Optimal Design and Characterization of Superparamagnetic Iron

14470

J. Phys. Chem. B 2008, 112, 14470–14481

Optimal Design and Characterization of Superparamagnetic Iron Oxide Nanoparticles Coated with Polyvinyl Alcohol for Targeted Delivery and Imaging† Morteza Mahmoudi,‡ Abdolreza Simchi,‡,§ Mohammad Imani,| Abbas S. Milani,⊥ and Pieter Stroeve*,# Institute for Nanoscience and Nanotechnology, Sharif UniVersity of Technology, Tehran, Iran, Department of Materials Science and Engineering, Sharif UniVersity of Technology, Tehran, Iran, NoVel Drug DeliVery Systems Department, Iran Polymer and Petrochemical Institute, Tehran, Iran, School of Engineering, UniVersity of British Columbia Okanagan, Kelowna, Canada, and Department of Chemical Engineering and Materials Science, UniVersity of California DaVis, DaVis, CA ReceiVed: April 7, 2008; ReVised Manuscript ReceiVed: May 23, 2008

Superparamagnetic iron oxide nanoparticles (SPION) with narrow size distribution and stabilized by polyvinyl alcohol (PVA) were synthesized. The particles were prepared by a coprecipitation technique using ferric and ferrous salts with a molar Fe3+/Fe2+ ratio of 2. Using a design of experiments (DOE) approach, the effect of different synthesis parameters (stirring rate and base molarity) on the structure, morphology, saturation magnetization, purity, size, and size distribution of the synthesized magnetite nanoparticles was studied by various analysis techniques including X-ray powder diffraction (XRD), thermogravimetric analysis (TGA) with differential scanning calorimetry (DSC) measurements, vibrating-sample magnetometer (VSM), transmission electron microscopy (TEM), UV-visible, and Fourier transform infrared (FT-IR) spectrometer. PVA not only stabilized the colloid but also played a role in preventing further growth of SPION followed by the formation of large agglomerates by chemisorption on the surface of particles. A rich behavior in particle size, particle formation, and super paramagnetic properties is observed as a function of molarity and stirring conditions. The particle size and the magnetic properties as well as particle shape and aggregation (individual nanoparticles, magnetic beads, and magnetite colloidal nanocrystal clusters (CNCs)) are found to be influenced by changes in the stirring rate and the base molarity. The formation of magnetic beads results in a decrease in the saturation magnetization, while CNCs lead to an increase in saturation magnetization. On the basis of the DOE methodology and the resulting 3-D response surfaces for particle size and magnetic properties, it is shown that optimum regions for stirring rate and molarity can be obtained to achieve coated SPION with desirable size, purity, magnetization, and shape. 1. Introduction Due to their ultrafine size, biocompatibility, and superparamagnetic properties, iron oxide nanoparticles are emerging as promising candidates for various biomedical applications such as enhanced resolution magnetic resonance imaging, drug delivery, tissue repair, cell and tissue targeting and transfection, etc. Especially for in vivo applications, such as drug delivery, superparamagnetism is essential as an activation mechanism because once the external magnetic field is removed, the magnetization disappears, and thus the agglomeration, and hence the possible embolization of the capillary vessels can be avoided.1–6 However, two major shortcomings encountered in the in vivo application of these particles include their destabilization due to the adsorption of plasma proteins and the nonspecific uptake by the reticulum-endothelial system (RES).7,8 Due to their high specific surface area of the nanosized particles, plasma proteins interact with the particles which can cause an † Part of the “Janos H. Fendler Memorial Issue”. * Corresponding author. E-mail: [email protected]. ‡ Institute for Nanoscience and Nanotechnology, Sharif University of Technology. § Department of Materials Science and Engineering, Sharif University of Technology. | Iran Polymer and Petrochemical Institute. ⊥ University of British Columbia Okanagan. # University of California Davis.

Figure 1. U18(62) design layout used in the experiments. The molarity is in moles per liter, M, and the stirring rate is given in revolutions per minute (rpm).

increase in the particle size and often results in agglomeration. The particles are also considered as an intruder by the innate immune systems and can be readily recognized and engulfed by macrophage cells that may cause agglomeration. In both cases, the particles will be removed from the blood circulation which will yield a decrease in their effectiveness, leading to a reduction in efficiency of nanoparticle-based diagnostics and therapeutics. To inhibit both phenomena and provide longer circulation times, the particles are usually coated with hydro-

10.1021/jp803016n CCC: $40.75  2008 American Chemical Society Published on Web 08/26/2008

Optimal Design and Characterization of SPION

J. Phys. Chem. B, Vol. 112, No. 46, 2008 14471

Figure 2. Variation of TEM results for three different molarities given at constant stirring rate, and vice versa. Size bars are far left 25 nm, middle top 16 nm, middle center 20 nm, middle bottom 20 nm, and far right 80 nm.

Figure 3. TEM images of (a) rod nanoparticles (S5400M1.3), bar size is 80 nm, (b) CNCs (S5400M1.3), bar size is 80 nm, and (c) CNCs (S9000M1.2), bar size is 20 nm. “S” refers to the stirring rate in rpm, and “M” refers to the value of base molarity.

philic and biocompatible polymers/molecules, such as polyethylene glycol (PEG), dextran, polyvinyl alcohol (PVA), poly(acrylic acid), poly(lactide-co-glycolide) (PLGA), chitosan, pullulan, and poly(ethyleneimine) (PEI).9,10 Polyvinyl alcohol has excellent film forming, emulsifying, and adhesive properties. Coating of particle surfaces with PVA prevents their agglomeration, giving rise to monodisperse particles.11 During the past decade, mainly two types of iron oxide nanoparticles including magnetite and maghemite have been used for various biomedical applications. Of the two, magnetite is a very promising candidate because of its proven biocompatibility.12 Since the particles with various sizes exhibit different flow rates in the same environment (same capillary size), it is essential to use particles of a desirable size for targeted drug delivery and/or imaging. Particles of different sizes may be exposed to different viscosities and behave differently, particularly with regard to their velocities as they move through

capillaries. As a result, severe agglomeration may occur with particles of nonoptimized size for desired sites.13 It has been recognized that internalization of particles as well as take-up by specific cells depends strongly on the size of the magnetic particles.14 A particle is a composite heterogeneous material comprised of a magnetic inner core with a characteristic “core” size and modifying outer coating. The individual nanoparticles and the nanoparticle agglomerates can be characterized both by the hydrodynamic diameter and by the magnetic core size. Both parameters are very important for targeting purposes. The first parameter is responsible for the “magnetic response” in applied inhomogeneous magnetic fields, and the second parameter is important for targeting and cell interactions.15 A coating layer should prevent the agglomeration of the particles, increase the circulation time, and provide biocompatibility. Therefore, monodispersed particles with a high saturation magnetization

14472

J. Phys. Chem. B, Vol. 112, No. 46, 2008

Mahmoudi et al. It should be added that different types of problems can be realized during DOE practice.18,19 For the first type, one may be interested in a screening procedure in which a small number of factors (called “main effects”) are extracted from a larger pool of factors. For the second type, one may aim at finding a description of how factors affect the response (the input-output relation). Eventually, using such a relation, the goal can be to optimize the response surface function. The third type is when the experiments are tuned to give an estimation of testing errors (i.e., the robustness of the solution is of interest rather than its optimality). The fourth type relies on obtaining a mathematical model for the input-output relation and also to estimate the typical size and structure of errors. The present study intends to show such an application for the synthesis of superparamagnetic iron oxide nanoparticles (SPION) and the stabilization by nanoencapsulation using PVA. Specifically, a uniform design (UD) of experiments, with a type-II objective (i.e., finding a description of how the input factors affect the output response and optimizing the response), is adapted to study the effects of stirring rate and chemical concentration of the synthesis media on particle size and distribution, the magnetic properties, as well as the morphology of the synthesized particles. More details of the design are presented in the next section.

Figure 4. Formation of (a) magnetic beads and (b) CNCs due to the presence of PVA.

and functionalized with suitable coatings are required for targeting and imaging in hyperthermia, transfection, and MRI.15 The aim of the present work is to study the effect of the synthesis parameters on the particle size, purity, shape and configuration (individual particles and magnetic beads, i.e., random dispersion of iron oxide nanoparticles in polymeric beads due to film formation of polyvinyl alcohol and magnetite colloidal nanocrystal clusters), and magnetic saturation for targeting drug delivery, hyperthermia, and/or imaging. In this work, magnetite nanoparticles with specific size and saturation magnetization are synthesized. Since blood flow is different in various capillaries, specific sizes and saturation magnetizations are needed to prevent particle agglomeration as well as particles withstanding the drag of blood flow.16 The nanoparticles considered in this work were coated by polyvinyl alcohol (PVA) to prevent coagulation of particles during bioapplications.17 Design of Experiments (DOE) Approach. To reduce the large number of experiments due to a number of synthesis parameters, or factors, a design of experiments (DOE) approach18,19 is used here. The application of DOE methods in optimization of nanomaterials processing is beneficial, particularly given the high cost of experimentation and complex random error structures.

2. Materials and Methods Polyvinyl alcohol (Mw ) 30 000-40 000) was purchased from Fluka. Iron chloride and sodium hydroxide (NaOH) of analytical grades were supplied by Merck Inc. (Darmstadt, Germany) and used without further purification. Solutions were prepared using deionized (DI) water after 30 min of bubbling with argon for deoxygenation. The iron salts were dissolved in DI water containing 0.5 M HCL where the mole fraction of Fe2+ to Fe3+ was adjusted to 2:1 for all samples. The precipitation was performed by dropwise addition of iron salt solutions to NaOH solutions under an argon atmosphere. In order to control mass transfer, which may allow particles to combine and build larger polycrystalline particles, turbulent flow was created by placing the reaction flask in an ultrasonic bath and changing the homogenization rates between 3600 and 9000 rpm (in the first 2 min of the reaction). Various molarities of the NaOH solution were also examined. After 30 min, a PVA solution was added by syringe as a stabilizer, and the reaction mixture was stirred at a constant temperature of 35 °C for an additional 30 min at 3600 rpm. The particles were collected by centrifugation at 6000 rpm for 10 min and redispersed in DI water, and the ferrofluid was kept at 4 °C for future usage. The synthesized nanoparticles were characterized as follows. The morphology of the particles was investigated by TEM

Figure 5. TEM images of samples showing the changing of the shape by increasing molarity: (a) S5400M1.1, bar size is 16 nm, (b) S5400M1.3, bar size is 80 nm, and (c) S5400M1.5, bar size is 16 nm.



Figure 6. TEM images of (a) S3600M1.4, bar size is 20 nm, (b) S3600M1.2, bar size is 16 nm, (c) S12600M1.1, bar size is 16 nm, (d) S12600M1.4, bar size is 80 nm, (e) S9000M1.6, bar size is 16 nm, and (f) PVA nanorods in S3600M1.4, bar size is 25 nm.

(transmission electron microscopy, ZEISS, EM-10C, Germany) operating at 100 kV. To prepare samples for TEM, a drop of the suspension was placed on a copper grid and dried. Phase characterization was accomplished using the XRD (X-ray powder diffraction, Siemens, D5000, Germany) technique with Cu KR radiation, and the Scherrer method was used for particle size determination. The XRD samples were prepared by drying the particles in a vacuum oven at 40 °C for 12 h after centrifugation. Fourier transform infrared (FT-IR) spectra of the samples were taken with KBr pellets using an ABB Bomem MB-100 FT-IR spectrophotometer. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) measurements were performed with a DSC/TGA thermal analysis system (SDT Q600, USA). The mass loss of the dried samples was monitored under an inert atmosphere (argon gas) from 30 to 750 °C at a rate of 10 °C min-1. Ultraviolet visible spectroscopy (UV-vis) of the samples was performed with a Lambda 950 spectrophotometer (Perkin-Elmer, USA) from 200 to 1000 nm wavelength. The magnetization of the samples in a variable

Figure 7. XRD patterns of magnetite nanoparticles: (a) S12600M1.1, (b) S12600M1.4, and (c) S12600M1.5.

14474


Mahmoudi et al.

Figure 8. TEM and XRD results for (a) S9000M1.2, bar size is 20 nm, and (b) S3600M1.2, bar size is 16 nm.

magnetic field was measured using a vibrating-sample magnetometer (VSM) with a sensitivity of 10-3 emu and a maximum magnetic field of 10 kOe. The magnetic field was changed uniformly at a rate of 66 Oe/s. Experimental Design. For the design of experiments (DOE) approach, the values of the operational factors (molarity and stirring rate) are chosen on the basis of a two-dimensional uniformly scattered experimental space, as shown Figure 1. A conventional U18(62) is employed which refers to a uniform design with a total of 18 experiments containing 2 factors and 6 levels for each. Note that for a full factorial design one would require to perform 62 ) 36 experiments, whereas U18 is a fractional factorial design with a 50% reduction in the number of necessary experiments. Figure 1 represents such a design with the two factors being stirring rate and molarity. The main motivations in using a UD are the following: (a) It can reduce the total number of necessary experiments to perform the study (in the current case from 36 to 18). (b) It ensures a balance of properties in all dimensions of the experimental space; i.e., the number of repeats of each factor level is identical. (c) For a given number of runs and factors, a set of experiments can be chosen that yields a minimum discrepancy among all possible designs. (d) It is not required to have a precise knowledge of the input-output model to select a design. (e) The analysis becomes more robust against potential biased errors.20 In the present work, the NaOH molarity and the stirring rate are considered as independent variables and their effects on the optimization of nanoparticle size and magnetization are studied. The proposed DOE is also used to explore the variation of the response with a parameter while setting the other parameter at different constant levels (i.e., interaction effects). In what

follows, in order to refer to the synthesis parameters of a sample during the analysis, the samples are labeled by SxMy, where S is the value of the stirring rate, x is the stirring rate, M is the molarity, and y is the value of the molarity. For instance, S3600M1.2 refers to a sample prepared with a stirring rate of 3600 rpm and a NaOH molarity of 1.2. 3. Results and Discussion 3.1. TEM. Transmission electron microscopy (TEM) of the 18 prepared magnetite nanoparticles reveals either spherically shaped, or rod-shaped, or magnetic bead iron oxide nanoparticles (see Figures 2–6; the figures will be discussed in more detail in the following sections). On the basis of the TEM results in Figure 2, for a given stirring rate and molarity, it appears that fixing the stirring rate and increasing the molarity favors the formation of spherically shaped particles. In addition, formation of polymeric nanocomposites (magnetic beads) and rod-shaped nanoparticles is also detected, e.g., sample S5400M1.3 in Figure 3a. The rod-shaped particles seem to be similar to the magnetic nanoparticles synthesized by magnetotactic bacteria.21 The formation of rods may be related to the presence of PVA, since rod-shaped PVA without the presence of SPION has been detected by TEM (Figure 6f). The magnetic beads may form into two possible types of aggregates shown in Figure 4. The first type of aggregate may form by the aggregation of PVA coated nanoparticles, while the second type of aggregate may contain magnetite colloidal nanocrystal clusters (CNCs) coated with PVA, as seen in Figure 3b and c. The CNCs can be dispersed in water and remain stable in solution for at least several months, similar to dispersed individual Fe3O4 nanoparticles. Each CNC is composed of many magnetite crystallites


J. Phys. Chem. B, Vol. 112, No. 46, 2008 14475 consistent with results obtained from TGA and FT-IR), PVA can bind to the particle’s surface by strong chemisorption. Since the CNCs are composed of small primary nanocrystals, the CNCs retain superparamagnetic behavior at room temperature and can have bigger saturation magnetization than individual nanoparticles of the same size. Recently, Ge et al.22 developed a method for the preparation of polyacrylate-capped superparamagnetic magnetite CNCs with tunable sizes from 30 to 180 nm, using a high-temperature hydrolysis process, and applied them as photonic crystals. Our experimental results indicate that, with increasing molarity, spherical particles are obtained instead of rod- and cubicshaped crystals (Figure 5). The reason may be due to the reaction mechanism of magnetite:23

Fe3+ + 3OH- ) Fe(OH)3

(1)

Fe(OH)3 ) FeO(OH) + H2O

(2)

2+

Fe

-

+ 2OH ) Fe(OH)2

2FeO(OH) + Fe(OH)2 ) Fe3O4 + 2H2O

(3) (4)

An increase in molarity affects reactions 1 and 3 and may affect the kinetics of the nucleation and growth of Fe3O4 particles. In order to define the polydispersity of the spherical nanoparticles, sample statistics was applied. The variance is calculated by considering 100 particles in the TEM images via24 k

Figure 9. (a) 2-D and (b) 3-D graphs showing particle size for different synthesis conditions (molarity is given by M; dots are data points with error bars from the experiments).

of approximately 4 nm (consistent with our XRD findings - see later discussion). As illustrated in the TEM images (and

s2 )

1 ny2 fjyj2 n - 1 j)1 n-1

∑

(5)

where n is the number of measurements, yj is the size of each individual nanoparticle, jy is the sample mean, fj is the absolute frequency of the value yj, and s is the standard deviation. The

Figure 10. Magnetization curves for samples (a) S12600M1.5, (b) S5400M1.5, and (c) S7200M1.5.

14476


Figure 11. (a) 2-D and (b) 3-D graphs of the magnetization due to different synthesis conditions (dots are data points with error bars from the estimated response surface).

minimum standard deviation, indicating the least particle size variation, is obtained at S9000M1.2 (relatively low molarity and medium stirring rate). The maximum standard deviation is achieved at S3600M1.6, indicating the importance of the stirring rate on mass transfer and crystal growth rate.

Mahmoudi et al. 3.2. XRD. Figure 7 shows examples of XRD spectra for S12600M1.1, S12600M1.4, and S12600M1.5. We find (see the Supporting Information) that the XRD spectra for samples S9000M1.4, S10800M1.2, S10800M1.3, S12600M1.1, S12600M1.4, and S12600M1.5 match well with magnetite (Fe3O4, reference JCPDS No. 821533), indicating that the samples have a cubic crystal system.26,27 On the other hand, for some samples, extra peaks are observed which show the existence of impurities such as Fe2O3 and FeO(OH). Small traces of FeO(OH) and Fe2O3 were observed in samples S5400M1.3 and S5400M1.1, while in the remaining samples a negligible amount of Fe2O3 was detected (for example, at 2θ ) 54°). The XRD spectra are affected by the particle size and also by the magnetite content, both of which are influenced by the synthesis parameters. A possible effect may be the generation of bubbles in the reaction solution due to the use of high stirring rates and the effect of the ultrasonic bath which may cause magnetite to be oxidized.28 However, we find that the samples prepared at high stirring rates (12600 rpm) are pure magnetite. In fact, by increasing the stirring rate to 12600 rpm, the effect of molarity on the formation of pure magnetite is insignificant. In contrast, the effect of molarity on sample purity can be important at lower stirring rates (9000 and 10800 rpm). The particles prepared using the homogenizer with the highest stirring rate (12600 rpm) show a significant peak broadening and an amorphous-like pattern which may be explained by the formation of extremely small particles (Figure 7). The defective structure of the SPION may also diffuse the X-ray reflections. In Figure 8, an extra peak is visible around 18°. It is worth noting that magnetite nanoparticles produced by magnetotactic bacteria also exhibit this extra peak.28 However, Testa et al.29 mention that this peak is related to the matrix peak. Interestingly, the magnetic beads exhibit the peak in their XRD spectra (Figure 8). The full width at half-maximum (fwhm) of the (311) reflection was used to determine the average crystallite size of

Figure 12. Magnetization curves for samples (a) S12600M1.1, (b) S12600M1.4, and (c) S12600M1.5.



Figure 13. Magnetization curves for samples (a) S9000M1.2 (size 4.4 nm) and (b) S7200M1.5 (size 5 nm).

Figure 14. Simultaneous TGA and differential thermal analysis (DTA) results for (a) coated nanoparticles and (b) uncoated nanoparticles.

14478


Mahmoudi et al.

Figure 15. FT-IR spectrum of (a) uncoated SPION, (b) PVA, and (c) coated SPION.

the nanoparticles by using the Scherrer method. The size response due to the synthesis parameters is shown in Figure 9. Sun et al.27 reported that the size of magnetite nanoparticles is decreased by increasing the stirring rate due to the increase of the energy transferred to the suspension medium creating smaller droplets (reaction solution). Figure 9a reveals that the size may increase or decrease depending on both the molarity and mass transfer effects (stirring rate). Considering the size response in the context of the experimental design, it can be concluded that molarity and homogenization rate are interactive in nature. For

instance, in samples with M ) 1.5 there is an increase in the particle size from 7200 to 12600 rpm, whereas in samples with M ) 1.1 the trend is reversed, which means a simultaneous accounting of both factors (rpm and M) in the optimization process of the particles size is necessary. The dependence of particle size on M and rpm can be better understood from the existence of the multiple optima and minima in the 3-D response surface shown in Figure 9b, which is calculated from the 18 experimental points in Figure 1 using least-squares estimation.31 For the most part, a higher value of stirring rate (7200-9000



Figure 16. Absorbance spectrum for (a) Fe2+, (b) Fe3+, and (c) PVA coated SPION dispersion (note: PVA coated SPION is blown up 30 times for better comparison).

rpm) results in a smaller particle size for a given level of base concentration. The interaction effects within this range are less insignificant, and stirring rate can be used as the control variable for obtaining the desired size of the particles. The minimum particle size is found for a low molarity (M ) 1.1) and a low stirring rate (∼5400 rpm). However, in this neighborhood, the sensitivity of the response can be high and each small variation in the rpm and M values can lead to a major change in the size response, which is not recommended for a robust design.30 The maximum sensitivity to stirring rate is seen at the molarity values of 1.1 and 1.2. Molarities between 1.3-1.4 and homogenization rates between 7200-9000 rpm appear to provide a stable range for producing magnetite nanoparticles with similar particle size and coated with PVA. 3.3. VSM. All 18 samples were analyzed by VSM and showed superparamagnetic behavior with different magnetic saturations. Figure 10 illustrates hysteresis loops of the synthesized nanoparticles showing a negligible remanence and coercivity in the hysteresis loops. The effect of the synthesis parameters on the magnetic saturation of the samples is illustrated in Figure 11a. The interaction between stirring rate and base concentration is again evident, which can be shown in the 3-D magnetization response surface of Figure 11b. For the most part, a value of molarity larger than 1.4 tends to result in higher magnetization values for a given stirring rate. It appears that the value of M ) 1.4 is a critical molarity level for the magnetization response, resulting in a sharp sensitivity with rpm. The optimal maximum magnetization response is seen around a low molarity (1.1-1.2) and medium stirring rates (7200-9000 rpm). In this neighborhood, the magnetization is notably insensitive to the synthesis parameters, which is favorable for a robust design. Comparable optimal conditions were also observed at the points of 12600 rpm and 1.5 M or 3600 rpm and 1.6 M. Considering the size and magnetization responses simultaneously, it can be concluded that, for a biobjective optimization

of these two objective functions, a solution in the neighborhood of 7200-9000 rpm and 1.1 M may be considered as a compromise solution. The exact solution of such a biobjective optimization process remains to be investigated using a weighted additive global response. Several researchers have reported that the magnetic saturation of superparamagnetic magnetite increases when the size of the magnetite increases, which can be attributed to the increase of the weight and volume of magnetite nanoparticles.28,32 According to these studies, the magnetic properties can be lower than that of the bulk phase which is 88 emu/g.33–35 The sizes of pure magnetite for samples S12600M1.1, S12600M1.4, and S12600M1.5 are 4.7, 4.9, and 5.5 nm, respectively. Considering the abovementioned literature reports, the magnetic saturation is expected to increase due to an increase in the size; however, Figure 12 clearly reveals that the existence of magnetic beads consisting of larger primary particles decreases the magnetic saturation in comparison with individual smaller nanoparticles due to a decline in exchange penetration as well as dipolar interactions. On the contrary, the existence of CNCs considerably enhances the magnetic saturation for the same larger individual nanoparticles (Figure 13). Since CNCs are composed of many small nanocrystals, they retain the superparamagnetic properties at room temperature and show much higher saturation magnetization than individual nanocrystals of the same size.23 Since high saturation magnetization values are required to achieve high contrast, CNCs may be considered to be promising candidates for MRI applications. 3.4. TGA and DSC. Results of simultaneous thermal analysis (TGA and DSC) on the PVA coated and uncoated nanoparticles are shown in Figure 14. A three-step mass loss is observed for the coated nanoparticles in contrast to a two-step mass loss for the uncoated nanoparticles. The first mass loss occurred at 50-500 °C for the uncoated particles and may be related to the elimination of hydroxyl groups formed on the particle surfaces during the synthesis process. In the coated

14480


nanoparticles, three significant mass losses occur in the temperature ranges 150-200, 500-700, and 800-900 °C. The first loss can be related to removal of the PVA layer. The second loss can be attributed to the elimination of chemisorbed PVA on the particle surfaces.36 The last peak, which is noticed at relatively high temperatures, could be due to removal of OH groups from contaminant FeO(OH) and/or reduction of the magnetite to Fe-O or Fe by decomposed PVA. 3.5. FT-IR. Researchers have investigated the interaction between a coating polymer and Fe3O4 particles.36–40 For instance, polymer interactions were studied in Fe3O4/polypyrrole nanocomposites37 and in Fe3O4/polyaniline nanocomposites.38 They assumed interactions exist between the lone pair electrons of the N atom in the polypyrrole chain or in the polyaniline chain with the 3d orbital of the Fe atom to form a coordinate bond. Li et al. reported that the interactive mechanism of oleic acid adsorption on the surface of Fe3O4 nanoparticles could be due to a hydrogen bonding or a coordination linkage.38 Zhang et al. reported that poly(methacrylic acid) could adhere to Fe3O4 nanoparticles via coordination linkages between the carboxyl groups and iron.39 Figure 15 shows FT-IR spectra of Fe3O4 uncoated nanoparticles, pure PVA, and coated nanoparticles. The FT-IR spectra of iron oxide exhibit strong bands in the low-frequency region (1000-500 cm-1) due to the iron oxide skeleton. This pattern is consistent with the magnetite (Fe3O4) spectrum (band between 570-580 cm-1) or the maghemite (γ-Fe2O3) spectrum (broadband 520-610 cm-1).40 The characteristic band of Fe-O at 572 cm-1 shows that the particles consist mainly of Fe3O4. The peaks around 3400 cm-1 in both the uncoated nanoparticles and the pure PVA are related to hydroxyl groups. Since no peak at ∼3400 cm-1 is detected for PVA coated nanoparticles, PVA is probably chemisorbed to the particle surface. In order to show that PVA can also retard crystal growth by preventing iron ion diffusion to the surfaces of nanoparticles, UV visible spectroscopy (UV-vis) has been applied. Figure 16 shows the absorbance pattern of Fe2+ and Fe3+ (both in FeCl2 and FeCl3 solutions, respectively) and the solution after synthesis (obtained ferrofluids before centrifugation). Fe2+ and Fe3+ show an absorbance edge around 200 and 400 cm-1, respectively. From the UV-vis result of the synthesized samples, the trace of iron ions, especially for Fe2+, is observed. Since the pH of the solution is appropriate to reduce Fe2+ on the surface of the nanoparticles, further crystal growth could be achieved with longer incubation times. However, the growth in nanoparticle size is prevented by PVA. 4. Conclusions Both uniform-spherical magnetite nanoparticles, nanorods, and cubic polymeric nanocomposite/magnetic beads were synthesized by an aqueous coprecipitation process. A rich behavior in particle size, particle type formation, and superparamagnetic properties was observed as a function of molarity and stirring conditions. From the size, shape, and magnetic measurements, it was shown that nanoparticles, magnetic beads, and the CNCs exhibit well-defined superparamagnetic behavior with different magnetic saturations. The XRD patterns show that the purity and the size of prepared magnetite needs to be considered as a function of the synthesis parameters. Using a uniform DOE approach with 3-D response surfaces, it is found that desirable size and/or magnetic saturation can be achieved by choosing appropriate parameters for targeting delivery or imaging applications. From the investigations on particle size and magnetization responses, one can conclude that, for a

Mahmoudi et al. biobjective optimization of these two functions, a solution in the neighborhood of 7200-9000 rpm and 1.1 M may be considered as a possible robust solution. Magnetic beads consisting of larger primary particles showed a significant decrease in saturation magnetization in comparison with individual smaller nanoparticles. The CNCs, composed of smaller primary particles, showed a considerable saturation magnetization with respect to larger individual nanoparticles. Acknowledgment. The authors would like to thank Prof. Urs Hafeli from the University of British Columbia for his constructive comments on an earlier version of the manuscript. Supporting Information Available: XRD spectra. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Mornet, S.; Vasseur, S.; Grasset, F.; Duguet, E. Magnetic Nanoparticle Design for Medical Diagnosis and Therapy. J. Mater. Chem. 2004, 14, 2161–2175. (2) Hafeli, U.; Schu tt, W.; Teller, J.; Zborowski, M. Scientific and Clinical Applications of Magnetic Microspheres; Plenum Publication: New York, 1997. (3) Mahmoudi, M.; Simchi, A.; Milani A. S.; Stroeve, P. Surface Engineering of Iron Oxide Nanoparticles for Drug Delivery. Proceedings of The 2nd Conference on Nanostructures, Kish Island, Iran, 2008; pp 6668. (4) Corr, S. A.; Rakovich, Y. P.; Gun’ko, Y. K. Multifunctional Magnetic-fluorescent Nanocomposites for Biomedical Applications. Nanoscale Res. Lett. 2008, 3, 87–104. (5) Zaitsev, V. S.; Filimonov, D. S.; Presnyakov, I. A.; Gambino, R. J.; Chu, B. Physical and Chemical Properties of Magnetite and MagnetitePolymer Nanoparticles and Their Colloidal Dispersions. J. Colloid Interface Sci. 1999, 212, 49–57. (6) Kang, Y. S.; Risbud, S.; Rabolt, J. F.; Stroeve, P. Synthesis and Characterization of Nanometer-Size Fe3O4 and γ-Fe2O3 Particles. J. Chem. Mater. 1996, 8 (9), 2209–2211. (7) Berry, C. C.; Curtis, A. S. G. Magnetic Nanoparticles for Applications in Biomedicine. J. Phys. D: Appl. Phys. 2003, 36, 198. (8) Moghimi, S. M.; Hunter, A. C.; Murray, J. C. Long-circulating and Target-specific Nanoparticles: Theory to Practice. Pharm. ReV. 2001, 53, 283. (9) Gupta, A. K.; Gupta, M. Synthesis and Surface Engineering of Iron Oxide Nanoparticles for Biomedical Applications. Biomaterials 2005, 26, 3995–4021. (10) Harris, J. M.; Martin, N. E.; Modi, M. Pegylation: a Novel Process for Modifying. Clin. Pharmacokinet. 2001, 40, 539–551. (11) Xue, B.; Sun, Y. Protein Adsorption Equilibria and Kinetics to a Poly(vinyl alcohol)-based Magnetic Affinity Support. J. Chromatogr., A 2001, 921, 109–119. (12) Schwertmann, U.; Cornel, R. M. Iron Oxides in the Laboratory; VCH: Weinheim, Germany, Cambridge, U.K., 1991. (13) Hong, R.; Li; J.; Wang, J.; Li, H. Comparison of Schemes for Preparing Magnetic Fe3O4 Nanoparticles. Particuology 2007, 5, 186–191. (14) Gupta, A. K.; Gupta, M. Synthesis and Surface Engineering of Superparamagnetic Iron Oxide Nanoparticles for Drug Delivery and Cellular Targeting Handbook of particulate drug deliVery; American Scientific Publishers: Steven Ranch, CA, 2007. (15) Gupta, A. K.; Naregalkar, R. R.; Vaidya, V. D.; Gupta, M. Recent Advances on Surface Engineering of Iron Oxide Nanoparticles and Their Biomedical Applications. Nanomedicine 2007, 2 (1), 23–29. (16) Pankhurst, Q. A.; Connolly, J.; Jones, S. K.; Dobson, J. Applications of Magnetic Nanoparticles in Biomedicine. J. Phys. D: Appl. Phys. 2003, 36, 167–181. (17) Shan, G. B.; Xing, J. M.; Luo, M. F.; Liu, H. Z.; Chen, J. Y. Immobilization of Pseudomonas Delafieldii with Magnetic Polyvinyl Alcohol Beads and Its Application in Biodesulfurization. Biotechnol. Lett. 2003, 25, 1977–1983. (18) Fowlkes, W. Y.; Creveling, C. M. Engineering Methods for Robust Product Design; Prentice Hall: New York, 1995. (19) Robinson, G. K. Practical Strategies for Experimenting; Wiley: New York, 2000. (20) Fang, K. T.; Lin, D. K. J.; Winker, P.; Zhang, Y. Uniform Design: Theory and Applications. Technometrics 2000, 42 (3), 237–248. (21) Vali, H.; Kirschvink, J. L. Iron Biominerals; Plenum Press: New York and London, 1991; pp 97-116.

Optimal Design and Characterization of SPION (22) Ge, J.; Hu, Y.; Biasini, M.; Beyermann, W. P.; Yin, Y. Highly Tunable Superparamagnetic Colloidal Photonic Crystals. Angew. Chem., Int. Ed. 2007, 46, 1–5. (23) Lian, S.; Wang, E.; Kang, Z.; Bai, Y.; Gao, L.; Jiang, M.; Hu, C.; Xu, L. Synthesis of Magnetite Nanorods and Porous Hematite nanorods. Solid State Commun. 2004, 129, 485–490. (24) Biles, W. E.; Swain, J. J. Optimization and Industrial Experimentation; Wiley: New York, 1980. (25) Itoh, H.; Sugimoto, T. Systematic Control of Size, Shape, Structure, and Magnetic Properties of Uniform Magnetite and Maghemite Particles. J. Colloid Interface Sci. 2003, 265, 283–295. (26) Voit, W.; Kim, D. K.; Zapka, W.; Muhammed, M.; Rao, K. V. Magnetic Behavior of Coated Superparamagnetic Iron Oxide Nanoparticles in Ferrofluids. Mater. Res. Soc. 2001, 676, 781–786. (27) Sun, J.; Zhou, S.; Hou, P.; Yang, Y.; Weng, J.; Li, X.; Li, M. Synthesis and Characterization of Biocompatible Fe3O4 Nanoparticles. J. Biomed. Mater. Res., Part A 2007, 80, 333–341. (28) Yeary, L. W.; Moon, J. W.; Love, L. J.; Thompson, J. R.; Rawn, C. J.; Phelps, T. J. Magnetic Properties of Biosynthesized Magnetite Nanoparticles. IEEE Trans. Magn. 2005, 41, 4384–4389. (29) Testa, A. M.; Foglia, S.; Suber, L.; Fiorani, D.; Casas, L.; Roig, A.; Molins, E.; Greneche, J. M.; Tejada, J. Unconventional Magnetic Behavior of Iron-Oxide Nanoparticles in Polymeric Matrices. J. Appl. Phys. 2001, 90, 1534–1539. (30) Myers, R. H.; Montgomery, D. C. Response Surface Methodology: Process and Product Optimization Using Designed Experiments, 2nd ed.; Wiley: New York, 2002. (31) Haddad, P. S.; Martins, T. M.; Li, L. D.; Li, L. M.; Metze, K.; Adam, R. L.; Knobel, M.; Zanchet, D. Structural and Morphological Investigation of Magnetic Nanoparticles Based on Iron Oxides for Biomedical Applications. Mater. Sci. Eng., C [Online early access]. DOI: 10.1016/ j.msec.2007.04.014. Published Online: Apr 16, 2007.

J. Phys. Chem. B, Vol. 112, No. 46, 2008 14481 (32) Huang, Z.; Tang, F. Preparation, Structure and Magnetic Properties of Mesoporous Magnetite Hollow Spheres. J. Colloid Interface Sci. 2005, 281, 432–436. (33) Gomez-Lopera, S. A.; Plaza, R. C.; Delgado, A. V. Synthesis and Characterization of Spherical Magnetite/Biodegradable Polymer Composite Particles. J. Colloid Interface Sci. 2001, 240 (1), 40–47. (34) Jiang, W.; Yang, H. C.; Yang, S. Y.; Horng, H. E.; Hung, J. C.; Chen, Y. C.; Hong, C. Y. Preparation and Properties of Superparamagnetic Nanoparticles with Narrow Size Distribution and Biocompatible. J. Magn. Magn. Mater. 2004, 283, 210–214. (35) Lan, Q.; Liu, C.; Yang, F.; Liu, Sh.; Xu, J.; Sun, D. Synthesis of Bilayer Oleic acid-coated Fe3O4 Nanoparticles and Their Application in pH-responsive Pickering Emulsions. Colloid Interface Sci. 2007, 310, 260– 269. (36) Deng, J. G.; Peng, Y. X.; He, C. L.; Long, X. P.; Li, P.; Chan, A. S. C. Magnetic and Conducting Fe3O4-polypyrrole Nanoparticles with Core-Shell Structure. Polym. Int. 2003, 52 (7), 1182–1187. (37) Deng, J.; Ding, X.; Zhang, W.; Peng, Y.; Wang, J.; Long, X.; Li, P.; Chan, A. S. C. Magnetic and Conducting Fe3O4-cross-linked Polyaniline Nanoparticles with Core-Shell Structure. Polymer 2002, 43, 2179–2184. (38) Li, P.; Yu, B.; Wei, X. Synthesis and Characterization of a High Oil-Absorbing Magnetic Composite Material. J. Appl. Polym. Sci. 2004, 93, 894–900. (39) Zhang, H.; Wang, R.; Zhang, G.; Yang, B. A Covalently Attached Film Based on Poly(methacrylic acid)-capped Fe3O4 Nanoparticles. J. Appl. Polym. Sci. 2003, 429, 167–173. (40) Ma, H. L.; Qia, X. R. Preparation and Characterization of Superparamagnetic Iron Oxide Nanoparticles Stabilized by Alginate. Int. J. Pharmacol. 2007, 333, 177–186.

JP803016N

Optimal Design and Characterization of Superparamagnetic Iron

Recommend Documents