-wound, hollow-fiber, and disk modules.

In all RO desalination plants in operation, the great effort involved in the pretreatment of water is conspicuous. The extent of treatment depends, first, on the composition of the raw water and, second, on the module concept employed. The required pretreatment is directly proportional to the sensitivity to fouling and scaling of the module type employed. In disk modules with a relatively open feed channel, adjustment of the pH, chlorination, and separation of the suspended solids by sand filtration are sufficient. In spiral-wound or hollow-fiber modules, additional fine filtration by cartridge filters is required. In general, treatment includes the following steps:

1. Chlorination to prevent growth of algae

2. Filtration (e.g., by a combination “sand filter and subsequent precoat filter”) 3. Adjustment of pH 4. Addition of hexametaphosphate to prevent CaSO4 precipitation in the modules 5. Fine filtration with a 10-µm cartridge filter (5-µm cartridge filter in the case of hollow-fiber modules)

Based on a large-scale operating plant, main design features of a reverse osmosis plant for the desalination of seawater are discussed below.

Seawater Desalination Plant in Las Palmas, Gran Canaria. This seawater desalination plant was put into operation in 1989 (Fig. 51). It has a capacity of about 36 000 m3/d and consists of six parallel trains of identical design. Each train consists of 128 pressure vessels, which contain six spiral-wound modules each. These are arranged in such a way that eight pressure vessels form one unit. This unit is divided into five pressure pipes in the first block and three pressure pipes in the second block (tapered design).

The entire plant achieves a water yield of 45 %. The product has a concentration of dissolved substances of 450 – 500 mg/L (seawater: 38 300 mg/L). The specific energy consumption of the plant is 6.16 kW · h/m3. A turbine at the brine discharge side of the reverse osmosis plant recovers 1.65 kW · h/m3. The total costs of the plant including pretreatment amounted to ca. $ 47×106 in 1989.

Apart from large-scale plants of this type, reverse osmosis is used especially in desalination plants with lower capacities. Reverse osmosis plants with disk modules have, for instance, fully replaced the evaporators previously used on passenger ships (cruise ships) and warships for the following reasons: (1) savings in floor space, and, even more important, in height, and (2) independence on heat sources such as ship's diesel engines.

Figure 51. Seawater desalination plant in Las Palmas, Gran Canaria

9.6.2. Recovery of -Caprolactam -Caprolactam, C6H11ON, M = 113.69 g/mol, is the monomer used for the production of polyamide 6. In the production of

polyamide 6, a wastewater stream containing 5 wt % of -caprolactam must be treated. Besides the treatment of the water, the objective is to recover the -caprolactam as a concentrated solution. This cannot be achieved with reverse osmosis because osmotic pressure limits the water recovery rate. Assuming 45 bar as the maximum acceptable osmotic pressure of the concentrate ( < p), the concentration of the concentrate is limited to 20 wt % and correspondingly, the water recovery to 75 %. In the example considered here, 254 m3 of -caprolactam solution accumulates per day so that after the membrane unit a concentrate flow rate of 63.5 m3/d is obtained which would have to be further concentrated (e.g., by evaporation). In addition, elaborate pretreatment would be required because wastewater has a high fouling potential. Thus, evaporation is carried out before reverse osmosis. Figure 52 shows the flow diagram of the process. The recovered -caprolactam can be reused for the production of polyamide 6. The condensate, which still contains ca. 0.1 % of -caprolactam is concentrated to 5 % in a reverse osmosis stage and recycled to the evaporation stage. The reverse osmosis permeate can be used in the producton.

The membrane plant in use is equipped with 15 pressure vessels, which contain four 8-inch spiral-wound modules each. The water recovery is 98.4 %, and the rejection of -caprolactam is >99 %.

Figure 52. Flow diagram of a plant for separation of -caprolactam from evaporator condensate

9.6.3. Leachates from Dump Sites

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 64 of 82

Page 65: Water Ullmans

According to the latest version of the German water management law (Wasserhaushaltsgesetz) [259] and successive decrees, leachates from dump sites must be treated before discharge.

Since garbage, industrial solid waste, and special waste have different constituents, the composition of leachates from dump sites depends not only on the age, mode of operation, and construction of the disposal site, but especially on the disposed substances. The average composition of leachates from garbage disposal sites (operating in the stable methane phase) is given in Figure 53.

In last ten years, the following process variants have proved especially favorable for the treatment of leachates from dump sites:

1. Combination of biological treatment and adsorption

2. Combination of biological treatment and nanofiltration with recirculation of the concentrate from nanofiltration into the bioreactor, partial removal of organics from the NF concentrate by adsorption or ozonation, if necessary.

3. Combination of membrane processes, e.g., RO – high pressure RO – drying of the concentrate – nitrogen removal

Biological Processes. In treating leachates from disposal sites, membrane bioreactors are increasingly replacing traditional techniques (i.e., open-tank bioreactors combined with sedimentation). Instead of sedimentation, ultrafiltration in combination with an aerobic – anoxic biological treatment stage (Fig. 54) has the following advantages [260]: (1) 100 % retention of biomass, especially of the nitrifying bacteria that are problematic in sedimentation; (2) operation at significantly increased biomass concentrations (solids up to 20 – 30 mg/L), resulting in significantly increased volume-specific capacities (space – time yield).

With 1-inch tubular modules and feed flow rates of ca. 4 m/s, a constant permeate flux is obtained over a period of weeks (ca. 100 L m–2 h–1). An ultrafiltration unit with a membrane area of 44 m2 combined with anoxic – aerobic closed-vessel bioreactors (plant capacity: 50 – 65 m3/h) is shown in Figure 55.

In the biological stage, easily degradable organics (BOD) and ammonium – ammonia are degraded preferentially [261]. The discharge from the bioreactor still contains

1. Recalcitrants

2. Nonbiodegradable organic compounds 3. Nonbiodegradable inorganic compounds (e.g., chlorides or heavy-metal salts)

Substances that are nonbiodegradable or poorly biodegradable are adsorbed conventionally on activated carbon. However, a combination of a membrane bioreactor with nanofiltration is an interesting alternative when the concentrate is recycled to the bioreactor, possibly after partial removal of organics from the NF concentrate by adsorption or ozonation. The mostly higher molar mass recalcitrants are rejected by NF membranes, whereas degradation products and low molar mass salts permeate the membrane with the water. If the concentrate from nanofiltration is recycled to the bioreactor, the hydraulic residence time and the residence time of the recalcitrants are decoupled, in contrast with classic biological treatment. The bioreactor then operates at much higher concentration of recalcitrants than in the raw water. As a consequence the rate of degradation increases according to the Monod law. This rate increases with increasing concentration until inhibition by toxic substances occurs.

In some cases, the amount of water discharged with excess sludge is sufficient to guarantee stable operation of the bioreactor at high concentration (prevention of inadmissible concentration of nondegradable substances in the biological treatment stage). If not, intolerable concentrations in the loop can be reduced via adsorption or ozonation. Compared to a

Figure 53. Average composition of dumpsite leachate

Figure 54. Flow diagram of “BioMembrat Plus” process

Figure 55. Ultrafiltration plant for separation of biomass from biologically clarified leachates

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 65 of 82

Page 66: Water Ullmans

state-of-the-art combination “bioreactors – adsorption” or “bioreactor – ozonation”, adsorption and ozonation in combination with nanofiltration are much more efficient, because these combinations can operate at favorably high concentration levels instead of the low discharge concentration level.

Reverse Osmosis – Evaporation or Reverse Osmosis – Nanofiltration. A successful alternative to the biological treatment of dumpsite leachate is treatment by reverse osmosis in combination with further concentration steps.

In most cases the RO unit is designed as a two-stage cascade with tubular or disk modules in the first and spiral wound or disk modules in the second stage. Using a two-stage cascade with the permeate of the first stage cascade serving as feed for the second stage the legal limiting values for discharge can be safely guaranteed [262].

Reverse Osmosis – >Evaporation. Usually the RO concentrate is concentrated further by evaporation and, finally, dried when the disposal of a dry residue is stipulated. Furthermore, the process must contain a step for the separation of either a stripping unit or a bioreactor. This process combination can be considered as state of the art. Disadvantageous are the high specific energy consumption and the high specific investment costs of evaporation – dryer stages.

From the composition of leachate (Fig. 53) follows that the osmotic pressure is a limiting factor for concentration by RO. Modules are now available for operation pressures of 120 and 200 bar. By implementation of these modules into the process the evaporation stage can be eliminated. Instead of concentration factors of 2 – 3, concentration factors of up to 10 can be achieved by the RO stage.

Since only small amounts of water are removed in the drying step, which, however, gives rise to 35 – 38 % of the overall treatment costs, it must be investigated whether this process combination can be improved. A simple further increase of the water recovery rate by RO cannot be expected for two reasons:

1. Danger of calcium sulfate scaling

2. Unacceptably high membrane compaction at transmembrane pressure differences above 200 bar

As shown in Figure 53, chlorides form a major part of the inorganic components of leachates. Due to the high rejection of RO membranes for chlorides, they remain almost completely in the concentrate, resulting in a high osmotic pressure and causing disposal problem.

Reverse Osmosis – Nanofiltration. The implementation of a nanofiltration stage into the treatment process can considerably extend the limits of RO set by scaling and/or osmotic pressure. Figure 56 shows the flow diagram of the treatment process. Essentially, the concentrate of the 120 bar RO stage is treated by a combination of NF and crystallization. At moderate transmembrane pressure differences (20 – 50 bar) the nanofiltration unit produces a permeate containing mainly chlorides. This permeate can be concentrated further without danger of scaling by a 200-bar high-pressure reverse osmosis stage. Essential for the process is that the NF – crystallizer cycle is able to operate in the supersaturation range with respect to calcium sulfate and at high concentrations of organics. For this reason the success of the process depends to a large extent on a proper module design — the modules must be insensitive to fouling and to the presence of crystals. Such modules have been developed by Rochem, Germany. Stacks of rectangular membrane cushions and matching spacer plates are arranged in series in the pressure vessel in such a way that stagnant areas are avoided and internal friction losses are minimized. Feed flow is strictly parallel to the vessel axis, the velocity in the feedsite module channels is ca. 1.5 m/s.

The process has been installed and tested on a technical scale at the Ihlenberg dumpsite. A nanofiltration stage consisting of four blocks with nine modules each and a total membrane area of 180 m2 has been added to the existing reverse osmosis – high-pressure reverse osmosis combination and comissioned in September 1994 [263]. The unit is designed for the treatment of about 4 m3/h RO concentrate with a volumetric concentration factor of CFV RO = 10. Depending on the

concentration factor of the 60 and 120 bar RO stage, the NF stage achieves a further concentration of CFV NF = 10 – 20. A sludge consisting of organics, precipitated inorganics, and water is discontinuously withdrawn from the bottom of the crystallizer – sedimentation tank.

Together with the concentrate of the subsequent 200 bar high-pressure reverse osmosis stage this concentrate is solidified with fly ash and disposed of at the dumpsite.

Cleaning of the NF modules is performed by flushing with feed at zero transmembrane pressure difference for 30 s every hour and an alkaline cleaning every 250 – 300 h. The NF rejection rate for sulfates is 92 – 95 %, for the dissolved solids 20 – 35 %.

At the Ihlenberg dumpsite the specific energy consumption of the NF and the subsequent 200 bar RO stage is 32 kW · h per cubic meter permeate. With figures of 8.5 kW · h per cubic meter of totally produced permeate and an overall water recovery rate of 97 %, the overall specific power consumption of the process is extremely low compared to other processes. With respect to residues the process can be regarded as an almost zero discharge process.

Figure 56. Flow diagram of the combination reverse osmosis – nanofiltration – high pressue reverse osmosis for leachate treatment (Ihlenberg dumpsite, Germany)

9.6.4. Wastewater from the Dye Industry The production of textile dyes involves a precipitation by salting-out as the final stage. After separation of the dye, the

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 66 of 82

Page 67: Water Ullmans

wastewater of the process contains the salt and, unavoidably, small amounts of dye. Treatment of this wastewater by reverse osmosis is not possible because of the high osmotic pressure and because the resulting concentrate would contain both salts and organics. However, the ability of nanofiltration to fractionate low molar mass salts and higher molar mass dyes can be used here.

A plant installed in a Swiss chemical company is shown in Figure 57. It has a total membrane area of 960 m2 and is equipped with tubular modules because of the high fouling potential of feed. It operates at a volumetric concentration factor CFV = 10 and has an average capacity of ca. 16 m3/h. Since the wastewater of several dye-production plants is treated simultaneously, the NF concentrate cannot be recycled, but must be disposed of. To this end, the concentrate is subjected to wet oxidation. The permeate, containing the major part of the salts, can be discharged into the Rhine river. Here, nanofiltration has two advantages:

1. The mass flow rate of the stream subjected to wet oxidation is significantly lower.

2. The corrosion potential of this stream is strongly reduced, because the major part of the salts is removed.

Obviously, nanofiltration offers a possibility of recovering the dyes, at least in principle. This would be of economic interest where large amounts of only one dye are produced.

[Top of Page]

10. Oxidation Processes in Water Treatment William H. Glaze

Oxidation processes are used in water treatment for disinfection and removal of obnoxious or potentially toxic contaminants in the water, and for some industrial purposes such as pulp bleaching and purification of wastewater before discharge into the environment. The most familiar chemical oxidant is oxygen from air; other examples of oxidants used in water treatment are chlorine, chloroamines, ozone, chlorine dioxide, hydrogen peroxide, and potassium permanganate. Sometimes these oxidants are used in combination with each other (e.g., ozone with hydrogen peroxide) or with photons from irradiation lamps. Oxidation in water treatment is often carried out with the assistance of microbiological organisms, in which case the oxidant is oxygen ( Wastewater – Biological Principles, Wastewater – Basic Biotechnological Considerations,

Wastewater). Only chemical oxidation is considered in this chapter.

10.1. Physical and Chemical Properties of Chemical Oxidants The power of a chemical oxidant to cause a chemical oxidation reaction is measured by two properties: (1) the thermodynamic driving force and (2) the rate constant of oxidation. The thermodynamic driving force of the oxidation process for one oxidant compared to another is determined by its oxidation potential. Table 31 is a list of the standard potentials E 0 for common chemical oxidants in aqueous solution at 298 K [265], [266].

Table 31. Standard electrode potentials for some common oxidizing agents in aqueous solution at 298 K [265], [266]

The oxidation potential is a measure of a chemical's oxidizing (or reducing) power; the more positive the value of E 0, the more likely is the chemical reaction to take place as written. The parameter E 0 is related to the free energy G 0 by the

Figure 57. Nanofiltration plant for treatment of wastewater from the production of textile dyes

Oxidant Half reaction E 0, V

Oxygen [7782-44-7] O2 + 4 H+ + 4 e– 2 H2O 1.23

Ozone [10028-15-6] O3(g) + 2 H+ + 2 e– O2(g) + H2O 2.08

Chlorine [7782-50-5] Cl2 + 2 e– 2 Cl– 1.36

Hypochlorous acid [7790-92-3] HOCl + H+ + 2 e– Cl– + H2O 1.50

Hypochlorite ClO– + H2O + 2 e– Cl– + 2 OH– 0.90

Monochloroamine [10599-90-3] ClNH2 + H2O + 2 e– Cl– + NH3 + OH– 0.75

Chlorine dioxide [10049-04-4] ClO2 + 2 H2O + 5 e– Cl– + 4 OH– 1.71

Permanganate 1.68Hydrogen peroxide [7722-84-1] H2O2 + 2 H+ + 2 e– 2 H2O 1.78

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 67 of 82

Page 68: Water Ullmans

formula G 0 = – nFE 0, where n is the number of electrons transferred and F is the Faraday constant (9.648×104 J/V). For example, the large positive value of E 0 (2.08 V) for Equation (10.1) below indicates that ozone is a powerful oxidizing agent.

The values in Table 31 are written as half reactions; thus, another half reaction is needed to complete the chemical equation. If the compound being oxidized by ozone is formate (Equation 10.2), the overall process would be as follows:

The overall chemical free energy change for the process represented by Equation (10.3) (i.e., the driving force) is calculated from

The large negative value of G 0 indicates that the reaction should proceed spontaneously.

All common chemical oxidants shown in Table 31 are relatively powerful and should oxidize most organic compounds to carbon dioxide and water. The fact that this does not always occur (e.g., organic compounds on our planet are stable to oxidation by oxygen in most cases) indicates that thermodynamics is not the only important factor in determining whether a chemical reaction will occur spontaneously.

The other factor that is important in rating chemical oxidants is how fast they cause a chemical oxidation to proceed. This is the domain of chemical kinetics. In fact, in very few cases do the oxidants listed in Table 31 oxidize environmental contaminants as completely as would be expected from thermodynamic considerations. Chemical reactions that should occur, as indicated by the thermodynamic driving force, may do so very slowly if the rate constant for the process is low. For example, the reaction of formate with ozone is actually a very slow reaction, whereas that of another common pollutant, phenol, is very fast [267]. Another example is the rate of oxidation of two important natural products that contribute to bad taste and odor in water supplies: geosmin and methylisoborneol. The rates of chemical oxidation of geosmin and methylisoborneol by chemical oxidants vary substantially, making some impractical for use in water treatment. Of the common oxidants, ozone is more effective than others for the control of obnoxious taste and odor.

Chlorine. The solubility of chlorine in water is high compared to that of oxygen because in water, Cl2 hydrolyzes to form hypochlorous acid:

Hypochlorous acid is a weak acid with a pKa of 7.50 at 25 °C, dissociating to a hydrogen ion and the hypochlorite ion:

This means that at pH 7.50, equal amounts of HOCl and hypochlorite ion (OCl–) are present. At higher pH, OCl– predominates; at lower pH, HOCl. Hypochlorite ion is a weaker oxidant than HOCl (see Table 31) and also a weaker disinfectant. Both are powerful oxidizing agents as indicated by the E 0 values in Table 31, so aqueous chlorine is capable of oxidizing many organic and inorganic compounds [268], [269]. For example, chlorine has been used for years as a bleaching agent in pulp and paper treatment, for the oxidation of sulfides and cyanides, and for the oxidation of metals such as iron and manganese. Aqueous chlorine also reacts with organic compounds to substitute chlorine for hydrogen, which yields new chlorinated compounds (see Section Toxicology and Environmental Health). Chlorine is a toxic gas, and worker exposure must be avoided ( Chlorine – Toxicology).

Chloroamines. Chloroamines are formed by reaction of chlorine with ammonia [268], [269]. The reaction occurs in three steps as shown below, leading to mono-, di-, and trichloroamines. These have slightly different oxidation potential and disinfection power, but in general, chloroamines are not as potent in both respects as chlorine. However, chloroamines are more stable in water than chlorine so they can maintain an active disinfection capacity for longer periods of time.

Ozone is always produced when oxygen is decomposed in air, e.g., in an electric discharge or when shortwave (<190 nm) UV radiation is absorbed by oxygen. This decomposition occurs naturally in the atmosphere, but it may also be used to generate ozone for water treatment by the reaction shown below using either air, oxygen-enriched air, or pure oxygen.

O3(g) + 2H+ + 2e– O2(g) + H2O E 0 = 2.08 V (10.1)

HCOO– CO2(g) + H+ + 2e– –E 0 = 0.31 V (10.2)

O3(g) + HCOO– + H+ CO2(g) + O2(g) + H2O (10.3)

(10.4)

(10.5)

(10.6)

(10.7)

(10.8)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 68 of 82

Page 69: Water Ullmans

Ozone is more soluble in water than oxygen but less soluble than chlorine. In water it decomposes slowly, with a half-life that depends on the other substances present in water. Ozone is a toxic gas, and monitors are necessary where ozone is made and used, to minimize worker exposure.

Ozone is a powerful oxidant but reacts with different chemicals at various rates that range over several orders of magnitude. For example, the half-lifes of ozone's reactions with phenol and formate under similar conditions differ by a factor of about 100 000 [267]. Like chlorine, ozone has been used for the treatment of drinking water for disinfection, color and taste control and other purposes for over nine decades (ozone since 1906, installation at the Bon Voyage plant in Nice, France) [270]. Ozone has been more popular in Europe, especially France, than in North America, but this has been changing lately because of some of the environmental impacts of chlorine (see Section Toxicology and Environmental Health). Ozone alone is not always effective in oxidizing organic substances in water such as pollutants or taste and odor compounds, but its effectiveness may be enhanced by combining it with other oxidants (such as hydrogen peroxide) or with UV radiation [270].

Chlorine dioxide ( Chlorine Oxides and Chlorine Oxygen Acids – Chlorine Dioxide) is a clear, colorless gas that dissolves readily in water without dissociation. It is a powerful disinfectant and oxidant (Table 31) and has been used extensively for water treatment. Chlorine dioxide is becoming an increasingly important alternative to chlorine as a bleaching agent in pulp and paper production. However, chlorine dioxide reacts rapidly with many substances in natural water, such as natural organic matter, in a reduction – oxidation reaction that produces chlorite ion . Chlorite has some documented health effects such as irreversible binding to hemoglobin, so the total amount of chlorine dioxide used in drinking water treatment is regulated in some countries (e.g., United States). Also, residuals of chlorine dioxide in drinking water have an objectionable taste and odor to some people.

Potassium permanganate has been used in water treatment for many years, primarily for oxidation of manganese(II) and iron(II) to the corresponding insoluble oxides or hydroxides, for color removal, and for taste and odor control. Potassium permanganate is usually not recommended for drinking water disinfection. Also, care must be taken not to overdose and produce a pink-colored water, which is objectionable to consumers.

Hydrogen peroxide ( Hydrogen Peroxide) is a milder oxidant than chlorine or ozone. It is used in many industrial and medical applications as a mild disinfectant or oxidant, but worldwide its principal application is in wood pulping.

10.2. Production of Chemical Oxidants Chlorine is produced in very large quantities worldwide, only a smal fraction of which is used for water treatment. Chlorine gas is produced mainly by electrolysis of brine (sodium chloride solution) according to the chlor – alkali, the mercury cell, or the diaphragm process ( Chlorine, Chlorine – Mercury Cell Process, Chlorine – Diaphragm Process,

Chlorine – Membrane Process). Pure chlorine can be transported as liquefied gas and added to water under pressure, where it dissolves readily and hydrolyzes to hypochlorous acid and hypochlorite ion (see Eqs. 10.4 and 10.5). Commercial units are available to meter a specified amount of chlorine into water, usually into a slipstream which is then added to the total water stream to be treated so as to give the desired dose. Alternatively, aqueous chlorine (hypochlorite) may be prepared at the site of chlorine production and shipped to the location where it can be metered into the water to be treated. The latter is preferred in some highly populated areas where transport of gaseous chlorine is forbidden.

Chloroamines. The formation of chloroamines is carried out on-site by combining a chlorine solution with aqueous ammonia. To minimize byproducts, the chloroamines may be preformed by using an excess of ammonia so that no chlorine remains. A chlorine – ammonia weight ratio of 3: 1 corresponds roughly to the stoichiometric amount to form monochloroamine plus some dichloroamine, and a pH of 8.4 is optimal. Excess chlorine leads to the formation of trichloroamine, which has an objectionable taste.

Ozone is always produced on-site because it is an unstable gas that cannot be stored or shipped. Ozone is usually produced by an electric discharge in air, oxygen, or oxygen-enriched air ( Ozone – Production). The most commonly used generators are of the silent discharge type in which discharge occurs across a dielectric barrier [270]. Small ozone generators may utilize discharge between ceramic electrodes or irradiation lamps.

Chlorine dioxide is also formed on-site, usually by reaction of sodium chlorite and chlorine (hypochlorous acid):

Commercial generators are available to regulate the ratio of chlorine and chlorite carefully or to recycle the stream so as to minimize the content of chlorine and byproducts in the product and to consume all of the chlorite ion. However, chlorine dioxide used in water treatment always contains some chlorite and usually chlorate, as well.

Potassium permanganate is produced by electrolysis of potassium manganate(VI) (K2MnO4), which is obtained by melting manganese dioxide and potassium hydroxide in the presence of air ( Manganese Compounds – Fusion Processes ,

Manganese Compounds – Anodic Oxidation of Manganate(VI)).

Hydrogen peroxide is usually produced from oxygen by using a catalytic process involving anthraquinone ( Hydrogen Peroxide – Anthraquinone Process (AO Process)). It is marketed as a solution in water, usually containing 35 – 50 % H2O2. The compound is unstable, decomposing to oxygen and water by a reaction that is catalyzed by several substances, even

(10.11)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 69 of 82

Page 70: Water Ullmans

common soil. Hydrogen peroxide must therefore be shipped and stored in vessels made of passivated stainless steel, for example, in which it is more stable.

Mixed Oxidants. Oxidants are sometimes used in combination, the most common example being a mixture of ozone and hydrogen peroxide. These are always mixed at the site of application because the chemical reaction that occurs produces free-radical intermediates that have a very short lifetime (see Section Advanced Oxidation Processes).

10.3. Uses of Chemical Oxidants Oxidants are used in water treatment for a variety of purposes, including disinfection and oxidation, in the treatment of drinking water and wastewater, oxidation in pulp and paper treatment, and disinfection and scale control in the treatment of cooling tower water.

10.3.1. Drinking Water Treatment Chemical oxidants are used for treatment of drinking water, mostly for disinfection and removal of obnoxious chemical compounds or potential toxicants. The most common drinking water oxidant worldwide is chlorine (hypochlorous acid). In parts of Europe, ozone has been used for as long as chlorine, and it is growing in use worldwide.

10.3.1.1. Disinfection Treatment with chemical oxidants has become the traditional method for disinfection of water for distribution in municipal water systems. Ozone and chlorine were first used for this purpose at about the beginning of this century and have been used continuously since that time [268], [269]. Chlorine is the more preferred of the two because it is more stable and maintains its disinfection capability for longer periods of time. Ozone decomposes quickly but is a more powerful disinfectant.

Drinking water sources can be classified generally as groundwater or surface water. Groundwater often does not require treatment except for disinfection, usually by addition of a chemical oxidant such as chlorine. In a drinking water plant for treating surface water, disinfection is accompanied by other unit processes such as coagulation (see Chap. Flocculation), sedimentation, and filtration (see Chap. Filtration). Also, the chemical oxidant that is applied as a disinfectant may be used to achieve other treatment objectives such as chemical oxidation (see Section Chemical Oxidation). The oxidant – disinfectant may be added at the beginning of treatment (peroxidation) or at other points in the treatment train. The amount of oxidant added and its contact time with the water are chosen such that the kill of microorganisms is highly efficient. At the end of the treatment train, a smaller amount of disinfectant (e.g., chlorine) is usually added so that water being sent to the customer will have a residual amount of disinfection protection. In Western Europe, the amount of residual disinfectant used is lower (≈0.1 mg/L) than is traditional in the United States (≈ 1 mg/L).

Disinfection efficacy is usually measured in terms of the dose (mg/L) of oxidant required to kill a certain fraction of a microorganism. Since the kill is also determined by the time the microorganism is exposed to the dose, the product of dose (C ) times contact time (T ) is an important process control parameter. In the United States, the Environmental Protection Agency has published values of CT for each disinfectant required to kill a certain fraction of various microorganisms. For tests used in other countries, see Disinfectants – Efficacy Testing Methods, Requirements for Efficacy.

10.3.1.2. Chemical Oxidation Oxidants are also used in drinking water treatment to remove substances that give unwelcome color to the water, compounds that cause poor taste or odor, and micropollutants that may have deleterious health effects. In addition, oxidants are used to remove iron and manganese, which are not toxic but may cause discoloration of household appliances.

Color in natural water is caused by natural organic matter (NOM) and sometimes by metallic elements complexed with NOM. Color is measured by comparison with a standard solution of a colored chemical compound such as chloroplatinic acid. The color of the water is expressed as “color units” (i.e., the amount of chloroplatinic acid necessary to give the same color). In drinking water plants using water sources that are highly colored, the water is usually treated with a chemical oxidant so that product water does not have a discernible color. Oxidants destroy color by chemically oxidizing the color center in the macromolecular NOM. Of the common chemical oxidants, ozone appears to be most effective for removing color. However, since the chemical structures in NOM that cause color vary from source to source, another oxidant may be more or less successful than ozone in removing the color, depending on the water source.

Taste and odor in water supplies are usually caused by chemical compounds produced by algae and other organisms but may also result from pollution due to industrial discharge. Again, the source of the bad taste or odor may vary from source to source, so a particular chemical oxidant may vary in its ability to remove the problem. Two compounds produced by algae, geosmin and methylisoborneol, are particularly obnoxious in that they can be sensed by humans at the ng/L level. Ozone is virtually the only common chemical oxidant that is effective in removing geosmin and methylisoborneol.

Iron and manganese are usually present in natural water. Higher oxidation states of both metals form rather insoluble hydroxides or oxides [e.g., Fe(OH)3 and MnO2], but lower oxidation states are more soluble (Fe2+ and Mn2+). Therefore, these soluble forms are likely found under anoxic (reducing) conditions, such as in groundwater and below the hypolimnion in surface water. These substances are preferably removed in drinking water treatment before the water is sent to customers, where oxidation may cause precipitation of unwanted, colored oxides in plumbing. Oxidation of iron is a rapid reaction that can be accomplished easily in water treatment, even with molecular oxygen. All of the common oxidants can also easily oxidize Fe(II) to Fe(III), which is then precipitated and removed by filtration. Stronger oxidants are required for oxidation of Mn(II) to Mn(IV). Chlorine, potassium permanganate, and ozone are most commonly used, usually at the beginning of treatment prior to coagulation and sedimentation.

Micropollutants are synthetic organic or inorganic chemicals present in natural surface water sources because of discharges from industries, municipal waste treatment plants, and other sources. For example, surface waters in Europe and the United States are contaminated with agricultural chemicals, especially in the spring when herbicides are used in

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 70 of 82

Page 71: Water Ullmans

abundance. Among the common oxidants, ozone is the most effective for removal of these and other organic micropollutants, but addition of hydrogen peroxide just before ozonation increases removal. The chemical mechanism responsible for the synergy between ozone and hydrogen peroxide is discussed in Section Advanced Oxidation Processes.

Groundwater may be contaminated with solvents, petroleum distillates, pesticides or herbicides, and other compounds that migrate into aquifers from spills, unwise use, and other human activities. A variety of treatment methods have been developed to clean contaminated groundwater, including activated carbon adsorption and air stripping. Oxidants may also be used to treat contaminated groundwater, but again the efficacy of treatment will vary depending on the molecular form of the contaminant. With regard to the most common groundwater contaminants, oxidation has been most successful in treating tri- and tetrachloroethylene, solvents that have leaked into reservoirs from storage tank, and accidental spills. The most successful oxidation methods for this purpose are ozone with hydrogen peroxide and ozone or hydrogen peroxide with UV radiation.

10.3.2. Wastewater Treatment The use of chemical oxidants is not as common in wastewater treatment as in drinking water treatment because wastewater usually contains more impurities and, therefore, would consume more oxidants. Since oxidants are rather costly commodities, they are generally used for polishing (i.e., treating wastewater just before discharge) after most impurities have been removed [268], [269].

Chlorine has been used traditionally for treatment of domestic wastewater after biological treatment. The rationale was to minimize the concentration of microorganisms and ammonia being discharged into receiving streams. The chemistry of ammonia removal is essentially the same as that used to form chloroamines (see Section Physical and Chemical Properties of Chemical Oxidants), but in this case an excess of chlorine is added to oxidize the ammonia completely to nitrogen. This is referred to as breakpoint chlorination:

Breakpoint chlorination of wastewater is still practiced, but concern over the chlorine byproducts produced is increasing so alternative oxidants such as ozone are gaining in favor.

(10.12)

10.3.3. Oxidation in Pulp and Paper Treatment Chemical oxidants are used for bleaching pulp and paper ( Paper and Pulp – Conventional Bleaching, Paper and Pulp), and for decolorizing wastewater prior to discharge. Chlorine was the chemical oxidant of choice, but during the last two decades evidence of the formation of chlorinated byproducts including chlorinated dioxins and furans, has forced a reexamination of this practice. Of particular concern is the formation of chlorinated dioxins such as the highly carcinogenic 2,3,7,8-tetrachloro isomer. Modification of pulp and paper treatment is proceeding on a worldwide basis. Chlorine dioxide, oxygen, hydrogen peroxide, and ozone are being used instead of chlorine.

10.3.4. Advanced Oxidation Processes The term advanced oxidation processes (AOPs) is used for water treatment processes that involve the formation of highly reactive, short-lived chemical intermediates. These intermediates (e.g., the hydroxyl radical) are powerful oxidants — more powerful than the conventional oxidants discussed above — and are used to oxidize water contaminants that resist other forms of treatment. Several AOPs have been developed and commercialized including the following:

1. Ozone in combination with hydrogen peroxide (sometimes called Peroxone process) with UV radiation, or with both hydrogen peroxide and UV

2. Hydrogen peroxide in combination with iron(II) salts (Fenton process), with UV radiation, or with UV radiation and other modifiers such as iodine ion, iron(III) salts, etc.

3. Oxygen in combination with high-energy, high-frequency sound waves (sonication), electron-beam irradiation, or gamma radiation

Oxidation with AOPs is used worldwide in the treatment of drinking water for oxidation of herbicides such as atrazine and simazine, and for oxidation of taste and odor compounds that are formed in rivers, lakes, and reservoirs. AOPs are also used in groundwater treatment to oxidize contaminants such as halogenated solvents (trichloroethylene and tetrachloroethylene). In addition, AOPs are being investigated for treatment of industrial wastewater, but they appear to be substantially more expensive than conventional biological treatment. Thus, AOPs may be most useful for treating low-volume streams containing compounds that are not easily biodegraded. In this case, oxidizing the target compounds only partially (i.e., not completely to carbon dioxide) may be most economical. Partially oxidized compounds are likely to be more easily biodegradable than their precursors, and the stream may then be given to a biological treatment unit process for final polishing before discharge into the environment.

Advanced oxidation processes that involve UV radiation utilize lamps such as mercury arc lamps that produce radiation of various wavelengths, depending on the power density of the current in the lamp. Low-pressure mercury arc lamps produce UV radiation primarily at 254-nm wavelength, whereas high-pressure lamps (operated at higher current densities and therefore at higher internal pressure and temperature) produce UV radiation with variable intensity at several different wavelengths from the visible region down to less than 200 nm. However, lamps are usually housed in quartz envelopes that cut off all radiation below 200 nm or, in the case of highly purified quartz, below 185 nm.

Low-pressure mercury arc lamps are suitable for disinfection of water because the 254-nm wavelength is absorbed by the DNA in bacterial and viral cells. This wavelength is also suitable for decomposition of ozone in the ozone – UV AOP, since ozone absorbs strongly at 254 nm. The following chemical equations summarize this complex process, which undoubtedly involves several steps not shown:

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 71 of 82

Page 72: Water Ullmans

When ozone is used in combination with hydrogen peroxide, the last three equations summarize the chemical reaction.

High-pressure mercury arc sources are more suitable for other AOPs such as the H2O2 – UV process where UV radiation decomposes peroxide into OH radicals, which then oxidize water contaminants. The decomposition of hydrogen peroxide is described by the equations shown below:

Similarly, when hydrogen peroxide is used in combination with iron(II) salts, OH radicals are produced by an oxidation – reduction reaction:

The same radicals can be produced by direct decomposition of water by high-energy radiation or sonication:

As noted above, OH radicals in combination with oxygen oxidize organic contaminants in water eventually to form oxidized byproducts and, if enough oxidant is used, to form carbon dioxide and water. Whether the AOP is used to oxidize the contaminant completely is determined by economics.

10.4. Toxicology and Environmental Health Chlorine and other chemical oxidants have had a very positive influence on human health because they are the principal disinfectants used in drinking water treatment, disinfection of sanitary facilities, and other applications. The incidence of waterborne diseases such as cholera and typhoid fever decreased dramatically in the early decades of the 20th century as water disinfection was practiced more extensively. In parts of Europe, particularly France, ozone served as the principal water disinfectant, whereas chlorine was preferred worldwide.

Chemical oxidants are generally toxic compounds themselves, so workers and the general public should be protected from large doses. However, low levels of oxidants, particularly chorine and chloroamines, are consumed regularly in drinking water with no apparent ill effects.

Disinfection Byproducts. In 1974, workers in the Netherlands and the United States discovered that chlorination of water produces a set of previously undetected byproducts known as the trihalomethanes (THMs) [271]. Trihalomethanes in this context usually mean four compounds: CHCl3, CHCl2Br, CHClBr2, and CHBr3, but other analogous containing mixtures of iodine, bromine, and chlorine (but not fluorine) as substituents are produced in low quantities. As a class, these compounds are called disinfection byproducts (DBPs). Subsequent studies showed that these and many other chlorinated and brominated byproducts are formed from the reaction of aqueous chlorine and the organic matter present in all natural water.

Other chlorinated DBPs include halogenated acetic acids such as trichloroacetic acid and mixed chloro – bromo analogues, chlorinated acetonitriles such as Cl2HCCN, halogenated aldehydes and ketones, and many others. Studies have shown that the yield of these compounds is generally in the microgram – milligram per liter range and depends on the concentration of NOM in the water, the pH, temperature, and chlorine dose.

Concern over the formation of DBPs was immediate because some of the compounds had previously been shown to be carcinogenic in test animals or to have other toxic effects. As a result, THMs are now the subject of water quality regulations in many countries, and maximum concentration levels are recommended by the World Health Organization. Depending on the country, the limit of THMs allowed in finished drinking water is in the range of 100 – 200 µg/L.

As a result of these regulations, which may be strengthened in the future, the use of chlorine in water treatment is being

(10.14)

(10.15)

(10.16)

(10.17)

(10.16)

(10.17)

(10.19)

(10.21)

(10.22)

(10.23)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 72 of 82

Page 73: Water Ullmans

reexamined. Many water treatment facilities are now using lesser amounts of chlorine; others have shifted to an alternative disinfectant such as ozone, chloroamines, or chlorine dioxide. However, all chemical oxidants react with naturally occurring compounds in water and form chemical byproducts [272]. For example, ozone reacts with naturally occurring bromide ion to form bromate ion, a substance that causes cancer in animal bioassays. The oxidation of bromide proceeds through the formation of hypobromite ion, which also may form byproducts on conversion to its corresponding acid form:

Chlorine dioxide and chloroamines form byproducts on reaction with NOM, although these products have not been as well characterized.

Other Oxidation Byproducts. Chemical oxidants used in wastewater treatment also form chemical byproducts and, in some cases, may affect the natural species in the receiving streams into which these waters are discharged. For example, residual ammonia in domestic wastewater is chlorinated to form chloroamines (see Section Physical and Chemical Properties of Chemical Oxidants). Chloroamines are toxic to fish and other aquatic organisms. In addition, chlorine may react with unidentified substances to form chlorinated dioxins, furans, furanones, and thiophenes. Some of these may have direct toxic effects, others may bioaccumulate in organisms and have long-term effects.

Oxidation byproducts may be produced even when oxidants are used in groundwater treatment. For example, oxidation of tri-and tetrachloroethylene, common groundwater contaminants, produces small amounts of chlorinated acetic acids, which may be more potent carcinogens than their precursors. Also, oxidation of simazine and atrazine, two common contaminants in surface water, produces various partially oxidized derivatives of the original compound.

In summary, all chemical oxidants form byproducts when they are used in water treatment. In some cases, such as drinking water chlorination, these byproducts have been well characterized and their toxicological properties are well known. World health authorities agree that the levels of these byproducts should be minimized, and regulations to do so have been promulgated in many countries. For other oxidants, very little information on the formation of chemical oxidation byproducts is available and further research is needed.

Finally, the risks from chemical byproducts formed by the use of oxidants as water disinfectants are rather low. Toxicologists estimate that the risk of obtaining cancer from the consumption of water containing chloroform at a concentration of 100 µg/L for a lifetime is only about 1 out of 100 000 – 10 000. In contrast, the absence of water disinfection can result in large outbreaks of microbiologically mediated diseases such as cholera, typhoid fever, and gastroenteritis. The risks of disinfection byproducts should always be compared with the risks of the microbiological species that disinfection removes [273].

[Top of Page]

11. Acknowledgement

The entire topic was coordinated by Fritz H. Frimmel

[Top of Page]

(10.24)

(10.25)

(10.26)

(10.27)

References1. N. E. Dorsey: Properties of Ordinary Water Substance, Reinhold Publ. Co., New York 1968. 2. D. Eisenberg, W. Kauzmann: The Structure and Properties of Water, Oxford University Press, Oxford 1969. 3. F. Franks (ed.): Water. A Comprehensive Treatise, vol. 1, Plenum Press, New York 1972. 4. A. Rahman, F. H. Stillinger, J. Chem. Phys. 55 (1981) 3334. 5. E. U. Franck, K. Roth, Disscuss. Faraday Soc. 43 (1967) 108. 6. G. S. Kell in [1], chap. 7, p. 363. 7. V. A. Kirillin (ed.): Heavy Water, Thermophysical Properties, U.S. Department of Commerce, National Technical

Information Service, Springfield, Va., 1971. 8. L. Haar, J. S. Gallagher, G. S. Kell: “NBS/NRC Steam Tables: Thermodynamic and Transport Properties and

Computer Programs for Vapor and Liquid States of Water in SI Units,” Nat. Bur. Stand. (U.S.) Nat. Stand. Ref. Data Ser. 1984

9. U. Grigull, J. Straub, P. Schiebener: Steam Tables in SI-Units, 3rd ed., Springer Verlag, Berlin 1990. 10. F. Franks (ed.): Water. A Comprehensive Treatise, vol. 7: “Water and Aqueous Solutions at Subzero Temperatures,”

Plenum Press, New York 1982. 11. H. Weingärtner, Z. Phys. Chem. (Wiesbaden) 132 (1982) 129. 12. W. L. Marshall, E. U. Franck, J. Phys. Chem. Ref. Data 10 (1980) 295. 13. M. Uematsu, E. U. Franck, J. Phys. Chem. Ref. Data 9 (1980) 1291. Links 14. JANAF Thermochemical Tables, Nat. Stand. Ref. Data Ser. (U.S. Natl. Bur. Stand.) 37 (1971) 15. W. Gerrard: Gas Solubilities, Pergamon Press, Oxford 1980. 16. R. Fernandez-Prini, M. L. Japas, F. Corti: High Temperature Aqueous Solutions, CRC Press, Orlando 1992.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 73 of 82

Page 74: Water Ullmans

17. International Union of Pure and Applied Chemistry (ed.): Solubility Data Series, Pergamon Press, Oxford. 18. F. Franks (ed.): Water. A Comprehensive Treatise, vol. 4: “Aqueous Solutions of Amphiphiles and Macromolecules,”

Plenum Press, New York 1975. 19. C. Tanford: The Hydrophobic Effect, Wiley, New York 1973. 20. W. Kauzmann, Adv. Protein Chem. 14 (1959) 1. Links 21. H. S. Frank, M. W. Evans, J. Chem. Phys. 13 (1945) 507. Links 22. F. Franks (ed.): Water. A Comprehensive Treatise, vol. 2: “Water in Crystalline Hydrates and Aqueous Solutions of

Simple Nonelectrolytes,” Plenum Press, New York 1973. 23. J. S. Rowlinson, F. L. Swinton: Liquids and Liquid Mixtures, Butterworths, London 1984. 24. R. A. Robinson, R. H. Stokes: Electrolyte Solutons, Butterworths, London 1969. 25. J. E. Enderby, G. W. Neilson in F. Franks (ed.): Water. A Comprehensive Treatise, vol. 6: “Recent Advances,” Plenum

Press, New York 1979, chap. 1, p. 1. 26. G. Engel, H. G. Hertz, Ber. Bunsen Ges. Phys. Chem. 72 (1968) 808. 27. W. A. Luck, Ber. Bunsen Ges. Phys. Chem. 69 (1965) 69. 28. K. S. Pitzer in R. M. Pytkowicz (ed.): Activity Coefficients in Electrolyte Solutions, CRC Press, Boca Raton 1979. 29. C. W. F. T. Pistorius, E. Rapoport, J. B. Clark, J. Chem. Phys. 48 (1968) 5509 – 5514. Links 30. T. B. Bizhigitov, N. N. Sirota, JETP Lett. (Engl. Trans.) 44 (1986) 417 – 419. 31. N. N. Sirota, T. B. Bizhigitov, Russ. J. Phys. Chem. (Engl. Transl.) 62 (1988) 134 – 136. 32. Landolt-Börnstein, VI.4, pp. 16 – 18. 33. K. Tödheide, Ber. Bunsen. Ges. Phys. Chem. 70 (1966) 1022. 34. F. Franks: Water. A Comprehensive Treatise, vol. 1, Plenum Press, New York 1972. 35. N. E. Dorsey: Properties of Ordinary Water-Substance, Hafner Publ.Co., New York 1940. 36. O. V. Nagornov, V. E. Chizov, PMTF Zh. Prikl. Mekh. Tekh. Fiz. 1990, no. 3, 41 – 48. 37. Landolt-Börnstein, V.1b, pp. 482 – 507. 38. P. V. Hobbs: Ice Physics, Clarendon Press, Oxford 1974. 39. J. V. Leyendekkers, R. J. Hunter, J. Chem. Phys. 82 (1985) 1440 – 1446. Links 40. B. Bullemer et al., Solid State Commun. 6 (1968) 663. Links 41. A. Saul, W. Wagner, J. Phys. Chem. Ref. Data 18 (1989) 1537 – 1564. Links 42. D. Eisenberg, W. Kauzmann: The Structure and Properties of Water, Clarendon Press, Oxford 1969. 43. M. Christoforakos, E. U. Franck, Ber. Bunsen Ges. Phys. Chem. 90 (1986) 780 – 789. 44. U. Grigull, J. Straub, P. Schiebener: Steam Tables in SI-Units, 3rd ed., Springer Verlag, Berlin 1990. 45. U. Grigull: Properties of Water and Steam in SI-Units, 4th ed., Springer Verlag, Berlin 1989. 46. L. Haar, J. S. Gallagher, G. S. Kell: “NBS/NRC Steam Tables: Thermodynamic and Transport Properties and

Computer Programs for Vapor and Liquid States of Water in SI Units,” Nat. Stand. Ref. Data (U.S. Natl. Bur. Stand.) (1984).

47. R. A. Fine, F. J. Millero, J. Chem. Phys. 59 (1973) 5529 – 5536. Links 48. F. J. Dietz, J. J. de Groot, E. U. Franck, Ber. Bunsen Ges. Phys. Chem. 85 (1981) 1005 – 1009. 49. R. Tufeu, B. Le Neindre, Int. J. Thermophys. 8 (1987) 283 – 292. Links 50. K. H. Dudziak, E. U. Franck, Ber. Bunsen Ges. Phys. Chem. 70 (1966) 1120 – 1128. 51. W. J. Lamb, G. A. Hoffmann, J. Jonas, J. Chem. Phys. 74 (1981) 6875 – 6880. Links 52. W. L. Marhsall, E. U. Franck, J. Phys. Chem. Ref. Data 10 (1981) 295 – 304. 53. M. Uematsu, E. U. Franck, J. Phys. Chem. Ref. Data 9 (1980) 1291. Links 54. G. Wiegand, E. U. Franck, Ber. Bunsen Ges. Phys. Chem. 98 (1994) 809 – 817. 55. J. H. Dymond, E. B. Smith: The Virial Coefficients of Pure Gases and Mixtures. A Critical Compilation, Oxford

University Press, New York 1980. 56. E. U. Franck, J. Chem. Thermodynamics 19 (1987) 225. 57. R. W. Shaw et al., Chem. Eng. News (1991), December 23, 26 – 39. 58. L. A. Hütter: Wasser und Wasseruntersuchung, Salle + Sauerländer, Frankfurt 1994. 59. H. Hein, W. Kunze: Umweltanalytik mit Spektrometrie und Chromatographie. Von der Laborgestaltung bis zur

Dateninterpretation, VCH Verlagsgesellschaft, Weinheim, Germany 1994. 60. American Water Works Association (eds.): Standard Methods for the Examination of Water and Wastewater, Published

jointly by American Public Health Association (APHA), American Water Works Association (AWWA), Water Pollution Control Federation (WPCF), Washington 1994.

61. H. P. Hesse, F. Malz: “Die parameterspezifische Probenahme in der chemischen Abwasseranalytik,” Abwassertechnik 2 (1991) 21 – 26.

62. H. Gudernatsch: “Die Probenahme als wesentlicher Bestandteil der Wasser- und Abwasseranalytik,” Vom Wasser 60 (1983) 95 – 105.

63. C. Schöneborn: “Die Praxis der BSB-Bestimmung nach DIN 38 409-Teil 51,” Korresp. Abwasser 36 (1989) 424, 427 – 431.

64. M. Hahn et al.: “Eine neue Meßanordnung zur schnellen coulometrischen Bestimmung von extrahierbaren organischen Halogenverbindungen (EOX),” Z. Wasser Abwasser Forsch. 24 (1991) 191 –194.

65. G. Schwedt: Taschenatlas der Analytik, Thieme Verlag, Stuttgart 1996.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 74 of 82

Page 75: Water Ullmans

66. G. Schwedt: Analytische Chemie, Grundlagen, Methoden, Praxis, Thieme Verlag, Stuttgart 1995. 67. H. Berndt, U. Harms, M. Sonneborn: “Multielement-Spurenanreicherung aus Wässern an Aktivkohle zur

Probenvorbereitung für die Atomspektroskopie (Flammen-AAS, ICP/OES),” Fresenius Z. Anal. Chem. 322 (1985) 329 – 333.

68. M. Sperling, X. Yin, B. Welz: “Flow Injection On-line Separation and Preconcentration for Electrothermal Atomic Absorption Spectrometry,” Part 1: Determination of Ultratrace Amounts of Cadmium, Copper, Lead and Nickel in Water Samples,“ J. Anal. At. Spectrom. 6 (1991) 293 – 300.

69. R. Herzog, F. Dietz: “ICP-Massenspektrometrie — Erfahrungen mit einer neuen Analysenmethode in der Wasseruntersuchung,” Vom Wasser 73 (1989) 67 – 109.

70. M. Weidenauer, K. H. Lieser: “Bestimmung von Spurenelementen im Flußwasser mit Hilfe der Voltammetrie,” Fresenius Z. Anal. Chem. 320 (1985) 550 – 555.

71. F. H. Frimmel, T. Gremm: “The Importance of Element Speciation in Water Analysis — a Plea for Further Investigations,” Fresenius J. Anal. Chem. 350 (1994) 7 – 13. Links

72. K. Stein, G. Schwedt: “Speciation of Chromium in the Waste Water From a Tannery,” Fresenius J. Anal. Chem. 350 (1994) 38 – 43. Links

73. M. Sperling, Yin Yuefeng, B. Welz: “Differential Determination of Arsenic(III) and Total Arsenic Using Flow Injection On-line Separation and Preconcentration for Graphite Furnace Atomic Absorption Spectrometry,” Spectrochim. Acta Part B46 B (1991) 1798 – 1801. H. H. Dieter: “Vorkommen und Toxikologie von Arsen und seine gesundheitliche Bedeutung im Trinkwasser,” in K. Aurand, U. Hässelbarth, H. Lange-Asschenfeldt, W. Steuer (eds.): Die Trinkwasserverordnung. Einführung und Erläuterungen für Wasserversorgungsunternehmen und Überwachungsbehörden, 3rd revised ed., Erich Schmidt Verlag, Berlin 1991, pp. 154 – 171. F. W. Pontius, K. G. Brown, C.-J. Chen: “Health Implications of Arsenic in Drinking Water,” J. Am. Water Works Assoc. 86 (1994) no. 9, 52.

74. R. D. Wilken, J. Kuballa, E. Jantzen: “Organotins: Their Analysis and Assessment in the Elbe River System, Northern Germany,” Fresenius J. Anal. Chem. 350 (1994) 77 – 84. Links

75. J. A. C. Broekaert, S. Gücer, F. Adams (eds.): “Metal Speciation in the Environment,” NATO ASI Ser. G 23 (1990) 76. P. C. do Nascimento, G. Schwedt: “Comparative Studies of the Determination of Cyanide at Low Concentration Levels

in Waste Waters,” Anal. Chim. Acta 283 (1993) 755 – 761. Links 77. S. Hinkamp, G. Schwedt: “Online-Verfahren zur Bestimmung von löslichem Gesamt-Stickstoff in Abwässern durch

Kopplung eines UV-Aufschlusses mit der Fließinjektionsanalyse,” Z. Wasser Abwasser Forsch. 24 (1991) 60 – 65. 78. S. Hinkamp, G. Schwedt: “Determination of Total Phosphorus in Waters With Amperometric Detection by Coupling of

Flow-injection Analysis With Continuous Microwave Oven Digestion,” Anal. Chim. Acta 236 (1990) 345 – 350. Links 79. B. Rössner, J. Behnert, A. Kiplinger: “Die Ionenchromatographie in der routinemäßigen Wasser- und Umweltanalytik,”

Fresenius Z. Anal. Chem. 327 (1987) 698 – 700. 80. G. Schwedt, B. Kondratjonok: “Simultane Ionen-Chromatographie von Kationen und Anionen mit Komplexonen als

Elutionsmittel und selektiver UV-Detektion,” Fresenius Z. Anal. Chem. 332 (1989) 855 – 861. 81. G. Schwedt, B. Rössner: “Ionen-chromatographische Spurenanalyse amperometrisch detektierbarer Anionen in

Wässern,” Fresenius Z. Anal. Chem. 327 (1987) 499 – 502. 82. J. Harms, L. Dunemann, G. Schwedt: “Comparison Between Capillary Electrophoresis and IonChromatography for

Analysis of Inorganic Anions,” Mikrochim. Acta 108 (1992) 251 – 260. Links 83. A. Meyer, A. Kelling: “Schnelltestverfahren im Vorfeld der Umweltanalytik,” CLB Chem. Labor. Betr. 45 (1994) 414 –

417. 84. G. Schwedt: Chromatographische Trennmethoden, Thieme Verlag, Stuttgart 1993. 85. R. Götz: “Zur spektralphotometrischen Bestimmung von Ligninsulfonsäuren und Huminsäuren in Wässern,” Fresenius

Z. Anal. Chem. 296 (1979) 406 –407. 86. G. Abbt-Braun, F. H. Frimmel: “Spektroskopische Strukturaufklärung aquatischer Huminstoffe,” Vom Wasser 77 (1991)

292. 87. H. Hellmann: “Aufgabe und Einsatz der IR-Spektroskopie im Rahmen der Tensidanalytik in Abwasser und Gewässern,”

Vom Wasser 66 (1986) 111 – 125. 88. P. Belouschek et al.: “Bestimmung von chlorierten KW mit kombinierter Headspace- und GC/MS-Technik,” Vom

Wasser 79 (1992) 1 – 8. 89. H.-J. Neu, W. Ziemer, W. Merz: “Ein neuartiges Derivatisierungsverfahren zur Bestimmung von Spuren polarer

organischer Substanzen in Wasser mittels statischer Dampfraum-Gaschromatographie,” Fresenius J. Anal. Chem. 340 (1991) 65 – 70. Links

90. C. L. Arthur et al.: “Environmental Analysis of Organic Compounds in Water Using Solid Phase Microextraction,” J. High Resolut. Chromatogr. 15 (1992) 741 – 744.

91. C. L. Arthur et al.: “Solid Phase Microextraction for the Direct Analysis of Water: Theory and Practice,” LC-GC 10 (1992) 656, 658, 660 – 661.

92. H.-J. Brauch: “Gas Chromatography for Determination of Pesticides in Aquatic Systems — Analytical Basis,” Acta Hydrochim. Hydrobiol. 21 (1993) 84 – 88. Links

93. J. Brodesser, H. F. Schöler: “Die Analyse von polychlorierten Biphenylen in Wasseer im ng/l-Bereich,” Vom Wasser 72 (1989) 145 – 150.

94. L. Braunstein, K. Hochmüller, K. Spengler: “Analyse gaschromatographierbarer Wasserinhaltsstoffe durch Kopplung von GC und massenselektivem Detektor,” Vom Wasser 73 (1989) 167 – 189.

95. H. F. Schröder: “Hochdruck-Flüssigkeit-Chromatographie gekoppelt mit Tandemmassenspektrometrie — eine schnelle und zukunftsweisende Analysenmethode in der Wasser- und Abwasseranalytik,” Vom Wasser 73 (1989) 111 – 136.

96. R. Reupert, E. Plöger: “Bestimmung stickstoffhaltiger Pflanzenbehandlungsmittel im Trink-, Grund- und Oberflächenwasser, Analytik und Ergebnisse,” Vom Wasser 72 (1989) 211 – 233.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 75 of 82

Page 76: Water Ullmans

97. T. Gremm, F. H. Frimmel: “Systematische Untersuchung der PAK-Bestimmung mittels HPLC,” Vom Wasser 75 (1990) 171 – 182.

98. U. Obst, A. Holzapfel-Pschorn: Enzymatische Tests für die Wasseranalytik, Oldenbourg Verlag, München 1990. 99. B. Hock: “Immunchemische Verfahren in der Umweltanalytik und ihre Perspektiven,” Z. Umweltchem. Ökotox. 5 (1993)

309 – 319. 100. P. Herzsprung, L. Weil, K. E. Quentin, I. Zombola: “Bestimmung von Phosphorpestiziden und insektiziden Carbamaten

mittels Cholinesterasehemmung,” Vom Wasser 74 (1990) 339 – 350. 101. K. Stein, G. Schwedt: “Einsatz eines Acetylcholinesterase-Biosensors aus pH-Elektrode und Enzymmembran zum

Pestizid-Screening,” Vom Wasser 79 (1992) 211 – 224. 102. R. Kanne: “Aussagekraft ökotoxikologischer Testverfahren zur Beurteilung von Abwässern,” GWF Gas — Wasserfach:

Wasser/Abwasser 129 (1988) 394 – 397. 103. P. MacCarthy, R. W. Klusman, S. W. Cowling, J. A. Rice: “Water Analysis,” Anal. Chem. 65 (1993) 244R – 292R. Links104. A. Baumgartner, E. Reichel: Die Weltwasserbilanz, R. Oldenbourg Verlag, München – Wien 1975. 105. Umweltbundesamt (eds.): Daten zur Umwelt 1992/93, Erich Schmidt Verlag, Berlin 1994. 106. H. Bretschneider, K. Lecher, M. Schmidt (eds.): Taschenbuch der Wasserwirtschaft, Paul Parey, Hamburg – Berlin

1982. 107. U. de Haar (ed.): Hydrologischer Atlas der Bundesrepublik Deutschland im Auftrag der Deutschen

Forschungsgemeinschaft unter der Gesamtleitung von R. Keller, Boldt-Verlag, Boppard 1979. 108. Bundesverband der deutschen Gas- und Wasserwirtschaft e. V. (BGW) (eds.): 106. Wasserstatistik Bundesrepublik

Deutschland, Berichtsjahr 1994, Wirtschafts- und Verlagsgesellschaft Gas und Wasser, Bonn 1995. 109. Bundesverband der deutschen Gas- und Wasserwirtschaft e. V. (BGW) (eds.): 104. Wasserstatistik Bundesrepublik

Deutschland, Berichtsjahr 1992, Wirtschafts- und Verlagsgesellschaft Gas und Wasser, Bonn 1993. 110. R. Schleyer, H. Kerndorff: Die Grundwasserqualität westdeutscher Trinkwasserressourcen, VCH Verlagsgesellschaft,

Weinheim, Germany 1992. 111. St. P. Johnson, G. Corcelle: “Water,” The Environmental Policy of the European Communities, Graham & Trotman,

London 1989, pp. 25 – 108. 112. Council Directive of June 16, 1975 Concerning the Quality of Surface Water Intended for the Abstraction of Drinking

Water in the Member States (75/440/EEC), O. J. no. L 294, July 15, 1975. 113. Deutscher Verein des Gas- und Wasserfaches e. V. (DVGW): “Eignung von Oberflächenwasser als Rohstoff für die

Trinkwasserversorgung,” DVGW-Arbeitsblatt W 151, ZfGW-Verlag Frankfurt/Main, July 1975. 114. Deutscher Verein des Gas- und Wasserfaches e. V. (DVGW): “Eignung von Wasser aus Fließgewässern als Rohstoff

für die Trinkwasserversorgung,” DVGW-Merkblatt W 251, DVGW, Eschborn, June 1994. 115. Council Directive of July 15, 1980 Concerning the Quality of Water Intended for Human Consumption (80/778/EEC).

O. J. no. L 229, Aug. 30, 1980. Proposal for a Council Directive Concerning the Quality of Water Intended for Human Consumption (95/C 131/03). Presented by the Commission April 28, 1995. O. J. no. C 131, May 30, 1995.

116. Der Bundesminister für Jugend, Familie, Frauen und Gesundheit: Verordnung über Trinkwasser und über Wasser für Lebensmittelbetriebe (Trinkwasserverordnung — TrinkwV) vom 5.12.1990. BGBl. I, 2612 – 2629 (1990).

117. K. Aurand, U. Hässelbarth, H. Lange-Asschenfeldt, W. Steuer (eds.): Die Trinkwasserverordnung; Einführung und Erläuterungen für Wasserversorgungsunternehmen und Überwachungsbehörden, 3rd ed., Erich Schmidt Verlag, Berlin 1991.

118. Fachnormenausschuß Wasserwesen (FNW) im Deutschen Normenausschuß (DNA), DIN 2000: Zentrale Trinkwasserversorgung. Leitsätze für Anforderungen an Trinkwasser, Planung, Bau und Betrieb der Anlagen, November 1973, Beuth-Verlag, Berlin.

119. F. W. Pontius, J. Am. Water Works Assoc. (J. AWWA) 88 (1996) no. 3, 36 – 46. 120. S. D. Faust, O. M. Aly: Chemistry of Water Treatment, Butterworths, Boston 1983. 121. T. H. Y. Tebbutt: Principles of Water Quality Control, Pergamon Press, Oxford 1992. 122. Bundesminister des Inneren (eds.): Künstliche Grundwasseranreicherung, Stand der Technik und des Wissens in der

Bundesrepublik Deutschland. Bearbeitet vom BMI-Fachausschuß “Wasserversorgung und Uferfiltrat”. Erich Schmidt Verlag, Berlin 1985.

123. H. Sontheimer: “Trinkwasser aus dem Rhein?” Bericht über ein Verbundforschungsvorhaben zur Sicherheit der Trinkwassergewinnung aus Rheinuferfiltrat bei Stoßbelastungen, Academia Verlag, Sankt Augustin 1991.

124. Council Directive of December 17, 1979 Concerning the Protection of Groundwater Against Pollution by Certain Dangerous Substances (80/86/EEC). O. J. no. L 20, Jan. 26, 1980.

125. Der Bundesminister für Umwelt, Naturschutz und Reaktorsicherheit: Gesetz zur Ordnung des Wasserhaushalts (Wasserhaushaltsgesetz — WHG) vom 23.9.1986. BGBl. I, 1529 – 1544, ber. S. 1654 (1986), 6. Novelle (18. März 1994).

126. J. Mallevialle, I. H. Suffet, U. S. Chan (eds.): Influence and Removal of Organics in Drinking Water, Lewis Publ., Boca Raton 1992.

127. Rehabilitation of the River Rhine. Proceedings of the International Conference on Rehabilitation of the River Rhine March 15 – 19, 1993, Arnhem, The Netherlands, Water Sci. Technol. 29 (1994) no. 3, 1 – 394.

128. H. Bendlin: Reinstwasser von A bis Z, VCH Verlagsgesellschaft, Weinheim, Germany 1995. 129. H.-D. Held: Kühlwasser, Verfahren und Systeme der Aufbereitung, Behandlung und Kühlung von Süßwasser,

Brackwasser, Meerwasser in der Industrie, neubearbeitet von H. G. Schnell, 4th ed., Vulkan-Verlag, Essen 1995. 130. H. von Kienle, E. Bäder: Aktivkohle und ihre industrielle Anwendung, Enke Verlag, Stuttgart 1980. 131. H. Sontheimer, J. C. Crittenden, R. S. Summers: Activated Carbon for Water Treatment, 2nd ed., DVGW-

Forschungsstelle, Karlsruhe 1988.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 76 of 82

Page 77: Water Ullmans

132. R. M. Clark, B. W. Lykins: Granular Activated Carbon: Design, Operation and Cost, Lewis Publ., Michigan 1989. 133. R. Kümmel, E. Worch: Adsorption aus wäßrigen Lösungen, Deutscher Verlag für Grundstoffindustrie, Leipzig 1990. 134. V. L. Snoeyink in: Water Quality and Treatment: A Handbook of Community Water Supplies, McGraw Hill, New York

1990. 135. I. Langmuir, J. Am. Chem. Soc. 40 (1918) 1361 –1403. Links 136. H. Freundlich, Z. Phys. Chem. A 57 (1906) 385 – 470. 137. R. A. Dobbs, J. M. Cohen, U.S. Environ. Prot. Agency, Report EPA-600/8–80–023 (1980). 138. T. F. Speth, R. J. Miltner, J. Am. Water Works Assoc. 82 (1990) 72 – 75. 139. A. L. Myers, J. M. Prausnitz, Am. Inst. Chem. Eng. J. 11 (1965) no. 1, 121 – 127. 140. C. J. Radke, J. M. Prausnitz, Am. Inst. Chem. Eng. J. 18 (1972) no. 4, 761 – 768. 141. J. C. Crittenden et al., Environ. Sci. Technol. 19 (1985) no. 11, 1037 – 1043. 142. J. C. Crittenden, P. Luft, D. W. Hand, Water Res. 19 (1985) no. 12, 1537 – 1548. Links 143. D. W. Hand et al., J. Am. Water Works Assoc. 81 (1989) no. 1, 67 – 77. 144. D. R. Yonge, T. M. Keinath, J. Water. Pollut. Control. Fed. 58 (1986) no. 1, 77 – 81. 145. P. C. Singer, C.-Y. Yen in I. H. Suffet, M. J. McGuire (eds.): Activated Carbon Adsorption of Organics from the

Aqueous Phase, vol. I, Ann Arbor Science, Ann Arbor, Michigan 1980. 146. J. Fettig, H. Sontheimer, J. Environ. Eng. (N.Y.) 113 (1987) no. 4, 780 – 794. 147. B. Haist-Gulde: “Zur Adsorption von Spurenverunreinigungen aus Oberflächenwässern,” PhD Thesis, University

Karlsruhe (1991). 148. H. Sontheimer et al., Vom Wasser 75 (1990) 183 – 200. 149. B. Haist-Gulde et al., GWF Gas Wasserfach: Wasser/Abwasser 132 (1991) no. 1, 8 – 15. 150. I. N. Najm, V. L. Snoeyink, Y. Richard, J. Am. Water Works Assoc. 83 (1991) no. 8, 57 – 63. 151. E. Worch, Chem. Tech. (Leipzig) 43 (1991) no. 3, 111 – 114. 152. E. Worch, Chem. Tech. (Leipzig) 43 (1991) no. 6, 221 – 224. 153. R. Ford et al., Proc. Annu. Conf. Am. Water Works Assoc. June 18 – 22 (1989). 154. T. F. Speth, R. J. Miltner, J. Am. Water Works Assoc. 81 (1989) no. 4, 141 – 148. 155. R. S. Summers et al.: Standardized Protocol for the Evaluation of GAC, Amer. Water Works Assoc. Research Found.,

Denver, CO, 1992. 156. J. C. Crittenden, W. J. Weber, J. Environ. Eng. Div. (Am. Soc. Civ. Eng.) 104 (EE3) (1978) 433 – 443. 157. J. C. Crittenden, W. J. Weber, J. Environ. Eng. Div. (Am. Soc. Civ. Eng.) 104 (EE6) (1978) 1175 – 1195. 158. J. C. Crittenden, J. K. Berrigan, D. W. Hand, J. Water. Pollut. Control. Fed. 58 (1986) no. 4, 312 –319. 159. J. C. Crittenden et al., J. Environ. Eng. (N.Y.) 113 (1987) no. 2, 243 – 259. 160. J. C. Crittenden et al., J. Am. Water Works Assoc. 83 (1991) no. 1, 77 – 87. 161. L. Cummings, R. S. Summers, J. Am. Water Works Assoc. 86 (1994) 6. 162. R. Kunin: Ion Exchange Resins, J. Wiley & Sons, New York 1958. 163. S. B. Applebaum: Demineralization by Ion Exchange, Academic Press, New York – London 1968. 164. F. Helfferich: Ion Exchangers, Mc Graw-Hill, New York 1962. 165. K. Dorfner: Ion Exchangers, De Gruyter, Berlin 1991. 166. M. J. Slater: The Principles of Ion Exchange Technology, Butterworth-Heinemann Ltd., Oxford 1991. 167. B. A. Bolto, L. Pawlowski: Wastewater Treatment by Ion Exchange, E. & F. Spon, London – New York 1987. 168. T. Shedlowsky, D. A. MacInnes: “The First Dissociation Constant of of Carbonic Acid, 0 to 38 °C from

Conductivity Measurements,” J. Am. Chem. Soc. 57 (1935) 1705. 169. H. S. Harned, S. R. Sholes: “The Ionization Constant of from 0 to 50 °C,” J. Am. Chem. Soc. 63 (1941) 2030. 170. H. Sontheimer, P. Spindler, U. Rohmann: Wasserchemie für Ingenieure, ZfGW-Verlag, Frankfurt/Main 1980. 171. S. H. Eberle: “A Correctness Test of the Computation of Chemical Speciation for the Main Constituents of Natural

Waters,” Water Res. 23 (1989) 1373. Links 172. W. Höll: Der Einfluß von Ionenaustauschvorgängen auf das Gleichgewichtssystem der Kohlensäure im Wasser, KfK-

Report 4367, Kernforschungszentrum, Karlsruhe 1987. 173. J. Tillmans, O. Heublein: “Über die kohlensauren Kalk angreifende Kohlensäure natürlicher Wässer,” Gesund. Ing. 35

(1912) 669. 174. Degrémont: Water Treatment Handbook, 6th ed., Lavoisier Publ., Paris 1991, p. 262. 175. D. H. Caldwell, W. B. Lawrence: “Water Softening and Conditioning Problems,” Ind. Eng. Chem. 45 (1935) 535. 176. K. Dorfner: Ion Exchangers, De Gruyter, Berlin 1991, chap. 2.2. 177. R. Kunin: Amber-Hi-Lites 95 (1966); 96 (1966);97 (1967). 178. LEWATIT-LEWASORB-Handbook, Bayer AG, Leverkusen 1972. 179. K. R. Gray, H. L. Crosby, US 2 656 245, 1953. 180. R. Kunin: “Regeneration of Carboxylic Cation Exchange Resins With Carbon Dioxide,” Ind. Eng. Chem. Prod. Res.

Dev. 2 (1963) 1. 181. C. Berger-Wittmar: “Regeneration schwach saurer Ionenaustauscher mit Kohlendioxid,” Vom Wasser 47 (1976) 297. 182. in Ref. [176] , chap. 2.4. 183. D. A. Clifford, W. J. Weber: “Nitrate Removal From Water Supplies,” EPA Report 600/8 77 015 (1977) R-31,

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 77 of 82

Page 78: Water Ullmans

Cincinnatti, Ohio. 184. A. K. Sengupta, D. A. Clifford: “Chromate Ion Exchange Mechanism for Cooling Water,” Ind. Eng. Chem. Fundam. 25

(1988) 249. 185. D. A. Clifford: “Processes for Removal of Inorganic Contaminants from Water,” Water Eng. Manage. (1982) R-31. 186. G. A. Guter, US 4 479 877, 1980. 187. F. X. McGarvey, B. Bachs, S. M. Ziarkowski: “Removal of Nitrates from Natural Water Supplies,” paper presented at

American Chemical Society Meeting, Dallas, Texas, 1989. 188. J. P. v. d. Hoek, A. Klapwijk: “Nitrate Removal From Ground Water,” Water Res. 21 (1987) 989. 189. CHRIST AG, Technical Information TI-A-203. 190. R. Kunin, B. Vassiliou: “New Dionization Techniques Based Upon Weak Electrolyte Ion Exchange Resins,” Ind. Eng.

Chem. Process Des. Dev. 3 (1964) 404; Further Studies on the Weak Electrolyte Ion Exchange Resin Desalination Process (DESAL Process), Desalination 4 (1968) 38.

191. Kernforschungszentrum Karlsruhe, EP 81 109 498.6, 1984 (W. H. Höll, B. Kiehling). 192. W. H. Höll, B. Kiehling: “Regeneration eines Ionenaustauschermischbetts mit CO2 zur Teilentsalzung von

Trinkwasser,” Vom Wasser 59 (1982) 207. 193. W. H. Höll, J. Horst: “Simultane Elimination von Schwermetallen und Komplexbildnern aus Wässern mit

Anionenaustauschern in Hydrogencarbonatform,” Vom Wasser 74 (1990) 51. 194. D. E. Weiss et al.: “The Sirotherm Demineralization Process — An Ion Exchange Process with Thermal Regeneration,”

part I, J. Inst. Eng. Aust. 37 (1966) 193. 195. D. E. Weiss: “Pilot Plant Studies of Partial Demineralization of Brackish Waters by the ‘Sirotherm' Process,” Ion Exch.

Membr. 1 (1972) 109. 196. L. Sigg, W. Stumm: Aquatische Chemie, 3rd ed., B. G. Teubner Verlag, Stuttgart 1994. 197. Committee Report, J. Am. Water Works Assoc. 71 (1979) 588 – 603. 198. K. J. Ives: The Scientific Basis of Flocculation, Sijthoff & Noordhoff, Alphen aan den Rijn 1978. 199. J. Bratby: Coagulation and Flocculation, Uplands Press, Croydon 1980. 200. M. Jekel in DVGW (ed.): “Wasseraufbereitung für Ingenieure,” 2nd ed., DVGW-Schriftenr. Wasser 206 (1987). 201. M. Jekel, R. Ließfeld (eds.): “Die Flockung in der Wasseraufbereitung,” DVGW-Schriftenr. Wasser 42 (1985) 202. C. F. Baes, R. E. Mesmer: The Hydrolysis of Cations, J. Wiley & Sons, New York 1976. 203. M. Jekel, J. Water SRT-Aqua 40 (1991) 18 – 24. 204. A. Amirtharajah, K. M. Mills, J. Am. Water Works Assoc. 74 (1982) 210 – 216. 205. J. K. Edzwald, AICHE Symp. Ser. Water. 75 (1979) 54 – 62. 206. B. A. Dempsey, R. M. Ganko, C. R. O'Melia, J. Am. Water Works Assoc. 76 (1984) 141 – 150. 207. DIN:Standard 19 622, Polyacrylamides for Water Treatment, Beuth Verlag, Berlin – Köln 1977. 208. J. Gregory, J. Colloid Interface Sci. 55 (1976) 35 – 44. Links 209. C. R. O'Melia in W. J. Weber, Jr. (ed.): Physico-chemical Processes for Water Quality Control, Wiley-Interscience, New

York 1972. 210. D. S. Parker, W. J. Kaufman, D. Jenkins, J. Sanit. Eng. Div. Am. Soc. Civ. Eng. 98 (1972) 79 – 99. 211. R. Klute in [212] , 53 – 65. 212. A. Grohmann, H. H. Hahn, R. Klute: Chemical Water and Wastewater Treatment, G. Fischer Verlag, Stuttgart – New

York 1985, pp. 113 – 131. 213. G. Smethurst: Basic Water Treatment for Application World-wide, Thomas Telford Ltd., London 1979. 214. C. R. Schulz, D. A. Okun, J. Am. Water Works Assoc. 75 (1983) 212 – 219. 215. Degrémont: Water Treatment Handbook, 5th ed., J. Wiley & Sons, New York 1979. 216. N. Tambo, Y. Watanabe, Water Res. 13 (1979) 429 –448. Links 217. P. Wölfel in [201], 173 – 190. 218. H. Bernhardt, H. Schell, Z. Wasser Abwasser Forsch. 12 (1979) 123 – 133. 219. A. J. Rees, D. J. Rodman, T. F. Zabel, Aqua 8 (1980) 170 – 182. 220. M. Jekel, Ozone Sci. Eng. 16 (1994) 55 – 66. 221. H. E. Hudson, E. G. Wagner, Proc. Water Qual. Technol. Conf. 2A-3 (1979) 222. H. H. Hahn, R. Klute, Z. Wasser Abwasser Forsch. 12 (1979) 111 – 119. 223. M. Jekel, R. Gimbel (eds.): “Optimal Dosing of Coagulants and Flocculants,” Proceedings of the IWSA-IAWG Joint

Specialist Group on Coagulation, Flocculation, Filtration, Sedimentation and Flotation in Water and Wastewater Treatment Workshop, Jan. 12 – 13, 1994, Mülheim an der Ruhr, Germany. Berichte aus dem Rheinisch-Westfälischen Institut für Wasserchemie und Wassertechnologie GmbH, Institut an der Gerhard-Mercator-Universität Duisburg — Gesamthochschule Duisburg, vol. 10, 1994.

224. K. H. Schmidt, “Biologische Verfahren,” DVGW-Schriftenr. Wasser 206 (1983) 11.1 – 11.42. 225. Degrémont: Water Treatment Handbook, 5th ed., Halsted Press (Wiley and Sons), New York 1979. 226. N. J. D. Graham: Slow Sand Filtration: Recent Developments in Water Treatment Technology, Ellis Horwood,

Chichester 1988. 227. S. Vigneswaran, R. Ben Aim: Water, Wastewater, and Sludge Filtration, CRC Press, Boca Raton, FL, 1989. 228. R. Gimbel: “Abscheidung von Trübstoffen aus Flüssigkeiten in Tiefenfiltern,” Veröffentlichungen des Bereichs und des

Lehrstuhls für Wasserchemie am Engler-Bunte-Institut der Universität Karlsruhe, Heft 25, ZfGW-Verlag, Frankfurt 1984.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 78 of 82

Page 79: Water Ullmans

229. P. A. Schweitzer (ed.): Handbook of Separation Techniques for Chemical Engineers, McGraw-Hill, New York 1979. 230. L. Svarovsky: Solid-Liquid Separation, 2nd ed., Butterworths, London 1981. 231. Degrémont: Water Treatment Handbook, vols. 1 + 2, Lavoisier Publ., Paris 1991. 232. K. W. Roennefahrt: “Filtration,” DVGW-Schriftenr. 206 (1983) 6–1 – 6–34. 233. R. Gimbel: “Theoretical Approach to Deep Bed Filtration,” in: Water, Wastewater, and Sludge Filtration, CRC Press,

Boca Raton, FL, 1989. 234. C. Tien: Granular Filtration of Aerosols and Hydrosols, Butterworth Publish., Boston 1989. 235. T. Iwasaki: “Some Notes on Sand Filtration,” J. Am. Water Works Assoc. 29 (1937) 1591 ff. 236. J. P. Herzig, D. M. Leclerc, P. Le Goff: “Flow of Suspension through Porous Media – Application to Deep Filtration,”

Ind. Eng. Chem. 62 (1970) no. 5, 8 ff. 237. R. Rautenbach, R. Mellis, W. Dahm, J. St. Kollbach; “Membranverfahren zur Aufarbeitung organisch/anorganisch

hochbelasteter Abwässer — Kombinationen mit anderen Verfahrensstufen wie Biologie und Eindampfung,” Berichte der 27. Essener Tagung für Wasser und Abfallwirtschaft 1994.

238. A. Gröschl: “Umkehrosmose organisch/wäßriger Systeme — Stofftransport in Membranen und Verfahrensentwicklung,”PhD Thesis, RWTH Aachen 1991.

239. G. Schneider: “Trennverhalten von Nanofiltrationsmembranen,” PhD Thesis, RWTH Aachen 1993. 240. E. N. Sieder, G. E. Tate: “Heat Transfer and Pressure Drop of Liquids in Tubes,” Ind. Eng. Chem. 28 (1936) 1429. 241. W. H. Linton, T. K. Sherwood, Chem. Eng. Prog. 46 (1950) 258. Links 242. M. Mulder: Basic Principles of Membrane Technology, Kluwer Academic Publ., Dordrecht, 1991. 243. G. Menges: Werkstoffkunde Kunststoffe, 3rd ed., Hanser Verlag, München, Wien 1990. 244. E. Staude: Membranen und Membranprozesse, VCH, Weinheim, Germany 1992. 245. R. R. Bhave: Inorganic Membranes — Synthesis, Characteristics and Applications, Van Nostrand Reinhold, New York

1991. 246. Fachgruppe Wasserchemie in der Gesellschaft Deutscher Chemiker: Deutsche Einheitsverfahren zur Wasser-,

Abwasser- und Schlammuntersuchung. Insbesondere: Verfahren der Gruppe C: Physikalische und physikalisch-chemische Kenngrößen (DIN 38 404); Verfahren der Gruppe D: Anionen (DIN 38 405); Verfahren der Gruppe E: Kationen (DIN 38 406); Verfahren der Gruppe G: Gasförmige Bestandteile (DIN 38 408); Verfahren der Gruppe H: Summarische Wirkungs- und Stoffkenngrößen (DIN 38 409).

247. W. Liebig: “Grundlagen der biologischen Abwasserreinigung am Beispiel einer Anlage für hochbelastete Abwässer eines Chemiewerkes,” ChED Chem. Exp. Didakt. 1 (1975) no. 1, 239 – 246.

248. P. Schuler, R. Degner: Kleines Handbuch über die photometrische CSB-Bestimmung und Analysen von Wasserinhaltsstoffen, Wissenschaftl. Techn. Werkstätten GmbH.

249. Permasep Products, Engineering Manual 1982. 250. A. Grohmann: “pH-Wert und Calcitsättigung des Wassers,” in K. Aurand, U. Hässelbarth, H. Lange-Asschenfeldt, W.

Steuer (eds.): Die Trinkwasserverordnung, 3rd ed., Erich Schmidt Verlag, Berlin 1991, pp. 338 – 348. 251. H. E. Hömig: Seawater and Seawater Distillation, Vulkan-Verlag, Esssen 1978. 252. DIN 50 930, Korrosion der Metalle, Beuth Verlag, Berlin, Feb. 1993. 253. Babcock: Handbuch Wasser, 6th ed., Vulkan-Verlag, Essen 1982. 254. E. Heymann: “Filtration”, in DVGW Deutscher Verein des Gas- und Wasserfaches e. V. (eds.): “DVGW-

Fortbildungskurse Wasserversorgungstechnik für Ingenieure und Naturwissenschaftler, Kurs 6: Wasseraufbereitungstechnik für Ingenieure,” 3rd ed., DVGW-Schriftenr. Wasser 206 (1987) 5–1 – 5–21.

255. Ullmann, 4th ed., 11, 581 – 586. 256. E. Müller: Mechanische Trennverfahren, vol. I, Grundzüge der Verfahrenstechnk, Otto Salle Verlag/Verlag

Sauerländer, Frankfurt/Main 1980. 257. K. Marquardt: “Flocculation, Precipitation, Sedimentation and Floatation for Use as Pretreatment Stages for Brackish

Water and Sea Water in Desalination Plants,” in: H. G. Heitmann (ed.): Saline Water Processing, VCH Verlagsgesellschaft, Weinheim 1990, pp. 121 – 133.

258. R. Rautenbach, R. Albrecht: Membrantrennverfahren — Ultrafiltration und Umkehrosmose, Otto Salle Verlag/Sauerländer Verlag, Frankfurt/Main 1981.

259. Gesetz zur Ordnung des Wasserhaushalts (Wasserhaushaltsgesetz — WHG) vom 23. September 1986. BGBl. I, p. 1529 ff., Berichtigung p. 1654.sowieAllgemeine Rahmen-Verwaltungsvorschrift über Mindestanforderungen an das Einleiten von Abwasser in Gewässer (Rahmen-Abwasser-VwV — R-Abw.Vwv) vom 8. September 1990. Gemeinsames Ministerialblatt, Nr. 25, vom 22.9.1989, pp. 518 –520. Geändert am 29. Oktober 1992, Gemeinsames Ministerialblatt, Nr. 42, November, 1992, pp. 1065 –1066.

260. R. Mellis: “Zur Optimierung einer biologischen Abwasserreinigungsanlage mittels Nanofiltration,” Dissertation, RWTH Aachen 1994.

261. R. Rautenbach: Umkehrosmose auf der Deponie Schönberg — Betriebsergebnisse 1992/93, Interner Bericht, RWTH Aachen 1994.

262. T. Peters: “Reinigung von Deponie-Sickerwasser mit druckgestufter Umkehrosmose in DT-Modultechnik und fraktionierter Reststoff-Ausschleu-sung,” Preprints zum Aachener Membran Kolloquium, Aachen 1993, GVC-VDI — Gesellschaft Verfahrenstechnik und Chemieingenieurwesen, Düsseldorf 1993.

263. R. Rautenbach, T. Linn: “Aufarbeitung von Konzentraten aus der Deponiesickerwasser-Behandlung mittels Nanofiltration,” Entsorgungspraxis 9 (1995) 44 – 48.

General References264. J. M. Montgomery: Water Treatment Principles and Design, J. Wiley & Sons, New York 1985. F. M. M. Morel, J. G.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 79 of 82

Page 80: Water Ullmans

[Top of Page]

Symbols and Abbreviations

AAS: atomic absorption spectrometry

AMD: automated multiple development

AOP: advanced oxidation process

AOX: adsorbable organic halogens

BAC: biologically activated carbon

BiAS: bismuth-active substance

BV: throughput

CUR: carbon use ratio

DBP: disinfection byproduct

DOC: dissolved organic carbon

DOCl: dissolved organic chlorine

DOX: dissolved organic halogens

EBCT: empty-bed contact time

ECD: electron capture detector

EDTA: ethylenediaminetetraacetic acid

ELISA: enzyme-linked immunosorbent assay

EOX: extractable organic halogens

FIA: flow injection analysis

FID: flame ionization detector

FM: feed side of the membrane

G: mean shear gradient

GAC: granular activated carbon

I: ionic strength

IAS: ideal adsorbed solution

IC: ion chromatography

Hering: Principles and Applications of Aquatic Chemistry, Wiley-Interscience, New York 1993. C. H. Tate, K. F. Arnold: “Health and Aesthetic Aspects of Water Quality,” Water Quality and Treatment, 4th ed., American Water Works Association (AWWA) 1990, 63 – 156.

Specific References265. W. M. Latimer: The Oxidation States of the Elements and their Potentials in Aqueous Solution, 2nd ed., Prentice-Hall,

Englewood Cliffs, NJ, 1952. 266. A. J. Bard, R. Parsons, J. Jordan (eds.): Standard Potentials in Aqueous Solution, Marcel Dekker, New York 1985. 267. J. Hoigné, H. Bader: “Rate Constants of Reactions of Ozone with Organic and Inorganic Compounds in Water,” Water

Res. 17 (1983) 185. 268. J. C. Morris in S. D. Faust, J. V. Hunter (eds.): Principles and Application of Water Chemistry, J. Wiley and Sons, New

York 1967. 269. G. C. White: The Handbook of Chlorination, 2nd ed., Van Nostrand Reinhold, New York 1986. 270. W. Glaze : Ozone in Drinking Water Treatment, Environ. Sci. Technol. 21 (1987) 224 – 230. 271. J. J. Rook: Formation of Haloforms During Chlorination of NaturalWaters, Water Treat. Exam. 23 (1974) 234 – 243. 272. W. H. Glaze: “Reaction Products of Ozone: A Review,” Environ. Health Perspect. 69 (1986) 151 – 157. Links 273. American Water Works Association Research Foundation: Drinking Water and Health in the Year 2000, American

Water Works Association (AWWA), Denver, CO, 1993.

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 80 of 82

Page 81: Water Ullmans

ICP: inductively coupled plasma

Kb: ebullioscopic constant

Kf: cryoscopic constant

KF: Freundlich constant

l: length of the pipe

L: volume of solution (adsorption)

: mass flow

m alkalinity: methyl orange alkalinity

MBAS: methylene blue-active substance

MSD: mass spectrometric detector

n: Freundlich exponent (Chap. Adsorption Processes in Water Treatment); impeller speed (Chap. Ion Exchange)

: molar flow per unit area

N: particle concentration

NF: nanofiltration

NOM: natural organic matter

NTA: nitrilotriacetic acid

OES: optical atomic emission spectrometry

P: power input of stirrer

p alkalinity: phenolphthalein alkalinity

PAC: powdered activated carbon

PAH: polycyclic aromatic hydrocarbon

PCB: polychlorinated biphenyl

PFR: plug-flow reactor

PM: permeate side of the membrane

PND: phosphorus – nitrogen detector

POC: particulate (undissolved) organic carbon

POX: purgeable organic halogens

PSA: potentiometric stripping analysis

q: concentration in the solid phase (adsorption)

R: rejection

Re: Reynolds number

RO: reverse osmosis

Sc: Schmidt number

Sh: Sherwood number

SI: saturation index

TAC: titre alcalemetrique complet (total alkalinity)

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 81 of 82

Page 82: Water Ullmans

About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2007John Wiley & Sons, Inc. All Rights Reserved.

TCD: thermoionic detector

TDS: total dissolved solids

THM: trihalomethane

TIC: total inorganic carbon

TOC: total organic carbon

TOCl: total organic chlorine

TOX: total organic halogens

UF: ultrafiltration

: volumetric flow rate

VOC: volatile organic carbon

VOX: volatile organic halogens

w: concentration

z: filter layer depth, membrane thickness

Greek symbols: collision efficiency factor

: filtration coefficient, filter load

: equivalent factor, volume

: concentration of solids

[Top of Page]

Water : Ullmann's Encyclopedia of Industrial Chemistry : Wiley InterScience

page 82 of 82