Leibniz-Institutfr

Oberflchenmodifizierung e. V.

BIENNIAL REPORT 2010/2011

Executive BoardDirector:

Prof. Dr. Dr. h.c. Bernd Rauschenbach

Tel.: +49 341 235-2308 Fax.: +49 341 235-2313E-mail: bernd.rauschenbach@iom-leipzig.de

AddressPermoserstrasse 15

D-04303 Leipzig (postal address)D-04318 Leipzig (address for dispatch)

WWWhttp://www.iom-leipzig.de

Members of the Board of TrusteesFreistaat Sachsen, Minister fr Wissenschaft und Kunstvertreten durch Frau RORin Cathrin Liebner

Bundesrepublik Deutschland, Bundesminister fr Bildung und Forschungvertreten durch Herrn MinR Dr. Herbert Zeisel

Verein des Leibniz-Instituts fr Oberflchenmodifizierung e. V.vertreten durch Prof. Dr. Rdiger SzarganUniversitt Leipzig, Wilhelm-Ostwald-Institut fr Physikalische und TheoretischeChemie

Members of the Scientific Advisory BoardDr.-Ing. Markus Roth (Chairman)Osram GmbH, Business Unit UV/IR, Wipperfrth

Prof. Dr. Andre AndersUniversity of California and Lawrence Berkeley National Laboratory, PlasmaApplication Group, USA

Prof. Dr. Klaus-Dieter AsmusAdam-Mickiewicz-University Poznan, Poland

Dr. K.-F. BeckstetteCarl-Zeiss Oberkochen

Dr. Kurt DietlikerETH Zrich and BASF Schweiz AG, Switzerland

Prof. Dr. Roger GlserUniversitt Leipzig, Fakultt fr Chemie und Mineralogie

Prof. Dr. Marius GrundmannUniversitt Leipzig, Fakultt fr Physik und Geowissenschaften

Prof. Dr. A. TnnermannInstitut fr Angewandte Physik der Universitt Jena undFraunhofer Institut fr Angewandte Optik und Feinmechanik, Jena

Organisation of the Institute

MITGLIEDERVERSAMMLUNG

KuratoriumRORin C. Liebner / SMWK (Vorsitzende)

MinR Dr. H. Zeisel / BMBFProf. Dr. R. Szargan / Uni. Leipzig

WissenschaftlicherBeirat

Vors.: Dr. M. RothOsram GmbH

Wissenschaftlich-Technischer RatVors.: Dr. A. Schulze

Physikalische Abt.Prof. B. Rauschenbach

Sekretariat: A. Wedemann

Chemische Abt.N.N.

Sekretariat: M. Zuchhold

OmbudsmannDr. D. Meinhard

Gleichstellungsbe-auftragteDr. U. Helmstedt

Direktor / VorstandProf. Dr. Dr. .Rauschenbach

Administration /Infrastruktur

BetriebsratDipl.-Chem. D. Hirsch

GertetechnischeGrundlagenDipl.-Phys. H. Neumann

Laserstrukturierungund -ablationDr. K. Zimmer

Schichtabscheidungund StrukturierungDr. F. Frost

Anorg. / org. Grenz-flchen u. SchichtenProf. Dr. S.G. Mayr

Oberflchenpr-zisionsbearbeitungPD Dr. A. Schindler

ProzessentwicklungPlasmajetDr. T. Arnold

Grundlagenunters.zu PolymerschichtenDr. T. Scherzer

FunktionsschichtenDr. L. Prager

PolymermodifizierungDr. A. Schulze

BiofunktionalePolymereDr. C. Elsner

VerwaltungDipl.-k. V. ZellinSekretariat: N. Frster

Forschungswerk-stattS. Daum

Technologietransfer /ffentlichkeitsarbeitDipl.-Ing. Y. Bohne

ContentsPreface 7

Scientific and Technology Results 9

Reports 10

Laser processing for CIGS thin film photovoltaics 10

Atomic particle beam techniques progress in ultra-

precision surface finishing 14

Ion beam assisted synthesis and characterization of ferro-

magnetic shape memory alloys for medical applications 17

Development of a new rotogravure printing technology 21

Quantum-chemical modelling of primary processes 25

Membrane hydrophilization using electron beam and plasma

techniques 29

In-line monitoring of the thickness of thin printed layers by

near-infrared reflection spectroscopy: An innovative method

for process control 33

Selected Results 38

Early stages of GaN film growth by ion-beam assisted epitaxy 38

Self-organized pattering on Si by ion sputtering with

simultaneous metal incorporation 39

Advanced electron microscopy in material science at IOM 40

Numerically controlled local plasma jet oxidation of silicon 41

In-situ temperature distribution measurement on electric

propulsion thruster 42

Time dependent decomposition of metastable expanded

austenite phases in FeCrNi and CoCr alloys 43

Investigation of UV- and e-beam curing and properties of

waterborne urethane acrylate nanodispersions 44

Wavelength dependence of the photochemical conversion of

(meth)acrylates in the range of 172-222 nm (VUV-UVC) 45

Aspects of photochemical-based fabrication of gas barriers 46

Glycidol functionalization of plasma-treated polymer surfaces 47

Magnetic particles: A simple approach for the evaluation of

surfaces for bio applications 48

Synthesis and functionalization of porous polymeric materials 49

Personal Activities 51Doctoral Theses 52

Diploma and Master Theses 53

Bachelor Theses 54

Activities in Scientific Organisations 55

Honours and Awards 57

Scientific Events 59Scientific Meetings 60

Institute Colloquia 60

Lectures 63

Seminars 65

Publications and Presentations 67Publications in Journals and Books 68

Conference Proceedings 81

Talks 84

Posters 98

Patents 109

PrefaceThe Leibniz Institute of Surface Modification (IOM), a member of the LeibnizAssociation, combines basic research and application-oriented studies in thefields of surfaces and thin films modified by low-energy ion bombardment, laserand electron irradiation and plasma treatment.

Research and development areas of the Institute are Ion and plasma assisted ultra-precision shaping and smoothing Micro- and nanodimensional structuring and structure transfer Thin film deposition and nanostructures Fundamental principles of polymeric coatings Manufacture of functional coatings Functional nano- and microstructured systems

In its research, the IOM puts strong emphasis on collaborations with industry,small and medium enterprises, universities, and other research laboratories. TheIOM also participates in joint projects directly funded by industry or FederalAgencies such as the BMBF or by the Free State of Saxony. Among extensiveresearch other activities, the participation in DFG research units, the excellencecompetition and main focus programs should be mentioned. The successfulcooperation with chemical, optical, and semiconductor industry was continued.

In this biennial report the IOM presents its scientific activities and majorachievements in the years 2010 and 2011. In this context, the report presentedhere gives a comprehensive summary of our results. In the first part, overviewson selected projects are given, arranged according to the structure of the IOMresearch program. These overviews are supplemented by feature articles onselected topical highlights. Finally, the appendices give a full list of publications,talks, teaching activities, and other achievements of the IOM staff.

The Institute would like to thank all friends and organisations who supported itsprogress in the last two years. Special thank is due to our Board of Trustees andScientific Advisory Board. Our partners from industry and other researchinstitutes play an essential role for the IOM. The Board of the Institute would liketo thank all members and guests of the institute for their active and excellentcontributions to a successful development

Leipzig, January 2012

Prof. B. Rauschenbach

The institute aims with the audit at thecreating sustainability of the existingoffers and the development of furthermeasures to the compatibility of careerand family.The offer presents a balance betweenofficial business and employee-interests.They should strengthen the work-contentment and motivation of the co-workers with it as well as should makepossible the development of theirachievement potential.

The certificate to the audit berufund-familie was handed out over to MsV. Zellin from the IOM in March 2011.

Institute awards in 2010Dr. Thomas. Arnold (research award)Dr. Rajendar Bandari (doctorate award)Dr. Marisa Mder (doctorate award)

Institute awards in 2011Dr. Sergej Naumov (research award)Dr. Johanna Lutz (doctorate award)

Every year the IOM Leipzig honours the best scientific and/or technologicalwork with the research award and the best thesis with a doctorate award.

Scientific and Technology Results

Reports

Selected Results

Reports

10

Laser processing for CIGS thin film photovoltaicsK. Zimmer, M. Ehrhardt, A. Wehrmann, H. Schulte-Huxel

IntroductionHigh-efficient solar cell modules are the key forfurther development of photovoltaics as onecomponent of the alternative energy generatingsystem that must be competitive with conven-tional or other green energy sources. To qualifyphotovoltaic generators for these applicationsboth the technical specifications as well as theeconomical requirements must be fulfilled. Cur-rently a number of very different technicaldevelopments are encouraged to achieve thisgoal. Among them laser processing is one topic,whereby laser application in photovoltaics areranging from laser scribing, laser drilling, andlaser texturing to laser recrystallisation, localdoping, and local contact formation.

Scaling-up issues of copper-indium-gallium-selenide (Cu(In,Ga)Se2, CIGS) thin-film solarcell (TFSC) production and the reduction ofproduction costs below $1/Wp are essentialrequirements for the future development of theentire fabrication process of CIGS modules. Tothis end, laser processing is a key technology toaddress this issues [1].

Thin-film solar modulesThe electrical interconnection technique for thephovoltaic module is essential for the modulefabrication separate solar cells as well as forintegrated approaches. Traditionally, single,entirely processed solar cells are connected formodule fabrication by making use of, e.g., metalribbons or conductive adhesives. Furthermore,other interconnection methods known from rigidsolar cell or electronics, such as soldering,welding, and wire bonding, are underinvestigation for flexible TFSC. However, theunique properties of the materials used for theCIGS solar cell stack as well as of thepolyimide (PI) substrate prevent a simpleadaption of these technologies to flexible TFSC.For instance, since the substrate of the CIGSTFSC is a non-conductive polyimide,conventional ways for interconnections, such asbonded stripes to the top and bottom (back-contact) of another cell are ruled out. Therefore,the choice of employed technolog for flexibleCIGS TFSC is limited and specificinterconnection techniques are mandatory.

Integrated interconnection techniques arebeneficial for thin-film solar cells as a cost-

effective approach but still challenging toachieve. Due to the different fabrication flow ofmonolithic integrated interconnection (MII) andexternal integrated interconnection (EII), seefigure 1, as well as the different interconnectiontechniques, specific optimized laser scribingparameters and processes are required [2-6].The alternative approach of increasing theeffective area of solar modules can be realizedby shingling of solar cells like for a roof [7].

Laser for CIGS TFSC interconnectionBesides their high quality, the efficiency offlexible CIGS solar cells, which was recentlyreported by EMPA to be 18.7% [8] is muchlower, on module level. On the other hand, therecord of a monolithic integrated, flexible CIGSsolar cell module is 15.9% with an aperture areaof 75.7 cm [2]. In contrast, this flexible moduleis made by mechanical scribing of P2 and P3onto a thin (300 m) ceramic substrate; hencethe industrial challenge of CIGS TFSC onpolyimide is not addressed.

Photolithography for patterning of CIGS TFSCworks well [3] but cannot be employed for largearea module fabrication [9]. Currentlymechanical tools are still in use to scribe CIGSTFSC for module fabrication. Especially forselective scribing of thermal sensitive films, suchas CIGS (P2) and the TCO film (P3), thesetechnologies are useful due to the less damageof the TFSC. However, the process suffers fromtool degradation and features limitations in theprecision and speed of the process.

Different approaches for the development oflaser scribing processes in CIGS solar module

P3P2P1 P1* P3P2*

MII EII

CACIGSPI IATCOMo

Figure 1: Sketches of monolithic integrated intercon-nection (MII) and external integrated interconnection(EII) with the scribes P1, P2, and P3. The asterisksdenote the different stages for scribing within thefabrication process. At EII all scribing are done afterthin film deposition. IA /CA: Insulating / conductiveadhesive.

Reports

11

fabrication by monolithic integrated interconnec-tion are known. The scribing of the molybdenumfilm (P1) by laser is well accepted [2-4].However, ns-Nd:YAG laser may produce cracksin the molybdenum that have to be avoided.

Since P2 is a connective scribe according toRef. [3] it is not necessary to remove theCIGS completely. In consequence, a micro-weldprocess was developed, that enables theelectrical interconnection by laser-inducedmodification of the CIGS. Therefore, a betterquality interface especially near the frontcontact can be achieved by depositing the TCOfilm without CIGS scribing [4]. The fabrication ofCIGS modules with a reasonable efficiency madeby laser scribing has been reported [5, 9].Within a close collaboration with industrialpartners the most prospective approaches oflaser utilization for the fabrication of CIGS thin-film solar modules were investigated.

Laser ablation for back side contactsA suitable technique for solar module fabricationis the shingle technique that is also called roof-tile method [7]. For this approach single solarcells are arranged like a shingle roof cladding, inwhich the front and back contact, overlap andare connected with a conductive material. Thisapproach is straightforwardly applicable forTFSC on conductive substrates such as copperor stainless steel foils. For CIGS TFSC on insula-tion polyimide substrates, that offer a number ofbenefits for application, the back contact needsto be exposed prior interconnection of the TFSCin the manner of shingles. The main benefit ofthe shingle technique, confirmed by the sketchin figure 2, is the large effective area of themodule.

For the development of shingled solar cellmodules a back side opening process (BSO) wasdeveloped that is based on the ablation of thepolyimide substrate [10] with UV laser( =248 nm, tp = 20 ns). This BSO process suitsthe requirement of a gentle selective laserablation of the polymer substrate and ischaracterized by: (i) the complete ablation ofthe polyimide substrate, (ii) an almost cleansurface of the exposed metal film, and (iii) anundamaged CIGS TFSC. The basic approach ofthe developed laser process is to perform laserablation in several steps with specificparameters that fit the needs to the process atdifferent stages of the substrate ablation. Thisimplies that not until high laserfluence (~ 800 mJ/cm) is reached, the bulk ofthe PI substrate is ablated whereas near theinterface the laser fluence is reduced to avoidthin-film damage due to melting or disruption.

The result of an exposed molybdenum backcontact fabricated with an optimized BSOprocess is shown in figure 3. The size of a singleexposed area is determined by the laser spotsize; within the optimization of the BSO processexposed areas of up to 800 x 75 m wereachieved.

For the fabrication of complete modules aconductive adhesive was applied to the exposedback contact to perform both the electricalinterconnection as well as the mechanical joiningof the serial interconnected solar cells. Thecontact resistance of such laser-exposed backside openings with a size of 0.1 x 0.1 mm areslightly above the contact resistance of large-area contacts. This effect might be the result ofthe confinement of the conductive adhesive withthe micrometer-sized silver flakes.

Shingled modules from 32 x 38 cm2 CIGS thinfilm solar cells were fabricated using this

Figure 3: Optical micrograph of exposedmolybdenum back contact fabricated with anoptimized BSO process.

Figure 2: Schematically sketch of the shingletechnique for CIGS TFSC on polyimide.

Reports

12

innovative BSO process. The measured V-C curve at a solar simulator proved theoperability of all solar cells of the shingled CIGSsolar module with a reasonable efficiency.

Laser scribing of CIGS for solarmodules in a role-to-role processThe mechanical tools for scribing of CIGS TFSCprovide evidence that laser scribing of such thin-film stacks is still challenging for the currentlaser technology. One reason is the demand onthe functional evaluation of the laser scribingprocess in relation of laser parameters and themethodology for the optimization of the laserscribing in relation to the module performance.Thus, in addition to geometrically demands,such as scribing width etc., the electricalcharacteristic of the thin solar films, includingshunts and contact resistance, are of fundamentlrelevance for the solar module characteristics.

The requirements on the laser scribing processfor external integrated interconnections are evenmore challenging compared to the scribingprocess for the integrated monolithicinterconnection as the damage of the TFSC mustbe avoided for all scribes from P1* to P3.

Quasi in situ I-V measurements of CIGSTFSC

An experimental set-up shown in figure 4 was

developed enabling quasi in situ measurementsat laser scribing. This allows the collection ofreliable and quick information on the changes ofthe solar cell characteristics by the laser scribes.Together with the capabilities of the laserscribing workstation now the scribe geometry,the optical image of the scribe as well as theelectronic properties extracted from the I-V curves can be evaluated quasi in situ for theassessment of the laser thin film interactionsas well as for the laser scribe optimization ofTFSC.

The sequential scribing of CIGS TFSC withultrashort laser pulses ( = 775 nm, tp = 150 fs)cause a sudden drop of the parallel resistancedue to shunt formation as shown in figure 4 b).Surprisingly this happens during the first scribethat results in a laser ablation depth of nearlyone micron but does not remove the CIGScompletely. However, further scribes do notchange the Rp substantially because theobservable Rp rise is caused by the recovery ofthe laser damaged cell areas of the first scribe.

Images of the edges from the laser scribes byoptical microscopy and SEM show differences inthe noticeable modifications: The modifiedregion outside the molten CIGS is onlyobservable in the optical image (see figure 5)whereas in the SEM the TCO film seems to beintact.

I-V characteristics together with the opticallyvisible damage at laser fluences below theablation threshold suggest that the defectmechanism is probably due to the destruction ofthe pn junction at the CIGS/TCO interface [11].This damage mechanism can be understood inthe following manner: Due to the Gaussianintensity distribution of the laser focus and thefound laser damage threshold fluences for775 nm 150 fs laser pulses intensities of0.55 J/cm and 0.25 J/cm are measured forTCO and CIGS, respectively the energydensity at the shoulder of the laser spot is belowthe ablation threshold of the TCO but above the

KeithleyI-V measurements

Laserscribing

PC-Matlab

a)

b)Figure 4: Quasi in situ measurements of CIGS TFSCfor the optimization of the scribing process withultrashort pulse laser (tp ~ 150 fs). a) Experimentalset-up, b) Parallel resistance during sequential laserscribing with a laser fluence of 0.786 J/cm and apulse overlap of 77%.

Molybdenum molten CIGS Modificationpristine

CIGS film

1 m1 m

TCO-film edge

Figure 5: Optical micrograph of the edge region froma laser scribe with the characteristic features.

Reports

13

damage threshold of the strong absorbing CIGSand CdS films. Hence, the damage of theinterface without ablation of the TCO can beexplained.

CIGS modules in a role-to-role process

With the aim to establish a consistent role-to-role (R2R) process for the demonstration ofexternal integrated interconnections of CIGSTFSC on polyimide for module fabrication, amanual R2R module for laser scribing has beendeveloped. This R2R module mounted into thepicosecond laser workstation and used for laserscribing is shown in figure 6. Due to theincorporation of the module in the laserworkstation the full functionality including imagereorganization, laser processing, and autofocusis available. These features can be combined toa fully automated process flow enabling a fullCNC control of laser scribing of CIGS thin solarstack material for the fabrication of P1*, P2*,and P3 scribes with different processingparameters. A relative precision and an overlayaccuracy to external markers of better than5 m and 10 m has been measured within theR2R module area of 20 x 20 cm, respectively.

The compete EII process for CIGS modulefabrication comprises (i) the deposition of thecomplete stack of CIGS SC material, (ii) theperforming of the laser scribing with ultrashortlaser pulses, and (iii) the external integratedinterconnection by screen printing of a silver-containing conductive adhesive. The benefits ofthis technology are the separation of the thin-film deposition and the laser scribing, wherebyboth the thin-film deposition and the laserscribing can be optimised without interference,the less demands on the scribing procedureconcerning the requested accuracy, and theopportunity of utilization the same process for

CIGS thin-film deposition as used for CIGS solarcell fabrication.

Laser-scribed CIGS thin-film modules, seen infigure 7, with output voltages of ~ 4.5 V and anefficiency of approx. 10% were fabricated. Therather extended interconnection area is due tothe limited precision of the screen printing andholds the potential for further increase of themodule efficiency in the future.

The results of this work are based on acollaboration with Ch. Scheit and A. Brau (SolarionAG, Leipzig).

Literature[1] N. G. Dhere, Solar Energy Materials and Solar

Cells 95 (2011) 277.[2] Shogo Ishizuka et al., Solar Energy Materials and

Solar Cells 94 (2010) 2052.[3] F. Kessler, D. Herrmann, M. Powalla, Thin Solid

Films 480 (2005) 491.[4] P. O. Westin, U. Zimmermann, et al., Solar

Energy Materials and Solar Cells 95 (2011) 1062.[5] S. Wiedeman et al., presented at the Photovoltaic

Specialists Conference, 2002. Conference Recordof the Twenty-Ninth IEEE, 2002.

[6] Sae Chae Jeoung et al., Opt. Express 19 (2011)16730.

[7] M. Winkler et al., Thin Solid Films 387 (2001) 86.[8] Ayodhya N. Tiwari, http://www.empa.ch/plugin

/template/empa/*/79159 (2011).[9] Johan Wennerberg, PhD Thesis, Uppsala

University, 2002.[10]K. Zimmer et al., Appl. Surf. Sci. 255 (2009)

9869.[11]A. Wehrmann, H. Schulte-Huxel, M. Ehrhardt, et

al., in Proc. of SPIE 7921 (2011) 79210T.

Scanner I

R2R-module

CIGS web

Camera +Focus

xy-table of the workstation

Scanner II

Exhaust

Figure 6: Optical image of the R2R tool mountedonto the laser workstations stages.

Figure 7: Images of a CIGS module made by EIIusing the R2R laser scribing process.

Reports

14

Atomic particle beam techniques - progress in ultra-precisionsurface finishing

A. Schindler, F. Pietag, A. Nickel, H. Paetzelt, T. Arnold, G. Bhm, T. Hnsel

IntroductionIn last two years in ultra-precision surfacemachining IOM focused on techniques to solvedeterministic correction of mid spatial frequencyroughness (MSFR) surface figure errors usingboth methods - ion beam figuring andatmospheric plasma jet machining (APJM).Amplitudes of those surface errors are in thesub-10 nm peak to valley (PV) range. Besides areduction of the Gaussian shape tool size downto 500 m full width of half maximum (FWHM)of the ion beam and the plasma jet the beamand jet powers had also to be adjustedaccordingly to provide for the necessarily highlong term stability of the respective removalfunctions. Other problems in such high spatialresolution deterministic surface shape correctionarise from the alignment of the tool and theworkpiece together with the measured errortopology of it. The alignment accuracy has to bebelow 10 m in x and y. Especially for largeworkpieces thermal expansion duo to the toolpower will be a problem. With the low powerpulsed microwave cold plasma jet with jettemperatures as low as 30C developed for suchdedicated surface processing including forthermal sensitive materials it was possible tostart research in the new field of plasmamedicine for dentistry application.

Ion Beam Figuring (IBF)Ion beam figuring (IBF) for ultra precisionsurface finishing is well established in high endoptics fabrication for lithography, space andbeam-line optics and advanced opticalinstruments, respectively [1, 2]. IBF standardtechnology uses a constant and stable ion beamthat moves computer controlled across theentire optic via a meander like scanning withvariable scan line velocity according to the localdwell time necessary for the specified removal ofmaterial at the certain place. The beam tool sizehas to be adjusted according to the spatialsurface error size processing. One seriousdisadvantage of this method is the continuousand constant beam and the limitations inmaximum speed and acceleration/decelerationof the mechanical multi axes motion system.This causes wasting of material removal at leastin that part of the surface where no materialshould be removed (absolute minimum of thesurface figure). Further additional surface errorcan be arise in cases of steep gradients andsmall surface feature sizes of the surfacetopology where the beam can not follow thevery small dwell times due to the limitations ofthe dynamics of the motion system.

The new solution uses a pulsed ion beaminstead of a direct current (dc) one combined

CAM-Table

PC I/O: Firewire

Beam on /off

PWM (EIA485) Uacc

Axes-controller

Beam switching unit

Axes motion control

Beam switching parameters:

f = 10 kHz;

tpuls = 2 98 s

PWM = 2 98%

DtCalc

tdwell1, tdwell2

CNanotec

Parameter

DtCalc

tdwell1, tdwell2

CNanotec

Parameter

FW

3-(5-)Axes-motion system

Initial andfinal (target)surface shapedata Beam profile data

Ion source

Ubeam

Figure 1: Scheme of the motion synchronized PWM ion beam controlled ion beam figuring technique.

Reports

15

with pulse width modulation (PWM) for variationof the mean beam power. PWM signal iscontrolled by the motion control of the multi-axes system that moves the ion beam (source)across the surface to be figured extend thelimited dwell time scale by the mechanicalmotion system by two orders of magnitude tolower values. Fig. 1 shows the scheme of thenew technique. New IBF processing software("DtCalc") has been developed using two dwelltimes instead of one as before. The first is thevery same as for the standard IBF techniquewhere the velocity of the motion system iscontrolled and the second is the additional newone which controls the PWM of the ion beam asfollows. For calculated dwell times lower than anadjustable minimum dwell time value(dependent on the maximum speed of themechanical motion system) the motion systemmoves further with a constant velocitycorresponding to this value and automatically atthe very same time the PWM control starts tolower the average beam power that takes effectto reduce the dwell time further. Thus the PWM

beam control allows extend the dwell time rangeby two orders of magnitude. Two hardwarecomponents have been added to a standard IBFfacility, (i) an intelligent axis controller with afast processor unit and a fast data transfer forcontrolling of the standard axes of the motionsystem and the virtual PWM axis and (ii) abeam switching unit for the pulse powering ofthe beam and accelerator voltages of the RF(13.56 MHz) ion source.

Fig. 2 shows results of the simulation for thenew ion beam figuring PWM technique of a lenssurface dominated by MSFR error characteristicswith effect of strong reduction of the baseremoval and the processing time compared tothe state of the art IBF. Fig. 3 showsexperimental test results of IBF processing ofsine-like nanostructures in SiO2 thin film by thestandard and by the new PWM techniques,respectively.

A new IBF PWM technique has been demon-strated with main advantages of strong reduc-tions of the wasting base removal and of the

a)

Initial Surface PV: 5.01 nm RMS: 0.51 nm [nm]

6

5

4

3

2

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

PWM IBF Simul. Result PV: 1,78 nm RMS: 0,15 nm[nm]

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

0,8

0,4

0

0,4

0,8

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

[ms]

30

20

10

0

Local dwell time distribution by axes motion

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

[ms]

2,0

1,5

1,0

0,5

0

Local dwell time distribution by PWM

pixeldistance:0,1 mm

Initial Surface PV: 5.01 nm RMS: 0.51 nm [nm]

6

5

4

3

2

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

Initial Surface PV: 5.01 nm RMS: 0.51 nm [nm]

6

5

4

3

2

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

PWM IBF Simul. Result PV: 1,78 nm RMS: 0,15 nm[nm]

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

0,8

0,4

0

0,4

0,8

PWM IBF Simul. Result PV: 1,78 nm RMS: 0,15 nm[nm]

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

0,8

0,4

0

0,4

0,8

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

[ms]

30

20

10

0

Local dwell time distribution by axes motion

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

[ms]

30

20

10

0

Local dwell time distribution by axes motion

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

[ms]

2,0

1,5

1,0

0,5

0

Local dwell time distribution by PWM

-80 -60 -40 -20 0 20 40 60 80

-60

-40

-20

0

20

40

60

X [mm]

y[mm]

[ms]

2,0

1,5

1,0

0,5

0

Local dwell time distribution by PWM

pixeldistance:0,1 mm

b) Dwell Time [ms]

New PWM-IBF

Standard IBF

Results of PWM IBF SimulationStandard IBF PWM IBF

Base removal 1,275 nm 0,025 nmProcess time 3:46 h 2:25 h

Dwell Time [ms]Dwell Time [ms]

New PWM-IBF

Standard IBF

Results of PWM IBF SimulationStandard IBF PWM IBF

Base removal 1,275 nm 0,025 nmProcess time 3:46 h 2:25 h

Figure 2: a) Upper row: IBF simulation of a 160 mm lens with MSFR sub-nm error features using new PWMtechnique, left: interferogram before IBF, right: residual error map as a result of the IBF simulation, bottom row:calculated local dwell time distributions of the PWM action only (left) and of the axes velocity (right); b) histo-grams of the calculated dwell times for the standard IBF with dc beam (black) and new PWM ion beam IBF (red).

a)

(1)

(2)

[nm]

(1)

(2)

[nm]

b)

- 40 -30 -20 -10 0 10 20 30 4 0-40

-35

-30

-25

-20

-15

-10

-5

0

5

10mean of neighbour ing 25 l ines, line distance 0.08205mm

de

pth

[nm

]

po sit ion [mm ]

w/o PW M

S inusoidal wave s w /o PW M (quarz) - cr oss section

base re mo val ~ 1. 5nm

de

pth

m ea n dep th = 26.5 nm 0.6nm

0 20 40 60 80990

1000

1010

1020

1030

1040me a n of n e igh bou r ing 5 l in es , l in e di s tanc e 0. 1m m

de

pth

[nm

]

position [mm]

Sine with PW M

Sinuso id al w aves w ith P WM - cross sectio n

base removal = 0 !

de

pth

mean depth = 25 .69nm 1nm

-40 -30 -20 -10 0 10 20 30 4 0-40

-35

-30

-25

-20

-15

-10

-5

0

5

10mean of neighbour ing 25 l ines, line distance 0.08205mm

de

pth

[nm

]

po sit ion [mm ]

w/o PW M

S inusoidal wave s w /o PW M (quarz) - cr oss section

base re mo val ~ 1. 5nm

de

pth

m ea n dep th = 26.5 nm 0.6nm

0 20 40 60 80990

1000

1010

1020

1030

1040me a n of n e igh bou r ing 5 l in es , l in e di s tanc e 0. 1m m

de

pth

[nm

]

position [mm]

Sine with PW M

Sinuso id al w aves w ith P WM - cross sectio n

base removal = 0 !

de

pth

mean depth = 25 .69nm 1nm

Figure 3: a) Interferogram of sinusoidal structures etched by standard IBF (1) and PWM IBF (2) into a quartzplate; b) cuts of both profiles from a), showing the strong reduction of the base removal in PWM case.

Reports

16

IBF processing time, respectively furtherlowering the wear of the axes system andenabling precision figuring of separate areas ofsurfaces.

Atmospheric Plasma Jet MachiningAtmospheric Plasma Jet Machining (APJM) withreactive gas components has a greattechnological potential for the ultra-precisionmodification and figuring of optical surfaces.High rate and sub-surface damage freeprocessing are the main advantages of themethod. Figuring with sub-mm lateral resolutionand sub-nm depth accuracy has been achievedfor different optical materials using fine focusedatmospheric plasma jets with sub-millimeterFWHM tool size. As an example, results for fusedsilica are shown.

Two microwave (2.45 GHz) poweredatmospheric plasma jet sources have beendeveloped to perform APJM of silicon basedoptical materials [3, 4]. System I operates indc-mode up to 600 W while system II is in apulsed mode with an adjustable average outputpower of 3 to 10 W. Ar/He-plasma is used forboth systems with NF3, SF6 or CF4 as fluorinecontaining reactive gas and O2/N2 as additionalgases. Typical total flow is less than 10 slm forsystem I and less than 5 slm for system II. Bothsystems were optimized to achieve a rotationallysymmetric removal function with a nearGaussian shape (FWHM size of ~0.7 mm) andmaterial removal rates of 0.001 to 0.003mm3min-1 for fused silica. System II operates incontact mode with the plasma jet touching thesurface (Fig. 4) while system I uses the reactiveplasma species of the plasma jet afterglow forthe etching process. We used both systemsmounted on 5-axes CNC machines allowing thetreatment of large work pieces with sizes of upto 500 mm in diameter. Deterministic surfaceshaping is performed by the same dwell timealgorithm and processing software as forstandard IBF described above. Details on dwelltime calculation based on MATLAB-programscan be found in [1]. Most important

preconditions for ultra precision large areafiguring are (i) a very long term stable plasmajet removal function and (ii) a most accurateAPJM process simulation calculation fordetermining the dwell time distribution. The firstone was achieved by optimization of the processparameter sets and the source configurations,the second by extended measurements of thetool working function by etching of foot-prints,line profiles and meander-like scanned areas.APJM tests of samples (up to 150 mm ) withboth systems showed substantial reduction ofthe mid spatial figure errors by 50% and more.Fig. 5 shows the interferograms of a 30 x 30mm part of a fused silica lens before and afterAPJM using system II. Average scan velocity was6 mms-1, scan line distance 150 m and theprocessing time 17 min.

For the first time we have demonstrated APJMdeterministic fine correction of nanometric midspatial frequency surface figure errors for opticalfinishing technology. Since there are almost nogeometric restrictions for the contactless plasmajet tool there exist nearly no limits for theoptical surface correction and design. This opensa very interesting and powerful new technology.

The results of this work are based incollaboration with K. Nomura, Nikon Corp.Sagamihara, Japan and M. Weiser, N. Kaier, CarlZeiss SMT GmbH Oberkochen, Germany.

Literature[1] T. Hnsel, F. Frost, A. Nickel, A. Schindler,

Vakuum in Forschung und Praxis, 19 (2007), 24-30.

[2] M. Weiser, Nucl. Instr. Meth. B, 267 (2009),1390-1393.

[3] T. Arnold, G. Boehm, I. Eichentopf, M. Janietz, J.Meister, A. Schindler, Vakuum in Forschung undPraxis, 22/4 (2010), 10-16.

[4] T. Arnold, G. Bhm, R. Fechner, J. Meister, A.Nickel, F. Frost, T. Hnsel, and A. Schindler, A.,Nucl. Instr. Meth. A, 616 (2010), 147-156.Figure 4: Pulsed microwave (2.45 GHz) powered

APJM (system II) during fine correction.

+6.00

-4.00

[nm]

+6.00

-4.00

[nm]

Figure 5: Sub-mm spatial surface error correction of a160 mm SiO2 lens by APJM. Surface error maps of200 x 200 pixel (pixel size 150 m x 150 m) areshown: before APJM PV 9.13 nm, RMS 1.09 nm (left);after APJM PV 5.23 nm and RMS 0.39 nm (right).

Reports

17

(a)

(b)

(c)

Figure 1: Scheme of shape changing process atdifferent stages for Ni-Mn-Ga ((a) (b) (c) andvice versa). An external magnetic field inducesreorientation of twin variants due to aligningmagnetic moments.

Ion beam assisted synthesis and characterization offerromagnetic shape memory alloys for medical applications

Y. Ma, A. Arabi-Hashemi, A. Graumann, A.M. Jakob, I. Claussen, T. Edler, S.G. Mayr

IntroductionFerromagnetic shape memory alloys (FSMA) arevery promising materials for various kinds offuture applications. Compared to conventionalshape memory alloys, FSMAs are fast switchableby a moderate external magnetic field atconstant temperature. Such as in Ni-Mn-Gasystems [1], as shown in Fig. 1, high strains dueto easy moveable martensitic twin variants (Fig.1b) can be induced via magneto elastic coupling[2]. So far, the maximum strain values up to10% have been reported for Ni2MnGa singlecrystals [3], while the maximum strain of 5%can be yielded in Fe70Pd30 single crystals [4].However, Fe70Pd30 has better mechanicalproperties as it is more ductile after martensitictransformation and does not tend to fractureformation [5].

Superior straining characteristics at appliedexternal magnetic fields enable a high potentialfor applications such as actuators, sensors,valves and switches. Especially thin film FSMAare highly promising for developing miniaturizedlow power switching devices applicable in micromechanics and medicine. This requires moredetailed knowledge of their mechanical

properties especially at the micro and nanoscale. Previous investigations, however, onlyfocused on determining mechanicalcharacteristics of large bulk samples. Measuringtechniques utilized for this purpose are notsuitable for our requirements. By means ofcontact resonance atomic force microscopy (CR-AFM), one can overcome this lack in order to geta deeper understanding in nano mechanicalprocesses behind the FSM effect.

Additionally, to gain more information on phasetransition conditions of Fe70Pd30,superconducting quantum interference device(SQUID) as well as temperature dependent X-ray diffraction (XRD) measurements wereperformed.

External controllability at constant temperaturesand sufficiently high strains makesferromagnetic shape memory alloys alsoexcellent candidates for biomedical actuation,such as surgical implant material. However thebiocompatibility and in vivo behaviour of thismaterial must be well confirmed before it can besafely used as an implant material. Ni2MnGa isnot a suitable material for such kind ofapplications due to its nickel content which ispoisonous at higher concentrations. For Fe70Pd30biocompatibility pro-perties have been examinedwith in vitro assessments in cooperation with theSoft Matter Physics Division (Physik der weichenMaterie) of the University Leipzig [6].

ExperimentalIron-Palladium thin films are synthesized byphysical vapour deposition by means of electronbeam evaporators in optional presence ofenergetic ion beams. Utilizing two separatelyrate controlled evaporation sources for Fe andPd allows a very precise adjustment of the filmcomposition of At-% Fe 70 and At-% Pd 30(Fig. 2).

High crystalline quality is achieved while heteroepitaxial growth of Fe70Pd30 on MgO (100)substrates is carried out under ultra highvacuum (UHV) conditions with a base pressureof less than 10-9 mbar. In addition, the substrateis kept at 850 C to provide good diffusivity ofthe deposited atoms on the surface, which isalso enhance by bombardment with energetic

Reports

18

Heater

Evaporator

Oscillatingcrystal MgO

Substrate

Fe Pd

Fig. 2: Schematic of the evaporation chamber.

Figure 3: TEM characterization of the MgO/ Fe-Pdinterface.

Fig. 4: XRD spectra before and after lift-off.

ions. Thin film single crystallinity is confirmed byhigh resolution transmission electron microscopy(TEM, Fig. 3) [7, 8] and pole figuremeasurements published in previous literature[9, 10].

Fe and Pd are deposited with a total depositionrate of around 1,5 A/s. Typically 500 nmthickness films are prepared.

That Fe70Pd30 is not stable at room temperatureis an additional reason for deposition at hightemperature. Using a liquid nitrogen shieldsupports on the one hand the ion getter pumpsand on the other hand it ensures a quick coolingdown after epitaxial thin film growth. Therefore,Fe70Pd30 films reach a metastable state at roomtemperature without decomposing into its stablephases alpha-Fe and Fe50Pd50 [11].

In order to eliminate the substrate constraints infunctionalization and application purposes, thesefilms must be lifted off from the MgO substrate.

Our results show, freestanding thin films can bereadily prepared by chemically dissolving theMgO with a NaHCO3 solution. Meanwhile, thecrystal structure, phase and composition arekept intact after lift off which is confirmed byEDX and XRD measurements (Fig. 4). Analternative lift off method is etching the MgOsubstrate using an EDTA solution [9].

Ni2MnGa thin films were epitaxially grown byDC-magnetron-sputtering of a pre-alloyedtarget. The base pressure is below 10-6 mbarwhile during the process pressure is kept at 10-3

mbar with flowing Argon. The MgO substrate isheld at 600C during deposition. After 35 min ofsputtering at a mean power of 50W thesubstrate heater is switched off. Anapproximately 400 nm thick thin film is obtainedwhich has the orthorhombic 7M phase aftersubsequently cooldown.

Mechanical properties of Ni-Mn-Ga thin

films

CR-AFM measurements were performed on Ni-Mn-Ga samples to gain information onindentation modulus M at the nano scale.Compared to theoretical predictions based onDensity Functional Theory (DFT) calculations,experimentally observed values are significantlyreduced, indicating movement of twin variants(see Fig. 1) due to stresses applied by thescanning probe. This behaviour could beaccompanied by intermartensitic phasetransitions. Further investigations are necessaryto get a deeper understanding. Moreover,mechanical contrast imaging revealed variationsof M up to 10% near boundaries between twinvariants. We therefore suggest that movementof existing twin variants during straining ispreferred over generating boundaries withinsingle twin variants.

Reports

19

200 250 300 350 400116

120

124

128

132

136

140

144

1.00

1.04

1.08

1.12

1.16

1.20H = 300 Oe

In plane

300-400 K

400-200 K

200-400 K

Ms

= 324 K

No

rmaliz

ed

Mag

netiz

atio

n

Af= 326 K

Mag

neti

zati

on

[em

u/g

]

Temperature [K]

Fig. 5: Temperature dependent magnetizationmeasurements on martensitic Fe-Pd film on MgO(100) substrate.

Magnetic properties of Fe-PdAttempts to investigate the phasetransformation temperature in substrate-attached and freestanding Fe-Pd thin films weresuccessful with temperature dependentmagnetization measurements (200 K to 400 K;temperature sweep rate approx. 2 K/min). Bycollaboration with Prof. Esquinazi from theDivision of Superconductivity and Magnetism ofthe Faculty of Physics and Earth Sciences of theUniversity of Leipzig, these measurements wereperformed by a SQUID magnetometer MPMS-7from Quantum Design in presence of an in-planeexternal magnetic field.

Fig. 5 shows the absolute and relative (norma-lized to 200 K) temperature dependentmagnetization of martensitic Fe-Pd filmsattached on substrate measured in a magneticfield of 300 Oe; arrows mark the direction oftemperature variation. It is clearly visible thatmagnetization almost reversibly increases as afunction of temperature up to 326 K, whilefollowing a Curie-Weiss behavior We interpretthis reversible phase transformation in terms ofthe favorable martensite variants. As the easyaxis of magnetization is established to coincidewith the longer a axis in Fe-Pd system [12],shape anisotropy of the film enforces in planemagnetization. Furthermore, the externalmagnetic field is too weak to enforcereorientation, only half of the variants will havetheir magnetization directed parallel to theapplied magnetic field deep in the martensitestate. This behavior changes dramatically duringthe martensite -> austenite transition, whereboth variants equivalently become cubic withone of their three easy axes (and thusmagnetization) directed along the externallyapplied field. While magnetic domain structurechanges favoring alignment of magnetizationalong the external field, this is reflected by anincreasing magnetization. From the location ofmaximum magnetizations, which reveal a slighthysteresis, the martensite start temperature(Ms) and austenite finish temperature (Af) aredetermined by the tangential method to be 324K and 326 K, respectively. Consequently, thephase transformation temperature is confirmedby XRD (T) and SQUID independently to occurat about 326 K for the martensitic Fe-Pd thinfilms attached on MgO substrates.

Additionally, the same procedure of temperaturedependent magnetization measurements hasbeen performed on austenitic and martensiticfreestanding Fe-Pd thin films. The martensite

variants selection is observed from the M(T)curves measured at a magnetic field of onlysome tens of Oe. These findings pave the wayfor application of Fe-Pd based membranes inminiaturized devices.

BiocompatibilityAt the present state our work focused on in vitrobiocompatibility assessments of single crystallineFe70Pd30 FSMA films grown on MgO (100) singlecrystal substrates. As the first step, a simulatedbody fluid (SBF) approach to mimic the nominalion concentrations in human blood plasma, asproposed by Kokubo and Takadama [13], wasemployed at the IOM. The SEM images of themorphological changes on the surface of Fe-Pdfilms during SBF test showed that bonelikeapatites with granular microstructure formedafter soaking in SBF, these findings clearlyindicate, that Fe70Pd30 film can, in fact, inducematerials aggregation from the SBF on itssurface, which is very important for bone-bonding between the live tissues and implants.Moreover, the Ca/P ratio of 1.32 was determinedfrom the EDX spectra. Fe and Pd concentrationswithin SBF before and after specimen removalwere measured with inductively coupled plasmaoptical emission spectroscopy (ICP-OES). Theresults indicate that Fe concentration increased,while the concentration of Pd did not changeduring the test.

As the second step, viability cell tests wereperformed to investigate the interactions of NIH3T3 embryonic mouse fibroblast cells with thesample material. This work was conducted byDr. Zink from the Soft Matter Physics Division ofthe University of Leipzig.

Fig. 6 presents the optical morphologicalfeatures of NIH 3T3 cells after 65 h growing on a500 nm Fe70Pd30 film and staining with calceinacetoxymethylester (AM) and propidium

Reports

20

Figure 6: Optical morphological features of greenfluorescent mouse fibroblast cells from calcein stainingon the 500 nm FePd film (a) and plastic culturedish(b). Cells in the white circle had a red fluorescentcore from PI which signal is too weak to overshine thecalcein fluorescence.

iodide (PI). It turned out that cells adhered andproliferated on the surface of the film, somesamples exhibited a lower cell density comparedto the cells on the surface of the surroundingculture dish. Additionally, cells on the film wereslightly smaller than cells on the culture dish. Itturned out that after staining with calcein AMand PI, cells on the film surface fluoresced ingreen. Only a few red cell cores weredetermined, whereas these cells also exhibited astrong green calcein signal and adhered to thefilm surface. This behavior indicates that thesecells were not apoptotic but pathologicallytransformed. In contrast, only viable cellsfluorescing in green were obtained in thesurrounding culture dish without any redfluorescing cores.

Anyway, these results suggested that Fe70Pd30films are biocompatible with little restrictions.Therefore, in vivo mouse test are planed tofurther test the biocompatibility of Fe70Pd30 films.Additionally, film coatings will be employed toimprove biocompatibility.

ConclusionIn summary, single crystalline Fe70Pd30 andNi2MnGa thin film shape memory alloys weresynthesized using electron beam evaporationand magnetron sputtering, respectively.Regarding nano mechanical characterization ofNi2MnGa, CR-AFM revealed the important role oftwin boundaries for the FSM effect. Further CR-AFM investigations will deal with comparableobservations of Fe70Pd30 thin films.

Furthermore, magnetic properties of Fe70Pd30were examined with SQUID measurements. Theaustenite-martensite transitiontemperature wasdetermined at about 326 K, which is alsoconfirmed with XRD (T) measurements.

The in vitro tests show the biocompatibility ofFe70Pd30 with little restrictions. Bonelike apatites

with granular microstructure formed on it aftersoaking in simulated body fluid which makes itan interesting candidate for a wide range ofmedical purposes. Key future work will befocused on film surface modification forimproving the biocompatibility of Fe70Pd30 films.

AcknowledgementsThis project is funded in parts by the GermanFederal Ministry of Education and Research(BMBF, PTJ-BIO, 0313909), the LeipzigGraduate School of Natural Sciences Buildingwith Molecules and Nano Objects (BuildMoNa)the DFG as well as the SMWK/ESF.

The results of this work originated incooperation with M. Zink A. Setzer and P.Esquinazi, University of Leipzig, Germany.

Literature[1] G. J. Mahnke, M. Seibt, S.G. Mayr, Phys. Rev. B

78 (2008) 012101.[2] R. D. James, M. Wuttig, Philos. Mag. A 77 (1998)

1273.[3] A. Sozinov, A. A. Likhachev, N. Lanska, K.

Ullakko, Appl. Phys. Lett. 80 (2002) 1746.[4] T. Kakeshita, T. Fukuda, Mater. Sci. Forum 394

(2002) 531.[5] J. Cui, T. W. Shield, R. D. James, Acta Mater. 52

(2004) 35.[6] Y. Ma, M. Zink, S.G. Mayr, Appl. Phys. Lett. 96

(2010) 213703.[7] T. Edler, J. Buschbeck, C. Mickel, S. Fhler, S.G.

Mayr, New J. Phys. 10 (2008) 063007.[8] L. Khnemund, T. Edler, I. Kock, M. Seibt, S.G.

Mayr, New. J. Phys. 11 (2009) 113054.[9] T. Edler, S.G. Mayr, Adv. Mater. 22 (2010) 4969.[10]T. Edler, S. Hamann, A. Ludwig, S.G. Mayr,

Scripta Mater. 64 (2011) 89.[11]I. Kock, T. Edler, S.G. Mayr, J. App. Phys. 103

(2008) 046108.[12]J. Cui, R. D. James, IEEE Transactions on

magnetics 37 (2001) 2675.[13]T. Kokubo, H. Takadama, Biomaterials 27 (2006)

2907.

Reports

21

Development of a new rotogravure printing technologyY. Bohne, U. Helmstedt, A. Freyer, M. Naumann, B. Rauschenbach

IntroductionRotogravure printing is a type of direct printingfor which the image is engraved onto a cylinder.It combines cost-efficient production of prints inhigh numbers of copies with best quality andeasy-to-handle printing. These advantages areopposed by a complex and time-consumingimage carrier production. It compriseselectrodeposition of various metal layerspartially with highly toxic galvanic baths toprovide the printing surface as well as digitalelectromechanical engraving of the image with adiamond stylus.

To improve the technology recently differentstudies focused on laser imaging, but have notyet been realized technically. In the majority ofcases a direct laser engraving of the cylinderssurface is discussed. Replacement byalternatives to metal, e.g. ceramics or polymerswere in discussion, each with different laserablation mechanisms, interpreted as thermal ornon thermal ablation. Various laser sources such as cw-, pulsed- or UV-Excimer-lasers were used, respectively.

Combining IOMs expertise in UV-curablenanocomposite materials and laser ablationtechniques seemed a promising way to realizethese ideas technically. One has to keep in mind

the technical requirements arising from aprinting cylinder of about 4 metres length, whichrotates with a regular speed of 10 m/s.

Here, it is reported about a successfuldevelopment of the following technology on alaboratory scale ([1], figure 1): The cylinder iscoated with a nanocomposite in a dip-coatingprocess followed by UV-curing. High-qualitygrinding provides the dimensional accuracy andsurface roughness required for the printingprocess. An ultra-short pulse UV-laser is used toengrave cells which in combination with thenanocomposites surface gave satisfying resultsfor the printed image. After the printing processthe cylinder can be prepared for the next writingstep either by a new grinding process or byrefilling the engraved images followed by a newdimensional grinding.

Three fields had to be worked on, which arediscussed in their order within the technologychain:

1. Coating material

A hard, scratch- and solvent-resistant polymericmaterial had to be found which providessufficient adhesion to the companies printingcylinders surface. After curing it had to beprocessable to the needed surface accuracy androughness as well as engravable by the laserefficiently and without remaining ablationproducts. Upscaling for real-machine tests hadto be managed.

2. Printing cylinder preparation

An efficient and technologically feasible processfor applying the nanocomposite layer onto thecompanies printing cylinders was needed. Forthis purpose the technology was developed inlaboratory scale for cylinders of about 30 cmslengths. A technology was sought to achieve therequired dimensional accuracy and roughness ofthe printing surface without defects.

3. Laser imaging

An efficient way to engrave the image into thenanocomposite surface which results in printedimages of the required rotogravure qualityneeded to be developed. The latter could beverified by test prints on a laboratory printingunit, provided by the industrial partner.

Cylinder Workflow

Dip-Coating

UV-Curing

GrindingDirect LaserEngraving

RIP(Raster Image

Processor)

PostScript

Data File

PrintingProcess

Cleaning andStoring

Cylinder

SurfacePreparation

Figure 1: Cylinder workflow in the new rotogravureprinting technology.

Reports

22

Material DevelopmentUp to now the polymer materials tested in theliterature for image carrier coating(thermoplastic polymers, elastomers, resins,waxes and others) do not meet the abovementioned criteria. Organic-inorganicnanocomposites possess properties to fulfilthese specific demands. In the scope of thisproject an UV-curable nanocomposite wasdeveloped, which is based on radicalically UV-curable acrylates and filled withaluminiummaleate nanoparticles (ALMAL). UV-curable compositions have the advantage ofbeing free of solvents. This fact and theabolishment of the handling of galvanic bathswould lead to a safer working environment forthe employees.

Adhesion and solvent resistance

At the beginning of the project six differentacrylate formulations were identified out of alarge number of commercial available UV-curable acrylates, which showed both,acceptable solvent resistance and adhesion onthe cylinders surface. Those two properties arecontrary qualities. A high degree of cross-linkingyields high solvent resistance but low adhesiondue to high volume contraction during curing.Above mentioned six formulations exhibited anacceptable compromise.

Performance in the laser processEngraving of the image into the coating has tobe highly efficient and precise. Therefore theinfluence of nanoparticle fillers on the laserprocess was tested. Nanoparticles with differentchemical composition and surface modification[2], were dispersed in selected formulations.The resulting coatings were tested in a modellaser process. The efficiency of laser ablation forthe different materials is shown in figure 2 and

identifies formulations with ALMAL and TiO2nanoparticle fillers to perform best. According tothe other relevant criteria, ALMAL coatings wereidentified as superior within the process.

Barrier coating

As could be shown by stress tests in thelaboratory printing unit (see next chapter) thematerial meets the abrasion and solventresistance criteria on the laboratory scale. To beprepared for an eventual failure in real scalemachines preliminary tests for a barrier coatingfor the image carrier were performed. A layer ofSiOx was deposited on the surface of the grindednanocomposite by UV-induced conversion ofpolysilazanes [3]. Mechanical studies showed anoticeable reduction of abrasion (Figure 3).

Printing cylinder preparationCoating

For the manufacture of recyclable cylinderslaboratory-scale printing cylinders were coatedwith composite material using a custom-builtcoating machine combining coating and UV-curing pro-cesses, both executed in rotationmode (Fig. 4).

The coating technology was chosen due to thehigh viscosity of the coating material. Itcomprises dip-coating with a coating blade toensure consistent layer thickness adjustmentand homogeneous distribution of the material

a)a) b)b)

Figure 4: Custom-built coating machine for preparationof a recyclable cylinder with part a) coating process andpart b) UV-curing arranged in rotation.

0 200 400 600 800 1000 1200 1400

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0no

2

ALMAL2

SiO2

TiO2

ZiO2

abla

tion

depth

/puls

ein

m

laser fluence in mJ/cm

Figure 2: Ablation depth per pulse in nanocompositematerials with different inorganic fillers.

0 1000 2000 3000 4000

outside

inside

Squeegee Range

Width [m]

b)

0 1000 2000 3000 4000-20

-15

-10

-5

0

5

a)

Width [m]

Depth [m]

Squeegee Rangeinside

outside

0 1000 2000 3000 4000

outside

inside

Squeegee Range

Width [m]

b)

0 1000 2000 3000 4000-20

-15

-10

-5

0

5

a)

Width [m]

Depth [m]

Squeegee Rangeinside

outside

Figure 3: Wear marks measured with profilometer: a)without protective coating, b) with protective coating.

Reports

23

during the coating process.

For layer deposition the dipped cylinder rotateson its axis resulting in formation of smoothsurfaces with marginal roughness after UV-curing (Fig. 5). The thickness of the depositedcomposite layers upon the printing cylindersvaried between 100 to 300 m.

Surface processing

After the coating process a surface finish of thecoating is required to eliminate irregularities inthickness, to produce a plane-parallel surfaceand to exhibit the characteristic surfaceroughness for printing.

Utilisation of a cylindrical grinding machineaffords the generation of plane surfaces( 10 m) with roughness 1 m necessary forprinting surfaces that do not tone if no image isengraved.

The stability of a coated printing cylinderregarding mechanical stress during handling andprinting is heavily determined by sufficientadhesion between the cylinders surface and thecoating material. To support adhesion bychemical modification of the coating material amechanical pre-treatment of the cylinderssurface was developed. The stability of coatedand polished cylinders tested in a laboratoryprinting unit (Fig. 6) indicated optimal resultsusing a surface roughness of Rz 10 m for theuncoated cylinder.

A 15 h long-term test could affirm the suitableabrasion resistance of the material. Onlymarginal roping was observed on the surface.

Recycling

Another technological step is the recycling of thecoated cylinder after print. Different recyclingprocesses were proven to be acceptable optionsfor the process. Re-coating of surface afterprinting was tested using an already coated andpolished cylinder after print. During the secondcoating process a homogeneous distribution ofuncured compound on the UV-cured layer couldbe observed as a result of good wettability of

the cleaned surface. The two-layer cylinderswere subsequently grinded and tested in thelaboratory printing unit with the result that nofailure or layer separation was observed.

Laser ImagingLaser ablation of polymers uses an intenseabsorption of those materials in the ultravioletregion. Using laser fluences near the ablationthreshold chemical bonds can be brokenphotochemically without thermal damage [4].Based on this well known ablation-mechanismthe formulated polymers were investigated.

Material testingTo test the developed materials firstly planecoated samples (film-thickness approx. 100 m)have been studied using UV laser radiation ( =248 nm, tp = 20 ns). Both, laser fluence andpulse number were varied. Ablation depths up to0.6 m/pulse could be achieved. Heat affecteddamage such as cracks, burr or deformation wasnot detected. Carbon related ablation products deposited at the surroundings of cells could beeasily removed. To provide insight into theinteraction of the engraved surface with ink,printability was tested for a grey scale wedge(depth and diameter of cells vary). Hence,varying laser spot sizes (20-100 m) were usedto expose simple pyramidal stepped cells. A testprint showed satisfying preliminary results.

Direct laser ablation of gravure cells

Large dimensions as well as high resolutions ofgravure print forms require an efficientengraving process. Concerning this, high pulserepetition rates are crucial for large scale microstructuring [5].

For this purpose an industrial ultra-short pulseUV laser was applied ( = 355 nm, tp= 10 ps,frep = 1 MHz), whose reduced pulse durationleads to less heat diffusion into the polymer [6].The picosecond-workstation was adapted to therequirements for the imaging of the laboratory-

Figure 6: Long-term test (15 h) of a coated andpolished cylinder in a laboratory printing unit(MOSER HS-157) using cyan ink.

Figure 5: Laboratory-scale printing cylinder aftercoating and UV-curing using composite material.

Reports

24

scale cylinders suitable for the coating machineand the laboratory printing unit. The pulse-defined programming of the workstation enableseach geometrical cell shape by overlappingseveral laser pulses (spot diameter about15 m). Typical cell volumes produced werebetween 2 x 103 and 300 x 103 m.

Provided cylinders have been structured byfluences in the range from 3 J/cm to 7 J/cmwithout any thermal damage or debris. Process-related ablation particles were completelyremoved by exhaustion (see figure 7).

To engrave rotogravure characteristic printedwedges coated cylinders were structured usingpyramidal, cylindrical, prismatic, elliptic andconic cell geometries in five gradations. Theresult for all imaged areas 5 cm per wedge showed excellent ink transfer to the paper asconfirmed by the printout in figure 8.

To scale up the technology optimizedparameters were fixed. Regarding processperformance on real scale further characteristics such as engraving efficiency, repeatability andan 8-bit grey scale will have to be taken intoaccount in further investigation.

ConclusionIn the laboratory scale the full technology chainhas been developed and showed to work in areproducible manner:

Application of an abrasion-resistant and laser-engravable nanocomposite layer makes galvanicprocess steps redundant, releases no volatileorganic solvents and therefore will reduceenvironmental management costs. A dip-coatingtechnology could be established which results inlayers of about 200 m thickness, which are UV-curable and processable to result indimensionally accurate and printable surfaces.The developed laser engraving technology iscompetitive to state-of-the-art electro-mechanical engraving of images regardingaccuracy.

For first tests on real production cylinders thematerial was produced in a 10 kg scale bysubcontractors (Convertex Chemie GmbH andCetelon Nanotechnik GmbH). Prinovis Ltd. & Co.KG could upscale the dip-coating and surfaceprocessing. The first coated production cylinderpassed the first dry stress test in the realprinting machine successfully.

In addition to these promising results thedeveloped laser-imaging process opens up newpossibilities concerning shape-variability of theprinting cells. This will lead to higher printingstandards and could open up new markets forrotogravure printing.

The results of this work are based incollaboration with C. Jahn-Wolf, O. Krmer, A.Schenk, W. Berthold, Prinovis Ltd. & Co. KG,Dresden.

The authors thank the Schsische Aufbaubank(SAB) for funding.

Literature[1] Y. Bohne, C. Elsner, B. Rauschenbach, C. Jahn;

EP 2 151 324 A3.[2] F. Bauer, V. Sauerland, H. Ernst, H.J. Glsel, S.

Naumov, R. Mehnert; Macromolecular Chemistryand Physics 204 (2003) 375.

[3] L. Prager, A. Dierdorf, H. Liebe, S. Naumov, S.Stojanovic, R. Heller, L. Wennrich, M. R.Buchmeiser, Chemistry A European Journal 13(2007) 8522.

[4] D. Buerle; Laser Processing and Chemistry,Springer Berlin Heidelberg, 3rd rev. (2000).

[5] G. Hennig, K.H. Selbmann, S. Brning; LTJ, No.3,5, (2008) 52.

[6] A. Serafetinides, M.I. Makropoulo, C.D.Skordoulis, A.K. Kar; Applied Surface Science 180(2001) 42.

a)

b)

Figure 7: SEM-micrograph of pyramidal shaped cells(perspective 45 tilted), structured by ultra-shortpulse UV laser (laser fluence 3.5 J/cm, pulse overlap75%); a) section of the wedge, b) edge of a cell.

Figure 8: Optical-micrograph of a wedge (pyramidalcells, 5-tone grey scale), printed by a laboratoryprinting unit, using cyan ink and HWC-paper.

Reports

25

Quantum-chemical modelling of primary processesS. Naumov

IntroductionFor a better understanding of ongoing reactions,the reactivity and other properties ofintermediates must be known. Here, quantumchemical calculations can be of a considerablehelp. Polymerisation reactions, for example,may be induced upon electron- and photo-excitation or by free radicals. Ad initio, densityfunctional theory and other calculations allow usto compute molecular geometries and energiesof excited states and of intermediates such asneutral radicals, radical cations and radicalanions. The interpretation of ESR spectra thatmay have been obtained in the course of thereaction can be assisted by calculating thecoupling constants of such intermediates. Withthe increasing complexity of the systems thatare investigated, programs capable of dealingwith larger molecules at high level of theory atreasonable computation times, e.g., Jaguar(Schrdinger), have to be introduced. Oftenquantum-chemical calculations are carried out inthe gas phase. This may lead to erroneousresults when the reaction has been carried outin highly polar solvents such as water, especiallyin the cases of charged molecules. Hence,solvent polarity has always been taken intoaccount in the simulations to be reported.

ExperimentalReactions of excited states

The initiation mechanism of the VUV-inducedconversion of polyorganosilazanes into methylSiOSi networks was studied by means ofmodel disilazane compounds. Quantum-chemicalcalculation of the various potential decay routes,starting from excited state both singlet andtriplet (Fig. 1) allowed rationalizing theexperimental observations [1].

Even though the smallest dissociation Gibbs freeenergy was calculated for the SiCH3 bondcleavage, SiNH bond scission is found to be thepreferred photochemically induced reactionpathway in condensed media, which can berationalised by the favourable change of electrondistribution along the SiNHSi bonds uponexcitation.

Quantum-chemical calculation of the variouspotential decay routes of acrylate formulations,starting from excited triplet state formed aftershort-wavelength vacuum UV (VUV) irradiation.The short-wavelength VUV irradiation ofacrylates results in radical formation and self-initiation of the photopolymerization, i.e.photoinitiator-free curing of acrylates [2]. TheUV-Vis excitation spectra were calculated at TDDFT method und shows very reasonableagreement with experiment (Fig. 2).

Figure 1: Quantum chemical calculations on the photolytic excitation of 1,1,3,3-tetramethyldisilazane (TMDSz)and possible fragmentation pathways.

Reports

26

After photoexcitation, the first excited singletstate of 1- and 2-thio-naphthols decay byradiationless internal conversion andintersystem crossing with pronounced S-Hphotodissociation. Fluorescence as adeactivation channel plays a minor role.

To understand the influence of a structuralvariation between the parentthiophenol and itslarger aromatic thionaphthol moiety on theirfirstexcited singlet states deactivation, the excitedstate energy properties (singlet and triplet) of 1-and 2-NpSH and their corresponding structuresas well as those of the parent thiophenol werecalculated using DFT method at the B3LYP/6-31+G(d,p) level [3]. From the comparisonbetween the molecular orbitals of differentaromatic thiols given in Figure 3 one canobserve that the extended aromatic moiety ofthionaphthols increases the -electrondelocalization and consequently lowerssignificantly the excited-state energies.Consequently, increasing -electrondelocalization ought to stabilize the S1 electronicstate of thionaphthols (nanosecond time scale)and cause a remarkable spectral red shift (about50-60 nm) of their fluorescence in comparisonwith those of the parent thiophenol molecule(ArSH(S1) in subpicosecond time scale). Thesedata are in line with our steady-stateexperimental results.

Formation and properties of free radicals

The OH radical, the solvated electron, eaq, andthe hydrogen atom, H, are the primary radicalsin the radiolysis of water. In basic solution, OHis present as O [pKa(

OH) = 11.8]. Of these,only OH does not react with O2. The reason forthis has now been elucidated by quantum

mechanical calculations, and based on thethermo-dynamic cycle (Fig. 4) the pKa value ofHO3

could be established at 2.0 [4].

Although O2 is widely used a scavenger of freeradicals, a fast and irreversible reaction occursonly with a very limited number of free-radicaltypes, notably the abundant group of carbon-centered free radicals. This is in contrast toozone, which reacts with most free radicals,including oxygen- nitrogen, and most halogen-centred free radicals [5]. For the calculations ofthe standard Gibbs free energies of thesereactions, the above mentions Jaguar programhas been used with advantage.

The OH radical is also generated in the reactionof ozone with OH and with HO2. The presentmechanistic concept has been revised base onthermokinetic and quantum-chemicalcalculations [6, 7]. In a pulse radiolysis studywith -OH-, -CH3-, or -NH2-substituted indolechalcones and hydroxy benzenoid chalconesinvolving OH and other oxidizing radicals, theobserved intermediates by optical detectionwere characterized by quantum-chemicalcalculation of the spectra of various conceivableintermediates [8]. A experimental and quantum

Figure 3: Electron-density distribution of the molecularorbitals for thiophenol and 1- and 2-thionaphthols inthe most stable structure.

140 160 180 200 220 240 260

wavelength [nm]

sig

nal

inte

nsi

ty[a

.u.]

TPGDA

acrylatescalc

172 nm UV

140 160 180 200 220 240 260

wavelength [nm]

sig

nal

inte

nsi

ty[a

.u.]

TPGDA

acrylatescalc

172 nm UV

Figure 2: Comparison of the UV spectrum oftripropylene glycol diacrylate (TPGDA) and a DFTcalculated spectrum of an acrylate with the emissionspectrum of the monochromatic 172 nm excimerlamp.

- H+

OH + O2 HO3

O + O2 O3

- H+

Figure 4: Thermodynamic cycle involving O2,

OH/O

and HO3

/O3

Reports

27

chemical study has been performed on thereactivity, formation and properties of transientsgenerated in the reaction of selected organicselenides with OH, O, eaq and

H in aqueoussolution [9].

In case of the neutral adduct radical Me2SeOH

the con-version into the three-electron bondeddimer species, Me2Se...SeMe2 proceeds, in part,via the molecular (Me2Se...OH2)+ radical cation.DFT calculations also revealed a generally higherthermodynamic stability of the Se-centeredradicals relative to the S-centered ones.

Properties of radical cations and electrontransfer reactions in condensed media havebeen reviewed [10-12]. As main argumentsresulting from the quantum-chemicalconsiderations the following points might bederived: -bending motions around the Ar-XHaxis (where X = O, S, Se, N) result in acontinuous change of the molecular structure.This concerns the electron distributions as wellas the geometry of the molecule. Within theBorn-Oppenheimer approximation the moleculargeometry remains stiff during the time requiredfor electron relaxation and he very fast(instantaneous) electron jump. The heteroatoms carry lone electron pairs and are thuspotential electron donors. In the planarstructure, due to the strong resonance with the-electron of the aromatic ring, the n-electrons

are shifted from the hetero atom to the aromaticmoiety. Hence, the (HOMO) molecular orbital inthe planar structure is strongly delocalized overthe whole molecule. Rotation around the Ar XHaxis disturbs this coupling between ring and loneelectron pairs. Therefore, the molecular orbitalin the perpendicular structure assumes n-symmetry and is almost entirely localized at thehetero atom (Fig. 5). The consequencesemerging from all the above show up in theinitial products of the FET involving the twoborder line conformer structures, namely, theplanar and the perpendicular (twisted) ones.

General chemical kinetics

In many aspects, OH radicals and ozone havesimilar properties. They are both stronglyelectro-philic, and in their reactions witharomatic compounds they both form adducts.The logarithm of adduct formation relates to the(calculated) standard Gibbs free energy ofadduct formation (Fig. 5) [13].

A similar correlation with the HOMO (Fig. 6) isthe quantum-chemical basis of this dramaticsubstituent effect.

Also for other ozone reactions, quantum-chemical calculations were essential forunderstanding mechanistic details [14, 15]. Inone case [15], the program indicated the pathtaken during structure optimization. Prior to

O H r o t a t io n , d e g r e e

- 9 0 - 6 0 - 3 0 0 3 0 6 0 9 0

L ( O H ) , A n g s t r o m

0 ,8 1 ,0 1 , 2 1 ,4 1 ,6 1 , 8

Figure 5: Potential diagrams describing the energetic situation of phenol as donor in the ground state and for thedifferent conformer radical cations. The left part shows the potentials dependent on the rotation angle, and theright part illustrates energy changes dependent on the bond length (L) of Ar-OH.

Reports

28

this, the reaction even had not been considered.

For the quantum chemical study of the systemscontaining the heavy atoms and the transitionmetals complexes, new hybrid meta exchange-correlation functional (M06, M06-2X etc) withthe basis sets, which uses different models ofthe relativistic effective core potentials (LACVP,Lanl2DZ etc.) for the inner core electrons andtreats the outer core and valence electrons witha 4s/4p/2d/2f basis set, will be applied.

Mass and energy spectra of negative andpositive ions in magnetron sputtering dischargeshave been investigated with an energy-dispersive mass spectrometer. A variety oftarget metal (Cu, In, and W) containing negativeand positive molecular ions were found in thedischarge. Calculations of bond dissociationenergies were required for understandingplasma processes [16, 17]. These energies werecorrelated to the electron affinity and the bondstrength of the molecules which have beencalculated by density functional theory. Theoccurrence of the different ions is explained in

the context of their bond strengths obtainedfrom DFT calculations.

The synthesis and characterization of water-soluble dispersions of Ag nanoparticles by thereduction of AgNO3 using tryptophan underalkaline synthesis conditions are reported [18].The binding of thryptophan to silvernanoparticles was studied both experimentaland by quantum chemical calculations. Ourresults suggest that the replacement of the BH4 ions adsorbed on the nanoparticle surface bytryptophan destabilizes the particles and furthercaused aggregation. A mechanism is proposedfor the formation of silver nanoparticles bytryptophan. The Ag nanoparticles werecharacterized by UVvis absorption, dynamiclight scattering and transmission electronmicroscopy techniques.

Literature[1] W. Knolle, L. Wennrich, S. Naumov, K. Czihal, L.

Prager, D. Decker, M.R. Buchmeiser, Phys. Chem.Chem. Phys. 12 (2010) 2380.

[2] F. Bauer, U. Decker, S. Naumov, C. Riedel, Progr.Org. Coatings 69 (2010) 287.

[3] Y.M. Riyad, S. Naumov, B. Abel, R. Hermann, J.Phys. Chem. A 115 (2011) 718.

[4] S. Naumov, C. von Sonntag, J. Phys. Org. Chem.24 (2011) 600.

[5] S. Naumov, C. von Sonntag, Environ. Sci.Technol. 45 (2011) 9195.

[6] G. Mernyi, J. Lind, S. Naumov, C. von Sonntag,Chemistry Eur. J. 10 (2010) 1372.

[7] G. Merenyi, J. Lind, S. Naumov, C. von Sonntag,Environ. Sci. Technol. 44 (2010) 3507.

[8] P. Gaikwad, K. I. Priyadarsini, S. Naumov, B. S.M. Rao, J. Phys. Chem. A 114 (2010) 7877.

[9] Th. Tobien, M. Bonifai, S. Naumov, K.-D.Asmus, Phys. Chem. Chem. Phys 12 (2010) 6750.

[10 O. Brede, S. Naumov, In: Recent Trends inRadiation Chemistry, J. F. Wishart, B.S.M. Rao(Eds.) World Scientific Publishing, Singapore(2010) pp. 411.

[11]O. Brede, S. Naumo, Chem Soc Rev 39 (2010)3057.

[12]O. Brede, S. Naumov, In: Charged Particle andPhoton Interactions with Matter - RecentAdvances, Applications, and Interface, Y. Hatano,Y. Katsumura, A. Mozumder (Eds.), (2010),Taylor & Francis, Boca Raton, pp. 237.

[13]S. Naumov, C. von Sonntag, Ozone: Sci. Eng. 32(2010) 61.

[14]S. Naumov, G. Mark, A. Jarocki, C. von Sonntag,Ozone: Sci. Eng. 32, (2010) 430.

[15]G. Mark, S. Naumov, C. von Sonntag, Ozone: SciEng 33 (2011) 37.

[16]Th. Welzel, S. Naumov, K. Ellmer, J. Appl. Phys.109 (2011) 073302.

[17]Th. Welzel, S. Naumov, K. Ellmer, J. Appl. Phys.109 (2011) 073303.

[18]J. A. Jacob, S. Naumov, T. Mukherjee, S. Kapoor,Colloid Surf. B: Biointerfaces 87 (2011) 498.

-20 -10 0 10 20

-2

0

2

4

6

8

10

exergonic endergonic

PhenolMethoxybenzene

Methylbenzene

BenzeneChlorobenzene

Nitrobenzene

1,3,5-Trimethoxybenzene

Phenolate ion

log(k

)

G0(ozone adduct formation) / kcal mol

-1

Figure 6: Correlation of the calculated standardGibbs free energies with the logarithms ofexperimental rate constants of the reactions ofozone with some aromatic compounds inaqueous solution.

-2 0 2 4 6 8 10-8.0

-7.5

-7.0

-6.5

-6.0

-5.5

-5.0

Nitrobenzene

Chlorobenzene

Benzene

Methylbenzene

Methoxybenzene

Phenol

Phenolate ion

HO

MO

/eV

log(k)

Figure 5: Correlation of the logarithm of the rateconstant with the energy of the HOMO.

Reports

29

Figure 2: Change of water contact angles aftermodification of PES (homemade), PVDF, PSf, andPAN (commercial pre-hydrophilized) membranes withmolecules 1-15. The water contact angles of theunmodified membranes are: PES: 76; PVDF: 67,PSf: 62; PAN: 52.

Membrane hydrophilization using electron beamand plasma techniques

A. Schulze, A. Boulares-Pender, M. Went, I. Thomas, B. Marquardt, A. Prager

IntroductionToday, micro- and ultrafiltration membranes arepredominantly fabricated from syntheticmembrane materials such as polyethersulfone(PES), polyvinylidene fluoride (PVDF),polysulfone (PSf), or polyacrylonitrile (PAN). Thediversity of applications requires themodification of the polymer in order to avoidany fouling of the hydrophobic membranesurfaces or to achieve a functionalizedmembrane.

Here, we present two different methods usingelectron beam irradiation (EB) [1-3] andcombined plasma and EB treatment forpermanent hydrophilization of polymermembranes. No catalysts, photoinitiators,organic solvents or other toxic reagents wereused, and no additional synthetic or purificationsteps are required.

Electron Beam ModificationMembrane modification was performed bydipping the membrane into an aqueous solutionof one of the functional molecules 1-15(Figure 1) followed by EB irradiation (dose:30-300 kGy). The modified membrane wasrinsed with water for 1.5 h and subsequentlydried at 20-100 C.

ResultsModification of hydrophobic polymer membranes(PES, PVDF, PSf, PAN) with one of the reagents1-15 resulted in considerably improved water

wettability which was observed by water contactangle measurements (Figure 2).

Water contact angles of homemade pure PESmembranes resulted in contact angle changes> 20 after modification. Commercial pre-hydrophilized PVDF, PSf, and PAN membranescould also be improved in water wettability butsince the respective starting contact angle wasalready small the effect was not as strongcompared to the pure PES membrane.

Fouling tests were performed by filtration of BSAsolution (1 g/l in PBS buffer at pH 7.0) throughthe membrane followed by back-washing withpure water. This treatment was repeated severaltimes resulting in significantly decreased flux inthe case of the unmodified PVDF membrane(Figure 3). Modification with several of the smallmolecules resulted in improved antifoulingmembrane properties. Especially aftermodification with molecules 12, 14, and 15(glucose, phosphocholine, taurine) a strongpositive impact on the resulting membraneperformance was observed.

This effect was also confirmed by SEMmeasurements after 5 filtration cycles(Figure 4). Here, it can clearly be seen that theunmodified PVDF membrane is completelycovered by a fouling layer of BSA, and theformer pore structure is rarely observable(Figure 4 top). When the same treatment isapplied on a membrane modified with glucose12, the membrane pore structure is open and

Figure 1: Functional molecules used for the EB-basedmembrane modification.

Reports

30

Figure 3: Fouling properties after treatment with BSAsolution of an unmodified PVDF membrane and aftermodification with different small molecules.

Figure 4: SEM pictures of PVDF membranes afterfouling treatment with BSA solution: unmodifiedmembrane (top) and membrane modified with 12(bottom).

Table 1: Protein adsorption [g/cm] before and aftersoxhlet extraction in boiling water (7 d).

PES PVDFreagent 14 1 12before extraction 3.61 1.87 0.86after extraction 3.68 1.79 0.91

nearly no filter cake has been formed(Figure 4 bottom).

Membrane performance in terms of pure waterflux was characterized for selected modificationsby dead end filtration using pure water. Here,the modified PES membranes show a higher flux(e.g., +25% after modification with 13) than theunmodified. Similar results were obtained forthe modified PVDF membranes. The fluxgenerally increased, e.g. 16 % after modificationwith 1 compared to the unmodified PVDFmembrane. The increase in flux can be

correlated to the improved wettability of thehydrophilic membrane surface, whereby contactangles decrease at the same time.

To demonstrate the stability of the modification,selected modified PES membranes were exposedto a continuous Soxhlet extraction for 7 days inboiling water. Albumin adsorption was measuredbefore and after extraction and was found to becomparable within the limits of experimentalerror (Table 1). Similar results were found forthe corresponding contact angles before andafter Soxhlet extraction.

However, due to the low concentration of thefunctional groups in the final modifiedmembrane, it was not feasible so far tounambiguously identify the exact connectivity ofthe functional molecules to the membranes. Thehigh stability of the modification indicates apermanent membrane functionalization, i.e. thatcovalent bonds are formed between the polymerbackbone and the functional molecules.

XPS analysis showed that the electron beam-based treatment did not result in any changes inthe PES structure and no additional S-basedsignals were observed. This result wasadditionally confirmed by bubble point andMWCO measurements before and afterirradiation confirming that the pore size was notenlarged by decomposition of the membranepolymer.

ConclusionWe successfully modified different types ofpolymer membranes (PES, PVDF, PSf, PAN)soaked in aqueous solutions of functionalmolecules and treated by EB. Furthermore, wehave shown that the modification method seemsto be generally applicable on different types ofpolymer membranes. With selected functionalmolecules, this simple modification procedureresulted in significantly improved waterwettability as well as reduced fouling behaviour.The procedure guarantees for a permanentfunctionalization of the membrane polymer thatneither requires the use ofcatalysts/photoinitiators nor of other toxicreagents. Further experiments are currently in

Reports

31

Figure 5: SEM images of PVDF membrane samples:unmodified (left), oxygen plasma treated (middle)and oxygen plasma followed by electron beamexposure (right).

Table 2: Water and diiodomethane (DIM) contactangles of the modified PVDF membrane.

ContactAngle

SurfaceEnergy Contributions

Water[]

DIM[] [mN/m] Disp. Polar

PVDF 138.3 77.8 22.1 18.6 3.5EB 128.6 67.8 26.8 24.1 2.7

N2 137.5 90.4 14.0 12.5 1.5N2-EB 110.8 80.9 17.2 17.0 0.1

N2/Ar 137.3 94.4 11.9 10.8 1.1N2/Ar-EB 126.5 77.7 19.8 18.7 1.1

Ar 132.9 100.0 8.9 8.7 0.2Ar-EB 129.6 89.3 13.6 13.0 0.6

O2/Ar 126.0 101.3 8.2 8.2 0.6O2/Ar-EB 102.2 85.0 16.7 15.0 1.7

O2 65.1 82.8 37.0 16.1 20.9O2-EB 68.9 71.0 36.1 22.3 13.8

O2/N2 127.4 72.6 23.3 21.4 1.9O2/N2-EB 116.8 77.4 19.0 18.9 0.1

progress to shed light on the mechanism, i.e.the reactive intermediates.

Combined Plasma and ElectronBeam TreatmentPlasma modification is a powerful method forhydrophilization of polymer surfaces. However,after common plasma treatment hydrophobicrecovery is very often observed leading toincreasing contact angles within several days oreven weeks.

In the present study, plasma modification wasfixed by subsequent electron beam (EB)