Size-tunable hydrophilic cerium oxide nanoparticles as a ‘turn-on’ fluorescence sensor for the rapid detection of ultralow concentrations of vitamin C

Abstract

The novel perspective of cerium dioxide as a fluorescence sensor has been demonstrated in the present study. The green coloured emission associated with the nano-dimensions of ceria has been adopted as an analytical tool to sense vitamin C, which is a biologically important molecule, in dilute concentrations. Ultrafine ceria nanoparticles of average size 2.2 nm have been fabricated by a surfactant assisted thermal decomposition strategy. The particular fashion of attachment of the oleic acid surfactant with ceria resulted in the surface hydrophobicity of the nanoparticles which in turn prevents their interaction with a hydrophilic molecule like vitamin C in the reaction media. In order to tackle the incompatibility of the nanoparticles with water, a hydrophilic surfactant coating has been grafted over their surface via bilayer surface functionalisation. The success of the accomplished strategy has been confirmed by thermogravimetric analyses, zeta potential and contact angle measurements. The redox properties of ceria and its optical properties served as a probe to quantify vitamin C in the concentration range 10⁻⁷ to 10⁻⁴ M with a very low limit of detection (LoD) of 500 nM. The designed sensor exhibits a rapid ‘turn on’ fluorescence response within 30 seconds and the reversibility of its fluorescence even after 5 cycles of vitamin C addition corroborates its reusability. The high selectivity of the sensor to detect vitamin C again highlights its suitability as an analytical tool. The realistic application of the sensor has also been displayed by the quantification of vitamin C in pharmaceutical formulations within acceptable error limits.

---

Introduction

In light of recent advances in science and technology, there is a desire to discover innovative materials as well as to improve existing functionalities capable of imparting a significant scientific breakthroughs. In this respect, exploring the undiscovered applications of current materials beyond their conventional uses is a hot topic of current research. Nanosized cerium dioxide or ceria is a functionally esteemed semiconductor employed for many applications by virtue of its characteristic features like high thermal and chemical stability, redox nature, oxygen storage capacity and other tuneable physiochemical properties.^1–4 Ceria crystallises in the face centered cubic fluorite mode with a Fm3m crystal structure in which Ce⁴⁺ cations occupy the cubic corners, which are in turn coordinated to O²⁻ ions present in the tetrahedral voids.⁵ Besides Ce⁴⁺, cerium ions also exist as Ce³⁺ in the crystal phase of ceria, owing to the lower interconversion energy between the two oxidation states⁶ and thus there is an ease of formation of oxygen vacancies in the crystal. In fact, the wide utility of ceria has mainly been accomplished due to its redox property, which is bestowed by the lattice oxygen vacancies. With respect to this unique redox property, ceria has been traditionally exploited as an automobile exhaust catalyst, anti-oxidant agent, in solid oxide fuel cells (SOFC), as an oxygen gas sensor, in oxygen permeation membrane systems, as a catalyst for reactions involving oxidation etc.^7–10

The foremost enthralment of the redox property of ceria is its tunability with the Ce³⁺ to Ce⁴⁺ ratio in the crystal. With reference to the literature, both the size and morphology of ceria nanostructures are the main factors which determine the amount of Ce³⁺ present in the crystal facets.^5,11 a reduction in the overall crystal dimensions favours a higher Ce³⁺ to Ce⁴⁺ ratio in connection with a high surface to volume ratio. Deshpande et al. observed a 15% increase in Ce³⁺ concentration when the particle size was decreased from 6 nm to 3 nm.⁵ Lee et al. demonstrated that the anti-oxidant properties as well as the ROS scavenging ability of ceria nanoparticles are a function of the crystal diameter.¹² The oxygen vacancy in crystals varies with the exposed crystal planes resulting from the alignment and growth of crystals under different synthetic conditions.¹³ According to previous studies, ceria nanostructures with exposed (100) and (111) crystal facets possess higher oxygen storage capacity and hence offer better catalytic activity.¹⁴ Studies by our group have also implied that ceria nanostructures with a mixed morphology of rods and cubes showed higher activity towards the combustion of diesel soot owing to the higher texture co-efficient of exposed active planes.¹³ In short, ceria is endowed with a tunable redox centre, possessing scope for future practical applications.

Apart from its customary uses, there are few additional examples of the utility of ceria for distinct applications. Hirst et al. revealed the anti-inflammatory properties of ceria nanoparticles by means of their ability to scavenge nitric oxide (NO) free radicals, the critical mediator of inflammation.¹⁵ A ceria based hierarchical mesoporous structure fabricated by Corma et al. exhibited a linear photovoltaic response, offering it as a photoactive semiconductor candidate for the construction of solar-cells.¹⁶ Patil et al. designed ceria as a drug carrier for transporting and releasing human carbonic anhydrase (hCAII) inhibitors which are capable of controlling glaucoma, a medical condition causing blindness.¹⁷ Inverse opal 3DOM ceria films engineered by Waterhouse et al. using a colloidal crystal template approach exhibited photonic band gaps in the visible region, suggesting their application as an optical sensor.¹⁸ Suresh babu et al. tuned ceria as a strong upconversion material capable of killing lung cancer cells and projected its versatility as an anti-cancer agent.¹⁹ On account of the superoxide dismutase (SOD) mimetic activity of ceria, there are also a few reports on its potential for theranostic and therapeutic applications.²⁰ Besides, ceria has also been explored as a sensor for the determination of various chemical moieties which are tagged as either detrimental or indispensable substances.^21–26 With reference to recent literature on ceria, it is apparent that efforts are still in progress to unveil the concealed aspects of ceria so as to widen its technological and commercial importance.

On the subject of the impact of nanotechnology in ceria based research, there have been numerous adoptable synthetic strategies for nanoscale ceria which could successfully fabricate crystals with dimensions as small as 1.8 nm.²⁷ The green coloured emission exhibited by ceria with diminutive dimensions, owing to the augmented Ce³⁺ concentration, further accentuates its versatility as a next generation material.²⁸ In the present context, ceria had been designed as a fluorescence sensing probe for the detection of vitamin C, based on the size induced green emission and its dependence on Ce³⁺ content. Vitamin C or ascorbic acid is a biologically important molecule with multifunctional roles in human beings, such as metabolism, neurotransmission, nutrition, immunisation, electron transport, wound healing etc.^29–31 The human body is not able to synthesise vitamin C and consequently it has to be provided externally through diet, otherwise its deficiency causes serious disorders like scurvy, gingival bleeding etc.³⁰ Due to its medical relevance, there have been plenty of pharmaceutical supplements for vitamin C and as it is prone to easy degradation,³² accurate quantification of this molecule in these supplements is vital to assure their quality. Many analytical techniques have been developed so far towards this goal including electrochemical, spectrophotometric and chromatographic methods, each of them having their own flaws and faults.^31,33,34 Although many of these methods are capable of providing a precise and accurate estimation of vitamin C, the related drawbacks cannot be disregarded when the ultimate goal of an ideal method is being considered. Due to the lack of specificity to vitamin C, some of these methods are inappropriate in the presence of other reducing agents.³⁵ Some methods display drawbacks in terms of cost effectiveness.³¹ The crucial weakness of all these methods is their inability to quantify vitamin C at low analyte concentrations,³⁶ thus restricting their execution in pharmaceutical and food industries as well as for biological evaluation which entails trace analyses. Though electrochemical analysis is one of the prominent techniques in the detection of vitamin C under dilute conditions, it is also not free from flaws. The major drawback of this method is the high overpotential, which causes fouling by the adsorption of oxidation products on the electrode surface and thus decreases the reproducibility and sensitivity of the electrode.³⁰ On the basis of the high industrial demand for an appropriate analytical system capable of fast, sensitive, selective, accurate, miniaturisable and low-cost assessment of vitamin C at analyte concentrations as low as possible, research is still ongoing to meet the requirements for devising an ideal commercial sensor. The numerous research articles on vitamin C sensing in recent years support the above statement and the present study is an attempt to address the quandary based on the use of ceria nanoparticles.

The major challenge in implementing ceria as a sensor for an extremely water soluble molecule like vitamin C is its inherent hydrophobicity. In the absence of any chemical modification, the surface of ceria is hydrophobic due to its unique electronic structure which prevents it from hydrogen bonding with interfacial water.³⁷ Herein we have adopted a bilayer surface modification strategy using oleic acid to render the surface of ceria nanoparticles hydrophilic. The as synthesised water dispersible nanoparticles exhibited green coloured fluorescence in connection with their ultrafine size and their realistic perspective as a vitamin C sensor by modulating their Ce³⁺ content has been demonstrated. To date, there is only one study exploring the capability of ceria for sensing vitamin C by Sharpe et al. and was based on the colorimetric change of ceria nanoparticles after their interaction with antioxidants.³⁸ The detection range was 20 to 400 mM implying their limitation to sense molecules in dilute conditions, which is highly desirable for pharmaceutical estimations. Moreover, certain issues like selectivity have not been addressed, raising concerns over practical applicability. The present work is expected to receive great attention due to the ease and efficiency of the method to sense vitamin C at a concentration as low as 6 μM with high reproducibility, selectivity and rapidity. Besides, to the best of our knowledge, this is the first report to date that projects ceria as a ‘turn-on’ fluorescence sensor by making use of its redox properties.

Experimental

Materials and synthesis

All the chemicals were used as received without further purification. Cerium acetate (99.9%), ascorbic acid (99%), potassium permanganate (99%), citric acid anhydride (99%) and calcium carbonate (99%) were purchased from Merck (India), titanium dioxide (99%) was purchased from CDH Laboratory Reagents, India, D-glucose (>99.5%), diphenyl ether (99%) and oleyl amine (70%) were procured from Sigma Aldrich, oleic acid (90%) was obtained from Alfa Aesar (UK), ammonium hydroxide (25%, analytical grade) was bought from Qualigens Fine Chemicals, India. Common solvents such as acetone, cyclohexane and toluene (analytical grade) were acquired from Merck, India. Branded pharmaceutical formulations as well as supplements of vitamin C in the form of tablets have been supplied by GlaxoSmithKline Pharmaceuticals Limited, India and Wipro Care India, the amount of the vitamin in each tablet was 500 mg and 1 mg, respectively. Double-distilled water using a quartz glass distillation unit was used for the synthesis procedure.

Synthesis of oleophilic ceria nanoparticles

Size controlled synthesis of cerium oxide was carried out by executing a thermal decomposition method. The precursor, cerium acetate, dissolved in diphenyl ether solvent (boiling point ∼ 260 °C), upon heating will undergo decomposition to form the corresponding oxide. Oleic acid, by virtue of the steric effect provided by its long alkyl chain, was used as a surfactant to control the size of the nanocrystals. In a typical synthesis, 0.005 moles of cerium acetate were dissolved in 100 mL diphenyl ether in a round bottom flask. About 0.02 moles oleic acid and 0.023 moles oleylamine were added to the reaction mixture and refluxed for 1 h. Oleic acid, in the presence of oleyl amine, undergoes ionisation to form the corresponding oleate ion which is capable of coordinating with the positive core of the nanoparticles formed. As the reaction progresses, the solution turned brown, indicating the formation of ceria crystals. The mixture was allowed to cool to room temperature 1 h after commencing the thermal decomposition. Acetone was added to the reaction mixture to precipitate the oleic acid coated nanoparticles, which were later separated by centrifugation. The separated nanoparticles were washed thoroughly with acetone to remove excess oleic acid and finally dried in an oven to obtain a pale brown powder which is denoted as MLNP. The successful surface modification was confirmed by dispersing the nanoparticles in toluene, which yielded a transparent and stable dispersion.

Bilayer surface functionalisation of the oleophilic ceria nanoparticles

Oleic acid was again employed as a secondary surfactant to provide a bilayer coating over the nanoparticles. About 1 g of MLNP was combined with 20 mL of distilled water, which was almost immiscible due to the hydrophobic nature of the NP surface. 20 mL of 10% (w/v) ammonium oleate in water was added drop-wise to this and stirred vigorously. Ammonium oleate was prepared by adding an adequate amount of ammonium hydroxide to oleic acid so that the pH of the resultant solution was ∼9, which is mandatory for the ionisation of oleic acid. After stirring the initial mixture of nanoparticles and oleate salt in water for about 3 h, the mixture transformed into a stable suspension signifying the successful bilayer surface modification of oleate ions over the nanoparticles’ surface. The water in the suspension was allowed to evaporate slowly by using a constantly boiling water bath as the heat source. The nanoparticles obtained were washed thoroughly with cyclohexane, followed by acetone to remove any uncoated particles as well as unreacted oleate ions, which were later dried and redispersed in water to yield a stable dispersion. The sample is named as BLNP henceforth.

Preparation of ceria dispersions for sensing vitamin C at different concentrations

A parent dispersion of the nanoparticles in water was prepared with a concentration of 0.0008 M, the photoluminescence spectra of which was recorded. To 2 mL of this dispersion, different volumes (0.1 mL to 0.5 mL) of 0.5 mM potassium permanganate solution were added and its photoluminescence was continuously monitored to optimise the volume of permanganate solution required to quench its fluorescence. A set of vitamin C solutions in water at concentrations in the range of μM to mM were prepared and 2 mL of each solution was added to 2 mL of the ceria dispersion whose fluorescence had been quenched by the addition of permanganate. The quantitative estimation of vitamin C was carried out by the acquisition of the photoluminescence spectra of the respective samples. For demonstrating the selectivity of the sensor towards vitamin C, different concentrations of common pharmaceutical ingredients such as citric acid, calcium hydroxide, glucose and titanium dioxide in water were added to the BLNP dispersion and the PL spectra of these were acquired. For the analysis of commercial samples, the tablets under the study were dissolved in water so that their concentration fell in the μM range and 2 mL of this solution was mixed with 2 mL of the fluorescence quenched BLNP dispersion, whose photoluminescence intensity was also acquired.

Instrumental techniques

The X-ray diffraction (XRD) patterns of the powder specimens were obtained using a Philips X’PERT PRO diffractometer with Ni-filtered CuKα1 radiation (λ = 1.5406 Å) using a 30 mA current at 40 kV. The continuous scan was performed over the 5–100 degree (2θ) range at a scanning speed of 2 degree per min and a step size of 0.04°. The thermogravimetric analyses (TGA) of the powders were carried out using a Perkin Elmer, STA 6000 simultaneous DTA–TGA in an ambient atmosphere heated at a constant ramp of 10 °C min⁻¹ under an air purge. The morphology and average size of the nanocrystals were investigated by high resolution transmission electron microscopy (HR-TEM) using a FEI Tecnai 30 G2 S-Twin microscope operated at 300 kV and equipped with a Gatan CCD camera. The zeta potential as well as size of the cerium oxide nanoparticles in the suspension were measured by photon correlation spectroscopy (PCS) at 25 °C on a Zetasizer 3000 HSA, Malvern Instruments, Worcestershire, UK using a 60 mW He–Ne laser producing a wavelength of 633 nm with General Purpose algorithm with Dispersion Technology Software (v. 1.61) at a 90° detection angle. A minimum of seven measurements for each sample were taken to ensure statistical significance. The water contact angle measurements were made on glass substrates coated with aqueous and nonaqueous nanoparticle dispersions of the same concentration by sessile drop method in a Data Physics DCAT21 Dynamic Contact Angle Meter which is equipped with a contact angle goniometer and high resolution cameras and software to capture the profile of a pure liquid on a solid substrate. A micro syringe steel needle (Hamilton) of capacity 500 μL was positioned above the surface of the coated glass slide, and a drop of the test liquid (3 μL) was dispensed at a rate of 1 μL s⁻¹. After dispensing the drop, its shape was monitored with a digital camera, and the contact angle was determined by aligning the tangent of the sessile drop profile at the contact point with the surface. The angle formed between the liquid/solid interface and the liquid/vapour interface is the contact angle and was recorded as an average of 5 measurements. The colloidal stability of the nanoparticle dispersion was monitored with a nephelometer (CL 52D, ELICO, India) as the intensity of transmitted visible light through the fluid against time. The absorption spectra of the samples were obtained using a UV-visible 2401 PC spectrophotometer (Shimadzu, Japan) in the wavelength range 200–800 nm. The photoluminescence (PL) spectra of the nanoparticle dispersions (MLNP in toluene and BLNP in water) were taken at room temperature using a Cary Eclipse spectrofluorometer (Varian, Australia) with an excitation wavelength of 400 nm, unless otherwise specified. For each measurement, three replicates were carried out and the final value was represented in terms of standard deviation of the replicates. X-ray photoelectron spectroscopy (XPS) spectra of the particle powders were recorded on a VG Microtech Multilab ESCA 3000 spectrometer, maintaining a base pressure of the analysis chamber in the range of 3 to 6 × 10⁻¹⁰ Torr. Mg Kα (1253.6 eV) was used as the X-ray source. Binding energy (BE) calibration was performed with Au 4f_7/2 core level at 83.9 eV. The XPS data was deconvoluted with XPSPEAK 4.1 software which produced stable and almost superimposable baselines, confirming the stability of the fits and helping to validate the interpretation.

Results and discussion

Synthesis and surface functionalisation of nanoparticles

Cerium oxide nanoparticles were prepared by thermally decomposing cerium acetate in diphenyl ether at 260 °C, which is the natural boiling point of the solvent. The choice of the synthetic strategy is due to its ability to produce highly crystalline and monodisperse nanoparticles compared to other methods.³⁹ As already reported, the decomposition of cerium acetate proceeds via the formation of free radical intermediates, as shown in the equations below.⁴⁰

M–OOCR → M˙ + R–COO˙ (1)

M–OOCR → MO˙ + R–CO˙ (2)

M˙ + MO˙ → M–O–M (3)

The association between Ce˙ and CeO˙ free radicals resulted in Ce–O–Ce linkages, which ended up with the nucleation of CeO₂ crystals.⁴⁰ The function of the surfactant, oleic acid, is to control the growth of nanocrystals after nucleation and thus ensure the ultrafine dimensions of the formed nanoparticles. The oleate ions formed due to the ionisation of oleic acid in the presence of oleyl amine, possess a negatively charged head, corresponding to the carboxylate group and a non-polar tail comprising the longer alkyl chain.²⁸ During the course of the reaction, the negative oleate head becomes attached to the nucleated metal oxide core, which is usually positive in nature.⁴¹ According to previous studies, restriction of crystal growth, imparted by the bulky as well as the kinked structure of oleic acid, triggers the size controlled formation of nanoparticles.⁴² The particular mode of attachment of the oleate ion with the nanoparticle resulted in the protrusion of the non-polar alkyl chains over the particle surface, thus inducing surface hydrophobicity. As the dispersibility of the as-prepared nanoparticles is limited to organic solvents, their incompatibility with water is in turn a hurdle for applications intended in aqueous media. In order to make the particles attuned to hydrophilic surroundings, slight polarity has been induced over the surface of the nanoparticles by an oleic acid mediated bilayer surface encapsulation. The conceptual pathway of the synthetic strategy is depicted in Scheme 1.

Scheme 1 Synthetic protocol of the bilayer surface functionalised hydrophilic nanoparticles.

The long, non-polar alkyl chain of the oleate ion in the monolayer intertwines with those of the oleate ions that have been added during the bilayer surface modification process. This interlink has been effected by the hydrophobic–hydrophobic interaction between the two alkyl chains⁴³ which ultimately resulted in the outward projection of the carboxylate group of the oleate ion over the nanoparticle surface. The polarity brought about by the C–O– bond in the carboxylate ions reversed the characteristics of the nanoparticle surface from hydrophobic to hydrophilic. The visual change in the appearance of the nanoparticles after bilayer modification is depicted in Fig. S1 of the ESI.† It could be observed that whereas MLNP has a sticky nature, BLNP is more powder like. This serves as initial evidence for the change in the nanoparticle surface from hydrophobic to hydrophilic. The dispersibility of BLNP in water (Fig. S1†) could be considered as confirmation of the success of the synthetic strategy.

Preliminary characterisation

The XRD pattern of BLNP shown in Fig. 1 reveals that the core of the nanoparticle belongs to the pure face-centered cubic fluorite phase of ceria with the space group Fm3̄m (JCPDS no. 43-1002).⁴⁴ The characteristic peaks in the diffractograms have been indexed to the reflections from the crystal planes in the (111), (220), (311) and (331) directions as denoted in the figure.⁴⁵ The extremely broadened XRD peaks indicate the nanoscale of the crystallites and the D_XRD calculated using the Scherrer equation is 1.6 nm.

Fig. 1 X-ray diffraction pattern of BLNP showing the characteristics peaks corresponding to the crystal facets of ceria.

HR-TEM images of BLNP, as shown in Fig. 2, illustrate the spherical morphology of the particles and the calculated average size of 100 particles is 2.2 ± 0.2 nm. The relatively low contrast in the TEM image of the crystal is apparently due to the presence of the surfactant, oleic acid, over the surface of the nanostructure.²⁸ The uniform size distribution and good separation of the particles without any apparent interaction as revealed by TEM is credited to the surface modification facilitated by oleic acid, thus substantiating its role in the size controlled synthesis. The visual examination of the exposed crystal facets of BLNP indicates that the lattice fringes correspond to predominant (111) and (100) planes of ceria with corresponding interplanar spacings of 0.31 nm and 0.27 nm respectively, indicating the phase purity of the crystal structure.

Fig. 2 HR-TEM images of BLNP (a) showing the size of particles (b) indicating the exposed facets of the crystal.

The hydrodynamic sizes of the particles in aqueous (12.6 nm) and non-aqueous dispersions are shown in Fig. 3. The D_PCS value of the MLNP nanoparticles (4.5 nm) is slightly larger than the D_TEM value because D_TEM provides only the size of the core nanoparticle, and D_PCS considers the size of the core–shell structure comprising both nanoparticle and surfactant.⁴⁶ The value of D_PCS is in good agreement with D_TEM, when the chain length of the surfactant on the NP surface is considered.

Fig. 3 Photon correlation spectra of MLNP and BLNP in toluene and water, respectively, for deriving hydrodynamic diameters.

The D_PCS of the nanoparticle in aqueous dispersion (BLNP) was higher than that in non-aqueous dispersion and the increase is brought about by the double layer of oleic acid over the nanoparticle surface in the aqueous media compared to the monolayer of surfactant in oleophilic MLNP.

Surface properties of the nanoparticles

The thermogravimetric profiles of the samples and the zeta potential of the aqueous BLNP dispersion (pH ∼ 7) are shown in Fig. 4. Panel (a) shows a significant loss in weight between 240 and 450 °C on account of the removal of surfactants from the surface of the nanoparticles. MLNP showed a single step weight loss corresponding to the monolayer of chemisorbed oleate ions over the nanoparticles.

Fig. 4 (a) Thermogravimetric profiles of MLNP and BLNP, and (b) zeta potential measurement of BLNP in water.

The bilayers of surfactant in BLNP resulted in a two step weight loss in the thermal stability data, the first and the second step being related to the outer and inner layers, respectively. As the outer layer is linked to the inner layer by relatively weak hydrophobic–hydrophobic interactions due to partial interpenetration of the hydrocarbon tails, the oleate ions from the outer layer could get rid of the attractive force and undergo decomposition at a lower temperature (316 °C) compared to those in the inner layer (400 °C). In the primary layer, owing to the strong chemisorption between the oleate ions and the nanoparticle surface, the surfactant molecules undergo decomposition at a slightly higher temperature, as implied by the TGA. In fact, the two step weight loss in BLNP implies the effectiveness of surface functionalisation offered by the adopted synthetic strategy. In order to evaluate the surface charge of the particles after bilayer surface functionalisation, the zeta potential of the aqueous BLNP dispersion (pH ∼ 7) has been measured (Fig. 4, panel b). The particles show a negative zeta potential of about −27 mV, which is attributed to the negative carboxylate groups projecting out of the nanoparticle surface after bilayer modification. The value, which in turn implies the polarity over the surface, confirms the successful grafting of the hydrophilic layer over the nanoparticles by functionalisation.

The change in the nature of the nanoparticles from hydrophobic to hydrophilic has also been demonstrated in Fig. 5, which shows the water contact angle measurements of glass slides coated with dispersions of MLNP and BLNP. The wetting behaviour exhibited by the particles before and after bilayer functionalisation is drastically different. Whereas the MLNP coated surface made a contact angle of 97° with water, which falls in the hydrophobic range, the glass slide with BLNP over its surface exhibited a hydrophilic contact angle of 50°. The difference in wettability of the two samples is due to the change in the functional groups over the nanoparticle surfaces brought about by surface modification. In MLNP, due to the non-polar hydrocarbon chain over the nanoparticles, the surface is rather hydrophobic, as shown in Fig. 5a.

Fig. 5 Sessile drop water contact angle measurements of glass slide coated with (a) MLNP and (b) BLNP.

On the other hand, the surface polarity of BLNP as endowed by the carboxylate group, due to the reverse manner of attachment of the alkyl chains in the two layers, rendered the glass surface hydrophilic. The hydrophilic surface of the nanoparticles facilitated their dispersion in water (Fig. S1†), which was stable over a period of several weeks. The dispersion did not show any visible change in appearance or formation of a precipitate during ageing. The zeta potential value of the dispersion, −27 mV, also indicates the stability of the dispersion, as a value in the range of 25–30 mV signifies the electrostatic stabilisation of the nanoparticles.⁴⁷

The turbidity measurement data of the dispersion for more than a month are presented in Table 1. Although, the initial value of turbidity i.e., 22 NTU slightly decreased to 19 NTU after 1 week, later, the dispersion showed no further precipitation for more than 1 month. Thus the turbidity data, along with the zeta potential of the dispersion, corroborates the higher stability offered by the hydrophilic nanoparticles in water.

Table 1 Turbidity measurement data of BLNP aqueous dispersion

Ageing time (days)	Turbidity (NTU)	Decrease in turbidity (%)
1	22	0
7	19	13
14	19	13
21	19	13
28	19	13
35	19	13

---

Optical properties of the nanoparticles

The absorption and emission spectra of BLNP are depicted in Fig. 6. The particles showed an absorption edge at ∼400 nm and the corresponding emission was in the visible region with an emission maximum at 515 nm, which is in the green region. Ceria usually absorbs in the UV region on account of the charge-transfer transition from O₂ (2p) to Ce⁴⁺ (4f) orbitals.¹⁶ Hence, the optical properties of BLNP show a redshift with respect to bulk ceria towards the visible region. Ultrafine ceria has already been reported to show a size related redshift in absorption and associated green coloured emission.²⁸ The redshift is attributed to the higher amount of Ce³⁺ in the crystals associated with the higher surface to volume ratio.

Fig. 6 Absorption and emission spectra of the aqueous dispersion of BLNP showing a redshift from the UV region. The inset shows a visual image of the green coloured emission.

In ceria, the surface entropy of Ce³⁺ ions is higher than that of the bulk which induces most of the surface atoms to exist in the +3 oxidation state. With an increase in surface area, there will be a subsequent increase in Ce³⁺ to Ce⁴⁺ ratio. The Ce³⁺ ions in the crystal lattice create an intermediate band in between the valence and conduction bands of ceria. As a result, the excitation of electrons from the valence band occurs to this new band, which is of lower energy than the conduction band. Thus the overall energy of absorbance was lowered, resulting in a redshift in the optical properties. Consequently, the absorption edge of BLNP is shifted to 400 nm with a strong emission in the visible region. Thus the size dependent green coloured fluorescence exhibited by the particles is due to the presence of Ce³⁺ in the crystal. The higher Ce³⁺ content of ∼38% as revealed by the XPS spectra of BLNP (Fig. S2†) apparently supports the proposed conjecture.

Sensing of vitamin C by the nanoparticles

With respect to the optical properties of the nanoparticles, the role of Ce³⁺ is pivotal in the fluorescence exhibited by the nanoparticles. The dependence of fluorescence on the Ce³⁺ content has been verified by monitoring the emission intensity of the nanoparticles at various Ce³⁺ concentrations. Being an oxidising agent, KMnO₄ has been used as a mediator to vary the amount of Ce³⁺ in the BLNP dispersion. KMnO₄ converts Ce³⁺ to Ce⁴⁺ according to the equation⁴⁸

3Ce³⁺ + MnO₄⁻ + 2H₂O → 3Ce⁴⁺ + Mn⁴⁺ + 4OH⁻ (4)

Different amounts of 0.0005 M KMnO₄ were added to BLNP dispersions, the PL spectra of which are shown in Fig. 7. It is observed that the PL spectra show a relative decrease in intensity with respect to the amount of KMnO₄ added. On the basis of available literature mentioning the role of Ce³⁺ in the fluorescence properties of ceria,^28,49,50 it could be assumed that as Ce³⁺ is oxidised to Ce⁴⁺ by KMnO₄, the abundance of the intermediate energy level created by Ce³⁺ in the band gap will be inversely affected.

Fig. 7 The PL spectra of the BLNP dispersion (in water) showing quenching of fluorescence with the addition of different volumes of KMnO₄ solution.

The consequential decrease in exciton transfer to the Ce³⁺ band is responsible for the diminishing of fluorescence exhibited by the nanoparticles. The oxidation of Ce³⁺ to Ce⁴⁺ could also be observed visually by examining the colour change of the BLNP dispersion. The BLNP dispersion which was initially bluish-white in colour gradually changed to yellowish, after the addition of KMnO₄, indicating the formation of Ce⁴⁺ in the dispersion. The optical photograph of the dispersions before and after the addition of KMnO₄ is supplied as Fig. S3.† The amount of KMnO₄ needed to quench the fluorescence of the BLNP dispersion under the preferred experimental conditions has been optimised as 0.5 mL (0.25 μmoles of KMnO₄), the addition of which extinguished the fluorescence of the nanoparticles almost completely.

The principle of the present nanoparticle based sensor is based on the fact that the reverted conversion of Ce⁴⁺ to Ce³⁺ by vitamin C can restore the quenched fluorescence to its initial vigour. The structure of vitamin C is such that the two enolic hydroxyl groups attached to the C3 and C2 carbons of the lactone ring are prone to release electrons under favourable conditions so as to form a comparatively stable and oxidised deprotonated intermediate.⁵¹ The sensing of vitamin C by nanoparticles stems from its ability to convert the oxidised Ce⁴⁺ back to Ce³⁺ by the donation of an electron, as represented in Scheme 2. The stoichiometric equation displaying the interaction between vitamin C and cerium ions is shown as eqn (5)⁴⁸

2Ce⁴⁺ + C₆H₈O₆ (vitamin C) → 2Ce³⁺ + C₆H₆O₆ + 2H⁺ (5)

Scheme 2 Representation of the principle behind the sensing of vitamin C by ceria nanoparticles.

According to the literature,^28,49,50 as the creation of the intermediate band (IB) is based on the concentration of Ce³⁺, we could propose that when Ce⁴⁺ is converted back to Ce³⁺, the IB in the band gap is regenerated, which establishes the initial fluorescence. Different amounts of vitamin C in water at concentrations in the range of μM to mM have been added to BLNP dispersions, whose fluorescence has been quenched by the oxidation of Ce³⁺. Upon addition of vitamin C, the oxidised cerium ions (Ce⁴⁺) will return to their initial +3 state along with the return of the fluorescence, the emission intensity being decided by the amount of vitamin C added. The analysis of the PL spectra of the respective dispersions in fact confirms that the intensity of emission by the nanoparticles is a function of the concentration of vitamin C present in the dispersion.

Variation in PL intensity of BLNP and the linear fit of PL intensity plotted against vitamin C concentration in the low concentration range are shown in Fig. 8. The PL intensity of the dispersion showed a linear correlation with the concentration of vitamin C. Linearity is observed over a wide range of concentrations of vitamin C solutions (μM to mM). However, there is no observable increase in PL intensity upon addition of vitamin C solutions with concentrations above ∼2.6 × 10⁻⁴ M. It is noteworthy that, at this concentration, the PL spectra have almost regained their initial intensity implying complete retrieval of Ce⁴⁺ to Ce³⁺. The variation in PL intensity with the addition of ultra dilute vitamin C solutions (μM) is perfectly linear as seen from the linear regression (panel b) with an R² value of 0.9963. The linearity implies the efficiency of the proposed sensor to quantify vitamin C under ultra dilute conditions. The limit of detection (LoD) of the sensor, which is the lowest analyte concentration that can be measured reliably by the proposed method, has been calculated as 3 times the standard deviation (σ) for replicates of the blank.⁴⁸ The combination of very small scatter (σ), 0.17, and the ability to detect reproducibly at much lower concentration levels than already reported, positions our nanoparticle based sensor as very promising to estimate vitamin C content in solutions as dilute as 500 nM. In short, the dynamic range of detection of the sensor is assigned as 10⁻⁷ to 10⁻⁴ M (0.02 to 17.6 ppm).

Fig. 8 (a) Variation in PL intensity of BLNP with the addition of different concentrations of vitamin C solution, and (b) linear fit of the PL intensity plotted against molarity of vitamin C in the low concentration range.

The selectivity of the sensor towards vitamin C in pharmaceutical formulations and the response time of the sensor detecting vitamin C are demonstrated in Fig. 9. Common pharmaceutical ingredients which are used as sweetening agents, stabilisers and binders have been employed along with BLNP and the PL spectra of these were acquired. It was observed that almost all of the ingredients did not show any interaction with the nanoparticles so as to create any impact on their PL spectra. It is also revealed that none of the ingredients interferes with the interaction between vitamin C and the nanoparticles, as implied by the retention of the linear progression of PL intensity with vitamin C concentration in the presence of all ingredients.

Fig. 9 (a) Variation in PL intensity of BLNP with the addition of different pharmaceutical ingredients at different concentrations and with the addition of vitamin C along with all ingredients. (b) Time dependent variation in PL intensity of the BLNP dispersion with the addition of KMnO₄ and vitamin C.

The PL intensity of the dispersion as a function of time has been monitored and during the approach, the fluorescence was quenched and reactivated by the addition of KMnO₄ and vitamin C. It is observed that the PL intensity which was diminished by the addition of KMnO₄ could be resumed within 30 seconds after the addition of vitamin C (panel b of Fig. 9). In order to highlight the reusability of the sensor, a study involving five consecutive additions of KMnO₄ followed by vitamin C addition to the dispersion was executed and the resultant PL intensity data are featured in Fig. 10. The turn on/off behaviour of the nanoparticle emission intensity was retained even after 5 cycles, without much difference in the initial intensities. Thus the reversibility of fluorescence exhibited by the nanoparticles enhances its long term usability, which is one of the prerequisites for a desirable sensor. An attempt was made to elucidate the amount of vitamin C in some pharmaceutical formulations by means of the described sensing procedure (see the ESI†). The results for the evaluation of vitamin C in real samples are tabulated in Table 2.

Fig. 10 Variation in PL intensity of BLNP with the alternating addition of KMnO₄ followed by vitamin C.

Table 2 Summary of the quantitative estimation of vitamin C in commercial samples

Sample	Certified amount of vitamin C per tablet (mg)	Evaluated amount of vitamin C per tablet (mg)	Recovery (%)	RSD (%)
Tablet 1	500	528	105	0.107
Tablet 2	1	0.98	98	0.092

---

The satisfactory RSD (relative standard deviation) and % recovery substantiate the reliable estimation of vitamin C in commercial formulations within acceptable error limits. The practical applicability of the method has been validated by the analysis of real samples. In short, without any tedious methodology, expensive instrumentation or storage protocols, the presented method could rapidly and repeatedly sense vitamin C with high accuracy.

Conclusions

A novel turn-on fluorescence sensor based on cerium dioxide has been devised by a simple two step synthetic strategy. The principle of the sensor is based on the size induced redox property of ceria and its fluorescence. Nanoscale ceria with an average size ∼2.2 nm has been engineered by thermally decomposing cerium acetate in diphenyl ether in the presence of oleic acid as surfactant. The oleophilic surface of the as synthesised nanoparticles has been converted to hydrophilic by bilayer surface functionalisation. The consequent change in its properties was reflected in its surface polarity as well as wetting behaviour as revealed by the zeta potential and contact angle measurements. One of the appealing achievements of the demonstrated method is that, while the bilayer surface functionalisation allowed the surface of the particles to become hydrophilic, there was no inverse effect on the behaviour of the central ceria core, including its optical and redox properties. The implemented hydrophilic features enabled the water compatibility of the nanoparticles to yield its aqueous dispersion, which exhibited long term stability. The emission intensity of the nanoparticles showed a quenching behaviour with respect to the oxidation of Ce³⁺ to Ce⁴⁺ in the nanoparticle dispersion by KMnO₄. The reversal of the oxidation state from Ce⁴⁺ to Ce³⁺ with the addition of vitamin C showed a linear increase in fluorescence intensity with its concentration, thus enabling its quantitative assessment. The fascination of the proposed method is its efficiency to sense vitamin C at a concentration as low as 500 nM with high reproducibility, selectivity and rapidity, without any tedious protocol. The intriguing possibilities of the method for practical application have also been demonstrated by satisfactorily quantifying vitamin C in commercial supplements with high accuracy.

Acknowledgements

The authors thank CSIR-NIIST for providing the necessary facilities for the work and the CSIR–Central Glass & Ceramic Research Institute for the same. Miss. Srividhya J. Iyengar of CSIR-CGCRI, Kolkata is acknowledged for the XPS measurements. Mr M. Kiran and Ms Remya are acknowledged for HR-TEM imaging and zeta potential measurements. Authors gratefully acknowledge Dr Sailaja G. S. of CUSAT, Dr Prabha D. Nair and Ms Nimi of SCTIMST for water contact angle measurements. Authors A. K. and T. S. S. acknowledge CSIR for the CSIR Fellowships. This work was partly funded by the Indian Rare Earths Limited Technology Development Council (IRELTDC), DAE, India.

Notes and references

1. A. Ahniyaz, Y. Sakamoto and L. Bergstrom, Cryst. Growth Des., 2008, 8, 1798.
2. A. Trovarelli, C. de Leitenburg, M. Boaro and G. Dolcetti, Catal. Today, 1999, 50, 353.
3. M. Yamashita, K. Kameyama, S. Yabe, S. Yoshida, Y. Fujishiro, T. Kawai and T. Sato, J. Mater. Sci., 2002, 37, 683.
4. H. Imagawa, A. Suda, K. Yamamura and S. Sun, J. Phys. Chem. C, 2011, 115, 1740.
5. S. Deshpande, S. Patil, S. Kuchibhatla and S. Seal, Appl. Phys. Lett., 2005, 87, 133113.
6. C. Sun, H. Li and L. Chen, Energy Environ. Sci., 2012, 5, 8475.
7. T. Garcia, B. Solsona and S. H. Taylor, Catal. Lett., 2005, 105, 183.
8. S. Maensiri, C. Masingboon, P. Laokul, W. Jareonboon, V. Promarak, P. Anderson and S. Seraphin, Cryst. Growth Des., 2007, 7, 950.
9. P. Palmisano, N. Russo, P. Fino, D. Fino and C. Badini, Appl. Catal., B, 2006, 69, 85.
10. M. G. Bellino, D. G. Lamas and N. E. de Reca, Adv. Funct. Mater., 2006, 16, 107.
11. H. X. Mai, H. L. D. Sun, Y. W. Zhang, R. Si, W. Feng, H. P. Zhang, H. C. Liu and C. H. Yan, J. Phys. Chem. B, 2005, 109, 24380.
12. S. S. Lee, W. Song, M. Cho, H. L. Puppala, N. Phuc, H. Zhu, L. Segatori and V. L. Colvin, ACS Nano, 2013, 7, 9693.
13. T. S. Sreeremya, A. Krishnan, K. C. Remani, K. R. Patil, D. F. Brougham and S. Ghosh, ACS Appl. Mater. Interfaces, 2015, 7, 8545.
14. A. Younis, D. Chu, Y. V. Kaneti and S. Li, Nanoscale, 2016, 8, 378.
15. S. M. Hirst, A. S. Karakoti, R. D. Tyler, N. Sriranganathan, S. Seal and C. M. Reilly, Small, 2009, 5, 2848.
16. A. Corma, P. Atienzar, H. Garcia and J. Y. Chane-Ching, Nat. Mater., 2004, 3, 394.
17. S. Patil, S. Reshetnikov, M. K. Haldar, S. Seal and S. Mallik, J. Phys. Chem. C, 2007, 111, 8437.
18. G. I. N. Waterhouse, J. B. Metson, H. Idriss and D. Waterhouse, Chem. Mater., 2008, 20, 1183.
19. S. Babu, J. H. Cho, J. M. Dowding, E. Heckert, C. Komanski, S. Das, J. Colon, C. H. Baker, M. Bass, W. T. Self and S. Seal, Chem. Commun., 2010, 46, 6915.
20. I. Celardo, J. Z. Pedersen, E. Traversa and L. Ghibelli, Nanoscale, 2011, 3, 1411.
21. P. Jasinski, T. Suzuki and H. U. Anderson, Sens. Actuators, B, 2003, 95, 73–77.
22. S. Chakraborthy and A. Sen, Sens. Lett., 2009, 7, 91.
23. M. B. Gumpu, N. Nesakumar, S. Sethuraman, U. M. Krishnan and J. B. B. Rayappan, Sens. Actuators, B, 2014, 199, 330.
24. M. Hosseini, M. R. K. Pur, P. Norouzi, M. R. Moghaddam, F. Faridbod, M. R. Ganjali and J. Shamsi, Anal. Methods, 2015, 7, 1936.
25. M. Singh, N. Nesakumar, S. Sethuraman, U. M. Krishnan and J. B. B. Rayappan, J. Colloid Interface Sci., 2014, 425, 52.
26. D. Zhang, W. Wu, X. Ni, X. Cao, X. Zhang, X. Xu, S. Li, G. Han, A. Ying and Z. Tong, J. Mater. Sci., 2009, 44, 3344.
27. S. H. Yu, H. Colfen and A. Fischer, Colloids Surf., A, 2004, 243, 49.
28. A. Krishnan, T. S. Sreeremya, E. Murray and S. Ghosh, J. Colloid Interface Sci., 2013, 389, 16.
29. A. Barberis, G. Bazzu, G. Calia, G. M. G. Puggioni, G. G. Rocchitta, R. Migheli, M. Schirra, M. S. Desole and P. A. Serra, Anal. Chem., 2010, 82, 5134.
30. S. N. Faisal, M. M. Hossain and H. J. Lee, J. Electrochem. Sci. Technol., 2010, 1, 121.
31. Z. Gazdik, O. Zitka, J. Petrlova, V. Adam, J. Zehnalek, A. Horna, V. Reznicek, M. Beklova and R. Kizek, Sensors, 2008, 8, 7097.
32. Y. T. Chen, F. A. Isherwood and L. W. Mapson, Biochem. J., 1953, 55, 821.
33. H. Beitollahi, A. Mohadesi, M. Mostafavi, H. Karimi-Maleh, M. Baghayeri and A. Akbari, Ionics, 2013, 20, 729.
34. S. S. Mitic, D. A. Kostic, D. C. Naskovic-Dokic and M. N. Mitic, Trop. J. Pharm. Res., 2011, 10, 105.
35. P. J. O’Connell, C. Gormally, M. Pravda and G. G. Guilbault, Anal. Chim. Acta, 2001, 431, 239.
36. Y. Andreu, S. de Marcos, J. R. Castillo and J. Galban, Talanta, 2005, 65, 1045.
37. G. Azimi, R. Dhiman, H. M. Kwon, A. T. Paxson and K. K. Varanasi, Nat. Mater., 2013, 12, 315.
38. E. Sharpe, T. Frasco, D. Andreescu and S. Andreescu, Analyst, 2013, 138, 249.
39. Y. W. Jun, J. S. Choi and J. Cheon, Angew. Chem., Int. Ed., 2006, 45, 3414.
40. T. S. Sreeremya, A. Krishnan, A. P. Mohamed, U. S. Hareesh and S. Ghosh, Chem. Eng. J., 2014, 255, 282.
41. S. Y. Lee and M. T. Harris, J. Colloid Interface Sci., 2006, 293, 401.
42. A. Krishnan, T. S. Sreeremya and S. Ghosh, CrystEngComm, 2015, 17, 7094.
43. L. F. Shen, P. E. Laibinis and T. A. Hatton, Langmuir, 1999, 15, 447.
44. P. Maitarad, D. Zhang, R. Gao, L. Shi, H. Li, L. Huang, T. Rungrotmongkol and J. Zhang, J. Phys. Chem. C, 2013, 117, 9999.
45. M. Mari, B. Mueller, K. Landfester and R. Munoz-Espi, ACS Appl. Mater. Interfaces, 2014, 7, 10727.
46. T. S. Sreeremya, K. M. Thulasi, A. Krishnan and S. Ghosh, Ind. Eng. Chem. Res., 2012, 51, 318.
47. M. Bloemen, W. Brullot, L. Tai Thien, N. Geukens, A. Gils and T. Verbiest, J. Nanopart. Res., 2012, 14, 1100.
48. W. Di, N. Shirahata, H. Zeng and Y. Sakka, Nanotechnology, 2010, 21, 365501.
49. X. H. Lu, X. Huang, S. L. Xie, D. Z. Zheng, Z. Q. Liu, C. L. Liang and Y. X. Tong, Langmuir, 2010, 26, 7569.
50. P. Patsalas, S. Logothetidis and C. Metaxa, Appl. Phys. Lett., 2002, 81, 466.
51. R. P. Tripathi, B. Singh, S. S. Bisht and J. Pandey, Curr. Org. Chem., 2009, 13, 99.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra07504e

---