Deciphering the Parameters to Produce Highly Reproducible and Scalable Iron Oxide Nanoparticles

Reaction Chemistry & Engineering PAPER Cite this: React. Chem. Eng.,2023, 8,1638 Received 23rd November 2022, Accepted 23rd February 2023 DOI: 10.1039/d2re00516f rsc.li/reaction-engineering Deciphering the parameters to produce highly reproducible and scalable iron oxide nanoparticles† Ashish Avasthi, a Carlos Caro, a María Luisa Garcia-Martin* ab and Manuel Pernia Leal * c Nanomedicine has been long hailed as a game changer for treating several ailments, but its translation from bench to bedside is facing some hurdles. Over the past few decades, there have been a plethora of reports regarding the synthesis of nanomaterials and, in particular, of iron oxide nanoparticles. However, very few reports discuss the role of stirring speed, reproducibility, and scalability. This work attempts to comprehensively revisit the most widely used existing protocols and discuss how the particle size or shape varies when certain parameters are altered and different precursors and solvents are used. It also discusses the probability of reproducing and scaling up the reactions while deciphering the effect of the ramp rate on size and shape. Lastly, it upgrades the existing methods and suggests a modification to produce highly reproducible and scalable nanoparticles of ∼4 nm, which can be further tuned to ∼2nmbymerely modifying the stirring speed. 1. Introduction Over the last decade, nanomedicine has shown exceedingly promising results, which, if translated to be used as common clinical practices, hold the capability of revolutionizing the healthcare industry in the future. This successful translation from bench to bedside requires, in the first place, a very efficient synthesis process. There are numerous reports on nanoparticles (NPs) synthesis, but rarely do they mention the scalability and/or reproducibility to make large-scale synthesis plausible, making these two factors the main roadblocks in bringing nanomedicine to the patient. Iron oxide, in particular its nanosized counterpart, is one such material that has been revolutionizing several different industries such as sensor components, 1–3 contrast/imaging agents, 4–10 targeted drug delivery, 11–14 magnetic nanoprobes, 15 nanocatalysts 16,17 and magnetically responsive materials. 18,19 In the field of medicine, iron oxide nanoparticles (IONPs) hold an advantageous place due to their non-toxic nature, biocompatibility, relative ease of functionalization, and, in some cases, their ability to act as magnetic hyperthermia mediators. Up until mid-2022, close to 7000 studies (PubMed database) had been published on magnetic NPs, showcasing IONPs as potent contrast agents (CAs) for magnetic resonance imaging (MRI), but extremely low clinical translation has been achieved so far. 20 Some of the reasons for this low number include the possible generation of ROS species upon the interaction between bare IONPs and cellular structures, 21–24 and, in the absence of proper coating, it could also show increased toxicity, or it could lead to oxidation of IONPs, thereafter forming a more toxic oxide of iron (Fe 2 O 3 ), which would lead to a reduction in viability of cells upon prolonged exposure. 25–28 Despite the abundant use and applications of IONPs, their mechanism of synthesis is still in debate after decades of research. This complexity in defining the mechanism is illustrated by recent reports challenging the pre-established mechanism by LaMer while other studies continue to support the mechanism favoring nucleation. 29–34 The continuous efforts placed in decoding the synthesis mechanism imply the importance of its knowledge to aptly tailor the NPs to desired size and shape according to the need of the application. Most properties, such as saturation magnetization, biodistribution, tumor targeting, or cell uptake, depend on the size and shape of the NPs. 27,35–40 Moreover, in some reports, it has also been observed that shape is even more relevant than size. 41–48 The importance of controlling size and shape is further highlighted by the fact 1638 |React. Chem. Eng.,2023,8,1638–1653 This journal is © The Royal Society of Chemistry 2023 a Instituto de Investigación Biomédica de Málaga y Plataforma en Nanomedicina (IBIMA Plataforma BIONAND), Universidad de Málaga, C/Severo Ochoa, 35, 29590 Málaga, Spain. E-mail: [email protected] b Biomedical Research Networking Center in Bioengineering, Biomaterials & Nanomedicine (CIBER-BBN), Málaga, Spain c Departamento de Química Orgánica y Farmacéutica, Facultad de Farmacia, Universidad de Sevilla, 41012 Seville, Spain. E-mail: [email protected] †Electronic supplementary information (ESI) available. See DOI: https://doi.org/ 10.1039/d2re00516f Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online View Journal | View Issue React. Chem. Eng.,2023,8,1638–1653 | 1639This journal is © The Royal Society of Chemistry 2023 that magnetic properties such as magnetic anisotropy and magnetic susceptibility are particularly important for IONPs to be used as MRI CAs. The low translation of IONPs in clinical products can also be attributed to the failure to reproduce the results obtained inside the laboratory on an industrial scale. It also calls for a need to re-evaluate our reporting of such application-oriented NPs. 49,50 In this work, we evaluate how size and shape are affected when using different precursors and solvents. The thermal decomposition method was chosen for the synthesis as it is the most commonly used method for the preparation of IONPs for biomedical applications due to its several advantages, including narrow size distribution, good crystallinity, and high saturation magnetization values. 4 In addition to using different solvents (hexadecene, octadecene, benzyl ether, or diphenyl ether), different temperature ramp rates were also tested since the size and shape of the NPs are shown to be affected by them as well. 51 Finally, magnetic relaxivity was determined at clinical magnetic fields since one of the main applications of the NPs synthesized here is as MRI CAs. The results from this extensive study revealed a few loopholes in different synthesis methods, which are discussed in detail throughout the article. In addition to discussing these different results based on parametric differences, a modification to the synthesis method is introduced that helps yield highly reproducible and scalable NPs of ∼4 nm. These NPs have high relaxivity (the highest among all the work presented throughout this article), making them ideal as MRI contrast agents, particularly for difficult-to-reach tumors, e.g., glioblastoma multiforme. 52 This size can further be reduced to ∼2 nm. 2. Experimental 2.1 Materials Iron(III) chloride hexahydrate (CAS –10025-77-1), hexadecene (CAS –629-73-2), and diphenyl ether (CAS –101-84-8) were purchased from Merck. Oleic acid (CAS –112-80-1), iron acetylacetonate (CAS –14024-18-1), 1,2-hexadecanediol (CAS – 6920-24-7), oleyl alcohol (CAS –143-28-2), oleylamine (CAS – 112-90-3), benzyl ether (CAS –103-50-4), and octadecene (CAS –112-88-9) were procured from Sigma Aldrich, and sodium oleate (CAS –143-19-1) from Sigma Life Science. Ethanol (99.9% purity) (CAS –64-17-5) and hexane (99% purity) (CAS –110-54-3) were purchased from PanReac AppliChem, acetone (99.5% purity) (CAS –67-64-1) from Arcos Organics, and toluene (99% purity) (CAS –108-88-3) from Alfa Aesar. All the chemicals were used without further purification. 2.2 Synthesis of IONPs 2.2.1 Using iron oleate. Prior to the synthesis of IONPs, the precursor, iron oleate, was synthesized following the method reported by Park et al. 53 Briefly, 10.8 g of iron chloride were mixed with 36.5 g of sodium oleate in a solution containing hexane (160 mL), ethanol (80 mL), and distilled water (60 mL). The mixture was stirred under inert conditions (using N 2 ), and the temperature was increased to reach reflux. The reaction was continued for 4 h and then allowed to naturally cool down to room temperature. Then, the mixture was washed in a separation funnel with distilled water 3 times to obtain a slurry. The slurry was then subjected to rotavapor to remove solvents and attain a viscous reddish-brown slurry. Later, IONPs were synthesized following the method by Park et al., 53 wherein 3.6 g of iron oleate and 0.57 g of oleic acid were mixed with 20 g of octadecene in a flask. They were mixed in an inert atmosphere (N 2 purging through a Schlenk line connected to the reaction flask). The setup was then heated at a ramp rate of 3.3 °C min −1 (using an automatic programmable temperature controller) up to the reflux temperature and allowed to stay there for 1 h. The color of the solution turned black. Then, it was allowed to cool down to room temperature. The NPs were collected by adding a mixture of acetone and ethanol (1 :1) to precipitate the particles and centrifuging them thrice (6000 rpm). Finally, they were redispersed in toluene. 2.2.2 Using iron acetylacetonate. For the synthesis of IONPs using iron acetylacetonate (Fe(acac) 3 ), previously described methods were used and modified. 54–56 1gof Fe(acac) 3 was mixed in phenyl ether (20 mL) with 2.584 g of 1,2-hexadecanediol, 1.69 g oleic acid, and 2.5 g oleylamine under N 2 atmosphere, and heated to reflux for 30 min. Then, the mixture was allowed to cool down to room temperature. The NPs were precipitated using ethanol:acetone mixture (1: 1), and collected by centrifugation. Particles were redispersed in toluene. Further, the method reported by Takahashi et al. 57 was also used and modified. 0.17 g of Fe(acac) was mixed along with 1.04 g of 1,2-hexadecandiol, 2.25 g of oleic acid, 0.53 g of oleyl amine, and 10 mL of benzyl ether. The reaction mixture was heated to 110 °C for 1 h with N 2 flow. The temperature was then raised to 200 °C and kept at that temperature for 30 min. Under N 2 blanket, the mixture was refluxed with a heating rate of 15 °C min −1 and kept refluxing for 1 h. The solution mixture was cooled down to room temperature, and the NPs were precipitated out by adding ethanol. The product was washed with ethanol and resuspended in the mixture of toluene and hexane. 2.2.3 By the addition of oleyl alcohol. IONPs using oleyl alcohol were synthesized using a previous report by Kim et al. 7 with slight modifications. 1.8 g of iron oleate, 0.57 g of oleic acid, and 1.61 g of oleyl alcohol were mixed with 10 g of diphenyl ether. The reaction mixture was heated to 250 °Cat a ramp rate of 10 °C min −1 and allowed to stay at this temperature for 30 min under inert atmosphere. They were then collected following the method described above. 2.2.4 Small-sized IONPs synthesized using modified protocols. To obtain NPs of 2–4 nm, some modifications to previously reported procedures were made. 55,56,58 1 g of Fe(acac) 3 was mixed in phenyl ether (20 mL) along with 2.58 g of 1,2-hexadecanediol, 1.69 g of oleic acid, 1 g of sodium oleate, 0.9 g of oleyl alcohol and 5 mL of oleylamine Reaction Chemistry & Engineering Paper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 1640 |React. Chem. Eng.,2023,8,1638–1653 This journal is © The Royal Society of Chemistry 2023 under N 2 , and heated to reflux for 1 h. The mixture was allowed to cool down to room temperature. The NPs were precipitated using ethanol: acetone mixture (1 :1), and then collected by centrifugation. The NPs were redispersed in toluene. 2.3 Characterization 2.3.1 Transmission electron microscopy (TEM). TEM images were acquired on a FEI Tecnai G2 Twin microscope operated at an accelerating voltage of 100 kV. TEM samples were prepared by dropping a solution of the IONPs at ∼1g L −1 of Fe on a carbon-coated copper grid and letting the solvent evaporate. For each synthesis, an average of 100 NPs were measured for size calculation. 2.3.2 X-ray diffraction (XRD) pattern. XRD was studied with a PANalytical X'pert PRO MPD diffractometer using CuKα 1 radiation in the 2θrange of 10–80 degrees. 2.3.3 Fourier transform infra-red (FTIR) spectroscopy. FTIR spectra were recorded with an FTIR-4100 Jasco using a single reflection ATR accessory (MIRacle ATR, PIKE Technologies) coupled to a liquid nitrogen-cooled mercury cadmium telluride (MCT) detector. All spectra were recorded in the 4000 to 500 cm −1 range at 4 cm −1 resolution and 50 scans. 2.3.4 Inductively coupled plasma high-resolution mass spectroscopy (ICP-HRMS). Fe concentration was determined on an ICP-HRMS. 2 mL of aqua regia was added to 25 μLofa solution of NPs in a volumetric flask. The mixture was left to digest the particles overnight. Then, Milli-Q water was added to complete the total volume of 25 mL. 2.3.5 Dynamic light scattering (DLS). The size distribution and zeta potential were measured on a Zetasizer Nano ZS90 (Malvern, USA). The nanoparticles were dispersed in Milli-Q water, phosphate buffer solution (PBS), and fetal bovine serum (FBS) at a concentration of 200 mg L −1 of Fe. To Fig. 1 IONPs synthesized using iron oleate with octadecene (scale bar –50 nm) [a–d–protocol from ref. 44 was followed; e–g–concentration of iron oleate was varied; h –synthesis in the absence of oleic acid; i and j –synthesis with increased and decreased ramp rate respectively; k and l reproducibility at low ramp rates; m–o–attempts made with different molar ratios of the reaction with different ramp rates; p –synthesis conducted at increased stirring speed; q–s–synthesis scaled up 5 times; t –synthesis scaled up at a higher stirring speed]. Reaction Chemistry & EngineeringPaper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online React. Chem. Eng., 2023, 8,1638–1653 | 1641This journal is © The Royal Society of Chemistry 2023 properly measure the DLS in FBS, we set the following values for the dispersant: 1.3568 as refractive index, 0.94 cP as viscosity, and 80 as dielectric constant. For the size distribution measurement, the number mean was selected. 2.3.6 Transverse relaxivities (r 2 ). The r 2 values were calculated at 1.44 T on a TD-NMR system (Bruker Minispec) using concentrations of magnetic NPs ranging between 2 and 0.0125 mM of Fe under physiological conditions. 3. Results 3.1 Synthesis of IONPs 3.1.1 IONPs synthesized using iron oleate as precursor and octadecene as solvent. The experiments were conducted based on the previously published article. 53 The synthesized NPs were relatively homogenous with high monodispersity. The size of the NPs varied from 8.5 ± 0.7 nm to 9.8 ± 0.5 nm (Fig. 1(a–d)). Subsequently, some alterations were made to check if they had any effect on the size or shape of the NPs. A detailed table containing the sizes of all NPs depicted in Fig. 1 to 4 is included in the ESI†(Tables S1–S4). First, the amount of the precursor added was varied, while the amount of other reactants was kept the same as in the aforementioned protocol. Three different variations to iron oleate quantity were attempted –1.8 g, 3 g, and 4.5 g. Interestingly, the particle size remained ∼9 nm in all cases, as shown in Fig. 1(e–g). Next, the amount of the surfactant oleic acid was varied. It was observed that even at a concentration of oleic acid three times that of the original protocol, the NPs were still ∼9nm (Fig. S1†). Additionally, the synthesis of IONPs in the absence of oleic acid was also tested, but it resulted in high polydispersity (11 ± 3.0 nm) (Fig. 1h). Since the impact of the ramp rate has been stressed upon, leading to various reports on the inverse relationship between size and ramp rate, changes to it were made. 51,59 The ramp rate was initially increased from 3.3 °Cmin −1 up to 15 °Cmin −1 , but, surprisingly, no discernable change in the size and shape of the particles could be appreciated (Fig. 1i). However, when the ramp rate was decreased to 1 °Cmin −1 , the size increased as predicted to 17.7 ± 1.1 nm (Fig. 1j). Notably, monodisperse NPs of this size were only obtained in 1 out of 7 attempts. The other attempts rendered highly polydisperse NPs, albeit with considerably larger sizes. These results indicate that particles with larger sizes can be formed at low ramp rates but with very low reproducibility (Fig. 1k and l). Furthermore, when the molar ratio of the whole reaction was varied instead of one reactant (amount of iron oleate, oleic acid, and octadecene were all varied) at the same ramp rate, neither the size nor the shape of the NPs changed noticeably (between 8.4 ± 0.7 nm and 9.2 ± 0.7 nm). On the other hand, changes in the ramp rate from the initial 3.3 °C min −1 to 5 °C min −1 ,8°C min −1 ,or15°C min −1 , produced NPs with similar size (∼9.5 ± 1.2 nm) and shape with varying Fig. 2 IONPs synthesized using iron oleate and hexadecene (scale bar –50 nm) [a and b –synthesis using hexadecene as a solvent following the aforementioned protocol; c–e–synthesis after changing the surfactant/precursor ratio; f –synthesis at increased ramp rate (7 °C min −1 ); g – synthesis at lower ramp rate (1 °C min −1 ); h –synthesis at a higher stirring speed (∼1500 rpm); i –synthesis when scaled up to 5 times]. Reaction Chemistry & Engineering Paper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 1642 |React. Chem. Eng.,2023,8,1638–1653 This journal is © The Royal Society of Chemistry 2023 degrees of polydispersity (Fig. 1m–o). There were minor differences in size, which are discussed later. The results achieved are conflicting to the widely accepted postulates that establish a clear relationship between the NPs size and the reaction ramp rate. 51,59 Varying the amount of precursor at a lower ramp rate (1 °C) led to some variations in the size and shape of NPs (Fig. S2†). The particles were bimodal, cuboidal or elongated. Another factor that was evaluated was the stirring speed. Initially, an attempt was made to homogenize the nanoparticles at the lower ramp rate of 1 °C min −1 by increasing the stirring speed. However, it did not help homogenize the particles or develop proper facets (Fig. S3†). Later, the stirring speeds were studied as a variant, with speeds varying from ∼250 rpm up to ∼1600 rpm. Astonishingly, the size of the NPs remained almost invariable (∼8.4 ± 1.1 nm). However, NPs of ∼7 nm were obtained when the stirring speed was increased from ∼800 to ∼1600 rpm in a couple of experiments (Fig. 1p). Additionally, in the reactions where iron oleate concentration was varied, no change was observed in polydispersity. It is worth noting that the vacuum/N 2 cycle was used at the beginning of all experiments, but later some experiments were conducted in the absence of vacuum, with the purging of N 2 starting 15 min before the ramp was initiated. However, it did not impact the shape or size of the particles (Fig. S4†). Finally, the scalability of this synthesis with octadecene was evaluated by performing several scale-up (5×) experiments. The resulting NPs were usually either polydisperse or had bimodal distribution (Fig. 1q–s). On the other hand, increasing the stirring speed to ∼1600 rpm produced bi-pyramidal-like structures for the scale-up experiments (Fig. 1t). 3.1.2 IONPs synthesized using iron oleate as precursor and hexadecene as solvent. In order to evaluate the influence of the solvent in the synthesis of NPs, octadecene was replaced by hexadecane, and the synthesis of NPs was repeated. 53,60 The size of the NPs turned out to be between ∼7–8.5 nm the majority of the time (Fig. 2a and b). The size of the NPs remained similar (∼7.5 nm) when the precursor concentration was varied (Fig. 2c–e), suggesting that variation in the precursor/surfactant ratio has no direct effect on the size. As for the effect of the surfactant, the change in just the oleic acid concentration did not produce any change either. Additionally, no remarkable change was observed after increasing the ramp rate from 3 °C min −1 to 5 °C min −1 or 7 °C min −1 , resulting in NPs between 7.4–8.5 nm (Fig. 2f). Moreover, no drastic change in size and morphology was observed when the NPs were synthesized in hexadecene at a ramp rate as low as 1 °C min −1 , contrary to NPs synthesized using octadecene (Fig. 2g). Interestingly, an increase in the stirring speed up to ∼1500 rpm at higher ramp rates was suspected of causing a further decrease in size down to ∼6.5 nm (Fig. 2h). However, when the stirring speed was similarly increased at the lower ramp rate, it gave slightly larger NPs, of 9.6 ± 0.8 nm (Fig. S5†). When the stirring speed was decreased to the minimum, polydisperse NPs were formed that were difficult to separate (Fig. S6†). Reactions conducted using hexadecene showed reasonable scalability, with the particle size being around 7.5 nm. However, these syntheses resulted in mainly spherical NPs instead of the cubic counterparts previously synthesized at a low scale (Fig. 2i). Another important aspect to mention here is that only 1 in 3 scale-up reactions gave the desired product having the same size and shape (Fig. S7†). Fig. 3 IONPs synthesized using iron acetylacetonate as a precursor with benzyl ether and diphenyl ether (scale bar –50 nm) [a –synthesis using the mentioned protocol with benzyl ether; b –synthesis with increased oleylamine concentration; c –synthesis performed at a ramp rate of 15 °C min −1 (adapted from ref. 46); d –synthesis performed at a lower ramp rate of 5 °C min −1 .; e –synthesis when the solvent was diphenyl ether; f – synthesis with diphenyl ether with the addition of sodium oleate]. Reaction Chemistry & EngineeringPaper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online React. Chem. Eng., 2023, 8, 1638–1653 | 1643This journal is © The Royal Society of Chemistry 2023 Later influenced by the literature, other solvents, such as benzyl ether, a mixture of benzyl ether and octadecene, were also evaluated while using iron oleate as the iron precursor, resulting in highly polydisperse particles 39 (Fig. S8†). 3.1.3 IONPs synthesized using iron acetylacetonate as precursor. Reactions proceeded in a pattern similar to those described in the above sections. Previously established protocols were used for the primary experiments, 54 and various changes were then made to these synthesis protocols, as described below. The syntheses with benzyl ether, following the protocol by Sun et al. 54 and Yang et al., 57 yielded NPs of around 7 nm (6.7 ± 1.4 nm) (Fig. 3a), which were reproducible albeit with some persistent polydispersity. The size of these NPs slightly decreased to ∼5 nm upon further increasing the amount of the surfactant, oleylamine, from 1.6 g to 5 g (Fig. 3b) as previously reported but at the cost of monodispersity. 61 As for the ramp rate, the size of the NPs was affected by its increase leading to the synthesis of NPs up to 18.4 ± 4.7 nm at a ramp rate of 15 °C min −1 . However, these NPs presented random shapes, showing the lack of reproducibility under these conditions (Fig. 3c). Reactions at ramp rates as low as 1 °C min −1 were not attempted because at the relatively low ramp rate of 4 °C min −1 , the size of the NPs had already increased drastically to 37.1 ± 5.5 nm (Fig. 3d). Subsequently, diphenyl ether was used as solvent in the synthesis of NPs in place of benzyl ether, resulting in mainly spherical NPs of 4.4 ± 1.9 nm with some polydispersity (Fig. 3e). In this case, no significant change was observed when the ramp rate was increased or decreased. The reproducibility and scalability worked better for reactions with diphenyl ether in comparison to reactions conducted using benzyl ether as solvent; however, a few experiments suffered from large polydispersity issues. Later, in an attempt to reduce the polydispersity, sodium oleate, a well-known shape-directing agent, 62 was added to the above reaction, resulting in quasicubic NPs with high monodispersity and size of 4.3 ± 0.5 nm (Fig. 3f). However, the particles had somewhat different shapes indicating a lack of control over facet growth. 3.1.4 IONPs synthesis with the addition of oleyl alcohol. Previous reports indicate that the incorporation of oleyl alcohol in the reaction mixture as a co-solvent could affect the size and shape of the produced NPs. 7 Therefore, oleyl alcohol was incorporated into the synthesis of NPs. First, olelyl alcohol was incorporated with octadecene since it has the highest boiling point compared to all the solvents used above. Irrespective of the ramp rate used, the NPs thus produced were highly polydisperse, as well as heterogeneous, with sizes ranging from 5.3 ± 1.6 nm to 7.8 ± 5.8 nm (Fig. 4a and b). Fig. 4 Synthesis of IONPs with oleyl alcohol as co-solvent (scale bar –50 nm) [a and b –use of oleyl alcohol with octadecene; c and d –use of oleyl alcohol with hexadecene; e and f –use of oleyl alcohol with benzyl ether; g –use of oleyl alcohol with diphenyl ether; h –synthesis with diphenyl ether and oleyl alcohol at lower stirring speed (∼100 rpm); i –modified protocol with higher stirring speed (∼1500 rpm) showing a size of ∼4 nm; j –modified protocol illustrating a change in size upon decreasing the stirring speed (∼1000 rpm) to ∼2 nm; k and l –synthesis using modified protocol when scaled up to 5 times]. Reaction Chemistry & Engineering Paper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 1644 |React. Chem. Eng., 2023, 8,1638–1653 This journal is © The Royal Society of Chemistry 2023 Then, hexadecene was utilized with oleyl alcohol and exhibited a similar fate, with NPs synthesized between 4.8 ± 1.4 nm and 6.9 ± 2.5 nm (Fig. 4c and d). Next, oleyl alcohol was used in combination with benzyl ether, which also led to polydispersity and heterogeneity issues, with sizes ranging from 5.7 ± 2.9 nm to 8.5 ± 6 nm (Fig. 4e and f), and a similar trend at different ramp rates or stirring speeds. In addition, oleyl alcohol was also used in combination with benzyl ether and octadecene, producing very polydisperse NPs (Fig. S9†). Diphenyl ether was last used as the solvent for the synthesis of the NPs, resulting in sizes of 3.9 ± 0.5 nm (Fig. 4g) and 5.6 ± 2.7 nm (Fig. S10†) in different trials. This variation in size or polydispersity indicates the reproducibility issues with this synthesis. Moreover, the polydispersity of the NPs further increased when the stirring speed was reduced to ∼100 rpm, leading to NPs of 15 ± 7 nm (Fig. 4h). Eventually, a new synthesis protocol was attempted based on the method recently published by Vangijzegem et al., 58 which combines oleyl alcohol and iron acetylacetonate. This new modification combines iron acetylacetonate along with sodium oleate and oleyl alcohol in order to produce a narrow size distribution, from 4.0 ± 0.6 nm to 4.3 ± 0.7 nm (Fig. 4i). Furthermore, decreasing the stirring speed to ∼1000 rpm elicited a decrease in the size of the NPs, ranging from 2.6 ± 0.3 nm to 2.7 ± 0.5 nm (Fig. 4j) in all experiments conducted, demonstrating the high reproducibility of this protocol. Notably, the scale-up of this new synthesis protocol rendered NPs of 4.5 ± 1.1 nm, regardless of the stirring speed (Fig. 4k and l). These results clearly demonstrate the superior performance of our new protocol in terms of reproducibility and scalability. 3.2 Characterization of IONPs 3.2.1 XRD. The XRD results show the synthesis of magnetite NPs as evident from the (220), (311), (400), (511), (440) peaks at 30.96, 35.43, 41.6, 57.2, and 62.5 degrees, respectively, as verified with literature and JCPDS data (Fig. 5). 3.2.2 Magnetic relaxivity. The magnetic relaxivity of NPs was measured to determine if the NPs synthesized using the Fig. 5 XRD spectra (clockwise) of IONPs synthesized using (A) octadecene with iron oleate, (B) hexadecene with iron oleate, (C) with iron acetylacetonate, (D) with the use of oleyl alcohol (also includes the modified protocol). Reaction Chemistry & EngineeringPaper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online React. Chem. Eng., 2023, 8, 1638–1653 | 1645This journal is © The Royal Society of Chemistry 2023 new protocol were at par or better than NPs synthesized using the protocols described throughout this article. Fig. 6 shows that the NPs synthesized using our method gave comparable or slightly higher relaxivity (r 2 = 100.24 mM −1 s −1 ). 3.2.3 Hydrodynamic diameter (HDD) and stability. Table 1 depicts the comparison of sizes of different NPs measured using TEM and DLS. HDD was measured using DLS (post phase transfer of IONPs using a PEGylated ligand) in different solvents –water, PBS, and FBS to verify the dispersibility and size of NPs for in vivo applications. Only 3 particles were chosen based on the desired application, and their sizes in all three solvents were very similar. Table 2 represents the stability of the 3 selected NPs in PBS over 168 h (1 week). As can be observed, all the NPs were very stable and did not show any sign of aggregation. 4. Discussion There is a plethora of articles on the synthesis of IONPs with different shapes and sizes. However, to produce these particles on a large scale for a specific application, it is imperative to choose the correct method. For this purpose, the right chemicals must be chosen to cut down the cost, scale-up must be efficient, reproducibility should be high, and the NPs thus synthesized must be uniform. Unfortunately, many synthesis protocols published to date fail to fulfill all these requirements. The literature indicates that the mechanism of the reaction depends on multiple conditions, and therefore, it becomes difficult to precisely identify, with the current technology, the main responsible variables. Thus, a comprehensive study has been conducted on how different variables, such as the precursor, solvent, temperature, ramp rate, and stirring speed, influence the size, shape, reproducibility, and scalability of IONPs synthesis. All the precursors used for the synthesis of IONPs have different decomposition temperatures as well as different breakdown mechanisms. These different decomposition mechanisms in the precursors have been noted as the prime reason to use different solvents and surfactants with different precursors. 63 For instance, the growth of IONPs from iron oleate was initially thought to be directly from iron oleate, but improvements in scientific techniques unveiled the presence of iron-oxo clusters. 64 Meanwhile, iron acetylacetonate decomposes slightly differently when compared to iron oleate. It needs to be reduced prior to its decomposition to Fe 2+ ions, so 1,2-hexadecanediol was used for this purpose. Moreover, the theories behind the mechanism for the formation of nanoparticles have also evolved from simple LaMer's model of nucleation and growth 31 to a continuous process 30 and defect-mediated ripening. 29 Iron oleate was used in this study with two different solvents, viz. octadecene (b.p. 315 °C) and hexadecene (b.p. 274 °C). In the experiments conducted using the protocol established by Park et al., 53,60 the synthesized particles were monodispersed and homogenous, with a size around 9 nm. They also displayed acceptable reproducibility. However, these NPs showed a cubic shape instead of the predicted spherical one. On the other hand, the experiments conducted with iron oleate and a mixture of solvents, such as octadecene and benzyl ether, yielded a bimodal distribution, with smaller NPs being monodisperse while larger NPs were heterogeneous (Fig. S8†). Further, since it has been noted in the literature that the structure of iron oleate plays an important part in attaining uniform NPs, FTIR spectroscopy was performed prior to all the reactions to assure the stability of the precursor (Fig. S11†). 50,65,66 It is worth noting that the reactions with 3 months, 6 months, and 1 year-old iron oleate with octadecene rendered similar sizes and shapes, suggesting that just like its amount, the age of the precursor is not a critical factor once iron oleate has been properly prepared and preserved. Another important variable in the synthesis that leads to changes in NPs size and shape is the residence time in reflux. 54,58 Thus, time-dependent experiments were performed to study this effect using iron oleate as precursor with both solvents –octadecene and hexadecene. Aliquots were taken out after 5, 15, 30, 45, and 60 min and were analyzed by TEM (Fig. S12†). The reaction with octadecene began to show monodisperse particles within the first 5 min, with no change in shape or size after 60 min. However, the Fig. 6 Transverse relaxivity at 1.44 T of NPs with different sizes. r 2 values correspond to the slope of the linear fit of 1/T 2 vs. Fe concentration. Table 1 TEM size and hydrodynamic diameters of different NPs in water, PBS, and FBS Sr. no. TEM size (nm) HDD in water (nm) HDD in PBS (nm) HDD in FBS (nm) 1 4.3 ± 0.7 24.8 ± 2.1 27.5 ± 6.4 23.1 ± 4.1 2 7.6 ± 0.7 27.8 ± 5.6 37.5 ± 4.6 25.2 ± 7.5 3 9.3 ± 0.8 24.3 ± 2.1 31.1 ± 2.9 25.9 ± 3.6 Reaction Chemistry & Engineering Paper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 1646 |React. Chem. Eng., 2023, 8,1638–1653 This journal is © The Royal Society of Chemistry 2023 cubical facets became slightly more prominent after 15 min, which can be due to the preferential growth of the crystal (110 to 111). The polydispersity stayed below 1 nm in all aliquots. However, when aliquots were taken out of the reaction with hexadecene, the homogenous and monodisperse particles were not formed until after 60 min (Fig. S13†). After the first 5 min, almost no NPs were formed. In subsequent aliquots, very few NPs (9.5 ± 3 nm) were formed, but after 60 min, the NPs population in a similar aliquot increased dramatically, and size became 8.4 ± 0.7 nm, which suggests that the mechanisms of synthesis defined thus far may need to be redefined for every solvent. Following the study on residence time, a reaction with octadecene as solvent was conducted at 250 °C and it was allowed to stay for 4 h. It produced very monodisperse quasicubic NPs of ∼16 nm (Fig. S14†). This suggests that the importance of residence time is not only solvent-dependent but also temperature-dependent, in particular at temperatures closer to the nucleation temperature. On the other hand, temperature fluctuations during nucleation and growth have also been postulated to cause changes in size and shape. 51 Therefore, we performed experiments where the real-time temperature changes were followed while the ramp was set to increase at a constant temperature. The real-time temperature was noted every minute, and the average change was calculated for temperatures ranging from 200 to 250 °C, 250 to 290 °C, and 290 to 320 °C, to determine which part of the reaction – nucleation, continued nucleation or growth –is the driving force for the change in size. However, as can be seen in Table 3 and Fig. 7, no definite conclusion could be drawn from the gathered data. Iron acetylacetonate was used with benzyl ether (b.p. 298 °C) and diphenyl ether (b.p. 258 °C.) as solvents. Particles synthesized with iron acetylacetonate exhibited a size range of 2–7 nm, as mentioned in previous articles. 54,63 Moreover, it was observed that the crystallinity of particles was better when synthesized using iron acetylacetonate, as well as their relative reproducibility (Fig. 5). In most experiments, NPs with narrow size distribution were produced, albeit with slight polydispersity. Keeping in mind the difficulty of producing monodisperse NPs below 5 nm with clear facets, it can be concluded that the syntheses worked well. On the other hand, the reproducibility of the reaction with benzyl ether was relatively low if only monodisperse particle synthesis was concerned. In the case of diphenyl ether, the size of NPs was smaller, ∼4.5 nm. This smaller size was expected since the reflux temperature for the reaction was relatively closer to the nucleation temperature due to the lower boiling point of the solvent. It has been accentuated in the literature that the rate at which temperature is increased directs the size and shape of the NPs. Therefore, we have thoroughly analyzed the effect of this parameter in our experiments. First, it was observed that the starting point of the ramp (room temperature, 100 °Cor 200 °C) is not a critical factor in determining the size and shape of NPs as long as it is not beyond the nucleation Table 2 Stability of NPs over time in PBS Sr. no. TEM size (nm) 0 h 24 h 48 h 72 h 168 h 1 4.3 ± 0.7 27.5 ± 6.4 28.1 ± 5.5 29.4 ± 5.1 38.0 ± 4.5 25.6 ± 3.3 2 7.6 ± 0.7 37.5 ± 4.6 29.5 ± 4.8 39.9 ± 4.7 32.2 ± 7.3 29.8 ± 4.3 3 9.3 ± 0.8 31.1 ± 2.9 31.0 ± 4.4 35.9 ± 2.1 34.4 ± 2.8 33.4 ± 3.3 Table 3 Change in temperature (real-time) during suspected nucleation (200–250 °C), continued nucleation (250–290 °C) and growth (300–320 °C), and size of the final NPs (only experiments with octadecene are mentioned) Experiment 200–250 °C (°C min −1 ) 250–290 °C (°C min −1 ) 290–320 °C (°C min −1 ) ∼Size (nm) 1 3.8 2 3.3 9.8 2 5 8 1.1 17.8 3 4.1 2.1 2.3 9.3 4 3.1 3 3 8.7 5 3.3 2.8 2.5 9 6 2.9 3 1.6 9.6 7 3.3 3 2.1 8.9 8 3.8 4 4 8.6 9 3.3 3.3 1.7 8.5 10 3.1 3 2 9.3 11 3.1 4 2.5 11.5 12 3.3 2.8 3.3 9.4 13 3.1 2.8 3 7.1 14 2.9 3.6 1.8 18.3 15 20 16 4 8.9 Fig. 7 Temperature variation during IONPs synthesis using iron oleate with octadecene at different times (the numbers on top of each rectangle is the size of the NPs formed). Reaction Chemistry & EngineeringPaper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online React. Chem. Eng., 2023, 8,1638–1653 | 1653This journal is © The Royal Society of Chemistry 2023 values of specific absorption rate for cancer cell hyperthermia treatment, ACS Nano, 2012, 6(4), 3080–3091. 60 J. Park, E. Lee, N. M. Hwang, M. Kang, S. C. Kim, Y. Hwang, J. G. Park, H. J. Noh, J. Y. Kim, J. H. Park and T. Hyeon, Onenanometer-scale size-controlled synthesis of monodisperse magnetic iron oxide nanoparticles, Angew. Chem., Int. Ed., 2005, 44(19), 2873–2877. 61 Z. Xu, C. Shen, Y. Hou, H. Gao and S. Sun, Oleylamine as both reducing agent and stabilizer in a facile synthesis of magnetite nanoparticles, Chem. Mater., 2009, 21(9), 1778–1780. 62 M. V. Kovalenko, M. I. Bodnarchuk, R. T. Lechner, G. Hesser, F. Schäffler and W. Heiss, Fatty acid salts as stabilizers in size-and shape-controlled nanocrystal synthesis: the case of inverse spinel iron oxide, J. Am. Chem. Soc., 2007, 129(20), 6352–6353. 63 A. Roca, M. Morales and C. Serna, Synthesis of monodispersed magnetite particles from different organometallic precursors, IEEE Trans. Magn., 2006, 42(10), 3025–3029. 64 S. G. Kwon, Y. Piao, J. Park, S. Angappane, Y. Jo, N. M. Hwang, J. G. Park and T. Hyeon, Kinetics of monodisperse iron oxide nanocrystal formation by “heating-up”process, J. Am. Chem. Soc., 2007, 129(41), 12571–12584. 65 E. Wetterskog, M. Agthe, A. Mayence, J. Grins, D. Wang, S. Rana, A. Ahniyaz, G. Salazar-Alvarez and L. Bergström, Precise control over shape and size of iron oxide nanocrystals suitable for assembly into ordered particle arrays, Sci. Technol. Adv. Mater., 2014, 15(5), 055010. 66 L. M. Bronstein, X. Huang, J. Retrum, A. Schmucker, M. Pink, B. D. Stein and B. Dragnea, Influence of iron oleate complex structure on iron oxide nanoparticle formation, Chem. Mater., 2007, 19(15), 3624–3632. 67 D. Wilson and M. Langell, XPS analysis of oleylamine/oleic acid capped Fe3O4 nanoparticles as a function of temperature, Appl. Surf. Sci., 2014, 303,6–13. 68 R. A. Harris, P. M. Shumbula and H. T. van der Walt, Analysis of the interaction of surfactants oleic acid and oleylamine with iron oxide nanoparticles through molecular mechanics modeling, Langmuir, 2015, 31(13), 3934–3943. 69 S. Mourdikoudis, M. Menelaou, N. Fiuza-Maneiro, G. Zheng, S. Wei, J. Perez-Juste, L. Polavarapu and Z. Sofer, Oleic acid/ oleylamine ligand pair: a versatile combination in the synthesis of colloidal nanoparticles, Nanoscale Horiz., 2022, 7(9), 941–1015. 70 G. Salas, C. Casado, F. J. Teran, R. Miranda, C. J. Serna and M. P. Morales, Controlled synthesis of uniform magnetite nanocrystals with high-quality properties for biomedical applications, J. Mater. Chem., 2012, 22(39), 21065–21075. 71 Y. Bao, J. Sherwood and Z. Sun, Magnetic iron oxide nanoparticles as T 1 contrast agents for magnetic resonance imaging, J. Mater. Chem. C, 2018, 6(6), 1280–1290. 72 S. Ali, S. A. Khan, Z. H. Yamani, M. T. Qamar, M. A. Morsy and S. Sarfraz, Shape-and size-controlled superparamagnetic iron oxide nanoparticles using various reducing agents and their relaxometric properties by Xigo acorn area, Appl. Nanosci., 2019, 9(4), 479–489. 73 W. Wu, Q. He and C. Jiang, Magnetic iron oxide nanoparticles: synthesis and surface functionalization strategies, Nanoscale Res. Lett., 2008, 3(11), 397–415. Reaction Chemistry & Engineering Paper Open Access Article. Published on 11 April 2023. Downloaded on 1/10/2024 2:13:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online