Effect of Surface Chemistry on the Fluorescence of Detonation

Oct 31, 2017 - KEYWORDS: detonation nanodiamonds, fluorescence, functionalization, nitrogen-vacancy center, diamond, sp2 carbon, carbon dots...
1 downloads 0 Views 5MB Size
www.acsnano.org

Effect of Surface Chemistry on the Fluorescence of Detonation Nanodiamonds Philipp Reineck,*,† Desmond W. M. Lau,† Emma R. Wilson,† Kate Fox,‡ Matthew R. Field,§ Cholaphan Deeleepojananan,∥ Vadym N. Mochalin,*,∥,⊥ and Brant C. Gibson†

Downloaded via KAOHSIUNG MEDICAL UNIV on July 1, 2018 at 13:32:08 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



ARC Centre of Excellence for Nanoscale BioPhotonics & School of Science, ‡School of Engineering, and §RMIT Microscopy and Microanalysis Facility (RMMF), RMIT University, Melbourne, VIC 3001, Australia ∥ Department of Chemistry and ⊥Department of Materials Science & Engineering, Missouri University of Science & Technology, Rolla, Missouri 65409, United States S Supporting Information *

ABSTRACT: Detonation nanodiamonds (DNDs) have unique physical and chemical properties that make them invaluable in many applications. However, DNDs are generally assumed to show weak fluorescence, if any, unless chemically modified with organic molecules. We demonstrate that detonation nanodiamonds exhibit significant and excitation-wavelength-dependent fluorescence from the visible to the near-infrared spectral region above 800 nm, even without the engraftment of organic molecules to their surfaces. We show that this fluorescence depends on the surface functionality of the DND particles. The investigated functionalized DNDs, produced from the same purified DND as well as the as-received polyfunctional starting material, are hydrogen, hydroxyl, carboxyl, ethylenediamine, and octadecylamine-terminated. All DNDs are investigated in solution and on a silicon wafer substrate and compared to fluorescent high-pressure high-temperature nanodiamonds. The brightest fluorescence is observed from octadecylaminefunctionalized particles and is more than 100 times brighter than the least fluorescent particles, carboxylated DNDs. The majority of photons emitted by all particle types likely originates from non-diamond carbon. However, we locally find bright and photostable fluorescence from nitrogen-vacancy centers in diamond in hydrogenated, hydroxylated, and carboxylated detonation nanodiamonds. Our results contribute to understanding the effects of surface chemistry on the fluorescence of DNDs and enable the exploration of the fluorescent properties of DNDs for applications in theranostics as nontoxic fluorescent labels, sensors, nanoscale tracers, and many others where chemically stable and brightly fluorescent nanoparticles with tailorable surface chemistry are needed. KEYWORDS: detonation nanodiamonds, fluorescence, functionalization, nitrogen-vacancy center, diamond, sp2 carbon, carbon dots organic molecules10,18 or fragments of sp2 carbon, and only weak NV emission was observed in a small fraction of purified and deaggregated DNDs.8 However, studies showed that NV fluorescence in DND can be strongly enhanced via sintering,19 and NV fluorescence was studied in irradiated and annealed material11,20 and can even be found in mostly unprocessed DNDs.9,14 However, the proximity of NV centers to the diamond surface has been shown to strongly inhibit NV emission in diamond particles of less than 10 nm.21 In general, carbon can give rise to fluorescence through three mechanisms: (1) Delocalized π electrons in aromatic hydrocarbons. This is the cause for strong absorption and re-emission

D

etonation nanodiamond (DND) is a unique material: it consists of diamond particles of about 5 nm in size, can be produced in commercial quantities, and offers many exceptional mechanical, chemical, electronic, and optical properties of bulk diamond at the nanoscale.1 DND has a wide range of applications from tribology2 to polymer- and metalmatrix composites3,4 to theranostics.5,6 While many physical and chemical properties of detonation nanodiamond have been studied extensively,1,7 its fluorescence is still poorly understood although it has been investigated in several publications.8−17 This is in stark contrast to generally larger (>10 nm) milled high-pressure high-temperature (HPHT) nanodiamonds, which can be processed to contain high concentrations (>10 ppm) of the most widely known fluorescent defect in diamondthe nitrogen-vacancy (NV) color center. In general, DNDs are believed to be mostly nonfluorescent unless modified with © 2017 American Chemical Society

Received: July 3, 2017 Accepted: October 31, 2017 Published: October 31, 2017 10924

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

Cite This: ACS Nano 2017, 11, 10924-10934

Article

ACS Nano

Figure 1. Overview of the as-received DND and functionalized detonation nanodiamond materials produced from DND-COOH. “Diameter” refers to the diameter of the particles in dispersion as determined via dynamic light scattering.

Figure 2. (A) Particle size distribution of DND particles dispersed in water (all except DND-ODA) and chloroform (DND-ODA) determined by dynamic light scattering. (B) FTIR spectra of the DND powders recorded in KBr pellets. (C) Raman spectra of particles on a silicon substrate obtained using 532 nm laser excitation.

of photons in organic fluorophores, which remain among the brightest fluorescent materials known.22 It has been shown that the stacking of simple aromatic molecules like anthracene or pyrene into nanoparticulate structures also reproduces many of the optical properties of so-called carbon dots,23 which generally consist of graphitic (sp2) or amorphous carbon. Conjugated carbon systems in isolated sp2-hybridized islands have also been used to explain the fluorescence properties of single-layered graphene oxide.24 (2) Defects in sp2 carbon. Graphene is a semimetal without an optical band gap. When treated with an oxygen plasma it becomes fluorescent25,26 and shows emission properties similar to those observed for graphene oxide produced via wet chemical methods.27 Oxygen, nitrogen, and other surface defects in sp2 carbon have been suggested to be the main cause for the bright and excitationwavelength-dependent fluorescence of carbon dots.28 Surface defects in sp2 carbon-based nanoparticles are routinely used to “explain” the fluorescence properties of carbon dots. However, the underlying photophysics are not yet understood. (3) Optically active crystal defects in diamond. A large number of such defects have been identified.29 Most of these have been investigated in far less detail compared to the NV center, mainly because their creation is challenging and large-scale fabrication not feasible in many cases.29

Here, we investigate the fluorescence properties of hydrogen-, hydroxyl-, carboxyl-, ethylenediamine-, and octadecylamine-functionalized DNDs as well as the as-received polyfunctional starting material. The moieties attached to the DND surface as well as hydrogen are known to be nonfluorescent on their own and are therefore well-suited to study the surface chemistry effects on the intrinsic DND fluorescence. We demonstrate that the fluorescence properties strongly depend on the surface chemistry of the DND particles. All particle types are investigated in solution and in a dry state on a silicon wafer substrate using fluorescence spectroscopy and confocal microscopy. Results are compared to ca. 100 nm sized HPHT nanodiamonds containing high concentrations of nitrogenvacancy centers. Most DND particle types show strongly excitation wavelength dependent fluorescence from the visible to the near-infrared (NIR) spectral region. Our results suggest that most of the observed fluorescence originates from residual non-diamond carbon. However, we locally also find significant fluorescence from NV centers in hydrogenated, hydroxylated, and carboxylated DNDs. Our results highlight the need for a better understanding of the photophysics of fluorescence originating from non-diamond carbon and the effects of nondiamond carbon and impurities therein on defect-based diamond core fluorescence. 10925

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano

RESULTS AND DISCUSSION Materials Characterization. Commercially available DND particles were used as received (UD90 grade, NanoBlox Inc., USA) and functionalized according to reported protocols to obtain hydrogenated (DND-H),30 hydroxylated (DND-OH),31 carboxylated (DND-COOH),32 as well as ethylenediamine (DND-EDA)33 and octadecylamine (DND-ODA)18 functionalized particles. The as-received particles are referred to as DND throughout the manuscript. The particles were characterized using Fourier transform infrared spectroscopy (FTIR), UV−vis absorption spectroscopy, dynamic light scattering (DLS) and zeta potential measurements, and Raman spectroscopy. A summary of their basic properties is shown in Figure 1. All particles were dispersed in water except DND-ODA, which is strongly hydrophobic and was dispersed in chloroform. The primary particle size of DNDs is 4−6 nm. In dispersion, all particles were present in aggregates of 40 nm or larger in size and showed polydisperse size distributions (Figure 2A). All particles except DND-EDA showed good colloidal stability after the removal of larger aggregates via centrifugation (1000 rcf for 5 min, see the Methods for details). DND-EDA particles were not stable in solution and sedimented within less than 1 h, most likely due to the reduction of the initial highly negative zeta potential of DNDCOOH during the wet chemical attachment of ethylene diamine (EDA) to produce DND-EDA.33 The surfaces of all other particles dispersed in water carry strong positive (DNDH) or negative charges (DND-OH, DND-COOH, DND), while DND-EDA is not as strongly charged, which is consistent with the particles’ colloidal instability. DND-ODA is only weakly charged and sterically stabilized in chloroform. Despite the visibly different colors of the DND powders, the extinction spectra of all particles are dominated only by scattering from small particles. See Figure S1 for extinction spectra of all samples. It is important to note that in this study DND-COOH was a precursor for all other chemically modified nanodiamonds. As-received DNDs have graphitic (sp2 carbon) or amorphous carbon shells with highly polyfunctional surfaces exhibiting carboxylic acid, ester, ether, and ketone groups1,7,32 among others. Many chemical and physical processes are known to remove non-diamond carbon including strong acid treatments, annealing in air,1,7,32 and ozone treatments.34 However, even with highly efficient processes, up to several weight percent of non-diamond carbon generally remains in detonation nanodiamond samples.32 In this study, the sp2 non-diamond and sp3 diamond content of all samples was estimated on the basis of electron energy loss spectroscopy (EELS). The results are summarized in Table 1. Details on the derivation of the shown

values from EELS spectra as well as TEM images can be found in Figures S2−S6. DND-COOH has the highest diamond content of about 96 wt %, while all other samples contain significantly more sp2-bonded carbon (between 19 and 30 wt %). ODA and EDA both contain sp3-hybridized non-diamond carbon in the form of hydrocarbon chains. However, the EELS signal of the hydrocarbon chains could not be separated from the sp2 carbon peak in our samples (see the SI for details). Therefore, we note that the EELS-derived values for sp2bonded carbon quoted for DND-ODA and DND-EDA throughout the manuscript include contributions from sp3 hydrocarbon chains. The presence of the various surface functionalities was verified using Fourier transform infrared spectroscopy (FTIR) as shown in Figure 2B. DND-ODA exhibits peaks at 1650 cm−1 and between 2800 and 3000 cm−1. These correspond to the presence of amides and C−H stretch vibrations, respectively. The presence of amides is direct evidence of the covalent attachment of ODA to the DND particle, while the C−H peaks are characteristic of long hydrocarbon chains in DND-ODA. The same amide peak is present in the DND-EDA spectrum, again indicating the formation of a covalent bond between EDA and the DND surface. However, the C−H peak is much less pronounced since the EDA molecule only features a two carbon atom long aliphatic chain between the two end amino groups. A peak at 1630 cm cm−1 is also characteristic of O−H bending, which is clearly present in DND-COOH but most pronounced in DND-OH. DND-OH, which is produced by reduction of DND-COOH, also has a greatly reduced signal in the region of the CO stretch, evidence of CO conversion into C-OH. The presence of a clear peak around 1750 cm−1 indicates the presence of CO bonds in DND-ODA, DND-EDA, and DND-COOH. Its absence in the DND-H spectrum and the appearance of the C−H peaks between 2800 and 3000 cm−1 show the presence of CH2 and CH3 bonds, evidence of successful hydrogenation of the diamond surface. Peaks indicative of OH groups (bending at 1650 cm−1 and stretching at 3300 cm−1) in the DND-H sample are most likely due to water adsorbed on the surface and covalently bonded OH groups that could not be converted into C−H under these conditions.30 We also investigated the samples using Raman spectroscopy. All particles except DND-ODA show a clear sp3 diamond peak around 1330 cm−1 as well as a broader peak around 1620 cm−1. All particle spectra exhibit a more or less pronounced upward slope toward longer wavenumbers, which is caused by fluorescence. A peak close to 1620 cm−1 is often attributed to the G-band of graphite-like sp2 carbon. However, in the presence of surface functional groups, the analysis of the sp2 contribution to the overall Raman signal is challenging since various oxygen-containing functional groups (e.g., O−H) can significantly contribute in this spectral region.35 No diamond sp3 carbon peak is present in the Raman spectrum of the DNDODA sample. However, characteristic diamond lattice features can be clearly seen in high-resolution TEM images (see Figure S6 for TEM images of all samples), and 74% sp3 diamond carbon content was estimated from EELS experiments in DND-ODA (see Table 1). In-Solution Spectroscopy. Normalized fluorescence spectra of all particles in solution are shown in Figure 3 for excitation wavelengths from 400 to 700 nm. Spectra for longer excitation wavelengths are omitted in the cases of DND-EDA, DND-OH, and DND-COOH due to very low fluorescence

Table 1. Estimates of the Fraction of sp3 Diamond and Nondiamond Carbon Present in the Different Samples Based on EELS sample

sp3 diamond (wt %)

sp2 and non-diamond sp3 carbon (wt %)

DND DND-ODA DND-EDA DND-H DND-OH DND-COOH

81 74 83 70 76 94

19 26 17 30 24 6 10926

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano

Figure 3. Normalized fluorescence spectra of all particles in solution. Particles were excited at different wavelengths between 400 and 700 nm as indicated in each graph. Raman spectra of the solvent have been subtracted from all spectra (see main text and SI for more details).

Figure 4. (A) Fluorescence peak position λem as a function of excitation wavelength λex. (B) Relative fluorescence brightness of DND and HPHT particles in solution for the different excitation wavelengths λex. For the brightness comparison, all spectra were corrected for differences in NP concentration and excitation intensities at different λex. (C) Time-resolved fluorescence decay traces of DND-ODA, HPHT ND, and the instrument response function (IRF) with and without a long-pass filter. Dashed lines represent single exponential fits to the data and the decay time constants are given in the graph.

shows λem as a function of λex for HPHT nanodiamonds with a high concentration of NV centers. Here, λem also shifts to shorter wavelengths with decreasing λex. This is caused by a more pronounced conversion of the NV− charge-state to the neutral NV0 charge-state with increasing photon energy.9 See Figure S11 for HPHT ND fluorescence spectra. All DND particles are most efficiently excited at λex = 450 nm, except DND-EDA, which shows a continued increase in fluorescence brightness with decreasing λex down to 400 nm (the shortest λex in our experiments, Figure 4B). The fluorescence brightness strongly decreases for all particles toward longer excitation wavelengths. HPHT ND particles are most efficiently excited at 550 nm, which is close to the absorption maximum of the NV center in diamond.36 The fluorescence brightness of the particles varies by more than 2 orders of magnitude between the brightest DND-ODA and the very weakly fluorescing DND-COOH. DND-OH is about two times brighter than DND-COOH, while DND, DND-H, and DND-EDA are all of the same brightness upon 450 nm excitation and more than 10 times brighter than DND-COOH.

signals. For all samples except DND-ODA, the Raman signal of the solvent was on the same order of magnitude and, in the case of DND-OH and DND-COOH, significantly stronger than the fluorescence of the particles. Therefore, the Raman spectrum of the solvent was subtracted from all spectra, which particularly in the case of DND-OH and DND-COOH causes slight distortions of the resulting curves. See Figure S9 for the raw spectra and details on the data acquisition and analysis. For DND, DND-ODA, and DND-EDA, the spectral position of the emission peak (λem) increases approximately linearly with the excitation wavelength (λex) between 400 and 600 nm (Figure 4A). At longer excitation wavelengths, the emission peak position cannot be clearly identified in any sample due to a decrease in the Stokes shift with increasing λex. DND-H, on the other hand, shows only a very weak increase in λem with increasing excitation wavelength. Due to a very weak overall fluorescence signal, a clear excitation wavelength dependence cannot be reliably inferred from the spectra of DND-OH and DND-COOH. However, both types of nanodiamond also show a trend toward longer λem with increasing λex. Figure 4A also 10927

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano

Figure 5. Confocal fluorescence images of all particle types drop-cast onto a silicon wafer. All images were acquired under identical imaging conditions (λex = 500 nm, 100 μW, fluorescence collected above 540 nm). Note that the intensity scales differ between images. Only 10% of photons were used for imaging and 90% for spectroscopy. Hence, “counts per second” (CPS) values in this figure represent CPS values detected by the avalanche photodiode multiplied by 10. Circles in panels E and F indicate locations where NV fluorescence was observed.

Time-resolved fluorescence decay traces of all particles in solution were acquired, and the decay curves of DND-ODA and HPHT ND are shown in Figure 4C together with two instrument response functions (IRF). Due to the strong scattering of all particles in solution and weak fluorescence intensity of some samples, spectral filtering was required. The introduction of a long-pass filter leads to a significant broadening of the IRF. All nanodiamond types except DNDODA showed a relatively fast dominant fluorescence lifetime of less than 1 ns that could not be deconvoluted from the IRF. See Figure S12 for decay traces of all particles. The longer components of the decay for DND-ODA and HPHT ND were estimated using a single exponential fit as indicated by the dashed lines in Figure 4C to be 5 and 13 ns. The number of publications investigating the fluorescence of DND particles in solution is very limited to date.17,18,37−39 It must be stressed that the properties of detonation nanodiamonds strongly vary between different providers of the “raw” material12 and that small variations in the processing protocols used for purification and functionalization can have significant effects on the particles’ physical and chemical properties and their dispersion behavior. Hence, one must be cautious when comparing the properties of materials produced from different raw materials and following different protocols. To avoid this, we have used one batch of DND to produce DND-COOH, and all other chemically modified DNDs were produced from this DND-COOH. The synthesis and bright fluorescence of DND-ODA was first reported by Mochalin et al.18 However, only blue fluorescence has been demonstrated, and neither the strong shift of λem as a function λex observed here nor the fluorescence lifetime is reported. Vervald et al. show fluorescence spectra of DND-COOH using 405 nm excitation, which are slightly blueshifted compared to the spectrum shown in Figure 3F.17 In another recent study, polyfunctional and fluorescent particles have been produced from DND particles in water via laser

ablation.37 An excitation wavelength dependence of the emission spectra was also observed (only for λex = 360−500 nm) and rationalized with different “surface states” associated with the various functional groups present on the particle surfaces. Dolenko and co-workers have investigated the fluorescence of fluorinated, hydroxylated, and carboxylated DND particles in water (λex= 488 nm) and report weak fluorescence (relative to the water Raman signal) for all functionalizations.16 Two other reports mostly focus on using fluorescence from DND NPs in solution-based applications.38,39 To the best of our knowledge, the fluorescence of DND, DND-EDA, DND-H, and DND-OH synthesized from the same purified and well-characterized nanodiamond have not been investigated in solution to date. Confocal Fluorescence Imaging and Spectroscopy. The optical properties of all samples were also analyzed on a silicon wafer substrate using a custom-built scanning confocal fluorescence microscope (see the Methods for details). A 40 μL portion of nanoparticle (NP) solution was drop-cast onto a silicon wafer at a temperature of 100 °C on a hot plate and allowed to dry. All imaging and spectroscopy were carried out using 500 nm pulsed laser excitation at 5 MHz repetition rate, and fluorescence was collected at wavelengths >540 nm. Confocal fluorescence images of all samples are shown in Figure 5 (note the different intensity scales in the images). Due to the surface properties of the differently functionalized particles, the morphology and thickness of the resulting dry particle layers are significantly different, making a quantitative assessment of the fluorescence brightness challenging. However, it is worth noting that the average fluorescence intensity in the confocal images shown in Figure 5 is in good qualitative agreement with the in-solution results for the fluorescence brightness: DND-ODA exhibits the strongest fluorescence, almost 2 orders of magnitude stronger than DND-COOH. DND−OH is about twice as bright as DND-COOH, while all other particles are between these extremes. See Figure S14 for 10928

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano

Figure 6. Fluorescence spectra (A, C) and time-resolved fluorescence decay traces (B, D) of different DNDs drop-cast onto a silicon wafer. A typical NV fluorescence spectrum of HPHT NDs is shown in panel A and C for comparison. (A, B) Average fluorescence spectra and decay traces observed for most particles on the substrate. (C, D) Fluorescence spectra and decay traces of individual bright spots that show typical characteristics of NV fluorescence. Note the different x-axis scaling in panel B and D. λex = 500 nm at 100 μW excitation intensity in all cases.

Table 2. Average Fluorescence Lifetimes and Relative Amplitudes for All Samples Drop-Cast on a Silicon Substrate DND τ1 (ns) amplitude 1 (%) τ2 (ns) amplitude 2 (%)

1 μm are found in agreement with the large agglomerate sizes obtained from DLS measurements. Both DND-ODA and DND-H particles also strongly cluster during the drying process and form highly fluorescent regions. The confocal fluorescence images in Figure 5 were analyzed in two ways: (1) by analyzing individual locations of interest and (2) via averages obtained while scanning part of an image. Fluorescence spectra and time-resolved fluorescence decay traces obtained in both ways are shown in Figure 6. The individual locations investigated were bright spots that can easily be identified in Figure 5E,F (examples are highlighted by circles). In all other images, these spots are difficult to identify. In the DND sample they are very rarely found, while in Figure 5B−D the large number of brightly fluorescing spots and the high overall fluorescence makes their identification difficult.

To obtain results averaged over many particles, regions of homogeneous fluorescence were imaged (10 × 10 μm, λex = 500 nm) with spectra and fluorescence decays recorded while scanning. These spectra and decay traces are shown in parts A and B, respectively, of Figure 6. The spectra of all particle types exhibit a similar full-width-half-maximum of about 150 nm. The Stokes shift varies and is smallest for DND-ODA and DNDEDA and largest for DND-COOH and DND particles. All emission spectra are blue-shifted with respect to a typical NV emission spectrum from HPHT NDs. A direct comparison of the fluorescence spectra obtained in dispersion vs on a silicon substrate is shown in Figure S15. All particle types show a multiexponential fluorescence decay (Figure 6B). This implies the presence of more than one fluorescence decay pathway for optically excited states. Using deconvolution with the instrument response function (IRF), we find that all traces are well approximated by biexponential decays. The fluorescence lifetimes obtained from this fitting process are shown in Table 2. DND showed a decay component well below 1 ns, which was too short to resolve. All other decay traces are dominated by a short lifetime component of between 0.8 ns (DND-EDA) and 1.4 ns (DNDODA). DND-ODA has a significant contribution from a longer 10929

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano lived state with τ2 = 5.8 ns, which is close to the value obtained for the particles in solution. All other particles (except DND) show comparable τ2 values between 3.9 and 6.5 ns, but with a significantly lower amplitude. The HPHT ND particles also show a fast (9.1 ns) and slow (44 ns) lifetime component, which contribute almost equally to the overall decay with amplitudes of 44% and 56%, respectively. In each image, 20 individual brightly fluorescing spots were randomly chosen and their spectra and fluorescence decay recorded. The results for DND, DND-ODA, and DND-EDA were identical to the average results shown in Figure 6 A,B. Importantly, in the cases of DND-H, DND-OH, and DNDCOOH, a subset of these particles showed clear signatures of NV fluorescence. Representative spectra and fluorescence decay traces are shown in parts C and D, respectively, of Figure 6, together with a typical NV spectrum and fluorescence decay collected from HPHT NDs. For DND-COOH, 17 out of 20 investigated spots showed clear NV0 zero-phonon line fluorescence at 575 nm and long fluorescence lifetimes >10 ns. For DND-H only 1 and DND-OH 3 out of 20 showed these characteristics. Broad fluorescence spectra from DND particles on glass or Si substrates upon excitation with green or blue light, similar to the ones in Figure 6A, have been reported.8,9,11,13,14,20,40 Several studies use proton- or electron-irradiated and vacuum-annealed DND particles to increase the number of NV defects in the material after chemical or physical purification.11,20,40 Other studies investigate as-received9 or chemically purified8,14 material without prior irradiation. Smith and colleagues11 attribute a fast fluorescence decay, similar to the one seen in Figure 6B for DND, to graphitic carbon. A longer lifetime component >10 ns was attributed to NV fluorescence in timegated spectroscopy experiments. The spectra show a featureless emission from sp2 carbon in the first few nanoseconds after excitation. Nitrogen-vacancy zero-phonon lines appear in the spectrum when recorded with a 20 ns delay. However, Vlasov et al. suggested that the clear NV fluorescence originates from larger diamond crystals in DND of >30 nm and not from individual or clustered 3−6 nm sized DND particles.40 Kirmani et al. further illuminate the chemical structure of non-diamond carbon around the diamond core and its role in the quenching of core-originating fluorescence.14 Clear NV signatures like the ones shown in Figure 6C,D have also been reported in DND particles. However, most of these studies used irradiated and annealed DND material.11,20,40 Our group has previously reported NV fluorescence of comparable brightness from as-received DND material (from a different provider than the material used in the present study). The photostability of the DND particles was investigated by positioning the beam in an area of average fluorescence brightness and recording the collected fluorescence over time. The resulting photobleaching traces, normalized at time t = 0, are shown in Figure 7. In general, all DND particle types show photobleaching behavior. However, the individual spots showing NV fluorescence characteristics (Figure 6C,D) were found to be highly photostable, in agreement with our previous report9 (not shown here). The fluorescence of DND-ODA decreases by more than 75% within less than 30 s. DND, DNDEDA, DND-H, and DND-OH show better photostability characteristics, while DND-COOH shows the highest photostability with a reduction in fluorescence of about 20% during the first 5 min of illumination. All photostability traces were well fitted by a single (DND-COOH) or biexponential decay

Figure 7. Photostability of DND particles: fluorescence intensity as a function of time. All traces were normalized to the intensity at t = 0 s to show the different bleaching dynamics.

(all others). The characteristic photobleaching times for all DND particles are summarized in Table 3. It is worth noting Table 3. Bleaching Time Constants Obtained from an Exponential (COOH) or Biexponential (for All Other Samples) Fit to the Data Shown in Figure 7 DND-COOH DND-OH DND-H DND-EDA DND DND-ODA

τB1 (s)

τB2 (s)

121 15 13 10 8.3 3.5

144 142 104 117 102

that the photostability is approximately anticorrelated with the fluorescence brightness shown in Figure 4; i.e., the brighter a particle is, the less photostable it is, with the exception of HPHT ND, which exhibits perfect photostability (not shown here). All traces were background-corrected. Hence, background effects can be excluded as an explanation for the observed differences in photobleaching behavior. No changes in fluorescence spectra obtained before and after photobleaching were found for any sample. The excitation intensity of 100 μW used here (λex = 500 nm) is at least 1 order of magnitude lower than the excitation intensities used in Raman experiments performed on the same samples at a similar wavelength (532 nm). Raman spectra were found to remain unchanged upon repeated acquisition. The illuminated areas were also visually inspected for any laser light induced changes before and after the Raman experiments, and no changes were found. This suggests that the light intensities used in the photostability experiments are also unlikely to induce significant changes to the DND particles such as graphitization. However, it is likely that the employed light intensity of about 1.4 × 105 W cm−2 in the focal spot is high enough to oxidize functional surface groups and molecules of illuminated particles, which may cause the observed photobleaching. Origins of Fluorescence. We discuss two distinct types of fluorescence that were found to originate from our detonation nanodiamond samples: (A) carbon-dot-like fluorescence from graphitic or amorphous carbon and (B) fluorescence from NV centers in diamond. 10930

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano 1. Carbon-Dot-like Fluorescence. The particles investigated here are clusters >40 nm of individual 4−6 nm sized DND particles (see Figure 8). The amount of non-diamond carbon

frequently encountered in the DND-COOH sample. However, the main reason for this could be the low background of the majority of the DND-COOH, which only weakly fluoresces and facilitates the identification. Therefore, occurrence alone is insufficient to draw conclusions regarding the effect of the particle’s surface functionalization on the fluorescence of the embedded NV defects. In the DND-COOH sample we find significant variations in the optical properties of these spots. Their brightness varies by 1 order of magnitude, and we find spectra that are strongly dominated by NV− emission as well as the ones dominated by NV0 emission (see Figure S16) under identical imaging conditions. The slow decay fluorescence components of the NV spots show slightly less variation between 24 and 34 ns, but the relative contribution of fast and slow decay also differs by 1 order of magnitude (see Figure S16). These large variations within a single sample necessitate the analysis of a significant number of particles to identify systematic differences between different particle types. Identifying the effect of the surface groups on NV emission and understanding why only a small subset of particle aggregates shows significant NV fluorescence will be the focus of future investigations. The starting material for all particles studied in this work is the same. None of the processing steps can lead to a removal or creation of NV centers from the diamond particles unless the structure of the diamond crystal lattice itself is compromised. Since we know that all DND particles consist of ≥70% diamond, NV centers are highly likely to be present in all particle types. However, although many of them can be located in solid state experiments, most of these centers are optically inactive or cannot be identified amidst strong carbon-dot-like fluorescence.

Figure 8. Generalized schematic illustration of the structure of the DND particle aggregates, the location of surface modifications, and color centers as well as the different sources of fluorescence.

present on the surface or around these clusters strongly depends on the processing and functionalization of the particles. We find the particles with the highest sp3 diamond content (DND-COOH) to show the weakest overall fluorescence both in solution and on a silicon substrate. The non-diamond carbon content in all other particles varies maximally by about 11%. However, their fluorescence properties vary dramatically, which emphasizes the importance of the surface functionalization. DND-ODA, -EDA, -OH, and -H are all created from DND-COOH, i.e., from a very weakly fluorescing material consisting of about 96% diamond. During the various functionalization processes the chemical structure of only the outermost carbon layers is changed and results in an increase of non-diamond carbon. This graphitic (sp2) and amorphous carbon causes the relatively fast decay (fluorescence lifetime) and quick photobleaching fluorescence observed from DND-ODA and DND-EDA. This fluorescence and its strong λex dependence are similar to the as-received polyfunctional and sp2-containing DND material and typical for carbon dots.28 Octadecylamine41 and ethylenediamine42,43 have both been used in the literature previously for the synthesis of carbon dots. Nitrogen impurities have been identified as one of the causes of carbon dot fluorescence.44 In this sense, DND-H stands out as it is also dominated by short-lived fluorescence decay and of comparable brightness to DND-EDA but lacks the λex dependence and shows a significant contribution from a longer-lived emission above 10 ns (see Figure 6B). In addition, the surface chemistry of DND-H is much simpler than that of DND-ODA/EDA and does not involve nitrogen-rich molecules. DND-OH shows a slightly stronger λex dependence but also significantly weaker overall fluorescence with a short fluorescence lifetime, which most likely also originates form non-diamond carbon. The significant NIR fluorescence of DND, DND-ODA, and DND-H in a spectral region above 800 nm is noteworthy (Figure 2A,B,D), particularly since only a handful of carbon dots fluorescing in the NIR spectral region have been reported to date.44 However, we also find the fluorescence of DND, DND-ODA, and DND-H to be most pronounced in the visible light range. 2. Nitrogen-Vacancy Fluorescence. NV fluorescence from DND-H, DND-OH, and DND-COOH particles deposited on a substrate can clearly be identified in individual locations. These individual, generally diffraction limited spots are most

CONCLUSION We have systematically characterized the fluorescence of 5 differently functionalized detonation nanodiamond particles produced from the same carefully purified and characterized starting material as well as the as-received material in solution and on a silicon wafer substrate. We find a strong excitation wavelength dependence and significant fluorescence ranging from the visible to the NIR spectral region from DND, DNDODA, and DND-EDA particles in solution. Of the simple surface terminations (H, OH, COOH) we find DND-H to show the brightest fluorescence, which only weakly depends on the excitation wavelength. DND-OH and DND-COOH only show weak fluorescence. Based on the characteristic decay time, we attribute the fluorescence observed in solution to residual non-diamond carbon present in all samples. Fluorescence from NV centers in diamond is found in DND-H, OH and COOH samples. A summary of the key findings is presented in Figure 9. Pure diamond particles are optically transparent and completely nonfluorescent. The introduction of defects and impurities is known to create fluorescent color centers like the NV center. Our results suggest that the surface chemistry of detonation nanodiamonds is critical for their fluorescence properties in two ways: first, through the creation of fluorescent defects on or close to the particle surface via the local reconfiguration of carbon atoms as a result of the functionalization process and, second, through the interaction of the modified particle surface with the optical defects and color centers inside the particles’ diamond core. The fundamental photophysics of both effects are currently poorly 10931

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.7b04647. Extinction spectra, EELS spectra and analysis, TEM images, raw in-solution fluorescence spectra, schematic illustration of the fluorescence spectroscopy setup, fluorescence spectra of fluorescent HPHT nanodiamonds in water, fluorescence decay traces of all DND types in solution, bright-field microscopy images, brightness comparison of DND particles on Si wafer, and comparison of fluorescence spectra of particles in solution and particles on a Si wafer (PDF)

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected].

Figure 9. Summary of the key findings of this study.

ORCID

Philipp Reineck: 0000-0003-1549-937X Desmond W. M. Lau: 0000-0002-1437-8990 Emma R. Wilson: 0000-0002-4066-5862 Kate Fox: 0000-0001-8090-3215 Cholaphan Deeleepojananan: 0000-0002-6369-3911 Vadym N. Mochalin: 0000-0001-7403-1043 Brant C. Gibson: 0000-0002-7109-2796

understood, and we expect our results to spark further research into the fluorescence of DND. Furthermore, our results may enable fluorescence-based imaging or detection of structures and materials in applications such as nanomedicine or materials technology where DND has thus far only been used for its superior physical (mainly mechanical) and chemical properties.

Author Contributions

METHODS

V.N.M., P.R., and B.C.G. conceived and planned the study. V.N.M. and C.D. purified, modified, and characterized the detonation nanodiamond samples using FTIR. P.R. performed all optical characterizations, conducted all data analysis (except for EELS), generated all graphs, and created all illustrations. D.W.M.L. performed all TEM and EELS experiments and data analysis. E.R.W. performed DLS and zeta-potential measurements. K.F. and M.R.F. contributed to the TEM and EELS data analysis and interpretation.

Sample Preparation. H-, COOH-, OH-, and EDA-functionalized DND powders were dissolved in deionized water at a concentration of 1 mg/mL and sonicated for 1 h at 150 W using a horn sonicator with a 66% duty cycle. DND-H, DND-COOH, DND-OH, and DND-ODA solutions were centrifuged at 1000 rcf for 5 min, and the supernatant was used for experiments. DND-ODA is well dispersible in chloroform, and the centrifugation did not change the concentration significantly. The DND-ODA and DND-EDA solutions were diluted to a concentration of 0.2 and 0.25 mg/mL in chloroform and water, respectively. The concentrations of all other particles in solution were estimated using nanoparticle tracking analysis (NanoSight NS300, Malvern Instruments) to be 0.4 ± 0.2 mg/mL (DND, DND-COOH, and DND-OH) and 0.2 ± 0.1 mg/mL (DND-H). These solutions were used for in-solution spectroscopy experiments. For confocal microscopy/spectroscopy experiments, 40 μL of NP solution was drop-cast on a silicon wafer substrate on a hot plate at 100 °C and allowed to dry. HPHT ND particles were purchased from Nabond Technologies and processed as described in the SI. In-Solution Spectroscopy. A collimated laser beam (Fianium, WhiteLaseSC 400) was weakly focused into a cuvette with an optical path of 10 mm, 2 mm diameter (FireflySci, Inc., 18FL Micro). The fluorescence was collected, fiber coupled, and analyzed with a spectrometer (Princeton Instruments, SpectraPro with a PIXIS CCD camera) to obtain fluorescence spectra. For time-resolved direct fluorescence decay traces, photons were collected with avalanche photo diodes (APD, Excelitas, SPCM-AQRH-14) and analyzed with a correlator card (Picoquant, TimeHarp 260). See the SI for a schematic drawing of the setup. Confocal Imaging and Spectroscopy. The same light source and detectors used for in-solution spectroscopy were used in a custombuilt confocal fluorescence microscope setup. The excitation beam (λex = 500 nm) was focused onto the sample using a 100× air objective (0.9 NA). Fluorescence was filtered by a 532 nm dichroic and longpass filter (Semrock) and fiber coupled for detection by the above APD and spectrometer.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by the ARC Centre of Excellence for Nanoscale BioPhotonics and ARC grants (FT110100225, LE140100131, CE140100003). B.C.G. acknowledges the support of an ARC Future Fellowship. We are thankful to Dr. K. Turcheniuk for initial help with synthesis of DND samples. We acknowledge the use of the RMIT Microscopy and Microanalysis Facility (RMMF) and the MicroNano Research Facility (MNRF) at RMIT University. REFERENCES (1) Mochalin, V. N.; Shenderova, O.; Ho, D.; Gogotsi, Y. The Properties and Applications of Nanodiamonds. Nat. Nanotechnol. 2011, 7, 11−23. (2) Chou, C. C.; Lee, S. H. Tribological Behavior of NanodiamondDispersed Lubricants on Carbon Steels and Aluminum Alloy. Wear 2010, 269, 757−762. (3) Mochalin, V. N.; Gogotsi, Y. Nanodiamond-Polymer Composites. Diamond Relat. Mater. 2015, 58, 161−171. (4) Turcheniuk, K.; Mochalin, V. N. Adsorption Behavior and Reduction of Copper (II) Acetate on the Surface of Detonation 10932

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano Nanodiamond with Well Defined Surface Chemistry. Carbon 2016, 109, 98−105. (5) Zhu, Y.; Li, J.; Li, W.; Zhang, Y.; Yang, X.; Chen, N.; Sun, Y.; Zhao, Y.; Fan, C.; Huang, Q. The Biocompatibility of Nanodiamonds and Their Application in Drug Delivery Systems. Theranostics 2012, 2, 302−312. (6) Turcheniuk, K.; Mochalin, V. N. Biomedical Applications of Nanodiamond (Review). Nanotechnology 2017, 28, 252001. (7) Shenderova, O. A.; McGuire, G. E. Science and Engineering of Nanodiamond Particle Surfaces for Biological Applications (Review). Biointerphases 2015, 10, 30802. (8) Bradac, C.; Gaebel, T.; Naidoo, N.; Sellars, M. J.; Twamley, J.; Brown, L. J.; Barnard, A. S.; Plakhotnik, T.; Zvyagin, A. V.; Rabeau, J. R. Observation and Control of Blinking Nitrogen-Vacancy Centres in Discrete Nanodiamonds. Nat. Nanotechnol. 2010, 5, 345−349. (9) Reineck, P.; Capelli, M.; Lau, D. W. M.; Jeske, J.; Field, M. R.; Ohshima, T.; Greentree, A. D.; Gibson, B. C. Bright and Photostable Nitrogen-Vacancy Fluorescence from Unprocessed Detonation Nanodiamond. Nanoscale 2017, 9, 497−502. (10) Wang, Z.; Xu, C.; Liu, C. Surface Modification and Intrinsic Green Fluorescence Emission of a Detonation Nanodiamond. J. Mater. Chem. C 2013, 1, 6630. (11) Smith, B. R.; Gruber, D.; Plakhotnik, T. The Effects of Surface Oxidation on Luminescence of Nano Diamonds. Diamond Relat. Mater. 2010, 19, 314−318. (12) Shenderova, O. A.; Vlasov, I. I.; Turner, S.; Van Tendeloo, G.; Orlinskii, S. B.; Shiryaev, A. A.; Khomich, A. A.; Sulyanov, S. N.; Jelezko, F.; Wrachtrup, J. Nitrogen Control in Nanodiamond Produced by Detonation Shock-Wave-Assisted Synthesis. J. Phys. Chem. C 2011, 115, 14014−14024. (13) Borjanovic, V.; Lawrence, W. G.; Hens, S.; Jaksic, M.; Zamboni, I.; Edson, C.; Vlasov, I.; Shenderova, O.; McGuire, G. E. Effect of Proton Irradiation on Photoluminescent Properties of PDMSNanodiamond Composites. Nanotechnology 2008, 19, 455701. (14) Kirmani, A. R.; Peng, W.; Mahfouz, R.; Amassian, A.; Losovyj, Y.; Idriss, H.; Katsiev, K. On the Relation between Chemical Composition and Optical Properties of Detonation Nanodiamonds. Carbon 2015, 94, 79−84. (15) Chung, P. H.; Perevedentseva, E.; Cheng, C. L. The Particle Size-Dependent Photoluminescence of Nanodiamonds. Surf. Sci. 2007, 601, 3866−3870. (16) Dolenko, T. A.; Burikov, S. A.; Rosenholm, J. M.; Shenderova, O. A.; Vlasov, I. I. Diamond − Water Coupling Effects in Raman and Photoluminescence Spectra of Nanodiamond Colloidal Suspensions. J. Phys. Chem. C 2012, 116, 24314−24319. (17) Vervald, A. M.; Burikov, S. A.; Shenderova, O. A.; Nunn, N.; Podkopaev, D. O.; Vlasov, I. I.; Dolenko, T. A. Relationship Between Fluorescent and Vibronic Properties of Detonation Nanodiamonds and Strength of Hydrogen Bonds in Suspensions. J. Phys. Chem. C 2016, 120, 19375−19383. (18) Mochalin, V. N.; Gogotsi, Y. Wet Chemistry Route to Hydrophobic Blue Fluorescent Nanodiamond. J. Am. Chem. Soc. 2009, 131, 4594−4595. (19) Baranov, P. G.; Soltamova, A. A.; Tolmachev, D. O.; Romanov, N. G.; Babunts, R. A.; Shakhov, F. M.; Kidalov, S. V.; Vul’, A. Y.; Mamin, G. V.; Orlinskii, S. B.; Silkin, N. I. Enormously High Concentrations of Fluorescent Nitrogen-Vacancy Centers Fabricated by Sintering of Detonation Nanodiamonds. Small 2011, 7, 1533− 1537. (20) Smith, B. R.; Inglis, D. W.; Sandnes, B.; Rabeau, J. R.; Zvyagin, A. V.; Gruber, D.; Noble, C. J.; Vogel, R.; Osawa, E.; Plakhotnik, T. Five-Nanometer Diamond with Luminescent Nitrogen-Vacancy Defect Centers. Small 2009, 5, 1649−1653. (21) Bradac, C.; Gaebel, T.; Pakes, C. I.; Say, J. M.; Zvyagin, A. V.; Rabeau, J. R. Effect of the Nanodiamond Host on a Nitrogen-Vacancy Color-Centre Emission State. Small 2013, 9, 132−139. (22) Reineck, P.; Francis, A.; Orth, A.; Lau, D. W. M.; Nixon-Luke, R. D. V.; Rastogi, I. D.; Razali, W. A. W.; Parker, L. M.; Sreenivasan, V. K. A.; Brown, L. J.; Gibson, B. C. Brightness and Photostability of

Emerging Red and near-IR Fluorescent Nanomaterials for Bioimaging. Adv. Opt. Mater. 2016, 4, 1549−1557. (23) Fu, M.; Ehrat, F.; Wang, Y.; Milowska, K. Z.; Reckmeier, C.; Rogach, A. L.; Stolarczyk, J. K.; Urban, A. S.; Feldmann, J. Carbon Dots: A Unique Fluorescent Cocktail of Polycyclic Aromatic Hydrocarbons. Nano Lett. 2015, 15, 6030−6035. (24) Eda, G.; Lin, Y. Y.; Mattevi, C.; Yamaguchi, H.; Chen, H. A.; Chen, I. S.; Chen, C. W.; Chhowalla, M. Blue Photoluminescence from Chemically Derived Graphene Oxide. Adv. Mater. 2010, 22, 505−509. (25) Nourbakhsh, A.; Cantoro, M.; Vosch, T.; Pourtois, G.; Clemente, F.; van der Veen, M. H.; Hofkens, J.; Heyns, M. M.; De Gendt, S.; Sels, B. F. Bandgap Opening in Oxygen Plasma-Treated Graphene. Nanotechnology 2010, 21, 435203. (26) Gokus, T.; Nair, R. R.; Bonetti, A.; Bohmler, M.; Lombardo, A.; Novoselov, K. S.; Geim, A. K.; Ferrari, A. C.; Hartschuh, A. Making Graphene Luminescent by Oxygen Plasma Treatment. ACS Nano 2009, 3, 3963−3968. (27) Shang, J.; Ma, L.; Li, J.; Ai, W.; Yu, T.; Gurzadyan, G. G. The Origin of Fluorescence from Graphene Oxide. Sci. Rep. 2012, 2, 792. (28) Lim, S. Y.; Shen, W.; Gao, Z. Carbon Quantum Dots and Their Applications. Chem. Soc. Rev. 2015, 44, 362−381. (29) Zaitsev, A. M. Vibronic Spectra of Impurity-Related Optical Centers in Diamond. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61, 12909−12922. (30) Kondo, T.; Neitzel, I.; Mochalin, V. N.; Urai, J.; Yuasa, M.; Gogotsi, Y. Electrical Conductivity of Thermally Hydrogenated Nanodiamond Powders. J. Appl. Phys. 2013, 113, 214307. (31) Zheng, W.; Hsieh, Y.; Chiu, Y.; Cai, S.; Cheng, C.; Chen, C. Organic Functionalization of Ultradispersed Nanodiamond: Synthesis and Applications. J. Mater. Chem. 2009, 19, 8432. (32) Osswald, S.; Yushin, G.; Mochalin, V.; Kucheyev, S. O.; Gogotsi, Y. Control of sp2/sp3 Carbon Ratio and Surface Chemistry of Nanodiamond Powders by Selective Oxidation in Air. J. Am. Chem. Soc. 2006, 128, 11635−11642. (33) Mochalin, V. N.; Neitzel, I.; Etzold, B. J. M.; Peterson, A.; Palmese, G.; Gogotsi, Y. Covalent Incorporation of Aminated Nanodiamond into an Epoxy Polymer Network. ACS Nano 2011, 5, 7494−7502. (34) Shenderova, O.; Koscheev, A.; Zaripov, N.; Petrov, I.; Skryabin, Y.; Detkov, P.; Turner, S.; Van Tendeloo, G. Surface Chemistry and Properties of Ozone-Purified Detonation Nanodiamonds. J. Phys. Chem. C 2011, 115, 9827−9837. (35) Mochalin, V.; Osswald, S.; Gogotsi, Y. Contribution of Functional Groups to the Raman Spectrum of Nanodiamond Powders. Chem. Mater. 2009, 21, 273−279. (36) Acosta, V. M.; Bauch, E.; Ledbetter, M. P.; Santori, C.; Fu, K. M. C.; Barclay, P. E.; Beausoleil, R. G.; Linget, H.; Roch, J. F.; Treussart, F.; Chemerisov, S.; Gawlik, W.; Budker, D. Diamonds with a High Density of Nitrogen-Vacancy Centers for Magnetometry Applications. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80, 1−15. (37) Xiao, J.; Liu, P.; Li, L.; Yang, G. Fluorescence Origin of Nanodiamonds. J. Phys. Chem. C 2015, 119, 2239−2248. (38) Mkandawire, M.; Pohl, A.; Gubarevich, T.; Lapina, V.; Appelhans, D.; Rödel, G.; Pompe, W.; Schreiber, J.; Opitz, J. Selective Targeting of Green Fluorescent Nanodiamond Conjugates to Mitochondria in HeLa Cells. J. Biophotonics 2009, 2, 596−606. (39) Su, S.; Wei, J.; Zhang, K.; Qiu, J.; Wang, S. Thermo- and pHResponsive Fluorescence Behaviors of Sulfur-Functionalized Detonation Nanodiamond-poly(N-Isopropylacrylamide). Colloid Polym. Sci. 2015, 293, 1299−1305. (40) Vlasov, I. I.; Shenderova, O.; Turner, S.; Lebedev, O. I.; Basov, A. A.; Sildos, I.; Rähn, M.; Shiryaev, A. A.; Van Tendeloo, G. Nitrogen and Luminescent Nitrogen-Vacancy Defects in Detonation Nanodiamond. Small 2010, 6, 687−694. (41) Wang, Y.; Hu, A. Carbon Quantum Dots: Synthesis, Properties and Applications. J. Mater. Chem. C 2014, 2, 6921−6939. 10933

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934

Article

ACS Nano (42) Dong, W.; Zhou, S.; Dong, Y.; Wang, J.; Ge, X.; Sui, L. The Preparation of Ethylenediamine-Modified Fluorescent Carbon Dots and Their Use in Imaging of Cells. Luminescence 2015, 30, 867−871. (43) Niu, W.-J.; Li, Y.; Zhu, R.-H.; Shan, D.; Fan, Y.-R.; Zhang, X.-J. Ethylenediamine-Assisted Hydrothermal Synthesis of Nitrogen-Doped Carbon Quantum Dots as Fluorescent Probes for Sensitive Biosensing and Bioimaging. Sens. Actuators, B 2015, 218, 229−236. (44) Hu, S.; Trinchi, A.; Atkin, P.; Cole, I. Tunable Photoluminescence Across the Entire Visible Spectrum from Carbon Dots Excited by White Light. Angew. Chem., Int. Ed. 2015, 54, 2970−2974.

10934

DOI: 10.1021/acsnano.7b04647 ACS Nano 2017, 11, 10924−10934