Download Chemical Processes at the Water-Manganite (γ-MnOOH

Chemical Processes at the
Water-Manganite (γγ-MnOOH)
Interface
by
Madeleine Ramstedt
Akademisk avhandling
Som med tillstånd av rektorsämbetet vid Umeå Universitet för erhållande av
Filosofie Doktorsexamen framlägges till offentlig granskning vid Kemiska institutionen,
sal KB3B1, KBC-huset, onsdagen den 19 maj 2004, kl 13.00.
Fakultetsopponent: Professor Steven A. Banwart, Groundwater Protection and Restoration
Group, Department of Civil & Structural Engineering, University of Sheffield, England
i
TITLE
Chemical Processes at the Water-Manganite (γ-MnOOH) Interface
AUTHOR
Madeleine Ramstedt
ADDRESS
Department of Chemistry; Inorganic Chemistry, Umeå University,
SE-901 87 Umeå, Sweden
ABSTRACT
The chemistry of mineral surfaces is of great importance in many different areas
including natural processes occurring in oceans, rivers, lakes and soils. Manganese
(hydr)oxides are one important group to these natural processes, and the thermodynamically
most stable trivalent manganese (hydr)oxide, manganite (γ-MnOOH), is studied in this thesis.
This thesis summarises six papers in which the surface chemistry of synthetic manganite
has been investigated with respect to surface acid-base properties, dissolution, and adsorption
of Cd(II) and the herbicide N-(phosphonomethyl)glycine (glyphosate, PMG). In these papers,
a wide range of analysis techniques were used, including X-ray photoelectron spectroscopy
(XPS), extended X-ray absorption fine structure (EXAFS) spectroscopy, Fourier transform
infra-red (FTIR) spectroscopy, atomic force microscopy (AFM), scanning electron
microscopy (SEM), X-ray diffraction (XRD), potentiometry, electrophoretic mobility
measurements and wet chemical techniques, in order to obtain a more complete understanding
of the different processes occurring at the manganite-water interface.
From the combined use of these techniques, a 1-pKa acid-base model was established
that is valid at pH>6. The model includes a Na+ interaction with the surface:
=MnOH2+½
log β0 (intr.) = -8.20 = -pHiep
=MnOH-½ + H+
=MnOHNa+½ + H+
log β0 (intr.) = -9.64
=MnOH2+½ + Na+
At pH<6 the manganite crystals dissolve and disproportionate into pyrolusite (β-MnO2) and
Mn(II)-ions in solution according to:
2 γ-MnOOH + 2H+ β-MnO2 + Mn2+ + 2H2O
log K0 = 7.61 ± 0.10
The adsorption and co-adsorption of Cd(II) and glyphosate at the manganite surface was
studied at pH>6. Cd(II) adsorption displays an adsorption edge at pH~8.5. Glyphosate adsorbs
over the entire pH range, but the adsorption decreases with increasing pH. When the two
substances are co-adsorbed, the adsorption of Cd(II) is increased at low pH but decreased at
high pH. The adsorption of glyphosate is increased in the entire pH range in the presence of
Cd(II). From XPS, FTIR and EXAFS it was found that glyphosate and Cd(II) form inner
sphere complexes. The binary Cd(II)-surface complex is bonded by edge sharing of Mn and
Cd octahedra on the (010) plane of manganite. Glyphosate forms inner-sphere complexes
through an interaction between the phosphonate group and the manganite surface. The largest
fraction of this binary glyphosate complex is protonated throughout the pH range. A ternary
surface complex is also present, and its structure is explained as type B ternary surface
complex (surface-glyphosate-Cd(II)). The chelating rings between the Cd(II) and glyphosate,
found in aqueous complexes, are maintained at the surface, and the ternary complex is bound
to the surface through the phosphonate group of the ligand.
KEYWORDS
manganite, γ-MnOOH, mineral surface, acid-base, adsorption, Cd(II),
N-phosphonomethylglycine, glyphosate, PMG, surface complex,
dissolution, disproportionation, XPS, EXAFS, infrared spectroscopy,
SEM, AFM, potentiometry, electrophoresis
ISBN
91-7305-634-0
74 pages and 6 papers
ii
How can I say thanks
for the things you have done for me?
Things so undeserved, that you gave
to prove your love for me.
The voices of a million angels
could not express my gratitude.
All that I am and ever hope to be,
I owe it all to Thee…
(Andrae Crouch)
iii
Printed by:
Print & Media, Umeå University,
Umeå, Sweden
ISBN
91-7305-634-0
iv
Chemical Processes at the Water-Manganite (γ- MnOOH) Interface
Madeleine Ramstedt
Department of Chemistry, Inorganic chemistry
Umeå University
SE-901 87 Umeå
SWEDEN
This thesis contains a summary and a discussion of the following papers, referred to in the
text by their numbers 1-6.
1.
Characterization of hydrous manganite (γ-MnOOH) surfaces – an XPS study.
Madeleine Ramstedt, Andrei Shchukarev, Staffan Sjöberg,
Surface and Interface Analysis, 34, 632-636, 2002
2.
Surface Properties of Hydrous Manganite (γ-MnOOH) – A potentiometric, electroacoustic and
XPS study.
Madeleine Ramstedt, Britt M Andersson, Andrei Shchukarev, Staffan Sjöberg,
Manuscript submitted to Langmuir
3.
Low pH Phase Transformation and Proton Promoted Dissolution of Hydrous Manganite
(γ-MnOOH).
Madeleine Ramstedt, Staffan Sjöberg,
Manuscript
4.
Chemical speciation of N-(phosphonomethyl)glycine, PMG, in solution and at mineral surfaces.
Madeleine Ramstedt, Caroline Norgren, Julia Sheals, Andrei Shchukarev, Staffan Sjöberg,
Surface and Interface analysis, In Press, 2004
5.
Thermodynamic and Spectroscopic Studies of Cadmium(II) – N- (phosphonomethyl)glycine
(PMG) complexes.
Madeleine Ramstedt, Caroline Norgren, Julia Sheals, Dan Boström, Staffan Sjöberg, Per
Persson,
Inorganica Chimica Acta, 357, 1185-1192, 2004
6.
Co-adsorption of Cadmium(II) and Glyphosate at the Water-Manganite (γ-MnOOH) Interface.
Madeleine Ramstedt, Caroline Norgren, Staffan Sjöberg, Per Persson,
Manuscript
Publications of interest but not included in the thesis:
•
A study of the Aqueous Geochemistry in the Udden Pit lake – Complete Dataset.
Madeleine Ramstedt, Erik Carlsson, Lars Lövgren,
MiMi-Report 2002:1, Stockholm, 2002, (ISSN 1403-9478, ISBN 91-89350-19-7)
•
Aqueous geochemistry in the Udden pit lake, northern Sweden.
Madeleine Ramstedt, Erik Carlsson, Lars Lövgren,
Applied Geochemistry, 18, 97-108, 2003
v
Index
Populärvetenskaplig sammanfattning.............................................................................................. 1
Introduction .................................................................................................................................. 3
Aim of the Thesis .......................................................................................................................... 4
Manganese - General Chemistry .................................................................................................... 5
Manganese - Geochemistry ........................................................................................................... 7
Manganese - Geochemical Cycle........................................................................................... 8
Manganese Oxides - Structures ..................................................................................................... 9
Manganite (γ-MnOOH) ................................................................................................................ 12
Surface Structure ................................................................................................................. 13
Mineral Synthesis................................................................................................................. 15
Methods ..................................................................................................................................... 17
Spectroscopy and Diffraction .............................................................................................. 17
X-ray Photoelectron Spectroscopy (XPS) ....................................................................... 18
X-ray Absorption Spectroscopy....................................................................................... 29
Infrared Spectroscopy ...................................................................................................... 32
X-ray Diffraction.............................................................................................................. 34
Microscopy........................................................................................................................... 36
Equilibrium Analysis............................................................................................................ 38
Potentiometric Titrations ................................................................................................. 38
Batch Adsorption Experiments ........................................................................................ 39
Electrophoretic Mobility Measurements.......................................................................... 39
Modelling......................................................................................................................... 40
Manganite Surface Processes...................................................................................................... 42
Surface Charge and Acid-base Properties .......................................................................... 42
Dissolution at Low pH ......................................................................................................... 47
Surface Complexes with Glyphosate and Cd(II).................................................................. 52
Manganite-Glyphosate ..................................................................................................... 52
Manganite-Glyphosate-Cadmium(II)............................................................................... 56
Summary.................................................................................................................................... 65
Future Research ......................................................................................................................... 67
References................................................................................................................................. 69
Acknowledgements………………………………………………………………………………………………74
vi
Populärvetenskaplig sammanfattning
Syftet med denna avhandling är att beskriva ytkemin hos ett manganmineral med namnet
manganit ( -MnOOH). Ytkemin beskrivs med avseende på det rena mineralets* syra-bas
egenskaper* och dess upplösning i sur miljö. Dessutom beskrivs hur manganit reagerar med
kadmiumjoner och glyfosat. Glyfosat och kadmium är båda intressanta ämnen i ett
miljöperspektiv, glyfosat eftersom det är en aktiv ingrediens i vissa växtbekämpningsmedel
och kadmium pga att det är en giftig tungmetall som förekommer i vår miljö.
Fördelningen i naturen av till expempel tungmetaller och näringsämnen påverkas av ett flertal
processer, bland annat kan de adsorberas* på ytan av mikroskopiskt små mineralkorn. I
vattendrag följer i sådana fall dessa kemikalier med mineralerna när de sedimenterar eller
transporteras iväg i vattenströmmar. I naturliga vattendrag finns ofta stora mängder av dessa
mineralpartiklar och de kan därför i stor grad påverka olika kemiska processer i vattnet. Ute i
naturen är mineraler som innehåller aluminium, järn och mangan vanliga. Därför valde vi att
studera ytkemin hos manganit som kan bildas i sjövatten och i jordar.
För att studera manganits beteende och egenskaper användes ett flertal metoder.
Spektroskopiska metoder ger information om ett provs sammansättning med avseende på
mängd och struktur på molekylnivå. Mikroskopi beskriver den yttre formen hos
manganitkristallerna och våtkemiska metoder användes för att undersöka kemiska
förändringar i lösningarna och på ytan av manganit.
I avhandlingen visas att manganitpartiklar får en negativ ytladdning vid alkaliska pH-värden.
Ytans negativa laddning minskar då pH sjunker och vid pH 8.2 är ytan oladdad. Under detta
pH-värde är ytan positivt laddad och under pH 6 löser manganit upp sig och omvandlas till ett
mineral med namnet pyrolusit (β-MnO2).
Vid alkaliska pH-värden adsorberar manganit gärna metalljoner men i närvaro av glyfosat
minskar denna adsorption. Däremot ökar glyfosat metalljonadsorptionen vid neutrala pH
värden. Adsorptionen av glyfosat ökade både vid neutrala och alkaliska pH-värden. I ett
*
mineral ~ sten- eller lermaterial
syra-bas egenskaper ~ ett materials egenskaper vid olika pH
adsorbera ~ fastna på ytan
1
naturligt vatten med pH 6-7, skulle detta till exempel kunna betyda att manganit adsorberar
mer metalljoner om föreningar som glyfosat finns närvarande. Tidigare undersökningar har
visat att nedbrytningen av glyfosat blir långsammare hos adsorberade molekyler. Våra resultat
kan därför tyda på att glyfosatmolekylerna troligen kommer att brytas ner långsammare i
närvaro av manganit. Dessutom, om även kadmiumjoner finns närvarade, kan nedbrytningen
minska ännu mer eftersom kadmiumjoner binder upp den del av molekylen där nedbrytningen
börjar. Däremot kommer adsorptionen att leda till en minskad rörlighet av både glyfosat och
kadmiumjoner i jordar som innehåller manganit.
2
Introduction
The chemistry of mineral surfaces is of great importance in many different areas including
natural processes occurring in rivers, lakes and soils. Metal ions and other compounds can
adsorb onto mineral surfaces and, thus, the fate of these substances can be greatly influenced
by the surface chemistry of the mineral. In water, mineral particles can settle to the bottom of
e.g. a lake, or be transported with water currents. In soils the adsorption onto stationary
mineral surfaces can reduce the transport of different substances and also affect the
bioavailability and degradation of these substances. Minerals can also dissolve and thereby
release environmental contaminants or nutrients that previously were part of their structure or
adsorbed onto the mineral surface.1,2,3
The scavenging of metal ions and other substances by mineral surfaces can proceed by
surface complexation, surface precipitation, and formation of solid solutions. In surface
complexation processes, adsorbents can form bonds to specific sites on the surface or be
electrostatically attracted to the surface. Surface precipitation can be preceded by formation of
surface complexes, and as the coverage of adsorbents is increased, the species at the particle
surface polymerise to form a surface coating. The formation of solid solutions involves an
exchange or insertion of substances into the crystal lattice of a mineral, thereby altering the
composition of the solid.2,4
The dissolution of minerals also affects the content and distribution of different compounds in
water solutions. This process can be beneficial if the dissolved ions are nutrients but harmful
if they are environmentally hazardous substances. Many minerals dissolve through redox
processes, such as the reductive dissolution of iron and manganese (hydr)oxides by the
degradation of organic material.5,6,7
Fe, Al and Mn are abundant elements in the Earth’s crust and readily form (hydr)oxide
minerals. Consequently, minerals such as iron(II, III), manganese(II, III, IV) and
aluminium(III) (hydr)oxides form one important group in all of the processes mentioned
above1. Although manganese is the least abundant of these three, it is reported to be the most
mobile in some areas of the world8. This mobility leads to relatively high concentrations of
3
Mn(II) in surface waters9. The oxidation of Mn(II) and the formation of different
manganese(hydr)oxides can effect the distribution of other substances. For example, in
ferromanganese nodules, some heavy metals are enriched within the manganese fraction of
the nodule instead of the iron fraction, indicating that these metals have a stronger affinity for
manganese (hydr)oxides8. Thus, knowledge of the chemistry of manganese (hydr)oxides is
essential to an understanding of different processes occurring in the environment.
Manganese (hydr)oxides are a large and heterogeneous group of minerals with manganese
oxidation states ranging from +II to +IV under natural conditions8. The oxidation state
depends on the ambient conditions, and the chemical characteristics of the different
manganese valences are reflected in the chemistry of the manganese (hydr)oxides. The ability
to change oxidation state gives manganese (hydr)oxides the possibility to interact with other
redox sensitive substances and participate in processes such as degradation of organic
material6,7.
The most stable of the trivalent manganese hydroxides, manganite (γ-MnOOH), is also a
naturally occurring mineral8. Due to its composition, it can function both as an oxidising and a
reducing agent, making it an interesting mineral in geochemical processes.
Aim of the Thesis
The objective of this thesis was to investigate the surface characteristics and reactivity of
synthetically prepared manganite with respect to acid-base properties and metal/ligand
adsorption reactions. Glyphosate (N-(phosphonomethyl)glycine) and cadmium(II) were
chosen as model substances to investigate the reactivity of the manganite surface with small
organic ligands and heavy metals, respectively. The two model substances are interesting
from an environmental perspective and, furthermore, the molecular structure of glyphosate
allows for the study of surface and metal ion interactions of different functional groups within
the same molecule.
4
The experimental strategy employed here was to combine a range of different techniques to
obtain a more complete understanding of the system studied. Various spectroscopic methods
were used, together with x-ray diffraction, microscopy techniques, potentiometry,
electrophoresis measurements, and batch adsorption experiments. From these techniques,
information about the molecular structure of adsorbates, morphological structure of the
mineral particles, and quantitative thermodynamic data were acquired and used to describe the
chemical processes at the water-manganite interface.
Manganese - General Chemistry
Manganese is a 3d transition metal with an electronic configuration of [Ar] 3d5 4s2. In nature
the oxidation states of (II), (III) and (IV) predominate10,11, but Mn can display any oxidation
state from -III to +VII10. Mn shows typical behaviour of the 3d transition metals10, which is
partly explained by the crystal field theory, with splitting of the d orbitals to form nongenerate energy levels10. The high spin species of Mn(IV) and Mn(II) predominate in nature
and have electron configurations (d3 and d5 respectively) that give Mn(IV) but not Mn(II) a
large stabilisation energy. Mn(III) generally displays Jahn-Teller distorted octahedra with two
long apical distances and four short axial distances due to its d4 valence electron
configuration10.
Mn(II) is the most stable oxidation state of manganese in solution and it can form a number of
different metal complexes with ligands12. Mn(III) ions hydrolyse readily according to
(eq. 1-2)13 and are thermodynamically unstable in aqueous solutions (eq. 3)10, unless
stabilised by a strong complexing ligand10 such as pyrophosphate14, EDTA or citrate15.
Mn(III) can also be stabilised in the form of trivalent hydroxides at higher pH. Mn(IV) mainly
occurs in the form of solid manganese oxides11.
Mn3+ + H2O
Mn3+ + 2H2O
MnOH2+ + H+
Mn(OH)2+ + 2H+
5
log K = 0.4
(1)
log β = 0.1
(2)
2Mn3+ + 2H2O
Mn2+ + MnO2(s) + 4H+
log K = 9.5
(3)
The thermodynamic stability of the different manganese (hydr)oxides is determined by
oxygen partial pressure, redox conditions (pe), pH and temperature. Figure 1a shows the
distribution of predominant oxidation states and compositions with pH and redox
conditions1,11 and Figure 1b shows the disproportionation of manganite at low pH.
a)
b)
pe
120
MnOOH(s)
3
2-
mmoles/dm
MnO4
Mn(OH)3
80
2+
Mn
& MnO2(s)
40
-
0
2
3
4
5
6
7
8
pH
pH
-5
-3
Figure 1. a) pe – pH diagram for the Mn-H2O-CO2 system at 25 °C. Mntot = 10 M, Ctot = 10 M, the
dashed line represents the equilibrium: H2O
+
-
1
½ O2 + 2H + 2e (modified from Stumm and Morgan ).
b) Distribution of Mn species with pH showing the disproportionation of manganite forming β-MnO2
2+
11
and Mn at low pH (paper 3).
The oxidation of Mn(II) is complex and the final product is dependent on the prevailing
conditions during the oxidation process. Extensive efforts have been made to describe the
oxidation from Mn(II) to Mn(IV) 16,17,18,19,20. These studies show that under most laboratory
conditions, manganese does not form Mn(IV), but more often the end product is one of the
Mn(III) hydroxide polymorphs11,21,22. In lake water, a mixture of hausmannite (Mn3O4) and
feitknechite (β-MnOOH) has been reported to precipitate, which with time transforms into
manganite 20,21,16,19,22. This indicates that also in natural waters the oxidation of Mn(II) forms
MnOOH instead of MnO2. However, at low manganese concentrations and in the presence of
bacteria, amorphous Mn(IV) oxides have also been reported to form30,18,23.
6
The Mn(II) oxidation process is also affected by different ligands. In the presence of weak
complexing ligands, such as nitrate or chloride, a mixture of hausmannite and feitknechite has
been shown to form with pyrochroite (Mn(OH)2) as a precursor, but if sulphate was present,
the precipitate consisted of manganite16. To explain this phenomena, it was suggested that the
ion pair MnSO40 inhibits the formation of hexagonal Mn(OH)6 plates and, in this manner,
prevents β-MnOOH from being formed16. In contrast, oxalate has not been shown to influence
the solid phase speciation, although it slows down the oxidation process due to the strong
complex formation22. High concentrations of oxalate were also shown to decrease the
transformation of hausmannite and feitknechite into manganite22.
Manganese - Geochemistry
One of the reasons for the environmental interest in manganese (hydr)oxides is that they act
as scavengers for heavy metals and other substances in natural waters and soils1,8,24,25.
Furthermore, Mn can form oxide coatings on other minerals and bacteria and in this manner
form reactive surfaces24. It has been reported that heavy metals, such as Cd or Co, have a
higher affinity for manganese oxides compared to iron- and aluminium oxides26,27,28. Also,
accumulated amounts of heavy metals have been reported in ferromanganese nodules, where
some metals in particular, Ni, Co, Pb and Mo, seem to be concentrated within the manganese
fraction of the nodules8. These findings emphasise the importance of studies of metal
adsorption onto the different types of manganese (hydr)oxides.
Manganese occurs mainly in sedimentary-, metamorphic- and igneous rock, but Mn is also
present in the form of colloidal manganese oxides in most soils and waters. The average rock
contains about 0.1 % manganese which makes manganese the tenth most abundant element in
the Earth’s crust24,25. Although it is less abundant than both iron and aluminium (1/50 and
1/25 respectively), it is the most mobile of the three8. The mobility can partly be explained by
the higher solubility of the Mn (hydr)oxides compared to Al or Fe (hydr)oxides under ambient
conditions10, but also by the slower oxidation rate for Mn(II) compared to Fe(II)1,8. This
mobility is highest in sub-arctic and temperate soils, which have high organic content and
acidity8. In these regions, Mn leaches from the soil and is transported in dissolved forms or as
7
small colloids in lakes and rivers9. Tropical or arid soils are generally more alkaline and have
a lower relative content of organic material. Thus, in these regions, Mn(II) oxidises within the
soil or forms immobile coatings on rock surfaces on top of the soil, which results in low
mobility of Mn8.
Apart from weathering of Mn containing rock, Mn also originates from anthropogenic or
biogenic sources, although to a much smaller extent. The metal industry and combustion of
fossil fuel are the main anthropogenic sources25, while the biogenic sources are mainly
decomposed plants and animals8,25. The manganese content in plants is due to its catalytic
position in photo system II where it oxidises water to form oxygen during the
photosynthesis29. For animals, manganese is an essential nutrient25 since it forms reactive sites
in different enzymes29, e.g. in superoxide dismutas that react with oxygen radicals in the
body1.
Manganese - Geochemical Cycle
In the geochemical cycle of manganese, dissolved Mn(II) is produced mainly from weathering
of metamorphic and igneous rock24,25. Mn(II) ions are transported in water where they form
complexes with different ligands, adsorb onto particle surfaces, or become oxidised to
different manganese oxides or hydroxides8,25 (Figure 2). The oxidation process can be
autocatalytic, but most commonly the oxidation of manganese is mediated by different
microbes8,25,30. When the particles grow in water they eventually settle to the bottom of the
lake and become part of the sediments. However, if the redox conditions are altered, e.g. in
the presence of a redoxcline, the manganese particles are reduced back into Mn(II) ions.
These ions will diffuse along the redox gradient until they reach oxidising conditions where
they will be reoxidised and form (hydr)oxide particles8. This cycling around the redox
boundary leads to an enrichment of Mn at specific depths, which can be detected in natural
lakes25. The redox boundary is usually found close to or within the sediments in shallow lakes
while, in deeper stratified lakes or lakes with high concentration of organic matter, it is
present within the water column25. In rivers, where the water is well oxygenated and
constantly moving, these redox boundaries are rare and the manganese is mainly present in
the form of colloidal (hydr)oxide particles, although, dissolved Mn(II)† can also be present
depending on season9,31. In the Kalix river in northern Sweden dissolved Mn(II) dominated
†
filtered through 45µm filter
8
during winter and particulate Mn during summer9,31. The manganese (hydr)oxides settle to
become part of the sediments, and with time they form metamorphic or igneous rock as a
consequence of sediment compression, which completes the geochemical cycle for Mn.
Uptake by biota
MnL(aq)
Mn(II)
Mn2+
MnL(s)
X:Mn2+
MnOx
MnOx
redoxcline
Mn2+
Figure 2. The fate of manganese in natural water where manganese can form aqueous complexes
2+
(MnL(aq)) or solid phases(MnL(s)) with different ligands, adsorb onto particulate matter (X:Mn ),
oxidise to form different manganese (hydr)oxides (MnOx), or be consumed by biota. The solids formed
slowly settle and end up in the sediment.
Manganese Oxides - Structures
Manganese oxides and hydroxides are among the most common manganese containing
minerals32. More than 30 manganese oxides are known, although not all of their structures
have been determined24. The oxidation of Mn results in precipitation of different oxides
depending on surrounding conditions, and often an intergrowth of two or more oxides are
found in natural deposits24. The variety of different structures in the manganese oxides is
illustrated in Figure 3. Most manganese minerals have a framework constructed of MnO6
octahedra sharing corners or edges8. From these different architectures, two large groups can
9
be formed that include most of the structures i.e. tunnel and chain structures and layered
structures (Figure 3 and Table 1)24.
a
d
b
e
c
f
Figure 3. The arrangement of MnO6 octahedra in a few selected manganese oxides; a) pyrolusite,
b) ramsdellite, c) hollandite, d) romanechite, e) birnessite (layers of MnO2 with metal ions in the
interspace), f) lithophorite (layers of (Al,Li)(OH)2 and MnO2 alternating in the structure). In c) and d)
counter ions and water have been omitted for clarity.
10
The structure with the smallest tunnel size is found in pyrolusite (Figure 3a), which is the
most thermodynamically stable Mn(IV) oxide under ambient conditions11. Pyrolusite has a
rutile (TiO2) type structure with rows of MnIVO6 octahedra sharing edges and corners.
Manganite has a similar structure. However, the MnIIIO6 octahedra are Jahn-Teller
distorted8,24, and half of the oxygen atoms are replaced by OH-groups38,33,34. Ramsdellite is
isostructural with groutite (α-MnOOH), goethite (α-FeOOH) and diaspore (AlOOH) and
displays tunnels which are slightly larger compared to pyrolusite (Figure 3b)24. An
intergrowth between the pyrolusite and ramsdellite structures is found in the nsutite mineral,
which has single and double chains randomly alternating in the structure, forming a mixture
of small tunnel sizes35. Larger tunnels are found in e.g. todorokite, romanechite and the
minerals in the hollandite group (e.g. hollandite, cryptomelane, coronadite, manjiroite)
(Figure 3c & d). In these latter minerals, the tunnels are built from Mn octahedra with a
mixture of Mn valences and, thus, the tunnels have cations present inside the tunnels to
balance the charge from the manganese framework24.
Lithophorite, chalcophanite and the minerals in the birnessite group are examples of
manganese minerals that have a layered structure (Fig 3e & f). They are all constructed from
charged layers of manganese octahedra and have different counterions in the interspace
between the layers. Pyrochroite also has a layered structure but without cations in the
interspace and with sheets constructed from MnII(OH)6 octahedra. It has not been possible to
determine the crystal structure of several of these layered minerals by the analysis of natural
samples24. Instead, the existing structures have been determined using synthetic analogues.
For example, birnessite, which is present in most soils, has been structurally determined in the
form of a synthetic potassium rich birnessite.
There also exist Mn oxides that have different structures from the ones discussed above. The
most common, hausmannite, has a spinel type structure. This structure consists of oxygen
atoms in cubic close packed arrangement, Mn(III) in ½ of the octahedral holes, and Mn(II) in
¼ of the tetrahedral holes. Apart from the layered pyrochroite, two Mn(II) structures are
included in Table 1. Manganosite has a halite (NaCl) type structure. It is a mineral that can be
found in metamorphic ore and can be produced by heating rhodochrosite. Rhodocrosite has a
calcite (CaCO3) type structure36, and although it is not a manganese oxide, it is included here
since it is commonly present in sediments and rearranges to form manganese (hydr)oxides11.
11
Table 1. Some manganese minerals with chemical composition and structure24,8.
Mineral name Chemical formula
Alternative name Mn valence
Manganosite
MnO
+II
Pyrochroite
Mn(OH)2
+II
Rhodocrosite MnCO3
+II
Hausmannite Mn3O4
+II, +III
Groutite
+III
α-MnOOH
Feitknechite# β-MnOOH
+III
Manganite
+III
γ-MnOOH
Birnessite
(Na,K,Ca)4Mn7O14 ×2.8 H2O δ-MnO2,
+II, +III +IV
7Å manganate
Vernadite* #
MnO2 ×nH2O
Lithophorite
(Al,Li)(OH)2 × MnO2
+III, +IV
Chalcophanite (Zn, Mn, Fe)Mn3O7 × 3H2O
+II, +IV
Todorokite
(H2O,…)2-xMn8(O,OH)16
10Å manganate +III, +IV
+III, +IV
Romanechite (Ba,Mn2+,…)3(O,OH)6Mn8O16 Psilomelane
Hollandite
Ba2-xMn8O16
+III, +IV
Cryptomelane K2-xMn8O16
+III, +IV
α-MnO2
Coronadite
Pb 2-xMn8O16
+III, +IV
Manjiroite
Na 2-xMn8O16
+III, +IV
+IV (+II, +III,
Nsutite
Mn(O,OH)2 × H2O,
γ-MnO2
+IV)35
(Mn4+1-xMn2+xO2-2x(OH)2x)
Ramsdellite
MnO2
+IV
Pyrolusite
+IV
β-MnO2
Str
L
T
L?
T
L
L?
L
L
T
T
T
T
T
T
T
T
T
*) Synthetic δ-MnO2 analogue24.
#
) no crystal structure has been determined for this mineral.
Str = structure, T = tunnel structure, L = layered structure, L?=layered structure is suggested in the literature
Manganite (γ-MnOOH)
Manganite (γ-MnOOH), is the most stable of the trivalent manganese hydroxides8 and is the
hydroxide most likely to be present in lakes and rivers with high manganese concentrations
and/or high sulphate concentrations22. It occurs in the form of colloidal particles in water and
sediments. However, due to the difficulties of sampling colloidal particles, it has been
suggested that the presence of manganite has been overlooked in many routine samplings37. In
sediments, manganite is commonly found as a layer between the anoxic MnCO3 phase and the
fully oxidised pyrolusite (β-MnO2) layer8. Careful studies of the manganite and pyrolusite
boundary show that the manganite is pseudomorphously replaced by pyrolusite during
oxidation8,11. A similar process can be observed when manganite is heated in air and oxidises
to pyrolusite. The only morphological change observed from this heat induced process is a
12
decrease in the unit cell length38 corresponding to the contraction of the elongated MnIIIO6
octahedra.
Proton promoted dissolution of manganite at low pH has been observed by several
authors5,39,40. The dissolution occurs through disproportionation of Mn(III) into Mn(II) ions in
solution and one of the solid Mn(IV) oxide polymorphs (eq. 4).
2 MnOOH(s) + 2 H+
MnO2(s) + Mn2+ + 2 H2O
log K~ 7.0-8.6
(4)
To our knowledge only two studies have been performed to understand this proton promoted
dissolution11,41 and in many low pH adsorption studies onto manganite the dissolution has not
been addressed although the adsorption experiments have been conducted in a pH region
where dissolution is significant.
Surface Structure
The adsorption of different substances onto a mineral surface is often interpreted using the
structure and composition of the surface. The structure of the mineral surface in aqueous
solution is often not known in detail but a fairly good assumption is that it resembles the bulk
structure42. Since the three dimensional framework in the bulk is lost at the surface, metal ions
will be under-saturated with respect to bond formation2. Thus, oxygen atoms positioned at the
metal hydroxide surface can be pictured as having fractional charges according to Pauling’s
valence bond theory43. According to this theory, the metal ion displays a fraction of its charge
at each coordinating atom or group. For example, a trivalent metal ion within an octahedron
of oxygen atoms would have a formal charge of +½ at each bond. Due to the presence of
water, the surface of a hydroxide or oxide mineral will not be coordinatively unsaturated but
will be covered by hydroxyl groups once immersed into a pure aqueous environment. At the
surface this results in hydroxyl groups, displaying a charge of -½ and doubly protonated
oxygen atoms with a charge of +½ (since OH2 is neutral). If the hydroxyl groups are
coordinated to two metal centres instead of one (i.e. doubly coordinated) they will display a
charge of 0 or, when protonated, +1. The net charge of the surface will be positive when the
number of protonated groups is higher than the number of deprotonated groups, and vice
versa. The pH at which the total positive charge is equal to the negative charge is called pHpzc
13
(pzc = point of zero charge). Thus, the change in charge at the surface can be seen as a result
of protonation and deprotonation of hydroxyl groups.
Assuming that the surface displays the same structure as the bulk, a theoretical number of
adsorption sites can be estimated for specific planes of the crystals. The dominating planes for
small manganite crystals have not been determined and may differ compared to larger
crystals44. However, since the (010) plane has been shown to be predominant in large
manganite crystals and since it also constitutes a plane of cleavage34,45, we have assumed that
the small manganite crystals of this work are dominated by the (010) plane.
On the (010) plane, the surface will display singly coordinated oxygen (=O-3/2) or hydroxide
groups (=OH-½) (Figure 4). In water suspensions these groups will be protonated44 to form
=OH-½ or =OH2+½, depending on surrounding pH (paper 2). The singly coordinated oxygen
atoms can be assumed to be responsible for the reactivity at the (010) plane42 since the doubly
coordinated oxygens display lower charge. The triply coordinated oxygens are positioned
further down in the structure, and are here assumed to be less reactive than the singly
coordinated oxygens. The (010) plane hosts 7.9 singly coordinated oxygens /nm2
(13 µmol/m2).
doubly
singly
triply
(001)
(010)
34
Figure 4. The manganite structure
with the (010) plane visualised by dotted lines. The plane of the
paper is (010) in the left structure and (001) in the right structure. The arrows show singly-, doublyand triply coordinated oxygen atoms at the surface.
14
This density of crystallographic sites on the manganite surface was used in the work presented
here due to the difficulties and absence of good methods for experimentally determining site
densities. In the literature there is, to our knowledge, only one previously reported study on
manganite site density. This was a fluoride adsorption study, which yielded 5 sites /nm2 or
8.3 µmol sites /m2 of manganite5.
Mineral Synthesis
The manganite used in this thesis work was synthesised according to Giovanoli and
Leuenberger46, by adding 300 cm3 of 0.2 M NH3 (Merck p.a.) to a solution of 20.4 cm3 30 %
H2O2 (JT Baker p.a.) and 1 dm3 0.06 M MnSO4 (BDH). The mixture was rapidly heated and
thereafter kept at a temperature of 95 °C for six hours with constant stirring. The formed
suspension was filtered, still hot, through a G4 glass filter, washed with equal amounts (with
respect to the initial solution) of hot deionised water, collected on watch glass and dried under
reduced pressure in presence of P2O5 (Riedel-de Haën). The dried manganite was ground to a
powder, the presence of crystalline phases checked using X-ray powder diffraction and the
crystallinity with High Resolution Transmission Electron Microscopy (HRTEM) (Figure 5).
4 nm
Figure 5. Transmission electron microscopy image of part of a manganite crystal showing the inner
crystalline structure. The scale bar in the upper left corner represents 4 nm. The width of the needle in
this picture is ~20 nm (paper 1).
15
XRD of the synthesised manganite gives the typical d-spacings for manganite, at 3.41, 2.64,
2.52, 2.41, 2.20, 1.78, 1.70, 1.67, 1.50, 1.44 and 1.32 Å. The synthesis usually yielded only
pure manganite but, at a few occasions, a trace of hausmannite was found in the diffraction
patterns. When this happened, the specific batches were discarded and not used further. The
manganite was stored as a powder, from which suspensions with known amounts of solid
were prepared. Suspensions were placed in an ultrasonic bath for a few minutes where they
were shaken from time to time to avoid over heating and to keep the particles from settling.
The suspensions were thoroughly bubbled with Ar(g) at the time of preparation and were
stored with Ar atmosphere in the bottles. Typically 10 g of manganite was used per dm3
suspension. The manganite crystals seem to disperse well in suspensions, although they form
aggregates when dried as seen in microscopy images (Figure 6). During the work of this
thesis it was found that the potentiometric titrations were reproducible if the suspensions were
allowed to age for a few months, which most likely is due to Oswald ripening processes i.e.
the growth of larger crystals on the expense of smaller ones.
a)
b)
Figure 6. Scanning electron microscopy images of dried manganite crystals. The scale bars represent
3µm in image a and 750 nm in image b. Length of scale bar:
The surface area of manganite was experimentally determined after synthesis using the BET
N2 adsorption method47 at 70 °C. This temperature was chosen in order to prevent phase
transformations40. It was discovered that heating manganite under a nitrogen atmosphere
converts it into hausmannite. However, if heated in air, then pyrolusite is formed, which
supports the findings of e.g. Giovanoli and Leuenberger46. Since the tests did not show any
changes in XRD patterns below 100 °C, a temperature of 70 °C was chosen as the degassing
16
temperature to ensure that no phase transformation occurred. The BET area varied between 30
and 51 m2/g for the manganite synthesised in this work, and the difference in surface area
between the batches is ascribed to the size difference between the manganite crystals46. The
synthetic manganite is crystalline and needle-shaped with an average length of approximately
500 nm and a width of about 20 - 40 nm (Figure 5 & 6).
Methods
In this study, chemical processes at the manganite surface were investigated using a number
of different analysis methods. This was done in order to have a broader understanding of the
manganite surface and reactions occurring at the manganite-water interface. Spectroscopy,
x-ray diffraction, microscopy and wet chemical techniques were used to examine the different
properties of the material and to interpret the chemical processes at the surface both
quantitatively and at a molecular level. The combination of different techniques gives the
possibility to validate data and also a better understanding of the surface processes can be
obtained.
Spectroscopy and Diffraction
Different spectroscopic and diffraction techniques can be used to obtain molecular level
information about a sample. The common theme in most spectroscopic methods is that
electromagnetic radiation is used and the effect from the interaction between the
electromagnetic radiation and the sample is recorded48. Three different types of spectroscopic
techniques were used in this thesis, X-ray Photoelectron Spectroscopy (XPS), X-ray
Absorption Spectroscopy (XAS) and Fourier Transform Infrared (FTIR) Spectroscopy. From
these three techniques, quantitative and qualitative information about the composition of
surface species, the molecular structure and coordination around a specific atom, and the
molecular structure and identity of compounds can be obtained. Furthermore, X-ray
diffraction (XRD) was used to identify and characterise the structure of the crystalline phases.
17
X-ray Photoelectron Spectroscopy (XPS)
XPS is used to determine and quantify the chemical composition of a sample surface.
Theory
When X-rays strike a surface, high energy photons cause ejection of electrons from the atoms
in the sample. By measuring the kinetic energy of the emitted electrons and knowing the
energy of the incident beam, the electron binding energy can be calculated using eq. 5.
EB = hν - EK - φ
(5)
where EB is the binding energy in eV
hν is the energy of the incoming photon
EK is the kinetic energy in eV and
φ is the spectrometer work function, determined in separate measurements.
The binding energy for core level electrons is specific for each element and orbital. This
makes it possible to elucidate the elemental composition of a surface from the binding
energies obtained in XPS. In addition, the binding energy of an element is influenced by the
chemical surroundings and oxidation state of the atom, making it possible to distinguish
between different chemical forms of an element e.g. between protonated and deprotonated
amine groups (Figure 7). All elements except H and He can be directly detected using XPS.
However, this does not mean that no information about H can be obtained, since the presence
of H will shift the binding energy of atoms bound to H (Figure 7). The quantitative estimate
of the different surface components are given in atomic % of the surface (< ± 10 %) and the
detection limit is approximately 0.1 atomic %49.
18
x 10
120
1
Name
NH2 +
NH
Pos.
402.0
399.9
FWHM
1.38
1.56
Area
%
659.6 48.5
701.0 51.5
110
Counts/s
100
90
80
412
408
404
400
Binding Energy (eV)
396
Figure 7. Nitrogen 1s spectrum of aqueous glyphosate at pH 9 showing the chemical shifts as a result
of the protonation (paper 4).
XPS spectra are created by electrons emitted from the top atomic layers of a sample, making
it a surface sensitive technique. Electrons ejected from atoms deeper down in the sample
collide with atoms or electrons on their way to the surface and loose some of their specific
kinetic energy. These electrons form the background intensity in XPS spectra. The analysis
depth (or information depth) is dependent on the kinetic energy of the measured photoelectron together with the sample density and the type of elements present. For an electron
with a kinetic energy around that of O 1s, the analysis depth would be roughly 4.5 nm for a
mineral such as manganite, 3 nm for Mn metal and 6.5 nm for an organic substance such as
glyphosate‡,50. However, since the likelihood of electrons escaping the surface falls off
exponentially to the depth, the major part (63 %) of the signal originates from a layer about
1/3 of the analysis depth51.
‡
Estimated as 3 × attenuation length for an electron with KE=650 eV in the material in question. Attenuation
length calculated using NIST database50
19
x 104
40
35
O
30
Mn
CPS
25
20
15
10
5
1000
800
600
400
Binding Energy (eV)
200
0
Figure 8. Wide spectrum for a sample of manganite with glyphosate adsorbed at the surface. The
arrows show the broader Auger lines from Mn and O atoms. The narrow lines represent photoelectron
lines from the different elements present in the sample.
In addition to the specific photoelectron lines in the spectra, Auger lines can be seen
(Figure 8). These lines originate from rearrangements of electrons within the atoms. When an
electron is ejected from a core level, a vacancy is created, which can be filled by a relaxation
process. An electron from a higher orbital moves to fill the empty site (Figure 9), and energy
is released either in the form of X-rays or as another ejected electron, i.e. an Auger electron.
The Auger lines are usually broader and have a more complex substructure, compared to
photoelectron lines. From Auger lines information about oxidation state of elements can be
extracted.
20
(a)
(b)
h
h
4
4
1
3
2
Figure 9. Schematic drawing of an oxygen atom exposed to X-rays (h ). (a) shows the photo-ejection
of an electron from the O 1s orbital. (b) shows a subsequent reorganisation resulting in emission of
X-rays or an Auger electron. The sequence of the processes is shown by the numbers 1-4.
Instrument
For this work, monochromatic Al Kα X-rays (h = 1486.6 eV) have been used as incident
radiation, but it is also common to use non monochromatic Mg Kα or Al Kα radiation for
XPS analyses. The non monochromatic X-rays have a higher intensity, but can contain other
X-ray lines apart from Kα, which can give rise to additional peaks in the spectra.
Furthermore, a monochromatic X-ray beam gives better energy resolution. In this work, the
monochromator used for the Al Kα X-rays consists of a set of quartz crystals positioned to
reflect the beam so that only a specific wavelength of the beam reaches the sample surface.
If the surface is not electrically conductive, it will build up a positive charge as a consequence
of the ejection of electrons. This charge will influence the electrons leaving the surface and to
prevent this, charge neutralising equipment is used. In this work, one of the most common
techniques for charge neutralisation was used, which is to flood the surface with low energy
electrons (< 20 eV). Despite this treatment, there are differences in charging between samples,
and to be able to compare data, the binding energy scale has to be calibrated. Most surfaces
analysed with XPS display an amount of hydrocarbons due to surface contamination. This
adventitious carbon can be used as an internal standard, and the difference in measured EB for
the aliphatic carbon component and the literature value (285.0 eV) is used to shift all the
21
spectra to obtain a more accurate scale. This approach assumes that the surface contamination
exhibits the same charging behaviour as the rest of the sample and that it does not interact
with sample surface. This is not always the case and suggestions to deposit a small gold dot
on the surface, or to use a well characterised component from the sample matrix as internal
standard have been made51,52. The first suggestion requires a gold evaporator and careful
control of the thickness of the gold since different thickness of gold films have been proven to
influence the BE for the Au lines. The second of these suggestions requires that the reference
component always has the same BE independent of the surrounding conditions in the sample.
In this thesis, the approach with the adventitious carbon has been used. The absence of
interactions between the aliphatic carbon and the manganite cannot be guaranteed, but should
be more or less identical in all samples. Thus, this internal calibration enables comparisons to
be made between samples. Furthermore, the low concentration of carbon at the surface
indicates that the manganite does not have a strong affinity for hydrocarbons.
XPS measurements take place under ultra-high vacuum conditions, i.e. at pressures below
10-8 Torr. There are several reasons for this; one is to prevent the electrons leaving the sample
surface from colliding with gaseous atoms and molecules. These collisions would result in
loss of the specific kinetic energy of the electrons and, thus, loss of signal intensity. Another
reason is that the high vacuum keeps the sample and the analyzer (f-i in Figure 10) clean from
surface contaminations. Unfortunately, the vacuum conditions result in “disappearence” of
volatile substances at room temperature. However, this disadvantage can often be overcome
by freezing the samples using N2(l) in order to reduce the evaporation process.
22
_
h
+
g
i
f
e
d
c
b
a
Figure 10. Schematic picture of the pathway of the electrons leaving the sample in XPS
a)
b)
c)
d)
e)
f)
Sample
Magnetic lens
Iris
Scan plates
Spot size aperture
Electrostatic lens
g) Retention grid
h) Hemispherical electron
energy analyzer
i) Detector system –
electron multiplier
= electron path
In this work, the kinetic energy of the electrons is detected in a concentric hemispherical
analyser (Figure 10h). The analyser consists of two hemispheres, one positive and one
negative, concentrically located on an insulating pedestal. On one end there is an entrance slit
and a retarding grid, and on the other there is an exit slit and an electron multiplier. By
changing the potential between the two hemispheres, a path between them is created which
only allows electrons with a specific kinetic energy (EK) to reach the electron multiplier,
while electrons with different EK will collide with surfaces on either side of the path. The
energy an electron must have in order to pass through the analyser is called the pass energy,
and this energy is proportional to the voltage difference between the two hemispheres. The
number of electrons passing through the path is, thus, decided by the pass energy. Using a low
pass energy, e.g. 20 eV, results in lower intensity of the output signal but better spectral
resolution. The retarding grid serves to maintain constant analyser energy, by accelerating
23
electrons in the beginning of a scan, when the electrons have low energy, and decelerating
electrons in the end of the scan.
The electron multiplier consists (for the work presented here) of a set of 8 channeltrons
(channel electron multipliers). The channeltrons are constructed of glass tubes coated on the
inside with a resistive material that has a high secondary emission coefficient. Thus, an
electron entering a tube generates many more electrons when colliding with the tube walls.
The secondary electrons produced continue to multiply in a cascade manner, and at the end of
the channeltron the resulting electrical current is measured and fed to software where it is
displayed as intensity in a spectrum.
To control the analysis area, different apertures (e in Figure 10) are used. For the work
presented here, a slot was employed that gives an analysis area of 0.3 × 0.7 mm. The energy
resolution for the spectra is estimated from the width of the Ag 3d5/2 line measured on a Ag
standard sample. At 20 eV pass energy and with the above slot, this line had a FWHM (full
width at half maximum) of 0.63 eV for the instrument used in this study.
For more information about the XPS technique, the reader is suggested to consult reviews by
Hochella51 Paterson & Swaffield53 and Ratner & Castner49.
Experimental procedure
For the XPS analyses, a relatively thin layer of the wet paste, or a small drop of solution, was
applied onto a sample holder and directly placed on the pre-cooled stage (-170 °C) in the
sample introduction chamber of the spectrometer. The samples visibly froze within
approximately 5 - 10 seconds and, subsequently, the pressure in the sample compartment was
reduced (while maintaining the sample at approximately -170 °C). When required pressure
was obtained (~5 × 10-7 Torr), the sample was introduced into the analysis chamber, placed
onto the pre-cooled stage and the measurements started. The temperature of the sample
manipulator was kept at approximately -155 °C or below throughout the entire measurement.
This treatment maintained water in the sample and kept the pressure in the analysis chamber
below 10-8 Torr. Next, the frozen paste was left overnight inside the spectrometer, with the
main valve open (to increase pumping capacity) and high voltage switches off (for safety).
After reaching room temperature, the sample was remeasured the following day in the exact
same spot (Figure 11).
24
x 102
x 102
20
Name Pos. FWHM
O
1.10
530.3
OH
531.5
1.35
H2O
533.5
1.60
18
Area
1774.7
2350.0
573.2
%
37.8
50.0
12.2
30
Name Pos.
O
530.2
OH
531.5
OH2 532.8
a)
FWHM
1.07
1.38
1.72
Area
2707.7
3775.3
242.7
%
40.3
56.1
3.61
b)
25
14
20
12
15
10
8
10
6
4
5
2
544
540
536
532
544
528
540
Binding Energy (eV)
536
532
528
Binding Energy (eV)
Figure 11. O 1s XPS spectra of a) frozen manganite paste and b) room temperature paste the
subsequent day, showing that water is maintained at the interface in the frozen samples (paper 1).
NH
Intensity
Intensity (counts/s)
16
+
NH2
407
405
403
401
399
397
395
Binding Energy (eV)
Figure 12. Spectrum showing the changes in N 1s with loss of water. The solid line represents
glyphosate adsorbed onto goethite analysed as a frozen paste. The dotted line represent the same
sample but at room temperature (paper 4).
25
This approach enables observations of changes caused by the disappearance of water (Figure
11 & 12) and the ability to estimate the water layer thickness. Typically this water layer was
estimated to 0.6 nm for the manganite paste.
The rough estimate of the water layer thickness on manganite was made by using an equation
for calculating the thickness of SiO2 on Si54, shown in eq. 6.
Ipaste = Iwarmed sample × e-d(aq)/(λ sinθ)
(6)
Ipaste is the intensity of the Mn 2p peaks in the frozen paste, Iwarmed sample is the intensity of the
same peaks when the water layer is removed, d(aq) is the thickness of the water layer, λ is the
attenuation length of Mn 2p electrons in water and θ the take-off angle (θ = 90 and,
consequently, sin θ =1). The attenuation length, λ, was calculated using the NIST standard
database50. However, this thickness is only a rough approximation since eq. 6 assumes a flat
surface, but the samples used in this thesis are in the form of powders.
In XPS analyses of frozen droplets of solution, a large concentration effect was observed in
relation to the actual concentration of the solution. In an aqueous solution containing 1 mM
glyphosate, NaOH (for pH adjustment) and H2O, the ratio between the area of the water peak
(~55 M) and the components of the glyphosate molecule (1 mM) should be very large
(55/0.001=55000). However, XPS spectra displayed a much lower ratio between water and
glyphosate suggesting that the glyphosate was concentrated through the freezing procedure
and/or analysis procedure. A third option could be that the solution-air interface displays a
different composition than the bulk of the droplet.
The concentration effect might arise through the evaporation of water from the frozen sample
droplet. This would leave a higher percentage of glyphosate at the surface than originally was
present. XPS spectra supported this hypothesis. It was observed that an increase in the
analysis time decreased the O/N ratio, which indicates that water molecules leave the sample
in the vacuum even at these low temperatures (-155 °C). This evaporation could also be
detected as a slight increase of pressure in the analysis chamber from ~5×10-10 to
~5× 10-9 Torr.
26
In spectra of frozen pastes, a contribution from the small amounts of remaining bulk solution
in the paste is possible. This has been observed for the ionic medium and raises the question if
non-adsorbed ligands in solution can contribute to the XPS spectra. The problem with NaCl
originating from the solution can be quite easily removed by subtracting e.g. the Na+
concentration by the Cl- concentration, if Na+ is in excess in the sample and Cl- does not
interact with the surface this would give the excess of Na+ at the surface. However, since the
concentration of adsorbing ligands is much lower than the concentration of ionic media,
(usually 100 mM ionic medium compared to < 5 mM ligand), the contribution in spectra from
non-adsorbed ligands should be much smaller smaller.
In paper 4, the concentration of protonated and deprotonated glyphosate was investigated as a
function of pH using XPS. It was observed that at higher surface loadings the concentration of
protonated glyphosate at the manganite surface increased, while the concentration of
deprotonated glyphosate stayed at the same level (Figure 13). This was suggested to be an
effect from surface saturation with respect to the deprotonated species, similar to previous
observations of glyphosate adsorption onto goethite at neutral pH values94 (paper 4).
However, this increase could also be due to an increased amount of protonated glyphosate in
the solution that contributed to the XPS spectra.
Atomic ratio N/Mn
0.08
0.06
0.04
0.02
0
5
6
7
8
9
10
pH
Figure 13. Atomic ratio of N 1s components in glyphosate adsorbed to manganite. The amount of
protonated ( ) and deprotonated ( ) amine at the surface of: manganite with 14.6 µmol glyphosate
(empty symbols), 9.8 µmol glyphosate (shaded symbols) and 2.4 µmol glyphosate (filled symbols)
2
added per m of manganite (paper 4).
27
To check this, closer examinations were made of the glyphosate-MnOOH system. Throughout
the pH range, the XPS spectra indicated the presence of both protonated and deprotonated
amine groups (Figure 13). This could also be observed in FTIR spectra (in the form of two
types of vibrations for the carboxylic group and the amine group). Since the contribution of
solutes in FTIR can be removed, these infrared data represent only the glyphosate adsorbed
onto the manganite surface. The difference in the concentration of protonated amine groups
observed in the XPS spectra should also be seen by subtracting infared spectra of different
surface loadings of glyphosate at the same pH (Figure 14). If the intensity due to protonated
glyphosate is greater in one spectrum, then this would give rise to a subtraction spectrum
identical to the spectrum of the protonated glyphosate (Figure 14c). However, no difference
was found in the regions for the two carboxylate-amine vibrations (~1600 & ~1400 cm-1).
Thus, the increase in protonated amine detected in the XPS spectra can also be explained by a
contribution from remaining solution in the paste. This solution effect is increased since
solutes are concentrated by the vacuum (as described above).
a)
b)
c)
1700
1600
1500
1400
/cm
1300
1200
-1
Figure 14. Difference between FTIR spectra of different surface loadings of glyphosate onto
manganite at pH 6.4. Vibrations indicating the protonation of the glyphosate should be observed
-1
-1
2
around 1600 cm and 1400 cm . Spectrum (a) represents 14.6 µmoles glyphosate/m manganite,
2
spectrum (b) 9.8 µmoles glyphosate/m manganite and (c) the difference between spectra a and b.
The contribution from solution should increase in importance as the concentration of free
ligand is increased in solution. In paper 6, the Cd(II)-glyphosate-manganite system was
28
studied using XPS. The concentration of free protonated glyphosate was 0.5 mM in these
experiments, but no contribution from the solution could be detected in XPS spectra. For the
pure glyphosate-manganite system, the concentrations of free ligand were similar to this for
the lowest surface loading reported in paper 4 but increased to 1.5 mM for the higher surface
loading. In these latter spectra, the contribution from solution was seen, which indicates that
samples should be prepared so that the concentration of free ligand in solution is
0.5 mM.
Despite the above mentioned problems, the low temperature XPS method can be used
successfully when using mineral systems with low concentration of ligand in solution. This is
clear from data for the Cd-glyphosate-manganite system presented in paper 6. By comparing
solution data with XPS data of the adsorbing species, it can be seen that these two methods
are in good agreement (Figure 42).
X-ray Absorption Spectroscopy
Extended X-ray Absorption Fine Structure (EXAFS) spectroscopy and X-ray Absorption
Near Edge Structure (XANES) spectroscopy can be used to characterise the local chemical
structure and geometry of a compound. A great advantage with these techniques over XRD
(described below) is that they can be used to determine the local structure in non crystalline
samples, as was done for the different Cd-glyphosate complexes in papers 5 and 6. XANES is
sensitive to changes in oxidation state of atoms and coordination geometry around the atom,
but this method has not been used in this thesis.
In both XANES and EXAFS, information about local structure is obtained from the fine
structure at the high energy side of the so called absorption edge. When atoms are irradiated
by high energy x-ray photons, electrons are ejected out of the inner orbitals. If a range of xray energies are scanned and the absorbance plotted vs. energy, this ejection is seen as an
absorption edge at the binding energy of the electron (Figure 15). On the high energy side of
the edge, an oscillating fine structure can be seen in the absorption curve. This fine structure,
extending from ~50 eV up to as much as ~1000 eV, is used in EXAFS. It is a result of
interactions between the ejected electron wave and atoms present around the absorbing atom.
By analyzing the oscillations in the spectrum, the local structure around the absorbing atom
can be determined in the form of number of coordinating atoms around the absorber and
29
distances (up to ~5 Å) between the absorbing atom and its neighbours. XANES, on the other
hand, uses the first region around the absorption edge, i.e. from < 10 eV below the absorption
edge to ~50 eV after the edge55.
Figure 15. Graph showing the absorption edge and the regions used in XANES and EXAFS
55
spectroscopies (modified from Brown et al. ).
Since the fine structure in XANES and EXAFS is fairly weak, intense x-ray radiation is
needed to obtain a good signal-to-noise ratio and, thus, synchrotron radiation is used.
Synchrotron radiation is emitted by electrons (or positrons) that travel close to the speed of
light in a curved pathway. It has a high degree of collimation and is very intense over a broad
and continuous energy range. Synchrotron x-rays are monochromated and the intensity of the
beam is measured before and after passage through a sample, which enables the amount of
absorption to be determined.
30
a)
c)
b)
d)
Figure 16. EXAFS data after different steps of the data treatment process. a) Absorption edge fitted
with a spline, b) normalisation of step height, c) isolated EXAFS oscillations, d) data in r-space
55
(modified from Brown et al. ).
The data treatment in EXAFS starts by extracting the weak EXAFS oscillations, which only
represent a few percentage of the total absorption. It is done by fitting the absorption edge
with a spline and normalising the intensity of the step height (Figure 16). This isolates the fine
structure oscillations at the high energy side of the edge. The pure oscillations can be seen as
a sum of several sinusoidal signals, and to make the data more illustrative the signals are often
Fourier transformed into r-space. This transformation results in a graph (radial distribution
function), and the peaks in this graph represent the different atomic shells surrounding the
absorbing atom (Figure 16d). The approximate distance from an atomic shell to the absorbing
atom is shown on the x-axis and the height of the peak represents the number of absorbing
atoms in that shell. When this graph is constructed, corrections for phase shift are not done
and this results in a slight error in the distances showed on the x-axis of the graph. The data
are analysed by curve fitting procedures and information about average coordination number
(± 0.2 - 0.5 atoms), distance to the absorbing atom (± 0.01 Å), and internal structural disorder
(Debye-Waller factor) is obtained55. In this fitting procedure the phase shift is taken into
account and, thus, the fitting gives a correct determination of the distances. Theoretical
31
scattering paths, calculated using the FEFF program code56, form the basis of the algorithm
used to fit of the data. These paths are calculated for a presumed model structure, e.g. a
fragment of a crystal structure.
Further reading about XAS can be found in the reviews by Jalilevhand57 and Brown et al.55.
Infrared Spectroscopy
Infrared (IR) light is used to study molecular vibrations and from these, deduce information
about bonds within molecules. Different molecular vibrations absorb different wavelengths of
light and, furthermore, the mass of the atoms participating in bonds and the strength of the
bonds will influence the wavelength absorbed. IR spectroscopy can be used qualitatively to
obtain a “fingerprint” of a substance48. Measured spectra are then compared to a library of
previously obtained spectra and the substance identified. IR can also be used to monitor
specific functional groups in a substance and how the structure of these groups changes as a
consequence of different chemical environments, as was done for the glyphosate molecule in
papers 5 and 6.
For small simple molecules the molecular vibrations and resulting IR-spectra can be
interpreted based on the molecular structure and symmetry. For more complex molecules this
is difficult to do, instead the experimental data can be evaluated by comparing them to
frequencies from ab initio calculations. But even without these calculations, structural
information can be derived from the spectra. When a part of a molecule forms bonds to a
surface or metal ion, the molecular vibrations will be affected and the changes are often
visible in the IR spectra. By monitoring these changes, information about structure can be
obtained.
Attenuated Total Reflectance - Fourier Transformed Infrared spectroscopy (ATR-FTIR) was
used in this thesis. The FTIR technique is applied in order to speed up the process of
collecting spectra and to obtain a better signal to noise ratio compared to conventional IR
spectroscopy. The ATR-cell is used since it allows spectra to be collected from strongly
absorbing samples or compounds in a strongly absorbing matrix such as aqueous solutions
and aqueous pastes.
32
(a)
(b)
(d)
(c)
(e)
(f)
Figure 17. Schematic diagram over a Michelson Interferometer
(a) = fixed mirror
(d) = monochromatic IR source
(b) = moving mirror
(e) = sample
(f) = detector
(c) = beam splitting mirror
The FTIR-instrument uses a device known as a Michelson Interferometer (Figure 17). The
incident IR beam is collimated by a concave mirror and divided into two beams at the
beamsplitter. One of these beams is reflected by a stationary mirror while the other is directed
to a mirror moving at constant speed. Reflections from the two beams are recombined at the
beamsplitter and the difference in path length between them will result in constructive and
destructive interference. In this manner a sinusoidal modulation in radiation intensity is
obtained that is passed through the sample and focused onto a detector. The detector records
intensity as a function of mirror position and this information is subsequently mathematically
converted, by the use of Fourier transformation, to intensity as a function of wavenumber.
33
IR light
Sample
To detector
Crystal
Figure 18. Schematic drawing showing the function of the ATR cell used in FTIR measurements. The
arrows represent the IR light, which is reflected several times within the diamond crystal on its way to
the detector.
The ATR cell used, in the present study, consists of a diamond crystal positioned horizontally
in the sample compartment and the sample is applied on top of the crystal (Figure 18). The
diamond crystal has a high refractive index, which enables IR light to be reflected at the
interface between the crystal and the sample (as long as the sample has a lower refractive
index). The geometry and size of the ATR crystal result in multiple reflections of the IR light
within the crystal. At each reflection, an evanescent field penetrates a few microns into the
sample. If the sample absorbs this evanescent wave, the intensity of the radiation is decreased
at the absorbing wavelengths giving IR spectra similar to that of transmission measurements
but from a path length that is only a few microns.
X-ray Diffraction
The structure and identity of crystalline samples can be determined using X-ray diffraction
(XRD). The X-ray wavelength is of the same order of magnitude as the bond distances
between atoms. As a result, the crystal array scatters the X-rays, and this scattering pattern is
unique for the particular atomic arrangement in a sample. In this work, X-ray powder
diffraction has mainly been used to check the outcome of the manganite synthesis but also to
identify phase transformations at low pH (paper 3). Furthermore, single crystal X-ray
diffraction was used in paper 5 to determine the crystal structure of a Cd(II)-glyphosate
compound, (Cd9(PMG)6(H2O)12 × 6 H2O, colour Figure on p 68).
The diffraction by crystalline material can be explained by the model by Bragg. In this model
the crystal structure is pictured as consisting of planes that reflect X-rays. The diffraction is
dependent on the angle of incidence of the X-rays and the distance between planes in the
crystal according to eq. 7.
34
2 d sinθ =n λ
Bragg’s law
(7)
where d is the distance between imaginary planes within the crystal,
θ the angle of the incident radiation,
the wavelength of the X-ray radiation and
n the order of the reflection, commonly n is set to 1.
Powder diffraction is often used to identify crystalline substances. Apart from identification of
crystalline phases, the one-dimensional powder diffraction pattern can give additional
information about the sample, e.g. small crystals can cause broadening of peaks, and the
presence of amorphous material can result in very broad peaks, so called halos. Unfortunately,
the identity of the amorphous phase can not be determined from the one dimensional
diffraction patterns and the presence of amorphous material frequently cause diffuse features
in the diffraction pattern.
To obtain a complete range of reflections during powder diffraction analysis, the angle θ
(theta) is altered stepwise or continuously. The sample is presented as a powder, which ideally
will expose all possible orientations (and planes) of crystals in the sample. The obtained one
dimensional diffraction pattern can be compared to reference patterns in databases and the
identity of the sample determined. In some cases the crystal structure can also be derived from
these patterns although single crystal diffraction is often preferred for structural
determinations.
In single crystal diffraction, a single crystal is synthesized and exposed to X-rays from
different angles. The diffraction pattern is recorded in a sphere around the crystal and from
this three dimensional pattern, the positions of different atoms within the crystal can be
determined.
Detailed descriptions of single crystal diffraction can be found in the books by Giacovazzo58
and Hammond59 and about powder diffraction in the book edited by Bish and Post60.
35
Microscopy
Images of a sample or a sample surface can be obtained using different microscopy
techniques. In conventional optical microscopy, the image is constructed by reflected or
transmitted light, which the eye interprets as an image. Thus, what we see is reflected light,
and the resolution is limited by the wavelength of light. Consequently, objects in the nm range
cannot be imaged using visible light. However, an image of these objects can be constructed
by using e.g. scattering of electrons. The three microscopic techniques described below are all
used to image particles in the nanometre or sub nanometre range. In this work, Scanning
Electron Microscopy (SEM) was used to determine morphological changes and formation of
new phases during the dissolution of manganite (paper 3). High Resolution - Transmission
Electron Microscopy (HR-TEM) was used to characterize the manganite crystals (Figure 5)
for morphology and crystallinity (paper 1). Atomic Force Microscopy (AFM) was used to
study morphological changes in dissolving manganite crystal in real-time (paper 3).
In Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) the
images are constructed from scattered or transmitted electrons. The image in SEM is obtained
by sweeping the surface with a finely focused beam of electrons. This focusing is obtained by
a set of magnetic lenses to form a small spot size. For the highest resolution, a special source
of electrons (field emission source) is used, resulting in a very small spot size and, thus, better
resolution (1 - 5 nm vs. 10 - 50 nm) 61. The scanning is obtained using electromagnetic coils
that deflect the beam in x and y directions. The detected image is a result of backscattered and
secondary electrons from the sample. In TEM, on the other hand, the image originates from
electrons transmitted through the sample and, as a consequence, the inner crystal array and
any structural defects will be visible in TEM images62.
By using an energy dispersive spectrometer (EDS) the elemental composition of the sample
can be determined from x-rays emitted from the sample. When the sample surface is subjected
to an electron beam, x-rays with characteristic energies are emitted from the sample, and these
x-rays are used to determine the elemental composition of the sample.
In AFM, images are obtained by scanning a sharp tip over a surface and recording the
deflection of the cantilever (in “contact mode”). The tip is typically constructed from Si or
Si3N4 and typically has a radius63 of 2 - 20 nm. The lower resolution limit is determined by
36
the radius of the tip and objects smaller than the tip will only show the geometry and structure
of the tip. The software of the instrument controls the movement and the force of the tip
against the surface. In order to obtain high resolution images, piezoelectric scanners are used,
which give very accurate movements along the x-y axis of the surface. Height changes in the
sample will cause deflection of the cantilever as the tip scans the surface. Consequently,
height information can be obtained by measuring this deflection. This is done by focusing a
laser beam on the back of the cantilever and as the cantilever moves up and down, the
reflection of the laser beam will be altered. These changes are measured by a photodiode
detector. Through a feedback loop the software maintains a constant deflection by increasing
or decreasing the force of tip at the surface. Accordingly, the used force will be directly
proportional to the height of the surface and plotting the force at the different points will
produce a topographic image of the surface.
Tapping Mode is another commonly used method of operation in AFM. In this mode the tip is
oscillated near the cantilevers resonance frequency, and as the tip approaches the surface it
will start to “tap” the surface. Height changes at the surface cause a decrease or increase in the
amplitude of the oscillations and through a feedback loop constant amplitude of the
oscillations is maintained. This enables the amplitude to be used as a measure for the
topography of the surface, similar to contact mode. Tapping Mode causes less damage to the
sample surface compared to contact mode and accordingly, it is often used for softer samples
or samples that are easily moved out of place by the tip.
SEM and AFM are two techniques that can give very similar information although the spatial
resolution is often better with AFM. However, they both have drawbacks and thus are often
used complementary to each other to obtain more information and for validation.
An advantage with AFM is that morphological changes can be monitored in real time, but in
SEM, a larger area can be imaged at once. Furthermore, a deeper field of view is obtained on
rough surfaces using SEM. AFM, on the other hand, can give additional physical information
of the surface, such as adhesion, friction, viscoelasticity and surface forces present. Also
sometimes the height is more easily obtained from AFM than from SEM. A slope in SEM
gives an increase of electrons but it might be difficult to say if the slope is directed upwards or
downwards. Furthermore, in SEM, non conductive samples will build up charge at the surface
due to the surface bombardment of electrons. This charge decreases the resolution in the
37
image and to obtain good electric conductivity, the surface is often covered with a thin carbon
or metal film (e.g. gold or iridium) prior to the measurements.
Equilibrium Analysis
In order to describe and predict the outcome of a chemical reaction, the equilibrium constant
for this particular reaction is important (together with knowledge of kinetic effects if such
information is available). From thermodynamic models, knowledge about solution
composition and concentrations at equilibrium can be obtained. Correspondingly, the exact
composition of a solution before and after a reaction enables determination of equilibrium
models64.
Potentiometric Titrations
The experimental setup for precise potentiometric titrations was designed and built at our
department and is thoroughly described in the literature65,66,67. In short, it consists of a titration
vessel immersed in a thermostated oil bath in order to maintain constant temperature
throughout the titration. The titration vessel is purged with Ar gas, to exclude interference
from CO2 and O2, and the addition of acid/base is accomplished by an automated burette. The
potential is measured using a glass electrode, a “Wilhelm” type bridge and a Ag/AgCl
reference electrode. Solutions are stirred with a magnetic stirrer and suspensions by the use of
a small propeller. The use of propellers avoids possible grinding of the solid material by the
stirrer.
The potential of the cell;
- Ag, AgCl(s) | ionic medium || equilibrium solution | glass electrode +
is measured and the concentration of H+ is calculated using Nernst equation (eq. 8);
[ ]
E = E0 + g log H + + E j
(8)
where E is the potential in mV, E0 an apparatus constant determined separately,
g=RTln10/nF=59.16 mV at 25°C and n=1, Ej is the liquid potential which occurs at the
interface between the “Wilhelm” bridge and the equilibrium solution.
38
Batch Adsorption Experiments
Batch experiments for equilibrium analyses were performed in test tubes, an appropriate
amount of ligand and/or metal ions were added to the manganite suspension and pH adjusted.
Care had to be taken not to lower the pH into the region of manganite dissolution. A
combination electrode was used that was filled with 3 M KCl solution. This electrode was
calibrated according to a concentration scale, obtaining -log[H+] instead of pH. A similar
approach as in the potentiometric titrations was used for the calibration, but E0 was
determined from the potential in two diluted H+ solutions with ionic medium. The test tubes
were left on an end-over-end-rotator to equilibrate. Thereafter -log[H+] was measured, the
solution and solid separated through centrifugation and filtration (0.22 µm filter), and
analyses were performed on the solid and the solution.
For adsorption and dissolution experiments, the concentration of metal ions or ligand in the
solution was measured using Atomic Absorption Spectroscopy or scintillation counting as
described in paper 6. The solid, in the form of wet paste, was analysed using the different
spectroscopic and microscopic techniques, described earlier, to obtain information about
quantities and molecular structures of the substances at the surface of the manganite.
Electrophoretic Mobility Measurements
The charge displayed at the surface of particles is important for their behaviour and
adsorption properties. Unfortunately, the surface charge is very difficult to measure
experimentally but it can be indirectly estimated using electrophoretic mobility techniques.
In the electrophoretic mobility measurements, movement of the particles are measured and
recalculated into a potential (zeta-potential) at the shear plane. In aqueous solutions, mineral
surfaces are covered by water molecules. These molecules are electrostatically attracted
towards the surface and exhibit somewhat different structural arrangement compared to the
bulk. In-between this layer and the bulk solution, a shear plane develops. The potential at this
plane reflects the surface charge and potential, but will usually have a smaller numeric value.
However, the change of this potential with pH can be used to determine the iso electric point
of a surface (pHiep), as was done for manganite in paper 2.
39
The potential at the shear plane was determined using the AcoustoSizer method developed by
O’Brian et al.68. This method uses the electroacoustic effect to determine electrophoretic
mobility of particles. When an alternating electric field is applied to a suspension of charged
particles, the particles start to oscillate and thereby creating sound waves. The oscillation is
influenced by the charge and size of the particles together with the frequency of the applied
electric field. The sound is detected as pressure waves at the electrodes in the AcoustoSizer
and recalculated into mobility, size and zeta potentials by the software68.
The experimental setup is explained in O’Brian et al68. In short, it consists of a container
equipped with a combination electrode for measuring pH, a probe for measuring conductivity,
four ceramic rods for maintaining temperature, and a glass propeller. The electrodes that
create the electric field are built into the container walls and are connected to two glass rods
extending on either side of the container. The suspension can be titrated with acid or base to
study changes in zeta potentials with pH.
Modelling
Different theoretical surface complexation models exist that aims to explain surface-ligand
interactions. The most simple model for surface complexation is the constant capacitance
model (CCM) 69 with the 1-pKa approach70,71, which also was the model used to describe the
acid-base properties of manganite in paper 2. The CCM treats the mineral-solution interface
as a parallel plate capacitor where the plates consist, on one side, of ionised surface groups
and, on the other, of counter ions from the electrolyte. The capacitance between the plates is
kept constant, and the model is valid for systems with small total charges or for systems with
high concentrations of the electrolyte3 i.e. where the double layer is compressed. In this
model, adsorbed ions are placed directly at the surface as inner sphere complexes (Figure 19).
.
40
a
surface
+
-
b
+
-
water molecule
Figure 19, Schematic picture of a metal ion in a) inner sphere coordination and b) outer sphere
coordination. Outer sphere complexes are usually due to electrostatic attraction between the surface
and the adsorbate while covalent bonds between the ligand and different surface sites are formed in
inner sphere complexes.
The 1-pKa approach involves one deprotonation step at the surface according to eq. 9, where
Me represents a metal centre at the metal hydroxide surface.
MeOH2+½
MeOH-½ + H+
(9)
An advantage of the 1-pKa model is that the logarithm of the equilibrium constant can be
directly related to experimental data since it corresponds to the pHpzc of the surface. This also
reduces the number of adjustable parameters in the modelling. If the surface species display
similar behaviour as corresponding aqueous groups, a second deprotonation step (resulting in
MeO-1½) should be approximately 10 pH units from the first72. Consequently, the second
deprotonation reaction should be of very little importance in most modelling scenarios.
There also exist more complex models that allow ionic strength effects to be modelled and
counter ions to be placed further away from the surface. These models also enable placement
of different charges of the adsorbate (if e.g. zwitterionic molecules or large organic molecules
are modelled) at different distances from the surface depending on expected interactions.
However, these models introduce more parameters that have to be adjusted to fit the data.
Thus, although the chemical description of the system might be more illustrative with a more
complex model, the use of a simple model with few parameters is often be preferred in many
41
systems. For more reading about surface complexation models, the paper by Lützenkirchen et
al.72 is suggested.
Experimental data were evaluated using the computer codes; LAKE73 and a modified version
of Fiteql 2.074. The modified Fiteql version correctly accounts for dilution factors on solid
concentrations and allow up to 9 surfaces to coexist in a system (J. Lützenkirchen pers.
comm.). The computer codes minimise the error sum of squares, U= (Hcalculated-Hexperimental)2,
to provide the best fit to the experimental data (H represent the total concentration of H+).
Diagrams showing the distribution of different species were calculated and plotted using
WinSGW75, a program package based on the SOLGASWATER algorithm76.
Manganite Surface Processes
The first part of the thesis work concerns the pure manganite mineral, its acid base and
surface charge properties at circum neutral and alkaline pH together with its dissolution at
acidic pH. The second part focuses on the surface complexation of manganite in the presence
of glyphosate and Cd(II).
Surface Charge and Acid-base Properties
When manganite is suspended in an aqueous ionic medium, an initial pH of approximately 6
is obtained. Since this pH is below the isoelectric point (pHiep) at pH 8.2 (see below), the
surface is positively charged and anions will be present at the surface to balance this charge.
In the synthesis MnSO4 is used as starting material and, consequently, sulphate is the most
likely anion at the surface. By analysing the sulphate concentration in the suspensions an
estimate for the positive charge, i.e. the protonation of the surface, could be made and resulted
in 0.55 µmol H+ /m2 manganite (4 % of crystallographic site concentration).
The charge of the manganite surface was investigated in paper 2 using the AcoustoSizer
technique and is illustrated in Figure 20 in the form of zeta potentials.
42
15
ζ-potential (mV)
10
5
pH
0
5.0
6.0
7.0
8.0
9.0
10.0
11.0
-5
-10
-15
-20
Figure 20. Experimentally determined potentials at the shear plane of manganite at three different
ionic strengths,
= 10 mM NaCl,
= 100 mM NaCl (paper 2).
= 30 mM NaCl and
No significant ionic strength dependence could be observed in the pH range 6 - 8. In other
words, the data at the three ionic strengths more or less overlapped within the experimental
uncertainties (Figure 20). The pHiep was recorded at pH 8.2 ± 0.1 in 10 mM NaCl with a small
shift to higher pH at the higher ionic strengths. In the pH range above the pHiep, more
pronounced ionic strength dependence was indicated suggesting surface interaction with Na+
from the ionic medium. This was supported by XPS data (paper 1 & 2), which confirmed the
presence of Na+ at the surface at high pH (Figure 21).
0.08
0.06
Ratio OH/OH
Ratio (Na-Cl)/Mn
2
0.04
0.02
1.6
1.2
0.8
0
5.5
6.5
7.5
8.5
9.5
5
10.5
pH
6
7
8
9
10
11
pH
+
Figure 21. The excess of Na at the manganite
surface versus pH as determined by XPS. The
Na present as NaCl is removed through the
+
Na-Cl subtraction. Na and Cl concentrations
are normalised to Mn. I = 10 mM NaCl. Error
49
bars represent a 10 % error in atomic
concentrations of the components (paper 1 &
2).
Figure 22. The ratio of OH/O at the manganite
surface versus pH as determined by XPS.
I = 10 mM NaCl. Error bars represent a 10 %
49
error in atomic concentrations of the
components (paper 1).
43
Furthermore, XPS data showed a slight increase in the ratio between hydroxide (OH) and
oxide (O) components with increasing pH (Figure 22). This was interpreted as due to the
deprotonation of surface OH2+½-groups forming OH-½ according to eq. 10 (paper 1).
=MnOH2+½
=MnOH-½ + H+
(10)
Potentiometric titrations of manganite suspensions (paper 2) also indicate a low buffer
capacity ( 10-5 M g-1 pH-1) at pH < 7.5 (Figure 23). Furthermore, no or very little ionic
strength dependence was observed in accordance to the zeta-potential measurements. At
higher pH, the buffer capacity increased with increasing ionic strength, which, again, could be
explained as Na+ interactions with the surface.
0.4
H (mM)
0.2
0
6
7
8
9
10
-0.2
-0.4
-0.6
pH
2
Figure 23. Potentiometric titration data for manganite suspensions (10 g/l, 51 m /g) with 100 mM NaCl
(♦) and 30 mM NaCl ( ) ionic medium. The solid lines represent the fit to the measured data by the
acid-base model presented in eq. 10 & 11 (paper 2).
The knowledge of the above mentioned surface characteristics enabled the acid-base
properties for the manganite to be modelled. Potentiometric data were modelled considering a
surface complex between Na+ and the negatively charged sites on the surface. The 1-pK
Model introduced by Bolt and van Riemsdijk70,71 was used, and the proton affinity constant
was set equal to the pHiep obtained from the electroacoustic measurements at 10 mM NaCl.
44
This low ionic strength was chosen to avoid the slight Na+ effect seen in pHiep. The constant
capacitance model applied contains only two adjustable parameters, the capacitance C and the
affinity constant for the interaction between Na+ and singly coordinated oxygen atoms at the
surface. The fit between the model and the data is shown in Figure 23 and the established
equilibria in eq. 10 and 11, (C = 0.39 F/m2).
=MnOH2+½
=MnOH-½ + H+
=MnOH2+½ + Na+
=MnOHNa+½ + H+
log β0 (intr.) = -8.20
(10)
log β0 (intr.) = -9.64
(11)
As expected, the Na+ complex is weak as indicated by the formation constant for
=MnOHNa+½ from =MnOH-½ , which is: logK0 (intr.) = -1.44. The distribution of surface
sites with pH can be seen in Figure 24. It can be noted that some 10 % of the surface site
concentration is found as the Na+ - surface complex at pH 10 in 100 mM NaCl. This model is
also supported by the XPS data showing a low affinity for Na+ at the surface. At pH =10 the
Nasurface/Mn ratio is approximately 5 in 10 mM NaCl (Figure 21), which is fairly close to the
percentage in the model (paper 2).
Fi
0.6
+½
=MnOH2
0.5
-½
0.4 =MnOH
0.3
0.2
+½
0.1
=MnOHNa
100 mM
0
0.6
+½
=MnOH2
0.5
0.4
=MnOH
-½
0.3
0.2
0.1
+½
=MnOHNa
10 mM
0
6
7
8
pH
9
10
11
Figure 24. The distribution of surface sites at the manganite surface in 100 mM NaCl and 10 mM NaCl
acquired from the acid-base model described in the text. The broken lines illustrate an extrapolation of
the model (paper 2).
45
To our knowledge, only one previous publication has been made describing the acid-base
characteristics of the manganite surface5. In that study the acid base properties were explained
using a 2-pKa model§ with a capacitance of 1.02 ± 0.1 F/m2 in 10 mM ionic medium. The
intrinsic protolysis constants were found to be pKa1 = 5.72 and pKa2 = 6.66, which resulted in
pHpzc at pH 6.2. This is much lower than the pHpzc found for the manganite material used in
this thesis.
Additionally, a small number of studies exist describing the location of pHpzc/iep of the
manganite surface (Table 2).
Table 2. Reported values for pHpzc or pHiep
Weaver et al.
77
Shaughnessy et al.
5
Xyla et al.
Matocha et al.
79
This work, paper 2
78
pHpzc/ iep
method
MnOOH origin
5.4
titration
natural
7.4
titration
natural
6.2
titration
synthetic
6.3
electrophoretic mobility
synthetic
8.2
electrophoretic mobility
synthetic
The previous data in Table 2 or the acid base model could not be supported by the new work
presented in this thesis, perhaps due to surface contaminations (e.g. sulphate) and differences
in origin of manganite material. However, the agreement between the different techniques
used in paper 2 gives us reason to believe that pHiep is situated at pH 8.2.
§
A model assuming the deprotonation of the surface to occur in two steps with the surface groups=OH2+, =OH
and =O-.
46
Dissolution at Low pH
Below pH 6, the dissolution of manganite becomes significant (Figure 25). During the
dissolution, manganite is proposed to disproportionate to form Mn(II) ions in solution and a
Mn(IV) oxide in the solid phase41 according to eq. 12.
2MnOOH(s) + 2H+
MnO2(s) + Mn2+ + 2H2O
(12)
3
2+
[Mn ] (mM)
2.5
2
1.5
1
0.5
0
3.5
4
4.5
5
5.5
6
pH
Figure 25. The dissolution of manganite at low pH, illustrated by the aqueous concentration of Mn(II)
after 24 h of reaction time (paper 1).
In paper 1, the remaining solid after the dissolution process was examined using XPS. It was
very difficult to distinguish Mn(IV) in spectra that contained large quantities of Mn(III).
However, in a manganite suspension at pH 1.5 (three days reaction time), the presence of
Mn(IV) could be detected. Mn(IV) was observed as a change in peak shape of the Mn 2p
peak, as well as an altered ratio between the hydroxide and oxide components of the O 1s
spectra. Comparing the XPS spectra at pH 1.5 with a mixture of MnOOH and β-MnO2
(~50/50 % by volume), a similar change is observed in Mn 2p and O 1s spectra (Figure 26 &
27), which supports the formation of a Mn(IV) phase in manganite suspensions at low pH.
47
Intensity (counts/s)
35
x 10
30
2
30
25
a)
25
2
b)
20
20
15
15
10
10
5
660
40
Intensity (counts/s)
x 10
35
30
x 10
650
640
660
650
640
2
30 x 10
2
25
c)
d)
20
25
15
20
15
10
10
5
660
650
660
640
Binding energy (eV)
650
640
Binding energy (eV)
Figure 26. Mn 2p XPS spectra of a) manganite, b) pyrolusite, c) 50/50 mixture manganite/pyrolusite
and d) manganite paste at pH 1.5, showing the changes in peak shape as a consequence of MnO2
formation (paper 1).
2
45 x 10
2
x 10
Intensity (counts/s)
30
a)
40
b)
35
25
30
20
25
20
15
15
10
10
5
5
544
540
536
532
528
544
540
Binding energy (eV)
48
536
532
528
2
50 x 10
45
2
x 10
30
c)
Intensity (counts/s)
40
d)
O
25
35
OH
20
30
25
15
20
15
H2O
10
10
5
5
544
540
536
532
544
528
540
536
532
528
Binding energy (eV)
Figure 27. O 1s XPS spectra of a) dry manganite, b) pyrolusite, c) mixture 50/50 manganite/pyrolusite
and d) manganite paste at pH 1.5, showing the altered OH/O ratio as a consequence of formation of
MnO2 (paper 1).
In paper 3, the dissolution of manganite at low pH was examined further. Using literature
equilibrium constants11 and the WinSGW computer code75,76, the pH dependency of the
dissolution of manganite was calculated (Figure 1b). It was found that for a suspension with
10 g of manganite/dm3, the phase transformation should be completed below pH 4.8.
Experiments were performed in the pH range 5.0 – 6.8 and although the
dissolution/disproportionation proved to be a slow process, equilibrium seemed to be reached
after one year. From solution data an equilibrium constant for eq. 12 was obtained
(Figure 28).
0
2+
log [Mn ]
log [Mn2+] = 7.40 + 2log[H+]
-2
-4
-6
-8
5
6
+
7
8
- log [H ]
+
Figure 28. Plot of log Mn vs –log[H ] showing the alignment of experimental data to a theoretical line
with a slope of 2. The line crosses the x-axis at 7.40, which represents the log K for eq. 12 in 100 mM
ionic strength (paper 3).
49
The formation of pyrolusite (β-MnO2) was after nine months observed in XRD patterns of a
sample at pH = 2.6, which suggests that eq. 12 can be written as:
2 γ-MnOOH + 2H+
β-MnO2 + Mn2+ + 2H2O
log K0 = 7.61 ± 0.10
(13)
where log K has been recalculated to zero ionic strength using the Davies equation. With
time, this phase transformation resulted in a colour change from brown to black, and as a
consequence, a colour gradient could be seen in the suspensions. The gradient represented the
ratio between the brown manganite and the black pyrolusite and appeared to be relative to the
initial amount of H+ added.
The morphological changes as the crystals dissolve were investigated using AFM. In real-time
AFM measurements, it was observed that the dissolution starts at the short edges of the
manganite crystals (Figure 29). Over a period of 2 hours, this results in a shortening of the
manganite crystals while their width is maintained.
Figure 29. Real-time AFM images of dissolving manganite crystals, showing the shortening of the
crystals while their width is maintained. The left picture shows a manganite crystal just before the
addition of acid, and the right picture after 120 minutes of reaction time (paper 3).
XRD patterns of partly dissolved manganite samples between pH 3.8 and 1.6 (reaction times
of one week) displayed a small peak at ~4.10 Å, suggesting the presence of ramsdellite
(MnO2) (Figure 30a & c). In SEM images of corresponding samples, needle shaped crystals
with smaller dimensions compared to the manganite crystals (Figure 30b & d) appeared,
50
which probably represent ramsdellite. In samples at pH < 1.6, the presence of pyrolusite was
indicated by an increasing reflection at 3.10 Å (Figure 30c & e).
#2
a)
10
Intensity (Counts) X 100
41- 1379
MANGANITE
b)
8
6
4
4.10 Å
2
10
20
30
40
50
2-Theta Angle (deg)
#5
41- 1379
MANGANITE
c)
d)
7
6
5
4
3.10 Å
Intensity (Counts) X 100
8
3
2
1
10
10
20
30
40
50
2-Theta Angle (deg)
#6
41- 1379
MANGANITE
e)
f)
4
4.40 Å
6
5.65 Å
Intensity (Counts) X 100
8
2
10
20
30
40
50
2-Theta Angle (deg)
Scale bar
Figure 30. XRD patterns and SEM images of samples at pH=3.8 (a & b), pH=1.6 (c, d & f) and pH=0.9
(e). The presence of manganite is indicated with lines in all XRD patterns and seen as the larger
needles in the SEM images. b) and d) show ramsdellite as small needles, f) shows manganese
chloride in the form of hexagons. Scale bar: 1.20 µm in b), 600 nm in d) 1.50 µm and in f) (paper 3).
51
However, the pyrolusite could not be seen in SEM images. The only “new” phase at pH 1.6
had a morphology similar to that of MnCl2 × 4H2O crystals (Figure 30f). This phase was
considered to be a precipitate that fell out of solution during the drying procedure. Small
amounts of MnCl2 were also seen in the XRD patterns (5.65 and 4.40 Å) at pH 0.9
(Figure 30e), which supports this hypothesis. The difficulties in recognising a pyrolusite
phase in the SEM images can suggest that the pyrolusite phase is formed through a
rearrangement of the ramsdellite or manganite phases. This hypothesis is supported by the
complete rearrangement into pyrolusite in the pH = 2.6 sample described above, and is in
agreement with previous observations38,80.
Surface Complexes with Glyphosate and Cd(II)
At pH < pHpzc, the manganite surface displays a net positive charge. In this pH range,
manganite has an increased possibility to attract negatively charged ligands such as oxalate5,
ascorbic acid7, phenolic compounds81, aminocarboxylates82 and arsenate83. The research
presented in papers 4 and 6 shows that glyphosate forms bonds to the manganite surface in
this pH region.
Manganite-Glyphosate
N-(phosphonomethyl)glycine (PMG, H3L) with the trivial name glyphosate is one of the most
commonly used organophosphorus herbicides in the world today due to its low toxicity to
mammals, fast degradation, and inactivation in soils84,85,86,87. In solid phase and in solution,
glyphosate molecules display a zwitterionic structure so that a proton is dislocated from the
phosphonate group to the amine group88,89 (Figure 31). The presence of three donor groups;
the carboxylate, the amine and the phosphonate group, leads to interesting complexation
behaviour with metal ions and mineral surfaces.
52
pKa3 = 10.1
-
+
pKa2 = 5.44
pKa1 = 2.22
H3L
Figure 31. Protonated glyphosate molecule showing the zwitterionic structure and the three
90
deprotonation constants (I=0.1M) . From the left the structure represents; HO3P-CH2-NH2-CH2-CO2H
91
(colour version on p 68, modified from Sheals et al ).
The inactivation of glyphosate in soils is mainly through adsorption onto mineral surfaces.
Although this adsorption decreases the toxicity to plants and the mobility of the herbicide, the
interactions can also affect the degradation of the glyphosate92,93,84. This can result in
accumulated amounts of glyphosate in soils and perhaps also in ground waters over time.
Previous studies show that the glyphosate mainly adsorbs through the phosphonate
group84,94,95. Due to the preferred interactions between hard acids and hard bases it is not
surprising that high affinities for Fe-94,95 and Al-minerals96 have been observed, since
phosphonate is a relatively hard base. As a consequence of the phosphonate interactions,
phosphate ions compete with glyphosate for binding sites in phosphate rich soils and thereby
increase the mobility of glyphosate 84,97.
Very little research has been performed concerning the adsorption of glyphosate onto
manganese (hydr)oxides, even though they are a common component in natural soils24. Mn2+
ions added to soil have been reported to increase the degradation of glyphosate, while Fe3+
and Al3+-ions added to soil decrease the degradation92. An explanation for this was suggested
to be due to changes in the bioavailability of glyphosate to micro-organisms. The difference
between these ions raises the question about possible differences in the reactivity of
glyphosate with the surfaces of FeOOH, AlOOH and MnOOH minerals.
In paper 4 and 6, the adsorption of glyphosate onto manganite was examined and compared to
the adsorption onto goethite. It was found that glyphosate adsorbed to the manganite surface
over the entire pH range studied (Figure 32) but that the concentration of glyphosate
decreased at the surface with increasing pH (paper 4 & 6). The glyphosate molecule displays
a net negative charge above pH ~2 (pKa(H3L) = 2.22) 90 and, thus, is expected to be more
53
attracted to a positively charged surface. The continued adsorption at high pH to the net
negatively charged surface could be a result of formation of covalent bonds between the
ligand and the surface, and/or electrostatic effects due to the zwitterionic properties of the
Atomic ratio N/Mn
molecule, which remain until pH~10 (pKa(HL2-) =10.1)90.
0.08
0.06
0.04
0.02
0
6
7
8
pH
9
10
Figure 32. Amount of adsorbed glyphosate on the surface of manganite. 2.4 µmol glyphosate were
2
added per m of manganite. Data points represent XPS N 1s spectra normalised against Mn 2p.
From the infrared spectra of glyphosate at the surface of manganite (Figure 33), it can be seen
that the carboxylate and amine groups do not form bonds to the surface since they display
spectra similar to aqueous glyphosate complexes. Thus glyphosate is likely bonded to the
manganite surface through the phosphonate group. For glyphosate adsorbed onto goethite,
the main information about the mineral – glyphosate interaction was derived from the
phosphonate bands 94. Unfortunately, these bands overlap with strong manganite peaks. Thus,
no direct evidence of bond formation between the phosphonate group and the surface could be
extracted from infrared measurements. However, the structure of glyphosate at the manganite
surface is here assumed to be similar to the structure of glyphosate at the surface of goethite.
54
Intensity
a)
b)
c)
1800
1400
1000
/ cm
-1
Figure 33. FTIR spectra of glyphosate; a) glyphosate at the surface of manganite, (pH 7.1 and
2
2[PMG]tot= 9.8 µmol/m manganite), b) aqueous glyphosate in the form of HL and c) fully deprotonated
3aqueous glyphosate (L ). The lines illustrate the presence and absence of carboxylate hydrogen
-1
bonded to the amine group of the glyphosate. In spectrum a) the vibrations below ~1200 cm were
excluded due to the strong overlapping manganite peaks.
From XPS and FTIR measurements, the presence of both protonated and deprotonated
glyphosate was observed at the manganite surface throughout the pH range (Figure 33 and 34)
studied. However, compared with the goethite surface, a higher percentage of the glyphosate
0.05
0.04
0.04
0.03
0.03
0.02
0.02
0.01
0.01
Atomic ratio N/Fe
Atomic ratio N/Mn
was protonated when bound to manganite, as shown in Figure 34.
0
0
6
7
8
pH
9
3
10
4
5
6 7
pH
8
9
Figure 34. XPS data showing the amount of protonated ( ) and deprotonated ( ) glyphosate at the
2
surface of: a) manganite with 2.4 µmol glyphosate added per m of manganite, b) goethite with 2.3
2
94
µmol glyphosate added per m of goethite (paper 4) .
Furthermore, the manganite surface did not adsorb glyphosate from the solution as efficiently
as the goethite surface at similar surface loadings. At pH 7, only ~50 % of the glyphosate was
55
adsorbed at the surface in the manganite-glyphosate system (unpublished data) compared to
~65 % in the goethite-glyphosate system (paper 4)94.
Manganite-Glyphosate-Cadmium(II)
In order to understand the chemistry at the mineral-water interface, it is very important to also
have an understanding of the processes occurring in the bulk solution. All the reactions
occurring at the interface are coupled with those in solution. Thus, the chemistry of the
surface species will be influenced by the chemistry in solution, and vice versa.
The speciation and equilibria of different Cd(II)-glyphosate complexes in solution are
presented in paper 5. Stability constants were determined and agree well with previous
formation constants in other ionic media98. The equilibrium analysis showed that Cd(II) exists
in solution in the form of 1:1 and 1:2 Cd(II)-glyphosate complexes at pH > 3.0 (Figure 35).
Fi 1
CdL-
Cd2+
0.8
CdL24-
0.6
0.4
CdHL
CdOHL2-
0.2
0
3
4
5
6
7
8
9
10
pH
+
Figure 35. Distribution of Cd(II) in the glyphosate-Cd(II)-H system (4 mM glyphosate, 2 mM Cd(II),
paper 5).
FTIR spectra of the two dominating complexes (CdL- & CdL24-) show an increased separation
between the symmetric and asymmetric carboxylate stretching vibrations of the glyphosate,
which has been shown to indicate monodentate coordination to a metal ion99 (Figure 36).
Furthermore, no broadening was observed in the asymmetric carboxylate peak
(at ~1600 cm-1), which is interpreted as a lack of hydrogen bonding between the carboxylate
and amine groups (Figure 36a). The changes in the phosphonate bands also indicate altered
bond structure, suggesting that the phosphonate group is coordinated to the Cd(II) ion.
56
a)
Intensity
b)
c)
d)
1800
1400
1000
/cm
-1
2-
3-
Figure 36. FTIR spectra of a) partly protonated glyphosate HL b) fully deprotonated glyphosate, L c)
4CdL d) CdL2 . Showing the increased split between carboxylate stretching vibrations and the changes
in phosphonate vibration in the Cd(II)-glyphosate complexes (paper 5).
EXAFS data show that the coordination environment around the Cd(II) ion is very similar for
the two dominating aqueous complexes (Figure 37). EXAFS data for both complexes could
be fitted with a model including single scattering paths from Cd-O/N, Cd-C and Cd-P together
with a multiple scattering path from Cd-C-O. In this model, the first shell consists of five
oxygen atoms and one nitrogen atom, the second shell consists of three Cd-C and one Cd-P,
and the third shell consists of a multiple scattering path corresponding to Cd-C-Odistal (Figure
38). In the Cd(II):glyphosate 1:2 complex, the second shell grew in magnitude due to the
presence of the two glyphosate molecules in the complex. Furthermore, an increased disorder
was observed in this complex, which is most probably caused by the large number of isomers
expected in these types of complexes100.
57
Cd-O/N
FT magnitude
Odistal
Cd-C
Cd-P
C
Cd(II)
N
C
Cd-C-O
P
C
0
2
4
6
R/Å
-
Figure 38. CdL complex. (paper 5)
version on p 68)
Figure. 37. Fourier transforms of EXAFS data
4for CdL (grey) and CdL2 (black) complexes.
The fit to the data is represented by the
broken lines (paper 5).
101
. (colour
Thus, it was found that the structure of the 1:1 complex displays two chelating rings formed
between the Cd(II) and the carboxylate, amine, and phosphonate groups in the glyphosate
(Figure 38). The 1:2 complex had a similar chelating structure, but with two glyphosate
molecules per Cd(II).
This chelating-ring structure is the most common way in which glyphosate coordinates to
metal ions (Cu(II)91,102, Pt(II)103, Co(III)100, Cr(III)104, La(III), Ce(III), Nd(III), Er(III)105,
Pt(IV)106, V(IV)107, U(VI)108). However, glyphosate coordinates to some metals through a
bidentate chelating structure where one of the functional groups in glyphosate does not
interact with the metal. This has been described for crystallised complexes with Fe(III)109,
Ba (II), Sr(II)110 or Na(I)111 where the amine group is uncoordinated, in alkaline solutions
with Pt(II) where the carboxylate group is not complexed, and in acidic solutions with Pt(II)
where the phosphonate group is free103. A third coordination mode, is where the metal ions
form bonds only through the phosphonate group, as for complexes with Fe(III) in solution109
or Ca(II)112,113 in the solid phase. Eu(III)114 and Ag(I)111 display polymeric structures with
glyphosate, where glyphosate molecules form bridges between metal ions through the
phosphonate group on one side and the carboxylate group on the other.
58
In paper 6, the co-adsorption of Cd(II) and glyphosate was studied at pH 6 -10 and compared
to adsorption in the corresponding binary systems (Cd(II)-manganite and glyphosatemanganite). In the binary Cd(II)-manganite system, Cd(II) adsorption increases with pH, and
an adsorption edge is found at pH 8 - 9 (Figure 39a). This agrees well with previous
adsorption studies of metals such as, Pb(II)79, Cd(II) 115, Zn(II)116,117 and Pu(VI)78 onto
manganite where the metal ions display an increased adsorption at neutral to alkaline pH
values. The increased adsorption also corresponds well with the development of a net
negative charge at the surface above the pHpzc = 8.2.
In the glyphosate-manganite system, glyphosate is adsorbed throughout the pH range studied
(as was discussed earlier), and this is shown in Figure 39a.
µmol ads/m2
8
a)
8
Cd
PMG
6
b)
Cd
Cd & PMG
6
4
2
2
2
1
7
8
pH
9
10
6
7
8
pH
PMG & Cd
PMG
3
4
6
c)
4
9
10
6
7
8
9
pH
Figure 39. Adsorption of glyphosate and Cd(II) on the surface of manganite as a function of pH.
2
[PMG]tot and/or [Cd(II)]tot = 14.6 µmol per m of manganite. Note the different scales on the y-axis
(paper 6).
In the presence of glyphosate, Cd(II) adsorption was increased at pH < pHpzc (Figure 39b).
This could be a consequence of several processes: i) an increased electrostatic attraction for
the surface due to the presence of adsorbed negatively charged glyphosate, ii) the formation of
ternary surface complexes and/or iii) the formation of a Cd(II) containing precipitate. In the
region pH > pHpzc, the adsorption of Cd(II) in the ternary system was less than the adsorption
observed in the binary system. This can be explained by the competing formation of
negatively charged Cd(II):glyphosate complexes in solution that display a decreased affinity
for the surface. The presence of Cd(II) increased the adsorption of glyphosate over the entire
pH range studied (Figure 39c), which also can be explained by the corresponding processes i)
- iii) described above.
59
10
Using FTIR and XPS, the speciation of the glyphosate at the surface of manganite could be
studied with respect to protonation and Cd(II) interactions. FTIR data have shown that both
the formation of aqueous Cd(II)-glyphosate complexes and the protonation of glyphosate
affect the glyphosate carboxylate frequencies. In XPS, the nitrogen spectra can indicate
whether the amine group is protonated or not. In the binary glyphosate-manganite system,
XPS results showed that the main part of the glyphosate was protonated throughout the pH
range investigated (Figure 34). Thus, the complete absence of such protonation of glyphosate
at the manganite surface indirectly suggests a Cd(II) interaction.
FTIR spectra of the Cd(II)-glyphosate-manganite system at high pH have a similar
appearance to the spectra of aqueous CdL- complex, indicating that glyphosate is coordinated
to Cd(II) in these surface species (Figure 40a & f). In the low pH samples, the carboxylate
bands were broadened compared to these bands at high pH (Figure 40a & c), which was also
observed in the spectra of the aqueous manganite-glyphosate complexes. Subtraction of the
spectrum of the binary glyphosate-manganite system (d) from the spectrum of the ternary
Cd(II)-glyphosate-manganite system at a similar pH (c) reveals the presence of a glyphosate
species coordinated to Cd(II) (Figure 40e & a). These results suggest the presence of a ternary
Cd(II)-glyphosate-manganite complex at high pH and two coexisting complexes at low pH,
which are believed to be a ternary Cd(II)-glyphosate-manganite and a binary glyphosatemanganite complex.
60
Abs
Ratio N/Mn
0.1
0.05
a
0
b
6
c
e
1500
1400
1300
9
10
Figure 40. (left) FTIR spectra of
a) Cd(II)-glyphosate-manganite at pH 10,
b) Cd(II)-glyphosate-manganite at pH 8.0,
c) Cd(II)-glyphosate-manganite at pH 6.7,
d) glyphosate-manganite at pH 6.5.
e) subtraction c-d,
f) aqueous CdL (paper 5 & 6).
f
1600
8
pH
Figure 41. (above) The speciation of the amine
group obtained from XPS analyses. The filled
triangles represent the protonated amine group
+
(>NH2 ) and the empty triangles represent the
deprotonated amine group (>NH). The atomic
percentages for N are normalised to the atomic
percentage of Mn (paper 6).
d
1700
7
1200
/ cm-1
An XPS study was conducted to determine the protonation of the amine group of surface
bound glyphosate in the ternary Cd(II)-glyphosate-manganite system, and the results are
shown in Figure 41. At high pH, XPS spectra only showed the presence of deprotonated
amine, which indirectly indicates glyphosate coordination to Cd(II). At pH < 8 the protonated
form is detectable in the XPS spectra. This indicates that a fraction (~30 - 40 %) of the
glyphosate is present in the form of a protonated surface species at low pH but not at high pH.
These results support the FTIR results discussed above.
XPS and batch adsorption experiments were performed to quantify the amount of Cd(II) and
glyphosate at the surface of the manganite. The total Cd(II) adsorbed is shown in Figure 42,
calculated both from solution and from XPS atomic ratios at the surface. These results show
that at pH 6.7 the ratio between Cd(II) and glyphosate adsorbed is 1:1, but with increasing pH
this ratio increases to about 2:1.
61
3
0.08
2
0.04
1
0
Atomic ratio
µmol/m2 adsorbed
0.12
0.00
6
7
8
pH
9
10
11
Figure 42. The concentration of Cd(II) and glyphosate at the surface of manganite.
represent Cd
concentration from XPS (normalised as Cd/Mn), × represent adsorbed Cd from solution data,
represent N concentrations from XPS i.e. the glyphosate concentration (normalised as N/Mn) and
represent adsorbed glyphosate from solution data. A error < 10% can exist in XPS atomic
49
2
concentrations . A total concentration 4.8 µmol of both glyphosate and Cd(II) was added per m of
manganite, which represents 37 % of the crystallographic site density (paper 6).
From batch adsorption, XPS and FTIR data, the presence of a Cd(II)-glyphosate complex was
observed at the surface of manganite throughout the pH range studied. An additional
protonated glyphosate complex was detected at low pH. Considering the 1:1 Cd(II)glyphosate ratio at low pH and the 2:1 ratio at high pH, the presence of a third surface
complex must be assumed. At low pH, the concentration of this third complex would be
similar to the concentration of the protonated glyphosate complex (~30 - 40 % of the Cd(II))
in order to generate the 1:1 Cd(II):glyphosate ratio at the surface.
To resolve this, the speciation from the perspective of Cd(II) was studied using EXAFS
spectroscopy. EXAFS data for the surface species in the ternary Cd(II)-glyphosate-manganite
system were compared with EXAFS data for the aqueous Cd(II)-glyphosate complexes
described above (and in paper 5). The general appearance of these datasets was very similar
but with a significantly decreased second shell in the surface species (Figure 43).
62
a
b
c
d
0
1
2
3
4
R/Å
5
6
7
8
Figure 43. Fourier transforms of a) Cd(II)-glyphosate-manganite at pH 6.7, b) Cd(II)-glyphosatemanganite at pH 8.0, c) Cd(II)-glyphosate-manganite at pH 10, and d) CdL (aq). The broken lines are
the results from the R-space fits.
The above hypothesis of the coexistence of Cd(II) containing ternary and binary surface
complexes implies that on average less than one glyphosate per cadmium is available for
coordination. This explains the decreased amplitude for the second shell compared to solution
data. The structure of the complexes was modelled assuming six O/N in the first shell and a
second shell with a combination of Cd-Mn back-scattering and Cd-C + Cd-P back-scattering.
The best fit to data was obtained for a scenario where 40 % of the Cd(II) was in the form of a
binary Cd(II)-manganite surface complex, and the remaining 60 % was in the form of a
ternary Cd(II)-glyphosate-manganite complex in which the Cd(II) is not directly bound to the
surface. The absence of a direct interaction between Cd(II) and the surface suggests that the
ternary complex is of type B (surface-glyphosate-Cd(II)) rather than of type A (surfaceCd(II)-glyphosate). Furthermore, the presence of P in the second coordination shell indicates
that the ternary complex displays a structure similar to the aqueous Cd(II)-glyphosate
complexes in that the two chelating rings between Cd(II) and glyphosate are intact. This
mode of surface coordination is also observed in Cu(II)-glyphosate-surface complexes on
aluminium (hydr)oxides118. The EXAFS results were also in good agreement with XPS data,
which indicated the presence of 30 - 40 % of the binary glyphosate-manganite complex. The
absence of Cd-Cd distances rule out the possibility of surface precipitation and the Cd-Mn
63
distances obtained suggest edge sharing between Cd and Mn octahedral in the binary Cd(II)surface complex, similar to what has been observed in previous studies115.
From the combined solution-, XPS-, FTIR- and EXAFS data, the presence of three different
surface complexes was established. At low pH approximately 60 % of glyphosate and Cd(II)
is present at the surface in the form of a ternary surface complex and 40 % in the form of
binary surface complexes (Figure 44). Calculations of the solution composition showed that a
1:1 ratio between Cd(II) and glyphosate exists both at the surface and in solution in the lower
pH range.
c)
a)
b)
Hydrogen
bonded
Figure 44. Possible structures for the identified surface complexes at the manganite surface (not to
scale); a) type B Cd(II)-glyphosate-manganite ternary surface complex, b) binary Cd(II)-surface
complex and c) binary glyphosate-surface complex (with a structure similar to the one described on
94
goethite by Sheals et al ). Hydrogen atoms have been omitted for clarity (paper 6).
With increasing pH the binary glyphosate-manganite complex desorbs from the surface
(complex c in Figure 44) while the concentration of e.g. the CdL24- complex in solution
64
increases. Apart from electrostatic effects between the complexes in solution and the surface,
these solution complexes successfully compete with the surface for the glyphosate, resulting
in a Cd(II):glyphosate ratio of 2:1 at the surface and 1:2 in solution.
The structure of the ternary complex can be compared to results from previous studies, e.g.
Cu(II)-glyphosate at the surface of goethite90, gibbsite and bayerite118. In the Cu(II)glyphosate-goethite system, the structure of the ternary complex changed with pH. At low pH,
a type B ternary complex was formed and at high pH it changed into to a type A ternary
complex90. The coordination around the metal ion was similar to the aqueous 1:1 Cu(II)glyphosate complex in that the bonds between the amine and carboxylate group were kept
intact at the surface of goethite. However, the chelating ring formed by bonds to the amine
and the phosphonate opened due to the phosphonate interaction with the surface. In the
Cu(II)-glyphosate-gibbsite and Cu(II)-glyphosate-bayerite systems, the structure of the
ternary complex was of type B and did not change with pH. Both of the chelating rings were
kept intact despite the phosphonate interaction with the surface118 similar to Cd(II)-glyphosate
complexes on manganite.
Summary
By combining XPS, potentiometry and electrophoretic mobility measurements, a simple
1-pKa acid-base model for the manganite surface was established that accounts for a weak
Na+-surface interaction. Furthermore, this acid-base model explains the low proton buffering
capacity of manganite at pH > 6, the net positive charge at pH < 8.2, and the net negative
charge at pH > 8.2 (pHiep = 8.2).
The dissolution of manganite at pH < 6 was studied at equilibrium (after one year) and in nonequilibrated samples. After one year of equilibration, stable [H+] readings were obtained.
Together with XRD data, this enabled an equilibrium constant for the manganite
disproportion into pyrolusite (β-MnO2) and Mn(II)-ions to be established. Non-equilibrium
microscopy measurements indicated that the dissolution initially starts at the apices of the
crystals. Furthermore, XRD measurements indicated that ramsdellite (MnO2) can be formed
in non-equilibrium processes, but that it rearranges into pyrolusite with time.
65
Increased Cd(II) adsorption at a net negative manganite surface (pH > 8.2) was observed in
wet chemical adsorption studies. EXAFS revealed that this adsorbed Cd(II) formed an inner
sphere complex by edge sharing between Mn and Cd octahedra. The adsorption of glyphosate,
which displays an overall negative charge at pH > 2, took place over the entire pH range
although it decreased with increasing pH. This can be explained by the strong interaction
between the phosphonate group and the surface, in combination with the zwitterionic
properties of the molecule. The structure of this binary complex was explained using FTIR
and XPS data. It was described as an inner sphere complex where the phosphonate group in
glyphosate forms bonds to the surface. This shows a preference for an interaction between the
surface and the phosphonate group, instead of the carboxylate or amine groups. From XPS
data it was observed that the largest fraction of the glyphosate-manganite surface complexes
displayed a protonated amine group up to pH 10.
In the presence of both Cd(II) and glyphosate, the overall adsorption at the manganite surface
was increased. At low pH, Cd(II) adsorption was increased in the presence of glyphosate. This
was explained by the formation of a ternary surface complex and altered electrostatics at the
mineral interface. At high pH, Cd(II) adsorption was decreased in the presence of glyphosate.
This is thought to be due to electrostatic repulsion between negatively charged species in
solution and a net negatively charged surface, as well as the formation of aqueous complexes
in solution that compete with the surface for Cd(II) (and glyphosate). The presence of Cd(II)
increased the concentration of surface-bound glyphosate throughout the pH range which was
explained by formation of ternary surface complexes and altered electrostatics at the interface.
By combining several techniques, the structure of the formed complexes was derived. XPS
and FTIR were used to study changes in the glyphosate, and EXAFS to study the Cd(II)
coordination environment. The data indicated the formation of a ternary type B surfaceglyphosate-Cd(II) complex throughout the pH range studied. EXAFS studies showed that the
chelating structure between Cd(II) and glyphosate, as seen in aqueous Cd(II)-glyphosate
complexes, was maintained at the surface. This structure involves the carboxylate, amine and
phosphonate groups coordinating to the Cd(II)-ion. The Cd(II)-glyphosate complex was
bound to the surface through the phosphonate group of the glyphosate molecule.
This research shows the importance of understanding both the chemistry in solution and at the
mineral surface in order to describe surface processes. Furthermore, it seems likely that not
66
only aluminium (hydr)oxides and iron (hydr)oxides, but also manganese (hydr)oxides have
the possibility to influence glyphosate mobility and degradation.
Future Research
•
Development of thermodynamic surface complexation models that give a more
complete understanding of the interactions between the manganite surface, glyphosate
and Cd(II).
•
Preliminary experiments have shown that although glyphosate inhibited Cu(II)hydroxide precipitation on the goethite surface, it did not do so at the manganite
surface. The nature of this precipitate (polymerisation) would be interesting to study
and compare with Cu(II)-glyphosate-goethite complexes. Also, can this precipitation
be avoided under different experimental conditions?
•
Further development of the low temperature XPS measurements, with respect to
sample pre-treatment, to avoid contributions from the solution in XPS spectra and to
minimise evaporation.
67
-
+
Figure 31 (page 53 in the thesis). Protonated
glyphosate molecule showing the zwitterionic
structure. Red atoms are oxygen, blue
nitrogen, black carbon, white hydrogen and
pink phosphorous
91
(modified from Sheals et al ).
Figure 6 (in paper 5). The molecular structure
of Cd9(PMG)6(H2O)12 × 6H2O. The coordination
geometries of the cadmium ions and the
phosphonate groups are represented by grey
octahedra and pink tetrahedra, respectively.
Hydrogens have been omitted for clarity. The
remaining atoms were colour coded; blue for
N, red for O and black for C.
-
Figure 38 (page 58 in the thesis). CdL
complex. Red atoms represent oxygen, blue
nitrogen, black carbon, pink phosphorous and
grey cadmium(II), hydrogens have been
omitted (also Figure 4 in paper 5).
68
References
1
Stumm W, Morgan JJ, Aquatic Chemistry, 3rd Edition, John Wiley & Sons, New York, 1996
Banwarth, S. A.Aqueous Speciation at the Interface Between Geological Solids and Groundwaters, Chapter 7
in Modelling in Aquatic Chemistry, Nuclear Energy Agency, Organisation for Economic Co-operation and
Development (OECD), Paris, France, 1997
3
Stumm, Werner. Chemistry of the solid-water interface – Processes at the mineral-water and particle-water
interface in natural systems, John Wiley & Sons, 1992,
4
Brown, G. E. Jr. Henrich, V.E. Casey, W. H. Clark, D. L. Eggleston, C. Felmy, A. Goodman, D.W. Grätzel,
M. Maciel, G. McCarthy, M. I. Nealson, K. H. Sverjensky, D. A. Toney, M. F. Zachara, J. M. Metal oxide
surfaces and their interactions with aqueous solutions and microbial organisms, Chem. Rev. 1999, 99, 77-174
5
Suter, D. Banwart, S. Stumm, W. Dissolution of Hydrous Iron(III) Oxides by Reductive Mechanisms,
Langmuir, 1991, 7, 809-813
6
Xyla AG, Sulzberger B, Luther GW III, Hering JG, Van Cappellen P, Stumm W, Reductive Dissolution of
Manganese (III,IV) (Hydr)oxides by oxalate: The effect of pH and Light, Langmuir, 1992, 8, 95-103
7
Jun Y-S, Martin ST, Microscopic Observations of Reductive manganite Dissolution under Oxic Conditions,
Environ. Sci. Technol., 2003,37,2363-2370
8
Varentsov IM, Grassely GY, Geology and Geochemistry of Manganese, Volume 1, E. Schweizerbart’she
Verlagsbuchhandlung, Stuttgart, 1980
9
Pontér, C. Ingri, J. Boström, K. Geochemistry of manganese in the Kalix River, northern Sweden, Geochim.
Cosmochim. Acta, 1992, 56, 1485-1494
10
Crerar. D. A. Cormick, R. K. Barnes, H. L. Geochemistry of manganese: an overview, in Geology and
Geochemistry of Manganese, Volume 1, Ed Varentsov IM, Grassely GY, E. Schweizerbart’she
Verlagsbuchhandlung, Stuttgart, 1980, 293-334
11
Bricker O, Some stability relations in the system Mn-O2-H2O at 25° and one atmosphere total pressure, Am.
Mineral., 1965, 50, 1296-1354
12
Lindsay, W.L. Chemical Equilibria in Soils, John Wiley and Sons, New York, 1979
13
Biedermann, G. Palombari, R. On the hydrolysis of the Manganese(III) Ion, Acta Chem. Scand. 1978, A32,
381-390
14
Kostka, J. E. Luther, G. W. III: Nealson, K. H. Chemical and biological reduction of Mn(III)-pyrophosphate
complexes: Potential importance of dissolved Mn(III) as an environmental oxidant, Geochim. Cosmochim. Acta,
1995, 59, 885-894
15
Klewicki, J. K. Morgan, J. J. Kinetic behaviour of Mn(III) complexes of pyrophosphate, EDTA, and Citrate,
Environ. Sci. Technol. 1998, 32, 2916-2922
16
Hem JD, Lind CJ, Nonequilibrium models for predicting forms of precipitated manganese oxides, Geochim.
Cosmochim. Acta, 1983, 47, 2037-2046
17
Hastings D, Emerson S, Oxidation of manganese by spores of a marine bacillus: Kinetic and thermodynamic
considerations, Geochim. Comsmochim. Acta, 1986, 50, 1819-1824
18
Tipping E, Thompson DW, Davison W, Oxidation products of Mn(II) in lake waters, Chem. Geol. 1984, 44,
359-383
19
Murray JW, Dillard JG, Giovanoli R, Moers H, Stumm W, Oxidation of Mn(II): Initial mineralogy ,
oxidation state and ageing, 1985, Geochim. Cosmochim. Acta., 49, 463-470
20
Giovanoli R, On Natural and Synthetic Manganese Nodules, in Geology and Geochemistry of Manganese,
Volume 1, Ed Varentsov IM, Grassely GY, E. Schweizerbart’she Verlagsbuchhandlung, Stuttgart, 1980, 159202
21
Kessick MA, Morgan JJ, Mechanisms of Autooxidation of manganese in Aqueous Soultion, Environ. Sci.
Technol. 1975, 9(2) 157-159
22
Lind CJ, Hausmannite (Mn3O4) Conversion to Manganite (γ-MnOOH) in Dilute Oxalate Solution, Environ.
Sci. Technol. 1988, 22, 62-70
23
Mandernack KW, Post J, Bradley MT, Manganese mineral formation by bacterial spores of the marine
Bacillus, strain SG-1: Evidence for the direct oxidation of Mn(II) to Mn(IV), Geochimica. Cosmochim. Acta.,
1995, 59(21) 4393-4408
24
Post JE, Manganese oxide minerals: Crystal structures and economic and environmental significance, 1999,
Proc. Natl. Acad. Sci. USA, Vol 96, 3447-3453
25
De Vitre RR, Davison W, Manganese particles in freshwaters, Chapter 7 in Environmental Particles Vol 2, Ed
Buffle J, van Leeuwen HP, 1993, Lewis Publishers, Boca Raton
2
69
26
Dong D, hua X, Li Y, Zhang J, Yan D, Cd adsorption properties of components in different freshwater surface
coatings: The important Role of Ferromanganese Oxides, Environ. Sci. Technol., 2003, 37, 4106-4112
27
Taillefert M, MacGregor BJ, Gaillard JF, Lienemann CP, Perret D, Stahl DA, Evidence for a dynamic cycle
between Mn and Co in the water column of a stratified lake, Environ. Sci. Technol., 2002, 36, 468-476
28
Pettersson, U. T. Ingri, J. The geochemistry of Co and Cu in the Kafue River as it drains the Copperbelt
mining area, Zambia, Chem. Geol. 2001, 177, 399-414
29
Lehninger AL, Nelson DL, Cox MM, Principles of Biochemistry, 2nd Edition, Worth Publishers, New York,
1993
30
Diem D, Stumm W, Is dissolved Mn2+ being oxidised by O2 in the absence of Mn-bacteria or surface
catalysts?, Geochim. Cosmochimica. Acta., 1984, 48, 1571-1573
31
Pontér, C. Ingri, J. Burman, J-O. Boström, K. Temporal variations in dissolved and suspended iron and
manganese in the Kalix river, northern Sweden, Chem. Geol. 1990, 81, 121-131
32
Marklund, K. Nationalencyclopedin, Band 13, Bokförlaget Bra Böcker, Höganäs, Sweden, 1994, p40-42,
33
Dachs von H, Neutronen- und Röntgenundersuchungen am Manganit, MnOOH, Zeitschrift für
Kristallographie, 1963, 118, 303-326
34
Buerger MJ, The symmetry and crystal structure of manganite, Mn(OH)O., Zeitschrift für Kristallographie,
1936, 95, 163-174
35
Chabre Y, Pannetier J, Structural and electrochemical properties of the proton/γ-MnO2 System, Prog. Solid
St. Chem., 1995, 23, 1-130
36
Maslen, B. E. N. Streltsov, V. A. Streltsova, N. R. Electron Density and Optical Anisotropy in Rhombohedral
Carbonates. III: Synchrotron X-ray Studies of CaCO3, MgCO3 and MnCO3, Acta Cryst, 1995, B51, 929-939
37
Stumm W, Giovanoli R, On the Nature of Particulate Manganese in Simulated Lake Waters, Chimica, 1976,
30(9) 423-425
38
Kohler T, Armbruster T, Libowitzky, E. Hydrogen Bonding and Jahn-Teller Distortion in groutite, αMnOOH, and manganite, γ-MnOOH, and Their Relation to the manganese Dioxides Ramsdellite and Pyrolusite,
J. Solid State Chem. 1997, 133, 486-500
39
Ramstedt M, Shchukarev AV, Sjöberg S, Characterization of hydrous manganite (γ-MnOOH) surfaces – an
XPS study, Surf. Interface Anal., 2002, 34, 632-636
40
Bochatay L, Persson P, Sjöberg S, Metal Ion Coordination at the Water-Manganite (γ-MnOOH) Interface. I.
An EXAFS study of cadmium(II), J. Colloid Interface Sci., 2000, 229, 584-592
41
Klewicki, J. K. Morgan, J. J. Dissolution of β-MnOOH particles by ligands: Pyrophosphate,
etylenediaminetetraacetate, and citrate, Geochim. Cosmochim. Acta, 1999, 63, 3017-3024
42
Hiemstra, T. Riemsdijk, van W. H. A Surface Structural Approach to Ion Adsorption: The Charge
Distribution (CD) Model, J. Colloid. Interface Sci. 1996, 179, 488-508
43
Pauling L, The principles determining the structure of complex ionic crystals, J. Am. Chem. Soc., 1929, 51,
1010-1026
44
Rakovan, John. Becker, Udo. Hochella, Michael F Jr. 1999, Aspects of goethite surface microtopography,
structure, chemistry, and reactivity, Am. Mineral. 84, 884-894
45
Berry, L. G. Mason, B. Mineralogy – Concepts, descriptions, determinations, 2nd Edition, Ed Dietrich, R.V.
W. H. Freeman and Company, San Fransisco, 1983, 318-319
46
Giovanoli, R. Leuenberger, U. Über die Oxydation von Manganoxidhydroxid, Helv. Chim. Acta, 1969, 52,
2333-2347
47
Brunauer, S., Emmet, P.H., Teller, E. Adsorption of Gases in Multimolecular Layers , J. Am. Chem. Soc.,
1938, 60, 309-319
48
Skog, D. A. Leary, J. J. Principles of Instrumental Analysis, 4th Ed, Hardcourt Brace College Publishers, Fort
Worth, USA, 1992
49
Ratner, B. D. Castner, D. G. Electron Spectroscopy for Chemical Analysis, Chapter 3 in Surface Analyses –
The Principal Techniques, Ed. Vickerman, Wiley & Sons, 1997
50
NIST Standard Reference Database 82, NIST Electron Effective-Attenuation-Length Database Version 1.0,
2001, U.S. Department of Commerce, Gaithersburg, Maryland, USA
51
Hochella, M. F. Jr. Auger Electron and X-ray Photoelectron Spectroscopies, Chapter 13 in Spectroscopic
Methods in Mineralogy and Geology, Volume 18 in Reviews in Mineralogy, , Ed Hawthorne, F.
C.,Mineralogical Society of America, Washington D.C. 1988, 573-637
52
Ilton, E. S. Moses, C. O. Veblen, D. R. Using X-ray photoelectron spectroscopy to discriminate among
different sorption sites of mica: with implications for heterogeneous reduction of chromate at the mica-water
interface, Geochim. Cosmochim. Acta, 2000, 64, 1437-1450
53
Paterson, E. Swaffield, R. X-ray photoelectron Spectroscopy, Chapter 6 in Clay mineralogy: Spectroscopic
and Chemical Determinative Methods, Ed. Wilson, M. J. Chapman & Hall, London, 1994, 226-259
70
54
Mitchell, D.F. Clark, K.B. Bardwell, J.A. Lennard, W.N. Massoumi, G.R. Mitchell, I.V., Film Thickness
Measurements of SiO2 by XPS, Surf. Interface Anal. 1994, 21, 44-50
55
Brown, G. E. Calas, G. Jr. Waychunas, G. A. Petiau, J. X-ray absorption spectroscopy and its application in
mineralogy and geochemistry, Chapter 11 in Spectroscopic Methods in Mineralogy and Geology, Volume 18 in
Reviews in mineralogy, , Ed Hawthorne, F. C.,Mineralogical Society of America, Washington D.C. 1988, 430512
56
Ankudinov, A.L., Ravel, B., Rehr, J.J., and Conradson, S.D. Real-space multiple-scattering calculation and
interpretation of x-ray-absorption near-edge structure, Phys. Rev. B 1998, 58, 7565-7576.
57
Jalilehvand, F. Structure of Hydrated Ions and Cyanide Complexes by X-ray Absorption Spectroscopy, PhD
Thesis, Inorganic Chemistry, Royal Institute of Technology, Stockholm, 2000, ISBN 91-7170-561-9
58
Giacovazzo, C. Fundamentals of Crystallography, Oxford Press, New York, 1992
59
Hammond, C. The basics of Crystallography and Diffraction, Oxford University Press, New York, 2001
60
Bish, D. L. Post, J. E. Modern Powder Diffraction, Volume 20 in Reviews of Mineralogy, Mineralogical
Society of America, Washington D.C. 1989
61
Goldstein, J. I. Lyman, C. E. Newbury, D.E. Lifshin, E. Echlin, P, Sawyer, L. Joy, D. C. Michael, J. R.
Scanning Electron Microscopy and X-ray Microanalysis, Kluwer Academic/Plenum Publishers, New York,
2003
62
Buseck, P. R. Minerals and Reactions at the atomic scale: transmission electron microscopy, Volume 27 in
Reviews in Mineralogy, Mineralogical Society of America, Washington D.C. 1992
63
Russell, P. Batchelor, D. Thornton, J. SEM and AFM: Complementary Techniques for High Resolution
Surface Investigations, Digital Instruments, Carlifornia, USA.
64
Öhman, L-O. Experimental determination of stability constants of aqueous complexes, Chem. Geol., 1998,
151, 41-50
65
Ginstrup, O. Experimental and Computational Methods for Studying Multicomponent Equilibria. II: An
Automated System for Precision emf Titrations, Chem. Instr. 1973, 4, 141-155
66
Sjöberg, S. Lövgren, L. The application of potentiometric techniques to study complexation reactions at the
mineral/water interface, Aq. Sci., 1993, 55, 324-335
67
Forsling, W. Hietanen, S. Sillén, L. G. Studies of the hydrolysis of metal ions. III. The hydrolysis of the
mercury(I) ion, Hg22+, Acta Chem Scand. 1952, 6, 901-909
68
O’Brien, R.W. Cannon, D.W. Rowlands, W.N. Electroacoustic Determination of Particle Size and Zeta
Potential, J. Colloid Interface Sci., 1995, 173, 406-418
69
Schindler, P. W. Gamsjäger, H. Acid-base reactions of the TiO2 (Anatase) – water interface and the point of
zero charge of TiO2 suspensions, Kolloid-Z. U. Z. Polymere, 1972, 250, 759-765
70
Van Riemsdijk, W. H., Bolt, G. H., Koopal, L. K. and Blaakmeer, J Electrolyte Adsorption on Heterogeneous
Surfaces – Adsorption Models J. Coll. Interf. Sci 1986, 109, 219-228
71
Bolt, G. H. Soil Chemistry, B. Physico-Chemical Models, Elsevier, Amsterdam, 1982
72
Lützenkirchen, J. Boily, J. F. Sjöberg, S. Current applications of and future requirements for surface
complexation models, Current topics in Colloid and Interface Science, 2002, 5, 157-190
73
Ingri, N., Andersson, I., Pettersson, L., Yagasaki, A., Andersson L., and Holmström K. LAKE- A Program
System for Equiliibrium Analytical Treatment of Multimethod Data, Especially Combined Potentiometry and
Nuclear Magnetic Resonance Data, Acta Chem. Scand. 1996, 50, 717-734.
74
Westall, J. C. FITEQL: A Computer Program for Determination of Chemical Equilibrium Constants from
Experimental Data, Version 2.0, Report 82-02, Department of Chemistry, Oregon State University, Corvallis,
OR, USA, 1982
75
http://www.chem.umu.se/dep/inorgchem/samarbeta/WinSGW_eng.stm
76
Eriksson, G. An Algorithm for the Computation of Aqueous Multicomponent, Multiphase Equilibria, Anal.
Chim. Acta 1979, 112, 375-383.
77
Weaver, R. M. Hochella, M. F. Ilton, E. S. Dynamic processes occurring at the CrIIIaq-manganite (γ-MnOOH)
interface: Simultaneous adsorption, microprecipitation, oxidation/reduction, and dissolution Geochim.
Cosmochim. Acta. 2002, 66(23), 4119-4132
78
Shaughnessy DA, Nitsche H, Booth CH, Shuh DK, Waychunas GA, Wilson RE, Gill H, Cantrell KJ, Serne
RJ, Molecular Interfacial Reactions between Pu(VI) and Manganese Oxide Minerals Manganite and
hausmannite, Environ. Sci. Technol., 2003, 37, 3367-3374
79
Matocha CJ, Elzinga EJ, Sparks DL, Reactivity of Pb(II) at the Mn(III,IV) (Oxyhydr)Oxide-Water Interface,
Environ. Sci. Technol., 2001, 35, 2967-2972
80
Yamada, N. Ohmasa, M. 1986, Textures in natural pyrolusite, β-MnO2, examined by 1 MV HRTEM, Acta
Cryst. B42, 58-61
81
Zhang H, Huang C, Oxidative Transformation of Triclosan and Chlorophene by Manganese Oxides, Environ.
Sci. Technol., 2003, 37, 2421-2430
71
82
Mcardell CS, Stone AT, Tian J, Reaction of EDTA and realted aminocarboxylate chelating agents with
CoIIIOOH (Heterogenite) and MnIIIOOH (Manganite), Environ. Sci. Technol. 1998, 32, 2923-2930
83
Chiu van Q, Hering JG, Arsenic Adsorption and Oxidation at Manganite Surfaces 1. Method for Simultaneous
Determination of Adsorbed and Dissolved Arsenic Species, Environ. Sci. Technol., 2000, 34, 2029-2034
84
Gimsing, A. L. Borggaard, O. K. Sestoft, P. Modeling the Kinetics of the Competitive Adsorption and
Desorption of Glyphosate and Phosphate on Goethite and Gibbsite and in soils, Environ. Sci. Technol. 2004, 38,
1718-1722
85
Carlisle SM, Trevors JT, Glyphosate in the Environment, Water, Air, and Soil Pollution, 1988, 39, 409-420
86
Sprankle P, Meggitt WF, Penner D, Adsorption, Mobility, and Microbial Degradation of Glyphosate in the
Soil, Weed Sci., 1975, 23(3) 229-234
87
Sprankle P, Meggit WF, Penner D, Rapid Inactivation of Glyphosate in the Soil, Weed Sci., 1975, 23(3) 224228
88
Sheldrick von WS, Morr M, N-phosphonomethylglycine, Acta Cryst. 1981, B37, 733-734
89
Shkol’nikova LM, Porai-Koshits MA, Dyatlova NM, Yaroshenko GF, Rudomino MV, Kolova EK, X-ray
structural study of organic ligands of the complexone type. III. Crystal and Molecular structure of
phosphonomethylglycine and iminodiacetic-monomethylphosphonic acid, Zhurnal Strukturnoi Khimii, 1982,
23(5), 98-107, (English translation)
90
Sheals J, Granström M, Sjöberg S, Persson P, Coadsorption of Cu(II) and glyphosate at the water-goethite (αFeOOH) interface: molecular structures from FTIR and EXAFS measurements, J. Colloid Interface Sci. 2003,
262, 38-47
91
Sheals J, Persson P, Hedman B, IR and EXAFS Spectroscopic Studies of Glyphosate protonation and
Copper(II) Complexes of Glyphosate in Aqueous Solution, Inorg. Chem. 2001, 40 (17) 4302-4309
92
Torstensson L, Behaviour of glyphosate in soils and its degradation, Chapter 9 in The Herbicide Glyphosate
Ed, Grossbard E & Atkinson D, 1985, Butterworths, London
93
Rueppel ML, Brightwell BB, Schaefer J, Marvel JT, Metabolism and Degradation of Glyphosate in Soil and
Water, J. Agric. Food Chem., 1977, 25, 517-528 (due to ads to soil)
94
Sheals J, Sjöberg S, Persson P, Adsorption of glyphosate in goethite: Molecular characterization of surface
complexes, Environ. Sci. Technol. 2002, 36, 3090-3095
95
McBride M, Kung K, Complexation of glyphosate and related lligands with iron, Soil Sci. Soc. Am. J, 1989,
53, 1668-1673
96
Gimsing, A. L. Borggaard, O. K. Bang, M. Influence of soil composition on adsorption of glyphosate and
phophate by contrasting Danish surface soils, Eu. J. Soil Sci., 2004, 55, 183-191
97
Jonge, H de. Jonge, L.W. de, Influence of pH and solution composition on the sorption of glyphosate and
prochloraz to a sandy loam soil, Chemosphere, 1999, 39(5), 753-763
98
Motekaitis RJ, Martell AE, Metal chelate formation by N-phosphonomethylglycine and related ligands, J.
Coord. Chem.,1985, 14, 139-149
99
Deacon, G.B. Philips, R.J. Relationships between the carbon-oxygen stretching frequencies of carboxylato
complexes and the type of carboxylate coordination, Coord. Chem. Rev. 1980, 33, 227-250
100
Heineke D, Franklin SJ, Raymond KN, Coordination Chemistry of Glyphosate: Structural and Spectroscopic
Characterization of Bis(glyphosate)metal(III) Complexes, Inorg. Chem. 1994, 33, 2413-2421
101
Ramstedt M, Norgren C, Sheals J, Sjöberg S, Persson P, Thermodynamic and Spectroscopic Studies of
Cadmium(II)-N-(phosphonomethyl)glycine (PMG) Complexes, Inorg. Chim. Acta, 2004, 357, 1185-1192
102
Clarke ET, Rudolf PR, Martell AE, Clearfield A, Structural Investigation of the Cu(II) Chelate of Nphosphonomethylglycine. X-ray Crystal structure of Cu(II)[O2CCH2NHCH2PO3]·Na(H2O)3.5, Inorg. Chim. Acta,
1989, 164, 59-63
103
Appleton TG, Hall JR, McMahon IJ, NMR Spectra of Iminobis(methylenephosphonic acid),
HN(CH2PO3H2)2, and related Ligands and their complexes with Platinum(II), Inorg. Chem. 1986, 25, 726-734
104
Rajendram CRA, Hoggard PE, The coordination of aminophosphonates to chromium(III), J. Coord. Chem.
1994, 33, 15-32
105
Ptaszy´nski B, Zwoli´nska A, Synthesis and Properties of Solid Complexes of Lanthanum, Cerium,
Neodynium and Erbium with N-Phosphonomethylglycine, Polish J. Environ. Studies, 2001, 10(4), 257-262
106
Appleton TG, Byriel KA, Hall JR, Kennard CHL, Lynch DE, Sinkinson JA, Smith G, Reactions of Di- and
Trimethylplatinum(IV) Complexes with N-(phosphonomethyl)glycine (glyphosate) and
Iminobis(methylenephosphonic acid). Crystal structures of three dimethyl(IV) complexes with Nphosphonomethyl)glycine coordinated facially, Inorg. Chem. 1994, 33, 444-455
107
Sanna D, Bódi I, Bouhsina S, Micera G, Kiss T, Oxovanadium(IV) complexes of phosphonic derivatives of
iminodiacetic and nitrilotriacetic acids, J. Chem. Soc. Dalton Trans, 1999, 3275-3282
108
Szabó Z, Structure, equilibrium and ligand exchange dynamics in the binary and ternary dioxouranium(VI)glyphosate-fluoride system. A multinuclear NMR study, J. Chem. Soc. Dalton Trans, 2002, 4242-4247
72
109
Barja BC, Herszage J, dos Santos Afonso M, Iron(III)-phosphonate complexes, 2001, Polyhedron, 20, 18211830
110
Sagatys DS, Dahlgren C, Smith G, Bott RC, Willis AC, Metal Complexes with N-(Phosphonomethyl)glycine
(Glyphosate): the Preparation and Characterization of the Group 2 Metal Complexes with Glyphosate and the
Crystal Structure of Barium Glyphosate Dihydrate, Aust. J. Chem., 2000, 53, 77-81
111
Sagatys DS, Dahlgren C, Smith G, Bott RC, White JM, The complex chemistry of N(phosphonomethyl)glycine (glyphosate):preparation and characterization of the ammonium, lithium, sodium (4
polymorphs) and silver(I) complexes, J. Soc. Dalton Trans, 2000, 3404-3410
112
Rudolf PR, Clarke ET, Martell AE, Clearfield A, Preparation and X-ray Structure of the Calcium(II)
Complex of N-Phosphonomethylglycine Ca[O2CCH2NH2CH2PO3]·2H2O at pH 7.2, Acta Cryst. 1988, C44, 796799
113
Smith PH, Raymond KN, Solid-State and solution Chemistry of Calcium N-(Phosphonomethyl)glycine,
Inorg. Chem. 1988, 27, 1056-1061
114
Galdecka E, Galdecki Z, Gawryszewska P, Legendziewicz J, Structure of a novel polynuclear europium
compound with n-phosphonomethylglycine: heptaaquaperchlorate monohydrate, [Eu2(HO3PCH2NH2CH2CO2)2(H2O)7(ClO4)]·3ClO4·H2O, New J. Chem., 2000, 24, 387-391
115
Bochatay L, Persson P, Sjöberg S, Metal Ion Coordination at the Water-Manganite (γ-MnOOH) Interface. I.
An EXAFS study of cadmium(II), J. Colloid Interface Sci., 2000, 229, 584-592
116
Bochatay L, Persson P, Metal Ion Coordination at the Water-Manganite (γ-MnOOH) Interface. II. An
EXAFS study of zinc(II), ), J. Colloid Interface Sci., 2000, 229, 593-599
117
Pan, G. Qin, Y. Li, X. Hu, T. Wu, Z. Xie, Y., EXAFS studies on adsorption-desorption reversibility at
manganese oxides-water interfaces I. Irreversible adsorption of zinc onto manganite (γ-MnOOH), J. Colloid.
Interface Sci. 2004, 271, 28-34
118
Sheals J, Molecular Characterisation of glyphosate complexes in aqueous solution and at the solutionmineral interface, PhD thesis, Umeå University, Sweden, 2002, ISBN 91-7305-343-0
73
Acknowledgements
A lot of absolutely wonderful people have surrounded me during these past years, and some of you I
would like to give a special “Thank you!” here in the end of my thesis.
To start with, I would like to thank my supervisor Staffan Sjöberg. You have many times showed me
that there actually does exist a way through the “swamps” when I thought I was about to give up. I am
very grateful for your support and guidance; you’ve been a great supervisor! I also would like to thank
Per Persson, my assistant supervisor, for introducing me to many fascinating spectroscopic techniques,
for helpful discussions and for being such a good colleague. I’ll miss your skånish “Tjena Madde!”
Lars Lövgren, my other assistant supervisor, I am very grateful that you introduced me into this field
of science and let me become MiMi-associated in some parts of my research. I have learned many
interesting things from you; from how to find the best restaurant to the composition of mining waste.
My unofficial assisting supervisor, Andrei Shchukarev, who have taught me almost everything I know
about XPS. I am very grateful to you for this knowledge and that you’ve shared your ideas and
chocolate with me. It’s been a lot of fun! Thank you!
Caroline and Kristina, my office mates; we have shared many things! Johan, Leif, Frasse, Laurence,
Jörgen, Maria, Agneta, Ingegärd – thank you so much for helping and coaching me during my first
years. Dan; thanks for answering all my questions about structural chemistry. Kristina, Julia, John and
Åsa for proof reading my texts. Unfortunately, I won’t be able to thank all of my colleagues and
friends at the department of inorganic chemistry individually on this page. All the same, I am very
grateful to all of you for making my time at the department as great as it has been. I’ve had so much
fun and I’ve learnt so many things!
I want to thank Bruce Johnson at La Trobe University for helping me arrange my time in Australia and
giving me the opportunity to work in the lab in Bendigo, but also for making me feel at home in
Australia from the very first beginning. Thanks to the rest of the colloid surface group in Bendigo for
your kindness and teaching me about Australian football. Paul Pigram, Narelle Brack and the students
in their group for really making me feel welcome in Melbourne and allowing me to use their nice
XPS. Ewen Silvester, thanks for all your help at CSIRO Clayton during and after my visit, and for
showing me a bit of your Australia. Patrick Hartley and co-workers, for allowing me to work in their
group and learn about AFM, especially I would like to thank Celesta Fong for all her help.
Gudrun Boström, min mentor och bollplank; jag har verkligen uppskattat våra samtal, de har gett mig
mycket glädje och varit som vitamininjektioner, Tusen tack!!
Min underbara familj, i flera generationer, som alltid stöttar och tror på mig. Särskilt vill jag förstås
tacka mamma, pappa och Natanael. Tack för att ni finns där!!! Och sist men inte minst är jag så
tacksam till alla mina vänner utanför jobbet. Ni ger mig perspektiv på tillvaron, tillför så mycket
glädje och betyder oerhört mycket för mig.
Tack allihopa!!
74