High-Throughput Thermal Decomposition Synthesis for Optimizing Magnetic Particle Imaging Tracers Study Guide

Research Context & Overview

• Project Title: High-throughput thermal decomposition synthesis for optimizing magnetic particle imaging tracers
• Institutional Affiliation: Department of Chemical Engineering, University of Florida
• Author: Eric Imhoff, Ph.D. Student, Rinaldi-Ramos Lab, Department of Chemical Engineering, University of Florida
• Proposal Date: April 18, 2024 ($04/18/24$)
• Core Nanomaterial Utility: Magnetic nanoparticles exhibit unique physical and magnetic properties due to their small physical dimensions and high surface area-to-volume ratio. These properties can be engineered and tuned for specialized performance across technological and medical applications.
• Biomedical Applications: Key fields of application include Magnetic Resonance Imaging (MRI), Magnetic Fluid Hyperthermia (MFH), magnetic drug targeting, magnetically triggered drug release, and Magnetic Particle Imaging (MPI).
• Synthesis Paradigms & Deficiencies:
  • Conventional synthesis efforts focus primarily on controlling physical diameter, shape, and size distribution monodispersity, while frequently neglecting magnetic properties.
  • Standard thermal decomposition under strictly anoxic conditions often results in phase-impure iron oxides containing nonmagnetic phases (such as wustite, $FeO$) or crystal structural defects, leading to suboptimal magnetic performance.
  • Prior work in the Rinaldi-Ramos Lab established a thermal decomposition method utilizing in-situ oxygen oxidation, producing phase-pure iron oxide nanoparticles with coinciding physical and magnetic diameters of approximately $20\text{--}22\text{ nm}$ (designated as tracer RL-1) that show strong MPI performance.
  • Alternative approaches, such as post-synthesis oxidation or non-aqueous redox solution tuning, have been demonstrated to improve phase purity and magnetic properties, but these synthesis modifications remain underutilized in the field.
• MPI Tracer Optimization Opportunity: Theoretical modeling and experimental data demonstrate that optimal iron oxide MPI tracer performance requires physical core diameters in the range of $25\text{--}35\text{ nm}$, which exceeds the size threshold reliably produced by standard thermal decomposition techniques.
• High-Throughput (HT) Platform Goals:
  • Address widespread reproducibility and reporting limitations in nanoparticle literature (where conclusions are frequently drawn from single-batch syntheses without statistical analysis).
  • Develop a parallelized high-throughput batch thermal decomposition reactor to systematically map the reaction parameter space for iron oxide and substituted ferrites (e.g., manganese ferrite, zinc ferrite).
  • Elucidate statistical correlations between synthesis parameters, physical/crystalline properties, and magnetic performance.
  • Utilize empirical parameter correlations to guide the continuous synthesis of optimized MPI tracers in a scalable millifluidic flow reactor.

Fundamental Physics & Theoretical Models

• Composition and Spin Structure:
  • Common iron oxide nanoparticles consist of magnetite ($Fe_3O_4$) or maghemite (\text{\gamma-}Fe_2O_3), both sharing an inverse-spinel crystal lattice.
  • Substituted ferrites follow the chemical formula $MFe_2O_4$, where $M$ represents a divalent transition metal ion such as $Mn^{2+}$, Zn^{2+}$, Co^{2+}$, or $Ni^{2+}$.
  • While bulk forms of these materials exhibit permanent ferrimagnetic or ferromagnetic ordering, domain size reduction below a critical threshold yields superparamagnetism.
• Superparamagnetism Dynamics:
  • Superparamagnetic nanoparticles possess a large single magnetic dipole. Thermal energy ($k_B T$) randomly reorients this dipole, resulting in zero net magnetization ($M = 0$) in the absence of an external magnetic field.
  • Application of an external field induces rapid dipole alignment, yielding high magnetic susceptibility and saturation at low field strengths without remanence or coercivity.
• Langevin Model for Superparamagnetism:
  • Assumes non-interacting, uniform spherical particles responding instantaneously to an applied magnetic field $H$.
  • Ensemble magnetization equation:         $M(H) = N m \mathcal{L}(k H)$
  • Dimensionless field parameter:         $k = \frac{\mu_0 m}{k_B T}$
  • Magnetic moment of an ideal spherical particle:         $m = \frac{M_{sat} \pi D^3}{6}$
  • Langevin function definition:         $\mathcal{L}(x) = \coth(x) - \frac{1}{x}$
  • Variables: $N$ is the total number of particles, $\mu_0$ is the permeability of free space ($4\pi \times 10^{-7}\,\text{T\,m\,A}^{-1}$), $k_B$ is the Boltzmann constant ($1.3806 \times 10^{-23}\,\text{J\,K}^{-1}$), $T$ is absolute temperature ($\text{K}$), $M_{sat}$ is bulk saturation magnetization ($\text{A\,m}^2\,\text{kg}^{-1}\text{ Fe}$ or $\text{A\,m}^{-1}$), and $D$ is the physical diameter of the nanoparticle core ($\text{m}$).
• Saturation Magnetization ($M_{sat}$):
  • Represents maximum achievable magnetization when all internal magnetic dipoles align completely parallel to the applied field.
  • Bulk magnetite saturation magnetization: $92\,\text{A\,m}^2\,\text{kg}^{-1}\text{ Fe}$.
  • Bulk maghemite saturation magnetization: $74\,\text{A\,m}^2\,\text{kg}^{-1}\text{ Fe}$.
  • Nanoparticle $M_{sat}$ reductions stem from surface defect states ("magnetically dead layer" composed of wustite, $FeO$), disordered mixed phase composition, or competing internal magnetic domain boundaries.
• Magnetic Diameter ($D_m$) Model:
  • Used to account for deviations from ideal Langevin behavior caused by surface dead layers, mixed iron oxide phases, or dipolar interactions.
  • Calculated by fitting experimental magnetization curves to a lognormal volume-weighted distribution of ideal Langevin particles:         $M(\alpha) = M_{sat} \int_0^\infty n_v(D_m) \mathcal{L}(\alpha)\,d D_m$$n_v(D_m) = \frac{1}{\sqrt{2\pi} D_m \ln(\sigma)} \exp\left(-\frac{\ln^2(D_m / D_{mv})}{2 \ln^2(\sigma)}\right)$$\alpha = \frac{\pi \mu_0 M_d D_m^3 H}{6 k_B T}$
  • Variables: $D_{mv}$ is the volume-weighted median magnetic diameter, and $\ln(\sigma)$ is the geometric standard deviation of the distribution. Discrepancies between physical diameter ($D$) and magnetic diameter ($D_m$) quantify structural or magnetic defects.
• Magnetic Relaxation Mechanisms:
  • Néel Relaxation: Internal rotation of the particle magnetic dipole relative to its crystal lattice axes:         $\tau_N = \tau_0 \exp\left(\frac{K V_C}{k_B T}\right)$         where attempt time $\tau_0 \approx 10^{-9}\,\text{s}$, $K$ is the effective anisotropy constant ($\text{J\,m}^{-3}$), and $V_C$ is the magnetic core volume ($\text{m}^3$).
  • Brownian Relaxation: Physical rotation of the entire nanoparticle inside its fluid medium:         $\tau_B = \frac{3 \eta V_H}{k_B T}$         where $\eta$ is the dynamic viscosity of the medium ($\text{Pa\,s}$) and $V_H$ is the hydrodynamic volume of the particle ($\text{m}^3$).
  • Anisotropy Sources: Magnetocrystalline anisotropy (crystallographic easy axis magnetization) and shape anisotropy (demagnetizing fields resulting from non-spherical particle geometries).
• Magnetic Particle Imaging (MPI) Principles:
  • First introduced by Gleich and Weizenecker in 2005. Images spatial distribution of superparamagnetic tracers by exploiting their non-linear magnetization response.
  • Excitation & Selection Fields: An Alternating Magnetic Field (AMF; typical amplitude $\sim 10\,\text{mT}$, frequency range $1\text{--}100\,\text{kHz}$) drives tracer magnetization into the non-linear regime. A strong static magnetic gradient field creates a Field-Free Region (FFR). Particles outside the FFR are saturated by the static field and yield no AC signal; only particles within the FFR respond to the AMF, generating an induced voltage in pickup coils.
  • Point Spread Function (PSF): Performance is governed by signal intensity ($I$) and spatial resolution (Full Width at Half Maximum, FWHM):         $I = \frac{N \pi M_{sat} D^3}{18}$$\text{FWHM} = \frac{24 k_B T}{\mu_0 \pi M_{sat} G D^3}$         where $G$ is the selection field magnetic gradient ($\text{T\,m}^{-1}$ or $\text{T\,m}^{-1}\,\mu_0^{-1}$).
  • The Relaxation Wall: Beyond physical core sizes of $25\text{--}35\text{ nm}$, relaxation time constants (Néel and Brownian) exceed the AMF oscillation frequency, introducing phase lag, diminishing signal intensity, and degrading spatial resolution.
  • Superferromagnetism: Strongly interacting linear nanoparticle chains (Tay et al.) respond superferromagnetically, providing order-of-magnitude increases in MPI intensity and spatial resolution.
  • Scanner Scale Challenges: Preclinical animal scanners operate at high gradients ($\sim 7\,\text{T\,m}^{-1}$), while human-scale clinical prototypes operate at lower gradients ($< 1\,\text{T\,m}^{-1}$), necessitating higher performing tracers to retain resolution.

Synthesis Pathways & Literature Limitations

• Thermal Decomposition Chemistry:
  • Involves breaking down organometallic precursors (e.g., iron oleate) in high-boiling organic solvents (e.g., 1-octadecene, squalane) with stabilizing surfactants (e.g., oleic acid) at temperatures ranging from $300\,^\circ\text{C}$ to $350\,^\circ\text{C}$ under inert atmosphere.
  • LaMer Nucleation and Growth Mechanism:
    • Stage I: Precursor thermally decomposes into monomer species; monomer concentration exceeds saturation.
    • Stage II: Monomer concentration reaches a critical supersaturation threshold, inducing rapid, self-limiting burst nucleation.
    • Stage III & IV: Monomer concentration falls below the nucleation threshold; remaining monomers deposit onto existing nuclei for controlled isotropic particle growth.
    • Ostwald Ripening: Over extended time scales, smaller high-energy particles dissolve to feed larger particles.
• Phase & Phase-Purity Tuning Methods:
  • Post-Synthesis Oxidation: Bubbling $O_2$ gas through hot nanoparticle suspensions for extended durations (e.g., LodeSpin Labs tracers yielding $1\,\text{mm}$ resolution). Processing time increases with core size.
  • In-Situ Oxidation: Adding stoichiometric oxygen during initial synthesis (Unni et al.), yielding phase-pure particles without secondary treatment steps.
  • Solution Redox Tuning: Utilizing redox-active solvent reactions without gaseous oxygen. 1-octadecene can undergo thermal reductive cleavage or oxidative decarboxylation (verified via GC-MS); addition of radical-generating solvents like benzyl ether eliminates structural crystal defects (Chen et al.).
• Gaps in Existing Literature:
  • Extreme parameter sensitivity: Minor shifts in ramp rate, peak temperature, surfactant ratios, or trace solvent impurities alter outcome metrics.
  • Single-batch evaluation: Most studies report properties from single synthesis runs, failing to assess reproducibility or batch-to-batch variance.
  • Kemp et al. statistical evaluation (>200 syntheses): Examined pre-synthesis vacuum treatment and surfactant ratios; observed a large variance ($\sim 5\,\text{nm}$ diameter spread for identical target conditions) and presented limited statistical modeling ($R^2$ values only).
  • Velazquez-Albino et al. evaluation: Parallelized batch thermal decomposition system evaluated physical and magnetic metrics using Pearson correlation matrices, but omitted continuous per-vessel temperature tracking.

Specific Aims

• Aim 1: Design and validate a high-throughput (HT) reactor system for thermal decomposition synthesis of magnetic nanoparticles.
• Aim 2: Characterize the reproducibility and control of magnetic nanoparticles synthesized in the HT platform.
• Aim 3: Identify reaction conditions that lead to optimal performance of iron oxide and manganese ferrite nanoparticles in MPI.
• Aim 4: Develop a millifluidic flow synthesis platform for the investigation of scalable synthesis of magnetic nanoparticles with optimized properties.

Preliminary Work & Experimental Observations

• HT Prototype Reactor Engineering & Deficiencies:
  • Initial System: High-temperature tube block ($400\,^\circ\text{C}$ limit) accommodating parallel test tubes under inert gas. Lacked active mixing and individual vessel temperature monitoring.
  • Prototype Modifications: Designed a 3D-printed 8-vessel tube holder array using high-temperature resin. Incorporated mechanical impeller drive motors, gas delivery lines, and thermocouple positioning guides.
  • Initial Validation Run: 8 distinct particle batches characterized via Transmission Electron Microscopy (TEM). A one-way ANOVA demonstrated significant batch-to-batch variance in mean physical diameter and polydispersity ($p < 0.001$).
  • Root Cause Identification: Poor thermal contact between glass test tubes and the block body generated a standard deviation of $SD \approx 3\,^\circ\text{C}$ across vessels at peak temperature (compared to $SD \approx 0.8\,^\circ\text{C}$ in a molten metal bath control). Infrared imaging revealed central block wells were hotter than perimeter wells. Solvent vapor condensation on the holder indicated loss of volatile species.
• Redesigned Custom HT Heating System:
  • Reflux Length Optimization: Reflux experiments in $30\,\text{cm}$ tubes indicated solvent condensation occurs at a maximum height of $10\text{--}11\,\text{cm}$. Test tube length was increased from $12\,\text{cm}$ to $15\,\text{cm}$ to ensure complete internal condensation and zero solvent loss.
  • Custom Block Architecture: Machined aluminum block holding sixteen $18\,\text{mm}$ diameter test tubes, powered by 8 distributed electric cartridge heaters ($41\,\text{W\,in}^{-2}$ heat flux rating) aligned axially with the reaction fluid height. Wrapped in 1-inch thick ceramic wool insulation.
  • COMSOL Multiphysics Modeling: Transient finite-element thermal simulations using oleic acid fluid properties demonstrated a heating ramp rate of $\sim 12\,^\circ\text{C\,min}^{-1}$ up to $350\,^\circ\text{C}$ (surpassing the required $5\,^\circ\text{C\,min}^{-1}$ threshold) and symmetric, uniform planar temperature profiles across all 16 sample wells at $t = 30\,\text{min}$.
• Pre-Reaction Vacuum Degassing Effects:
  • Procedure: Pre-reaction components (iron oleate, oleic acid, solvents) were subjected to a 30-minute vacuum degassing step until bubbling ceased, stripping dissolved oxygen, water, and volatile organics.
  • Statistical Comparison (12 Standard vs. 4 Degassed Batches):
    • An F-test confirmed significantly lower variance in physical diameter ($p = 0.029$) and polydispersity ($p = 0.083$) in the vacuum-treated group.
    • A two-sample t-test indicated a significant shift in mean physical diameter ($p = 0.098$) with no change in mean polydispersity ($p = 0.358$).
  • Extended Heated Degassing (20 Hours): System pressure plateaued at several Torr (vs. $< 1\,\text{Torr}$ blank). Particle synthesis following 20-hour degassing produced larger nanoparticles ($22\,\text{nm}$ diameter) after standard 4-hour reaction times.
  • Precursor Structural Analysis: SQUID magnetometry of untreated and 30-minute degassed iron oleate precursors revealed superparamagnetic behavior with an unexpected magnetic diameter $D_m = 10\,\text{nm}$, confirmed by TEM as premature nanoparticle nucleation in the precursor stock. The 20-hour vacuum-treated precursor exhibited purely paramagnetic behavior with zero premature nuclei.
  • Mechanism: Removal of volatile impurities and water suppresses local inhomogeneities, suppressing premature nucleation events. Fewer initial nuclei allow a higher monomer-to-nucleus ratio, yielding larger final particle core sizes.
• Nanoparticle Morphology Control:
  • Anoxic thermal decomposition syntheses produced smooth, highly spherical nanoparticles.
  • In-situ oxygen syntheses produced faceted nanoparticles.
  • Anoxic synthesis followed by post-synthesis oxidation preserved the spherical morphology.
  • Faceted geometry increases shape anisotropy, accelerating Néel relaxation. Post-synthesis oxidation of spherical particles minimizes shape anisotropy, enabling larger magnetic core growth before encountering the relaxation wall.

Comprehensive Safety Assessment

• Thermal Hazards:
  • Reaction operating temperatures: $300\text{--}350\,^\circ\text{C}$.
  • Personal Protective Equipment (PPE): Nonflammable lab coats, safety eyewear, heat-resistant gloves.
  • Handling: Manipulations performed at lower temperatures; boiling stones or enclosed array holders prevent thermal bumping.
• Combustion & Fire Hazards:
  • Solvent heating exceeds vapor flash points. Reactions are blanketed with inert gas (argon/nitrogen).
  • Headspace Oxygen Control: Kept below the Limiting Oxygen Concentration (LOC). LOC threshold target set below hydrogen's LOC ($5\%$ at STP), well below organic solvent limits (e.g., hexane $\approx 10\%$ LOC). Mass flow controllers combine $20\%$ $O_2$ in Ar with pure Ar to deliver a stoichiometric $1\%$ $O_2$ stream.
• Mechanical & Machining Hazards:
  • Use of impact-resistant safety glasses during cutting, drilling, and sanding of platform structural elements.
• Electrical Hazards:
  • Custom heat block circuits operate at $120\,\text{V\,AC}$ at multi-ampere currents.
  • All wiring work performed on unpowered circuits containing zero energy-storage devices (capacitors).
  • Enclosure of conductive components in grounded, insulated housings.
  • Ground Fault Circuit Interrupter (GFCI) protection trips at differential currents of $4\text{--}6\,\text{mA}$ within $\sim 1/40\,\text{s}$. Overcurrent protection via a $10\,\text{A}$ circuit breaker and inline thermal fuses.

Proposed Methodology & Aims Execution

Specific Aim 1: Design and Validate HT Reactor System

• Mechanical & Hardware Construction:
  • Fabrication of impellers via CNC milling to lower surface roughness, reduce fluid friction, and fit standard ball bearings for smooth agitation at lower motor power.
  • Hardware Interface: National Instruments (NI) USB-6001 Data Acquisition (DAQ) module connected to a PC running a custom LabVIEW PID control software loop. The PID signal controls Solid State Relays (SSRs) operating resistive heaters in parallel.
  • Per-Vessel Temperature Logging: High-capacity NI DAQ thermocouple module logging 16 individual vessel temperatures simultaneously.
• Calibration & Validation:
  • Thermocouple calibration at low temperatures (ice water $0\,^\circ\text{C}$, boiling water $100\,^\circ\text{C}$) and high temperatures (refluxing squalane $350\,^\circ\text{C}$).
  • PID tuning via the Ziegler-Nichols closed-loop method. Thermal uniformity across the block body verified via surface thermocouples, IR pyrometers, and thermal imaging cameras.
• Contingencies:
  • If large block thermal inertia induces control lag, continuous autotuning or fuzzy logic algorithms will replace simple PID logic.
  • If lateral thermal gradients emerge, multi-zone independent temperature control will be implemented using cartridge heaters with integrated internal thermocouples.

Specific Aim 2: Characterize Reproducibility and Control in HT Platform

• Analytical Characterization Suite:
  • TEM: Physical diameter ($D$), polydispersity ($\ln(\sigma)$), aspect ratio, circularity.
  • XRD: Phase identification and crystallite domain size via Scherrer equation:         $\tau = \frac{K \lambda}{\beta \cos\theta}$
  • SQUID Magnetometry: Saturation magnetization ($M_{sat}$), volume-weighted median magnetic diameter ($D_{mv}$), and geometric deviation ($\ln(\sigma)$).
  • Precursor Analytical Characterization: Fourier-Transform Infrared Spectroscopy (FTIR) to quantify free oleic acid percentage, permanganate titrations for $Fe^{2+}/Fe^{3+}$ ratios, Thermogravimetric Analysis (TGA), and SQUID magnetometry.
• Statistical Modeling:
  • 4-way Analysis of Covariance (ANCOVA) on a main-effects model evaluating physical diameter, magnetic diameter, and polydispersity against continuous and categorical factors:
    1. Average reaction temperature (continuous)
    2. Temperature ramp rate (continuous)
    3. Percentage of free oleic acid in precursor (continuous)
    4. Precursor batch ID (categorical across $\ge 3$ independent precursor synthesis runs)
• Workflow Throughput & Bottleneck Mitigation:
  • SQUID magnetometry requires $\sim 2\,\text{hours/sample}$ ($4\text{ days}$ per 16-sample batch), establishing a baseline throughput of $16\,\text{samples/week}$ (an 8-fold increase over the conventional 2-batch/week rate).
  • High-Throughput Magnetometry Alternative: Vibrating Sample Magnetometry (VSM) on solid/immobilized samples (dried powders or embedded in polystyrene-divinylbenzene / cotton matrix) to accelerate magnetic screening.

Specific Aim 3: Identify Conditions for Optimal MPI Performance

• Design of Experiments (DoE) Parameter Space:

• Factorial Exploration Logic:
  • A full 5-factor, 3-level factorial design in triplicate requires $3^5 \times 3 = 729$ syntheses ($\sim 45$ 16-vessel runs, $\sim 1\text{ year}$ duration).
  • Initial screening will vary factors independently relative to baseline controls, requiring 18 syntheses (2 16-vessel runs) to evaluate individual main effect sizes before running targeted partial factorial designs.
• MPI Screening & Characterization:
  • Tracer screening executed via Relaxation Measurement (RELAX) spectroscopy and 2D spatial point-source scans to quantify PSF peak height (Intensity $I$) and FWHM (Resolution).
  • Parallelized phase-transfer ligand exchange routines (up to 24 samples) shift the workflow bottleneck to TEM imaging ($1\text{--}2\text{ days}$ per batch).
• Substituted Ferrite Optimization ($MnFe_2O_4$, $ZnFe_2O_4$):
  • Rationale: Manganese ferrite possesses lower magnetocrystalline anisotropy ($K = 3\,\text{kJ\,m}^{-3}$ vs. $11\,\text{kJ\,m}^{-3}$ for magnetite) and higher saturation magnetization ($M_{sat} = 121\,\text{A\,m}^2\,\text{kg}^{-1}$ vs. $92\,\text{A\,m}^2\,\text{kg}^{-1}$ for magnetite), theoretically shifting the relaxation wall to larger core sizes.
  • Synthesis: Substituted precursor ratios ($Mn:Fe$ or $Zn:Fe$) varied from $1:3$ to $1:1$.
  • Elemental Characterization: Energy-Dispersive X-ray Spectroscopy (EDS) mapping in TEM to verify elemental distribution and rule out core-shell phase separation; Inductively Coupled Plasma Mass Spectrometry (ICP-MS) or Optical Emission Spectroscopy (ICP-OES) to quantify bulk elemental stoichiometry.

Specific Aim 4: Continuous Millifluidic Flow Synthesis Platform

• Flow Reactor Design:
  • Machined aluminum block modules with shallow channels accommodating millimeter-scale internal diameter stainless steel tubing.
  • Heated via embedded cartridge or plate heaters with thermocouple feedback; enclosed by an upper insulation retention plate.
  • Precursor fluids pumped via syringe pumps; segmented plug-flow maintained by injecting discrete $N_2$ gas bubbles via mass flow controllers.
  • Multi-block series configurations permit independent heating zones to recreate batch thermal ramp profiles.
• Fouling & Operational Contingencies:
  • Tubing interior treated with SilcoNert chemical vapor deposition (CVD) coatings to minimize particle adhesion; periodic acid/base cleaning flushes implemented.
  • 1-octadecene replacement with squalane to prevent high-temperature thermal polymerization/clogging (adjusting for lost 1-octadecene redox activity).
  • Multi-stage downstream precursor injection to grow core diameters beyond single-pass size limits ($> 17\text{ nm}$).

Summary & Broad Impact

• Establishes a high-throughput thermal decomposition platform that systematically links chemical synthesis parameters to fine magnetic particle properties.
• Provides systematic datasets suitable for machine learning algorithms to predict nanoparticle performance from synthesis conditions.
• Provides scalable synthesis pathways for optimized iron oxide and substituted ferrite tracers essential for human-scale clinical MPI translation.
• Offers transferable synthesis parameters for broader applications requiring strict magnetic property engineering, such as Magnetic Fluid Hyperthermia (MFH) and targeted drug delivery.