1. Introduction

At the last decades, with the rising of the nanoworld, the technology has been guided to nanometric dimensions exploring the unique properties offered by nanoparticles, in special, semiconducting nanoparticles. In 2004 a pioneer work realized by Xu et al. [1] has disclosed a new kind of nanoparticles called carbon nanodots (C-dots). The most remarkable point of their finding is that, unlike other complex structures of carbon nanoparticles, C-dots are found as the simplest routes involving products graphitization or graphite fragmentation [2,3]. C-dots have been intensively investigated at the last decade due to their low cost and low toxicity, and their fascinating photoluminescent properties, high photostability, and photobleaching resistance [4–6]. They present an excellent potential for applications as gain media, in opto-electronic devices and in organic light emitting diode [7–11]. After a decade of investigation regarding the origin of these unique properties, a fundamental insight has not yet been achieved, mainly because of the remarkable role played by the surface passivated layer [12,13]. The presence of the surface states greatly modifies the optical response of such colloids, as expected for nanoparticles. Such states present either at the semiconductor volume and or at his surface can also be originated by the presence of vacancies, impurities, dopants, and, for the surface, also chemisorbed species [14]. Also the thermo-optical properties are of great importance for carbon nanoparticles due to their interaction with the host, by defining the suitability regarding local interactions if considered for markers, energy transducers, heat sources or optical limiters [15,16].

In this work, a method is developed to quantify the amount of C-dots prepared by laser ablation of graphite diluted in a host solution by absorbance spectra analysis. This method produces an ensemble of C-dots, which has singular reproducible characteristics. This nanoparticle possess a graphitic structure, as determined by HRTEM (high resolution transmission microscopy), a well-defined polymeric layer adsorbed at the C-dots surface and an assumed impurity/dopant free structure as processed from the pure graphite powder [3]. Further evaluations of their optical responses are added, namely its characteristic fluorescence and respective lifetime as well as thermo-optical parameters regarding the energy transfer to poly(ethylene glycol), PEG₂₀₀, solution.

2. Synthesis

In order to produce C-dots in PEG₂₀₀ by the laser ablation method, a Nd:YAG laser at 1064 nm was operated at 20 Hz repetition rate of 8 ns pulses, each with an approximate energy of 100 mJ. The unfocused 5 mm beam diameter of this laser irradiated over one hour solutions composed by 0.1 mg of graphite powder (in our case, purity > 99.8%, Micrograf HC11, Nacional de Grafite, Brazil) diluted in 1 ml of PEG₂₀₀ that were previously ultrasonically agitated during three hours. The original solution presented a gray coloration, and after this process it changed its color to yellow due to the optical comminution promoted by ablation of graphite powder forming carbon nanodots. The produced colloid has demonstrated a good stability and we could not observe any precipitation of C-dots even after 2 year of storage in normal conditions, which indicates that mostly all original particles of the solution have been turned into C-dots.

3. Molar absorption coefficient

In order to investigate the dependence of the linear absorption as a function of the C-dots concentration, the original solution was then carefully diluted in PEG₂₀₀ yielding more two samples of known concentration. The evaluation of the molar spectral absorption coefficient, ε(λ), of C-dots was performed in a Cary 5000 UV-Vis-NIR from Agilent with 1.0 mm quartz cuvettes.

C-dot colloidal absorption spectra in PEG₂₀₀

In Fig. 1(a) a decrease of the spectral absorption is observed related to the dilution. To analyze the dependence of the absorption with the concentration, the corresponding data was selected in a specific wavelength range around 300 nm and the results presented in the inset of Fig. 1(a), where a linear behavior on dilution is observed for the absorption spectra of the ablated graphite solutions. In fact, it serves to estimate the amount of the C-dots in solution by a simple analysis of the linear absorption in the visible region, similar to the method intensively applied to estimate the filling factor of the gold nanoparticles in colloidal suspensions [17]. This spectrum of ε(λ) is presented in Fig. 1(b) for particles with an average size of 3 nm as reported in [3] similar to the results also presented in [18].

 

Fig. 1 (a) Absorption spectra of three different concentrations: 0.1, 0.05 and 0.025 mg/mL. In the inset is plotted the dependence of the absorbance at 300 nm as a function of the filling factor. (b) Spectral molar absorption coefficient, ε(λ).

Download Full Size | PDF

4. Photoluminescence characterization

Excitation spectra and quantum yield

C-dots suspensions displayed a tunable broadband fluorescence as presented in Fig. 2(a), presenting a maximum intensity of fluorescence under an excitation wavelength at 390 nm. Their tunable characteristics are attributed to different factors: inhomogeneous particle size distribution, which yields different quantum confining states; different trapping sites and mechanisms or a characteristic of the C-dots functionalization. The quantum yield was calculated as reported previously [3], by using quinine sulfate as reference system, presenting a low quantum yield of 1.7%.

 

Fig. 2 (a) Colormap of fluorescence; the vertical and horizontal axis corresponds to the excitation and emission wavelength, respectively. (b) Fluorescence spectra under an excitation wavelength of 310 nm. (c) Fluorescence lifetime with fitting curve with corresponding residuals in inset. (d) IR spectrum.

Download Full Size | PDF

Fluorescence lifetime

To perform the lifetime measurements the colloid was excited with an EPLED operating at 310 nm (repetition rate of 5 MHz) and the emission was acquired at 480 nm, corresponding in Fig. 2(b) to the high absorption region and the emission spectrum maximum, respectively. The emission was collected by a time-correlated single-photon counting (TCSPC). For the decay curve, the data is well fitted by a bi-exponential behavior as presented in Fig. 2(c). The fluorescence decay lifetimes were 3.35 ns and 13.78 ns, which yields an average lifetime of 9.93 ns. In general, the multi-exponential decay of the fluorescence lifetime indicates that the fluorescence arises from a recombination of multi-excitons, but different process cannot be discarded. As observed in the IR absorption spectrum in Fig. 2(d), it reveals an additional peak in 1643 cm⁻¹, as compared to the pure host, which has been attributed to carboxylate groups on the C-dots surface and claimed as the fluorescence origin in this system [18].

5. Thermo-optical properties

Thermally managed eclipsed Z-scan

Measurements of nonlinear optical properties of materials, particularly the fast third-order nonlinear refraction and absorption, are widely performed by the Z-scan technique [19], where a substantial sensitivity enhancement is achieved in the eclipse configuration (EZ-scan) [20]. The use of high-repetition-rate pulsed lasers introduces, in general, a cumulative thermo-optical contribution to the measured signal, and a thermal management approach is usually performed to separate the fast (e.g., orientational or electronic) and slow (thermal) nonlinear mechanisms [21]. The nonlinear optical characterization of PEG₂₀₀ and C-dots suspensions were realized based on this thermally managed EZ-scan method.

Nonlinear optical measurements on C-dots are present in literature and most data are evaluated for near-resonant conditions (λ = 532nm) and for picosecond and nanosecond duration pulses [22–24] and open-aperture Z-scan measurements [25, 26], where the optical limiting feature is the explored effect. In the work of [27] nonlinear optical properties were investigated for C-dots in the 800 nm region and also here only the optical limiting property was addressed.

The experimental setup consists in a chopped laser beam of a mode-locked Ti:Sapphire laser oscillator (76 MHz, 100 fs, 800 nm), which is focused by a lens and the analyzed sample is scanned over a range of a few Rayleigh lengths along the focus region, where a beam waist radius of 35 µm was measured. After traversing the sample, a disc blocks the central portion of the beam in the far field plane, performing a spatial selection of the beam edge (blocking factor S_d = 98.3% [21]).

A time evolution trace of the transmitted intensity, for each sample position relative to the focal plane, was detected by a fast Si photodiode and recorded by a digital oscilloscope. The thermal management approach employed in our measurements was based on the mechanical modulation of the laser beam at 9 Hz, generating a time exposure (illumination) window of 2.66 ms (2.4% duty cycle) with an opening rise time of about 15 μs. A small duty cycle is a requirement for the sample to recover its original temperature between two consecutive exposure windows, and as a result, for the thermally-induced refractive index being negligible at the very beginning of a new exposure (t = 0).

Measurements performed for the pure solvent PEG₂₀₀ are showed in Fig. 3. Two time evolution traces, at a prefocal (−0.52z₀) and a postfocal position (0.52z₀), are presented in Fig. 3(a). An EZ-scan curve retrieved from all time evolution traces (different sample positions), at t = 2.35 ms, is showed in Fig. 3(b). Within the uncertainty due to the measurement noise, the presence and determination (quantification) of a fast (electronic) nonlinearity could not be performed when the temporal traces are extrapolated to t = 0.

 

Fig. 3 Thermally managed EZ-scan measurements of solvent (PEG₂₀₀) with intensity equal to 0.3 GW/cm², P_aver = 50 mW. (a) time evolution and (b) retrieved curve at t = 2.35 ms. Inset: open-aperture measurement at I₀ = 2.3 GW/cm².

Download Full Size | PDF

As observed for the case of C-Dots suspensions, Fig. 4, at relatively low power, the colloids present increased heating when compared with the pure solvent. The enhanced thermal effect by the presence of the inclusions makes even more difficult to determine a fast nonlinearity at the beginning of the exposure window. In fact, the thermal management approach is limited when the sample accumulates a significant amount of energy right from the start of the illumination. Increasing the pulse peak power, by reducing the pulse duration, for example, is not always enough to overcome the thermal effect caused by the elevated average laser power, due to the high repetition rate.

 

Fig. 4 Thermally managed EZ-scan measurements: (a) Time evolution at prefocal (descending curves) and postfocal positions (ascending curves) of CQD-A (red square), CQD-B (blue square) and CQD-C (black square) samples, under the same laser intensity of 64 MW/cm², P_aver = 9.5 mW. (b) Retrieved curve for CQD-A at t = 2.35 ms. Inset: open-aperture measurement at I₀ = 350 MW/cm².

Download Full Size | PDF

Insets of Fig. 3 and Fig. 4 show the results of the measurements for the solvent used, PEG₂₀₀, and the CQD-A, respectively, in the case of unblocked beam (open-aperture configuration). Due to the insensibility to thermal effects, unblocked beam data for higher power are displayed, and even at these elevated intensities at the focal plane, these samples do not present nonlinear absorption. Other two samples, CQD-B and CQD-C, manifest the same behavior concerning nonlinear absorption. These results are in agreement with the data presented by [27] where nonlinear optical limiting was observed for fluences of orders of magnitude higher than the one used here.

Evaluation of thermo-optical properties

In order to investigate only thermo-optical properties, eliminating any possible fast contribution, a subtraction procedure was employed, which consists in the subtraction of two normalized EZ-scan curves retrieved from the measured temporal evolution traces at two distinct times.

To perform the modeling of the measured EZ-scan normalized transmittance signals, the Huygens-Fresnel propagation integral was employed to obtain the field distribution at the disk, after propagating from the exit plane of the sample. The exit field, in turn, is obtained in the thin sample approximation [28] as the product of the incident excitation field and a phase factor proportional to the temperature profile induced by the laser beam. The model for the temperature distribution within the medium [28, 29] is dependent on the thermal properties of the material, as thermal conductivity and specific heat. Usually, a fraction of the laser beam energy absorbed by the sample is converted into heat, i.e., a thermal conversion efficiency (η) less than unity. This in general implies on a difference between the absorption coefficient measured by UV-VIS (α) and the value obtained with the thermo-optical (thermal lens) technique.

The temporal evolution of the EZ-scan signals displayed for PEG₂₀₀ alone, as well as for all samples containing C-dots, present the same thermal characteristic time t_c = w²/4D, where w is the laser beam radius and D the material’s thermal diffusion coefficient. Therefore, fitting data to the model allows the adjustment of two parameters according to the concentration variation observed: the efficiency times the C-dots absorption coefficient, ηα, and the temperature dependent refractive index, dn/dT. Optimal values for these parameters are listed in the Table 1. According to the behavior observed for these two parameters, both present a linear dependence on concentration within this concentration range. Linear coefficients were obtained from the linear adjust of each parameter, (ηα)_COL - (ηα)_PEG = µ·C and (dn/dT)_COL - (dn/dT)_PEG = γ ·C, where COL an PEG refer to colloid and PEG₂₀₀ parameters, C is the C-dots concentration (mg/ml) and µ = 0.344 mL·mg⁻¹·cm⁻¹ and γ = 2.33 x 10⁻³ mL·mg⁻¹·K⁻¹ are the fitted coefficients.

Table 1. Thermo-optical parameters adjusted for CQD samples and PEG₂₀₀ @ 800 nm.

View Table

Thermal lensing dependence on concentration has also been observed before for colloidal solutions of metallic nanoparticles [31, 32] with an increasing value on nanoparticle concentration. In these studies, an enhancement of the absolute value of thermo-optical coefficient, dn/dT, was assigned to the colloids as the concentration is increased. Differently, in the case of our CQD samples, measurements and modeling point to a decrease in the absolute value of this thermo-optical property as the concentration is increased. As can be seen in Table 1, the most concentrated sample presents a variation about 50% of the dn/dT value relative to the pure solvent, PEG₂₀₀.

In order to perform a simple verification of this behavior, we employed an Abbe refractometer to measure the refractive index of the pure solvent and the most concentrated C-dots sample at two different temperatures. Measurements demonstrated the same qualitative tendency, i.e., the absolute value of the thermo-optical coefficient is decreased for the C-dots sample when compared to the pure PEG₂₀₀, although this measurement showed a variation about 25%. These observations demonstrate the need for further investigation of the physical mechanism underlying this peculiar behavior, namely, the reduction of the absolute magnitude of dn/dT when C-dots are present. This behavior may be unnoticed at previous work on silver colloids [32], where |dn/dT| is reduced for determinate nanoparticle concentrations when compared to the value of the solvent used (distilled water).

6. Conclusion

The presented work aimed to establish the optical and thermo-optical characteristics of carbon nanodots produced by photoablation in PEG₂₀₀. C-dots produced in this way have the advantage of synthesis control with good reproducibility and therefore chosen as a cornerstone to investigate C-dots optical properties. The evaluation of the spectral absorption coefficient is here associated to photoluminescence excitation spectra, evaluated also by the quantum yield, as well as the lifetime of this emission. Furthermore, the thermally managed EZ-scan method allowed evaluation of the main characteristics of the thermal-lens presented by this colloidal solution, which relies on the concentration dependence of thermo-optical coefficient, dn/dT. The analysis of such behavior and all the related characterization is hence determinant to provide a coherent perspective of the local thermal interactions of nanoparticles with their host.