Dextran-low fraction for biochemistry (M_w∼90 kDa), iron(III) chloride hexahydrate, iron(II) chloride tetrahydrate, sodium borohydride, sodium hydroxide, 30% ammonium hydroxide, tris buffer (2 M), magnesium chloride hexahydrate and calcium chloride dihydrate were obtained from Fisher Scientific (Fair Lawn, NJ) and used as received. Pentaerythritol glycidyl ether was obtained from Frontier Scientific, Inc. (Logan, UT) and used as received. Water was double-deionized using a Millipore Milli-Q system to produce 18 MΩ deionized water. Sealable 5 ml borosilicate microwave vials were obtained from Chemglass, Inc. (Vineland, NJ) and used as received. Tangential flow filtration for purification of synthesized Dex-SPIONS was performed using a KrosFlo Research Iii TFF system from Spectrum Labs, Inc. (Rancho Dominguez, CA). Preparation of stock electrolyte and brine compositions used in evaluation of colloidal stability is detailed in the supporting information.

Dex-SPIONs were synthesized according to the cold gelation approach as previously referenced. Crosslinking of dextran around the periphery of the nanoparticle was carried out using the tetra-crosslinker pentaerythritol glycidyl ether. Purification of the resulting nanoparticles via tangential flow filtration afforded superparamagnetic nanoparticles with an intensity average hydrodynamic diameter centered around 50 nm (Figure 1B). Examination of the nanoparticles using cryo-TEM (Figures 1A,C) reveals that the nanoparticles are actually clusters of magnetite crystallites that average 7 nm in diameter. These clusters are presumably held together by the crosslinked dextran shell which also increases the hydrodynamic diameter of the assembly due to its extended conformation in aqueous media.

As ultimate reservoir deployment of Dex-SPIONs is contingent upon their compatibility with reservoir fluids, evaluation of their stability in representative multi-valent/component brines proves crucial. The colloidal stability of Dex-SPIONs was followed via DLS at 103°C in both seawater and brine over the course of 2 weeks. As shown in Figure 2A, the colloidal stability of Dex-SPIONs in brine is satisfactory, however; the case in seawater is different as the average hydrodynamic diameter of the nanoparticles increases as a function of time—an early indication of colloidal instability. Initially, we were surprised by this result as the total electrolyte content of seawater is substantially less than that of brine. To investigate further, we proceeded to conduct a root cause analysis to identify if a particular ion present in the seawater was responsible for the decreased stability. The results of this study are depicted in Figure 2B which displays the hydrodynamic diameter of the nanoparticles as a function of time at 103°C in various salt solutions. Each salt component is present within the seawater formulation and is of the concentration as it exists in seawater. As can be seen from the results, magnesium is the only ion that leads to an increase in hydrodynamic diameter as a function of incubation time. Similar to earlier observations, an initial salt- and/or heating-induced dehydration event occurs across all samples in the initial 2-day time period before individual salt influences take full effect.

This preliminary result was indeed surprising as magnesium is also present in brine (see Supplementary Table S1) yet the nanoparticles are quite stable in this media; however, brine also possesses a much higher concentration of CaCl₂ in comparison with seawater, suggesting that Ca²⁺ may serve as the mitigating ion present in brine that off-sets the instabilities caused by magnesium salts, thus stabilizing the Dex-SPIONs against their coating-dependent aggregation.

Divalent ions such as calcium and magnesium pose the greatest challenge to the colloidal stability of nanomaterials in natural fluids such as connate and seawater. Within subsurface environments, high salinities bear strong association with nanoparticle aggregation, thus hindering their mobility (Y. Li et al., 2008; Saleh et al., 2008). With the predominance of Ca²⁺ and Mg²⁺ ions in synthetic seawater and low-salinity brine, an investigation into their potential (de)stabilizing effects is necessary. Dex-SPIONs were dispersed into aqueous media containing varying concentrations of CaCl₂ and MgCl₂, and the resulting solutions were subjected to continuous heating. Initial evidence of sample precipitation occurs after 36 h of heating (Figure 3A, t = 36 h), at which point the left-most sample containing solely magnesium salts begins to precipitate. Our following observations suggest a definite correlation between colloidal stability and the amount of calcium present in solution, as prolonged heating induces sample precipitation occurring in a seemingly ordered manner consistent with the hypothesis that brines containing higher levels of Ca²⁺ enable preservation of colloidal integrity for longer periods of time; with increasing Ca²⁺ concentration comes greater delays in particle growth/aggregation, as monitored by DLS (Figure 3B). As the DLS instrument measures each sample in triplicate, average particle sizes were calculated and plotted, with the standard deviation(s) included as y-error. Notably, within the first 24 h (approx.) all 5 samples experience an initial reduction/stagnation in particle growth which may be attributed to a temporary dehydration effect that occurs in the early stages of heating, leading to reductions in measured hydrodynamic diameters of Dex-SPIONs.

Figure 4 summarizes the results of further investigations into the singular, possibly stabilizing role of Ca²⁺ ions. Tap and deionized water were included within the series of electrolyte solutions as references. All Dex-SPION-containing dispersions were found to be stable at room temperature (Figure 4A), both initially and over time; however, precipitation of the deionized water sample occurred after 23 days of heating at 103°C (Figure 4B). Under the premise that calcium ions provide stabilizing effects, the lack of calcium in deionized water implies that no ion-specific stabilization is afforded; survivability of the tap water sample, however, may be attributed to the intrinsic calcium content of tap water (Morr et al., 2006).

Real-time analysis of calcium-dextran binding activity was performed using a Ca²⁺- selective electrode, wherein dextran solution was gradually titrated into electrolyte solution under stirring and the changes in [Ca²⁺] measured. Experimental details regarding electrode usage are outlined in the supporting information. Prior to the addition of dextran solution, the initial measured ionic activity of Ca²⁺ was 10.8 mM (Figure 5); if we solely account for the effects of dilution, the resulting Ca²⁺ activity upon completion of dextran addition should be equal to 10.3 mM. The discrepancy between this value and the measured value of 9.7 mM, however, indicates the possibility of Ca²⁺ binding to the dextran periphery that would then lead to a decrease in the presence of free calcium ions able to be detected.

To better understand the extent of coating stabilization of our Dex-SPION nanoparticles, we chose to examine the degradation mechanics of dextran alone under exposure to reservoir conditions, both at atmosphere (in the presence of oxygen) and under deoxygenated conditions (N₂ purged). Aqueous gel permeation chromatography (GPC) provides a means to detect changes in molecular weight distributions of dextran and its degradative products based on size-exclusion principles. Our results show that the molecular weight distribution of dextran dissolved in both DI water and seawater (abbreviated as DI and SW, respectively) does not change (after 7 days) with respect to their respective room temperature controls, implying that degradation does not occur even at elevated temperatures and in the presence of oxygen (Figures 6A,B). Conversely, GPC analysis of dextran dissolved in both brine and tap water (abbreviated as LSB and Tap, respectively) shows changes in elution volume and detector response that are indicative of molecular weight fragmentation, in reference to the lowest elution curve in each plot that represents samples sealed under regular atmosphere and subjected to heating, labeled as LSB-100C-ATM and Tap-100C-ATM (Figures 6C,D). To explain this phenomenon, we postulate that elevated levels of Ca²⁺ present in brine in comparison with seawater lead to maximized ion-polysaccharide interactions, allowing the dextran polymer to maintain an extended conformation. In this state, the polymer is more prone to degradation via chain scission catalyzed by the presence of heat and oxygen, causing changes in its molecular weight distribution and resulting in the differing GPC elution/detector response; presumably, the presence of calcium in tap water also gives rise to similar results (Xue et al., 2014). Although calcium is present in seawater, the concentration is much less than in brine and as such, the relative abundance of magnesium salts may exert opposing effects, as suggested by previous work (Morr et al., 2006). The implications of this include poorer solvation of dextran that causes the polymer to remain in a shrunken conformation. In this condensed form, the probability of chain scission is minimized as the polymer backbone is shielded from its surrounding environment. In DI water, dextran maintains this same conformation, thus accounting for the lack of degradation observed.

The purpose of the sand pack column transport experiments is twofold—1) to assess the mobility and stability of the Dex-SPIONS as they traverse a porous medium; and 2) to study the effect of ionic composition on the irreversible retention and arrival time of the Dex-SPIONs. Firstly, the assessment of mobility is important with respect to the robustness of the dextran coating. The bulk of prior experiments herein are performed in batch settings and the Dex-SPIONs are not subjected to any shearing effects that may compromise the coating. Although Ottawa sand is a relatively homogenous porous medium compared to consolidated rocks in nature, the tortuosity of the flow paths is significant enough to be deemed suitable for evaluating the fate and transport of engineered nanomaterials (Morr et al., 2006; Saleh et al., 2008; Xue et al., 2014). Each experimental injection contains a Dex-SPION concentration of 625 ppm in their respective ionic solution and spans a pulse width of 3 pore volumes.

Figures 7A–C show the normalized effluent concentration curves of Dex-SPIONS in various ionic solutions compared to those of Dex-SPIONs in DI water. The most telling artifact of these experiments is the slight delay in the arrival of Dex-SPIONs in DI water (shown by the red curves) in comparison with that of Dex-SPIONs in ionic media, as well as the observation that parity with the injected concentration does not quite reach C/C₀ = 1.0 (100% of input concentration). Both the delays in arrival and lower peak effluent concentrations are indicative of retention and perhaps irreversible loss due to adsorption. The mass recovery of the Dex-SPIONs as a function of eluted pore volumes is shown in Figure 7D.

To quantify Dex-SPION transport as a function of ionic solution, a mass balance of the injected pulse and subsequent flush of Dex-SPION-free ionic solution was kept to track material not collected in effluent samples. Below in Table 1, a full summary of the experimental results can be found. Even though the results were considered to be extremely promising, they could not be completely appreciated until an external benchmark could be drawn for comparison. Thus, an experiment in API brine with a Dex-SPION pulse width of 3 PVs was conducted with a setup and conditions equal to that detailed in Kmetz et al., 2016. (Kmetz et al., 2016). In the direct comparison, it is appropriate to compare total SPION recovery given all other parameters were equal. Approximately 99% of Dex-SPIONs were recovered in effluent samples whereas the external study recovered only about 80% of the injected nanoparticles. Likewise, the solid phase irreversible retention was approximately 7x′s larger in the external study.