Introduction

Microstructures in the Formation of Chemical Gardens

C. Ignacio Sainz-Diaz, Bruno Escribano, and Julyan Cartwright Laboratorio de Estudios Cristalográficos, Instituto Andaluz de Ciencias de la Tierra, CSIC-Universidad de Granada, Av. Fuentenueva s/n, Granada, 18002, Spain

ABSTRACT

Chemical gardens are biomimetic plantlike growths formed by a mixture of salts which precipitate by a combination of convection forced by osmosis, free convection and chemical reactions. Chemical gardens may be implicated in other phenomena of industrial interest which involve precipitation across a colloidal gel membrane that separates two different aqueous solutions, for example, in cement technology and metal corrosion processes. However, the variation in chemical composition, morphology and mechanical properties of the different surfaces of these formations is not well known. Several salts in different concentrations and conditions have been explored under terrestrial gravity and microgravity. The chemical garden structures have been characterised by morphology analysis, scanning electron microscopy, chemical analysis and X-ray diffraction, correlating these data with the biomimetic growth and the physical-chemical nanoprocesses involved in it. This approach may also be useful for the analysis of biomaterials with interesting biomechanical properties.

INTRODUCTION

The morphology of a solid at a nanoscopic scale during its growth may be affected by physical phenomena of fluid dynamics and equilibria between processes of adsorption, desorption and diffusion. An interesting phenomenon in this respect is the formation of biomimetic growths called chemical gardens. Chemical gardens are curious structures in the form of plants formed by a mixture of salts which precipitate by a combination of convection forced by osmosis, free convection and chemical reactions [i]. On adding a crystal of a soluble salt to a solution of sodium silicate a reaction is provoked producing a hydrated silicate of the cation of that salt which is deposited as a colloidal gel around the crystal. The gel acts as a semipermeable membrane across which water and hydroxyl ions pass under osmotic pressure. The crystal continues to disolve and the gel membrane expands by osmotic pressure until it breaks producing a jet of fluid of the solution around the crystal; on entering the surounding medium its solubility changes and salt precipitates where the jets appear. Thus at each point of membrane rupture tubular growths are formed which can attain lengths of many centimeters, showing biomimetic forms and receiving the name of gardens. Beyond their purely scientific fascination as spectacular examples of pattern formation, chemical gardens may be implicated in other phenomena of industrial interest which involve precipitation across a colloidal gel membrane separating two different aqueous solutions. For example, in cement technology it is important to understand the hydration of Portland cement in which are found interpenetrated tubular nanofilaments of hydrated calcium silicates following a type of inverse silicate garden [ii]. Another example is the phenomenon of metal corrosion in which is formed a colloid producing nanotubes of metal oxide on the corroding surface following a similar process to chemical gardens [iii]. Recently it has also been speculated that the membranes of chemical gardens produced in submarine vents may be an ideal site for the origin of life on Earth [iv]. However, the chemical composition of the different surfaces of these formations is not well known and has only been studied in a few cases and only on Earth [v,vi]. An experiment with a chemical garden was performed in space, obtaining a morphology different to that on Earth [vii], with structures typical in mechanisms of Laplacian growth, such as viscous fingers, showing the role of density- driven convection as one as the mechanisms implicated in the reaction.

Chemical gardens are formed by adding crystals of a soluble metal salt to aqueous solutions containing anions such as silicates, phosphates or carbonates. The chemical reactions involved produce a large variety of biomimetic structures resembling plants, trees, mushrooms, worms, seashells, and so on. This similarity to living organisms provoked a great deal of interest among investigators searching for the origin of life at the beginning of the 20th century [viii, ix]. However, only a few studies have been reported related with the physics and chemistry of this phenomenon [x, xi, xii, xiii] and chemical gardens still remain incompletely understood and their applications to industrial processes such as hydration or corrosion justify further research. A deeper knowledge of the physics and chemistry of this phenomenon should enable us to control the morphology and the mechanical properties of these solids and eventually to design new materials.

Mechanism of growth

When the metal salt is immersed in the sodium silicate solution it begins to dissolve and a coating of metal silicate envelops the whole particle. This coating acts as a semipermeable membrane and, since the ion concentration is higher on the inside, the osmotic pressure forces the membrane to dilate and water enters from the outside into the envelope through this membrane and the metal salt inside continues to dissolve. At some point the membrane bursts and a jet of metal salt solution is ejected under pressure through the rupture. After a few seconds, the injection pressure is reduced and the liquid flow no longer behaves like a jet but like a buoyant plume instead; the ejected solution is less dense than the sodium silicate solution, so it tends to rise by buoyancy. The flow thus changes from forced convection driven by the osmotic pressure at the beginning of the growth to free convection dependent on buoyancy at the end. This is why the tubes can grow initially at random angles but are soon redirected towards the vertical.

When the metal salt solution in the jet or plume comes in contact with the sodium silicate solution, metal silicate starts to precipitate around the fluid flow forming the walls of the new tube; the pH of the liquid in the jet is lower and the solubility of the silicate is lower at this pH, so it precipitates around the jet. The tube grows from the base upwards, starting where the membrane has ruptured, and forming a rather cylindrical conduct. The osmotic pressure difference acts as a pump, ensuring that water continues entering through the membrane and the metal solution keeps flowing through the growing tube until the original crystal is depleted. During this process hydroxide ions will enter the tube through its walls. This reduces the pH of the silicate solution and makes silica precipitate over the outer surface of the tube. At the same time metal hydroxide precipitates on the inner wall as the metal ions react with the hydroxide ions that came through the wall (figure 1).

Figure 1. Formation mechanism of a chemical garden, from Cartwright et al. [5].

Materials and Methods

Crystals of CaCl2, and MnCl2 at analytical purity and a commercial saturated solution of sodium silicate were used in our experiments. The particle size was between 1 and 5 mm radius. The sodium silicate solutions were prepared from a concentrated solution composed of 27% SiO2 and 15% NaOH, diluted with bidistilled water to several concentrations between 6.25 M and 0.625 M. The growing process was followed for three months. After that time the grown tubes were removed from the solution. Some of them were washed with distilled water to remove excess silica that adhered to the exterior surface. They were then dried in air at 25ºC.

The micrographs of the samples were obtained using a FEI Quanta 400 Environmental Scanning Electron Microscope (ESEM) at high vacuum and room temperature.

Chemical analysis of the tubes was performed in-situ in the microscope using EDAX analysis. Powder X-ray Diffraction (XRD) analyses were performed in a Bruker diffractometer grinding the samples carefully. In order to explore the crystallinity of the external surface of some tubes single-crystal XRD analyses were performed in a Bruker diffractometer with area detector. The analysis of these diffractograms was performed with the XRD2DScan code [xiv]. The identification of crystallographic phases in the diffractograms was performed with the Xpowder code [xv].

Results

Calcium chloride

CaCl2 particles were submerged in sodium silicate solutions at concentrations from 0.3 M to 3M. With the 3 M solution, the formation of a large osmotic membrane was observed. The walls of this membrane had a high flexibility. The inner solution dilated the membrane under osmotic pressure but it did not break owing to its plasticity and only the formation of some broad fingers was observed. Only one tube was formed, catalysed by the formation of an air bubble (figure 2).

Figure 2. Chemical garden formation from calcium with sodium silicate solution at 3M in solution.

With lower concentrations of silicate several thinner tubes were formed starting homogeneously all over the osmotic membrane surrounding the initial solid particle. However, the osmotic membrane became less unstable when the concentration of silicate solution decreased, being completely flat at 0.3 M.

ESEM (figure 3) combined with EDAX analysis showed different morphologies and compositions on the inside and outside surfaces of the tubes, which varied also along the length of the tube. In figure 4a we can see the presence of a small crystal of

CaCl2 attached to the internal surface of the wall in the lower part of the tube. EDAX analysis of this area shows a mixture of CaCl2 from these crystals and calcium silicate from the bottom of the wall. The morphology of these crystals and their roughness suggest us that they do not come from a recrystallization of CaCl2 but from a disintegration of the initial seed without being dissolved (figure 4b).

Figure 3. Section of one of the thick tubes grown at 3M.

a b

Figure 4. ESEM picture of the internal surface of the lower part of the tube (a) along with EDAX chemical analysis (b). The white grains are CaCl2 that did not dissolve, while the back wall is made mainly of calcium silicate.

If we focus on the upper end of the tube, we find a different composition. In figure 5 we can observe a close-up of the inner surface of the tube. We can distinguish small crystals adhering to the inner surface. The high resolution of the EDAX analysis showed that they were made of NaCl. X-ray powder diffraction confirmed the EDAX analysis (figure 5b).

a b

Figure 5. ESEM image (a) and XRD diffractogram (b) of the internal surface of the upper end of the tube. We can distinguish small crystals of NaCl adhering to the inner side of the wall.

Manganese Chloride

Particles of MnCl2 were immersed in sodium silicate at concentrations ranging from 6.25 M to 0.625 M. The reactivity was much higher than in calcium samples. Very different morphologies were observed, from a single thick tube at the highest concentrations to multiple very thin tubes that grew almost instantly at lower concentrations. A special case is found at middling concentrations (2 M), where two or three thick tubes grew changing their direction several times, forming a twisted morphology (see figure 6).

a b

Figure 6. Chemical gardens from MnCl2 particles in sodium silicate at 1 M (a) and 2 M (b).

ESEM images show that the twisted tubes grown at 2 M are a composite of several thinner tubes, each with a diameter of ~50 μm (figure 7a). The thickness of these tubes is ten times greater than with CaCl2. The ESEM pictures taken with a back- scattering electron detector show a clear separation of phases with different compositions (figure 7b). The EDAX and XRD analyses indicate that the external surface is formed mainly by manganese silicate and the internal surface is hydrated manganese oxide.

a b

Figure 7. ESEM micrographs of a section of a multiple tube (a) and a thin tube (back scattering electron detector, b) of the MnCl2 chemical garden grown in sodium silicate at 2M.

A further analysis of a complete tube of MnCl2 grown at 2 M was performed with single-crystal XRD at the middle and the edge of this tube (figure 8). The analysis of these diffractograms was performed with the XRD2DScan code, and confirmed the results obtained with powder diffraction: the only crystalline structure present was hydrated manganese oxide and no difference in chemical composition was observed between the middle and the edge of the tube, only slight texture differences where some crystals of larger size were observed at the edge.

Figure 8. A tube from a chemical garden grown from MnCl2 with 2M silicate solution. The diffraction rings correspond mainly to hydrated manganese oxide (MnO∙H2O).

Conclusions

The formation of chemical gardens is due to a combination of forced convection from osmosis and free convection from buoyancy forces. The ratio between them is highly dependent on the chemical nature of the crystal seed and the concentration of silicate solution. We found thicker tubes growing slower at higher concentrations, whereas at lower concentrations multiple thinner tubes were grown.

The mechanical properties of the osmotic membrane and plumes depend on the chemical composition. The semipermeable membrane from CaCl2 is thinner and more flexible than those from Mn2+ and other salts.

The viscosity of the solution affects the shape and the rate of formation of chemical gardens. Higher viscosity -higher silicate concentration- produces a slower reaction, but very dilute silicate solutions also yield low reactivity. In general, MnCl2 is more reactive than CaCl2.

The combination of ESEM microscopy, EDAX analysis and XRD is found to be very effective for characterizing heterogeneous materials at the nanoscale.

REFERENCES

i T. Hazlehurst, J. Chem. Edu. 18, 286–289, (1941) ii D. D. Double. Phil. Trans. Roy. Soc. Lond. A, 310:55–63, 1983 iii M. G. Fontana. Corrosion Engineering. McGraw-Hill, 3rd ed, 1986 iv M. J. Russell, A. J. Hall. J. Geol. Soc. Lond., 154:377–402, 1997 v J. H. E. Cartwright, J. M. García-Ruiz, M. L. Novella, F. Otálora. J. Colloid Interface Sci. 256, 351-359, 2002. vi C. Collins, G. Mann, E. Hoppe, T. Duggal, T. L. Barr, J. Klinowski. Phys. Chem. Chem. Phys.,1:3685–3687, (1999). vii D. E. H. Jones, U. Walter (1997) J. Colloid Interface Sci. 203, 286–293. viii S. Leduc. The Mechanism of Life. Rebman, London, 1911. ix A. L. Herrera. Nociones de Biología, Mexico, 1903. x R. D. Coatman, N. L. Thomas, and D. D. Double. Studies of the growth of “silicate gardens” and related phenomena. J. Mater. Sci., 15:2017–2026, 1980. xi C. Collins,W. Zhou, A. L. Mackay, and J. Klinowski. The ‘silica garden’: A hierarchical nanostructure. Chem. Phys.Lett., 286:88–92, 1998. xii C. Collins, W. Zhou, and J. Klinowski. A unique structure of Cu2(OH)3_NH3 crystals in the ‘silica garden’ and their degradation under electron beam irradiation. Chem. Phys. Lett., 306:145–148, 1999. xiii M. Zeleny, J. Klir, and K. D. Hufford. Precipitation membranes, osmotic growths and synthetic biology. In C. Langton, editor, Artificial Life, SFI Studies in the Sciences of Complexity, pages 125–139. Addison-Wesley, 1988. xiv A. B. Rodriguez-Navarro. XRD2DScan a new software for polycrystalline materials characterization using two-dimensional X-ray diffraction. Journal of Applied Crystallography 39, 905- 909 (2006). xv J.D. Martin http://www.XPowder.com/index.html (2004).