Modifying the thickness, pore size, and composition of diatom frustule in Craspedostauros sp. with Al3+ ions

Diatoms are unicellular photosynthetic algae that produce a silica exoskeleton (frustule) which exposes a highly ordered nano to micro scale morphology. In recent years there has been a growing interest in modifying diatom frustules for technological applications. This is achieved by adding non-essential metals to the growth medium of diatoms which in turn modifies morphology, composition, and resulting properties of the frustule. Here, we investigate the frustule formation in diatom Craspedostauros sp., including changes to overall morphology, silica thickness, and composition, in the presence of Al3+ ions at different concentrations. Our results show that in the presence of Al3+ the total silica uptake from the growth medium increases, although a decrease in the growth rate is observed. This leads to a higher inorganic content per diatom resulting in a decreased pore diameter and a thicker frustule as evidenced by electron microscopy. Furthermore, 27Al solid-state NMR, FIB-SEM, and EDS results confirm that Al3+ becomes incorporated into the frustule during the silicification process, thus, improving hydrolysis resistance. This approach may be extended to a broad range of elements and diatom species towards the scalable production of silica materials with tunable hierarchical morphology and chemical composition.


Mean diameter of areolae and small pores measurements
The mean diameter of areola was measured for 15 valves per culture in SEM micrographs using an in house MATLAB script 1 Figure S1 exhibits the selected axes for each areola. The average of the two-axes was used to determine areola diameter and error bars represent the standard deviations. Also, the mean diameter of small pore was determined via an in-house MATLAB script. Supplementary Figure S2 shows the procedure for determining the mean dimeter of small pore via the Matlab program. Figure S1. Selected axes for measuring the diameter of the areolae. Figure S2. Image analysis process for measuring the mean diameters of small pores.

Thickness mapping
The thickness of the valves was measured from TEM images in the following manner: The contrast in the TEM images can be approximated by mass-thickness contrast considerations. Diatoms frustules, are amorphous materials, loaded on the continuous carbon grid and were imaged at a low magnification (800×). To map the thickness of a single diatom, one flat field (FFD) TEM image only containing the unscattered incident electron flux 0 and one TEM image containing the electron flux transmitted through the sample were acquired. The thickness of diatoms D can be estimated (Equation (1)) based on Lambert-Beer Law: Where C is the thickness of the continuous carbon film in the TEM grid, and C and D are the elastic mean free path (EMFP, ) of the carbon and the diatom, respectively. EMFP calculations are based on Reimer's book 2 . All the thickness analysis and EMFP calculations were performed using in-house MATLAB scripts 3,4 . Important parameters used in the calculations were shown in Supplementary Table S2. Mapping the local thickness of the diatom was carried out using in-house MATLAB scripts. The detailed image analysis procedures are shown in Supplementary Figure S3.
Supplementary Figure S3. Image-processing procedures for single diatom and local area thickness mapping.

Internal volume of C.sp. and the frustule volume
The volume of the entire C.sp. was calculated using a simple assumption that C.sp. is a rectangular prism. V= L×W×H. Where V, L, W, and H are volume, length, width and height of the whole C.sp. (Supplementary Figure S4). Since, the length and width of the valves of the four cultures were comparable (SEM images) we used an average of C.sp. dimensions for determining the entire volume. The internal volume was calculated by subtracting the thickness of the frustule from the dimensions of the whole C.sp. Therefore, the volume of silica making up the frustule for each culture is the difference between the total volume and internal volume (Supplementary Table S3). It is worth noting that for this calculation, the thickness of the porous area was used as the thickness of the whole frustule.
Supplementary Figure S4. Overall structure of an intact C.sp. cell. Length, width, and height are shown.

FIB-SEM process of the embedded C.sp. cell
The embedded C.sp. cell was positioned at Eucentric height (10 mm) and tilted to 52° with the intention that the electron beam and the ion beam are focused at the coincidence point. Before the Serial Slice and View (SSV), a protective layer of platinum with dimensions of 20×10 ×1 μm was deposited on the surface of C.sp. using ion beam deposition at 30 keV and an ion beam current (IBC) of 0.3 nA. In order to remove material in front of the embedded cell, bulk milling was performed with dimensions 40× 20 × 10µm (IBC = 15 nA). Side trenches of 10× 20 × 5µm were created using an IBC of 7 nA. The fiducial marker was made with z ≥1µm (IBC = 3nA). Next the electron beam was focused on the cleaned block face and automated SSV operation was initiated using a milling IBC of 1 nA and a slice thickness of either 50 nm or 100 nm. Images of new revealed surface were taken in BSE mode.

The effect of Al 3+ on silica hydrolysis
Diatoms (C0 & C3) were stored in demineralized water at 90 °C for 6 day. The Si concentration was determined via atomic absorption spectrometry.
Supplementary Figure S10. Si concentration as a function of time for C0 and C3. Error bars represent standard deviation.