Organic Solvent-Free, One-Step Engineering of Graphene-Based Magnetic-Responsive Hybrids Using Design of Experiment-Driven Mechanochemistry

In this study, we propose an organic solvent-free, one-step mechanochemistry approach to engineer water-dispersible graphene oxide/superparamagnetic iron oxide (GO/SPIOs) hybrids, for biomedical applications. Although mechanochemistry has been proposed in the graphene field for applications such as drug loading, exfoliation or polymer-composite formation, this is the first study to report mechanochemistry for preparation of GO/SPIOs hybrids. The statistical design of experiment (DoE) was employed to control the process parameters. DoE has been used to control formulation processes of other types of nanomaterials. The implementation of DoE for controlling the formulation processes of graphene-based nanomaterials is, however, novel. DoE approach could be of advantage as one can tailor GO-based hybrids of predicted yields and compositions. Hybrids were characterized by TEM, AFM FT-IR, Raman spectroscopy, and TGA. The dose–response magnetic resonance (MR) properties were confirmed by MR imaging of phantoms. The biocompatibility of the hybrids with A549 and J774 cell lines was confirmed by the modified LDH assay.

3 temperature was monitored using a thermometer throughout the reaction. When the powders were fully dispersed, potassium permanganate powder (KMnO4, 12 g) was added slowly (portion--wise) to the suspension. The addition rate was carefully controlled so that the temperature of the suspension was kept below 20°C. After the addition of KMnO4 was complete, the suspension was stirred on an ice bath (0 ~ 10°C) for an additional 10 minutes. The ice bath was then removed and replaced with a water bath to allow the temperature to raise gradually to 35 ± 3°C. The temperature was maintained for at least 2 hr until the dispersion became a brown pasty mixture. The mixture was then heated to 60°C. When the temperature reached 60°C, 50 mL of DI H2O was added to the pasty mixture and stirred for 5 minutes. The temperature was then set to 80°C followed by the addition of 50 mL of DI H2O. When the temperature reached 80°C, 80 mL of DI H2O was added to the mixture. A white vapour was generated. When the water addition was complete, the temperature was maintained above 80°C (temperature fluctuated between 80--95°C) and kept stirring for another 30 minutes. To stop the reaction, hydrogen peroxide (H2O2) aqueous solution (35%, 10 mL) was added to the mixture, which was then kept stirring for 30 minutes at 80°C. H2O2 reduces the residual manganese dioxide (MnO2) and permanganate (MnO4 --) forming a brown suspension. The warm brown suspension from the previous step was transferred into 10 × 50 mL conical--bottom centrifuge tubes and centrifuged at 3,214 × g (3900 rpm) for 2 minutes. The supernatant was removed and the pellets were re--suspended in 1N HCl using a spatula with agitation. The centrifugation and washing steps were repeated several times with 1N HCl (1L) followed by 1N NaOH (0.5 ~ 1 L). The washing and centrifugation steps were monitored by observing the colour of the supernatant. The colour changed from yellow--green (early HCl washing stage), nearly transparent (late HCl washing stage), light brown (early NaOH washing stage) and finally to dark brown when NaOH washing was stopped. The pellets were re--suspended in DI H2O. Washing and centrifugation steps were repeated twice. The pellets were finally re--suspended in DI H2O (50 mL/tube) and bath sonicated for 1 hr (USC300TH, VWR, Belgium). The sonicated mixture was centrifuged once at 3,214 × g (3900 rpm) for 2 minutes. GO suspension was obtained by strongly hand--shaking the centrifuge tubes and the supernatant was collected. Oxidized--graphite remained as pellet no matter how hard the shaking was. The collected supernatant was then diluted with DI H2O (volume was not critical) forming a dark--brown GO suspension. The final concentration of GO in the dispersion was determined by TGA as described in the main text using Equation 6. The dispersion concentration for as--synthesised GO stock was 6 mg/mL

Hybrid preparation by ball milling
GO and SPIOs were loaded into the grinding jar with four 12 mm stainless steel grinding balls (Retsch ® , Germany). The milling process was performed via a high--speed vibrational ball--mill (MM 200 mixer mill, Retsch ® , Germany). The initial GO/SPIOs wt/wt% (RPInitial) was calculated as expressed in Equation 5, where MSInitial is the initial SPIOs weight (mg) loaded into the milling jar. The batch size of GO was fixed at 30 mg.
After milling, the mixture was recovered from the grinding jar by rinsing with 30 ml DI H2O, then transferred into a 50 mL self--standing conical bottomed centrifuge tube. The majority of insoluble SPIOs remained in the jar (firmly attached to the wall of the jar).
The dispersion was allowed to stand for 2 minutes to precipitate remaining free SPIOs. A magnet was then put in contact with the exterior of the tube at its middle part and left for 30 second. The magnetic hybrid was attracted to the magnet while non--magnetic GO remained in suspension. The non--magnetic fraction or any precipitates were discarded.
The magnetic fraction was collected and re--suspended in 30 mL DI H2O, and magnetic separation step was repeated once. After the final washing, the purified GO/SPIO hybrids was re--suspended in DI H2O and stored at room temperature.

Thermogravimetric analysis (TGA)
TGA (TGA Q5000, TA Instruments, USA) was used to analyse both the final weight of the collected hybrid and the SPIOs content in each sample. Samples were run isothermally at 120 °C under nitrogen atmosphere for 2 min to allow for water removal so that the weight of the dried hybrid can be calculated. The concentration of hybrid in the original dispersion was calculated accordingly using Equation 6, where C is the concentration (mg/mL), VL is the volume (μL) of dispersion being loaded on the TGA pan. The dried hybrid weight, generated from ball milling, was calculated using Equation 7, where MTotal is the total hybrid weight (mg) and VTotal is the total sample volume (mL).
The dried GO, SPIOs or GO/SPIO hybrids were then equilibrated at 100°C for 20 minutes before heating up with a ramp of 10°C /min from 100 to 900°C under compressed air with a flow rate of 90 mL/min. The residual weight (mg) after decomposition in compressed air was assigned as M2. The weight (mg) of SPIOs (MSFinal) and GO (MGFinal) in the hybrid was calculated using Equation 8a and 8b, respectively. As mentioned earlier, oleic acid coated SPIOs showed a weight loss of 31.48 ± 0.01 % at 800°C (n = 3).
The value of M2 was divided by 0.685 to account for oleic acid content in the coated

Atomic force microscopy (AFM)
The surface topography of GO samples deposited on a dry mica surface was studied with AFM using tapping mode. GO samples were prepared at a concentration of 0.

GO flake surface area analysis
The surface area of the GO flake was analysed using Image J 1.49i software from National Institute of Health (USA). Images from AFM were used for analysis with a total measurement number no less than 100 flakes. Histogram of surface area distribution and statistical calculation was generated and performed using DataGraph 3.1 (Visual Data Tools, Inc., USA).

Transmission electron microscopy (TEM)
TEM was performed on Philips CM 12 (FEI Electron Optics, The Netherlands) equipped with Tungsten filament and a Veleta --2k × 2k side--mounted TEM CCD Camera (Olympus, Japan). The accelerating voltage was 80 KV. The spot size was set at 3. Objective aperture was used with all samples. GO--SPIOs aqueous dispersions at a concentration of 0.2 mg/mL were deposited on carbon--film on 300 mesh copper grids or lacey carbon films on 300 mesh copper grids for the measurement.

Attenuated Total Reflectance Fourier Transform Infrared (ATR--FTIR)
ATR--FTIR was performed using PerkinElmer ® Frontier™ FT--IR equipped with ATR accessory (diamond ATR polarization accessory with 1 reflection top--plate and pressure arm). The pressure arm was used for all solid samples at a force gauge setting between 100--120 units; no compression was used for liquid/oil samples. The number of scans was set at 15. Samples were loaded on the reflection top--plate at a quantity sufficient enough to cover the entire diamond surface. GO/SPIO hybrids were dispersed in water and freeze dried into a loose sponge--like structure. The container was sealed immediately after freeze--drying. Samples were warmed to room temperature before the measurement to minimise moisture absorption, in order to obtain a good spectrum resolution with minimum interferences.

Raman spectroscopy
Raman spectroscopy was performed using Renishaw ® inVia--Reflex spectrometer (UK) with an excitation wavelength set to 785 nm and 0.1 to 50 % laser power. SPIOs and GO/SPIO hybrids were deposited on a calcium fluoride (CaF2) slide (Crystran Ltd, UK).
Acetone was used to disperse the sample on the slide and was removed by air--dry at room temperature. Measurements were performed from 500 cm --1 to 3200 cm --1 for each sample (n = 3). To boost the Raman signal of the SPIOs, higher laser power was used to transform the SPIOs into α--Fe2O3. In brief, GO/SPIO hybrids were firstly bleached with 50 % laser power for 30 seconds followed by 1 % laser power for 30 seconds. Raman spectra were then obtained using 1 % laser power with 5 accumulative scans for one sample.

Statistical design of experiments (DoE) using two--level factorial design
Two--level full factorial design was used to screen the effect of three factors namely: where "n" is the number of experiments at each level. Effect from the two--factor interactions (EAB, EBC, and EAC) were calculated using the same equation. 5 As shown in Table  S1, factors were coded as +1 or --1 for high or low levels, respectively. Coding was implemented so that the influence of the factors can be compared irrespective of the measurement unit. Three--replicated control runs i.e. center points (0) were introduced to estimate pure errors and to check for curvatures. A total of 11 experiments were performed (8 design points and 3 center points) in a random order to minimize bias.
The corded design layout is shown in Table  S2.

Analysis of response 1 (MTotal)
The main effects, 2FI and 3FI of response 1 were calculated using Equation S--1 and summarized in Table  S The estimated regression coefficients for response 1 (MTotal) are summarized in Table  S

Analysis of response 2, (RPFinal)
The half normal probability plot of the absolute standardized effects for MTotal is shown in Figure S       Despite a good number of 10--20 nm SPIOs were found in the hybrid (as shown in Figure   2), some larger aggregated particles were also found to be associated with GO. As SPIOs aggregate without solvents, this is expected and was acknowledged as one of the limitations of the ball--milling approach. Cryo--mill or ball--mills with higher grinding power may help to decrease the aggregation of SPIOs when using a solvent--free approach. 22