Influence of Oxidation Degree of Graphene Oxide on Its Nuclear Relaxivity and Contrast in MRI.

Graphene oxide (GO) serves as a versatile platform for various applications, with the oxygen content of GO playing an important role in governing its properties. In the present study, different GO types covering a wide range of oxidation degree were prepared using our newly developed two-step method involving ball milling of graphite followed by its oxidation to GO. In addition to the variations in their physicochemical properties, the different GO types exhibited differences in proton relaxivity due to their paramagnetic nature. Nuclear magnetic resonance spectroscopy studies showed that the degree of oxidation of GO perturbs its nuclear relaxation properties and, together with intercalated Mn2+ ions, provides large contrast variation in magnetic resonance imaging (MRI). The study for the first time reveals that the surface chemistry of GO affects its relaxivity and opens up new avenues for developing tunable GO-based contrast agents in magnetic resonance imaging for diagnostics and therapies.

Introduction

Graphene oxide (GO) serves as a versatile platform owing to its hydrophilicity, high aspect ratio, and high degree of functionalization capability.[1,2] The oxygen content of GO plays an important role in determining its physical and chemical properties such as photoluminescence, adsorption, ζ-potential, solubility, electrical conductivity, and thermal conductivity.[3−5] The degree of oxidation of GO is also important for determining its toxicity and interaction with biomolecules.[6,7] The focus, therefore, has been on development of methods for modulating the degree of oxidation of GO, which can be achieved by either modification of the oxidation conditions of graphite or reduction of as-synthesized graphene oxide.[8,9]

Among its various applications, GO has been used as a substrate for loading agents to improve contrast in magnetic resonance imaging (MRI),[10−12] a widely used technique for clinical diagnostics and imaging of soft tissue.[13,14] Using nonionizing radiation, MRI provides high spatial resolution, noninvasiveness, and real-time visualization that aides in the detection of tumors, cysts, microfractures, and anomalies in various parts of the body.[15−17] However, smaller lesion size and subtle changes in contrast property limit proper identification and diagnosis of the disease.[18,19] For enhanced sensitivity to detect subtle changes, exogenous contrast agents are injected.[20,21]

Our objective was to develop appropriately functionalized GO as a contrast agent for MRI applications. The contrast ability depends on the paramagnetic behavior of GO, which affects its nuclear relaxation property, termed as “relaxivity.”[22] Currently, the role of GO in the enhancement of proton relaxivity remains unclear. Previous studies have shown that GO exhibits paramagnetic behavior due to the presence of paramagnetic impurities such as Mn2+ ions;[23] these metal ions are entrapped between graphite sheets during the potassium permanganate-based oxidative procedure used for producing GO.[24] In the present study, different GO types covering a wide range of oxidation degree were prepared using a two-step method.[5] On probing the effect of GO on proton relaxivity using nuclear magnetic resonance (NMR) spectroscopy in combination with other techniques, we observed that GO with different degrees of oxidation exhibited varied relaxivity effects. It was found that the effect of GO on water proton relaxation depends on both the Mn2+-ion concentration and the degree of oxidation. The two factors together help in varying the contrast of images in MRI, which is required for accurate diagnosis. The approach, thus, for the first time provides a method to combine two parameters and helps in designing appropriately oxidized GO having a desired relaxivity for MRI applications.

Results and Discussion

Characterization of the Synthesized GO

The morphology of GO was studied using atomic force microscopy (AFM). Due to ball milling, the precursor undergoes morphological changes and various defects are introduced during the process.[25] During the initial hours of ball milling, shearing and breaking of the graphite particles occur. This phenomenon can be observed till 30 h of ball milling (under conditions of our setup). On further ball milling, these small particles start sintering and forming larger sheets, also known as cold welding.[26,27] All of the graphites were ball-milled separately to get ball-milled graphitic precursors. On oxidizing these precursors, we get the corresponding GO. Such variations in morphology of the GO are revealed in particle thickness measured by AFM, as shown in Figure (the AFM images are shown in Figure S1 of the Supporting Information). GO_0 was highly polydispersive and had the highest thickness among all GO. The sheet size, layer thickness, and dispersity decreased till GO_30. GO_50 had a larger sheet size, indicating that GO_50 consists of sintered particles of BMG_50. GO_80 and GO_100 had much more structured morphology due to sintering process. The particle thickness was found to decrease along the GO series (Figure ; the values are tabulated in Table S1).

Figure 1

Particle thickness of GO samples obtained using AFM.

The oxidation degree is calculated from: (i) the ratio of reacted graphite (GO) to the total product (unreacted graphite and GO), obtained by integrating the appropriate peaks in ssNMR spectra (Figure S2) and (ii) the ratio of nonaromatic carbon to total carbon obtained from the deconvolution of high-resolution C 1s X-ray photoelectron spectra (XPS) (Figure S3). The oxidation degree of the GO in Figure a indicates an overall increase in the oxidation degree along the GO series. The Fe3+/Fe2+ ion was initially expected to be present due to contamination from the ball milling jar, but on atomic absorption spectroscopic investigation, no trace of such contamination was observed. To measure the ζ-potential, which is an indirect measure for colloidal stability, electrophoretic light scattering (ELS) was used. The ζ-potential values for each GO are tabulated in Table S2 and plotted in Figure b. Similar to the oxidation degree, the ζ-potential increases along the GO series.

Figure 2

(a–c) Variations in oxidation degree, ζ-potential, and defect density along the GO series.

The density of defects for GO (ID/IG) shown in Figure c was calculated by deconvolution of the Raman spectra (Figure S4). The Raman band at 1350 cm–1 arises from A1g and disorder-induced mode and is assigned as the D band; the bands at 1532 and 1598 cm–1 are a combination of A1g, E1g, and E2g vibration modes and are assigned to the G band and amorphous carbon, respectively. The intensities of the D and G bands were used to calculate the defect density, and the results are tabulated in Table S3 and plotted in Figure c. The defect density, while decreasing along the GO series, was not found to follow any particular pattern.

Graphene oxide exhibits paramagnetic behavior. This has been attributed to the presence of Mn2+ ions in GO.[23,24] The intercalated Mn2+ ions exist in GO due to the use of potassium permanganate (KMnO4) for oxidation of graphitic precursor to obtain GO. The Mn2+ concentration is measured using electron paramagnetic resonance (EPR) spectroscopy. The Mn2+-ion concentration (Figure b) was calculated from the spins obtained from the EPR spectra (Figure a). From Figure b, we observe that GO_80 contains a higher amount of Mn2+ ions and that GO_50 has a low amount of Mn2+. Notably, GO_0 has a much higher amount of Mn2+ than GO_100, yet GO_100 exhibits better relaxivity than GO_0. This implies that there are other factors involved in governing the relaxation behavior of GO.

Figure 3

(a) EPR spectra recorded at 298 K of all GO showing six hyperfine lines at an operating frequency of 9.45 GHz and field band of 286–386 mT. (b) Variation of Mn2+ concentration along the GO series.

No peak for Mn lattice was detected in XRD; this indicates that the Mn is not present in any polycrystalline phase.[28] The observation of a well-resolved hyperfine structure (Figure a) indicates that Mn2+ ions exist in the compound as quite magnetically diluted paramagnetic complexes rather than in the form of some impurities of concentrated Mn salts.[24,29] The large hyperfine coupling is an indicator of the octahedral coordination of the Mn species. The narrow singlet line at the center of the EPR spectra can be attributed to carbon radicals.

Correlation of Oxidation Degree and Relaxivity in GO

We plotted defect density, ζ-potential, and Mn2+ concentration against oxidation degree to investigate the degree of correlation among them. On linearly fitting the plots, as shown in Figure , we found that “ζ-potential vs oxidation degree” and “defect density vs oxidation degree” were found to have Pearson’s correlation coefficient greater than 0.8, while the Mn2+ concentration vs oxidation degree plot had the correlation coefficient value of 0.52, indicating that Mn2+ concentration and oxidation degree are mutually exclusive to each other and are independent factors. The contribution of these two factors, Mn2+ concentration and oxidation degree, to the relaxation behavior of GO was further explored.

Figure 4

Correlation of oxidation degree of GO with (a) ζ-potential, (b) defect density, and (c) Mn2+ concentration and estimation of Pearson’s r (correlation coefficient) value. Color scheme of GO: blue square, GO_0; purple square, GO_10; red square, GO_20; green square, GO_30; peacock-green square, GO_50; blue square, GO_80; yellow square, GO_100.

Paramagnetic materials alter the nuclear spin relaxation (T1 and T2) of surrounding water protons, thereby facilitating a good image contrast for accurate diagnosis in magnetic resonance imaging.[30] The potency of such contrast agents termed as “relaxivity” is quantitatively measured by its effect on relaxation rates of the water protons. It is measured as the change in relaxation rate (inverse of relaxation time) per unit concentration of the CA.[31]

The relaxivity values for the different GOs are shown in Figure . The relaxivity curves for individual GO are shown in Supporting Information Figures S5 and S6. The values for each GO sample are provided in Supporting Information Table S4. No substantial difference was found between the Tnull method and inversion recovery experiment on comparing the values obtained from the Tnull method and inversion recovery experiments with 16 delay periods.

Figure 5

(a, b) Measured r1 and r2 relaxivities of different GOs. T1 and T2 measurements were carried out at a magnetic field strength of 9.4 T at a temperature of 298 K.

When the r1 and r2 relaxivities are plotted for the GO series, we observe an irregular pattern in the variation of the relaxivities along the series. A distinguishable observation is that GO_80 has a considerably higher relaxivity. GO_50 has the lowest r1 relaxivity followed by GO_0. GO_50 also has the lowest r2 relaxivity.

On plotting the r1 and r2 relaxivities against the corresponding Mn2+ concentration (Figure a,b), we noted that samples having different Mn2+ concentrations exhibit same relaxivity and vice versa. GO_80 having the highest Mn2+ concentration has the highest relaxivity; GO_0 has higher Mn2+ than most of the other GO samples yet the lowest relaxivity values. GO_50 has similar concentrations of Mn2+ to GO_100 but much lower relaxivity than the latter. At the same time, on plotting r1 and r2 relaxivities against the corresponding oxidation degree for each GO (Figure d,e), we noted similar variations. Samples having different oxidation degrees exhibited similar relaxivities; GO_0 showed lower relaxivity values than other GOs; GO_100 having the highest oxidation degree had higher relaxivity values than other GOs, except GO_80. From the above observations, it can be said that both oxidation degree and Mn2+ concentration contribute to the relaxation behavior.

Figure 6

(a, d) Plots of r1 and r2 relaxivities as a function of Mn2+ concentration; (b, c) plots of r1 and r2 relaxivities as a function of oxidation degree; (e, f) plots of r1 and r2 relaxivities as a function of the product of mean-normalized Mn2+ concentration and oxidation degree to evaluate their correlation; (g, h) three-dimensional (3D) plots of r1 and r2 relaxivities with respect to Mn2+ concentration and oxidation degree. Color scheme used: blue square, GO_0; purple square, GO_10; red square, GO_20; green square, GO_30; peacock-green square, GO_50; blue square, GO_80; yellow square, GO_100. T1 and T2 measurements were done at a magnetic field strength of 9.4 T at 298 K.

To decipher the combination of both these factors, the Mn2+ concentration and oxidation degree were normalized by their means and multiplied. The normalization was done by dividing the value of each Mn2+ concentration or the oxidation degree by their respective mean. The relaxivities were then plotted against this product (Figure c,f). The relaxivities were found to have a good correlation with the product of Mn2+ concentration and oxidation degree. From the 3D graphs (Figure g,h), we can see an overall increase in r1 and r2 relaxivities with respect to both the factors. Thus, in any given application, in addition to the concentration of Mn2+, it is important to take into account the oxidation degree of GO involved to tune its relaxivity. A variation of ∼30% in the oxidation degree in the GO shown translates to a 5-fold variation in r1 and r2 values.

The ability to change the relaxivity of GO was further evaluated for its contrast enhancement ability in MRI (Figure ). Phantom images were acquired and processed using RadiAnt DICOM Viewer with the intensities measured for the chosen region of interest (ROI) (as illustrated in Figure S8). The area taken (area of the ROI) was ∼60% of the total area of the MR image of the cross section of the tube containing the sample, to avoid the peripheral artifacts. The T1 contrast across the GO (Figure ) varied from 81.1 to 570, and the T2 contrast varied from 49.4 to 235.9 (Figure S9). When we compare the T1-weighted MR images (Figure a), we can clearly see better contrast in the case of GO_80 and very less contrast in the case of GO_50. Overall trends in the contrast along the GO series (Figure S9) with respect to blank are similar to their corresponding trend for r1 and r2 proton relaxivities.

Figure 7

Phantom MR images of GO suspensions in DI water acquired with a concentration of 500 μg mL–1. (a) T1-weighted MR images of GO suspensions and corresponding blank (T1-weighted MR image of water), TR = 300 ms, and TE = 20 ms; (b) T2-weighted MR images of GO suspensions and corresponding blank (T2-weighted MR image of water), TE = 160 ms, and TR = 3000 ms. MR imaging was performed at a magnetic field strength of 3 T at 296 K.

Kinetic inertness of a material is a very important factor that should be accounted while designing an MRI contrast agent. Toward this end, the r1 and r2 values of the aqueous solutions were measured on day 1 and after 14 days of making the sample. This is shown in Figure S10. No significant changes in the values could be observed, indicating that Mn2+ does not leach out of the GO with time. This also shows that Mn2+ ions are strongly embedded between the GO layers.

In recent years, a wide range of T1 and T2 contrast agents have been developed for MRI, such as gadolinium oxide nanoparticles,[32,33] iron oxide nanoparticles,[34,35] manganese-based nanoparticles,[36−38] and metallic core–shell nanoparticles.[39,40] Apart from the paramagnetic or superparamagnetic contrast agents, other types of CAs include particulate contrast agents (like fluorinated and nonfluorinated paramagnetic micelles or liposomes), diamagnetic hyperpolarization probes (gases and aerosols), metalloporphyrins,[41,42] and carbon nanotubes.[43] Wang et al. successfully synthesized water-dispersible Gd2O3/GO nanocomposites, which exhibited enhanced longitudinal relaxivity compared to previous Gadolinium-based MRI CAs.[11] Venkatesha et. al. (2016) illustrated CoFe2O4-decorated GO as T1 and T2 CA for MRI.[10] Fe3O4/GO composite has been exemplified as a T2 CA.[44] Peng et al. (2015) proposed a co-loading of GO with Mn-doped Fe3O4 (T2 agent) and MnO (T1 agent) magnetic nanoparticles, which exhibited T1 and T2 MRI contrast enhancement.[12]

High relaxivity values can be achieved when GO is doped with paramagnetic and superparamagnetic nanoparticles. The probable ways in which GO plays a significant role in these complexes is by providing an optimal arrangement to the loaded nanoparticles, allowing them to generate high magnetic field gradients or restricting the movement of water molecules by their entrapment between the layers, thus increasing particle and proton interaction. In the present study, the interlayer spacing of GO is ∼10 Å (based on XRD) and the size of water molecule is ∼2.75 Å. Thus, water can easily move between the layers,[45] allowing interaction of the water molecules with the intercalated paramagnetic Mn ions.

To compare the relaxivity values of the GO series prepared in this study with previously reported paramagnetic and superparamagnetic MRI contrast agents, the r1 and r2 relaxivities of the GO samples were found by fitting the graphs of inverse of the longitudinal and transverse relaxation times vs the Mn2+ concentration of the samples (Table S5). The range of r1 was 0.018–0.488 mM–1 s–1, and that of r2 was 0.193–4.159 mM–1 s–1 for the GO series. Table exhibits a comparison of GO_80, which has the highest relaxivity values in the GO series with a few previously reported MRI contrast agents. GO_80 has an r1 relaxivity value comparable to all of the contrast agents enlisted in Table . The r2 relaxivity value of GO_80 is comparable to the Mn-based contrast agents but lesser than ZnO@Gd2O3, ZnO@Gd2O3@FA, and PEG-PIONs/DOX complex. However, the relaxivity values of GO_80 indicate that GO can act as both T1 and T2 MRI contrast agent for detecting hepatic lesions and scanning gastrointestinal tract.[38] On plotting the r1 relaxivity values of the GO samples at 3T, we observed better r1 values than the values obtained at 9.4 T, as shown in Figure S11. GO_80 is expected to confer a good T1 contrast at 3 T.

Table 1

Comparison of r1 and r2 Relaxivities of GO_80 and Previously Reported Nanoparticle Systems for MRI

sample	r1 (mM–1 s–1)	r2 (mM–1 s–1)	field (T)	ref
GO_80	0.49	4.16	9.4
HMnO@mSiO2	0.99	11.02	11.7	(46)
MnO@PEG-phospholipid	0.11	6.16	11.7	(46)
MnO@mSiO2	0.65	9.5	11.7	(46)
MnO@dSiO2	0.08	2.27	11.7	(46)
ZnO@Gd2O3	4.2	192	9.4	(47)
ZnO@Gd2O3@FA	0.15	36	9.4	(47)
PEG-PIONs/DOX complex	1.4	179.07	9.4	(48)

The contribution of GO to relaxivity, as revealed by the presented work, arises from the contribution of oxidation degree. The morphology of GO that we have been able to attain is due to the oxidization of extensively ball-milled graphitic precursor. Although we observed that Mn2+ intercalation is a stochastic process, a very important observation is that the defect density of GO_80, prepared from BMG_80, which was ball-milled for 80 h, is highest among all of the ball-milled graphitic precursors. This may be presumably the reason for the capturing of Mn2+ ions during the oxidation process. The GO_80 sample of the GO series clearly shows the effect of Mn2+, which allows comparison with GO with a low Mn content and a high oxidation degree and GO with a low oxidation degree and a high Mn content. The tuning of oxidation degree and architecture using ball milling and oxidizing of graphite can be strategically used to vary the r1 and r2 relaxivities. BMG_100 has more in-plane defects by virtue of sintering of smaller particles, causing these defects to be sites for surface functional groups, namely, hydroxyl and ketone groups.

Conclusions

The study provides new insights into the origin of proton relaxivity in graphene oxide and demonstrates the combined effect of oxidation degree and Mn2+ concentration on its relaxivity. The surface chemistry of GO can be modulated using the two-step methodology of ball milling graphite followed by oxidation. A variation of ∼30% in the oxidation degree of the GO is shown to translate to 5-fold variation in its relaxivity and image contrast values. This opens up new avenues for developing tunable GO-based contrast agents for different applications in magnetic resonance imaging.

Experimental Section

Synthesis of GO with Varying Oxidation Degree

The strategy for the synthesis of GO samples is depicted in Scheme .[5] It involves a two-step methodology to prepare the GO samples; in the first step, the graphite powder is ball-milled for X hours (X = 10, 20, 30, 50, 80, 100), and then each graphitic precursor is oxidized using Tour’s method[49] of oxidation, which is an improvement over the method described by Hummers and co-workers.[50]

Scheme 1

Schematic Illustration of the Two-Step Methodology for Engineering GO Chemistry Using Ball Milling of Graphite Followed by Oxidation to Graphene Oxide Using Tour’s Method

The GO prepared from graphite is denoted as “GO_0”. The ball-milled graphitic precursors are denoted as “BMG_X”, and the corresponding graphene oxide samples are denoted as “GO_X”. Hence, the nomenclature for the precursor series is graphite, BMG_10, BMG_20, BMG_30, BMG_50, BMG_80, and BMG_100. Likewise, the GO series is GO_0, GO_10, GO_20, GO_30, GO_50, GO_80, and GO_100.

Planetary Ball Milling

In the planetary ball milling, a powder mixture is placed in a ball mill and exposed to high-energy collisions with stainless steel balls. Equal weight of graphite powder is taken in bowls and filled with hexane. Hexane is used as the ball milling medium as it has a low boiling point, and the ball-milled graphitic sample can be retrieved easily. The ball milling conditions used were: bowl speed, 700 rpm and plate speed, 300 rpm; each cycle consisted of 15 min ON time and 8 min OFF time.

Tour’s Method of Oxidation

A 1:6 w/w measure of graphitic precursor and potassium permanganate (KMnO4) and a 9:1 v/v measure of sulfuric acid and orthophosphoric acid (total volume = 134 mL) were taken in a beaker placed in an ice bath and continuously stirred for 12 h at 45 °C. The reaction mixture was slowly added to 140 mL of DI ice, followed by addition of 10 mL of hydrogen peroxide. The chrome yellow mixture was allowed to cool down, and 200 mL of deionized (DI) water was added. The final suspension was decanted for 2 days. The sediment was taken and washed thrice with DI water and twice with ethanol and then dried to obtain the graphene oxide powder.

Characterization of GO

Raman Spectroscopy

Raman spectra were acquired using a LabRam HR with 532 nm wavelength. The samples were drop-dried on glass slides to form thin films. The spectra were recorded in the range of 1000–3000 cm–1. The peaks were then deconvoluted to acquire the defect density of the GO samples.

Electrophoretic Light Scattering (ELS)

Zetasizer Nano (Malvern PanAnalytical) was used for the ELS study and to obtain ζ-potentials for the GO samples. Measurements were performed at 25 °C and averaged from three sequential runs. GO suspension in water with a final concentration of 50 μg mL–1 was used for the study.

Atomic Force Microscopy (AFM)

The morphology of the GO samples was investigated using the NX-10 (Park Systems) system. An Al back-coated Si probe (ACTA, AppNano, Inc.) with a resonance frequency of 300 kHz and a nominal spring constant of 40 N m–1 was used. The GO samples were dispersed in water, bath-sonicated, drop-dried on Si wafers, and then gently rinsed with DI water. The images were obtained at a scan rate of 0.8 Hz.

Solid-State NMR (ssNMR) Spectroscopy

An EXCII JEOL 400 MHz spectrometer with a 4 mm magic angle spinning (MAS) probe was used for ssNMR spectroscopy. To acquire the ssNMR spectra, a 13C single-pulse experiment was implemented with a relaxation delay period of 5 s between scans, 8192 scans, and a MAS speed of 10 KHz. The ssNMR spectra were obtained to determine the extent of oxidation in the GO samples. The intensity (area) of peaks pertaining to unreacted graphite and that of the carbon backbone peaks corresponding to GO were used for calculation of oxidation degrees of GO along with the data from XPS.

X-ray Photoelectron Spectroscopy (XPS)

The elemental composition of GO was determined using a Kratos X-ray photoelectron spectrometer with Al Kα X-ray source having an excitation energy of 1486.6 eV. For each set of sample, wide spectra and high-resolution C 1s spectra were recorded. The ratio of nonaromatic carbon to total carbon (sum of aromatic carbon and nonaromatic carbon) was calculated by deconvolution of the high-resolution C 1s spectra. This information was used to calculate the oxidation degree of the samples of interest.

From the deconvolution of high-resolution C 1s scans of the GO acquired using XPS, the percentage of aromatic and nonaromatic carbon was measured, and using solid-state NMR spectroscopy (ssNMR), the amount of oxidized graphite could be measured. These two parameters were then used to calculate the oxidation degree using the formula[51]where Iaromatic and Inonaromatic are calculated from XPS, and Igraphite and IGO are calculated from the total area under the peaks in the ssNMR spectrum corresponding to unreacted graphite and GO.

Atomic Absorption Spectroscopy (AAS)

A Thermo Electron Corporation M-Series machine was used for atomic absorption spectroscopy (AAS). Quantitative elemental estimation of Fe2+ and Fe3+ was obtained. Standard solutions of iron were prepared using ferrous ammonium sulfate (FAS) salt in double-distilled water; 1–5 ppm iron solutions were prepared by dilution. The GO samples with known weight were dissolved using nitric acid and then diluted with distilled water. The solutions were filtered using a microfilter to remove undissolved residue. The sample solution absorbance was recorded using an AAS instrument along with standard solutions.

Electron Paramagnetic Resonance (EPR) Spectroscopy

JES-X3 SERIES A SYSTEM was used for electron paramagnetic resonance (EPR) spectroscopy. All of the EPR spectra were acquired at a temperature of 298 K under identical conditions. The operating frequency was ∼9.45 GHz with a modulation frequency of 100 kHz and 0.99 mW power. Calibration was done using a solid sample of diphenyl picrylhydrazyl (DPPH). The field center was 336 mT with a sweep width of 50 mT for acquisition of the EPR signal. For quantification, a standard was prepared with MnCl2 in the KCl matrix with a 10% molar ratio. MnCl2 and KCl were dissolved in water and mixed well and then lyophilized to obtain a uniformly distributed solute in KCl. Three samples of the standard were taken to measure the average spin concentration of the standard. An equal amount of GO samples was taken to quantify the respective average spin concentration, and the Mn2+ concentration was calculated. The EPR spectra were double-integrated to obtain the intensities of the signals, and the spins for the standard sample and the samples were calculated.

Relaxivity Measurements

r1 Relaxivity

The r1 relaxivity was obtained from spin-lattice relaxation time (T1) measurements at 298 K on a Bruker Avance NMR spectrometer operating at a 1H resonance frequency of 400 MHz. The sample was prepared by weighing different amounts of GO and preparing eight dilutions in DI water. For each concentration of the sample, an independent experiment was performed. The T1 relaxation time was measured by a one-dimensional inversion recovery NMR experiment.[52] In this experiment, a 180° radio frequency (r.f.) pulse inverts the bulk magnetization from its equilibrium position along the +z-axis to the −z-axis, followed by a delay. During this delay period, due to T1 relaxation along the longitudinal axis, the magnetization recovers toward the equilibrium state. A 90° r.f. pulse is applied after the delay period to create a transverse magnetization. Following this, the acquisition of the signal is carried out. In each experiment, the value of the delay period is varied, ranging from 0 to 2 s. The delay period at which the recovery of the magnetization from −z-axis goes through a null was considered as “Tnull”, and T1 was calculated using the formula T1 = Tnull/ln 2,[52] and the longitudinal proton relaxation rate R1 was calculated by taking the inverse of T1.

r2 Relaxivity

The r2 relaxivity was obtained from spin–spin relaxation time (T2) measurements at 298 K on a Bruker Avance NMR spectrometer operating at a 1H resonance frequency of 400 MHz. The measurements were carried out using the one-dimensional Carr–Purcell–Meiboom–Gill sequence (CPMG) experiment.[53] In this experiment, a 90° r.f. pulse is applied to create the transverse magnetization, followed by a spin–echo period, which consists of a [delay-180° r.f. pulse-delay] block during the magnetization de-phases in the transverse plane due to T2 relaxation. The spin–echo block is repeated n times in sequence to obtain a total desired delay period for allowing the magnetization to relax. After the repetition of the spin–echo block for a desired number of times, the signal is acquired. A series of spectra are acquired with different relaxation delays (by varying the number of repetitions of the spin–echo block). R2 is obtained by fitting the measured intensities to the relaxation time periods (decreasing monoexponential graph) using the formula y = y0 + A1 × exp(−(x – x0) × R2), to obtain the transverse proton relaxation rate R2 (1/T2), where y is the intensity, y0 is the y intercept, A1 is the amplitude, x is the total evolution time, and x0 is x-offset.[54] The same sample as used for r1 relaxivity measurements was used for T2 measurements. The following relaxation delays were used: 6, 30, 60, 90, 120, 150, 180, 210, 240, 270, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, 2700, 3000 ms.

Magnetic Resonance Imaging (MRI)

The phantom imaging study was performed on a 3 T clinical MRI scanner (Ingenia, Philips, the Netherlands) using a 32-channel transmit-received head coil at 296 K. All of the phantom data were acquired in axial orientation using T1- and T2-weighted imaging sequences. Fixed echo time (TE = 20 ms) and repetition time (TR = 300 ms) were used to obtain the T1-weighted images for various GO suspensions. Similarly, for obtaining T2-weighted images, the fixed echo time was 160 ms and the repetition time was 3000 ms. The other acquisition parameters were slice thickness = 6 mm, FOV = 200 × 200, and matrix size = 128 × 128. The total acquisition time was ∼45 min/phantom study. For obtaining the r1 relaxivity, T1 images for all samples with concentrations varying from 100 to 500 μg mL–1 were acquired at TR = 300, 600, 1000, 1500, 2000, 3000. 5000, and 7000 ms. Information was extracted from the MR images, and the intensities were plotted against the TR values and fitted with exponential function to obtain the T1 relaxation constants. Inverse values of these T1 values were plotted against respective concentration for each sample to determine its r1 relaxivity value. The T2_CALC experiment was implemented with TR = 3000 ms and TE = 20–160 ms to find the T2 relaxation times of the samples, and the r2 relaxivity values for the GO samples were calculated.

Data Analysis

Data plotting and analysis were done using the OriginPro 8.5.1 and GraphPad Prism 5 softwares. Microsoft Excel has also been used for data management and statistical analysis.