Delft University of Technology

Toward high volume solution based roll-to-roll processing of OLEDs

Abbel Robert de Vries Ike Langen Arjan Kirchner Gerwin trsquoMannetje Hero Gorter Harrie WilsonJoanne Groen PimDOI101557jmr2017204Publication date2017Document VersionFinal published versionPublished inJournal of Materials Research

Citation (APA)Abbel R de Vries I Langen A Kirchner G trsquoMannetje H Gorter H Wilson J amp Groen P (2017)Toward high volume solution based roll-to-roll processing of OLEDs Journal of Materials Research 32(12)2219-2229 httpsdoiorg101557jmr2017204

Important noteTo cite this publication please use the final published version (if applicable)Please check the document version above

CopyrightOther than for strictly personal use it is not permitted to download forward or distribute the text or part of it without the consentof the author(s) andor copyright holder(s) unless the work is under an open content license such as Creative Commons

Takedown policyPlease contact us and provide details if you believe this document breaches copyrightsWe will remove access to the work immediately and investigate your claim

This work is downloaded from Delft University of TechnologyFor technical reasons the number of authors shown on this cover page is limited to a maximum of 10

INVITED PAPER

Toward high volume solution based roll-to-roll processing of OLEDs

Robert Abbel Ike de Vries Arjan Langen Gerwin Kirchner Hero trsquoMannetje Harrie Gorter andJoanne WilsonHolst CentreTNO Eindhoven 5656 AE the Netherlands

Pim Groena)

Holst CentreTNO Eindhoven 5656 AE the Netherlands and Novel Aerospace Materials Delft University ofTechnology Delft 2629 HS the Netherlands

(Received 10 March 2017 accepted 3 May 2017)

The large volume production of flexible electronics by solution based roll-to-roll (R2R)manufacturing technologies is a promising upscaling strategy for the organic electronics industryTypical optoelectronic devices like organic light emitting diodes (OLEDs) consist of a complexstack of functional layers Solution deposition of these structures eliminates the need forexpensive vacuum processing This contribution presents approaches for solution based R2Rproduction methods of functional OLED layers on flexible polymer substrates The developmentof a R2R line with two slot-die coating stations is discussed which can deposit two uniform layersconsecutively in a single run (ldquotandem coatingrdquo) at web speeds up to 30 mmin Furthermore itoffers the unique feature that there is no contact between the rollers and the top side of thesubstrate where the functional coating is deposited Thereby an important source of particlecontamination and other damage to the device is eliminated In addition to continuous depositionstripe and intermittent coating techniques have been developed allowing the production ofpatterned layers Finally examples will be shown of OLEDs where two functional materials aredeposited by R2R processing from solution

I INTRODUCTION

Organic and hybrid optoelectronic devices like perov-skite and organic photovoltaic cells organic light emit-ting diodes (OLEDs) and the like consist of a complexstack of thin layers of functional materials1ndash3 Relativelyefficient OLEDs have already been demonstrated somedecades ago where the organic semiconducting layerswere deposited as thin films by evaporation45 A typicalOLED device architecture is shown in Fig 1 displayingthe various materials necessary for efficient charge in-jection and transport and light emission by radiativerecombination This is an example of the so-calledbottom emissive OLED which is built on a transparentsubstrate and finished with a nontransparent reflectiveelectrode so that the generated light is emitted through thesubstrate When nontransparent (eg metallic) substratesare used the light is emitted directly from the deviceinstead of through the substrate and the so-called topemissive device is obtained All the work presented inthis study relates to bottom emissive stacks The com-mercial industrial production of these devices eg for

the consumer electronics market is currently dominatedby vacuum based ldquodryrdquo deposition techniques likesputtering and evaporation6 Although these processesare well-established device fabrication by solutionprocessing ie by printing and coating of functionalinks has emerged as a promising alternative routetoward industrial mass production of organic optoelec-tronic devices7ndash10 For solution processing on theacademic laboratory scale spin coating is frequentlythe method of choice for the thin layer deposition due toits ease of application and its good control over the layerthickness and homogeneity10 For industrial productionhowever it suffers from its lack of compatibility withhigh volume manufacturing and its inefficient use ofmaterials Research aimed at results more directlyapplicable in an industrially relevant production envi-ronment therefore focuses on other techniques like slot-die coating811ndash14

This approach exhibits its own specific challenges notencountered in vacuum processing but there are alsodistinct benefits associated with it which for certainapplication areas show good prospects for replacing drydeposition technologies on the mid-term Many of thesebeneficial characteristics are related to the possibility ofprocessing at atmospheric pressure with more efficientmaterials use high throughput and at low overall

Contributing Editor Gary L Messinga)Address all correspondence to this authore-mail pimgroentnonl

DOI 101557jmr2017204

J Mater Res 2017 Materials Research Society 2017 This is an Open Access article distributed under the terms of the

Creative Commons Attribution licence (httpcreativecommonsorglicensesby40) which permits

unrestricted re-use distribution and reproduction in anymedium provided the originalwork is properly cited

1

httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

production cost15 Several recent publications have dem-onstrated that slot-die coating can be used for the pro-duction of highly efficient OLED devices by solution basedtechniques11ndash14 These studies however have all made useof sheet-to-sheet (S2S) processing An especially interest-ing aspect of slot die coating and several other solutionprocessing methods is production upscaling by roll-to-roll(R2R) manufacturing techniques16ndash18 R2R slot-die coatingis a common technology which has been used for a longtime already for the production of photographic films andcoated papers It offers a way to apply a wet layer ona substrate in a very well controlled manner In the field offlexible electronics the technology is used for making theactive layers in OLEDs and organic or perovskite basedsolar cells16ndash18 Recently impressive progress has beenachieved in the field of organic photovoltaic (OPV) deviceproduction by the R2R solution based approach19ndash21

In this paper we report on our own results regarding theR2R processing of OLED devices by wet deposition offunctional inks for the hole injection and emissive layers

II RESULTS AND DISCUSSION

A Tandem R2R coating line for high yield printedelectronics production

Particle contamination is a serious problem in the pro-duction of organic electronic devices a o because thefunctional layers used are rather thin (in the order of20ndash200 nm) and any particles of a comparable or largersize pose a high risk for defects A typical source for dustparticles is top contact with the functional layers duringprocessing and sample handling eg by web handlingparts like nips or idler rollers When R2R fabricationtechniques are applied as is frequently the case in industrialproduction environments particle contamination can alsooccur by contact with the backside of the substrates duringrewinding The consequence can be malfunctional devices

decreased performances shortened lifetimes and lowerproduction yields Whereas a significant percentage ofdeficient products can frequently be tolerated in an aca-demic research environment this is absolutely prohibitivefor industrial manufacturing processes A strategy to reducethe risk of particle contamination due to backside contact inR2R processes is therefore the sequential processing ofseveral functional layers in a single run (ldquotandem R2Rcoatingrdquo) without rewinding between the coating steps Inaddition to the reduced particle contamination the relatedreduction in production costs is another benefit Its disad-vantage however is the need for extensive web handlingand thus frequent risk of surface top contact In addition toworking in a cleanroom environment a web handlingsystem which avoids top contacting while still enablingsequential coating is therefore a prerequisite for success

Whereas for the production of a complete organic solarcell or OLED by R2R solution processes at least threebut more typically five or even more sequential coatingsteps will be necessary the general concept can alreadybe demonstrated and investigated by two sequential stepscarried out in tandem A schematic drawing and photo-graphs of the system used in this study are shown inFig 2 which employs a novel concept that allows thetransport of the carrier substrate through two sequentialcoating stations and drying ovens including several 90degand 180deg turns without ever making contact to the topsurface To save cleanroom space only the winder re-winder and coating stations are located in a cleanroomenvironment whereas the dryers are located outside Thisseparation of the process parts between cleanroom andnoncleanroom areas drastically reduces the footprint ofthe needed cleanroom area compared to the total processarea In addition to the corresponding cost savings thisalso allows relatively easy control of the cleanroomenvironment and particle density Inside the ovens theuse of parts which move or touch the web on the rear sideis limited to reduce sources for particle generation Initialparticle level measurements (without a moving web)showed a cleanliness level between ISO 3 and ISO 4according to the cleanroom standard ISO 14644-1 Ifnecessary for the drying of air-sensitive materials theovens can be operated under nitrogen atmosphere withan oxygen concentration of below 80 ppm

B Patterned deposition of functional layers fromsolution

In addition to their sensitivity to dust particles thefunctional layers of organic electronic devices also needto be protected carefully against water and oxygeningression which quickly degrade many of the materialsused in them Whereas encapsulation of the top andbottom surfaces can be rather easily achieved by the useof barrier technologies22 ingression via the sides is

FIG 1 (a) A schematic cross-section of the OLED device stack used inthis study (b) A schematic cross-section of a bottom emissive OLEDdevice with unpatterned functional layers (c) Corresponding image fora device with patterned functional layers sandwiched between a top andbottom barrier Water ingression via the side is only possible for (b)

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

J Mater Res 20172httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

frequently a serious challenge This is especially relevantfor production processes that employ homogeneouscoating techniques with only very limited options forlocally selective deposition of the functional layers as isthe case for eg R2R slot-die coated OLEDs Once a rollof finished products is ready the individual OLEDmodules must be cut from it to achieve the requireddimensions As illustrated in Fig 1(b) this will inevitablyresult in the unpatterned coated layer becoming exposedto the ambient environment at the cut edges therebyopening an ingression route for moisture and oxygenThe consequence will be the immediate start of devicedeterioration via side leakage unless efficient sealing atthe device edges can be achieved Any option to patternthe functional layers will significantly ease an efficientencapsulation because some noncoated regions are avail-able next to the functional areas where the top andbottom barriers can be brought into direct contact witheach other [Fig 1(c)]

Whereas in the case of small scale production forresearch purposes especially when using sheet-to-sheet(S2S) techniques like spin coating the partial removal offunctional materials is commonly achieved by wipingthis is obviously no option in an industrial productionenvironment Selective removal of the dried layers bylaser ablation is a more viable route which has beendemonstrated in the case of OPV cell production23 but itis typically rather slow and creates debris particles whichcan redeposit on the functional area and cause problemswhen the following layer is coated Furthermore pre-patterning of the substrate by localized plasma treatment

to induce self-assembly of the coated layers is a possiblesolution24 but no satisfying results have been obtainedup to now in the field of OLED production The moststraightforward option is therefore to pattern the func-tional layers directly during the coating process ie toachieve locally selective deposition This removes theneed for the removal of excess material afterward in anadditional processing step and also avoids the risk ofcontamination with debris particles

C R2R stripe coating

Due to its nature as a basically homogeneous de-position technology slot-die coating as opposed to eginkjet printing is not suited to create complex patterns ina variety of shapes Rather by partially restricting the inkflow in the coating direction and opposite to it it allowsthe production of simple patterns like stripes and rec-tangles In many cases however these limited patterningoptions are already sufficient to prepare functional coat-ings that can be efficiently encapsulated to yield deviceswhich are stable against degradation by side leakage Asthe techniques to achieve patterned slot-die coating varyconsiderably depending on whether patterning in oropposite to the web direction is the goal they will inthe following be treated separately

For patterning in the coating direction stripe coating isa rather common method25 In conventional stripe coat-ing the ink flow through the slot-die onto the substrate ispartially prevented by a shim which is mounted into theslot gap of the device as shown in Fig 3(a) As the inkcan exit the slot-die only at the locations where its flow is

FIG 2 (a) A schematic drawing of the tandem coating line comprising two subsequent coating stations in one process (b) A schematic drawing ofthe web pass in the process which allows substrate handling without ever touching the top side (c) A photograph of the cleanroom part of the lineshowing the coating stations (d) A photograph of one of the drying ovens outside the cleanroom environment

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

3J Mater Res 2017httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

unobstructed by the presence of the shim the result isa stripe pattern the dimensions of which are determinedby the shim geometry This approach however has thedisadvantage that lateral wetting on the coating lip willoccur by capillary forces leading to a loss of control overthe exact width of the stripes in the pattern The resultingvalue of the stripe width will be larger than defined by theshim but the exact deviation is determined by a complexinteraction of many factors and therefore inherentlyunpredictable26 A solution to this problem as imple-mented in the coating line described in the present studyis the use of notched dies as is shown in Fig 3(b) In thisdesign the die lips are notched in accordance with thedesired stripe pattern and thus lateral spreading of thecoating solution on the lips is suppressed In this way thedeposited stripe widths correspond very closely to thevalue determined by the notch design as can be seen inFig 3(c)

Table I presents the results of a stripe coating studyusing various test fluids one based on water and one ona nonpolar organic solvent (xylenes) with differentrheological properties at two coating speeds The notcheswere prepared in such a way that patterns of stripesalternating with noncoated interspaces were producedThree different test patterns were examined comprisingparallel stripes with 9 8 and 7 mm design widths anda spacing of 1 2 and 3 mm respectively The actualmeasured widths obtained using these notched dies wereonly slightly larger than the targeted design width the

maximum deviation being 5 This good agreementbetween the designed stripe width and the experimentallyobtained values indicates the virtual absence of lateralspreading due to the capillary effect

As can be seen in Fig 3(d) stripe coating usingnotched slot dies also works with actual functional OLEDmaterials (emissive layer [EML] on hole injection layer[HIL] in this specific case) where another test patternwith broader stripes was used In agreement with theexperiments done using the test fluids also in this casegood agreement between target and obtained widths wasobserved (deviation 16)

D R2R intermittent coating

Whereas patterned slot-die coating in the web direc-tions (stripe coating) is rather straightforward patterningperpendicular to the coating direction is more compli-cated In this case the flow of the ink onto the substratedoes not simply need to be confined locally Instead itactually must be stopped and started in a controlledmanner without causing ill-defined edges by start or stopeffects This controlled interruption of the coating processis known as intermittent slot-die coating and has beendescribed for the production of lithium ion batteries27 Inthis particular case the thixotropic behavior of thecoating slurry in combination with its relatively highviscosity enabled intermittent coating to be initiatedsimply by stopping and starting the ink flow When wetried to transfer this process to the rather low viscosity

FIG 3 Options for R2R stripe coating (a) Slot-die with shim (b) Notched slot-die designed for accurate stripe coating and a detail of the notcheddie lip (inset) (c) Well-defined stripe coating using a notched slot-die Stripe pattern as defined in Table I total coating width 240 mm (d) Stripecoating of EML ink on the HIL under UV illumination

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

J Mater Res 20174httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

coating solutions typically used in the field of the solutionprocessing of OLED materials (1ndash10 mPa) the resultswere disappointing Under these circumstances thecapillary forces are dominant over the shear stress inthe coating bead The result of an initial coating exper-iment using an ink for the HIL is displayed in Fig 4(a)Upon starting of the coating flow an irregular patternwith streaks was observed which only stabilized intoa uniform closed layer after several centimeters awayfrom the starting position Similarly the interruption ofthe coating flow resulted in a gradual thinning of thecoated layer before final break-up of the trailing edgeswas achieved Again this process took several centi-meters until finally no ink was deposited Obviously thisis a situation which does not allow the reproducibleproduction of OLED devices

A much more desirable coating behaviour was obtainedwhen in addition to the start or stop of the ink flowa coating unit was developed in which the entire slot-diewas moved close to or away from the substrate [Figs 4(b)

and 4(c)] In this design the coating gap can be widenedfrom the coating position (50ndash100 lm away from the web)to a maximum distance of 4 mm within 25 ms In practicea stoke of 1 mm turned out to be sufficient to reliably breakup the coating bead and to enable well-defined coatingedges The process and its result are shown in Figs 4(d)and 4(e) where for reasons of visibility and to demonstratethe general applicability of the method also to nonpolarformulations a xylene-based test liquid containing a dyeis used In addition this liquid serves as a test model forthe EML ink which itself is also a solution of the emitterpolymer in xylene The vibrations induced by the highspeed mechanical motion of the slot-die were expected tobe carried forward into the stream of ink onto thesubstrate and thus to cause regular variations perpendic-ular to the coating direction known in literature asbarring28 To compensate for the movement and tosuppress these vibrations counter balance masses whichmove in the opposite direction as the retraction motion ofthe slot were included in the design In addition the

TABLE I Widths of spacings between stripes coated using notched slot-die lips with different designs (Agfa V109 PET film coating gap 100 lmwet coated layer thickness 10 lm polymer concentration 10 gL)

Coatingsolution

Surface tension(mNm)

Viscosity(mPa)

Coating speed(mmin)

Measured stripe width (mm) (relative deviation from designed target width)

Designed stripe width 90 mmwith 10 mm spacing

Designed stripe width 80 mmwith 20 mm spacing

Designed stripe width 70 mmwith 30 mm spacing

XylenesPS 300 10 50 93 (133) 83 (138) 72 (129)WaterPVA 280 20 50 92 (122) 83 (138) 72 (129)XylenesPS 300 10 100 91 (111) 81 (113) 73 (143)WaterPVA 280 20 100 93 (133) 84 (150) 71 (114)

FIG 4 R2R intermittent slot-die coating (a) Leading coating edge and trailing coating edge for an HIL ink during conventional intermittentcoating (without moving die) (b) A schematic representation of the principle of intermittent slot-die coating using a moveable die (c) A schematicdrawing of the moveable intermittent slot-die coating station (d) Intermittent coating of a test solution using this equipment (e) Leading coatingedge and trailing coating edge for this test ink and (f) for the HIL ink during intermittent coating using the moveable intermittent slot-die station

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

5J Mater Res 2017httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

moving parts of this process are mounted on leavesprings to further dampen any residual vibrations Takentogether these measures enabled us to significantlyimprove the straightness and thickness homogeneity ofthe leading and trailing edges during intermittent slot-diecoating as is displayed in Fig 4(f) for the HIL ink

E R2R coating of functional OLED layers

To arrive at a process allowing device coating underoptimized conditions the deposition procedures for eachof the functional inks were studied individually first TheHIL formulation used is based on a mixture of severalsolvents As a consequence it has a low surface tension(30ndash31 mNm) and thus spreads and wets well on the testsubstrates (pure poly(ethylene terephthalate) [PET] andpoly(ethylene naphthalate) [PEN] and PET coated withindium tin oxide [ITO]) (wetting angles 10deg) On theother hand good contact line pinning is also observedmeaning that the ink reliably stays on the coated areawithout spreading beyond To achieve dry HIL filmthicknesses of 30 nm or 60 nm at a line speed of5 mmin it was found that wet films in the order of8 lm or 16 lm respectively needed to be deposited Atfirst sight the resulting dry coatings looked uniform byvisual inspection but as the layers are thin and highlytransparent this is difficult to qualify reliably by the unaidedeye Indeed closer examination using contrast enhancement(by placing a white paper below the sample and increasingthe HIL thickness to 90 nm) revealed a slight mottle orspottiness within the layers [Figs 5(a)ndash5(d)] Using suchHILs in complete OLED devices revealed that this mottle

effect in the HIL even though hardly visible in the layeritself had a significant influence on the homogeneity ofthe light output in full OLED devices [Fig 5(e)] Asmottle is frequently caused by disturbances in the wetlayers caused by the air flow during drying29 theconditions in the conveyor oven were systematicallyvaried to study if the mottle could be reduced this wayAs is shown in Figs 5(a)ndash5(d) however it was found thatonly very little improvements were achieved in thismanner Another suspected reason for the observed mottlein thin coated films can be an inherent inhomogeneityalready present on the base substrate surface used Indeedonce the normally used PEN (Teonex Q65AH) or PET(Agfa V109) substrates were replaced by another com-mercial product (DuPont Teijin Films Melinex ST504)completely transparent coatings were obtained withoutany detectable mottle over a large area [Fig 5(f)]

The ink for the EML is based on a green light emittingpolymer dissolved in xylenes This results in a contactangle close to zero on plastic substrates coated with theHIL The ink can be easily coated on top of the HIL usingR2R techniques as is shown in Figs 6(a) and 6(b) [EMLmade visible in (a) by UV illumination] and does notremove or disrupt the underlying layer The control of thecoating width was not as good as with the HIL which isa direct consequence of its good wetting properties onboth web and slot-die lips During coating the ink isdrawn into the capillary gap between the side plates ofthe slot-die and the substrate This leads to a coatingwidth which is actually wider than expected solely basedon the slot-die design (200 mm) as shown in Fig 6(b) Itwas found that this occurred always regardless of the

FIG 5 (andashd) Optical photographs of the R2R coated HIL on the Teonex Q65HA transparent film under different drying conditions (e) Visualappearance of OLED emission using a mottled HIL (f) The coated HIL on Melinex ST504 In (andashd) and (f) the coated HIL visibility is improvedby contrast enhancement

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

J Mater Res 20176httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

web speed during coating and was particularly pro-nounced at low coating speeds [Fig 6(c)] In line withexpectation the effect also tended to aggravate withincreasing wet layer thickness Although an increase inthe distance between the slot-die and the web looks likean obvious solution to this problem its effect was foundto be only of a minor influence To suppress thisphenomenon a change in the slot-die design was neededin a manner completely analogous as described in themethod applied during stripe coating using notchededges (vide supra) The analysis of the dry thicknessesof both the HILs and the EMLs showed a good agree-ment with the targeted values and also displayed littlevariation within a sample of 80 80 mm2 (correspond-ing to the part which completely comprises the activeareas of all three R2R model devices [vide infra]) As canbe seen in Fig 6(d) which shows the measured layerthickness in the cross web direction both coatings appearwell on the target

When multiple layers are deposited on top of eachother from solution partial or even complete redissolu-tion of the underlying material during coating of the nextlayer can be a serious challenge Especially if the baselayer is not crosslinked as is the case with the HIL usedin this study this can lead to intermixing which typicallyhas negative influences on the device performanceHowever the very different polarities of the HIL andthe EML enabled us to use an orthogonal solvent(xylenes) which is a good solvent for the EML but

leaves the HIL unaffected By this choice of materialswell-defined layers without intermixing can be obtainedIt should be mentioned however that more complexstacks involving chemically more similar functionalmaterials will put restrictions on this approach In thesecases fixation by crosslinking before deposition of thefollowing layer is typically the method of choice

F OLED devices prepared by R2R coating

After the coating of both functional layers (HIL andEML) needed for a hybrid (partially solution processed)OLED device had been satisfactorily demonstrated thenext step was to apply these materials by R2R techniquesalso in actual functional devices For comparison iden-tical devices were also prepared purely using the S2Sslot-die coating equipment Drying of the S2S layers wascarried out on a hot plate (in a glovebox in the case of theEML) Thickness measurements revealed that within theexperimental error both HIL and EML had identicalthicknesses in the devices prepared by S2S and R2Rprocesses As at current state the R2R production ofplastic films with a patterned anode (structured ITO andmetallic bus bars) still poses an insurmountable chal-lenge all base substrates of 152 152 mm2 size wereprepared by S2S lithographic techniques One such sub-strate sheet contains three rectangular OLED deviceseach of which has an active area of 15 80 mm2 For thefurther S2S processed references HIL and EML were

FIG 6 (a) Coating of EML fluid visualized by photoluminescence under UV illumination Coating width 260 mm (b) A sample of the coatedHIL and EML on PETITO at an identical slot-die width (coating speed 5 mmin coating gap 250 lm) (c) The coated width of EML ink on V109PET as a function of the wet layer thickness for different line speeds (d) The measured coating thickness of the HIL and EML on V109 PET withinan area of 80 80 mm the target thickness for the HIL was 30 nm and for the EML 60 nm

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

7J Mater Res 2017httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

deposited sequentially by a S2S slot-die coating machineand dried on a hot plate The devices were then finishedby S2S evaporation of an electron injection layer anda cathode (BaAl) and sealed by a thin film encapsulationstack deposited by plasma deposition of silicon nitrideand an inkjet printed organic planarization layer

For R2R coating of the HIL and EML the base anodesubstrates were spliced together with long pieces of a barePET film acting as a leader and trailer resulting in a rollwith a total length in the order of 500 m This roll wasthen mounted on the R2R coating unit and overcoated ina continuous manner first with the HIL and subsequentlywith the EML Both layers were dried immediately aftercoating in a long hot air oven [Figs 2(a) and 2(d)] andrewound after coating As in our current setup oxygenand moisture levels are significantly higher (80 ppm)than in the glovebox used for the S2S experiments(1 ppm) the EML was dried at lower temperatures(75 degC for R2R as opposed to 120 degC for S2S) to limitoxidative damage This temperature is sufficient tocompletely remove all solvent Surface roughness meas-urements showed that the original value of the bare PETfilm with ITO (21 6 3 nm) was somewhat increased afterHIL (39 6 7 nm) and EML coating (39 6 4 nm) Fromthis roll consisting mainly of HIL and EML coated PETthe anode substrates were removed again and finished inan identical way as the S2S reference samples An OLEDdevice prepared by this approach using R2R coating isshown in Figs 7(a) and 7(b) Both the S2S reference

devices and the OLED prepared by R2R coating wereanalyzed with respect to their current densityndashvoltagendashluminance (IVL) characteristics [Figs 7(c) and 7(d)] It isimmediately obvious that the purely S2S prepared samplesperform significantly better in terms of the leakage current(order of magnitude 103 mAcm2 versus 101 mAcm2)and luminance (1000 cdm2 at 5ndash6 V versus 100 cdm2 at55ndash60 V) The efficiencies at 100 cdm2 by contrast werecomparable (5ndash7 cdA) for both types of devices mainlybecause the lower luminances of the R2R produced OLEDswere compensated for by the lower current densities alsoLifetime analysis revealed a higher stability for the S2Ssamples (LT50 2000 h) than for the R2R samples (LT50350 h) The higher oxygen and moisture contents in theR2R oven during drying as compared to the S2S equip-ment is an obvious source for EML degradation whichcould partially explain the lower performances of the R2Rproduced devices

Thermal imaging of both types of devices using aninfrared (IR) camera demonstrated the presence of hotspots already below turn-on voltage exclusively in theR2R prepared OLEDs [Fig 7(e)] After some time ofdriving the OLEDs these spots developed into slowlygrowing dark spots visible by the naked eye Subsequentoptical microscopic analysis revealed the presence of dustparticles in many of these spots [Fig 7(f)] It can besuspected that the presence of dust particles inducedimperfections at and around their locations leading toshort circuits which during operation degraded the

FIG 7 (a) Half-finished OLED devices spliced into a roll of plastic carrier films for R2R coating (b) A photograph of a functional OLED withR2R slot-die coated layers of the HIL and EML (c) Currentndashvoltage characteristics (solid lines) and luminance curves (dashed lines) of S2S andR2R slot-die coated OLEDs (d) Corresponding current efficiencies (solid lines) and power efficiencies (dashed lines) of these devices (e) Overlayof lock-in IR imaging at 2 V (ie below turn-on voltage) and the electroluminescent appearance of a R2R produced OLED observed after lifetimetesting (f) An optical micrograph (dark field mode OLED off) of a dust particle causing a defect

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

J Mater Res 20178httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

emissive material around them by local overheating Inaddition to the above mentioned oxidative EML degra-dation these defects are expected to contribute signifi-cantly to OLED performance deterioration and aresuspected to be a main cause for the observed differencesin device lifetimes This higher degree of contaminationin the R2R OLEDs compared to the S2S references isassumed to be a consequence of the much longerexposure of the coatings to the open environment duringdrying in the ovens Even though they constitute a clean-room environment the low density of particles presentthere is evidently still sufficient to deposit on the wetlayers in significant densities to result in a clearly lowerdevice performance

III CONCLUSIONS

Two functional layers in a bottom emissive OLEDdevice architecture (HIL and EML) have been depositedin a highly controlled and reproducible manner by R2Rslot-die coating Fully functional devices have beenprepared from these films Thereby the possibility hasbeen demonstrated to employ solution processing ina continuous fabrication mode on a scale relevant forindustrial manufacturing This achievement constitutesa decisive step towards the production of OLEDs andsimilar organic electronic devices in large volumes at lowcost Oxidative degradation during drying and contami-nation with dust particles during coating and dryingturned out to still pose significant limitations on theOLED performance as compared to S2S preparedreference devices However upon transfer from a researchto an actual industrial production environment theseissues are expected to be solvable

In addition patterned deposition of functional materi-als in the web direction (stripe coating) and perpendicularto it (intermittent coating) has been successfully demon-strated These patterning techniques offer a strategy thatenables complete OLED encapsulation and thus theprevention of premature device degradation by sideleakage of moisture and oxygen

IV EXPERIMENTAL SECTION

A Materials

All solvents were acquired from Sigma-Aldrich(Zwijndrecht The Netherlands) and used as receivedPolymer foils of 125 lm thickness have been obtainedfrom Agfa Gevaert (Melinex ST504 [PET] MortselBelgium V109 a hydrophilically coated PET) and fromDuPont Teijin Films (Teonex Q65HA [PEN] DumfriesUnited Kingdom) The HIL ink (XA-3215) was suppliedby Plexcore (Pittsburgh Pennsylvania) and is based ona mixture of nonaqueous polar solvents For slot-die

coating purposes the ink was diluted with diluentXA-3215-S (Plexcore) to solids contents of 045 wtneeded to be capable of coating dried layers in the range of30ndash60 nm The viscosity of the diluted ink is 66 mPas(determined by a WellsndashBrookfield DVLV coneplaterheometer WellsndashBrookfield Middleboro Massachusetts)and its surface tension is 30ndash31 mNm (Kruumlss EasydropHamburg Germany) The EML ink consists of a greenlight emitting polymer dissolved in xylenes (a mixture ofisomers) with a solid content of 075 wt The ink (gradeG13017X) was supplied by Sumitomo Chemical Co Ltd(Osaka Japan) and used as received The viscosity of thisink is 30 mPas (determined by a WellsndashBrookfieldDVLV coneplate rheometer)

B Substrate preparation

A barrier layer was applied on top of the substrates(size 15 15 cm2) by sheet-to-sheet plasma enhancedchemical vapor deposition PECVD of SiNx (OxfordInstruments Abingdon United Kingdom) in combinationwith igravenk jet printing of an organic planarization layer asdescribed in Ref 22 This resulted in a water vaportransmission rate of 106 gm2day ITO was depositedby RF-sputtering (KDF 744 NT Rockleigh New Jersey)to form a 135 nm thick layer with a sheet resistance of20 Xsq ITO patterning was achieved by photolithogra-phy Photoresist AZ1518 (Microchemicals GmbH UlmGermany) was spin coated and then exposed by UVillumination through a mask (Karl Suess MA6 GarchingGermany) The exposed resist was developed withAZ developer (Merck Performance Materials GmbHWiesbaden Germany) and rinsed with water to removethe illuminated parts ITO was etched in a FeCl3HClbath (035 M in 18 HCl 40 degC 3 min) Stripping of thephotoresist was done in acetone Bus bars and shuntinglines were prepared by sputtering [molybdenum (20 nm)aluminum (250 nm) and molybdenum (20 nm)] Struc-turing was achieved in an identical manner as with ITOexcept that another etchant (phosphoric acid acetic acidsulfuric acid ratio 74188 by volume) was used for metalremoval This resulted in substrates with three rectangularOLED devices each of which has an active area of 15 80 mm2

C Slot-die coating processes

S2S slot die coating was carried out using an nTactnRAD 1 (Dallas Texas) which can handle substrates ofup to 152 152 mm2 size HIL coating was done inambient atmosphere whereas the EML was depositedinside a glovebox with dry nitrogen environment (mois-ture and oxygen contents 1 ppm) Drying was per-formed on hotplates (HIL 7 min at 50 degC followed by15 min at 120 degC EML 15 min at 120 degC) in the sameenvironments as where the coating had taken place

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

9J Mater Res 2017httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

Initial R2R coating trials have been carried out ona R2R coating line (LC21 Coatema DormagenGermany) which is equipped with a slot-die unit and animpingement drying oven with three sections of 15 mlength each Final experiments including the coating ofthe OLED layers have been carried out on an advancedversion (built by VDL the Netherlands) which is pre-sented in Sec II (see Fig 2) It consists of two slot-diecoating units [TSE Troller and nTact see Fig 3(b)]coupled in one line with a common web transport systemand two impingement drying ovens with lengths of 20 mand 10 m respectively Inside the ovens zirconium oxidesensors can monitor the oxygen concentration when thesystem is operated under nitrogen atmosphere The webwidth used on both lines is 30 cm and the coating width ofthe slot-dies is 26 cm The advanced coating line candevelop speeds in the range of 3ndash30 mmin Coating andprinting steps were performed under cleanroom class10000 (ISO 14644 class 7) conditions Shims were usedto determine the coating width at 20 cm The HIL wascoated at a line speed of 5 mmin and an ink flow of 8 mLmin to achieve 30 nm of HIL thickness and 16 mLmin for60 nm thickness HIL drying was done without protectiveatmosphere at temperatures rising steadily from of60ndash140 degC in the twenty oven sections Patterning wasachieved by manual wiping The EML ink was coatedunder nitrogen at a rate of 10 mmin while dosing195 mLmin to achieve 60 nm and 26 mLmin to achieve80 nm of EML thickness It was dried at 75 degC ina nitrogen environment (water and oxygen levels80 ppm) but was shortly exposed to air after coatingand drying on its way to the rewinding station Thethicknesses and uniformities of the HIL and EML coatingswere determined using a Filmetrics F2 thin film analyzer(Filmetrics Unterhaching Germany) measuring ona spliced-in piece of a metal foil for enhanced contrast(measuring on plastic films is impossible due to the smalldifferences in refractive indices between the coatings andthe substrate material) Surface roughnesses were deter-mined by surface scans using a Bruker DektakXT profil-ometer (Bruker Corporation Billerica Massachusetts)

D Cathode and encapsulation

The cathode consisting of 2 nm Ba covered by 100 nmAl was applied by thermal evaporation using a home-built apparatus through a mask Finally the OLEDs wereencapsulated by thin film encapsulation of two SiNx

layers by PECVD through a mask separated by anorganic planarization layer printed by inkjet printingfollowing the procedure described in Ref 22

E OLED characterization

IVL curves were measured on a home-built characteriza-tion setup based on a SONY Cast Pro stage which was

equipped with a Konica Minolta Luminance Meter LS-110(Konica Minolta Tokyo Japan) a Keithley 2400 SourceMeter (Keithley Instruments Germering Germany) anda Keithley 6485 Picoammeter (Keithley Instruments) Life-time measurements were carried out under controlledenvironment (20 degC 50 relative humidity) at a constantcurrent (100 mA) Per R2R run four OLED substrates wereincluded resulting in twelve devices which were all finishedand characterized The reported values are the averages andstandard deviations of these twelve measurements

ACKNOWLEDGMENTS

The authors like to thank the VDL team An PreenenRene Aerts Huib van de Heuvel Jan van de Brink SorinStan and Auke van der Velde for the collaborationwith the new coating line Mitsuo Yaguchi MasahiroNakamura and Toshihiro Higuchi of Panasonic Corpora-tion Eco Solutions Company Core Technologies Develop-ment Center and Ken Yakura Norihito Ito and YoshihikoTsuchida of Sumitomo Chemical Co Ltd are gratefullyacknowledged for many fruitful discussions

Parts of the work presented here have received fundingfrom the European Commission from the EuropeanUnionrsquos Seventh Framework Programme (FP72007ndash2013) via the projects Clean for Yield (grant agreement281027) and LightTouchMatters (grant agreement31011) and from the European Unionrsquos Horizon 2020research and innovation programme (H20202014ndash2020)via the projects PI-SCALE (grant agreement 688093) andSOLEDlight (grant agreement 643791)

REFERENCES

1 MK Fung YQ Li and LS Liao Tandem organic light-emittingdiodes Adv Mater 28 10381 (2016)

2 G Luo X Ren S Zhang H Wu WCH Choy Z He andY Cao Recent advances in organic photovoltaics Devicestructure and optical engineering optimization on the nanoscaleSmall 12 1547 (2016)

3 S Kazim MK Nazeeruddin M Graumltzel and S Ahmad Perov-skite as light harvester A game changer in photovoltaics AngewChem Int Ed 53 2812 (2014)

4 CW Tang and SA van Slyke Organic electroluminescentdiodes Appl Phys Lett 51 913 (1987)

5 RH Friend RW Gymer AB Holmes JH BurroughesRN Marks C Taliani DDC Bradley DA Dos SantosJL Breacutedas M Loumlgdlund and WR Salaneck Electrolumines-cence in conjugated polymers Nature 397 121 (1999)

6 EK Lee Practical point source positioning for an OLED de-position system J Nanoelectron Optoelectron 11 148 (2016)

7 AR Duggal CM Heller JJ Shiang J Liu and LN LewisSolution-processed organic light-emitting diodes for lightingJ Disp Technol 3 184 (2007)

8 K Peters L Wengeler P Scharfer and W Schabel Liquid filmcoating of small molecule OLEDs J Coat Technol Res 11 75(2014)

9 N Aizawa YJ Pu M Watanabe T Chiba K Ideta N ToyotaM Igarashi Y Suzuri H Sasabe and J Kido Solution-processed

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

J Mater Res 201710httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at

multilayer small-molecule light-emitting devices with high-efficiency white-light emission Nat Commun 5 5756 (2014)

10 L Derue S Olivier D Tondelier T Maindron B Geffroy andE Ishow All-solution-processed organic light-emitting diodesbased on photostable photo-cross-linkable fluorescent small mol-ecules ACS Appl Mater Interfaces 8 16207 (2016)

11 KJ Choi JY Lee J Park and YS Seo Multilayer slot-diecoating of large-area organic light-emitting diodes Org Electron26 66 (2015)

12 KJ Choi JY Lee DK Shin and J Park Investigation on slot-die coating of hybrid material structure for OLED lightingsJ Phys Chem Solids 95 119 (2016)

13 S Raupp D Daume S Tekoglu L Merklein U LemmerG Hernandez-Sosa HM Sauer E Doumlrsam P Scharfer andW Schabel Slot die coated and flexo printed highly efficientSMOLEDs Adv Mater Technol 2 1600230 (2017)

14 SM Raupp L Merklein M Pathak P Scharfer and W SchabelAn experimental study on the reproducibility of different multi-layer OLED materials processed by slot die coating Chem EngSci 160 113 (2017)

15 A Sandstroumlm and L Edman Towards high-throughput coatingand printing of light-emitting electrochemical cells A review andcost analysis of current and future methods Energy Technol 3329 (2015)

16 J Hast M Tuomikoski R Suhonen KL Vaumlisaumlnen M VaumllimaumlkiT Maaninen P Apilo A Alastalo and A Maaninen Roll-to-rollmanufacturing of printed OLEDs In Proc SID 2013 (WileyVancouver 2013) p 192

17 T Minakata M Tanamura Y Mitamura M ImashiroA Horiguchi A Sugimoto M Yamashita K UjiiyeS Sunahiro Y Yada N Ibaraki and H Tomiyasu Fully R2R-processed flexible OLEDs for lighting In Proc SID 2015 (WileySan Joseacute 2015) p 219

18 RR Soslashndergaard M Houmlsel and FC Krebs Roll-to-roll fabri-cation of large area functional organic materials J Polym SciPart B Polym Phys 51 16 (2013)

19 C Koidis S Logothetidis A Ioakeimidis A Laskarakis andC Kapnopoulos Key factors to improve the efficiency of roll-to-roll printed organic photovoltaics Org Electron 14 1744 (2013)

20 Y Galagan IG de Vries AP Langen R AndriessenWJH Verhees SC Veenstra and JM Kroon Technologydevelopment for roll-to-roll production of organic photovoltaicsChem Eng Process 50 454 (2011)

21 CM Amb MR Craig U Koldemir J SubbiahKR Choudhury SA Gevorgyan M Joslashrgensen FC KrebsF So and JR Reynolds Aesthetically pleasing conjugatedpolymerfullerene blends for blue-green solar cells via roll-to-rollprocessing ACS Appl Mater Interfaces 4 1847 (2012)

22 P van de Weijer PCP Bouten S UnnikrishnanHB Akkerman JJ Michels and TMB van Mol High-performance thin-film encapsulation for organic light-emittingdiodes Org Electron 44 94 (2017)

23 Y Galagan H Fledderus H Gorter HH rsquot MannetjeS Shanmugam R Mandamparambil J BosmanJEJM Rubingh JP Teunissen A Salem IG de VriesR Andriessen and WA Groen Roll-to-roll slotndashdie coatedorganic photovoltaic (OPV) modules with high geometrical fillfactors Energy Technol 3 834 (2015)

24 P Svarnas L Yang M Munz AJ Edwards AG Shard andJW Bradley Highly-selective wettability on organic light-emitting-diodes patterns by sequential low-power plasmasJ Appl Phys 107 103313 (2010)

25 GH Han SH Lee WG Ahn J Nam and HW Jung Effect ofshim configuration on flow dynamics and operability windows instripe slot coating process J Coat Technol Res 11 19 (2014)

26 I de Vries and J Gabel Capillary velocity of liquids betweenheterogeneous surfaces In Proc 16th Int Coat Sci Techn Symp(International Society for Coating Science and TechnologyAtlanta 2012)

27 M Schmitt R Diehm P Scharfer and W Schabel An experi-mental and analytical study on intermittent slot-die coating ofviscoelastic battery slurries J Coat Technol Res 12 927 (2015)

28 SF Kistler and PM Schweizer eds Liquid Film Coating(Springer Science and Business Media BV Dordrecht theNetherlands 1997) ISBN 978-94-010-6246-6

29 RK Yonkoski and TJ Ludemann Effect of excess airfoil lengthon mottle In Proc ISampT 50th Ann Conf (held in Boston 1997)p 540

R Abbel et al Toward high volume solution based roll-to-roll processing of OLEDs

11J Mater Res 2017httpswwwcambridgeorgcoreterms httpsdoiorg101557jmr2017204Downloaded from httpswwwcambridgeorgcore Technische Universiteit Delft on 22 Jun 2017 at 084747 subject to the Cambridge Core terms of use available at