Cinetica de Reaccion Termo 3

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Silica Polymerization from Supersaturated Dilute
Aqueous Solutions in the Presence of Alkaline Earth Salts
Marina Kley, Andreas Kempter, Volodymyr Boyko, and Klaus Huber
Langmuir, Just Accepted Manuscript • Publication Date (Web): 23 May 2017
Downloaded from http://pubs.acs.org on May 25, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street

N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.
Page 1 of 40 Langmuir
1
2
3
4
5
6
7
8
9
10
11
Silica Polymerization from Supersaturated Dilute
12
13
14
15
Aqueous Solutions in the Presence of Alkaline Earth
16
17
18
19
Salts
20
21
22
23
24
M. Kley†, A. Kempter§, V. Boyko‡, K. Huber*,†
25
26 †
27 Physical Chemistry, University of Paderborn, Warburger Str. 100, 33098 Paderborn, Germany
28
29
§
30 BASF SE, Strategic Marketing and Product Development Automotive Fluids, Fuel and
31
32 Lubricant Solutions, 67056 Ludwigshafen, Germany
33
34
35 ‡
36 BASF SE, Material Physics and Analytics, Colloidal Systems, 67056 Ludwigshafen, Germany
37
38
39 ABSTRACT
40
41 The early stages of silica polymerization in aqueous solution proceeds according to a
42
43
44 mechanism based on three steps: nucleation, particle growth and agglomeration of the particles.
45
46 Application of time-resolved static and dynamic light scattering as a powerful in-situ technique
47
48 in combination with spectrophotometric analysis of the monomer consumption based on the
49
50
molybdenum blue method was carried out to further investigate this 3-step process. Experiments
51
52
53 were carried out at four different initial silicic acid contents covering a range between 350 ppm
54
55 and 750 ppm in the presence of either 10 mM NaCl or 5 mM of a mixture of CaCl2 and MgCl2.
56
57 The process in all cases was initiated with a drop of pH to 7. Addition of the salts made possible
58
59
60
1
ACS Paragon Plus Environment
Langmuir Page 2 of 40
1
2
3
4
5
6 an analysis of the impact of an electrolyte on the process. Independent of the presence or absence
7
8
9 of salt, particle growth in step two proceeded as a monomer-addition process without being
10
11 interfered significantly by Ostwald-ripening. The growing particles were compact with a
12
13 homogeneous density. The size of the particles approached final values between 5 nm and 20 nm
14
15
16
with the actual value increasing with decreasing initial silicic acid content. Above a certain
17
18 concentration of initial silica content, which depends on the level of added salt, particle-particle
19
20 interactions caused agglomeration. The presence of electrolyte shifted this level from
21
22
~ 2000 ppm to a range between 500 ppm and 750 ppm. The resulting agglomerates had a fractal
23
24
25 dimension of 2. Independent of the conditions, particle growth could be described with a simple
26
27 nucleation and growth model.
28
29
30
31
32 INTRODUCTION
33
34 Silica polymerization in aqueous media is an important feature in many natural and geological
35
36 processes including biomineralization, sinter formation and silica diagenesis. It becomes also
37
38
39
increasingly important in material science. Inspired by biological systems additives are to be
40
41 developed in order to direct silica formation into highly patterned and hierarchical structures.1,2
42
43 Last but not least, silica polymerization denoted as scaling or fouling is a challenging problem in
44
45
industry, since it causes decreased water recoveries in desalination plants or a lower economic
46
47
48 efficiency in geothermal power plants.3,4 A detailed knowledge of the mechanism and kinetics of
49
50 the silica polymerization at variable parameters like the pH, salinity or temperature of the
51
52 aqueous medium may indicate routes to particles with specific properties or to an inhibition of
53
54
55 silica precipitation. Accordingly, large efforts have been undertaken in the past to meet this
56
57 demand. An excellent overview on the progress accomplished by the late seventies was provided
58
59
60
2
1
2
3
4
5
6 by Iler5 in 1979. Based on a few pioneering publications6–11 Iler5 established a mechanism for the
7
8
9 polymerization of silica in supersaturated aqueous solution, which proceeds via three distinct
10
11 steps. The first step is an induction period where nucleation takes place. In the second step
12
13 particles grow. Monomers, oligomers and polymers are the species dominating throughout these
14
15
16
two steps. The two steps are characterized by a sigmoidal increase of monomer consumption
17
18 recorded by the molybdate-reactive silica. Using a particle size dependent solubility of silica,
19
20 Iler5 reconciled the data9,11 with the initial formation of small particles in the range of one nm
21
22
probably corresponding to nuclei, which successively grow in size up to a radius of a few
23
24
25 nanometers. Without providing direct evidence, data on the growth step could be best matched
26
27 with a monomer addition process. Monomer consumption is fastest in the range of 7 < pH < 9
28
29 with an optimum at pH = 8.3.7 At equilibrium monomeric silica approached concentrations in the
30
31
32 regime of 120-180 ppm.7 Most likely, part of the oligomers like polycyclic octamers and
33
34 decamers, which have also been observed in processes based on hydrolysis of silicic acid esters
35
36 by ESI-MS12–14 turn into nuclei and the nuclei grow by adding further monomers and small
37
38
39
oligomers, with those small oligomers acting as “monomers”. An unsettled question in this
40
41 context is whether Ostwald-ripening sets in at the later stages of particle growth. Once the initial
42
43 concentration of monomeric silica exceeds a certain level, the particles start to agglomerate8,
44
45
thus establishing the third step in Iler’s mechanistic scheme. At high enough silica concentrations
46
47
48 this agglomeration appears as a macroscopic gelation.6 As demonstrated by Merrill and
49
50 Spencer10, the exact level of the silica content where this agglomeration sets in depends on the
51
52 salinity and pH of the solution.
53
54
55 Despite the considerable number of publications meanwhile available, knowledge on the
56
57 mechanism of silica formation in aqueous media still suffers from subtle gaps or is awaiting clear
58
59
60
3
1
2
3
4
5
6 cut confirmation. Since the present work tries to solve unsettled issues of silica formation in
7
8
9 dilute aqueous solution close to neutral pH-values, some of the most important achievements in
10
11 this field shall be briefly summarized.
12
13 Aside from supporting the earlier findings mentioned above, two further studies focused on the
14
15
16
impact of added salts15,16, and revealed an acceleration of particle growth with increasing salt
17
18 concentration. Icopini et al.15 also identified a reaction rate for the monomer consumption of
19
20 fourth order in terms of the monomer concentration and took this as a hint for tetramers to act as
21
22
nuclei. Further insight into the origin of the induction period was provided by Noguera et al.17,
23
24
25 who published a computational approach based on the classical nucleation theory. They
26
27 successfully applied their model to experimental data from Rothbaum and Rhode18 and Tobler et
28
29 al.19,20. Silica polymerization was initiated via pre-established initial supersaturation, via
30
31
32 neutralization of a high pH solution of silica and via fast cooling. It turned out that the properties
33
34 of the particle population strongly depend on the experimental conditions. Further on they
35
36 postulated that the induction period has to be most likely attributed to a stage in which
37
38
39
nucleation, growth and dissolution compete with and compensate each other.
40
41 In two other recent contributions use could be made of modern scattering techniques in order
42
43 to answer some of the unsettled questions. Tobler et al.19 performed time-resolved small angle x-
44
45
ray scattering (SAXS) and dynamic light scattering at one angle together with the molybdate
46
47
48 yellow based analysis of the depletion of monomeric silica at two different silica concentrations.
49
50 From an extrapolation of the kinetic model fits to time zero, Tobler et al.19 inferred that
51
52 homogeneous nucleation leads to particles of 1-2 nm. Successively, particles grow to a size not
53
54
55 larger than 4 nm. It is the kinetic modelling of this growth (the second step) which establishes a
56
57 key feature of their work19 revealing a particle growth by surface controlled monomer
58
59
60
4
1
2
3
4
5
6 incorporation. However, no clear picture could be provided of the morphology of the particles
7
8
9 growing in the second step. Do those particles have a spherical shape or do they have a mass
10
11 fractal nature? The third stage was postulated to include Ostwald-ripening and growth via
12
13 particle-particle aggregation discernible by a steep increase of particle size values to radii close
14
15
16
to 30 nm. Kley et al.21 applied a combination of time-resolved static (SLS) and dynamic light
17
18 scattering (DLS) with the molybdate blue based monitoring of monomer consumption to analyze
19
20 the first stages of the particle formation mechanism covering a concentration regime of 300 –
21
22
3000 ppm of monomeric silica. Combination of these three methods provided excellent direct in-
23
24
25 situ data by yielding a time-resolved set of particle mass values in combination with the
26
27 evolution of the hydrodynamic radius of particles and once particles became larger than 20 nm
28
29 also the mean squared radius of gyration. In the regime of initial silica content of 350 ppm –
30
31
32 750 ppm, loss of monomers occurred simultaneously to the increase of size and mass of the
33
34 growing particles. Together with the power laws from a correlation of particle size with particle
35
36 mass, the results unambiguously proved a monomer-addition mechanism for the second step
37
38
39
leading to homogeneous particles most likely with a spherical shape. No Ostwald-ripening could
40
41 be observed in this second stage. A third step became only transparent at concentrations above
42
43 2000 ppm, with its threshold value depending on pH. In this third step, the compact particles
44
45
agglomerate according to a step-growth process where any particle may stick to any other
46
47
48 colliding particle resulting in entities with a fractal dimension close to 2.
49
50 Encouraged by those results, the present work extends this preceding study21 to the impact of
51
52 mixed (Mg1/5Ca4/5)Cl2 salt on the kinetics and mechanism of silica particle formation in the
53
54
55 regime of dilute silica solutions. This salt has been selected in order to represent a typical state
56
57 for hard water. In-situ SLS and DLS in combination with the molybdenum blue method is
58
59
60
5
1
2
3
4
5
6 applied in order to get direct access to particle formation and monomer consumption and to
7
8
9 answer the following questions. Does a change in mechanism occur upon addition of
10
11 (Mg1/5Ca4/5)Cl2? If a change does occur, is it ion specific or can it also be generated by the
12
13 addition of NaCl. Does the electrolyte change the tendency to form step-growth polymers in a
14
15
16
third step and does any evidence exist for Ostwald-ripening. Experiments are performed at four
17
18 different initial silica concentrations covering a regime of 350 – 750 ppm at a pH of 7 and at
19
20 [(Mg1/5Ca4/5)Cl2]= 5 mM. The evolution of particle mass shall be interpreted in terms of a simple
21
22
nucleation and growth model21 with its strength lying on a semi-quantitative interpretation of the
23
24
25 evolution of particle mass in the second step.
26
27
28
29 EXPERIMENTAL DETAILS
30
31
32
33 Materials. The preparation of solutions for light scattering experiments required three stock
34
35 solutions: a 0.2 M sodium silicate solution (Na2H2SiO4·8 H2O, assay ≥ 98%, Sigma Aldrich); a
36
37 solution with 0.2 M alkaline earth salt being composed of calcium chloride hexahydrate (CaCl2 ·
38
39
40
6 H2O, assay ≥ 98%, Sigma Aldrich) and magnesium chloride hexahydrate (MgCl2 · 6 H2O,
41
42 assay ≥ 99%, Sigma Aldrich) with a ratio of Ca : Mg = 4 : 1; a solution of sodium chloride
43
44 (NaCl, assay ≥ 99%, Sigma Aldrich) with a concentration of 0.4 M. All solutions were prepared
45
46
47
with water, which was purified by a Millipore system, leading to a conductivity of 0.055 Sm-1.
48
49
50
51 Molybdenum Blue Method. Observation of the decay of monomeric silica with time was
52
53
carried out by the molybdenum blue method22. The reaction of monomeric silica as Si(OH)4 with
54
55
56 acidified ammonium heptamolybdate to the blue silicomolybdic acid was used to determine the
57
58 concentration of monomeric silica via UV-vis-spectroscopy. For this purpose a small amount of
59
60
6
1
2
3
4
5
6 the silica solution which was prepared for a light scattering experiment was removed and diluted
7
8
9 with water to a volume of 50 mL. Addition of 1 mL of 50% (vol%) hydrochloric acid and 2 mL
10
11 of ammonium heptamolybdate solution with a concentration of 100 g/L while agitating led to the
12
13 formation of the yellow silicomolybdic acid after the solution was allowed to stand for seven
14
15
16
minutes. Addition of 2 mL oxalic acid with a concentration of 7.5 g/L and a second storage time
17
18 of two minutes eliminated a possible influence of phosphate on the reaction. In order to obtain
19
20 the blue silicomolybdic acid 2 mL of a reducing agent was added. The reducing agent was a
21
22
mixture of a solution of 23.08 g Na2S2O5 in 150 mL water with a solution of 500 mg 1-amino-2-
23
24
25 naphthol-4-sulfonic acid and 1 g Na2SO3 in 50 mL water. The absorbance of the blue sample was
26
27 measured at the latest 15 minutes after the addition of the reducing agent with a Lambda 19
28
29 spectrophotometer (Perkin Elmer) at a wavelength of 815 nm.
30
31
32
33
34 Preparation of Light Scattering Samples. The preparation of the light scattering samples
35
36 was performed in three steps. In a first step a certain amount of sodium silicate stock solution
37
38
39
depending on the desired concentration was diluted to a volume of 195 mL. Successively, the pH
40
41 value was adjusted to pH 7 by addition of 2 M HCl. The pH adjustment set the starting point of
42
43 the experiment. Right after the pH adjustment 5 mL of the respective stock solution of
44
45
Ca2++Mg2+ or Na+ were added. The sample was transferred into a dust free scattering cell
46
47
48 (20 mm in diameter) by filtration through a syringe filter (mixed cellulose ester) with a pore size
49
50 of 0.2 m. The scattering cells had been treated with a solution of 5% chlorotrimethylsilane
51
52
(98%, Janssen Chimica) in toluene, in order to avoid silica formation by the silicate solution at
53
54
55 the inner cell walls. All experiments were performed at 37 °C, the recording time of the
56
57 scattering intensity depended on the respective sample. Concentrations of silica denoted as c in
58
59
60
7
1
2
3
4
5
6 evaluation of data from light scattering experiments are given in units of g/l, where the
7
8
9 concentration of solid Na2H2SiO4·8 H2O is translated into a concentration of solid SiO2 with a
10
11 factor of 0.215. Alternatively, concentrations of silica are expressed as ppm of solid SiO2. All
12
13 solutions used for light scattering and for the molybdate blue method were 0.01 N with respect to
14
15
16
the added (Mg1/5Ca4/5)Cl2 or NaCl in addition to the NaCl inevitably formed during the initiation
17
18 of the process by neutralization. The contribution of NaCl from the neturalization amounted to
19
20 0.012 N at 350 ppm, 0.013 N at 400 ppm, 0.016 N at 500 ppm and 0.025 N at 750 ppm.
21
22
23
24
25 Scattering Setup. The ALV/CGS-3/MD-8 Multidetection Laser Light Scattering System from
26
27 ALV-Laservertriebsgesellschaft (Langen) makes possible a time-resolved recording of angular
28
29 dependent static light scattering (SLS) and dynamic light scattering (DLS) with a time resolution
30
31
32 limited by the time required for the evolution of a correlation function in DLS. It is equipped
33
34 with eight moveable detectors, which are positioned in angular increments of 8°. This set-up
35
36 covers an angle of 56°, which can be moved in an angular range of 15° ≤  ≤ 136° corresponding
37
38
39 to a q-range in water of 3.5·10-3 – 2.7·10-2 nm-1 with
40
41 4nsolv 
42 q sin  , (1)
43 0 2
44 the momentum transfer, nsolv = 1.332 (at T = 37°) the refractive index of water,  the scattering
45
46
47 angle and 0 the laser wavelength in vacuum. A He-Ne Laser with a wavelength of 632.8 nm and
48
49 a power of 35 mW is used as a light source.
50
51
52
53
54
55
Static Light Scattering. The static light scattering (SLS) signal provides the Rayleigh ratios at
56
57 a scattering angle , according to eq 2
58
59
60
8
1
2
3
4
5
6 r , sol  r , solv
7 R  R , std  , (2)
8 r , std
9
with R,std the tabulated Rayleigh ratio of the standard toluene, r,std the measured scattering
10
11
12 intensity of toluene, r,sol the scattering intensity of the solution and r,solv the scattering intensity
13
14 of the solvent.
15
16
17 For the determination of the weight averaged molar mass Mw and the z-averaged radius of
18
19 gyration Rg2 the Rayleigh ratios were processed according to Zimm’s approximation23
20
21 2
Kc 1 Rg
22   q 2  2 A2 c , (3)
23  R M w 3M w
24 with A2 the second virial coefficient and c the mass concentration in g/L of the solid in the
25
26
solution. As we were observing a reacting system an extrapolation of the scattering data to c = 0
27
28
29 was not possible. This forced us to neglect the influence of interactions among scattering
30
31 particles on the scattering signal by setting A2 = 0. Neglect was justified as the investigated
32
33 concentrations were very low (c ≤ 750 ppm). The contrast factor K introduced in eq 3 is given as
34
35 2
36 4 2 n 2  n 
K  4 tol   , (4)
37  0 N A  c 
38
39 with NA the Avogadro constant, ntol the refractive index of the bath liquid24 and ∂n/∂c the
40
41 refractive index increment of the solute in solution. The refractive index increment has been
42
43 determined for an aqueous solution of Na2H2SiO4 ·8 H2O to be ∂n/∂c = 0.223 mL/g.21 This value
44
45
is based on concentrations of Na2SiO3 as solid. If the monomer unit is defined as SiO2 and the
46
47
48 value for ∂n/∂c is calculated based on concentrations of SiO2, we get ∂n/∂c = 0.454 mL/g. This
49
50 value has been used in the present work for calculations of the weight-averaged molar mass. In
51
52 order to directly compare the resulting data with the weight-averaged molar mass values from the
53
54
55 previous publication21 these preceding values have to be multiplied with a factor of
56
57
58
59
60
9
1
2
3
4
5
6 0.241 = (0.223/0.454)2. This was done accordingly with the data21 shown in Figure 6, 7, 8A, S1
7
8
9 and S4a.
10
11 In order to directly compare the corresponding rate constants obtained from fits with the
12
13 kinetic model denoted as nucleation and growth model (NG-model) in Ref. 21 with the results
14
15
16
from model fits in the present study, the fits with the NG-model had to be re-done with the
17
18 corrected weight-averaged molar mass data of Ref. 21. The new rate constants established with
19
20 the NG-model reveal the same trends as discussed in Ref. 21 whereby kn is shifted by a factor of
21
22
1/0.241 or close to it, ke is shifted by a factor of 0.241 or close to it and kp remained unaffected.
23
24
25 An overview of all rate constants and the residuals of both kinetic models obtained from fits to
26
27 the light scattering data in the absence of added salt is given in Table S1 of the Supporting
28
29 Information.
30
31
32
33
34 Dynamic Light Scattering. Measurements of dynamic light scattering (DLS) gives the field-
35
36 time correlation function g1(). For polydisperse systems the field-time correlation function can
37
38
39 be described by a series of exponential functions, according to eq 5
40
41 n
42 g1 ( )    i  ei , (5)
43 i 1
44 with i the time constant of species i and i the weighting factor determined by the scattering
45
46
intensity of species i. According to the cumulant analysis25 the logarithm of the field-time
47
48
49 correlation function lng1 ( ) can be described by
50
51 1
52 ln[ g 1 ( )]  K 0  K 1  K 2 2  ... (6)
53 2!
54 where K0 is a constant, describing the signal to noise ratio and K1 is the z-average of the decay
55
56 time constant  z
and gives the z-averaged translational diffusion coefficient Dz according to
57
58
59
60
10
1
2
3
4
5
6
7
K1   z
 Dz q 2 . (7)
8
9
The variance of  is given by the coefficient K2      z

2
z
and provides information about
10
11
12 the polydispersity of the sample. A linear extrapolation of Dz(q) toward q2 → 0 and to c = 0
13
14 according to
15
16 Dz  D0 (1  q 2 Rg2C  k d c) (8)
17
18 gives the z-averaged diffusion coefficient D0. In eq 8 C is a dimensionless parameter depending
19
20 on the shape of the particles and kD accounts for the concentration dependence of Dz. As we
21
22
observe a reacting system, an extrapolation of D for c = 0 is not possible and in analogy to the
23
24
25 neglect of the concentration dependency in eq 3, a possible concentration dependency of Dz is
26
27 neglected. The z-averaged diffusion constant D0 can be converted into the hydrodynamic radius
28
29 Rh via the Stokes-Einstein equation
30
31
kT 1
32
Rh   (9)
33 6 D0
34
35 where k is the Boltzmann constant, T the temperature and the viscosity of the solvent.
36
37 The ratio  of the radius of gyration and the hydrodynamic radius is a structure-sensitive
38
39
40 parameter and is defined as
41
42 Rg
43 . (10)
44 Rh
45 For monodisperse linear polymer chains under -conditions denoted as unperturbed chains a
46
47 value of 1.504 is predicted by theory26,27, which turned out to be 17.5 % larger than values
48
49
50 revealed by experiment.28 The fractal dimensions of such unperturbed linear chains is 2. In case
51
52 of compact spheres  decreases drastically to 0.77.26 For rod-like structures a value of  ≥ 2.0 is
53
54
expected.29,30 Therefore a distinction between spheres, rods and fractal structures similar to
55
56
57 unperturbed polymer coils by means of the -parameter is possible.
58
59
60
11
1
2
3
4
5
6
7
8
9 RESULTS AND DISCUSSION
10
11
12 Effect of alkaline earth cations. The silica polymerization was investigated in the presence of
13
14 a mixture of divalent calcium and magnesium ions. A salt concentration of
15
16
17
[M2+] = [Ca2+]+[Mg2+] = 5 mmol/L at a ratio of Ca : Mg = 4 : 1 was selected in order to generate
18
19 conditions typical for hard water. Experiments of four samples with silica contents of 350 ppm,
20
21 400 ppm, 500 ppm and 750 ppm were performed. Investigation of the influence of sodium ions
22
23
was performed in order to enable comparison of the impact of the most simple and inert type of
24
25
26 salt with that of divalent alkaline earth cations. The Na+ concentration amounts to 10 mmol/L
27
28 introduced by the addition of NaCl. In case of divalent alkaline earth cations, Particle formation
29
30 was followed by means of time-resolved SLS and DLS experiments. The consumption of
31
32
33 monomeric silica is monitored by means of the molybdenum blue method.
34
35 Figure 1A represents experiments at pH = 7 in the presence of Ca2++Mg2+-ions at variable
36
37 silica contents. The growth rate as well as the finally reached molar mass strongly depends on
38
39
40
the initial silica content. The evolution of the molar mass determined by SLS of all four samples
41
42 correlates nicely with the consumption of the silica monomers determined by means of the
43
44 molybdenum blue method (Figure 1B). For all samples an equilibrium concentration of 170 ppm
45
46
SiO2 was reached at the end of the growth of silica particles. The three lowest concentrations
47
48
49 (350 ppm – 500 ppm) exhibit a distinct lag-time which decreases with increasing initial silica
50
51 content. During the lag-time the scattering contribution is comparable to the scattering signal of
52
53 the solvent and the monomer concentration keeps close to the initial silica concentration. After
54
55
56 this characteristic lag-time the decrease of the monomer concentration coincides with the
57
58 increase of the molar mass of the silica particles and both values reach an equilibrium value.
59
60
12
1
2
3
4
5
6 Only the highest silica content (750 ppm) shows a different behavior. In this case no lag-time
7
8
9 is observed. The consumption of the monomers up to the equilibrium limit of 170 ppm is
10
11 completed within t < 10 h, while the molar mass keeps increasing beyond this time without
12
13 reaching a plateau value. It is highly probable that particle formation at the silica content of 750
14
15
16
ppm during the first hours proceeds via monomer-addition in very much the same way as in case
17
18 of the lower silica contents. However, this process is completed during the first 10 hours and
19
20 smoothly followed by another growth step whereby the particles resulting from monomer-
21
22
addition establish the smallest units in this next growth step. As analysis of the joint SLS and
23
24
25 DLS data will show, fractal-like aggregates are formed in this third step.
26
27 25
Mw / 1E4 g mol-1
4
28 A
29 20
3
30
15
31
32 2
10
33
34 5 1
35
36 0 0
37 800
38
700
B
39
SiO2 / ppm
40 600
41 500
42 400
43 300
44
200
45
46 100
47 0 10 20 30 40 50 60 70
48
49 t/h
50
51 Figure 1: Formation of silica particles in aqueous solution as a function of time at pH 7 in the presence of
52 [Ca2+]+[Mg2+] = 5 mM (Ca2+/Mg2+ = 4/1). The silica contents are 750 ppm (black diamond), 500 ppm (red triangle),
53 400 ppm (blue circle) and 350 ppm (green square). (A) Weight-averaged molar mass from SLS. The experimental
54 mass values refer to two different scales with arrows pointing to the corresponding axis; (B) consumption of the
55 monomeric silica from the molybdenum blue method.
56
57
58
59
60
13
1
2
3
4
5
6 Similar trends for the evolution of the particle mass and the consumption of monomers is
7
8
9 observed for the silica polymerization in the presence of Na+-ions, which is shown by Figure 2.
10
11 A good correlation between the depletion of monomers and the increase of the particle mass is
12
13 observed for all four silica contents. In analogy to the measurements in the presence of
14
15
16
Ca2++Mg2+-ions the silica contents at 350 and 400 ppm exhibit a significant lag-time during
17
18 which no monomer is consumed and scattering intensity is comparable with the scattering
19
20 intensity of the solvent. After this lag-time which decreases with increasing silica content a fast
21
22
increase of the particle mass and simultaneously a decrease of the monomers is observed until
23
24
25 the monomers reach the equilibrium concentration of 170 ppm SiO2.
26
27 6 2.0
A
Mw / 1E4 g mol-1
28
5
29 1.6
30 4
31 1.2
3
32
33 0.8
2
34
0.4
35 1
36
0 0.0
37
38 700 B
39 600
SiO2 / ppm
40
500
41
42 400
43
300
44
45 200
46
100
47 0 20 40 60 80 150 300 450
48 t/h
49
50 Figure 2: Formation of silica particles in aqueous solution as a function of time at pH 7 in the presence of additional
51 Na+-ions at [Na+] = 10 mM. The silica contents are 750 ppm (black diamond), 500 ppm (red triangle), 400 ppm
52 (blue circle) and 350 ppm (green square). (A) Weight-averaged molar mass from SLS. The experimental mass
53 values refer to two different scales with arrows pointing to the corresponding axis; (B) consumption of the
54 monomeric silica from the molybdenum blue method.
55
56
57
58
59
60
14
1
2
3
4
5
6 Figure 3 compares the temporal evolution of the hydrodynamic radius of the samples in the
7
8
9 presence of Ca2++Mg2+-ions (Figure 3A) with the respective results collected for the samples in
10
11 the presence of Na+-ions (Figure 3B). As expected from the evolution of the particle mass in the
12
13 presence of Ca2++Mg2+, a distinct lag-time is followed by a short period of an increasing particle
14
15
16
size, which approaches a constant value for low silica contents (350 ppm-500 ppm). At 750 ppm
17
18 Rh increases almost linearly with t without approaching a plateau. Similar trends can be observed
19
20 in the presence of Na+-ions, only the absolute lag-times at the silica contents between 350 ppm
21
22
and 500 ppm are slightly longer than in the presence of Ca2++Mg2+. Final sizes of the particles
23
24
25 formed in the presence of Ca2++Mg2+-ions and in the presence of Na+-ions are comparable. The
26
27 evolution of Rh at 750 ppm SiO2 in the presence of Na+ shows the same linear increase with t as
28
29 in the presence of Ca2++Mg 2+.
30
31
32 In close analogy to the growth in water21 without added salt the final particle size decreases
33
34 with increasing initial silica content in the presence of metal cations (Figure 3). This indicates
35
36 that the nucleation rate increases considerably with increasing silica content. A faster nucleation
37
38
39
generates a larger number of nuclei, which have to share the respective amount of monomers
40
41 available. Hence, a lower number of nuclei, resulting from a slower nucleation, leads to larger
42
43 particles. As inferred from the lag-times (Figure 4), monovalent and divalent cations accelerate
44
45
46
the silica polymerization. The lag-time  obeys the following trend:  (salt
47
48 free) >  (Na+) >  (Ca2++Mg2+) whereby the differences between  (Na+) and  (Ca2++Mg2+)
49
50 remain close to the uncertainty of establishing ..
51
52
53 The trends discussed are nicely supported by a comparative discussion of the evolution of Mw
54
55 with time done separately at each silica content (Figure S1 in the Supporting Information).
56
57
58
59
60
15
1
2
3
4
5
6 35
7
8
30 A
9 25
10 20
11
15
12
13 10
14 5
15
Rh / nm
0
16
17 12
18 B
10
19
20 8
21 6
22
23 4
24 2
25
0
26
27 0 5 10 15 20 25 30 35 40 45 50
28 t/h
29
30
31
Figure 3: Hydrodynamic radius from DLS of silica particles as a function of time at pH 7 at variable silica contents:
32 (A) in the presence of [Ca2+] + [Mg2+] = 5 mM (full symbols) at 750 ppm (black diamond), 500 ppm (red triangle),
33 400 ppm (blue circle), 350 ppm (green square); (B) in the presence of additional Na+ at [Na+] = 10 mM (hollow
34 symbols) at 750 ppm (black diamond), 500 ppm (red triangle), 400 ppm (blue circle), 350 ppm (green square).
35
36
37
38 60
39
50
40
41
40
42
43 30
/h
44
45 20
46
47 10
48
49 0
50 300 400 500 600 700 800
51 SiO2 / ppm
52
53
54
55 Figure 4: Lag-time versus initial silica content for the measurement series without added salt (), in the presence of
56 additional Na+-ions () and in the presence of Ca2++Mg2+-ions (). The data recorded in the absence of added salt
57 are taken from Ref. 21. The lag-time has been defined as the intersection of two linear approximations: One
corresponding to the constant mass during the lag-time and a second one corresponding to the tangent adjusted to
58
steepest increase of Mw versus t.
59
60
16
1
2
3
4
5
6 As the observed particle size is very small for all measured samples no information about the
7
8
9 particle shape can be extracted directly from the scattering curves in terms of a formfactor.
10
11 However, the correlation of the hydrodynamic radius or the radius of gyration with the weight
12
13 averaged molar mass gives indirect information about the particle structure. The correlation
14
15
16
follows a power law
17

18 Rg  Rh  M w (11)
19
20 with a shape sensitive exponent  for self-similar structures. Compact, homogeneous spheres or
21
22 cubes lead to an exponent of  = 1/3 and polymer coils under ideal conditions give a value of 
23
24 = 1/2.31 Such an exponent would in fact be observed if any particle reacts with any other particle
25
26
27 in solution corresponding to a step-growth process and resulting in a monomodal size
28
29 distribution. Once particles are formed via a monomer-addition process like in a chain growth,
30
31 the weight-averaged molar particle mass, given by Mw = R= 0/Kc, describes an average of a
32
33
34 bimodal ensemble of monomers and growing particles if the entire silica concentration c0
35
36 corresponding to the concentration at t = 0 is used for the calculation. Knowledge of the
37
38 monomer concentration of silica cmon(t) would enable us to also calculate the concentration of
39
40
41 growing particles cpart(t) according to eq 12 and along with it to calculate an additional weight-
42
43 averaged mass for the particles only with cpart(t) used in eq 3.
44
45 cpart(t)  c0  cmon(t) (12)
46
47 However, if the mass values of an ensemble of growing particles and small monomers for a
48
49 monomer-addition process are based on c0 the topology based exponent is decreased by a factor
50
51
52
of 1/2.32,33 Accordingly, homogeneous spherical or cubic structures formed in a monomer-
53
54 addition process follow an exponent of =1/6.
55
56
57
58
59
60
17
1
2
3
4
5
6
7 A
8
9
10 10
11
12
13
14
15 Rh / nm
16 1
12
17
10
B
18
8
19
20 6
21
22 4
23
24
25
26 2
0.01 0.1 1 10 100
27 -1
28 Mw / 1E4 g mol
29
30 Figure 5: Correlation of the hydrodynamic radius Rh with the weight averaged molar mass Mw at variable silica
31 contents (A) in the presence of Ca2++Mg2+-ions (full symbols): 750 ppm (black diamond), 500 ppm (red triangle),
32 400 ppm (blue circle), 350 ppm (green square); (B) in presence of additional Na+-ions (hollow symbols): 750 ppm
33 (black diamond), 500 ppm (red triangle), 400 ppm (blue circle), 350 ppm (green square). The solid black line
34 indicates a slope of 1/2 and the dashed black line represents a slope of 1/6.
35
36
37 The correlation of Rh versus Mw for the measurement series in the presence of Ca2++Mg2-ions
38
39
40 are shown in Figure 5A and in the presence of Na+-ions in Figure 5B. The mass values shown
41
42 in Figure 5 have been calculated based on the initial silica concentration c0. The trends of the
43
44 correlations for the two types of salts are comparable. In both types of salts two different
45
46
47 exponents are observed. Silica contents of 350-500 ppm result in a slope of 1/6 and experiments
48
49 with 750 ppm SiO2 show a slope according to an exponent =1/2. As the smallest possible
50
51 exponent based on the topology is 1/3 corresponding to homogeneous spherical structures, the
52
53
54 lower exponent of 1/6 observed below 750 ppm is only explainable by the formation of compact
55
56 particles with homogeneous density (most likely of spherical shape) via a monomer-addition
57
58 process. The particle growth process at 750 ppm in presence of additional metal cations exhibits
59
60
18
1
2
3
4
5
6 no lag-time. Consumption of the monomers is fast and the weight-averaged molar mass is
7
8
9 increasing beyond the time when the equilibrium concentration of the monomers is reached. No
10
11 plateau value is reached for the particle size. The correlation of Rh with Mw gives an exponent of
12
13 1/2. The same exponent holds true for the correlation of Rg with Mw, which is only accessible for
14
15
16
the experiment in the presence of Ca2++Mg2+ as in this case the particles have reached a
17
18 sufficiently large size to detect a radius of gyration. The corresponding graph can be found in the
19
20 Supporting Information (Figure S2a).
21
22
For the exponent of 1/2 observed at 750 ppm two explanations are possible. (i) A particle
23
24
25 formation via monomer-addition is assumed, where we would have to multiply the observed
26
27 exponent by a factor of two to get the topology based exponent.32,33 The resulting exponent of 1
28
29 would suggest rod-like structures. However, given the state of knowledge on the formation of
30
31
32 colloidal silica under the present conditions2,5, a rod-like structure seems to be very unlikely.
33
34 Further arguments against a rod-like structure is provided by the shape-sensitive factor  = Rg/Rh
35
36 (eq 10) and by the dimensionless parameter C (eq 8), which are both represented for the
37
38
39 experiment at 750 ppm in the presence of Ca2++Mg2+ in the Supporting Information (Figure S2b
40
41 and Figure S2c). For rod-like structures a value of  ≥ 2 would be expected.29,30 Values in
42
43
44
Figure S2b are close to  ≈1.3 which is in agreement with the value observed for unperturbed
45
46 polymer chains28 but significantly smaller than  ≥ 2 expected for rod-like coils. The
47
48 dimensionless parameter C shown in Figure S2d approaches 0.18, which is close to 0.173
49
50
51 predicted for unperturbed Gaussian polymer chains26 and significantly larger than 0.033 < C <
52
53 0.044 anticipated for rod-like structures34. (ii) A growth based on a particle-particle
54
55 agglomeration also denoted as step-growth is assumed, where the observed exponent can be
56
57
58
related directly to the topology of the particles.32,33 Hence, an exponent of 1/2 indicates fractal-
59
60
19
1
2
3
4
5
6 like31,35 structures, with a fractal dimension of 2. In fact a comparison with Gaussian polymer
7
8
9 coils, which also have a fractal dimension of 2 and  ~ 1.3 and C ~ 0.175 teaches us that
10
11 explanation (ii) is appropriate. Most likely such structures are formed with the homogeneous
12
13 particles generated via monomer-addition, which succeedingly react via step-growth
14
15
16 polymerization.
17
18 These results together with the perfect correlation of the evolution of the particle mass with the
19
20 consumption of monomers, enable us to discriminate two concentration regimes with two
21
22
23 mechanisms of particle growth. The particle growth at 350 ppm-500 ppm SiO2 is characterized
24
25 by a distinct lag-time during which monomers are not consumed followed by particle growth via
26
27 monomer addition. Above a concentration of 500 ppm the particles aggregate to fractal-like
28
29
structures according to a step-growth process. These results are consistent with the findings for
30
31
32 silica polymerization in water without added salt, which have been already presented in a
33
34 previous publication21.
35
36
37
38
39 100
40
41
Rh / nm
42
43
44
10
45
46
47
48
49
50 0.1 1 10 100 1000
51 -1
52
Mw / E4 g mol
53
54 Figure 6: Correlation of the hydrodynamic radius Rh with the weight-averaged molar mass Mw=R/Kcpart(t) at the
55 silica contents 2000 ppm in pure water (empty, blue circle), 3000 ppm in pure water (full, pink circle), 750 ppm in
56 the presence of Ca2++Mg2+ (full, black diamond) and 750 ppm in the presence of additional Na+ (empty, black
57 diamond). Experimental data corresponding to silica polymerization in pure water is taken from Ref. 21. The dashed
58 line represents a slope of 1/2.
59
60
20
1
2
3
4
5
6 At this point, we would like to stress a comparative representation of the evolution of the
7
8
9 hydrodynamic radius with molecular weight for the growth processes which obey a step-growth
10
11 process under all three conditions. To this end values of the molecular weight of the fractals have
12
13 been calculated according to Mw = R/K·cpart(t) with cpart(t) = c0 - cmon(t) and cmon(t) ≈ 170 ppm
14
15
16 the (constant) equilibrium concentration of the non-consumed monomers. Justification for this
17
18 calculation of Mw is that the scattering contribution of the non-consumed monomers is negligible
19
20 which implies use of the (constant) concentration of growing fractals c0 – 170 ppm. As is
21
22
23 demonstrated in Figure 6, the corresponding correlations do not only follow a power law with an
24
25 exponent of 0.5 but also overlay perfectly. Such an overlay can be nicely reconciled with the fact
26
27 that independent on the actual conditions, i.e. independent on whether additional salt is absent or
28
29
present as Na+ or as Ca2++Mg2+, always the same fractals are formed and the same constituent
30
31
32 building units are incorporated with the building units corresponding to the homogeneous
33
34 particles formed respectively via the monomer-addition process. A variation of the size of the
35
36 constituent particles when changing the conditions does not affect the resulting exponent of the
37
38
39 power law in eq 11, but it would generate a parallel shift in the log-log plot as a result of a
40
41 variation of the prefactor. The coincidence of all curves is particularly interesting as this implies,
42
43 that the step-growth like agglomeration sets in always with constituent particles having a size in
44
45
46 the order of 1-3 nm.
47
48 Further support for the suggested mechanisms and morphologies is provided by the correlation
49
50 of the finally measured hydrodynamic radii with the corresponding mass values, as it is done in
51
52
Figure 7. Here, the mass values of the particles have been calculated according to eq 3 with the
53
54
55 mass concentration of the growing particles cpart(t) from eq 12 instead of using c0. In cases where
56
57 the growth process was still going on while the final measurement was done cmon(t) was
58
59
60
21
1
2
3
4
5
6 interpolated from the respective trend determined by the molybdenum blue method. In cases
7
8
9 where the growth process was completed cmon(t) corresponds to the equilibrium concentration of
10
11 170 ppm SiO2. Figure 7 is a comparison of data from three measurement series: silica
12
13 polymerization (i) in water without added salt from Ref. 21, (ii) in the presence of additional
14
15
16
Ca2++Mg2+ and (iii) in the presence of additional Na+. At all conditions the final values measured
17
18 for silica contents below 750 ppm follow a unique trend with a slope of 1/3. Noteworthy, these
19
20 are the conditions where no aggregation of particles takes place. This result confirms the
21
22
existence of homogeneous particles with a uniform characteristic density like for instance cubes
23
24
25 or spheres.
26
27
28
29
30
31
final Rh / nm
10
32
33
34
35
36
37
38
39
40 1
41 0.1 1 10 100
42 final Mw(cpart) /1E4 g mol-1

43
44
45
46 Figure 7: Correlation of Rh(t = tmax) with the final values of Mw=R/Kcpart(t = tmax) at variable silica contents:
47 750 ppm (black), 500 ppm (red), 400 ppm (blue), 350 ppm (green), empty symbols indicate the experiments in
48 water without added salt, full symbols the experiments in the presence of Ca2++Mg2+ and crosses indicate the
49 experiments in the presence of additional Na+. The straight line represents a slope of 1/3. For the sample 750 ppm in
50 the presence of Ca2++Mg2+ two data points are shown. The filled black diamond corresponds to the final size and
51 mass value, where the aggregation has been already started. The black star indicate one of the very early
52 measurements, where the aggregation process of this sample has not yet started.
53
54
55 As expected, the situation is different at later stages of the experiments with 750 ppm silica in
56
57
the presence of additional salts (black cross and black diamond in Figure 7). As no equilibrium
58
59
60
22
1
2
3
4
5
6 values for Rh and Mw is reached under those conditions, we choose data at a later time, when the
7
8
9 monomer concentration has reached its equilibrium value since long. At t = 50 h in the presence
10
11 of Ca2++Mg2+ the hydrodynamic radius amounts to Rh = 26 nm and the weight-averaged molar
12
13 mass is Mw = 1.40E+5 g/mol to give but an example. This data point lies far off the trend of the
14
15
16 correlation with the exponent  = 1/3 (Figure 7). This can be nicely reconciled with the growth
17
18 of fractal-like particles. However, if we consider the first data point of this TR-SLS/DLS
19
20 experiment the value of the hydrodynamic radius (Rh = 3 nm) and the weight-averaged molar
21
22
23 mass (Mw = 2.02E+3 g/mol) fits perfectly to the trend based on the exponent  = 1/3 (black star
24
25 in Figure 7). The mass value Mw = 2.02E+3 g/mol has been calculated under the assumption that
26
27 monomer concentration has already reached its equilibrium value. This is a hint that the early
28
29
30 intermediates still refer to spherical particles formed by a monomer-addition process and thus
31
32 nicely confirms that monomer-addition, which is observed for all lower silica concentrations,
33
34 also takes place at the initial period of the process at a silica content of 750 ppm in the presence
35
36
37
of an added salt. Only the rate of nucleation and growth is so fast that it is already completed
38
39 soon after the initiation. Due to an insufficient stabilization of the spherical particles, particle-
40
41 particle aggregation starts immediately and is dominating the growth process monitored by light
42
43
44
scattering measurements. The trend in  presented in Figure S2b of the Supporting Information
45
46 further confirms this aspect.
47
48 Although the final particle size as well as the lag-time slightly depend on the solution
49
50
condition the general trends of the evolution of the size are comparable for all conditions under
51
52
53 consideration i.e. in the presence of additional Ca2++Mg2+-ions or Na+-ions and in the absence of
54
55 any added salt. We thus can state that the presence of additional metal cations has no significant
56
57 effect on the general mechanism of the silica polymerization. For all three conditions, a change
58
59
60
23
1
2
3
4
5
6 from a monomer-addition mechanism generating compact particles with a homogeneous density
7
8
9 to a particle-particle aggregation leading to fractal-like particles with the homogeneous particles
10
11 as building units is observed once a certain silica content is reached. Independent on when this
12
13 change of the growth mechanism takes place, the homogeneous particles corresponding to the
14
15
16
building units have reached a size of Rh ≈ 1-3 nm. Only the concentration threshold at which the
17
18 aggregation of the particles becomes dominant is shifted to lower silica concentrations, if metal
19
20 cations are present. Whereas in the presence of metal cations this threshold is smaller than
21
22
750 ppm, the threshold concentration lies between 750 ppm and 2000 ppm SiO2 in solutions
23
24
25 without added salts.21 In none of the three steps characterizing the process of silica
26
27 polymerization does Ostwald ripening play a significant role. During the second step compact
28
29 homogeneous particles grow via monoer-addition as inferred from the exponent 0.5 in eq 11.
30
31
32 Growth by Ostwald ripening during this step can be excluded as this would have led to an
33
34 exponent of 1/3 indistinguishable from a step-growth process. During the third step fractals grow
35
36 with a fractal dimension of 2 thus excluding Ostwald ripening also during this step.
37
38
39
Kinetic Modelling. In a previous publication21 we introduced a kinetic model denoted as
40
41 nucleation and growth model (NG-model) in order to describe the process of silica
42
43 polymerization in distilled water. It is a simple process consisting of three successive reactions
44
45
(i) a precursor reaction, in which the reactive monomer is formed, (ii) a nucleation step and (iii)
46
47
48 an irreversible growth process via monomer-addition. The reactions (i) and (ii) establish the first
49
50 step of the 3-step mechanism. Unexpectedly, the rate constant of the precursor step varied over
51
52 several orders of magnitude as the initial silica content increased. This likely happened because
53
54
55 the nucleation step in which one monomer acts as a nucleus is too simple to account for the
56
57 complex nucleation or initiation step in reality. This deficiency may have been compensated by a
58
59
60
24
1
2
3
4
5
6 variation of the rate constant of the precursor step which helped to vary the induction period with
7
8
9 increasing initial silica content. Further on, the NG-model does not account for an equilibrium
10
11 concentration of the monomers, as only growth and no degradation of the particles is permitted.
12
13 An improvement of this model denoted as NGE-model shall be introduced in the present work.
14
15
16
The new model consists of the same three basic reactions, a precursor reaction, a nucleation and
17
18 a particle growth via monomer addition. However, the present model explicitly takes into
19
20 account a variable size of the nuclei indicated by the degree of polymerization n and it considers
21
22
explicitly depolymerization as the back reaction of monomer-addition.
23
24
25 kp
26
Precursor reaction A B (13)
27
28
29 kn
30 Nucleation nB Ci (14)
31
32
33 ke
34 Addition/Degradation B+Ci Ci+1 (i ≥ n =1,2,3...) (15)
35 k-e
36
37
38 Like for the original NG-model21 it is possible to set up five differential equations which
39
40 describe the first derivative with respect to time of the precursor concentration [A], the monomer
41
42
43 concentration [B], and the zeroth [C](0), first [C](1) and second moment [C](2) of the particle
44
45 ensemble. A detailed derivation of these differential equations is represented in the Supporting
46
47 Information.
48
49
50
Although the NG- and the NGE-model simulate the nucleation/initiation of the silica formation
51
52 process in a very simple way, they provide a clear benefit for the interpretation of time-resolved
53
54 light scattering data. Both models do not only allow to simulate the consumption of the
55
56
monomers and with that the changing supersaturation with time, but also to calculate the
57
58
59 variation of the moments of the particle ensemble with time, which makes possible for the first
60
25
1
2
3
4
5
6 time a direct comparison with the weight-averaged molar mass detected via SLS. Beyond doubt
7
8
9 the strength of the model lies on the description of particle growth, which is exactly what is
10
11 recorded by TR-SLS/DLS. The weight-averaged molar mass Mw accessible via light scattering is
12
13 an average over all species, including the precursor concentration, the monomer concentration
14
15
16
and all polymeric particles signified by the degree of polymerization i. Expressed in terms of
17
18 parameters and variables of the model this Mw reads
19
 
20
21 Mw (t) 
M0
A(t)  B(t)  C2 (t) (16)
22 A0
23
24 and is accessible by integrating numerically the differential equations for [A], [B] and [C](2)
25
26 (Supporting Information eqs S3, S6 and S7). M0 is the molar mass of the monomeric unit and is
27
28
29
set to 60.1 g/mol corresponding to a SiO2 unit and [A]0 is the initial concentration of the
30
31 precursor and is given by the respective experimental silica concentration.
32
33 The results of experiments under where no salt has been added21 will be used first to discuss
34
35
the performance of the NGE-model in comparison to that of the simple NG-model. Figure 8
36
37
38 represents a comparison of the fitting results obtained by the original NG-model and the new
39
40 NGE-model. A significant improvement could be achieved for the silica contents 350 and
41
42 400 ppm. The fit with the NGE-model is particularly improved in the time range of the initial
43
44
45 increase of the particle mass. In case of the highest silica contents of 500 and 750 ppm the NGE-
46
47 model does not significantly improve the fit quality achieved with the NG-model but helps to
48
49 confine kp to a range much narrower than in case of the NG-model. A more detailed presentation
50
51
52
of the fits with the NGE-model with regard to the weight-averaged molar mass, the monomer
53
54 concentration, the concentration of particles and 2 can be found in Chapter S4 of the Supporting
55
56 Information.
57
58
59
60
26
1
2
3
4
5
6 20 0.8
Mw / E4 g mol-1
7
8
18
A 0.7
16
0.6
9 14
10 12 0.5
11 10 0.4
12 8 0.3
13 6
0.2
14 4
0.1
15 2
16 0 0.0
cmon(SiO2) / ppm
17 700 B
18
600
19
20 500
21
400
22
23 300
24 200
25
(solubility limit 170 ppm)
26 100
0 20 40 60 80 100 120 140 160
27
28 t/h
29
30 Figure 8: Comparison of the experiments without added salt21 (750 ppm (black diamond), 500 ppm (red triangle),
31 400 ppm (blue circle), 350 ppm (green square)) with the fits based on the NG-model (dashed line) and the NGE-
32 model (solid line) in terms of the weight-averaged molar mass (A) and the monomer consumption (B). Note that the
33 mass values at a silica content of 500 ppm and 750 ppm are rescaled by a factor of 10. The arrows point to the
34 corresponding axis.
35
36 Classical nucleation theory36–39 predicts a decreasing nucleus size with increasing
37
38 supersaturation. In the light of this theory the concentration series has an increasing initial degree
39
40
of supersaturation in going from 350 ppm to 750 ppm SiO2. The NGE-model enables us to
41
42
43 consider this feature qualitatively by varying the size n of the critical nucleus. However, one has
44
45 to keep in mind that during the growth process obeying classical nucleation theory the critical
46
47 nucleus size increases with decreasing monomer concentration (with decreasing degree of
48
49
50 supersaturation), which is not accounted for by the NGE-model. In order to determine the
51
52 optimal nucleus size for each initial silica concentration, the fitting has been performed for
53
54 different nucleus sizes in the range of 1 ≤ n ≤ 30. The fit with the lowest χ2-value (Supporting
55
56
57
Information, eq S9), has been used to identify the optimal nucleus size. The resulting optimal n-
58
59 values are summarized in Table 1. Unfortunately, identification of n based on 2 values could
60
27
1
2
3
4
5
6 not be achieved with the same degree of significance for all four silica contents (Figure S4c of
7
8
9 the Supporting Information). However, the data show that the most suitable nucleus size n
10
11 (Table 1) is decreasing with increasing initial silica content, which is in line with the predictions
12
13 of the classical nucleation theory. It has to be stressed that the formation and growth of silica
14
15
16
particles is based on the incorporation of monomeric building units via chemical bonding
17
18 corresponding to a three dimensional polymerization or polycondensation reaction, which may
19
20 well differ considerable from a pure crystallization following a classical nucleation process.
21
22 Table 1: Rate constants at variable silica contents in the absence of salt at pH 7 from fits with the NGE-model to
23 data from Ref. 21. The value n describes the size of the theoretical nuclei in eq 14 which yields the smallest χ2–
24 value, kp is the rate constant for the precursor reaction, kn for the nucleation and ke for the growth reaction in eqs 13-
25 15.
26 c0, exp n n kp kn ke χ2
27
28
[ppm] (fixed) (optimal) [h-1] [L/mol)n-1h-1] [L(mol h)-1] [g mol-1]
29 350 8 3.15E-2 5.25E14 8.67E4 1.70E1
30 350 1 2.05E-2 1.42E-5 2.03E5 3.69E1
31 salt- 400 2 5.50E-2 1.04E-1 1.86E5 4.85E1
32 free 400 1 4.99E-2 1.06E-4 2.42E5 5.12E1
33
34
500 1 1.91E-1 1.03 6.15E6 3.92E1
35 750 1 1.94E-1 5.08 3.81E6 8.15E1
36
37
38 Unlike to the behavior in the presence of salt, silica particles generated in a solution of
39
40
41 750 ppm do not show particle-particle aggregation in the absence of added salt. In that case the
42
43 concentration threshold, which marks the beginning of a particle-particle aggregation lies well
44
45 above 750 ppm. However, as inferred from the poorer performance of the NG and NGE-model at
46
47
the two silica contents 500 ppm and 750 ppm in salt-free solution, those two higher silica
48
49
50 contents are much less dominated by particle growth. After a very fast increase the
51
52 hydrodynamic radius remains constant within experimental uncertainty at a value of 4-5.5 nm
53
54 while the weight-averaged molar mass is still slightly increasing. This behavior suggests that the
55
56
57 increase of Mw is dominated by the formation of further particles rather than by incorporation of
58
59 monomers to growing particles.
60
28
1
2
3
4
5
6 The experimentally observed generation of the particles with the size around 4-5.5 nm at
7
8
9 500 and 750 ppm SiO2 is interpreted as follows by the NGE-model. A very large nucleation rate
10
11 leads to the final amount of particles in a short time. These particles have to add the remaining
12
13 monomers necessary to reach the final mass values also very fast in order not to compete with
14
15
16
further nucleation since further nucleation would make the final particle mass too low.
17
18 Accordingly, the fits lead to a constant size and weight-averaged molar mass of the particles
19
20 soon after nucleation, which appears not to be fully compatible with data from 500 ppm and
21
22
750 ppm. A simple increase of the nucleation rate cannot describe gradually increasing mass at
23
24
25 later stages of the reaction.
26
27 Evaluation with the NGE-model shall be concluded with a focus on the different rate
28
29 constants. Table 1 summarizes the optimized rate constant of the precursor reaction kp derived
30
31
32 from the fits to the experimental data for silica in the absence of added salt solution at pH 7 for
33
34 different nucleus sizes n. The precursor reaction was introduced in order to model the induction
35
36 period, because a simple formation model consisting of a nucleation and a growth reaction with a
37
38
39
rate constant independent of the degree of polymerization is not able to reveal an induction
40
41 period. We expected a constant value for kp for all initial silica contents. In this respect the NG-
42
43 model failed, since the rate constant of the precursor reaction was increasing by several orders of
44
45
magnitude with increasing silica content21. The NGE-model with its nucleation/initiation
46
47
48 reaction of variable order n significantly improved this issue. The rate constants of the precursor
49
50 reaction kp retrieved with the NGE-model all fall within one order of magnitude. A comparison
51
52 of the rate constants for the precursor reaction, the nucleation and the growth reaction from fits
53
54
55 with the NGE-model and the NG-model are given in Table S1 of the Supporting Information.
56
57 The rate constant for the growth ke is in the order of 105 L(mol h-1) at 350 ppm and 400 ppm of
58
59
60
29
1
2
3
4
5
6 silica and increases by one order of magnitude at the higher silica contents. This jump is a hint
7
8
9 for deficiencies in the NGE-model as soon as the growth is not the dominating part any more.
10
11 We conclude that modelling of the silica polymerization could be improved by the NGE-model
12
13 with respect to the following aspects. The variation of the rate constant of the precursor reaction
14
15
16
could be confined to one order of magnitude, which is significantly narrower than the respective
17
18 variation over eleven orders of magnitude observed with the NG-model.21 If we consider only
19
20 the two lowest silica concentrations, for which we have unambiguously demonstrated, that a
21
22
monomer-addition mechanism is dominating, the deviations in kp are even lower. Variation of
23
24
25 the nucleus size has shown that the silica polymerization follows the trend of a decreasing
26
27 nucleus size with increasing degree of supersaturation, which is in accordance with classical
28
29 nucleation theory. The equilibrium state of the silica polymerization could be successfully
30
31
32 modelled by the implementation of a depolymerization reaction in which the polymers release
33
34 monomers. A possible explanation for the deficiencies of the NG and NGE-model observed with
35
36 the two higher silica contents would be an ongoing generation of nuclei as particles similar in
37
38
39
size during the entire experimental time period, with a single type of particle being formed
40
41 within a short period of time compared to the length of the overall period. Such a feature could in
42
43 fact also be described by the NGE-model if n would be fixed at values large enough to
44
45
correspond to 3-4 nm in size and if the rate constant of this nucleation would be very high in
46
47
48 comparison to the rate constant of the monomer-addition. This results in a transfer of the main
49
50 fraction of the active monomers into nuclei with a size large enough to fit Mw values in the order
51
52 of 4·103-8·103 g mol-1 and Rh values close to 4 nm. However, such size and mass values of a
53
54
55 nucleus which are much larger than the n values observed at low silica contents would not be
56
57 compatible with the classical nucleation theory because the size of nuclei are expected to get
58
59
60
30
1
2
3
4
5
6 smaller as the degree of supersaturation (i.e. c0) increases and not larger. A possible explanation
7
8
9 is that the initiation reaction as part of step 1 is not fully in accordance with the nucleation in the
10
11 sense of classical nucleation theory.
12
13 25
14 4
15 20
16 3
15
17
18 10
2
Mw / 1E4 g mol-1
19
20 5 1
21
22 0 0
23 5
1.5
24
25 4
26 3
1.0
27
28 2
0.5
29 1
30
31 0 0.0
32 0 10 20 30 40 50 60 70
33 t/h
34
35 Figure 9: Comparison of the experimental weight-averaged molar mass Mw with the corresponding fits with the
36 optimal nucleus size at variable silica contents (750 ppm (black diamond), 500 ppm (red triangle), 400 ppm (blue
37 circle), 350 ppm (green square)). (A) Experiments in the presence of Ca2++Mg2+-ions; (B) experiments in the
38 presence of additional Na+-ions. Note that the experimental data have different scales for a better overview. The
39 arrows point to the corresponding axis. The model calculations based on the NGE-model are represented as solid
40 lines.
41
42
43
After having evaluated the NGE-model with the data recorded under salt-free condition,
44
45
46 experiments in the presence of Ca2++Mg2+ and in the presence of additional Na+ shall now be
47
48 interpreted with the NGE-model. The experimental data corresponding to a silica content of
49
50 750 ppm has been excluded from the optimization process, as the correlation of the size with the
51
52
53 mass values (shown in Figure 5) revealed that the particle formation is dominated by a step-
54
55 growth like particle-particle aggregation. The fits with the NGE-model are represented in Figure
56
57 9 and the corresponding rate constants obtained from the optimization applied to the mass
58
59
60
31
1
2
3
4
5
6 evolution of the experiments in the presence of metal cations are given in Table 2. For each
7
8
9 silica content the results corresponding to the optimal nucleus size determined by the smallest χ2-
10
11 value are shown. Further details of the fits are given in the Supporting Information (Figure S5a
12
13 and S5b). The variation of the rate constant of the precursor reaction in the presence of added
14
15
16
salt extends over two orders of magnitude only and hence is similar to that of the salt-free case.
17
18 The rate constants of the growth reaction lie again in the order of magnitude ~ 105 L(mol h-1)
19
20 at silica contents lower than 500 ppm and can be considered as constant within the range of
21
22
uncertainty. We thus can conclude that the absence or presence of added Ca2++Mg2+ or Na+ does
23
24
25 not significantly affect the rate constant ke of the particle growth.
26
27 Table 2: Rate constants at variable silica contents in the presence of Ca2++Mg2+- ions and Na+-ions at pH 7 from the
28 optimization with the NGE-model. The value n describes the size of the theoretical nuclei which yields the smallest
29 χ2–value, kp is the rate constant for the precursor reaction, kn for the nucleation and ke for the growth reaction. The
30 star indicates cases when χ2 was continuously decreasing with n and where the values correspond to the fit with the
31 maximum possible n.
32 χ2
c0,exp n n kp kn ke
33 [g mol-
[ppm] (fixed) (optimal) [h-1] [(L/mol)n-1h-1] [L(mol h)-1] 1
34 ]
35
350 8 9.39E-2 2.26E14 2.08E5 7.91E1
36
37 350 1 5.97E-2 1.89E-5 3.83E5 1.43E2
38 Ca2++Mg2+ 400 20* 1.29E-1 5.07E51 1.18E5 6.66E1
39 400 1 9.89E-2 5.83E-4 2.50E5 8.98E1
40 500 1 2.94E-1 4.12 2.71E7 3.87E1
41 350 30* 1.16E-1 2.37E71 4.01E4 7.03
42
350 1 4.49E-2 1.37E-4 1.09E5 8.12
43 +
44 Na 400 10 2.16E-1 1.12E19 4.46E4 8.82E1
45 400 1 1.58E-1 8.00E-5 5.39E4 1.33E2
46 500 1 2.60E-1 5.23 5.90E7 7.29E1
47
48
49
50 The fits to data recorded in the presence of salt give an optimal nucleus size of n = 1 at the
51
52 silica content of 500 ppm. At lower silica contents, the optimal size of the nuclei is larger than 1.
53
54 Unfortunately in two cases a continuously decreasing χ2 has been observed with increasing n
55
56
57 (Table 2, Figure S5d and S5f) which prohibits identification of the optimal n in those two cases.
58
59 Also the significance of the optimal nucleus size is not always strong. However, the fits with the
60
32
1
2
3
4
5
6 NGE-model suggests that the size of the nuclei is decreasing with increasing initial
7
8
9 supersaturation, which is in line with the classical nucleation theory.
10
11 A characteristic magnitude for the nucleation/initiation is the number of generated particles,
12
13 which is described by the zeroth moment. Figure 10 compares the final value of the zeroth
14
15
16
moment, depending on the silica content and the reaction conditions. The particle number is
17
18 clearly increasing with increasing silica content, which confirms that the nucleation/initiation is
19
20 accelerated and plays a more dominating role with increasing supersaturation. The influence of
21
22
an added salt has only a minor effect on the number of formed particles and shows no clear
23
24
25 trend.
26
27
28
[C](0) / 1E-6 g mol-1
29
30
31 10
32
33
34
35
1
36
37
38
39
340 360 380 400 420 440 460 480 500
40
41 c0 / ppm
42
43
Figure 10: Final number of particles represented as zeroth moment based on the optimal fit with the NGE-model for
44
the corresponding silica content c0 for different conditions: Silica in water without added salt (); in the presence of
45 additional Na+ () and in the presence of Ca2++Mg2+ ().
46
47
48
49
50 To summarize, the NGE-model unlike to the NG-model can reproduce the evolution of mass at
51
52 variable initial silica concentration by an interplay of the rate constant of the precursor reaction
53
54 and the size of the nuclei, which is connected with a change in the reaction order of the
55
56
57
nucleation step. This works satisfactorily in case of the lowest silica concentrations 350 ppm and
58
59 400 ppm independent of the conditions, suggesting that the addition of salt does not influence the
60
33
1
2
3
4
5
6 general mechanistic features of the reaction. In line with this the kinetic parameters do not show
7
8
9 significant differences among the three conditions (absence of added salt, presence of
10
11 Ca2++Mg2+, presence of additional Na+) except for the rate constants of the precursor reaction
12
13 which are persistently lower if no salt had been added. It has to be emphasized that the precursor
14
15
16
reaction together with the initiation/nucleation reaction of the NG- and NGE-model establish the
17
18 first step of the 3-step mechanism. This first step usually denoted as nucleation step is
19
20 increasingly dominating the process of particle formation, once the silica content increases
21
22
beyond 400 ppm and the NGE-model like the NG-model performs progressively poorer. This
23
24
25 becomes obvious not only in terms of the decreasing fit quality but also in terms of the resulting
26
27 rate constants of the growth reaction, which at 500 ppm are deviating from those obtained at the
28
29 lower silica contents.
30
31
32
33
34 CONCLUSIONS
35
36 In a preceding work, the formation of silica particles in pure water at a pH = 7 covering a silica
37
38
39
concentration regime, which extends from 350 ppm to 750 ppm has been analysed by time-
40
41 resolved combined DLS/SLS. The work unambiguously revealed a short nucleation or initiation
42
43 step, which is followed by a particle growth based on a monomer-addition mechanism. A simple
44
45
kinetic scheme, the NG-model, could satisfactorily reproduce the evolution of the weight-
46
47
48 averaged particle mass and the consumption of monomers at 350 ppm and 400 ppm, however, its
49
50 performance turned out to be poorer at the higher silica contents of 500 ppm and 750 ppm. This
51
52 was attributed to an increasing influence of the nucleation/initiation step as the silica content
53
54
55 increases. The present work extends this knowledge in two different ways. First and foremost,
56
57 analogous experiments have been performed in the very same concentration regime of silica,
58
59
60
34
1
2
3
4
5
6 now in the presence of additional Ca2++Mg2+ or Na+ at a level of 0.01 N respectively in order to
7
8
9 analyze the impact of alkaline earth cations on the formation process of silica particles.
10
11 Comparison with Na+-ions does not reveal any ion specific effect for alkaline earth cations.
12
13 Second, the simple NG-model was refined with respect to two aspects, the variation of the size n
14
15
16
of the nuclei which are formed in a simple reaction of order n and the explicit consideration of a
17
18 monomer release step as the back reaction of the monomer addition. The refined model is named
19
20 NGE-model in the present work. Like for the NG-model a central part of the theoretical work
21
22
was to develop kinetic equations, which are capable of yielding the first three moments of the
23
24
25 resulting particle size distribution since it is this data, which is required for a comparison with
26
27 light scattering results.
28
29 In order to provide suitable reference data, the NGE-model was first verified with the data
30
31
32 from the preceding work collected in the absence of additional salt.21 Several features
33
34 demonstrated a better performance of the refined NGE-model. The fit improved particularly in
35
36 the regime of the lower silica contents. The rate constant of the precursor reaction could be
37
38
39
successfully confined to a range extending only over one order of magnitude which is much
40
41 better than the spread of the rate constant over eleven orders of magnitude observed for the NG-
42
43 model. Further on, the NGE-model indicated a decrease of the nucleus size with increasing silica
44
45
content in line with classical nucleation theory. With this in mind a more comprehensive
46
47
48 discussion of the new results recorded in the presence of an additional salt became possible.
49
50 Independent of the conditions, i.e. in the absence of added salt21 and in the presence of
51
52 additional Na+ and of Ca2++Mg2+, a three step process can be identified in agreement with
53
54
55 preceding work.5,19,21 The steps are (1) a nucleation/initiation step, (2) particle growth stage and
56
57 (3) agglomeration of the particles from step 2. The present work together with Ref. 21 unravels
58
59
60
35
1
2
3
4
5
6 further details on step 2 and 3 with combined time-resolved SLS/DLS as new in situ technique
7
8
9 and the molybdenum blue method.
10
11 Addition of salt reduced the lag-time preceding particle growth via monomer-addition. In
12
13 terms of the NGE-model this is captured by a slightly increased rate constant of the precursor
14
15
16
reaction and an increased/accelerated formation of nuclei/particles. The particles nucleated in
17
18 this first step grow according to a monomer-addition mechanism where the monomers may
19
20 include monomeric silicic acid and oligomers thereof. These features could unambiguously be
21
22
inferred from the coincidence of disappearing molybdate-reactive “monomers” with the particle
23
24
25 growth and from a correlation of particle size (from DLS) with particle mass (from SLS)
26
27 resulting in characteristic power laws of Rh ~ Mw1/6. Ostwald-ripening can be excluded during
28
29 step 2 as this would have modified the exponent of 1/6. The rate constant of the growth reaction
30
31
32 lies in the same order of magnitude for all three conditions (without added salt, with additional
33
34 Na+ at [Na+]= 0.01 N and with additional Ca2++Mg2+ at [Ca2+4/5Mg2+1/5]=0.01 N) as long as the
35
36 nucleation/initiation is not dominating the particle formation, which holds for silica contents
37
38
39
lower than 500 ppm.
40
41 The decrease in lag-time together with the increase of the precursor rate constant induced by
42
43 the addition of salt suggests that the added salt most significantly affects the first step of the
44
45
particle formation process. In line with this, no significant variation of the growth rate constant
46
47
48 could be discerned with the addition of salt. Although similar trends were reported in
49
50 literature15,16, we hesitate to attribute this effect to a lowering of the solubility of silica by salt
51
52 addition16 since we keep well below a salt content of 40 mM and vary the salt concentration only
53
54
55 within a very narrow range of 10 mM.
56
57
58
59
60
36
1
2
3
4
5
6 Agglomeration as the third step sets in once the concentration of silica is large enough. It
7
8
9 proceeds via a step-growth like particle-particle aggregation. As extracted from SLS/DLS-data,
10
11 the resulting agglomerates have a fractal dimension close to 2. Formation of fractals as step 3
12
13 discards Ostwald-ripening as mechanistic feature of this step. The onset of particle
14
15
16
agglomeration according to this step-growth process, which in the absence of added salt was first
17
18 observed at a silica content of 2000 ppm is shifted to lower values by adding salt. Agglomeration
19
20 takes place already at 750 ppm in the presence of both types of cations. Hence the presence of
21
22
cations destabilizes particles with respect to particle-particle agglomeration. This occurs most
23
24
25 likely due to screening of electrostatic interactions among the particles. However, the presence of
26
27 Ca2++Mg2+ and the presence of Na+ does not affect the fractal dimension of the growing
28
29 intermediates. Even more strikingly, the presence of an additional salt does not even affect the
30
31
32 nature of the constituting particles. Independent on whether salt is present or absent nucleation
33
34 succeeded by growth via monomer-addition lead to particles with the same critical size of a few
35
36 nanometers which then agglomerate to a unique type of fractal.
37
38
39
40
41 ASSOCIATED CONTENT
42
43 Supporting Information
44
45
The Supporting Information is available free of charge on the ACS Publications website.
46
47
48 Comparison of the silica particle growth in salt-free solution, in the presence of Ca2++Mg2+
49
50 and in the presence of additional Na+; detailed data of structural intermediates at 750 ppm SiO2
51
52 in the presence of Ca2++Mg2+; derivation of the NGE-model and representation of the
53
54
55 corresponding fit results for all three investigated reaction conditions.
56
57
58
59
60
37
1
2
3
4
5
6 AUTHOR INFORMATION
7
8
9
Corresponding Author
10
11
12 *E-mail: (K.H.) klaus.huber@upb.de. Telephone: (+49) 5251602125. Fax: (+49) 5251 604208.
13
14
15
16
17
18 Notes
19
20
21
The author declares no competing financial interest.
22
23
24
25
26 ACKNOWLEDGMENT
27
28
29 The BASF SE is acknowledged for funding this work within the project “Polymer Assisted Silica
30
31 Polymerization”.
32
33
34
35
36
37
38 References
39
40 (1) Patwardhan, S. V., Clarson, S. J., Perry, C. C. On the role(s) of additives in bioinspired
41 silicification. Chemical communications (Cambridge, England) 2005, 1113–1121.
42 (2) Belton, D. J., Deschaume, O., Perry, C. C. An overview of the fundamentals of the
43 chemistry of silica with relevance to biosilicification and technological advances. The FEBS
44 journal 2012, 279, 1710–1720.
45 (3) Gunnarsson I., Arnórsson S. Silica scaling:the main obstacles in efficient use of high-
46
47
temperature geothermal fluids. Int. Geoth. Conf. Reykjavik 2003, 30–36.
48 (4) Kempter, A., Gaedt, T., Boyko, V., Nied, S., Hirsch, K. New insights into silica scaling on
49 RO-membranes. Desalination and Water Treatment 2013, 51, 899–907.
50 (5) Iler R. K. The chemistry of silica: Solubility, polymerization, colloid and surface properties,
51 and biochemistry; Wiley: New York, 1979.
52 (6) Weitz, E., Franck, H.-G., Giller, M. Untersuchungen an Kieselsäuren. Z. Anorg. Allg. Chem
53
1964, 331, 249–255.
54
55 (7) E. Richardson, J. A. Waddams. Use of silico-molybdate reaction to investigate the
56 polymerization of low molecular weight silicic acids in dilute solutions. Res. Corresp. 1954,
57 S43.
58
59
60
38
1
2
3
4
5
6 (8) Greenberg, S. A., Sinclair, D. The Polymerization of Silicic Acid. J. Phys. Chem. 1955, 59,
7
435–440.
8
9 (9) Marsh A. R., III, Klein G., Vermeulen T. Polymerization kinetics and equilibria of silicic
10 acid in aqueous systems, 1975.
11 (10) Merrill, R. C., Spencer, R. W. The Gelation of Sodium Silicate. Effect of Sulfuric Acid,
12 Hydrochloric Acid, Ammonium Sulfate, and Sodium Aluminate. J. Phys. Chem. 1950, 54,
13 806–812.
14 (11) Baumann H. Polymerisation und Depolymerisation der Kieselsäure unter verschiedenen
15
16
Bedingungen. Kolloid Z. 1959, 28–35.
17 (12) Pelster, S. A., Schrader, W., Schuth, F. Monitoring temporal evolution of silicate species
18 during hydrolysis and condensation of silicates using mass spectrometry. Journal of the
19 American Chemical Society 2006, 128, 4310–4317.
20 (13) Pelster, S. A., Schuth, F., Schrader, W. Detailed study on the use of electrospray mass
21 spectrometry to investigate speciation in concentrated silicate solutions. Analytical
22
chemistry 2007, 79, 6005–6012.
23
24 (14) Pelster, S. A., Weimann, B., Schaack, B. B., Schrader, W., Schuth, F. Dynamics of silicate
25 species in solution studied by mass spectrometry with isotopically labeled compounds.
26 Angewandte Chemie (International ed. in English) 2007, 46, 6674–6677.
27 (15) Icopini, G. A., Brantley, S. L., Heaney, P. J. Kinetics of silica oligomerization and
28 nanocolloid formation as a function of pH and ionic strength at 25°C. Geochim.
29 Cosmochim. Acta 2005, 69, 293–303.
30
(16) Weres, O., Yee, A., Tsao, L. Kinetics of silica polymerization. J. Colloid Interface Sci.
31
32 1981, 84, 379–402.
33 (17) Noguera, C., Fritz, B., Clement, A. Precipitation mechanism of amorphous silica
34 nanoparticles: a simulation approach. Journal of colloid and interface science 2015, 448,
35 553–563.
36 (18) Rothbaum, H., Rohde, A. Kinetics of silica polymerization and deposition from dilute
37 solutions between 5 and 180°C. J. Colloid Interface Sci. 1979, 71, 533–559.
38
39
(19) Tobler, D. J., Shaw, S., Benning, L. G. Quantification of initial steps of nucleation and
40 growth of silica nanoparticles: An in-situ SAXS and DLS study. Geochim. Cosmochim.
41 Acta 2009, 73, 5377–5393.
42 (20) Tobler, D. J., Benning, L. G. In situ and time resolved nucleation and growth of silica
43 nanoparticles forming under simulated geothermal conditions. Geochimica et
44 Cosmochimica Acta 2013, 114, 156–168.
45
(21) Kley, M., Kempter, A., Boyko, V., Huber, K. Mechanistic Studies of Silica Polymerization
46
47 from Supersaturated Aqueous Solutions by Means of Time-Resolved Light Scattering.
48 Langmuir 2014, 30, 12664–12674.
49 (22) Adrew D. Eaton, Leonore S. Clesceri, Arnold E. Greenberg. Standard methods for the
50 examination of water and wastewater, 20th ed.; American Public Health Association:
51 Washington, D.C, 1998.
52 (23) Zimm, B. H. Apparatus and Methods for Measurement and Interpretation of the Angular
53
Variation of Light Scattering; Preliminary Results on Polystyrene Solutions. J. Chem. Phys.
54
55 1948, 16, 1099–1116.
56 (24) Hermans, J. J., Levinson, S. Some Geometrical Factors in Light-Scattering Apparatus. J.
57 Opt. Soc. Am. 1951, 41, 460.
58
59
60
39
1
2
3
4
5
6 (25) Koppel, D. E. Analysis off Macromoleecular Polyddispersity inn Intensity C Correlation
7
Speectroscopy: The Method of Cumullants. J. Cheem. Phys. 19972, 57, 48114–4820.
8
9 (26) Burrchard, W., Schmidt, M M., Stockmaayer, W. H. Information
I n on Polydisspersity and
10 Braanching from m Combinedd Quasi-Elaastic and Inttergrated Scattering. Maacromolecuules 1980,
11 13, 1265–12722.
12 (27) Burrchard W., Ed.
E Static and dynamicc light scatteering from bbranched poolymers andd
13 opolymers; S
biop Springer-Veerlag: Berlinn, 1983.
14 (28) Schhmidt, M., B Burchard, W.W Translatioonal diffusioon and hydrrodynamic radius
r of unpperturbed
15
16
flexxible chainss. Macromollecules 19811, 14, 210–2211.
17 (29) Schhmidt, M. C Combined inntegrated and dynamic llight scatterring by polyy(γ-benzyl glutamate)
18 in a helocogennic solvent. M Macromoleccules 1984, 17, 553–5660.
19 (30) Donkai, N., Inagaki, H., Kajiwara,
K K.., Urakawa, H., Schmiddt, M. Dilutee solution prroperties
20 of iimogolite. M
Makromol. C Chem. 19855, 186, 2623––2638.
21 (31) Gennnes, P. G. de, Witten, T. A. Scalinng Conceptss in Polymeer Physics. Physics
P Todday 1980,
22
33, 51–54.
23
24 (32) Liuu, J., Rieger,, J., Huber, K. Analysiss of the Nuccleation and Growth of Amorphouss CaCO 3
25 by Means of T Time-Resolvved Static Liight Scatteriing. Langmuir 2008, 244, 8262–82771.
26 (33) Liuu, J., Pancera, S., Boykoo, V., Shuklla, A., Narayyanan, T., H
Huber, K. Evvaluation off the
27 Parrticle Growtth of Amorpphous Calciuum Carbonaate in Waterr by Means of the Porodd
28 Invvariant from
m SAXS. Lanngmuir 20100, 26, 174055–17412.
29 (34) Schhmidt, M., S Stockmayer,, W. H. Quaasi-elastic liight scatterinng by semifflexible chaiins.
30
Maacromoleculles 1984, 177, 509–514.
31
32 (35) Debbye, P. Mollecular-weigght Determination by L Light Scatterring. J. Physs. Chem. 1947, 51,
33 18––32.
34 (36) Volmer, M., W Weber, A. K Keimbildung in übersättiigten Gebildden. Z. Physs. Chem. 1926, 277–
35 3011.
36 (37) Beccker, R., Dööring, W. Kiinetische Beehandlung der d Keimbildung in übeersättigten Dämpfen.
D
37 Annn. Phys. 19335, 719–7522.
38
39
(38) Robbb, D. T., PPrivman, V. Model of N Nanocrystal F Formation iin Solution bby Burst Nuucleation
40 andd Diffusionaal Growth. Langmuir
L 20008, 24, 26––35.
41 (39) Vekkilov, P. G. Nucleationn. Cryst. Groowth Des. 2010, 10, 50007–5019.
42
43
44
45 For Tab
ble of Conttents Only
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
40

Cinetica de Reaccion Termo 3

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cinetica de Reaccion Termo 3

Uploaded by

Copyright:

Available Formats

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street

42 final Mw(cpart) /1E4 g mol-1

You might also like