Biosilica as a New Stationary Phase in HILIC Mode

The aim of this work was to use the biosilica obtained from diatoms (microalgae) cultivated under laboratory conditions as a new stationary phase to fill the chromatographic column and test it in high-performance liquid chromatography. Biosilica is an inorganic polymer consisting of orthosilicate units formed by organisms such as diatoms or siliceous sponges. The results showed that the prepared columns were characterized by relatively high efficiency, comparable to a commercially available material HALO HILIC of dp = 2.7 μm. The retention of polar compounds under a high acetonitrile content was worse than that on HALO HILIC, but it was proportionally lower when the surface area of both materials was compared. As model test solutes, nucleosides and nucleobases have undergone testing. They were examined separately for retention, and attempts to separate test mixtures were successful.


Introduction
Hydrophilic interaction liquid chromatography (HILIC) is an alternative high-performance liquid chromatography (HPLC) mode for separating polar compounds.To mention for historical reasons, at first, it has been reported that HILIC is a variant of normal-phase liquid chromatography (NPLC), but it turned out that the explanation of the HILIC separation mechanism was more complicated than that in NPLC, so this approach was distinguished among LC modes and the acronym HILIC was first suggested by Alpert in 1990 [1].Compared to traditional normal-phase liquid chromatography and reverse-phase liquid chromatography (RPLC), HILIC provides several distinct advantages.For instance, it is appropriate to analyze substances in complicated systems that consistently elute close to the void in reserved-phase chromatography.Polar samples exhibit strong solubility in the aqueous mobile phase employed in HILIC, which outweighs the limitations of the frequently encountered poor solubility in NPLC.Since HILIC does not require expensive ion-pairing reagents, it is simple to couple it to mass spectrometry (MS), mainly when used in the electrospray ionization (ESI) mode.Gradient elution HILIC, in contrast to RPLC, starts with a low-polarity organic solvent and elutes polar analytes by increasing the polar aqueous concentration [2].For HILIC separations, any polar chromatographic surface may be utilized.Classical bare silica or silica gels that have been heavily modified with polar functional groups make up the majority of HILIC stationary phases [3].There is currently an increasing number of references in the literature reporting manufactured or homemade stationary phases that permit working in such environments [4].For example, amorphous silica is a common form of diatomaceous earth or diatomite in nature.It is a silica-rich substance made from dead diatoms, which are microscopic, unicellular photosynthesizing algae with amorphous hydrated silica frustules being their exoskeletons.In addition to silica, diatomaceous earth may also contain trace amounts of iron oxides, clay, and other minerals.Moreover, diatomaceous earth contains silica frustules of various forms and sizes from different species of diatoms.It is more frequently utilized in gas chromatography than it is as packing material for LC columns (e.g.Chromosorb stationary phase) [5].Cultivation diatoms of a selected species in laboratory conditions allow for obtaining uniform silica frustules (biosilica) with a 3D ordered structure and gives good prospects for the successful use of this material as a packing for chromatographic techniques.
Each diatom cell can be anywhere between 2 and 200 µm in size.Diatoms can be divided into two main groups based on the symmetry of their frustule morphology: centric diatoms, which are primarily discoid in shape and have ornamentation on their valves that is symmetrical in all directions, and pennate diatoms, which are boat-shaped and have bilaterally symmetrical valves.Diatom frustules have a distinctive double-sided shape like a Petri dish and are made up of two overlapping valves called thecae (top part: epitheca, lower part: hypotheca) [6][7][8][9].There are numerous silanol groups (Si-OH) and siloxane bridges (Si-O-Si) on the diatom frustule surface [7,9,10].Also, there are excellent opportunities to alter the physicochemical characteristics of diatom silica by applying modification techniques to create novel silica functional materials with retained distinctive 3D structures.According to their physicochemical characteristics, diatom frustules are a potential silica material for the chromatographic stationary phase and packing for liquid chromatographic columns.Therefore, this work aimed to prepare and use diatom frustules from cultivated diatoms as a packing material for capillary LC columns and test them in a HILIC mode.

Materials and Equipment
Water used as a mobile phase was purified using a Milli-Q system (Millipore, El Paso, TX, USA) in our laboratory.Acetonitrile, methanol, acetone, isopropanol, adenosine, adenine, cytosine, guanosine, and inosine were purchased from Sigma Aldrich (St. Louis, MO, USA).Fused silica capillaries with an internal diameter of 400 μm and outer diameter of 1/32" (794 μm) were purchased from Polymicro representative CM Scientific Ryefield (EU) Ltd. (Dublin, Ireland).The HALO HILIC 2.7 μm stationary phase (Advanced Materials Technology, Wilmington, DE, USA) was kindly gifted by MS Expert Ltd, Gdańsk, Poland.The 3D-structured diatom biosilica was obtained under laboratory conditions by the cultivation of the selected species of Pseudostaurosira trainorii.This diatom species was provided by the Culture Collection of Baltic Algae, Institute of Oceanography, University of Gdansk, Poland.Microalgae were cultivated using Erlenmeyer flasks (5 L) with a Guillard's growth medium that contained a silicon concentration of 7 mg/L under air aeration and a light regime of 12 h light/12 h darkness.After growing, diatom cells were separated by filtration using a vacuum pump and washed with distilled water.At first diatom biosilica (diatom exoskeletons) was isolated from dried diatom biomass by decomposition of the organic matter with hydrogen peroxide and hydrochloric acid followed by washing with water and dried in a vacuum oven at 70 °C.After this initial treatment, we decided to treat the biosilica with piranha solution (H 2 SO 4 /H 2 O 2 , 3:1) to eliminate some residual organic matter and increase the number of silanol groups on the surface.The piranha was applied for 40 min then the suspension was washed with distilled water, ethanol, and acetone, and the resulting material was dried in the vacuum oven at 70 °C.
The physicochemical properties of the studied biosilica obtained from Pseudostaurosira trainorii have already been characterized previously by Sprynskyy et al. [11,12].The chromatographic measurements were performed on a labmade capillary LC system consisting of Agilent 1260 pump with a 20 μL/min flow regulator, 10-port C-72MX Valco valve with a 200 nL capillary loop, Thermo Crystal-100 UV detector, and the PC with the Clarity chromatography station software (Data Apex Prague, Czech Republic).Micrographs of diatoms and silica gel particles were obtained using a scanning electron microscope (SEM/FIB Quanta 200 3D FEG).

Column Filling Procedure
A 20 cm section of the fused silica capillary (internal diameter of 400 μm and outer diameter of 1/32" (794 μm)) was consecutively flushed with 2 mL of each of the following solvents: dichloromethane, acetone, water, 1 M sodium hydroxide, water, and acetone, and dried in a stream of nitrogen.In each section, the outlet ceramic frit was synthesized using sodium water glass solution (750 μL) in formamide (120 μL), introduced by capillary forces to the 2.5 cm distance from the outlet.Then both ends of the capillary were plugged with pieces of silicone rubber and the whole was heated in the oven at 100 °C for 1 h.After that such prepared porous frit was flushed with water and acetone and dried in the stream of nitrogen.50 mg of biosilica was weighed and suspended in 470 μL of isopropanol.The mixture was then homogenized and degassed by sonication for 10 min.The slurry was transferred to a stainless-steel reservoir (100 × 2 mm) with the empty capillary column connected to its outlet.Such a prepared set-up was connected to a high-pressure air-driven pump (model DSF-122, Haskel, Burbank, USA).The slurry was pushed with methanol at a pressure of 20 MPa for 2 h.After that, the column was left to depressurize slowly.Finally, it was gently disconnected from the reservoir, and the outlet ceramic frit was cut to its final length of 0.5 cm.The same procedure was followed to prepare the HALO HILIC column.The columns were cut to the final length of 160 mm, unless otherwise stated.

SEM Imaging
Scanning electron microscopy gave a view of cleaned diatom frustules' (biosilica) morphological and structural features (biosilica) at different magnifications (Fig. 1).There are visible uniformity and well-preserved forms of the diatom frustules with walls of frustules perforated by periodic porous network.Diatom frustules can perform an analogous function as porous commercial silica particles.Their porous structure and the presence of polar silanol groups result in the retention of polar compounds.

Chromatographic Investigations
The adequately prepared diatom material (well-preserved frustules without debris) turned out to be a good material for a capillary chromatographic column.As this type of a silica material is new in chromatography, we decided to compare its chromatographic properties with a modern stationary phase that is HALO HILIC.As the amount of biosilica we had at our disposal was relatively small (it was connected with a laboratory scale of diatom cultivation tanks), we decided to prepare capillary columns (400 μm i.d.) and work in nano/microLC mode.The quality of the liquid chromatographic columns can be presented with such data as their efficiency at different flow rates (van Deemter plots) and permeability, as well as retention data for model test compounds.The manufacturer of HALO HILIC columns presents the example test chromatograms of acenaphthene (t 0 marker), adenosine, and cytosine separated with 90/10 ACN/ammonium formate (c = 0.1 M, pH 3.0).We compared two consecutive batches of biosilica with the HALO HILIC stationary phase.The separations of the above-mentioned test mixture are presented in Fig. 2.
The peak of acenaphthene and cytosine were used for the preparation of the van Deemter plots which are shown in Fig. 3, where HETP -height equivalent to a theoretical plate, A-eddy diffusion parameter, B u-longitudinal diffusion term, and C u-mass transfer term.The perme- ability (using acetonitrile as a test liquid) [13] of the two biosilica batches, namely batch 1 and batch 2 as well as HALO columns equaled K F = 8.02 × 10 -15 m 2 , 8.11 × 10 -15 m 2 and 9.65 × 10 -15 m 2 , respectively, and were calculated according to Darcy's law: where K F -permeability, F-flow rate, L-column length, η-dynamic viscosity of the fluid, ΔP-pressure drop, and r-column radius.
From Fig. 3 it is clearly seen that both biosilica columns were characterized by a relatively high efficiency, 14,500-15,000 plates per column, which was a little less than HALO HILIC plate count (N = 17,700).These values correspond to minimum plate heights of 10.9-11.2 and 9 μm, respectively.Relatively flat right branches of the plots stand for the excellent mass transfer in the described materials.The tested capillary columns (Biosilica 1, 2 and HALO HILIC) also showed linear relationships between the pressure drop and the flow rate (ΔP = f(F)) with R 2 values of, 0.9952, 0.997, and 0.9998 respectively, indicating that the stationary phase beds were stable over the whole range of the applied flow rates (from 1 to 10 μL/min) as shown in Fig. 4.
At the current state of our knowledge it is hard to explain surprisingly high efficiencies obtained on biosilica columns.
In our other paper (currently under revision), dealing with the utilization of biosilica modified with C18 bonded phase we discussed this phenomenon as the observed efficiencies were even higher than those reported here.Briefly, it is hard to classify the biosilica material, which is the population of frustules without any debris.On the one hand, from the classical chromatographic point of view they are not regular as they are not spherical.On the other hand, all the diatom exoskeletons are of the same shape and the same size as can be seen in Fig. 1.The frustules used in this work were of ca. 5 μm in diameter, their sidewall thickness was around 1 μm, and the thickness of the bottom is between 110 and 150 nm.Also, holes (pores) of 150-300 nm are present in the bottom of each frustule.Their distribution across the column is random, which could be seen in the SEM micrographs of the column cross sections (not shown here).The observed chromatographic behavior of the studied biosilica prepared The frustules positioned parallel to the long column axis would have a very narrow cross-section in terms of flow resistance.Those frustules in a position perpendicular to the long column axis would show the flow resistance cross-section equal to their diameter; however, the holes in a relatively thin frustule bottom make such a particle easily permeable.Such a particle, contrary to, for example, completely porous spherical silica, would exhibit a very low thickness of a stagnant mobile phase, so any diffusion-dependent effects on a band broadening would be reduced.So, in our opinion the cumulative effect of these properties make the biosilica columns so efficient, but whether exoskeletons of other diatom species could show similar properties would be the subject of our future work.The ability of a chromatographic stationary phase to efficiently separate polar compounds under HILIC mode depends on the preferential adsorption of water from a mobile phase which is rich in the organic solvent.Then, the retention of analytes is connected with their partition between the bulk mobile phase and the adsorbed water layer [14,15].As shown in Fig. 2, the biosilica material could separate polar test compounds in HILIC mode; however, their retention was worse than on commercially available HALO HILIC columns.The retention factors (k) for adenosine, and cytosine were calculated from the obtained results for Biosilica HILIC (batch 1 and 2) and HALO HILIC, and their values were as follows: 0.05 and 0.11, 0.06 and 0.12 as well as 0.51 and 1.01, respectively.Poorer retention was a result of the lower amount of the adsorbed water in the column as the surface area of biosilica (S BET = 16.9 m 2 /g) is lower than HALO material (S BET = 150 m 2 /g, according to manufacturer's data).The ratio of S BET,HALO /S BET,biosilica = 8.87, and it is consistent with the separation results as the retention factors ratios are: k adenosine,HALO /k adenosine,biosilica1 = 10.2, k cytosine,HALO /k cytosine,biosilica1 = 9.18, k adenosine,HALO /k adenosine,biosilica2 = 8.5, k cytosine,HALO /k cytosine,biosilica2 = 8.42.
The test mixture consisting of five compounds was also separated.The corresponding retention factors (k) as well as peak asymmetry(ƒAs) data of all the individually injected analytes are shown in Table 1 and the separations in isocratic and gradient elution are presented in Fig. 5A, B.
Analyzing the separation results presented in Table 1 and in Fig. 5 one can see that as a result of the change in the composition of the hydro-organic mobile phase a water-rich layer formed on the polar surface of the biosilica/diatoms plays a crucial role in the separation process.The polar analytes interact not only with silanol groups but mostly interact with waterrich layers near the silica surface [14,15].The dominant interactions here are hydrogen bonding (with an excess of water present close to the silica surface).Hence, polar individuals as nucleotides have strong interactions with the sorbed near-surface water layer.This contributes not only to the retention of separated analytes but also to peak symmetry and resolution.

Conclusions
Diatom biosilica frustules were successfully used as a stationary phase in the HILIC mode for the first time.The obtained stationary phase was effective in separating nucleic   bases and nucleosides, which were used as test compounds.However, compared to the commercial stationary phase, biosilica showed lower retention of the test analytes, which is most likely connected to its lower surface area.Nevertheless, the diatom-originated stationary phase may be an efficient separation medium after proper preparation of such material, and the subject is worth further studies.

Table 1
Retention coefficients and asymmetry factors of example test mixture at different compositions of the mobile phase (acetonitrile/ammonium formate pH 3.0, c = 0.1 M)