lable at ScienceDirect

Vacuum 83 (2009) 412–415

Contents lists avai

Vacuum

journal homepage: www.elsevier .com/locate/vacuum

Nonlinear optical properties of Ag nanocluster composite fabricatedby 200 keV negative ion implantation

Y.H. Wang a,*, S.J. Peng a, J.D. Lu a, R.W. Wang a, Y.G. Cheng b, Y.L. Mao b

a Department of Applied Physics, Wuhan University of Science and Technology, Wuhan 430081, Chinab Department of Physics, Henan University, Kaifeng 475001, China

a r t i c l e i n f o

Article history:Received 6 March 2008Received in revised form 4 May 2008Accepted 5 May 2008

Keywords:Ion implantationNanoclustersNonlinear opticsZ-scan

PACS:61.46.þW61.72.Ww42.65.-K

* Corresponding author.E-mail address: wyh61@163.com (Y.H. Wang).

0042-207X/$ – see front matter Published by Elseviedoi:10.1016/j.vacuum.2008.05.002

a b s t r a c t

Metal nanocluster composite glass prepared by 200 keV Ag ions’ implantation into silica with dose of2� 1017 ions/cm2 has been studied. The formation of sandwiched nanocluster-nanovoid-nanoclusterstructures has been evidenced by in situ transmission electron microscopy experiment (TEM). Fastnonlinear optical refraction and nonlinear optical absorption coefficients were measured at 532 nm and1064 nm of wavelength by the Z-scan technique. The third-order nonlinear susceptibility c(3) of this kindof sample was determined to be 4.0� 10�8 esu at 532 nm and 9.0� 10�8 esu at 1064 nm, respectively.

Published by Elsevier Ltd.

1. Introduction

Metal nanoclusters possess interesting linear and nonlinearoptical properties. Recently, there has been an increasing interest inthe third-order nonlinear susceptibility and the photorefractiveeffect of noble metal clusters embedded in dielectric matrices [1–4].Third-order nonlinearities of metal/dielectric composite materialswere influenced by the type and size of the embedded metalclusters, by the dielectric constant, thermal conductivity and heatcapacity of the dielectric matrices [1–7]. The most conspicuousmanifestation of confinement in optical properties of metal nano-cluster composite glasses (MNCGs) is the appearance of the surfaceplasmon resonance (SPR) that strongly enhances their linear andnonlinear responses around SPR wavelength [8–10]. Amongst thenanoclusters studied by earlier researchers, nonlinear absorptionand nonlinear refraction were found to be potential used fornonlinear optical devices in Ag and Ag containing nanomaterials[11–16].

Recently, many workers have observed the core/shell nano-clusters formed by single or double-element ion implantation [17–23]. Therefore, application aspects of the material are most relevantto the optical properties’ change versus the nanocluster structure.

r Ltd.

In this paper, we find that nanovoid-containing Ag nanoclusters witha set of satellite nanoclusters are formed when the dose reaches1�1017 ions/cm2, which changes to sandwiched nanocluster–nanovoid–nanocluster structure nanoclusters after increasing thedose to 2�1017 ions/cm2. We focused our interest on the study ofthe linear and nonlinear optical properties of this new structurenanoclusters. MNCGs were prepared by Agþ implantation into silica.Nonlinear optical properties were measured at the wavelength of532 nm and 1064 nm.

2. Experiment

Silica slides were implanted at room temperature by silver ionsat 200 keV. The current density of ion implantation was lower than2.5 mA/cm2. Optical absorption spectra were recorded at roomtemperature using a UV–vis dual-beam spectrophotometer withwavelengths from 1500 nm to 200 nm. Transmission electron mi-croscopy (TEM) observations were carried out with a JEOL JEM 2010(HT) microscope operated at 200 kV. TEM bright field images wereused to determine the size distribution and shape of nanoclusters.

The measurements of third-order optical nonlinearities of thesesamples were carried out using the standard Z-scan method. Theexcitation source is a mode-locked Nd:YAG laser (PY61-10, Con-tinnum), with a pulse duration of 38 ps and a repetition frequencyof 10 Hz. Wavelength (1064 nm) and doubled frequency (532 nm)

Fig. 2. SAED pattern for the sample implanted by 200 keV, 2�1017 Agþ ions/cm2.

Y.H. Wang et al. / Vacuum 83 (2009) 412–415 413

are used for excitation in the experiment. The detector is a dual-channels energy meter (EPM2000). With a converging lens off¼ 260 mm, the radiuses of the Gaussian beam spot at focal waist60 were about 45 mm and 26 mm for 1064 nm and 532 nm,respectively. In the Z-scan test, the sample was moved step by stepalong the propagation direction of the Gaussian beam under thecontrol of a PC. Meanwhile, a detector monitored the transmittedlaser power and the signals were sent back to the computer andrecorded. Nonlinear refraction and nonlinear absorption wereperformed by both open- and closed-aperture Z-scans of a series ofthe samples at room temperature.

3. Results and discussion

The cross-sectional and planar TEM images for the sampleimplanted by 2�1017 Agþ ions/cm2 are shown in Fig. 1(a) and (b).As can be seen from the image, the particle size distribution is notuniform. The size of nanoclusters varies from 5 nm to 50 nm.Interestingly, the largest nanoclusters show a sandwiched nano-cluster–nanovoid–nanocluster structure, which are indicated witharrows. Such sandwiched structure was predicted by Zhu [24]. Ourprevious study showed it was observed nanoclusters with brightcenters in Cu nanoclusters [5]. Spherical metal nanoclusters areformed for low ions’ implantation. Hollow metal nanoshells(nanovoids) were formed with increased ion doses. Appearance ofnanoclusters in nanovoids is possibly related to the further processof metal implantation. In Ag ions’ implantation process, nanovoidsalso could be formed [25]. With further implantation of Ag ions toAg nanovoids, some Ag atoms from the inner walls, which arebombarded into the nanovoids by subsequently implanted self-ions, were captured by the nanovoids and aggregated into Agnanoclusters. These small nanoclusters stay in the nanovoids byelectrostatic interaction with the inner walls of nanovoids. TheSAED pattern taken from the planar view, Fig. 2, also shows thesingle fcc phase of Ag. Therefore, the formation of new phases canbe excluded and the small nanoclusters in the nanovoids shouldalso be Ag nanoclusters. While the smallest ones are single-phase.The comparative size distributions of Ag nanoclusters are shown inFig. 3. The average size of nanoclusters in this sample is 13 nm.

Fig. 4 is the optical absorption spectra of the 2�1017 ions/cm2

Agþ implanted samples. For comparison, the doses of 1�1016 ions/cm2, 5�1016 ions/cm2, 1�1017 ions/cm2 Agþ implanted samplesare also shown in this figure. The spectra range from 200 nm to

Fig. 1. Cross-sectional (a) and planar (b) TEM images for the sample implanted by200 keV, 2�1017 Agþ ions/cm2.

1500 nm. The surface plasmon resonance (SPR) absorption peaks ofAg nanoclusters around 410 nm are observed. As can be seen fromthe figure, the implant dose and size distribution of Ag nanoclustershad important influence on the whole absorption spectra.

The nonlinear absorption in the sample can be described by b,which includes saturated absorption (SA) and reversed saturatedabsorption (RSA) [26]. The absorption of the sample is expressed bya¼ a0þ bI, where a0 is the linear absorption coefficient of thesample and I is the intensity of the laser. The third-order nonlinearabsorption and refraction are investigated by Z-scan techniques[27], which are simple and sensitive experimental techniques forthe study of nonlinear optical properties and allow determining thesign of the nonlinear refractive and absorption indices. The open-and closed-aperture Z-scan curves are theoretically fitted by [27]:

TðzÞ ¼XN

m¼0

½�q0ðzÞ�m�1þ x2

�mðmþ 1Þ3=2m � 0 (1)

TðzÞ ¼ 1þ 4DF0x�x2 þ 9

��x2 þ 1

� (2)

0 5 10 15 20 25 30 35 40 45 500

5

10

15

20

25

30

35

40

45

Nu

mb

er o

f n

an

oclu

ster

Diameter/nm

Fig. 3. Comparative size distribution profiles of 2�1017 Agþ ions/cm2 nanoclusters insilica samples.

Fig. 4. Optical absorption spectra of the Ag implanted samples to dose of 1�1016 ions/cm2, 5�1016 ions/cm2, 1�1017 ions/cm2 and 2�1017 ions/cm2.

-30 -20 -10 0 10 20 300.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

No

rm

alized

T

ran

sm

ittan

ce

Z (mm)

ExperimentTheoritical fit

Fig. 6. Z-scan experiment results for 1064 nm normalized close-aperture. Solid line:theoretical curve (2�1017 Agþ ions/cm2).

Y.H. Wang et al. / Vacuum 83 (2009) 412–415414

where x¼ z/z0, T is the normalized transmittance and z is the dis-tance along the lens axis in the far field. The nonlinear absorptioncoefficient b can be obtained from q0¼ bI0Leff, where I0 is the in-tensity of the laser beam at the focus (z¼ 0), Leff is the effectivethickness of the sample, which can be calculated from the realthickness L and the linear absorption coefficient a0, in the form ofLeff ¼ ½1� expð�a0LÞ�=a0. The nonlinear refractive index is calcu-lated by DF0 ¼ ð2p=lÞgI0Leff , where 2p/l is the wave vector of theincident laser.

Normalized open-aperture Z-scan of sample is displayed inFig. 5(a). The open-aperture measurement shows an obvious en-hanced transmittance near the focus, occurring due to the satura-tion of absorption. This reveals negative nonlinear absorptioncoefficient and indicates that the local field and collective oscilla-tion of free electrons on the surface of nanocluster decreased withthe increase of laser intensity. The nonlinear refraction wasobtained by dividing the closed-aperture data by the open-aperturedata [27,28]. In Fig. 5(b), the peak-valley configuration indicates thenegative sign of the nonlinear refractive index (n2< 0).

Third-order nonlinear optical properties of this kind of sampleof 1064 nm are shown in Fig. 6, the open-aperture Z-scan shows nononlinear signal, which indicates that the sample has no nonlinearabsorption at 1064 nm. A self-defocusing refraction is also foundfrom the peak-valley curve of closed-aperture data. The nonlinear

-20 -10 0 10 200.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5a

No

rm

alize

d T

ran

sm

ittan

ce

Z (mm)

ExperimentTheoritical fit

Fig. 5. Z-scan experiment results for 532 nm normalized open-aperture (a) and

properties of the bare silica substrates were measured and nodetectable change of the transmitted intensity under same Z-scanconditions was observed. We notice that when the laser peakintensity is larger than 15 GW/cm2, there is a probability for thehigh absorbing materials to be damaged at the tested point due toaccumulative heating, which will produce an ablation hole. Theclosed-aperture Z-scan curve of the ablation hole is symmetric andhas a peak-valley pattern similar to that of some materials withnegative nonlinear refraction [3]. So the peak intensity of 0.9 GW/cm2 was selected for the samples at 532 nm and 0.38 GW/cm2 at1064 nm. In our experiment, the asymmetric curve changes littlewhen repeated at the same point. Thus we suggest that therewas no formation of the ablation hole in these samples during theZ-scan.

In our experiments, Leff (nm) for sample is 115 nm. The solidcurve in Fig. 5(a) is fitted using Eq. (1) with the experimentparameters and the nonlinear absorption coefficient obtained isb¼�70 cm/GW for 532 nm. Fitting the Z-scan data of the closed-aperture with Eq. (2), we get values of g z�3.0�10�11 cm2/W for532 nm and g z�1.7�10�10 cm2/W for 1064 nm. The absolutevalues of third-order nonlinear susceptibility c(3) for Agþ implantedsamples are calculated from [27,28]:

DTp�v ¼ 0:406ð1� SÞ0:25jDf0j (3)

ExperimentTheoritical fit

-20 -10 0 10 200.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

b

No

rm

alized

T

ran

sm

ittan

ce

Z (mm)

the divided result (b). Solid line: theoretical curve (2�1017 Agþ ions/cm2).

Y.H. Wang et al. / Vacuum 83 (2009) 412–415 415

Recð3Þ ¼ 2n2030cg (4)

Imcð3Þ ¼ ðl=2pÞn2030cb (5)

cð3Þ ¼��

cð3ÞRe

�2þ�

cð3ÞIm

�2�1=2

(6)

Thus, we obtain the absolute value of c(3), 4.0�10�8 esu for532 nm and 1.8� 10�7 esu for 1064 nm.

The third-order nonlinearity observed is not induced by thermaleffect within the pulse temporal width. On one hand, the electronicnonlinearities arise very rapidly (within the 38 ps pulse duration).Refractive index changes due to thermal nonlinearities arise due todensity changes in the materials propagating with acoustic wavespeed caused by heating. If we estimate it to be on the order of3�10�3 m/s, the time to propagate a distance equal to the beamradius at focus is about 9 ns at 532 nm excitation and 15 ns at1064 nm excitation, about 230 times and 380 times longer than thepulse width, respectively. On the other hand, thermal heatinginduced by a single laser pulse persists over some characteristictime tc. As a result, when the time interval between consecutivelaser pulses is shorter than tc, the thermal effect increases. It isa common assumption that Z-scan measurements should be madewith repetition rate of few hertz in order to extract a nonlinearrefractive index influenced by only electronic effects. The time scaleof this cumulative process is given by tc ¼ 62

0=4D, where D is thethermal diffusion coefficient of the materials. Generally, the valueof D ranges from 1�10�7 m2/s to 6�10�7 m2/s. The magnitude ofthe calculated tc is within 10�3 s, which is much smaller than thetime interval between consecutive laser pulses, 0.1 s, used in ourexperiment [29].

Longer pulse duration of the excitation wavelength would leadto combined mechanisms including both optical processes of quickresponse time and slower ones. Under the picosecond pulse dura-tion and low repetition frequency laser in this experiment, only theelectronic effect with the response time in the order of picosecondin the semiconductor contribute to the nonlinear polarization.Excitation frequency is another cause for the variance. Variouswavelengths may induce different nonlinear optical transitions.Moreover, for diverse preparation techniques, Agþ implantednanostructured samples differ in size and shape as well as surfacestructure films, so do in the nonlinear optical behaviors.

Optical Kerr effect (in the case of picosecond and femtosecondpulses) and thermal effect (in the case of nanosecond pulses) arethe dominant processes responsible for the nonlinear refraction insuch structures. Compare the NLO responses with those measuredby other researchers for other MNCs. Yang et al. [13] recorded Z-scan curves of Ag/BaTiO3 nanocomposite films using a laser pulsewidth of 10 ns, thus in a duration time regime deducing cumulativeheating of the sample. Ganeev et al. [14] investigated with the Z-scan method the nonlinear optical parameters of Ag colloidal so-lutions at very low values of metal volume fraction; they observeda sign change at 1064 nm by varying the nanoparticle aggregationstage. The authors did not give an explanation for these results.However, for the determination of the nonlinear optical coefficientsthey used very large intensities (400 GW/cm2). Therefore, impor-tant contributions to n2 due to slow thermal mechanisms are rea-sonably possible. Takeda et al. [15] measured nonlinear absorptioncoefficient of Ag nanoparticles fabricated by 60 keV negative ionimplantation into silica. Our results show about 10–100 timeslarger than theirs. We think that was the effect of nanovoids on theNLO responses. Many small nanoclusters stay in the nanovoids by

electrostatic interaction with the inner walls of nanovoids [25]. Thisinfluenced the election distribution of nanovoids’ surface. The localfield of the sandwiched structure nanoclusters as a whole willenhance. Large third-order nonlinear optical properties of ourexperiments may attribute to the enhancement of the large localfield and the strong resonant absorption of surface plasma of newstructure nanoclusters. Thus the formation of sandwiched nano-cluster–nanovoid–nanocluster structure nanoclusters, which differin size and shape, may play an important role for nonlinear opticaleffect.

4. Conclusion

In summary, the formation of sandwiched nanocluster–nano-void–nanocluster structure in silica has been formed by Agþ ionimplantation. The nonlinear optical properties were investigated bythe Z-scan technique and exhibit exceptionally large third-ordernonlinearity. The c(3) measured at 532 nm manifests a real part of3.0�10�8 esu and an imaginary part of 2.6�10�8 esu. During the1064 nm excitation, the sample has no nonlinear absorption andc(3) is 1.8� 10�7 esu which all come from nonlinear refractioncontribution. The results indicated that this new structure metalnanoclusters gives a new circumstance in the study of the nonlinearoptical response. The further studies are in progress.

References

[1] Battaglin G, Calvelli P, Cattaruzza E, Gonella F, Polloni R, Mattei G, Mazzoldi P.Appl Phys Lett 2001;78:3953.

[2] Takeda Y, Zhao JP, Lee CG, Gritsyna VT, Kishimoto N. Nucl Instrum Methods B2000;116:877.

[3] Battaglin G, Calvelli P, Cattaruzza E, Polloni R, Borsella E, Cesca T, Gonella F,Mazzoldi P. J Opt Soc Am B 2000;17:213.

[4] Takeda Y, Lee CG, Kishimoto N. Nucl Instrum Methods B 2002;191:422.[5] Wang YH, Ren F, Wang QQ, Chen DJ, Fu DJ, Jiang CZ. Phys Lett A 2006;357:364.[6] Haglund Jr RF, Yang L, Magruder RH, Wittig JE, Becker K, Zuhr RA. Opt Let

1993;18:373.[7] Wang YH, Jiang CZ, Ren F, Wang QQ, Chen DJ, Fu DJ. Physica E 2006;33:244.[8] Cattaruzza E, Battaglin G, Calvelli P, Gonella F, Polloni R, Mattei G, et al.

Compos Sci Technol 2003;63:1203.[9] Tanahashi I, Inouye H, Tanaka K, Mito A. Jpn J Appl Phys 1999;38:5079.

[10] Wang P, Lu Y, Tang L, Zhang J, Ming H, Xie J, Ho FH, Chang HH, Lin HY, Tsai DP.Opt Commun 2004;229:425.

[11] Ganeev RA, Ryasnyansky AI, Stepanov AL, Marques C, da Silva RC, Alves E. OptCommun 2005;253:205.

[12] Qu Shiliang, Zhang Yawen, Li Huajun, Qiu Jianrong, Zhu Congshan. Opt Mater2006;28:259.

[13] Yang G, Wang W, Zhou Y, Lu H, Chen Z. Appl Phys Lett 2002;81:3969.[14] Ganeev RA, Ryasnyansky AI, Hamalov ShR, Kodirov MK, Usmanov T. J Phys D

Phys 2001;34:1602.[15] Takeda Y, Plaksin OA, Lu J, Kishimoto N. Vacuum 2006;80:776.[16] Cattaruzza E, Battaglin G, Gonella F, Mattei G, Mazzoldi P, Polloni R,

Scremin BF. Appl Surf Sci 2005;247:390.[17] de Julian Fernandez C, Tagliente MA, Mattei G, Sada C, Bello V, Maurizio C,

Battaglin G, Sangregorio C, Gatteschi D, Tapfer L, Mazzoldi P. Nucl InstrumMethods B 2004;216:245.

[18] Battaglin G, Cattaruzza E, De Marchi G, Gonella F, Mattei G, Maurizio C,Mazzoldi P, Parolin M, Sada C, Calliari I. Nucl Instrum Methods B 2002;191:392.

[19] de Julian Fernandez C, Sangregorio C, Mattei G, Maurizio C, Battaglin G,Gonella F, Lascialfri A, Lo Russo S, Gatteschi D, Mazzoldi P, Gonzalez JM,D’Acapito F. Nucl Instrum Methods B 2001;175–177:479.

[20] Meldrum A, Boatner LA, White CW. Nucl Instrum Methods B 2001;178:7.[21] Stepanov AL, Popok VN, Hole DE, Khaibullin IB. Appl Phys A 2002;74:441.[22] Ren F, Guo LP, Shi Y, Chen DL, Wu ZY, Jiang CZ. J Phys D Appl Phys 2006;39:

488–91.[23] Xiao XH, Ren F, Jiang CZ, Wang JB, Shi Y. Solid State Commun 2006;137:362.[24] Zhu XF. J Phys Condens Matter 2003;15:L253.[25] Ren F, Jiang CZ, Liu C, Wang Jb. Phys Rev L 2006;97:165501.[26] Couris S, Koudoumas E, Rutht AA, Leach S. J Phys B At Mol Opt Phys 1995;28:

4537.[27] Sheik-Bahae M, Said AA, Wei TH, Hagan DJ, Van Stryland EW. IEEE J

Quantum Electron 1990;26:760.[28] Sheik-Bahae M, Hagan DJ, Ban Stryland EW. Phys Rev Lett 1990;65:96.[29] He Tingchao, Cheng Yongguang, Du Yabing, Mo Yujun. Opt Commun 2007;

275:240.