ARTICLE pubs.acs.org/Langmuir
Interactions of Amines with Silicon Species in Undersaturated Solutions Leads to Dissolution and/or Precipitation of Silica Siddharth V. Patwardhan,†,§ Graham E. Tilburey,‡,§ and Carole C. Perry* Biomolecular and Materials Interface Research Group, School of Science and Technology, Nottingham Trent University, Clifton Lane, Nottingham NG11 8NS, U.K.
bS Supporting Information ABSTRACT: The biogeochemical silicon cycle is the focus for many researchers studying the dissolution of silicon species from quartz, amorphous, and biogenic silica. Furthermore, the precipitation of biogenic silica by diatoms, radiolarian, sponges, and plants is also a popular focus for research. The ornate silica structures created by these species has attracted interest from biomaterial scientists and biochemists who have studied mineral formation in an attempt to understand how biogenic silica is formed, often in the presence of proteins and long chain polyamines. This article is at the interface of these seemingly distinct research areas. Here we investigate the effect of a range of amines in globally undersaturated silicon environments. Results are presented on the effect of amine-containing molecules on the formation of silica from undersaturated solutions of orthosilicic acid and globally undersaturated silicon environments. We sought to address two questions: can silica be precipitated/harvested from undersaturated solutions, and can we identify the silicon species that are most active in silica formation? We demonstrate that none of the bioinspired additives investigated here (e.g., poly(allylamine hydrochloride), pentaethylenehexamine, and propylamines) have any influence on orthosilicic acid at undersaturated concentrations. However, under globally undersaturated silicon concentrations, small molecules and polymers containing amine groups were able to interact with oligomers of silicic acid to either generate aggregated materials that can be isolated from solution or increase rates of oligomer dissolution back to orthosilicic acid. Additional outcomes of this study include an extended understanding of how polyelectrolytes and small molecules can promote and/or inhibit silica dissolution and a new method to explore how (bio)organic molecules interact with a forming mineral phase.
1. INTRODUCTION The aqueous chemistry of silica and orthosilicic acid, including dissolution and precipitation in particular, is of interest to a wide scientific community due to its impact on the biological, geological, and chemical cycling of minerals.1 It is important to understand the stability of various silicon species under a range of conditions including silicon concentration, pH, and the presence of inorganic ions and organic compounds. Orthosilicic acid, the soluble form of silica and silicates, has been reported to occur at 515 ppm in seawater and 535 ppm in fresh water2 (for silicic acid, 100 ppm ≈ 1 mM). Together with the increased pressure and subambient temperatures that can be found in natural aqueous environments, these conditions are also highly undersaturated with respect to the condensation of orthosilicic acid, which generally occurs at concentrations greater than ca. 2 mM.2a The dissolution of amorphous silica, quartz, biogenic silica, and silicate rocks have been studied extensively to understand the transport of silica and minerals to aqueous reservoirs as part of the silicon biogeochemical cycle.2b The stability of the mineral phase is dependent on the molecular and physical characteristics of the material, including crystallinity, surface chemistry, surface area, and particle size.2a Dissolution of silica in aqueous solution is a depolymerization reaction that is typically catalyzed by hydroxide ions2a in an industrial or laboratory environment. In the aqueous reservoirs r 2011 American Chemical Society
of the earth, biogenic silica dissolution is catalyzed by alkali and alkaline earth cations,3 bacterial assemblages,4 and organic acids.5 In contrast, biogenic silica is stabilized by an organic coating present on the diatom frustule, and aluminum that has been found in biogenic silica is reported to inhibit dissolution.6 Despite a significant amount of research being performed in this area, it is still unclear what are the factors that control biogenic silica dissolution.5c,7 As well as the influence of silica dissolution on the silicon biogeochemical cycle, silica dissolution has also been studied to prevent or slow down mineral deposition (scaling) in engineering applications.8 Industrial processes that utilize membranes, heat exchangers, or pipelines are continually hampered by the detrimental effects of scaling.8,9 A number of additives such as dendrimers and polyelectrolytes have been shown to inhibit silica formation and enhance silica dissolution. For example, starburst polyaminoamide dendrimers have been shown to inhibit silica formation in 500 ppm solutions of sodium silicate at pH 7.9a,10 Importantly, due to the high cost of dendrimers, these additives are not commercially viable for industrial processes, and polyelectrolytes with similar functionality such as polyethylenimine (PEI) and polyvinylimidazole (PVI) have also been shown to Received: October 25, 2011 Published: November 15, 2011 15135
dx.doi.org/10.1021/la204180r | Langmuir 2011, 27, 15135–15145
Langmuir
ARTICLE
inhibit silica formation in 500 ppm solutions of sodium silicate.9a,11 In trying to understand the biogeochemical silicon cycle, researchers have largely studied the dissolution of silica from two distinct sources; the first source is amorphous biosilica formed primarily by diatoms, radiolarian, and sponges in aquatic environments, and the second source is abiotic silica, principally quartz, silicates, and amorphous silica. Typically, silica precipitation studies using bioinspired molecules as additives have been performed on solutions containing orthosilicic acid at supersaturated concentrations (typically 30 or 100 mM).12 However, no reports have been published studying the roles of additives in undersaturated orthosilicic acid solutions, even though the silicon concentration in natural waters is undersaturated (i.e., 98%) to yield silicic acid.15b The organic amines, when required, were dissolved in the acid containing solution and added as the hydrochloride to the 1 mM complex solution. The total starting volume of the system was kept constant at 1 L ( 1 cm3. The additives were studied using a molar ratio of silicon to nitrogen (denoted as Si:N) of 1:1 or 1:6, and the concentration of free orthosilicic acid was monitored over a period of 30 h using a modification to the molybdenum blue method reported previously15b and is briefly described below. The reason behind using these Si:N ratios was that it has been observed previously that the level of amine when varied from Si:N = 1 showed an effect on aggregation, gelation, and/or precipitation of silica.15c At predetermined time intervals, 200 μL portions of the reaction mixture were taken and placed into 16.5 mL of molybdic acid solution (no more than 1% of the total sample volume was removed from any solution for sampling purposes). The molybdic acid solution was left for 15 min to allow any dimers to dissociate to monomers and react with the molybdic acid to form a yellow silicomolybdic acid complex. The silicomolybdic acid complex was reduced using metol (4-methylaminophenol sulfate) to a blue silicomolybdous acid complex and left to stand at room temperature.15b After ∼24 h, an aliquot of this blue complex was transferred to a polymethacrylate cell, and the absorbance was measured at 810 nm using a Unicam UV/vis UV2 spectrometer. Calibration data obtained from standard silica solutions were used to calculate the concentration of orthosilicic acid in solution.
2. EXPERIMENTAL SECTION
orthosilicic acid solutions (i.e., no higher-mers are present). In order to prepare 1 mM solutions of “silicon”, which had a mixture of a range of silicon species, the following method was used. 30 mM solutions of orthosilicic acid at pH 6.8 ( 0.05 were prepared using dipotassium tris(1,2-benzene-diolato-O,O0 )silicate as described previously.15b The solution was allowed to condense for a predetermined time (3160 min) and then diluted to 1 mM (1 L) using distilled and deionized water. Experiments with or without the additives were conducted on these solutions. The experimental procedure involved an initial dilution of the 30 mM solution to 2 mM followed by an immediate dilution to 1 mM with a solution containing the chosen additive at the required concentration. The Si:N ratio studied was 1:6. Complete dilution from 30 to 1 mM took 40 s. To ensure homogeneity, the samples were stirred for the first 5 min following dilution. During this time the pH of the solution was measured. The batch was rejected if the pH was not pH 6.8 ( 0.1. The pH was measured periodically over the dissolution period, and while the pH is expected to rise for all experiments, the pH did not exceed pH 7 ( 0.2 during the dissolution period. The molybdenum blue method was used to monitor the dissolution of the silica species in solution as described in section 2.2. Initial rates of dissolution were assessed manually by fitting as many data points, including the first data point, as possible to a straight line. Analysis of the initial rates of dissolution for multiple experimental runs was performed in order to verify the reproducibility of the analysis. The data presented in Figure S1 show representative reproducibility with the initial slope analysis using the Spearman correlation coefficient F > 0.984, indicating that the linear fits were in excellent agreement. These rates were used to quantitatively compare and contrast the effects of additives. Similarly, a computational approach was used to model the silicon concentration versus dissolution time data. The following doubleexponential growth equation was used to fit the data with OriginLab:
2.1. Materials. Dipotassium tris(1,2-benzenediolato-O,O0 )silicate (97%) purchased from Aldrich and commercial silica (Gasil 23D) obtained from Ineos were used as the silica sources. The former was purified by dissolving in methanol followed by filtration to remove an insoluble fraction with removal of methanol by evaporation. The purity of the silicon catecholate complex was confirmed by 1H NMR and the molybdosilicate assay to be >99%. The additives used were poly-L-lysine (PLL, MW 22 100 g mol1), poly(allylamine hydrochloride) (PAH, MW 15 000 g mol1), poly(ethylenimine) (PEI, MW 25 000 g mol1), diaminoethane, 97% (DAE), diethylenetriamine, 99% (DETA), triethylenetetramine, 97% (TETA), tetraethylenepentamine, 98%, pentaethylenehexamine (PEHA), and 1,10-diaminodecane, 97% (1,10 DA), all of which were purchased from Sigma-Aldrich and used with no further purification. In addition, dipropylenetriamine (N3) and tetrapropylenepentamine (N5) were also used as additives. N3 and N5 were synthesized in house; the purity was confirmed to be above 99% by NMR.13,14 The additives selected have been shown to influence the formation of silica under bioinspired conditions and have been termed bioinspired/biomimetic additives previously.6a,12,14,15 These molecules have been studied because their chemical structures relate in some way to silaffins and LCPAs extracted from diatoms16 and sponges.17 Experiments to study the stability (dissolution and precipitation) of orthosilicic acid, poly(silicic acid)s, and silica species (together termed “silica species”) were undertaken using three methods (A, B, and C) as detailed below at 20 °C. 2.2. Method A: Studies Using Freshly Prepared Undersaturated Solutions. 1 mM solutions of orthosilicic acid were freshly prepared by dissolving 0.466 g (1 mM) of dipotassium tris(1,2-benzenediolato-O,O0 )silicate in 950 mL of dd (distilled and deionized) water. A predetermined amount of 2 M HCl was then placed into a separate plastic vial and the volume made up to 50 mL with deionized and distilled water. Addition of the acid solution to the silicate solution dissociated the complex to yield 1 L of a 1 mM solution of molybdate
2.3. Method B: A Study of Precondensed Undersaturated Solutions. The method described in the section 2.2 produces pure
½Si ¼ ½Si0 þ A1 ð1 ex=t1 Þ þ A2 ð1 ex=t2 Þ
ð1Þ
where [Si]0 is the concentration of silicon at the midpoint of the curve; A and t are amplitude and the width (duration) of a given step, respectively. 15136
dx.doi.org/10.1021/la204180r |Langmuir 2011, 27, 15135–15145
Langmuir
ARTICLE
At the midpoint, the slope of the tangent to the curve is A1/t1 + A2/t2; the equilibrium concentration (=1 mM) is the sum of two amplitudes (A1 and A2). This equation represents a two-stage exponential growth mechanism, which was found to model the process extremely well with R2 > 0.99 and reduced χ2 < 0.0001 (see Figure S5). In terms of the physical meaning of the model, in simple words, it represents a reaction of the type where the initial large silicate cluster of n silicon atoms dissociates into a medium size cluster with m silicon atoms which eventually converts into a monomeric or dimeric silicate that was detected by silicomolybdous blue assay. Sin f Sim f Si12
ð2Þ
where n . m . 1 are the number of silicon atoms present in a given silicate cluster in solution. The first derivative of this model at t = 0 can be used to calculate the initial rates of dissolution for a given system. Their trend compared extremely well with the experimental trend (Figure S6). Although the direct measurement of m-sized cluster is not possible, indirect data, shown in section 3.2, validate the feasibility of a two-step dissolution process and is discussed below. Silica precipitates, where formed, were isolated from 1 L batch by centrifugation at 3000 rpm using a Sorvall RT7 plus centrifuge and briefly washed 3 times with dd water and lyophilized prior to further analysis. The precipitation of silica could not be monitored by photon correlation spectroscopy due to an insufficient number of scattering species in solution. Instead, an alternative semiquantitative approach was undertaken, whereby the total amount of silicon remaining in solution (i.e., nonprecipitated silica) was measured. This concentration, when subtracted from the total silicon concentration (always = 1 mM), gives the concentration of silicon in the precipitatable material. Upon mixing of an additive solution to a diluted precondensed silica solution, 2 mL portions of the resulting solution were pipetted into 2 mL Eppendorf tubes and left to stand. At predetermined reaction times, samples were centrifuged for 1 min at 15 000 rpm in a Sanyo Micro Centaur in order to separate any particles aggregated by additives from the solution. The precipitated material was not measured. A 1 mL portion of the supernatant was taken and immediately mixed with 1 mL of 2 M NaOH and placed in a water bath at 80 °C for 60 min to ensure the remaining nonprecipitated silica species were converted to orthosilicic acid. The concentration of the silica that remained in solution was determined by taking a 200 μL aliquot of the digest and performing the molybdenum blue assay as previously described in section 2.2.
2.4. Method C: Dissolution of Fully Condensed Commercial Silica. A 1 mM suspension of commercial silica (pyrolysis silica
from Ineos; surface area 267 ( 9 m2 g1; primary particle size
