Large area inkjet-printed OLED fabrication with solution-processed TADF ink

This work demonstrates successful large area inkjet printing of a thermally activated delayed fluorescence (TADF) material as the emitting layer of organic light-emitting diodes (OLEDs). TADF materials enable efficient light emission without relying on heavy metals such as platinum or iridium. However, low-cost manufacturing of large-scale TADF OLEDs has been restricted due to their incompatibility with solution processing techniques. In this study, we develop ink formulation for a TADF material and show successful ink jet printing of intricate patterns over a large area (6400 mm2) without the use of any lithography. The stable ink is successfully achieved using a non-chlorinated binary solvent mixture for a solution processable TADF material, 3‐(9,9‐dimethylacridin‐10(9H)‐yl)‐9H‐xanthen‐9‐one dispersed in 4,4’-bis-(N-carbazolyl)-1,1’-biphenyl host. Using this ink, large area ink jet printed OLEDs with performance comparable to the control spin coated OLEDs are successfully achieved. In this work, we also show the impact of ink viscosity, density, and surface tension on the droplet formation and film quality as well as its potential for large-area roll-to-roll printing on a flexible substrate. The results represent a major step towards the use of TADF materials for large-area OLEDs without employing any lithography.


REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): In this manuscript, the author has developed large-area multilayer OLEDs with an inkjet-printed TADF emission layer and have successfully demonstrated that IJP as a single step maskless emission layer patterning technique to create intricate, high-resolution designs for signage, wearable electronics, and advertising without the use of typical lithography steps.However, this printing area is not very large and the strategy is not novel.
1. Why does the viscosity and surface tension decrease when two single solvents are mixed in ink solvents?Please the author provides an explanation.
2. In this article, the author prints a line from multiple points on the luminescent material, resulting in a thin film with very low surface roughness.Can the author explain in detail how to maintain the flatness of the film? 3.As is well known, temperature control can regulate the evaporation process.But the author only used three substrate temperatures of 30 ℃, 40 ℃ and 50 ℃.The author found that different substrate temperatures can lead to significant differences in patterns and processing the substrate at 40 ℃ can drastically reduce the coffee ring effect.However, the boiling point of the mixed ink solvent is very high, and it is difficult for smaller temperatures to have a significant impact in a short period of time.The author's experimental phenomenon contradicts this theory.Please provide a detailed explanation.4. Why is the ink solvent ratio 6:4?Would other ratios be better? 5.In addition, please improve the language used in this article.And also please check that all references are formatted according to the specified style.
Reviewer #2 (Remarks to the Author): The manuscript submitted to me for review is very well prepared.Despite the many articles devoted to printed emissive layers for OLED applications, there are still few examples of printed TADF systems.In my opinion, the greatest value of the work is the detailed description of the optimization of the ink and its full characterization, as well as the printing process and the evaluation of many parameters affecting the quality and thickness of the printed layers, which directly affects the performance of the devices.The authors have correctly selected the research methods, the analysis of the results presented is clear and consistent.The procedure presented in the paper can be considered as a kind of guide to optimize the process of printing thin organic layers.
In order to improve the quality of the manuscript, I recommend the authors: -to add information regarding the significance of the modification of ACRTX synthesis used, -to move Fig.S4 from Supporting Information to the manuscript, -to explain whether the difference in thickness of TADF layers (Table 3) in the compared devices (e.g., SP1 ca.36nm and IJP1 ca. 25 nm) can affect the characteristics shown in Fig. 7, -to explain in section of "OLED Device fabrication process" whether the solvent composition of the TADF ink used for printing and for the spin-cating process were the same, -to add a more detailed analysis of the characteristics shown in Fig. 7, especially the I-V and Luminance vs. Voltage curves.
In my opinion, after taking into account the above recommendations, the paper should be published in Nature Communications.

Reviewer #3 (Remarks to the Author):
This manuscript reports OLEDs with a printable TADF emissive layer.Given that are very few reports on printable TADF OLEDs, it is quite encouraging to see this manuscript.The authors have done extensive work on the ink formulation of the TADF emissive layer.While such study of ink formulation has been reported extensively for many other emissive materials, not much has been reported for TADF inks due to solubility issues of TADF emitters.
While the work is encouraging, the current manuscript needs more information and amendments for quality, clarity and reproducibility of the work.
Here are some suggestions and questions which may help in improving the quality and clarity of the manuscript: 1.While reading the abstract and introduction, the narrative felt the study is focussed on nonchlorinated solvents, as they are more industry friendly.However, in the results, the authors started with common chlorinated solvents and the narrative in the results section infers that the nonchlorinated solvents were chosen because the chlorinated solvents were not suitable for printing (low boiling point, etc.).Why were the chlorinated solvents studied if they were not intended to be used (which also coincidently turned out to be not suitable)?Changing either the narrative or moving the data for chlorinated solvents to supplementary information would be more suitable.
2. Were any other non-chlorinated solvents investigated?Include reason/s why Toluene and MB were specifically chosen.Why not any other non-chlorinated solvents?3. Why was the material ACRXTN chosen for this study?4. Provide supporting reference for the statement for optimal surface tension "…allowing for the final surface tension to be adjusted to the optimal range (28-32 mN m-1)." on page 6. OR is this a requirement for the ink-jet printer used here?5.In Figure 2c and 2f, what are the blue squares?6.On page 8, the authors have stated "We also discovered that reducing the number of firing nozzles from all (16) to few (5 or 1) reduced drop velocity."How many nozzles were fired for the printing of emissive layer?Are the images in Figure 4 from just one nozzle or many? 7. On page 11, the authors have stated "As shown in Figures 4c, d, & e, as the DPI was over 400, previously separate droplets began to overlap and combine to form a continuous film.Hence, we can conclude that resolution over 400 DPI are necessary to generate consistent films at 40 oC substrate temperature for our formulated ink."However, the film in 4d which corresponds to 500 DPI looks more like lines.The 3D image in supporting information doesn't add much information as the substrate and printed film have the same colour tone.Provide a cross sectional height profile similar to that of Figure S4f.4d and 4e, film of 500 DPI looks a lot more rough as compared to film of 600 DPI.However, the AFM height data indicates the 600 DPI film is less rough (from the height scale bar).The scan area of the AFM is quite small and would not reflect what is shown in Figure 4.A cross sectional height scan over a larger area should be included.This is all the more significant, given that the OLEDs reported here are 4x4 mm2.9. Why are the emissive layer films dried under nitrogen shower?Given that the authors have argued for ambient fabrication, wouldn't it be counter intuitive to dry in a nitrogen environment.What is the effect of drying in the absence of a nitrogen shower?10.Is the EL the same for all the OLED?Comment on that would be helpful.

Looking at films in Figures
11. Are the plots in Figure 7 for best performing OLEDs or for a typical representative device?12. Statistics for OLED performance should be provided.Only data of best performing devices are included currently.
13. Scale bar for figure 4Sa and d should be increased.

RESPONSES TO REVIEWERS' COMMENTS
We are thankful to the Reviewers who provided insightful evaluations and comments.
Accordingly, we have revised the manuscript in response to all the Reviewers' comments.All our modifications are shown in yellow highlighted texts while our responses to Reviewers' comments are written in blue colour font.

Reviewer #1 (Remarks to the Author):
In this manuscript, the author has developed large-area multilayer OLEDs with an inkjetprinted TADF emission layer and have successfully demonstrated that IJP as a single-step maskless emission layer patterning technique to create intricate, high-resolution designs for signage, wearable electronics, and advertising without the use of typical lithography steps.
However, this printing area is not very large, and the strategy is not novel.

Response:
We thank the Reviewer's comments.To address the Reviewer's comments on the printing areas, we have now successfully fabricated large-area IJP TADF devices with an active area of 6400 mm 2 (both non-patterned and patterned OLEDs).These are the largest OLED devices that we can fabricate using our facility as limited by our maximum evaporator deposition area.By using 600 DPI, we have positively demonstrated: i) large area inkjet-printed panel (LAP) OLEDs consisting of chromium gridlines on 120 × 120 mm 2 ITO/glass substrates with an active area of 80 × 80 mm 2 as shown in the new Figure 9a; ii) large area inkjet printed intricate design logos on 100 × 100 mm 2 ITO/glass substrates, namely, unpatterned and patterned large area OLEDs.
We have now added below new sections related to our new large-area IJP devices (both non-patterned and patterned OLEDs) to the manuscript (on pages 15-17):

"IJP large-area non-patterned OLEDs (with an active area of 80 × 80 mm 2 )
To demonstrate large area ink-jet printed (LAiP) OLEDs, we fabricated large-area panel (LAP) IJP TADF OLEDs on 120 × 120 mm 2 chromium/ITO/glass substrates with an active area of 80 × 80 mm 2 (Figure 9a) in a 10,000-class clean room at ambient conditions.We applied the chromium gridlines for the homogeneous current distribution as the sheet resistance of ITO starts to affect the large-area substrates.After performing an ozone surface treatment of the ITO substrate, PEDOT:PSS and PVK-TAPC layers were spin-coated on top.Substrates were then annealed at 120 °C for 15 min before the TADF ink was IJP at 600 DPI on top of the PVK-TAPC layer.This gave a device structure of ITO/PEDOT:PSS (40 nm)/PVK-TAPC (18 nm)/TADF (30 nm)/TPBi (32 nm)/Ca (20 nm)/Al (100 nm).The devices were then encapsulated for characterization and analysis.
Figure 9c shows an operational LAiP TADF OLED at 10.5 V.The device emitted homogenous light with a maximum luminescence of 622 cd m -2 .We noticed some non-uniformities of light emission at the lower voltages but they disappeared as the bias voltage increased.This may be due to non-uniformity of the spin-coated layers (PEDOT:PSS and PVK-TAPC) on a large substrate area.We also observed that the presence of dust particles on the substrate during the annealing process, which resulted in devices getting shorted.To minimize dust particles, substrates annealing was, therefore, performed under N2 shower.With few iterations, we successfully fabricated and tested the large-area devices.As shown in Figure 9c, the LAiP OLED had the maximum current efficiency of 13.7 cd A -1 @ 622 cd m -2 , which is close to those of spin-coated small area OLEDs.The J-V-L and current efficiency plot is shown in Supplementary Figure S11.First, we inkjet printed a SU-8 (a negative epoxy) based photoresist dielectric (DPI 800; thickness 1µm) to create the desired intricate template on 100 × 100 mm 2 ITO substrates.The resulted SU-8 patterned layer was then exposed with UV light to cross-link the photoresist.HIL (PEDOT:PSS) and HTL layers (PVK-TAPC) were then spin coated on top of the SU-8/ITO template.The TADF ink was IJP at 600 DPI on the PVK-TAPC layer to create the intricate light emission regions, matching with the dielectric template.Finally, the substrates were transferred to an evaporator for thermal deposition of TPBi (32 nm) and Ca/Al electrodes. 1.Why does the viscosity and surface tension decrease when two single solvents are mixed in ink solvents?Please, the author provides an explanation.

Response:
We thank the Reviewer's valid question.The viscosity and surface tension of an ink solution made from two separate solvents may differ from those of the individual solvents.This happens because of the intermolecular interaction of the two solvents, which might modify their physical characteristics.Both solvents utilized in this study have distinct viscosity and surface tension values.Namely, toluene has a viscosity of 0.8 cP, whereas MB has a viscosity value of 2.0 cP.When toluene and MB are combined in a ratio of 60:40, the solution's viscosity was determined to be 1.3 cP.The addition of 50 mg of material to the solution was not significant to impact the global solution viscosity.Furthermore, the surface tension at 20 °C for toluene and MB are 28 mN and 38 mN m -1 , respectively.As a binary solvent system, the surface tension falls between these two values.
The most commonly used model to predict the viscosity of a binary solvent mixture is the empirical equation known as the Grunberg-Nissan equation 1 : ηmix = Σχi*ηi, where ηmix is the viscosity of the mixture, ηi is the viscosity of each pure solvent, and χi is the volume fraction of each solvent in the mixture.The type of the solvents, the interactions between their molecules, and the concentration of the solutes are some examples of the factors that affect the size and direction of the change in surface tension that occurs during mixing.
Recently, Kleinheins et al. have compiled an overview of the most popular models and their ability to reproduce experimental data of ten binary aqueous solutions 2 .It is important to note that the specific behaviour of solvent mixing and its impact on viscosity and surface tension can vary depending on the solvents and their concentrations.Experimental measurements and theoretical models are typically used to understand and predict these changes 3,4 .

2.
In this article, the author prints a line from multiple points on the luminescent material, resulting in a thin film with very low surface roughness.Can the author explain in detail how to maintain the flatness of the film?

Response:
We are grateful for the Reviewer's remarks.First, we printed at lower resolution (i.e., lower DPI) so that the individual drop diameter under the printing conditions can be measured.In this study, at 200 DPI there was clear separation of each droplet, and a centre-tocentre (C2C) distance of 125 μm was measured (Figure R1a).The droplet diameter slightly varied with different printing parameters.Then we input these values into a droplet simulator to determine the DPI, where the droplets were overlapped.Insufficient droplet overlap (Figure R1b, c) led to non-uniform films and high surface roughness, while too much droplet overlaps wasted ink and increased drying time, leading to heterogeneous films.We have found that under the study conditions, >400 DPI, were the ideal printing resolution for uniform films with complete merger of droplets (Figure R1d) with a low surface roughness.

3.
As is well known, temperature control can regulate the evaporation process.But the author only used three substrate temperatures of 30, 40, and 50 ℃.The author found that different substrate temperatures can lead to significant differences in patterns and processing the substrate at 40 ℃ can drastically reduce the coffee ring effect.However, the boiling point of the mixed ink solvent is very high, and it is difficult for smaller temperatures to have a significant impact quickly.The author's experimental phenomenon contradicts this theory.
Please provide a detailed explanation.

Response:
We thank the Reviewer's comments.The droplets had a minimal volume of 10 pL (i.e., one millionth of a microlitre), and after ejecting from the nozzle, they were impinging on the substrate, which had an average diameter of 50-70 μm and a thickness of a few tens of nanometres.The droplets' thickness in nm and diameter in μm scale make them a large surface-to-volume ratio, meaning the droplets will be increasingly affected by slight variations in substrate temperature as well as the solvent vapour pressure.
The coffee ring effect originated from the capillary flow induced by the differential evaporation rates between the edge and interior of the TADF ink drop, directly affecting the uniformity of printed TADF films. 5To obtain uniform TADF films, the printing parameters require systematic optimization.Based on the formation mechanism of the coffee ring effect, there are three methods for inhibiting it: (i) weakening the capillary flow from the inside to the outside of the droplet, (ii) increasing the Marangoni flow from the outside inside the liquid, and (iii) controlling the movement of the three-phase contact line of liquid drops in the drying process.
To balance the capillary flow and eliminate the coffee ring effect, Sun et al. optimized the evaporation rate of their CsPbBr3 perovskite QD ink by altering the volume of two solvents (dodecane and toluene) to form appropriate Marangoni flow. 6Similarly to our case, at a lower substrate temperature of 30 o C, the capillary forces dominated, and the solvent did not evaporate sufficiently due to the high boiling point and low vapor pressure of MB.As we increased the substrate temperature to 40 o C, there was an optimal equilibrium between the Marangoni and viscous capillary forces, suppressing the coffee ring effect.As we increased the substrate temperature further to 50 o C, we obtained rougher films due to quick evaporation of the solvent post deposition.Namely, when the substrate temperature is too high, the droplets cannot merge in liquid form, and instead becomes a liquid ink addition to an already solid film.
We note that Amrut et al. 7 made similar observation of substrate temperature on final film quality.

4.
Why is the ink solvent ratio 6:4?Would other ratios be better?

Response:
We are thankful for the Reviewer's valid questions.We have achieved closer values concerning a printable range of surface tension of (28-32 mN m -1 ) for DMC Samba cartridges for inkjet printing specified by the company.Further increasing the MB ratio in the formulated ink leads to surface tension values of 35.4 mN m -1 (Figure R2) outside this ideal printable window.Response: We thank the Reviewer's helpful comment.We have now made the required update to the Supplementary Information as shown below.

Fig. 9
Fig. 9 Images of large-area ink jet panel substrates and OLEDs.a Handheld large-area panel (LAiP) substrate (120 × 120 mm 2 ) with an active area of 80 × 80 mm 2 through square chromium gridlines.b Inkjet patterned large area logos on a 100 × 100 mm 2 substrate, consisting of The University of Queensland and IIT Kanpur with two rectangular strips of 10

Figure 9d
Figure 9d shows the proof-of-concept device consisting of logos of The University of Queensland and IIT Kanpur as well as two rectangular strips of 10 × 70 and 5 × 70 mm 2 .The device achieved brightness of 500 cd m -2 at 10 V.The images show high contrast glowing edges revealing intricate features.It is important to note that our negatively IJP patterned SU-8

Fig. R1
Fig. R1 Simulation of TADF films for homogeneous film formation with 200, 300, and 400 and 500 DPI, respectively.a When DPI is 200, the centre-to-centre (C2C) distance is 125 μm.b only further increasing the DPI to 300 reduces the distance to 83.3 μm between the printed droplets.c Droplets starts to merge as C2C distance (62.5 μm) is less than the drop diameter (75 μm).d at 500 DPI C2C distance reduces to 50 μm and droplets merge with completely overlapped area.

Table 3 :
Summary of Von, LEmax, L, and CIE coordinates of the fabricated devices.Device pixel area 4 × 4 mm 2 .have added the following new Figure S10 to the Supplementary information.

Fig. S10
Fig. S10 Performance statistics of small area OLEDs.The device performance metrics were obtained from 3 to 4 pixels on one substrate.a Light turn-on voltage variation of spin coated and IJP devices measured at ≈1 cd m -2 .b Brightness from IJP and spin coated OLEDs.c Current

Fig. S4 a
Fig. S4 a 2D-image of the printed droplets with 200 DPI.d 2D image of the printed droplets at 600 DPI but after drying the film showing ununiform as many patches can be seen.