COM

MU

DOI: 10.1002/adma.200801508 N

ICATIO

N

Spatially Localized Photoluminescence at 1.5 MicrometersWavelength in Direct Laser Written OpticalNanostructures**

By Sean Wong,* Oliver Kiowski, Manfred Kappes, Jorg K. N. Lindner,

Nirajan Mandal, Frank C. Peiris, Geoffrey A. Ozin, Michael Thiel,

Markus Braun, Martin Wegener, and Georg von Freymann*

Direct laser writing (DLW) has proven to be a very versatile

tool for the fabrication of complex and high-quality structures

in all three spatial dimensions with nanometer precision and

accuracy.[1] This method enables the demonstration of

complex architectures that are otherwise difficult or impossible

to fabricate via traditional photolithographic techniques.[2]

The benefits of this technique is most evident in the field of

nanophotonics, allowing for challenging 3D structures to be

written such as suspended waveguides,[3] Mach–Zehnder

interferometers,[4] photonic crystals (PC),[5,6] and complex

heterostructure geometries.[7,8]

However, despite the advantages of DLW to fabricate

intricate optical nanostructures, a suitable photoresist material

which is able to derive the full benefits of this technique remains

elusive. A suitable photoresist for many optical applications

should simultaneously possess a high-refractive index (n) and

room temperature photoluminescence (PL) at the telecom-

munication wavelength of 1.5mm, which ideally should be

[*] S. H. Wong, Dr. G. von Freymann, O. Kiowski, Prof. M. Kappes,M. Braun, Prof. M. WegenerInstitut fur NanotechnologieForschungszentrum Karlsruhe in der Helmholtz-GemeinschaftPostfach 3640, Karlsruhe 76021 (Germany)E-mail: sean.wong@int.fzk.de, freymann@int.fzk.de

Dr. J. K. N. LindnerInstitut fur Physik, Universitat AugsburgAugsburg 86135 (Germany)

Dr. F. C. Peiris, N. MandalDepartment of Physics, Kenyon CollegeGambier, OH 43022 (USA)

Prof. G. A. OzinChemistry Department, University of TorontoToronto, Ont., M5S 3H6 (Canada)

M. Thiel, M. Braun, Dr. G. von Freymann, Prof. M. WegenerInstitut fur Angewandte Physik, Universitat Karlsruhe (TH)Wolfgang-Gaede-Strabe 1, Karlsruhe 76131 (Germany)

[**] We acknowledge financial support provided by the DeutscheForschungsgemeinschaft (DFG) and the State of Baden-Wurttembergthrough the DFG-Center for Functional Nanostructures (CFN) withinsubproject A1.4. The research of G. v. F. is further supported through aDFG Emmy–Noether fellowship (DFG-Fr 1671/4-3). GAO isGovernment of Canada Research Chair in Materials Chemistry. Heis grateful to the Natural Sciences and Engineering Research Councilof Canada (NSERC) for financial support of this work. The Ph.D.education of S. H. W. and M. T. is further supported by the Karlsruheschool of optics and photonics (KSOP).

Adv. Mater. 2008, 20, 4097–4102 � 2008 WILEY-VCH Verlag G

spatially localizable. Such a photoresist would enable more

compact optically active devices to be reduced to practice.

The current approach to overcome the low-refractive index

commonly encountered in photopolymers starts with the

polymeric structure as a template and converting it in

subsequent steps into high index of refraction materials like

silicon via a double-inversion[9] or single-inversion schemes.[10]

The remaining and to date unsolved problem is the precise

placement of photoluminescent material within the nanos-

tructure. While recent photoresist development demonstrated

the inclusion or generation of photoluminescent species inside

the bulk photoresist,[11,12] they will inevitably be lost during the

following conversion steps. Subsequent infiltration of the

high-index structures with, e.g., quantum dot solutions, fails

the requirement of precise placement inside the high-index

material.[13] Therefore, despite the ability for each separate

technique to perform well on its own, their mutual incompat-

ibility makes a total integration of techniques very challenging.

The starting point of our solution to this material problem is

the As2S3 inorganic photoresist system.[14,15] As2S3 is an ideal

material for fabricating 3D PC structures due to its high n of

2.45. Problematic absorption of the hydroxyl group in the

near-infrared region, usually found in silica glasses, is absent in

As2S3-based glasses. Hence, structures made fromAs2S3-based

glasses can support broadband access to the entire telecom-

munication window from 1.3 to 1.5mm. In our previous works,

we have already shown that this material can be formed into a

photoresist that is suitable for 3D DLW and 3D PC structures

possessing a full photonic bandgap were demonstrated. As2S3is also an excellent host for rare-earth dopants because it

possesses low-phonon energies reducing the non-radiative

decay of the rare-earth excited states.[16] The rare-earth

element erbium is routinely used as a dopant for providing PL

at 1.5mm in various optical devices such as doped fiber

amplifier systems.[17] Various studies on erbium doping of

As2S3 have demonstrated that the high-refractive index of the

host glass can be retained while providing room temperature

PL.[18–21] Although, structuring of this material using conven-

tional semiconductor-based fabrication techniques has been

demonstrated,[18] complex 3D structures have not been

realized.

In this work, we demonstrate a gas-phase direct doping

(GPDD) technique enabling the synthesis of erbium-doped

mbH & Co. KGaA, Weinheim 4097

COM

MUNIC

ATIO

N

Figure 1. a) The experimental (black curve) and fitting (red curve) of theRBS spectra of an Er/As2S3 photoresist film. b) The refractive indexdispersion of the doped film in the as-deposited (dotted red curve) andthe photoexposed state (solid red curve). The results are compared toun-doped As2S3 film that has been photopolymerized (black curve).

4098

arsenic trisulfide (Er/As2S3) that is suitable as a photoresist for

3D DLW. This Er/As2S3 photoresist exhibits both a high n of

2.45 and PL at 1.5mm wavelength. Furthermore, we show that

the photoluminescent region contains a homogenous incor-

poration of erbium, and can be spatially limited to layers as thin

as 139 nm, in contrast to systems where the bulk material is

photoluminescent. The crucial point here, is doping the As2S3host with Er while maintaining its photosensitive properties. In

the GPDD technique, a high-purity Er metal source and an

As2S3 source are arranged in a high-vacuum chamber such that

the gas-phase species emanating from the source overlap

during the deposition process thus enabling direct doping.

Here, we use high-purity (99.99%) Er metal as the dopant

source, providing accurate control over the stochiometry as

well as enabling stringent control of the final materials purity.

Tight control of both parameters is essential for preserving the

optical properties of the host, such as maintaining the desired

PL behavior and minimizing absorption from undesired

impurities.[16,18,20]

We have employed Rutherford backscattering spectroscopy

(RBS) to characterize the composition, purity, and homo-

geneity of Er/As2S3 thin films. The results of the RBS analysis

of a representative Er/As2S3 film is shown in Figure 1a. In the

experimental spectrum, the signal of each element can be

clearly resolved in its own channel region. The signal beginning

at channel 950 is assigned to erbium, 850 to arsenic, and 650 to

sulfur. The signals beginning at channel 475 and 275 belong to

aluminum and oxygen, respectively, and their presence is solely

attributed to the presence of the sapphire substrate used for

this analysis. A fitting of this experimental spectrum was

carried out to allow for the quantitative analysis of this film.

The analysis obtained from the fitting indicates that the film is

homogenously doped as each element present has a constant

concentration throughout the thickness of the film. A typical

film has a stochiometry of Er0.4As39.4S60.2 corresponding to an

Er concentration of 1.35wt %. The Er/As2S3 films produced

using the GPDD technique maintain the stochiometry of As to

S near the optimum ratio of 2 to 3 of the host glass.

Furthermore, this analysis also indicates that there is no

detectable oxygen inside the as-deposited film. These RBS

results clearly demonstrate the ability for this technique to

produce highly homogenous films of controlled stochiometry.

Thin film ellipsometery was used to investigate the behavior

of n for a 1.35wt % Er-doped and an un-doped As2S3photoresist film. The result is shown in Figure 1b. The

measurements show that after photoexposure, the Er/As2S3films attain an n¼ 2.455 at the wavelength of 1.5mm. This is

comparable to the un-doped and photopolymerized As2S3photoresist film. The increase in refractive index upon

photoexposure is a result of photodarkening caused by the

polymerization process.[18] A similar index of refraction before

and after the introduction of erbium atoms demonstrates that

the doping technique does not create any large disturbances of

the optical properties of the As2S3 host. Close index matching

of the doped material with the host material is desired because

drastic changes in the refractive index due to doping could limit

www.advmat.de � 2008 WILEY-VCH Verlag GmbH &

the design of possible photonic structures. These results show

that the Er/As2S3 photoresist film produced here can take full

advantage of the infinite design possibilities provided by 3D

DLW.

PL excitation spectroscopy (PLE) is used to characterize the

PL properties of the Er/As2S3 photoresist films. A PLE

mapping of the excitation behavior of this film is obtained by

scanning the sample with excitation wavelengths from 640 to

994 nm, in 1 nm steps, and collecting the intensity of the PL that

is generated between the wavelengths of 1.430–1.600 mm.

Three excitation resonances for the PL are identified in

Figure 2a. These resonances occur at pumpwavelengths of 660,

810, and 980 nm and correspond to the excitation from the4I15=2 ground state to the 4F9=2,

4I9=2, and 4I11=2 states,

respectively.[16] From this analysis, 980 nm is established as

being the most efficient wavelength for exciting the PL. In

Figure 2b, a single PL spectrum is analyzed for the behavior of

Co. KGaA, Weinheim Adv. Mater. 2008, 20, 4097–4102

COM

MUNIC

ATIO

N

Figure 2. a) The PLE map of a photoexposed Er/As2S3 photoresist film containing 1.35wt % Er.b) A single PL spectrum of this film (black curve) shows that PL is broadened due toinhomogeneous line broadening by the Stark splitting of the energy levels. c) A model is usedto account for the Stark splitting of the 4I13=2 and

4I15=2 energy levels of the Er atom inside thisamorphous matrix. The single PL spectrum shown in (b) can be effectively fitted (dotted redcurve) using the contributions of each Stark level from the model (color coded curves).

Figure 3. SEM image of the step-ladder structure with one of the support-ing walls removed via FIB milling. The position of the Er-doped layer ishighlighted with a red line as a guide to the eye. The bars are systematicallyshifted vertically in 200 nm steps relative to the doping layer, starting fromthe substrate (top left side of the SEM image). High-magnification SEMimaging of the Er/As2S3 layer (inset) shows that it has a thickness of139 nm (this value takes into account a sample stage tilt of 548).

the generated PL. The most intense feature of the PL spectra is

centered at 1536 nm and has a FWHM of 46 nm, irrespective of

the pump wavelengths used. This emission is characteristic of

an atomic transition from the 4I13=2 excited state to the 4I15=2ground state of the Er3þ ion.[16,22] The PL spectrum displays an

inhomogeneous line broadening brought about by the Stark

splitting of the energy levels of the well-dispersed Er atoms

inside the amorphous As2S3 matrix.[23] The agglomeration of

the Er dopants into microscopic clusters inside the

as-deposited materials is not observed since, the PL spectrum

collected here does not exhibit any sharp spectral features

characteristic of those samples where polycrystalline d-Er2S3are directly embedded into the As2S3 matrix.[24] Figure 2c

shows amodel which is used to account for the Stark splitting of

the 4I13=2 and 4I15=2 energy levels of the Er atom inside this

amorphous matrix. The single PL spectrum shown in Figure 2b

can be effectively fitted (dotted red curve) using the

contributions of each Stark level from the model (color coded

curves).

In order to fabricate a 3D structure where spatially localized

PL can be easily demonstrated, a tri-layer photoresist stack is

first prepared by sequentially evaporating alternating layers of

Er:As2S3 and AssS3 layers onto the substrate using the GPDD

technique. This tri-layer photoresist stack consists of a 139 nm

thin, 1.35wt%Er-doped layer that is precisely placed between

the un-doped As2S3 layers. Then, 3D DLW is used to

photopattern a 3D structure consisting of suspended bars that

are vertically staggered by 200 nm and laterally separated by

Adv. Mater. 2008, 20, 4097–4102 � 2008 WILEY-VCH Verlag GmbH & Co. KGaA,

5mm into this tri-layer stack forming a

step-ladder structure.[15] The design of this

3D structure allows open-visual access to the

features that are inside the structure, which

provides an unobstructed view of the Er

doping layer thus enabling a clear systematic

analysis of its progression and spatially

selective inclusion. After the photopattern-

ing process, the tri-layer photoresist stack

is developed using a highly selective

wet-etchant containing the molecule

N-4-methoxybenzyl-pyren-1-yl-amine that was

developed for As2S3-based photoresists.[15]

SEM images of this 3D structure are

shown in Figure 3. One of the supporting

walls of the step-ladder structure has been

removed via focused ion beam (FIB) milling

to allow for an unimpeded side-view of the

Er-doped layer as it progresses through the

individual bars inside the structure. The thin

Er-doped layer can be seen in the image as

the thin straight line (highlighted in red)

which conforms to the walls of the structure

and traverses through the individual bars. As

each bar is displaced vertically, some of the

bars intersect the Er-doped layer and Er is

incorporated into its structure, while bars

that lie above the doped layer contain no Er

doping. Careful examination of the SEM images in Figure 3

provides direct visual evidence that the Er-doped layer is only

incorporated into the individual bars intersecting this doping

layer but not in the bars above. The inset image in Figure 3

shows a high-magnification SEM image of the Er-doped As2S3layer which runs parallel to the substrate. From this image the

Er-doped As2S3 layer is measured to be 139 nm thick.

To examine the PL behavior of this 3D structure, we have

employed a spatially resolved confocal PL scanning technique

to image each bar of the step-ladder structure individually. An

excitation laser tuned to the wavelength of 980 nm is focused

using a microscope objective into a focal spot on the order of

Weinheim www.advmat.de 4099

COM

MUNIC

ATIO

N

Figure 4. a) A schematic of the step-ladder structure (centre), and the results of the spatiallyresolved confocal PL imaging technique in the 1D line-scanmode (bottom) and 2D imagingmode(inset). b) A representative cross-section SEM image of the bars (left) that were examined toconstruct a model (centre) used to show the progression of the Er doping layer in the bars situatedbetween at the positions between 5 and 65mm of the step-ladder structure (right). c) SEM imagesof bars positioned above the erbium doping layer (highlighted in red). Part of the highlighting isremoved to show actual Er/As2S3 layer (indicated with a red-arrow).

4100

500 nm. The PL between the wavelengths of 1475–1600 nmwas

collected with the same objective, detected using an InGaAs

detector array, and the intensity integrated. The spatially

resolved imaging was performed in both line-scanning mode

and 2D imaging mode. A schematic of the spatially resolved

confocal PL imaging technique is shown in Figure 4.

In the line-scanning mode, the focal spot of the microscope

objective is scanned through the entire structure at 200 nm

steps, in a single straight line bisecting the center of the bars.

The lower image in Figure 4 is a plot of the line-scan recorded

as the integrated PL intensity collected at each position of the

structure. To interpret the collected line-scan, a simple

geometric model of the bar, based on the SEM observations

that are made from each of the individual bar, is presented in

Figure 4b to aid in the analysis.

The geometric model of the progression of the Er-doped

layer inside bars positioned at different heights predicts that

strong PL signals should only be collected from bars that

contain the Er-doped layer, starting from the first bar

positioned at 5mm to the bar at 65mm. All subsequent bars

should not exhibit any PL signals as they lie above the doped

layer. Furthermore, since the volume of the Er-doped layer

changes as it moves through the profile of the bar, the intensity

should also vary between them. These predictions are clearly

observed in the measured line scan Figure 4a. In the line scan,

strong PL is only collected in the first 13 bars, positioned at

5–65mm. Beyond these positions only background PL can be

www.advmat.de � 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

detected. In fact, at the precise positions of

the bars which do not contain the Er doping

layer, dips in the collected background PL

indicate blocking of the collection through

the microscope objective by these bars

positioned between 70 and 100 mm. In

Figure 4c, SEM images of the bars at

positions from 90 to 100mm clearly confirm

that they do not interest the Er doping layer

and therefore no Er is incorporated in its

structure.

In the imaging mode, the focal spot of the

detection apparatus is scanned through part

of the 3D structure to excite and collect any

PL that is emitted from the features situated

between 22.5 and 47.5mm. The PL intensity

collected from this spatially resolved scan is

used to construct a two-dimensional image

of the 3D structure, which is presented in the

inset of Figure 4a. This image indicates that

PL is confined to the individual bars as well

as in the bracing wall of the structure, since

both of these contain the erbium-doping

layer. This PL image conforms to the

analysis of the SEM image in Figure 3,

and demonstrates that PL is observed

wherever the Er-doped layer is present.

These results demonstrate that spatially

selective PL at the important telecommu-

nication wavelength of 1.5mm can be achieved directly in a 3D

nanostructure using Er/As2S3 photoresist prepared by GPDD

and 3D DLW. Furthermore, these results provide three

important structural aspects of this photoresist that is worth

mentioning at this point. First, the surfaces of the individual

bars remain very smooth after etching, therefore and light-

scattering due to surface roughness in a fabricated photonic

structure is prevented. Second, since the photoluminescent

species are embedded inside the bars, the doping process does

not alter the dielectric volume filling fraction of the final

structure from its intended design. Third, the lateral resolution

of the bar produced here is on the order of 180 nm, which is the

same resolution that is obtained when un-doped As2S3 is used

as the photoresist material, therefore the resolution of As2S3 is

not altered due to the Er doping process.[14]

These material properties present obvious advantages over

current schemes used to provide a high-refractive index and PL

in 3D structures fabricated with organic photopolymers. In

fact, since the GPDD technique allows for the position and

thickness of the doped layer to be precisely tuned, complex

photonic structures with multi-level designs schemes can also

be implemented.

In conclusion, we have described GPDD of As2S3 with Er

to produce homogenous, high-refractive index, photosensitive

thin films that display room temperature PL at 1.5mm

wavelength. Fabrication of a 3D nanostructure containing a

spatially confined active emitting layer via 3D DLW is

Adv. Mater. 2008, 20, 4097–4102

COM

MUNIC

ATIO

N

demonstrated in a tri-layer photoresist stack prepared using

the GPDD technique. The material properties of this

photoresist system will enable the direct fabrication of both

active and passive nanophotonic components in one exposure

step, which will present a cornucopia of opportunities for the

facile integration of new and exciting photonic designs on a

single substrate.

Experimental

Preparation of Photoresist Thin Films: Except where indicated,silica glass cover-slips (Menzel Glaser, 170� 10mm) were used assubstrates. The substrates were cleaned in a Piranha solution and thendried under a stream of nitrogen before use. Solid As2S3 glass of>99.99% purity (Amorphous Materials, Garland, Texas) was used asthe chalcogenide precursor material. Pure erbium metal of 99.99%purity (2mm pieces, Alfa Aesar) was used as the erbium precursormaterial. The thin film preparation was performed using a high vacuum(P< 2� 10�6mbar, 2� 10�1 Pa), thermal evaporation depositionmethod. The chalcogenide glass is crushed and placed in an aluminacrucible and the Er pieces are placed directly in a tungsten boat. Thetwo sources were arranged in a co-evaporation setup, where the twovapor cones emanating from their respective thermal sources werearranged so that they provide an area of overlap at the substrate. Thesubstrates are mounted onto a water cooled, rotating substrate holderdirectly opposite at the substrate position. The evaporation of the twomaterials is carried out simultaneously, and their rates are monitoredby two quartz microbalances mounted in a staggered formation inorder to avoid cross-contamination. The entire deposition process isfully automated to ensure reproducibility. The as-deposited films areoptically clear and exhibit transparent light orange color. Films rangingfrom 500nm to 10mm thicknesses are routinely obtained using thisprocess.

Rutherford Backscattering Spectroscopy: Thin films of Er/As2S3were deposited on 2mm thick sapphire substrates. The samples weremeasured, in a standard backscattering geometry, using 2.0MeVHeþ

ions impinging at normal incidence and being backscattered at an angleof 1708. The collected spectra were analyzed by computer simulationusing the RUMP code (www.genplot/RUMP/index.htm) [25].

Photoluminescence and Photoluminescence Excitation Spectroscopy:PL and PLE maps were measured using a home-built near-infraredlaser microscope equipped with wavelength-tuneable excitationsources. The latter could be automatically changed (in 1 nm steps)within the wavelength ranges of 613–707, 693–863, and 832–994nm asprovided by CW dye and Ti:Sapphire lasers (Spectra Physics). Anear-infrared microscope objective (15�, NA¼ 0.4 for PLE mapping,and 100�, NA¼ 0.95 using 980 nm excitation for PL imaging) was usedto focus the excitation beam and to collect PL light. Excitation power inall measurements was�0.8mW.Emission spectra were acquired with aliquid nitrogen-cooled InGaAs photodiode array (sensitive in therange of �800–1600nm) attached to a 30 cm near-infrared spectro-graph (Roper Scientific). The excitation lasers and spectrograph werecoupled to the microscope via optical fibers, with the emissioncollection fiber also serving as a confocal pinhole.

Thin-film Ellipsometery Measurements: As-deposited thin filmswere measured directly without further preparation. For the prepara-tion of the photoexposed samples, the as-deposited photoresist filmswere placed in a quartz tube and flood exposed with a broadband UVsource while under dynamic vacuum (1� 10�3mbar) for 45min. Theellipsometery measurements were performed using a variable anglespectroscopic ellipsometer (JA Woollam). The spectral range of themeasurements was 200–1800nm and three angles of incidence (65, 70,and 758) were used in the measurements. The ellipsometery data weremodeled using a four layer model consisting of the substrate (silicaglass) and Er-doped As2S3 layer, sandwiched between two rough

Adv. Mater. 2008, 20, 4097–4102 � 2008 WILEY-VCH Verl

surface layers (�1 nm) in order to simulate any surface roughness at theinterfaces of the sample. The spectral data below the bandgap of thematerial were fitted using a Cauchy model while a single-oscillatormodel was used to fit the data above the bandgap region [26]. The fitswere optimized using a fitting algorithm based on a nonlinear leastsquares fitting method.

3D Direct Laser Writing and Etching of Structures: 3D DLW wasperformed using a regenerative-amplified Ti:Sapphire laser system(Spectra Physics Hurricane) with a pulse duration of 120 fs and arepetition rate of 1 kHz. The wavelength is tuned to 800 nm, where theone-photon absorption of the chalcogenide glass is negligible. Theoutput beam is attenuated by a half-wave plate/polarizer combinationand then passed through a quarter-wave plate, and after beamexpansion, coupled into an inverted microscope (Leica) and focusedinto the chalcogenide film by a 100� oil immersion objective withhigh-numerical aperture (NA¼ 1.4, Leica). The sample is placed on acapacitively controlled three-axis piezo scanning stage, which isoperated in closed loop and provides a resolution of better than5 nm at a full scanning range of 200mm� 200mm� 20mm (PhysikInstrumente). A computer controls the scanning operation of the piezoand synchronizes its movements with the output of the laser system.

The etchant was prepared and performed using N-4-methoxybenzyl-pyren-1-yl-amine as the etching molecule [15]. The 3D patternedphotoresist was immersed in the etchant until the unexposed areas havecompletely dissolved.

Received: June 2, 2008Revised: July 25, 2008

Published online: October 6, 2008

[1] C. N. LaFratta, J. T. Fourkas, T. Baldacchini, R. Farrer, Angew. Chem.

2007, 46, 6238.

[2] S. Kawata, H.-B. Sun, T. Tanaka, K. Takada, Nature 2001, 412,

697.

[3] S. Klein, A. Barsella, H. Leblond, H. Bulou, A. Fort, C. Andraud,

G. Lemercier, J. C. Mulatier, K. Dorkenoo, Appl. Phys. Lett. 2005, 86,

211118.

[4] R. Guo, S. Xiao, X. Zhai, J. Li, A. Xia, W. Huang, Opt. Express. 2006,

14, 810.

[5] K. K. Seet, V. Mizeikis, S. Matsuo, S. Juodkazis, H. Misawa, Adv.

Mater. 2005, 17, 541.

[6] M.Deubel, G. von Freymann,M.Wegener, S. Pereira, K. Busch, C.M.

Soukoulis, Nat. Mater. 2004, 3, 444.

[7] M. Deubel, M. Wegener, S. Linden, G. von Freymann, S. John, Opt.

Lett. 2001, 31, 805.

[8] M. Florescu, S. John, Phys. Rev. A 2004, 69, 053810.

[9] N. Tetreault, G. von Freymann, M. Deubel, M. Hermatschweiler, F.

Perez-Willard, S. John, M.Wegener, G. A. Ozin, Adv. Mater. 2006, 18,

457.

[10] M. Hermatschweiler, A. Ledermann, G. A. Ozin, M.Wegener, G. von

Freymann, Adv. Func. Mater. 2007, 17, 2273.

[11] M. J. Ventura, C. Bullen, G. Min, Opt. Express 2007, 15, 1817.

[12] Z.-B. Sun, X.-Z. Dong, W.-Q. Chen, S. Nakanishi, X.-M. Duan,

S. Kawata, Adv. Mater. 2008, 20, 914.

[13] P. Lodal, A. Floris van Driel, I. S. Nikolaev, A. Irman, K. Overgaag,

D. Vanmaekelbergh, W. L. Vos, Nature 2004, 430, 654.

[14] S. Wong, M. Deubel, F. Perez-Willard, S. John, G. A. Ozin, M.

Wegener, G. von Freymann, Adv. Mater. 2006, 18, 265.

[15] S. Wong, M. Thiel, P. Brodersen, D. Fenske, G. A. Ozin, M. Wegener,

G. von Freymann, Chem. Mater. 2007, 19, 4213.

[16] A. J. Kenyon, Prog. Quantum Electron. 2002, 26, 225.

[17] E. Desurvire, J. R. Simpson, P. C. Becker, Opt. Lett. 1987, 12,

888.

ag GmbH & Co. KGaA, Weinheim www.advmat.de 4101

COM

MUNIC

ATIO

N

4102

[18] V. Lyubin, M. Klebanov, B. Sfez, B. Askinadze, Mater. Lett. 2004, 58,

1706.

[19] V. Lyubin, M. Klebanov, B. Sfez, M. Veinger, R. Dror, I. Lyubina,

J. Non-Cryst. Solids 2006, 352, 1599.

[20] V. Kh. Kudoyarova, S. A. Kozyukhin, K. D. Tsendin, V. M. Lebedev,

Semiconductors 2007, 41, 914.

[21] J. Fick, E. J. Knystautas, A. Villeneuve, F. Schiettekatte, S. Roorda,

K. A. Richardson, J. Non-Cryst. Solids 2000, 272, 200.

www.advmat.de � 2008 WILEY-VCH Verlag GmbH &

[22] Y. D. Huang, M. Portier, F. Auzel, Opt. Mater. 2001, 15, 243.

[23] L. Bigot, A.-M. Jurdyc, B. Jacquier, J.-L. Adam, Opt. Mater. 2003, 24,

97.

[24] S. Q. Gu, S. Ramachandran, E. E. Reuter, D. A. Turnbull, J. T.

Verdeyen, S. G. Bishop, J. Appl. Phys. 1995, 77, 3365.

[25] L. R. Doolittle, Nucl. Instrum. Methods Phys. Res, Sect. B 1986, 15,

227.

[26] G. E. Jellison, F. A. Modine, Appl. Phys. Lett. 1996, 69, 371.

Co. KGaA, Weinheim Adv. Mater. 2008, 20, 4097–4102