The Effects of a Varied Gold Shell Thickness on Iron Oxide Nanoparticle Cores in Magnetic Manipulation, T1 and T2 MRI Contrasting, and Magnetic Hyperthermia

Fe3O4–Au core–shell magnetic-plasmonic nanoparticles are expected to combine both magnetic and light responsivity into a single nanosystem, facilitating combined optical and magnetic-based nanotheranostic (therapeutic and diagnostic) applications, for example, photothermal therapy in conjunction with magnetic resonance imaging (MRI) imaging. To date, the effects of a plasmonic gold shell on an iron oxide nanoparticle core in magnetic-based applications remains largely unexplored. For this study, we quantified the efficacy of magnetic iron oxide cores with various gold shell thicknesses in a number of popular magnetic-based nanotheranostic applications; these included magnetic sorting and targeting (quantifying magnetic manipulability and magnetophoresis), MRI contrasting (quantifying benchtop nuclear magnetic resonance (NMR)-based T1 and T2 relaxivity), and magnetic hyperthermia therapy (quantifying alternating magnetic-field heating). We observed a general decrease in magnetic response and efficacy with an increase of the gold shell thickness, and herein we discuss possible reasons for this reduction. The magnetophoresis speed of iron oxide nanoparticles coated with the thickest gold shell tested here (ca. 42 nm) was only ca. 1% of the non-coated bare magnetic nanoparticle, demonstrating reduced magnetic manipulability. The T1 relaxivity, r1, of the thick gold-shelled magnetic particle was ca. 22% of the purely magnetic counterpart, whereas the T2 relaxivity, r2, was 42%, indicating a reduced MRI contrasting. Lastly, the magnetic hyperthermia heating efficiency (intrinsic loss power parameter) was reduced to ca. 14% for the thickest gold shell. For all applications, the efficiency decayed exponentially with increased gold shell thickness; therefore, if the primary application of the nanostructure is magnetic-based, this work suggests that it is preferable to use a thinner gold shell or higher levels of stimuli to compensate for losses associated with the addition of the gold shell. Moreover, as thinner gold shells have better magnetic properties, have previously demonstrated superior optical properties, and are more economical than thick gold shells, it can be said that “less is more”.


Introduction
Inorganic nanomaterials such as gadolinium [1], iron oxide [2], silica [3], and gold [4] are becoming increasingly popular in nanotheranostic applications (combined therapeutics and diagnostics), particularly as imaging agents, drug deliverers, and hyperthermia agents. It is hypothesized that combining materials into a single nanoplatform strengthens the physical properties of the nanosystem. simultaneous optical imaging and magnetic hyperthermia, or magnetic separation combined with SERS. From our previous work [28], we found that thicker gold shells led to increased light scattering, and we would also expect decreased cytotoxicity [32] and circulation half-life [33] for this nanostructure. Meanwhile, thin-to-moderate gold shells facilitated spectrally selective photothermal activity due to the emergence of a spectral drift [28]. For this study, we explored the effects of the plasmonic gold shell in three notable magnetic-based applications, by comparing five different stages of magnetic-plasmonic nanoparticle with varied gold shell thickness. The three popular nanotheranostic applications considered were magnetic separation and targeting ability (by quantifying magnetophoresis), T 1 and T 2 MRI contrasting (by quantifying NMR relaxivity), and magnetic hyperthermia activity (by monitoring the heat generation in an alternating magnetic field).

Synthesis of Multistage Magnetic-Plasmonic Nanoparticles
First, 20.5 ± 1.3 nm diameter commercial oleic acid-capped iron oxide nanoparticles (Ocean NanoTech, San Diego, CA, USA) were amine functionalized by ligand exchange using (3-aminopropyl) triethoxysilane (APTES) and using a protocol modified from [34], to aid further functionalization and improved biocompatibility [35] (denoted O, black in graphs). Next, tetrakis(hydroxymethyl) phosphonium chloride (THPC)-reduced gold nanoseeds prepared by the Duff method [36] were grafted onto the iron oxide core giving a core-satellite or seeded stage (Os, red). The gold-seeded iron oxide nanoparticles were then subjected to three iterative gold reductions using a gold-plating solution and formaldehyde reduction modified from [37], yielding three further stages of gold growth called R1 (green), R2 (blue), and R3 (orange) throughout, and had a diameter determined by TEM of 28.5 ± 2.2, 42.1 ± 10.3, and 103.7 ± 16.9 nm, respectively. To read further details about the synthesis of these multistage magnetic-plasmonic nanostructures and see transmission electron micrographs of the various stages, please see our earlier work in [24]. To aid the comparison between gold-coated stages and uncoated iron oxide, the mass of iron oxide (i.e., the number of nanoparticles) was kept constant in this work.

Magnetophoresis
To conduct optical tracking of magnetophoresis, a deuterium-halogen UV-VIS-NIR light source (DH2000-BAL, Ocean Optics, Orlando, FL, USA) using only halogen as the light source was passed through a quartz cuvette (CV10Q3500, Thorlabs, Newton, NJ, USA) containing the suspended nanoparticles. The transmitted light was focused onto a flame spectrometer (Ocean Optics), which was paired with OceanView software (Version 1.6.7, Ocean Optics). The probed area was a cylindrical region in the cuvette with a 5 mm diameter and a 10 mm height, centered 70 mm above a magnet (neodymium, 10 mm × 10 mm × 10 mm). The magnet's distance-dependent magnetic flux density was measured using a gaussmeter. The extinction peak was monitored temporally, and a linear fit of the peak extinction is used to monitor magnetophoresis.

NMR T 1 and T 2 Relaxation Measurements
To conduct T 1 and T 2 measurements of the multistage magnetic-plasmonic nanostructures, a benchtop NMR ( 1 H Spinsolve 60 MHz, Magritek, Malvern, PA, USA) was used. The magnetic field strength was 1.5 Tesla, and the proton Larmor frequency was 60MHz. To track the effect of the gold shell on the magnetic nanoparticle relaxivity, the 1 H (water) peak at ≈4.8 ppm was monitored.
T 1 measurements were conducted using an inversion recovery pulse sequence, starting with the shortest inversion delay; a series of experiments were conducted with the inversion delay stepped linearly to the maximum inversion time. Two scans of 1.6 s acquisition time (8192 points, dwell time of 200 µs), repetition time of 7 s, and a varied inversion time of ≈5 s with 21 steps were taken for the T 1 measurements. T 2 measurements were made using a Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence, starting with the shortest echo time τ; a series of measurements were conducted where 180-degree pulses were stepped linearly, ending with the final (variable) echo time. Four scans with an acquisition time of 0.4 s (8192 points, 50 µs dwell time), 4 s repetition time, 20 steps with a CPMG echo time of 0.2 ms, and a varied final echo time of ≈2 s were taken for the T 2 measurements. The 90 • pulse width was 7 ms for both measurements.

Magnetic Hyperthermia
In this research, a Flexitune induction heater (Radyne, Wokingham, UK) was used with a water-cooled homemade copper coil, generating a 100 mT alternating magnetic field, with a frequency of 20 kHz. Induction heaters are typically used to heat a conductive material by eddy currents. Induction heaters are simply AC passed through an inductor (e.g., coil), thus producing an alternating magnetic field, per Ampere's law. A calibrated A655sc (FLIR, Wilsonville, OR, USA) infrared camera was used to monitor the magnetic field-induced hyperthermal activity, noncontact. Aqueous solutions of the various stages of magnetic-plasmonic nanoparticles in an Eppendorf tube (Eppendorf, Hamburg, Germany) (at a fixed concentration of 0.2 mg/mL of iron oxide) were suspended inside the cooled copper coil (noncontact) using a plastic mount. It is important to note that these measurements were not conducted under adiabatic conditions (i.e., there was a loss to the environment) unlike many other magnetic hyperthermia measurements [38], which are thermally insulated.

Results
To determine whether a gold coating affects the efficacy of magnetic nanoparticles in various magnetic-based nanotheranostic applications, Fe 3 O 4 -Au nanoparticles were synthesized with various stages of gold growth (see Section 2.1 for details). An illustration of the various nanoparticle stages is shown in Figure 1.
Nanomaterials 2020, 10, x FOR PEER REVIEW 4 of 16 T2 measurements were made using a Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence, starting with the shortest echo time τ; a series of measurements were conducted where 180-degree pulses were stepped linearly, ending with the final (variable) echo time. Four scans with an acquisition time of 0.4 s (8192 points, 50 μs dwell time), 4 s repetition time, 20 steps with a CPMG echo time of 0.2 ms, and a varied final echo time of ≈2 s were taken for the T2 measurements. The 90° pulse width was 7 ms for both measurements.

Magnetic Hyperthermia
In this research, a Flexitune induction heater (Radyne, Wokingham, UK) was used with a watercooled homemade copper coil, generating a 100 mT alternating magnetic field, with a frequency of 20 kHz. Induction heaters are typically used to heat a conductive material by eddy currents. Induction heaters are simply AC passed through an inductor (e.g., coil), thus producing an alternating magnetic field, per Ampere's law. A calibrated A655sc (FLIR, Wilsonville, OR, USA) infrared camera was used to monitor the magnetic field-induced hyperthermal activity, noncontact. Aqueous solutions of the various stages of magnetic-plasmonic nanoparticles in an Eppendorf tube (Eppendorf, Hamburg, Germany) (at a fixed concentration of 0.2 mg/mL of iron oxide) were suspended inside the cooled copper coil (noncontact) using a plastic mount. It is important to note that these measurements were not conducted under adiabatic conditions (i.e., there was a loss to the environment) unlike many other magnetic hyperthermia measurements [38], which are thermally insulated.

Results
To determine whether a gold coating affects the efficacy of magnetic nanoparticles in various magnetic-based nanotheranostic applications, Fe3O4-Au nanoparticles were synthesized with various stages of gold growth (see Section 2.1 for details). An illustration of the various nanoparticle stages is shown in Figure 1. Illustration of the various forms and thicknesses of magnetic-plasmonic nanoparticles studied for this paper. From left, the (3-aminopropyl) triethoxysilane (APTES) functionalized iron oxide (Fe3O4), called O throughout. Next, the Fe3O4 nanoparticles functionalized with gold seeds, Os. Middle, the Fe3O4-Au gold-plated nanoparticle after one gold reduction, R1. R2 is the magneticplasmonic nanoparticle after two gold reductions and, lastly, R3, after three gold reductions. Reproduced with permission from [28]; Copyright The Royal Society of Chemistry, 2020.

Magnetophoresis
Magnetic guidance or translation of magnetic (or hybrid magnetic) nanoparticles is vital in facilitating magnetic flux-driven drug delivery, magnetic cell or blood sorting, and magnetic fieldguided therapy. Magnetic nanoparticles suspended in fluid travel toward the highest magnetic field gradient; this movement is called magnetophoresis. A higher magnetophoresis speed is preferable for most applications and has been achieved by using a stronger magnetic field gradient, higher nanoparticle magnetization or concentration, nanoparticle shape anisotropy [39], and surface charge [40]. Considerable work in the magnetophoresis of various magnetic (and magnetic-plasmonic) structures has been carried out [41][42][43]. Furthermore, the addition of a nonmagnetic silica [44,45] and polymer [43] shell has been shown to decrease the magnetophoretic mobility of magnetite Figure 1. Illustration of the various forms and thicknesses of magnetic-plasmonic nanoparticles studied for this paper. From left, the (3-aminopropyl) triethoxysilane (APTES) functionalized iron oxide (Fe 3 O 4 ), called O throughout. Next, the Fe 3 O 4 nanoparticles functionalized with gold seeds, Os. Middle, the Fe 3 O 4 -Au gold-plated nanoparticle after one gold reduction, R1. R2 is the magnetic-plasmonic nanoparticle after two gold reductions and, lastly, R3, after three gold reductions. Reproduced with permission from [28]; Copyright The Royal Society of Chemistry, 2020.

Magnetophoresis
Magnetic guidance or translation of magnetic (or hybrid magnetic) nanoparticles is vital in facilitating magnetic flux-driven drug delivery, magnetic cell or blood sorting, and magnetic field-guided therapy. Magnetic nanoparticles suspended in fluid travel toward the highest magnetic field gradient; this movement is called magnetophoresis. A higher magnetophoresis speed is preferable for most applications and has been achieved by using a stronger magnetic field gradient, higher nanoparticle magnetization or concentration, nanoparticle shape anisotropy [39], and surface charge [40]. Considerable work in the magnetophoresis of various magnetic (and magnetic-plasmonic) structures has been carried out [41][42][43]. Furthermore, the addition of a nonmagnetic silica [44,45] and polymer [43] shell has been shown to decrease the magnetophoretic mobility of magnetite nanoparticles. Thus, if a polymer/silica shell affects the magnetophoresis, it is reasonable to hypothesize that a gold shell would also hamper this motion.
The magnetophoretic force, F M , exerted on a magnetic particle by an external magnetic field is dependent on the magnetic particle volume, V p , the magnetization of the nanoparticle, M p , the permeability of free space, µ 0 , and the magnetic field strength, H, at the particle location per [46,47]: Equation (1) is sufficient for non-interacting (noncooperative) high magnetic field gradient magnetophoresis. Drag and magnetic interaction forces should be considered in low magnetic field gradient magnetophoresis. For low magnetic field gradient (<100 T/m) magnetophoresis, as is the case here, typically magnetic particles cooperate by clustering together into chain-like structures while moving in the gradient, and this assembly accelerates the magnetophoresis transport [48,49]. This is likely due to the coupling of the individual magnetic moments of the nanoparticles into a new and larger effective magnetic moment. Furthermore, the drag force (friction) experienced by the assembled nanoparticles would be reduced per the slender-body theory [50]. Magnetic dipolar interaction forces that facilitate this magnetic assembly are increased when the nanoparticles are not colloidally stable and are in high concentrations (due to increased probability of proximity).
The magnetic saturation of Fe 3 O 4 -Au nanoparticles has been reported to be decreased compared to uncoated Fe 3 O 4 by varying degrees from ≈10% to 70% [51][52][53][54][55][56], likely a result of increased gold-to-iron oxide mass ratio and magnetic shielding by the diamagnetic gold shell [57][58][59]. This reduced magnetic saturation has also been reported for other materials with increased shell thickness, including gold shells on cobalt nanoparticles [60] and silica shells on FeNi nanoparticles [61]. On the contrary, there are reports of increased magnetic saturation with the addition of a gold shell on Fe 3 O 4 [62,63], and the authors hypothesized that conduction electrons are trapped, which induces an orbital moment. A reduction of magnetic saturation reduces the value M p can reach, thus leading to a reduction in the magnetic force causing the translation of the nanoparticles to the magnet. As the gold shell on the Fe 3 O 4 core was associated with a reduced saturation, we anticipated a reduced magnetic force and magnetophoresis.
In this work, the magnetophoresis of the multistage magnetic-plasmonic nanoparticles was monitored by optical extinction, per the experimental schematic seen in Figure 2A. In this experiment, the concentration of iron oxide (and thus, the number of nanoparticles) was kept constant. Therefore, the volume of magnetic material is constant, but the percentage per particle varies with each stage (see Table 1). Collimated light is passed through the cuvette containing the nanoparticles, which is then focused onto a fiber spectrometer. A permanent magnet is placed below the cuvette, with a distance-dependent magnetic flux density seen in Figure 2B. The probed area is a cylindrical region in the cuvette with a 5 mm diameter and a 10 mm height, centered 70 mm above the magnet. As the nanoparticles migrate toward the magnet (due to magnetic forces) and leave the probe area, the optical extinction decreases. The peak optical extinction was tracked during magnetophoresis and a linear fit was added (see Figure 2C). Further experimental details are given in the experimental Section 2.1.
From Figure 2C and Table 1, increasing the gold shell thickness leads to a significantly longer time to magnetically separate, i.e., decreased speed of magnetophoresis. The iron oxide nanoparticles without any gold coating take approximately 12 min to migrate out of the probed area of the solution. With the addition of gold, the magnetic separation time increases, eventually reaching 1421 min for the thickest gold shell. Thus, thinner gold shells are likely preferable for applications that require reasonable responsivity to the external magnetic field. Alternatively, much larger magnetic field gradients would be required to use these thick gold-shelled nanoparticles in such applications. It is likely a combination of a reduced magnetic saturation and loss of cooperative behavior of the thickly gold-coated nanoparticles that leads to reduced magnetophoresis. Furthermore, an increased drag force would be experienced by larger nanoparticles. Nanomaterials 2020, 10, x FOR PEER REVIEW 6 of 16  It should be noted that the surface charge of the nanoparticles may also affect the interaction between nanoparticles. Zeta potential measurements of the various stages of nanoparticle are given in Table 1, indicating good colloidal stability for R1, R2, and R3. Temperature and pH were constant throughout.  It should be noted that the surface charge of the nanoparticles may also affect the interaction between nanoparticles. Zeta potential measurements of the various stages of nanoparticle are given in Table 1, indicating good colloidal stability for R1, R2, and R3. Temperature and pH were constant throughout.

NMR/MRI Contrast Agent
MRI is a well-known non-ionizing technique that uses radio waves along with strong magnetic fields and gradients to conduct biological diagnostic imaging. MRI imaging is based on NMR principles, where certain atomic nuclei (typically hydrogen) undergo Larmor precession in a strong external magnetic field and, when the Larmor precession frequency of the nuclear magnet resonates with that of an incoming radio wave (RF pulse), energy is absorbed from the radio wave. Protons may align in two energy eigenstates, low and high energy, separated by a small splitting energy. At equilibrium, most protons reside in the low energy state (aligned with the external magnetic field) and a net polarization parallel to the external field. An incident resonant RF pulse may lead to the polarization Nanomaterials 2020, 10, 2424 7 of 16 vector either tipping sideways (90 • pulse) or reversing (180 • pulse), and the protons come into phase with the pulse and one another (phase coherence). In response to the RF pulse, we may consider two relaxations to equilibrium, namely spin-lattice (also known as T 1 or longitudinal relaxation) and spin-spin (also known as T 2 or transverse relaxation). The recovery of the longitudinal magnetization relaxation (M z ) to its thermal equilibrium value (M z,eq ) is called the T 1 (spin-lattice) relaxation; as the T 1 relaxation is related to the redistribution of the spins to the low energy state (thermal equilibrium), energy is dissipated to the surroundings (lattice). T 2 relaxation involves the recovery of the equilibrium state dephased spin; therefore, the magnetization vector perpendicular (transverse) to the static magnetic field is considered, M xy .
A downfall of MRI biological imaging is its low sensitivity. To combat low sensitivity, a contrast agent is commonly employed to selectively alter the relaxation time of the nuclei. As the magnetic relaxation of the nuclei varies with the square of the magnetic dipole moment, gadolinium (Gd 3+ ) with its seven unpaired electrons (high magnetic moment) is often used in chelated form to enhance MRI T 1 contrast. However, there are concerns regarding the safety of these gadolinium-based contrast agents [64]. Superparamagnetic nanoparticles have been proposed as an alternative, but rather mostly for T 2 image contrast.
In terms of MRI, a local reduction (shorter time) in either the T 1 relaxation (positive contrast) or T 2 relaxation (negative contrast) results in better image contrast. As the T 1 and T 2 times are concentration-dependent, we define relaxivities as a function of iron concentration, i.e., r 1 and r 2 . The longitudinal relaxivity r 1 is said to depend on the molecular tumbling time, proton residence lifetime, and coordinating number, while the transverse relaxivity r 2 is proportional to the square of the magnetic nanoparticle radius and the magnetic saturation [65]. As the T 1 relaxation is associated with inner-sphere mechanisms (chemical energy exchange), superparamagnetic nanoparticles are instead more effective in T 2 applications.
In the case of magnetic core-nonmagnetic shell particles, the magnetic field experienced by the water protons (T 2 ) and the degree of chemical exchange (T 1 ) lessens with increased shell thickness. Furthermore, if the nanoparticles are aggregated or assembled, worse r 1 contrast due to reduced surface area and improved r 2 contrast owing to the coupling of the magnetic moments are expected [66]. A relaxivity ratio (r 2 /r 1 ) is often used as another measure of contrast agent usefulness. An optimal negative contrast agent (T 2 ) would have a high r 2 /r 1 ratio along with a high r 2 value. If this ratio is less than 5, the contrast would instead be considered a good candidate for a T 1 contrasting; if larger than 5, the agent may be more suitable for T 2 contrasting [67].
Several r 1 and r 2 relaxivities from the literature are presented in Table 2, as well as the values for the multistage magnetic-plasmonic structures from this work. Marangoni et al. [68] studied the T 1 relaxation of gold-core nanoparticles with a Gd(III) ion-silica layer, followed by a varied gold shell. Interestingly, T 1 contrast was better for the gold-based nanoparticles compared to the typical Gd-based chelating agents. They also found that contrast was optimal for a gold-seeded shell but was seen to decrease with additional gold growth. The authors suggest that the added mass from the shell decreases the tumbling rate, hence decreasing the relaxivity for the thicker shelled nanoparticles (worse contrast). Pinho et al. [69] studied the effect of silica shell thickness on the r 1 and r 2 relaxivities of γ-Fe 2 O 3 (see Table 2 for details). The thickness of silica had a significant impact on both relaxivities, with increased shell thickness leading to a decrease in the relaxivity value (decreased contrast). Others show the decrease in r 1 [70] or decrease in both r 1 and r 2 [69,71] with the addition of a SiO 2 shell, similar to observations in this research with the addition of a gold shell. Moreover, Park et al. noted a reduced r 2 value for a diamagnetic avidin-biotin coating on Fe 3 O 4 nanoparticles and attributed this to diamagnetic shielding effects [72]. Assembled magnetic structures were also seen to increase the r 2 relaxation [73]. In this work, benchtop 1 H NMR spectroscopy was used to monitor the effects of the multistage magnetic-plasmonic nanoparticles on the relaxation of the water associated peak at ca. 4.8 ppm. Information on the T 1 and T 2 measurement sequences can be found in the experimental Section 2.3. From the data presented in Figure 3 and Table 2, we can conclude that the added gold has a considerable impact on the relaxation properties of the iron oxide nanoparticles. Beginning with T 1 properties, the iron oxide nanoparticles (O) have a r 1 relaxivity of 0.92 mM −1 s −1 ; this reduces with added gold to 0.24 mM −1 s −1 for the thick gold-shelled iron oxide nanoparticle (R3). In the case of with T 2 relaxation, there is also a reduction in the r 2 relaxivity, which is 32.70 mM −1 s −1 for the bare iron oxide (O) and 13.58 mM −1 s −1 for the iron oxide core with the thick gold shell (R3). Interestingly, the r 2 /r 1 ratio is minimum for the iron oxide nanoparticle with a thin gold shell (23.27) and maximum for the thick gold-shelled iron oxide nanoparticle (55.48). This large ratio for the thick gold shell could suggest that this stage is a good candidate for T 2 imaging, but it is not paired with a high r 2 value. In summary, increasing the thickness of the gold shell reduces the relaxivity (increases the relaxation time) for both T 1 and T 2 MRI imaging, hence reducing the contrasting capability.
shell, similar to observations in this research with the addition of a gold shell. Moreover, Park et al. noted a reduced r2 value for a diamagnetic avidin-biotin coating on Fe3O4 nanoparticles and attributed this to diamagnetic shielding effects [72]. Assembled magnetic structures were also seen to increase the r2 relaxation [73]. In this work, benchtop 1 H NMR spectroscopy was used to monitor the effects of the multistage magnetic-plasmonic nanoparticles on the relaxation of the water associated peak at ca. 4.8 ppm. Information on the T1 and T2 measurement sequences can be found in the experimental Section 2.3.
From the data presented in Figure 3 and Table 2, we can conclude that the added gold has a considerable impact on the relaxation properties of the iron oxide nanoparticles. Beginning with T1 properties, the iron oxide nanoparticles (O) have a r1 relaxivity of 0.92 mM −1 s −1 ; this reduces with added gold to 0.24 mM −1 s −1 for the thick gold-shelled iron oxide nanoparticle (R3). In the case of with T2 relaxation, there is also a reduction in the r2 relaxivity, which is 32.70 mM −1 s −1 for the bare iron oxide (O) and 13.58 mM −1 s −1 for the iron oxide core with the thick gold shell (R3). Interestingly, the r2/r1 ratio is minimum for the iron oxide nanoparticle with a thin gold shell (23.27) and maximum for the thick gold-shelled iron oxide nanoparticle (55.48). This large ratio for the thick gold shell could suggest that this stage is a good candidate for T2 imaging, but it is not paired with a high r2 value. In summary, increasing the thickness of the gold shell reduces the relaxivity (increases the relaxation time) for both T1 and T2 MRI imaging, hence reducing the contrasting capability.

Magnetic Hyperthermia
Hyperthermia therapy is a medical treatment that involves raising the temperature of the body in an attempt to treat diseases such as cancer. Standard hyperthermia therapy may be delivered using infrared light, microwaves, or radiofrequency. Like with other cancer treatments, there are adverse effects when non-targeted tissue is affected, which can lead to blood clots or even cardiovascular issues [77]. The temperature to achieve cell death is debated, although most agree that 40-43 • C must be reached, about 3-6 • C above healthy body temperature (37 • C) [77]. It is important to note that such therapy is not a conventional therapy, and would rarely be used as the first choice for treatment due to non-selective heating. To combat issues with non-selective heating, nanoparticle agents have been proposed as a technique for localizing the heat to targeted areas with minimal heating of the surrounding healthy tissue. One such nanoparticle agent is magnetic nanoparticles. Magnetic hyperthermia combines magnetic nanoparticles with an external high-frequency alternating magnetic field. To achieve selective heating, the nanoparticle should target the tumor site, using either passive or, preferably, active targeting techniques (using targeting ligands, magnetic field).
Typically, superparamagnetic nanoparticles are used as magnetic hyperthermia agents. Superparamagnetic nanoparticles flip to align their magnetic dipole to the external magnetic field orientation. When this magnetic field is alternating, and at a high frequency, these nanoparticles flip to orient with the field but lose power due to hysteresis, resulting in heat dissipation.
In an AC magnetic field, the magnetization of the nanoparticle may not be responsive enough to keep up with the AC frequency, resulting in a phase difference. This phase difference requires a complex magnetic susceptibility explanation. The concentration, size, and material of the nanomaterial impact the hyperthermia activity along with the field strength and frequency. Furthermore, magnetic assembly leads to higher hysteresis losses, thus increasing the heat generated [78]. A silica shell [79,80] and a polyphosphazene-gold shell [81] have also been shown to decrease magnetic hyperthermia activity.
To benchmark the heating efficiency of the magnetic-plasmonic nanoparticles, the specific power absorption (SPA) (W/g), also known as the specific absorption rate (SAR), is calculated per [82]. Furthermore, to account for the magnetic field strength and frequency used, the intrinsic loss power parameter (ILP) (H m 2 kg −1 ) is calculated per [82].
From Figure 4 and Table 1, we can deduce that the addition of gold leads to a reduction in the magnetic hyperthermia activity. As previously discussed, it is likely that with the addition of gold there is a reduction of interparticle assembly as this formation is linked closely with an increased magnetic hyperthermia activity. The heating efficiency of commercial magnetic colloids has been reported to be between 0.15 and 3.12 nH m 2 /kg Fe [12]. Therefore, the magnetic-plasmonic nanoparticles in this study are comparable until thicker gold shells are added (R2 and R3).
infrared light, microwaves, or radiofrequency. Like with other cancer treatments, there are adverse effects when non-targeted tissue is affected, which can lead to blood clots or even cardiovascular issues [77]. The temperature to achieve cell death is debated, although most agree that 40-43 °C must be reached, about 3-6 °C above healthy body temperature (37 °C) [77]. It is important to note that such therapy is not a conventional therapy, and would rarely be used as the first choice for treatment due to non-selective heating. To combat issues with non-selective heating, nanoparticle agents have been proposed as a technique for localizing the heat to targeted areas with minimal heating of the surrounding healthy tissue. One such nanoparticle agent is magnetic nanoparticles. Magnetic hyperthermia combines magnetic nanoparticles with an external high-frequency alternating magnetic field. To achieve selective heating, the nanoparticle should target the tumor site, using either passive or, preferably, active targeting techniques (using targeting ligands, magnetic field).
Typically, superparamagnetic nanoparticles are used as magnetic hyperthermia agents. Superparamagnetic nanoparticles flip to align their magnetic dipole to the external magnetic field orientation. When this magnetic field is alternating, and at a high frequency, these nanoparticles flip to orient with the field but lose power due to hysteresis, resulting in heat dissipation.
In an AC magnetic field, the magnetization of the nanoparticle may not be responsive enough to keep up with the AC frequency, resulting in a phase difference. This phase difference requires a complex magnetic susceptibility explanation. The concentration, size, and material of the nanomaterial impact the hyperthermia activity along with the field strength and frequency. Furthermore, magnetic assembly leads to higher hysteresis losses, thus increasing the heat generated [78]. A silica shell [79,80] and a polyphosphazene-gold shell [81] have also been shown to decrease magnetic hyperthermia activity.
To benchmark the heating efficiency of the magnetic-plasmonic nanoparticles, the specific power absorption (SPA) (W/g), also known as the specific absorption rate (SAR), is calculated per [82]. Furthermore, to account for the magnetic field strength and frequency used, the intrinsic loss power parameter (ILP) (H m 2 kg −1 ) is calculated per [82].
From Figure 4 and Table 1, we can deduce that the addition of gold leads to a reduction in the magnetic hyperthermia activity. As previously discussed, it is likely that with the addition of gold there is a reduction of interparticle assembly as this formation is linked closely with an increased magnetic hyperthermia activity. The heating efficiency of commercial magnetic colloids has been reported to be between 0.15 and 3.12 nH m 2 /kg Fe [12]. Therefore, the magnetic-plasmonic nanoparticles in this study are comparable until thicker gold shells are added (R2 and R3).

Discussion
In this work, the following five stages of magnetic-plasmonic nanoparticles were studied: The magnetophoresis rate of magnetic-based nanostructures is key in magnetic field-driven drug delivery, magnetic sorting, and other applications requiring magnetic manipulation. It was found that increased gold coating leads to a reduction in the magnetophoresis rate, rendering the nanoparticle less responsive to magnetic fields. The magnetophoresis speed is reduced to ca. 1% for the thickest gold shell compared to the uncoated iron oxide nanoparticle (see Table 1 for further details). The magnetophoresis speed decreases exponentially (see Figure 5) with increased gold shell thickness; thus, if responsivity is required for the application, a thinner gold coating is preferable. We attribute this reduced responsivity of the increased gold shell thickness to a reduced magnetic saturation, loss of cooperative behavior, and increased drag forces. saturation, loss of cooperative behavior, and increased drag forces.
Iron oxide nanoparticles are commonly used as MRI contrast agents, specifically as T2 agents. In this work, the T1 and T2 relaxivities (r1 and r2) of the various magnetic-plasmonic nanostructures were measured. As with the magnetophoresis, the T1 and T2 relaxivities, r1 and r2, decreased exponentially with additional gold growth (see Table 1). Compared to the purely magnetic counterpart, the thickest gold shell has a reduced relaxivity, to 22.2% and 41.6% for r1 and r2, respectively. These relaxivities also decreased exponentially (see Figure 5) with increased gold shell thickness (see Figure 5). We attribute this decreased relaxivity of thicker gold shells to reduced magnetic saturation, loss of cooperative behavior, and reduced exchange interaction with water nuclei.
Lastly, increasing the gold content of the magnetic-plasmonic hybrid nanoparticles led to an exponential decrease (see Figure 5) in the specific power absorption of the nanoparticles undergoing magnetic hyperthermia. A reduction to 14.3% was seen for the intrinsic loss power parameter, a measurement of heating efficiency that considers magnetic field strength and frequency, for the thickest gold shell on the magnetic core compared to the purely magnetic nanoparticle. This reduced heating efficiency with gold addition is likely a result of reduced cooperative behavior along with reduced magnetic saturation. Figure 5. Exponential dependency of the normalized NMR relaxivities r1 and r2, the specific power absorption (SPA), and the magnetophoresis speed (vmag) as a function of gold shell thickness. See Table 1 for values. Figure 5. Exponential dependency of the normalized NMR relaxivities r 1 and r 2 , the specific power absorption (SPA), and the magnetophoresis speed (v mag ) as a function of gold shell thickness. See Table 1 for values.
Iron oxide nanoparticles are commonly used as MRI contrast agents, specifically as T 2 agents. In this work, the T 1 and T 2 relaxivities (r 1 and r 2 ) of the various magnetic-plasmonic nanostructures were measured. As with the magnetophoresis, the T 1 and T 2 relaxivities, r 1 and r 2 , decreased exponentially with additional gold growth (see Table 1). Compared to the purely magnetic counterpart, the thickest gold shell has a reduced relaxivity, to 22.2% and 41.6% for r 1 and r 2 , respectively. These relaxivities also decreased exponentially (see Figure 5) with increased gold shell thickness (see Figure 5). We attribute this decreased relaxivity of thicker gold shells to reduced magnetic saturation, loss of cooperative behavior, and reduced exchange interaction with water nuclei.
Lastly, increasing the gold content of the magnetic-plasmonic hybrid nanoparticles led to an exponential decrease (see Figure 5) in the specific power absorption of the nanoparticles undergoing magnetic hyperthermia. A reduction to 14.3% was seen for the intrinsic loss power parameter, a measurement of heating efficiency that considers magnetic field strength and frequency, for the thickest gold shell on the magnetic core compared to the purely magnetic nanoparticle. This reduced heating efficiency with gold addition is likely a result of reduced cooperative behavior along with reduced magnetic saturation.

Conclusions
In conclusion, increasing the gold shell thickness has a significant reduction in the magnetophoresis rate, T 1 and T 2 contrast, and magnetic hyperthermia activity of iron oxide nanoparticles. A number of factors are likely responsible for this reduced activity with increased gold shell thickness. We hypothesize that less cooperation due to reduced dipole-dipole magnetic interaction forces and a reduced magnetic saturation owing to the decreased magnetic volume and possible diamagnetic magnetic shielding by the gold shell are predominantly responsible.
In a previous work ( [28]), we demonstrated that the thinner gold shells (R1 and R2) exhibit a large spectral drift between the optical absorption (highest photothermal activity) and the optical scattering (strongest optical imaging contrast). We demonstrate in this work that these thinner gold shells are also better for magnetic-based applications. Hence, this work further emphasizes that a thick gold shell is not necessary to achieve a highly functional nanotheranostic nanoplatform. In other words, thinner gold coating appears preferable in an extending series of both magnetic-and light-based applications.