Fi

Pa

b

c

a

ARRAA

KCPCC

1

fo[ot(a[lcv

Ifnta

f

0d

Colloids and Surfaces B: Biointerfaces 91 (2012) 34– 40

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces

j our na l ho me p age: www.elsev ier .com/ locate /co lsur fb

luorescence behavior of non-functionalized carbon nanoparticles and theirn vitro applications in imaging and cytotoxic analysis of cancer cells

radip Kumara, Ramavtar Meenab, R. Paulrajb, A. Chanchalc, A.K. Vermac, H.B. Bohidara,∗

School of Physical Sciences, Jawaharlal Nehru University, New Delhi 110067, IndiaSchool of Environmental Sciences, Jawaharlal Nehru University, New Delhi 110067, IndiaNanobiotech Lab, Department of Zoology, Kirori Mal College, University of Delhi, New Delhi 110007, India

r t i c l e i n f o

rticle history:eceived 28 September 2011eceived in revised form 14 October 2011ccepted 18 October 2011vailable online 25 October 2011

eywords:arbon nanoparticles

a b s t r a c t

We report fluorescence behavior in non-functionalized carbon nanoparticles (NCNP) prepared from lampsoot and their application in imaging of normal and cancer cells. Structural characterization of these parti-cles by Raman spectroscopy showed characteristic peaks located at 1350 and 1590 cm−1 corresponding tothe diamond-like (D) and graphite-like (G) bands of the carbon allotropes respectively with the character-istic ratio ID/IG = 2.24. X-ray diffraction study confirmed the presence of amorphous as well as graphitizedcarbon in these nanostructures with minimum grain size ≈2 nm. A typical luminescence lifetime mea-sured by time resolved fluorescence spectroscopy was obtained 3.54 ns. The photoluminescence behavior

hotoluminescenceell imagingytotoxicity

of these particles was excitation dependent and gave off blue, green and red fluorescence under UV,blue and green excitation, respectively. Cellular uptake of these NCNP yielded excellent results for cellimaging of human embryonic kidney, lung carcinoma and breast adenocarcinoma cells. Cell imaging wasfurther correlated with cytotoxicity in the above mentioned cell lines and also in leukemia cell lines. Dosedependant cytotoxicity was observed after 24 h up to 48 h of incubation of nanoparticles. Fluorescencemicroscopy of nanoparticle-cell interaction clearly indicated aggregation of the particles.

. Introduction

Carbonaceous nanostructures, like nano-colloids, onions,ullerenes, graphene sheets, single and multiwall carbon nan-tubes have promising applications in various biological fields1,2]. Understanding their physical, chemical, electronic andptical properties to explore their application potential have ledo an explosion of extensive research. The carbon nanoparticlesCNP) are normally synthesized, purified or functionalized in

solvent medium, although non-solvent-based methods exist3–7]. Various experimental methods [8–11] for example, pulseaser deposition, carbon arc technique and microwave-plasmahemical-vapor deposition were developed to produce CNP ofarious conformations and sizes.

Carbon soot is a source of polydisperse and ultrafine particles.t has been widely used as black ink in paints and in fountain pensor ages. Recently soot originated CNP has been rediscovered as a

ew class of carbonaceous nanostructures with interesting proper-ies [12,13]. In vitro testing of CNP revealed significant antimicrobialctivity against bacteria called “Klebsiella pneumonia” [13]. The syn-

∗ Corresponding author. Tel.: +91 11 2670 4637/2671 7562;ax: +91 11 2674 1837.

E-mail address: bohi007@mail.jnu.ac.in (H.B. Bohidar).

927-7765/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.colsurfb.2011.10.034

© 2011 Elsevier B.V. All rights reserved.

thesis of carbon soot nanostructures is an easy and inexpensivemethod. However, in order to improve their aqueous dispersibilityand fluorescence yield, these particles are often treated with oxida-tive acid protocol. This produces negatively charged hydrophilicCNP [12]. Fluorescent CNP are not deeply studied due to lack of reli-able preparative methods. Common route for making fluorescentCNP include laser ablation of graphite [14], creation of point defectsin diamond [15] and wet electrochemical methods [16]. Liu et al.[12] have reported the soot-based synthesis produces particles ofdifferent colors and used gel electrophoresis technique for isola-tion of different particle fractions, which was a difficult task. In alltechniques used up to now, the CNP surface was modified in orderto achieve fluorescence. On the other hand, the understanding offluorescence emanating from CNP can be considered incomplete.For example, information on the microstructure and specifics ofsurface ligand binding has remained unclear. The organic surfacepassivation details are not sufficient to aid understanding of thesurface states beneficial for fluorescence emission. The origin offluorescence in organic solvents is due to the trapping of excitationenergy on the nano-structure surface has been reported [12,14,17].

The fluorescent carbon nanotubes are known for a long time;

these samples are normally prepared through rigorous proto-cols that make the product expensive. Therefore, the search forbenign fluorescent CNP through an easy and inexpensive routehas become an urgent challenge. Herein, we report an easy and

faces B

ewoc(etfltnnbt

2

2

m[mcctpcttduissis

2

ftIssiaW0wcaPTwgF

2

slau

P. Kumar et al. / Colloids and Sur

conomical approach to prepare fluorescent CNP from lamp sootithout any requirement for surface functionlization. The objective

f the present work was: (i) to undertake a structural and opti-al characterization of non-functionalized carbon nanoparticlesNCNP), obtained from home-made carbon soot using an array ofxperimental techniques like X-ray diffraction (XRD), Raman spec-roscopy, transmission electron microscopy (TEM), time resolveduorescence spectroscopy (TRFS), UV–vis and fluorescence spec-roscopy, and (ii) to apply these fluorescent NCNP in imagingormal Human embryonic kidney (HEK-297), Human breast ade-ocarcinomm (MCF-7) and Human lung carcinoma (H-1299) cellsy using the confocal laser scanning microscopy, and (iii) to assesshe cytotoxicity of these NCNP in cancerous and normal cells.

. Experimental

.1. Synthesis of fluorescent CNP

The carbon nanoparticles were synthesized by lamp sootethod which has been broadly described in our previous works

13,18] and an exact protocol is provided elsewhere [19]. In brief,ustard oil (commercial) was used as fuel to burn a lamp with a

otton (sterilized surgical grade) wick. A degreased and thoroughlyleaned metal (Cu, Ag and Au, 99% pure grade) plate was kept overhe flame at a distance of 10 cm. The flame soot deposited on thelate surface was scratched with a previously cleaned spatula andollected in a sterilized glass bottle. Following method was usedo prepare the aqueous dispersion of fluorescent NCNP at roomemperature (20 ◦C). Typically, 0.1 g of the powdered material wasispersed in 100 ml of de-ionized water and continuously mixedsing a magnetic stirrer approximately for 20 h. The dispersibil-

ty was observed to be very poor, so we allowed the dispersion totand for 2 h following mixing, to enable unsuspended particles toediment and settle down. The supernatant was collected and son-cated for 10 min to break the big clusters and was used for furthertudies.

.2. Cell imaging

HEK-297 cells, cancerous cells MCF-7 and H-1299 were obtainedrom National Center for Cell Science, Pune, India. They were cul-ured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Biologicalndustries, Israel) supplemented with 10% heat inactivated fetal calferum (FCS) (Biological Industries, Israel) and IX Penstrep antibioticolution (Biological Industries, Israel) and incubated in fully humid-fied 5% CO2 incubator at 37 ◦C. All cells were seeded in culture flasksnd they were divided into treatment group and control group.hen ∼70% of the growth occurred, the cells were washed with

.1 M phosphate buffer saline (PBS) and old media were replacedith fresh media. Culture plates were treated with various con-

entrations (0.1, 1, 1.5 and 2 mg/ml) of NCNP and were incubatedt 37 ◦C and 5% CO2 for 24 h. All plates were washed with 0.1 MBS and cells were collected by trypsinization (0.05% trypsinase).he cell pellets were dissolved in 1 ml of 0.1 M PBS solution andere imaged under bright field, UV (405 nm), blue (488 nm), and

reen (543 nm) excitation wavelengths with Olympus Fluo ViewTM

V1000 laser scanning confocal microscope.

.3. Cell uptake

Cells were seeded at a density of 104 cells/ml in a glass chamber

lide (Nalge Nunc International, NY) and maintained with regu-ar changes in media. The uptake experiments were conductedfter cell reached confluence in a chamber slide. For nanoparticleptake, the cells were washed with fresh medium and medium

: Biointerfaces 91 (2012) 34– 40 35

was replaced with NCNP dispersion. The cells were then incu-bated at 37 ◦C in a humidified 5% CO2 and 95% air atmosphere.After 4 h post-incubation, the glass slide chambers were completelywashed with Hank’s buffered salt solution (HBSS) buffer to removethe nonspecific binding particles. The cells were fixed with 4%paraformaldehyde for 10 min and after that the cover slip waswashed with PBS (pH 7.4) twice, after which it was mounted in Dis-tyrene Plasticizer and Xylene (DPX). The fixed cells were subjectedto microscopic analysis on a Nikon Eclipse 90i Epi-fluorescenceupright microscope equipped with a Nikon DXM 1200 digital cam-era and viewed at 20× magnifications.

2.4. Cytotoxicity assay

For cytotoxicity assay studies, the cells were maintainedin Iscove’s modified Dulbecco’s medium (Invitrogen/Gibco; Cat.No. 21980-032) supplemented with 2 mM GlutaMAX (Invitro-gen/Gibco; Cat. No. 35050-038), 100 �g/ml streptomycin, 100 U/mlpenicillin and 10% FCS and incubated at 37 ◦C in an atmosphereof 95% air and 5% CO2 at 90% relative humidity. Cytotoxic effecton the HL-60 (acute promyelocytic leukemia) and K-562 (chronicmyelogenous leukemia) cell lines was also assessed. Briefly,5 × 103cells/well were incubated in 100 �l of RPMI-1640 supple-mented with 10%FCS, 2 mM l-glutamine and various concentrationof NCNP. The cytotoxic effects of NCNP were tested using a stan-dard MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide) assay, in a 96-well microtiter plate for 24 h. MTT is anon-radioactive assay done routinely to assess the viability of thecell culture. After the incubation period of 24 h and 48 h, 20 �l ofMTT dye solution (5 mg/ml in PBS pH 7.4) was added to each well.After 4 h of further incubation the formazan crystals formed by thecellular reduction of MTT were dissolved in 150 �l of DMSO andplates were read on an ELISA-reader using 570 nm filter. All mea-surements were done in triplicates. The relative cell viability (%)related to control wells containing cells without nanoparticles wascalculated by

[A]test

[A]control× 100 (1)

where [A]test is absorbance of the test sample and [A]control isthe absorbance of the control sample. By using the non-radioactiveassay for assessing the proliferation of cells, we were able to quan-tify the amount of MTT cleaved, which is directly proportional tothe viable cell population.

2.5. Instrumentation

Average particle size and morphology information was obtainedby using a Fei-Philips, Morgagni TEM (Digital TEM with imageanalysis system and maximum magnification = 280,000×) operat-ing at a voltage 100 kV. The aqueous dispersion was drop-cast ontoa carbon-coated copper grid, and the grid was air dried at roomtemperature (20 ◦C) before loading into the microscope. The XRDpatterns (Bragg peaks) were recorded with a PANalytical X’Pert PROdiffractometer using a solid state detector with a monochromatizedCu k�1 (�Cu = 1.54060 A) radiation source at 45 kV. A Raman spec-trum of carbon soot powder was recorded using a Renishaw Ramanmicroscope with Ar-ion laser excitation at 514 nm, and at 50 mWpower. UV–vis absorption spectra were collected by using a Cecilmodel CE-7200 (Cecil Instrument, UK) spectrophotometer. All thefluorescent spectra were obtained by using a Varian Cary EclipseFluorescence spectrophotometer. For lifetime measurement, we

used TRFS which is a very good technique to probe lifetime ofnanoparticles, macromolecules, etc. in solution. The lifetime exper-iment was performed using time-correlated single photon counting(TCSPC) setup (FL920, Edinburgh Instrument, UK) with 375 nm

3 faces B: Biointerfaces 91 (2012) 34– 40

dp(dwt

F

w(vc

〈

rcswbciUssTH

3

3

meospiastapotl4iAoT(ua

Cwfiectw

Fig. 1. (A) XRD pattern of carbon soot, which indicate that crystallites were presentin our NCNP material. See Table 2 for analysis. (B) Raman spectra of carbon sootparticles. The Raman spectra fitted with Gaussian function and I(D)/I(G) ≈ 2.24.

6 P. Kumar et al. / Colloids and Sur

iode laser and the fluorescence decay was collected at magic angleolarization (55◦). The time resolution for TCSPC setup was ∼120 psmeasured with LUDOX solution). Further details about TRFS andata analysis can be found elsewhere [20]. The fluorescence decayas fitted with sum of-three-exponentials to model the decay pat-

ern given by

(t) = a0 + a1 exp(−t

�1

)+ a2 exp

(−t

�2

)+ a3 exp

(−t

�3

)(2)

here a0 is the time shift between instrument response functionIRF) and sample. The relaxation times �1, �2, and �3 correspond toarious lifetimes of characteristic excited states.The average timeonstant or mean lifetime is given as

�〉 =∑

i

ai�i (3)

A Fluorescence microscope (Olympus IX71) was used for fluo-escent imaging. For imaging, the nanoparticle dispersion was dropast on a cleaned glass slide and the sample was sealed with coverlip. The microscope was used in inverted mode and the imagesere captured under bright field and blue excitation (470–495 nm)

ands derived from a 100 W mercury lamp. For cell imaging appli-ation, we followed the same procedure for slide preparation andmages were taken by confocal laser scanning microscope underV, blue and green excitations. An epi-fluorescent Upright Micro-

cope from Nikon (model Eclipse 90i) loaded with NIS Element ARoftware was used for capturing and analyzing the cellular images.he MTT plates were read at a wavelength 570 nm in the Synergy-T instrument (Biotek, USA).

. Results and discussion

.1. Effect of substrates

Carbon soot was deposited on the surface of three transitionetal substrates of Cu, Ag and Au using identical protocol. How-

ver, fluorescent CNP could be only obtained from soot depositedn copper surface (typical TEM picture of such nanoparticles ishown in inset of Fig. 1A). In order to understand this behavior, thehysical characteristics of these transition metals vis-a-vis carbon

s compared and presented in Table 1. For most transition met-ls the energy in d-orbitals tends to be higher than that in the-orbitals of a higher quantum number. Therefore, valence elec-rons will occupy the latter first. However, an element will acquiren extra stability if the five d-orbitals are either half full or com-letely full. Sometimes this condition can be satisfied by movingne of the two s-electrons to d-orbital. Such elements with s1 elec-ron configuration are Cu, Ag and Au. Unlike 3d elements that haveittle overlap between higher energy 3d-orbitals and lower energys orbitals, the electron-rich 5d elements possess higher energies

n 6s orbitals due to the increased electron-electron repulsion. Cu,g and Au have already moved one electron from d-orbitals to s-rbitals. As a result, their d-orbitals are full. The data presented inable 1 distinguish Cu as compared to Ag and Au in three aspects:i) Cu is associated with interatomic distance that is ≈13% less, (ii)nit cell of Cu is ≈8% smaller and (iii) Cu provides better solubility,s compared to Ag and Au.

Now the question arises: why nanoparticles obtained only fromu substrate exhibit fluorescence properties? NCNP dispersed inater yielded similar UV–vis absorption spectra (data not shown)

or nanoparticles derived from all the three metal surfaces whereasn case of organic solvents soot collected only on Cu surface

xhibited fluorescence explicitly. A detail fluorescence study in allommon organic solvents has been reported [21]. The soot con-ained amorphous as well crystalline carbon (Fig. 1A), thus thereas no lattice mismatch mechanism as reported for graphene. On

Inset of Fig. 1A shows the TEM snapshots of NCNP dispersed in water. Particle size≈30–40 nm.

the basis of our experimental results, we could not explain exactmechanism for this anomalous, but interesting observation. How-ever, the following deserves attention. The reactivity of carbon withdifferent transition metal surfaces might be one of the possible rea-sons that can be attributed to the observation of this anomalousbehavior. The transition metal tends to react with carbon by over-lapping its d-orbitals with carbon’s p-orbitals. Hence, the reactivityof transition metal toward carbon is primarily determined by itselectronic state, specifically, the number of electron vacancy of d-orbitals. The reactivity of transition metal with carbon increaseswith its number of electron vacancies in d-orbitals. Element Cu,Ag and Au have similar electronic configuration 3d104s1, 4d105s1

and 5d106s1, respectively with no d-vacancies and are inert rela-tive to carbon (see Table 1). In the absence of sufficient numberof d-vacancies, the reactivity of these metals toward carbon is notstrong enough to form carbide. In this case, carbon will not be tiedup in fixed places, but would form free-moving solution. However,the amount of such solutes is also dependent on the reactivity oftransition metals. Thus, elements Cu, Ag and Au with no vacanciesin d-orbitals can dissolve only trace amount of carbon. It is knownthat coinage metals, in particular Cu, are found to be more favor-able for the growth of carbon nanostructures [22]. The case of Cuis particularly noteworthy for the following reasons. First, the low-est carbon diffusion barriers, suggesting that synthesis could takeplace at much lower temperatures. Secondly, Cu has low chemi-cal potentials than Ag and Au show that graphitic fragments bindindifferently to surface.

The synthesis and luminescence studies of various allotropes

of carbon nanostructures, which includes soot derived particles,is discussed in excellent details in a review by Baker and Baker[24]. For example, Lu et al. [12], Ray et al. [25] and Wang et al.

P. Kumar et al. / Colloids and Surfaces B: Biointerfaces 91 (2012) 34– 40 37

Table 1Physical characteristics of metals used as substrate in depositing carbon soot [23].

Substrate Crystal structure Atomic No. Melting point (◦C) Vacancy d-orbitals Carbon solubility d/A (Interatomic distance) Unit cell size (A)

[patcr

3

bopmifms

b[iFatf

d

wl�s

tararp1ppprt

L

wa2n

a2oa

[17]. A widely accepted mechanism is the radiative recombinationof excitons from carbon nanoparticles of different sizes, and thecontribution arising from distribution of different emissive trap

0.0

0.4

0.8

1.2

750600450300

7006506005505004504000

30

60

90

120

150

Wavelength (n m)

254 nm

A

Abs

orba

nce

5505004504003503000

50

100

150

PL In

ten.

(a.u

.)

Emi. Wav e.(nm)

395 nm

B

cb

PL In

tens

ity (a

.u.)

Emission Wavele ngth (nm )

a

Ag fcc 47 962 0

Au fcc 79 1050 0

Cu fcc 29 1110 0

26], collected soot by placing a piece of aluminum foil or a glasslate atop a burning candle. The collected soot was followed byn oxidative acid treatment to introduce OH and CO2H groups tohe surfaces to make them negatively charged and hydrophilic. Inontrast, we adopted a single step preparation protocol that did notequire surface modification.

.2. Structural and optical characterization

The average nanoparticle cluster size and zeta potential haveeen reported in our previous work [18]. The average cluster sizef CNP determined from light scattering data was ≈165 nm and zetaotential ≈ −18 mV in de-ionized water. Soot contained mainly ele-ental carbon (93%) and oxygen (7%) atoms [13]. The TEM picture

s shown in inset of Fig. 1A depicts 30 nm diameter nanoparticlesormed into big clusters. However cluster size cannot be deter-

ined exactly from this picture, as this corresponds to dehydratedample where solvent-mediated dispersion character has been lost.

Flame generated CNP are known to contain a variety of car-on nanostructures such as graphite, carbon nanorods, tubes, etc.27,28]. Thus, it was felt imperative to deduce the crystalline behav-or of these particles. The XRD spectrum of carbon soot is shown inig. 1A, which indicated that soot contained amorphous as wells crystalline allotropes of carbon. The average crystallite sizes ofhe nanoparticles (d) were determined by using Debye–Scherrerormula [29] given by

= K�Cu

cos �(4)

here K is a constant with approximate value = 0.9, �Cu is the wave-ength of radiation (1.54060 A), is full width at half maxima and

is the position of the maxima. Using this formula, the crystalliteize was calculated which is given in Table 2.

Raman spectra of carbon soot and the fitting of D and G bandso Gaussian function are shown in Fig. 1B. The two bands, labeled Dnd G, refer to the diamond and graphite allotropes of carbon mate-ial, respectively [35]. The D band was found located at 1350 cm−1

nd it owes its origin to the breathing modes of sp2 atoms inings which refers to disordered graphitic phase and indicates theresence of nanocrystalline graphite. The G band was found near590 cm−1 and this band arises due to the bond stretching of allairs of sp2 atoms in both rings and chains, and indicates theresence of single crystal graphite. The grain size of graphite wasroposed to be inversely proportional to the integrated intensityatio between the D peak and G peak [36]. The relationship betweenhe grain size and the intensity ratio is given by

a = 44[I(D)/I(G)]

(in A) (5)

here La is the grain size and I(D) and I(G) are integrated intensityssociated with the D peak and G peak, respectively (the ratio was.24). Using the above formula, the calculated grain size of carbonanoparticles ≈2 nm.

UV–vis absorption spectra of NCNP in water showed a strong

bsorption band at 254 nm (see Fig. 2A). The absorption band at17 nm is attributed to �–�* band transition in small graphiticr amorphous carbon grains whereas the absorbance band seenround 216–225 nm was argued to be originating from the

0.01 2.95 4.080.01 2.94 4.070.04 2.61 3.68

physical size of the graphitic structure [37]. De Heer and Ugartealso observed the 264 nm absorption band in aqueous dispersionof carbon soot [38]. The absorption peak noticed at 254 nm may bearising due to the presence of graphitic or amorphous grains. Flu-orescence spectra of NCNP, dispersed in water are shown in Fig. 2.The spectra taken at 405, 488, and 543 nm excitation are shown inFig. 2B. A clear emission peak was observed at 300 nm excitationwavelength which is shown in the inset of Fig. 2A. No significantemission peak was obtained at excitations 488 and 543 nm and thisis may be due the presence of less emissive trap states at higherexcitation wavelength.

In a related study Wang et al. [39] have reported band-gaplike strong fluorescence in functionalized carbon nanodots. Theyobserved fluorescence quantum yield variations were due tochanges in the competing nonradiative processes from fractionto fraction (in size). This suggested a relatively uniform fluo-rescence radiative process in their samples. It has been furtherargued that the origin of fluorescence is associated with passivatedsurface defects of carbon nanoparticles [14]. The luminescence phe-nomenon of carbonaceous nanoparticles is not clearly understood

Fig. 2. (A) UV–vis absorption spectra and (B) fluorescence spectra of NCNP dispersedin water. In Fig. 2B, a, b and c represent the emission spectrum under UV, blue andgreen excitations, respectively. Inset in Fig. 2A shows the fluorescence emissionspectra at 300 nm excitation wavelength.

38 P. Kumar et al. / Colloids and Surfaces B: Biointerfaces 91 (2012) 34– 40

Table 2Typical physical attributes of carbon soot nanoparticles observed in our sample deduced from XRD data.

Peak position (2 �) Area (%) h k l plane Crystal structure Crystallite size (nm)

24.57 59.3 008 Graphitic carbon [30] 2.042.07 3.06 100 Cliftonite carbon [31] 2843.62 15.3 002 Carbon [32] 1349.12 5.60 110 Carbon [33] 19

sefltiw(npl

afeaettwpflc

tthaitt

strbw

Fe

50.97 16.6 102

ites alike what is observed in silicon nanocrystals [40,41]. Brunot al. [42] reported fluorescent carbon nanoparticles in premixedames with smaller particles (d = 1–1.5 nm) mostly fluorescing inhe UV (� ≈ 300–470 nm), and larger particles (d > 2 nm) fluoresc-ng in the visible region (� ≈ 490–580 nm). In the present study,

e observed clear emission peak at low excitation wavelength300 nm) but at higher excitation wavelength such behavior wasot discernible. Thus, the data in Fig. 2B may be arising due to theresence of fewer emissive trap states at higher excitation wave-

ength.The UV–vis absorption spectra show that there is significant

bsorption in visible region of spectrum and this gives possibilityor the nanoparticles to be fluorescent. For 405 nm excitation clearmission peaks were seen at 475 and 525 nm. For 488 nm excitationn emission peak was observed at 525 nm. However for the 543 nmxcitation, only an emission edge was observed. The novelty ofhese particles is that they are fluorescent without any surface func-ionalization. The solubility of these carbon nanoparticles in wateras poor which did not allow the solute concentration to be knownrecisely. So the exact quantum yield could not be quantified. Theuorescent properties of these were used for fluorescence-basedell imaging which will be discussed later.

Time resolved fluorescence decay measurement of the nanopar-icles dispersed in water is shown in Fig. 3. The decay was fitted tohe three component decay model as given in Eq. (2) with (�2 = 1.1)aving lifetimes �1 = 0.81 ns, �2 = 4.29 ns and �3 = 20.70 ns with rel-tive amplitudes a1 = 0.58, a2 = 0.33 and a3 = 0.08, respectively. Thisndicates that the multiple luminescence species were present inhe preparations. The mean lifetime time calculated by Eq. (3) wasypically 3.54 ns.

Wang et al. [39] successfully established correlation betweenize of carbon nanodots, fluorescence quantum yield and life-ime through a series of investigations. The estimated life-time

anged between ≈2 and 5 ns. CNP (diameter ≈3–4 nm) preparedy microwave pyrolysis have shown luminescence life-time ≈9 nshich was attributed to the radiative recombination nature of

ig. 3. Time resolved fluorescence decay curve of NCNP dispersed in water at 375 nmxcitation and 430 nm emission wavelengths.

Graphitic carbon [34] 10

excitations [43]. Zhu et al. [43] further argued that there must bequantum confinement of emissive energy traps on the particle sur-face that allow red shift with increase in particle size, similar tothose in quantum dots. Carbon dots with passivated surfaces haveshown life times in few nanoseconds region [44,45]. The lifetimeestimated for our particles in water was comparable to the valuesreported by Wang et al. [39] and Zhu et al. [43].

3.3. Cell imaging

Confocal laser scanning microscopy is amongst the most widelyused technique for high resolution and fluorescence-based cellimaging. Fluorescent CNP is an ideal cell-imaging probe with non-toxic characteristics [46]. Interestingly, we found that fluorescentNCNP were internalized by HEK-297 cells and were also adsorbedon the cell membrane. There was no fluorescence observed in thecontrol sample where as cells treated with NCNP (1 mg/ml) becomeblue, green, and red under UV, blue, and green excitation respec-tively as shown in Fig. 4. This clearly indicated penetration of NCNPinto the cells. Cell membranes were seen to be damaged at higherdose of nanopraticle treatment (2 mg/ml) which may be due to theproduction of reactive oxygen species [47,48].

The comparative imaging of NCNP was observed betweennormal cell lines HEK-297 and cancerous cell lines (MCF-7 andH-1299). It was found that the adsorption of NCNP was more inMCF-7 and H-1299 cells as compared to HEK-297 cells treatedwith same dose of particles (see Supporting Information, Figs. S2and S3). The difference in internalization of NCNP between can-cer and normal cells could be attributed to the different degree ofcell surface charge, thin cell membrane and higher turnover rate ofcancer cells than normal cells [49]. This study demonstrated thatthe fluorescent property of these NCNP can be used effectively forfluorescence-based cell imaging applications.

3.4. Cell uptake

For cell uptake and cytotoxicity studies, NCNP were added incell according to protocol as described in experimental section.The bright field image clearly shows aggregates of NCNP post 4 hof incubation (see Supporting Information, Fig. S4). The aggregateswere large in size and hydrophobic in nature which may be thereason of its adsorption on the cell membrane. Since the cellswere imaged post-incubation with low concentration (1 mg/ml) ofNCNP there was uptake by 80% of the HEK-297 cells. The highestfluorescence intensity was achieved at the excitation wavelengthof 450 nm, which shows an emission maximum at 520 nm, verysimilar to Fluorescein isothiocyanate. This was consistent to ourconfocal studies where we observed less uptake of NCNP by HEK-297 cells. It may be attributed to the hydrophobicity of the NCNPenhancing the adsorption of the cells after prolonged incubation.Therefore, it was imperative to understand the cytotoxic effects

of the nanoparticales. The cytotoxicity of these NCNP in HEK-297cells observed was 7% in 24 h and increased up to 20% in 48 h for1 mg/ml NCNP treatment. Detail cytotoxicity studies in cancerousand Leukemia cell lines are discussed in the next section.

P. Kumar et al. / Colloids and Surfaces B: Biointerfaces 91 (2012) 34– 40 39

Fig. 4. Fluorescent NCNP based cell imaging of HEK-297 cells. The first, second, third, and fourth columns represent the images in bright field, UV, blue and green excitation,respectively. First row images correspond to the control experiment where no NCNP was used and second row images correspond to the HEK-297 cells treated with NCNP.The light color in control sample is due to the well-known auto-fluorescence of cells. (For interpretation of the references to color in this figure legend, the reader is referredto the web version of the article.)

0

20

40

60

80

100

48 hr48 hr48 hr 24 hr24 hr

% C

ytot

oxic

ity

H-129 9 MCF-7

24 hr

MTT Assay of NCNP s

A

0

20

40

60

80

100

0.1mg/ml1mg/ml2mg/ ml

B

MTT Assay o f NCNP s

% C

ytot

oxic

ty

48 hr48 hr 48 hr24 hr24 hr

0.1mg/ml1mg/ml2mg/ ml

24 hr

HL-60 K-56 2

F showc . Perce

3

phiaIaTsocAo

ig. 5. Dose dependent cytotoxicity of NCNP after 24 h and 48 h incubation. Fig. 5Aell lines (HL-60 and K-562). Data is expressed as mean ± S.D. of three experiments

.5. Cytotoxicity assay

The MTT assay is widely used for its fast, inexpensive, and sim-le procedure for screening viability in large number of samples;owever, the assay is subject to variability and does not discrim-

nate between the routes of cell death. Leukemia cell lines, HL-60nd K-562 showed ∼70% cytotoxicity at a dose of 1 mg/ml (Fig. 5B).t is evident that the cytotoxicity is not only dose dependent butlso time dependent as increased toxicity was observed after 48 h.his is an interesting observation that haemopoeitic cells are moreusceptible to carbon nanoparticles. We also tested the efficacy

f the NCNP in p53 null H-1299, and p53 mutated breast cancerell line MCF-7 which clearly gave us different results (Fig. 5A).round 40% cytotoxicity was observed at a higher dose [∼2 mg/ml]f NCNP treatment. But, the intact cell morphology can be

s in cancerous cells lines (H-1299 and MCF-7 cells), and Fig. 5B shows in Leukemicnt cytotoxicity is expressed relative to controls.

correlated with no loss of membrane integrity and reduced cyto-toxicity. The hydrophobicity of the aggregated particles may beresponsible for enhanced uptake leading to cytotoxicity. The aboveobservations warrant further investigation into the mode of celldeath.

4. Conclusions

In conclusion, we have reported the synthesis as well asstructural and optical characterization of fluorescent NCNP withminimum crystallite size observed ≈2 nm by XRD and Raman mea-

surements. Our preparation method is very easy, economical andnon-chemical based, and can be used for milligram level synthesisof fluorescent NCNP. These particles could be synthesized follow-ing a simple protocol and without the need for sophisticated and

4 faces B

eootvtdssotfcehco

A

fALt

A

t

R

[[

[[

[

[

[[

[

[

[

[[[[[[[[[[

[[[[

[[[[[

[[

[

[

[[

[

[47] G. Jia, H. Wang, L. Yan, X. Wang, R. Pei, T. Yan, Y. Zhao, X. Guo, Environ. Sci.Technol. 39 (2005) 1378–1383.

0 P. Kumar et al. / Colloids and Sur

xpensive instruments. These particles exhibited considerable flu-rescence with average emission life-time of 3.54 ns. The originf fluorescence from NCNP even when these were not subjectedo acid or any other functionalization treatment (surface passi-ated) is not clearly understood. One possibility may be is due tohe presence of solvent induced different emissive trap states. Oxi-ation of nanoparticles by dispersing in de-ionized water inducesurface groups on CNP and this may be responsible for light emis-ion. Another possibility of origin of fluorescence is the presencef the polycyclic aromatic compounds. It was clearly observed thathese nanoparticles entered inside cells though there was no sur-ace functionalization and some of these were seen adsorbed toell boundary. Bio-imaging application of these particles has beenxplicitly shown for three types of cells with excellent results. Weave further elucidated the effects on cells by directly combiningell viability with imaging. We envision that further developmentf this novel material will lead to enhanced biological applications.

cknowledgements

PKG thanks University Grants Commission, Government of Indiaor providing a fellowship. The authors would like to thanksdvanced Research Instruments Facility (AIRF) centre, JNU for XRD,aser Scanning Confocal microscopy and TRFS facility and alsohanks to Fouran Singh, IUAC for help in the Raman measurement.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.colsurfb.2011.10.034.

eferences

[1] H. Yanhong, O.A. Shenderova, H. Zushou, C.W. Padgett, D.W. Brenner, Rep. Prog.Phys. 69 (2006) 1847–1895.

[2] P.C. Ke, R. Qiao, J. Phys. Condens. Matter 19 (2007) 373101–373125.[3] G.X. Chen, M.H. Hong, T.C. Chong, H.I. Elim, G.H. Ma, W. Ji, J. App. Phys. 95 (2004)

1455–1459.[4] H. Hu, B. Zhao, M.E. Itkis, R.C. Haddon, J. Phys. Chem. B 107 (2003) 13838–13842.[5] A. Evelyn, S. Mannick, P.A. Sermon, Nano Lett. 3 (2003) 63–64.[6] R. Bandyopadhyay, E. Nativ-Roth, O. Regev, R. Yerushalmi-Rozen, Nano Lett. 2

(2002) 25–28.[7] J. Gavillet, A. Loiseau, C. Journet, F. Willaime, F. Ducastelle, J.C. Charlier, Phys.

Rev. Lett. 87 (2001) 275504–275507.[8] S. Iijima, Nature 354 (1991) 56–58.

[9] C. Journet, W.K. Maser, P. Bernier, A. Loiseau, M.L. de la Chapelle, S. Lafrant, P.

Deniard, R. Leek, J.E. Fischer, Nature 388 (1997) 756–758.10] Y. Zhang, S. Iijima, Appl. Phys. Lett. 75 (1999) 3087–3089.11] A.M. Cassell, J.A. Raymakers, J. Kong, H.J. Dai, J. Phys. Chem. B 103 (1999)

6484–6492.

[

[

: Biointerfaces 91 (2012) 34– 40

12] H. Liu, T. Ye, C. Mao, Angew. Chem. Int. Ed. 46 (2007) 6473–6475.13] B. Mohanty, A.K. Verma, P. Claesson, H.B. Bohidar, Nanotechnology 18 (2007)

445102–445110.14] Y.P. Sun, B. Zhou, Y. Lin, W. Wang, K.A.S. Fernando, P. Pathak, M.J. Meziani, B.A.

Harruff, X. Wang, H. Wang, P.G. Luo, H. Yang, M.E. Kose, B. Chen, L.M. Veca, S.Y.Xie, J. Am. Chem. Soc. 128 (2006) 7756–7757.

15] A. Gruber, A. Dräbenstedt, C. Tietz, L. Fleury, J. Wrachtrup, C. Von Borczyskowski,Science 276 (1997) 2012–2014.

16] J.E. Riggs, Z. Guo, D.L. Carroll, Y.P. Sun, J. Am. Chem. Soc. 122 (2000) 5879–5880.17] J. Zhou, C. Booker, R. Li, X. Zhou, T.K. Sham, X. Sun, Z. Ding, J. Am. Chem. Soc.

129 (2007) 744–745.18] P. Kumar, S. Karmakar, H.B. Bohidar, J. Phys. Chem. C 112 (2008)

15113–15121.19] H.B. Bohidar, P. Kumar, Non-functionalized carbon nanoparticles having

fluorescence characteristics, method of preparation thereof, and their useas bioimaging and solvent sensing agents, Indian Patent Application no.2184/DEL/2010.

20] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd ed., Kluwer Aca-demic/Plenum, New York, 2002.

21] P. Kumar, H.B. Bohidar (communicated).22] O.V. Yazyev, A. Pasquarello, Phys. Rev. Lett. 100 (2008) 156102–156104.23] C.M. Sung, M.F. Tai, Int. J. Refract. Met. Hard Mater. 15 (1997) 237–256.24] S.N. Baker, G.A. Baker, Angew. Chem. Int. Ed. 49 (2010) 6726–6744.25] S.C. Ray, A. Saha, N.R. Jana, R. Sarkar, J. Phys. Chem. C 113 (2009) 18546–18551.26] X. Wang, K. Qu, B. Xu, J. Ren, X. Qu, Nano Res. 4 (2011) 908–920.27] S.D. Stasio, Carbon 39 (2001) 109–118.28] L.E. Murr, P.A. Guerrero, Atmos. Sci. Lett. 7 (2006) 93–95.29] E.F. Kaelble, Handbook of X-Rays, McGraw-Hill, New York, 1976.30] D.E. Nixon, G.S. Parry, A.R. Ubbelohde, Proc. R. Soc. London, Ser. A 291 (1996)

324–339.31] G.P. Vdovykin, Space Sci. Rev. 14 (1973) 758–831.32] C. Yeh, Z.W. Lu, S. Froyen, A. Zunger, Phys. Rev. B 46 (1992) 10086–10097.33] H. Lipson, A.R. Stokes, Nature (London) 149 (1942) 328.34] M. Read, Belltelephone Laboratories, Murry Hill, NJ, USA, 1960, Private Com-

munication.35] A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 14095–14107.36] D.S. Knight, W.B. White, J. Mater. Res. 4 (1989) 385–393.37] Y. Kimura, T. Sato, C. Kaito, Carbon 42 (2004) 33–38.38] W.A. De Heer, D. Ugarte, Chem. Phys. Lett. 207 (1993) 480–486.39] X. Wang, L. Cao, S.T. Yang, F. Lu, M.J. Meziani, L. Tian, K.W. Sun, M.A. Bloodgood,

Y.P. Sun, Angew. Chem. Int. Ed. 49 (2010) 5310–5314.40] W.L. Wilson, P.F. Szajowski, L.E. Brus, Science 262 (1993) 1242–1244.41] Z. Ding, B.M. Quinn, S.K. Haram, P.E. Pell, B.A. Korgel, A.J. Bard, Science 296

(2002) 1293–1297.42] A. Bruno, C.D. Lisio, P. Minutolo, A. D’Alessio, Combust. Flame 151 (2007)

472–481.43] H. Zhu, X. Wang, Y. Li, Z. Wang, F. Yang, X. Yang, Chem. Commun. 34 (2009)

5118–5120.44] H. Goncalves, J.C.G. Esteves da Silva, J. Fluoresc. 20 (2010) 1023–1028.45] X.J. Mao, H.Z. Zheng, Y.J. Long, J. Du, J.Y. Hao, L.L. Wang, D.B. Zhou, Spectrochim.

Acta A: Mol. Biomol. Spectrosc. 75 (2010) 553–557.46] S.T. Yang, L. Cao, P.G. Luo, F. Lu, X. Wang, H. Wang, M.J. Meziani, Y. Liu, G. Qi,

Y.P. Sun, J. Am. Chem. Soc. 131 (2009) 11308–11309.

48] T. Xia, M. Kovochich, J. Brant, M. Hotze, J. Sempf, T. Oberley, C. Sioutas, J.I. Yeh,M.R. Wiesner, A.E. Nel, Nano Lett. 6 (2006) 1794–1807.

49] N.R. Aiken, R.J. Gillies, Anticancer Res. 16 (1996) 1393–1397.