Ultraviolet light blocking optically clear adhesives for foldable displays via highly efficient visible-light curing

In developing an organic light-emitting diode (OLED) panel for a foldable smartphone (specifically, a color filter on encapsulation) aimed at reducing power consumption, the use of a new optically clear adhesive (OCA) that blocks UV light was crucial. However, the incorporation of a UV-blocking agent within the OCA presented a challenge, as it restricted the traditional UV-curing methods commonly used in the manufacturing process. Although a visible-light curing technique for producing UV-blocking OCA was proposed, its slow curing speed posed a barrier to commercialization. Our study introduces a highly efficient photo-initiating system (PIS) for the rapid production of UV-blocking OCAs utilizing visible light. We have carefully selected the photocatalyst (PC) to minimize electron and energy transfer to UV-blocking agents and have chosen co-initiators that allow for faster electron transfer and more rapid PC regeneration compared to previously established amine-based co-initiators. This advancement enabled a tenfold increase in the production speed of UV-blocking OCAs, while maintaining their essential protective, transparent, and flexible properties. When applied to OLED devices, this OCA demonstrated UV protection, suggesting its potential for broader application in the safeguarding of various smart devices.


Reviewer #2 (Remarks to the Author):
To address the need for reducing power consumpfion in foldable smartphones, the development of an advanced opfically clear adhesive (OCA) capable of blocking UV light while preserving crucial properfies like adhesion and strain relief is imperafive.The authors introduce a novel photo-inifiafing system (PIS) that enables UV-blocking OCA producfion at a rate 10 fimes faster than their previous work when exposed to visible-light irradiafion.While the overall design is interesfing, the manuscript's presentafion lacks clarity in conveying the innovafions, potenfially leading to reader misinterpretafion.

My comments are outlined below:
1.In the Abstract and Introducfion, the authors assert that they employed a newly designed PC (PhotoCatalyst), specifically 4DP-IPN.However, it is worth nofing that this PC, 4DP-IPN, has been previously explored in their earlier research.Throughout the manuscript, substanfial effort is invested in comparing 4DP-IPN with 4Cz-IPN, both in the main text and Figure 1.This extensive comparison may inadvertently obscure the innovafive aspects and significance of the current study.It is essenfial to emphasize that the primary disfincfion between the photo-inifiafing systems (PIS) developed in this work and their prior research lies in the choice of the photoinifiator.This pivotal disfincfion could be highlighted more prominently in the present manuscript.Addifionally, Figures 1 and 2 contain a large amount of redundant informafion from their earlier work, warranfing aftenfion for a more concise and focused presentafion.
2. The molecules are inifially excited to the singlet state and subsequently transifion into the triplet state in the photoexcitafion.The emphasis in this work primarily centers on the design principles pertaining to the triplet state of various species within the photo-inifiafing system (PIS).However, the quesfion arises: why is the singlet state of these materials not equally crifical and deserving of discussion? 3. It's befter to differenfiate more for the terms "triplet-triplet energy transfer (EnT)" and "energy transfer (ET)".
4. The authors achieved developing UV-blocking OCAs at a rate approximately 10 fimes faster than before.And claimed that the technique could be immediately commercialized.However, the manuscript's current presentafion may leave readers seeking more compelling evidence to fully substanfiate this claim.
5. What about the UV blocking ability of the layers "BM and CF", "ToE", and "TFE"?In the dome of Fig. 4, it might be befter to include those layers for the comparison.6.For the adhesive test, please provide the tesfing condifions and the related standard.The peel strength in the newly proposed OCA is lower than in their previous work, what could be the potenfial reasons?
7. What advantages does the design of the OCA with UV-blocking ability offer over alternafive strategies, like incorporafing UVAs into the BM-CF layer?This alternafive approach might potenfially expedite OCA development significantly, given its potenfially reduced demands on the choice of PC and inifiators.
8. How does the UV-blocking ability depend on the thickness of the OCA? 9.The misspelling of 'TFE' is found in the Fig. 1 legend.

Reviewers' comments to author:
We are grateful to the reviewers for their constructive comments and suggestions, which greatly helped to clarify a number of important points, avoid misunderstandings, and improve the clarity and referencing of the paper.The reports of the reviewers were overall very positive on the significance and novelty, stressing the suitability of our work for publication in Nature Communications.We have carefully revised our manuscript in line with all the reviewers' comments.The point-by-point response is given below.Our responses to reviewers' comments are highlighted in blue color with a highlight in the revised manuscript and SI.

Reviewer #1 (Remarks to the Author):
In their work, the authors greatly improved the visible-light curing speed by developing a highly efficient photoinitiating system, based upon their previous research work on the synthesis of ultraviolet light blocking OCA.The key mechanical properties of OCA that are utmost important to the structural integrity of OCA bonded foldable screens, such as adhesion strength, stress relaxation, and strain recovery, have been analyzed and showed good results comparable to traditional commercial OCAs.This work presents an important progress in developing UV-blocking soft adhesives for foldable screens.
The manuscript is well written.Below are some minor questions and a few typos.
1) What are the advantages of the stress relaxation followed by strain recovery control mode used in this work, compared to the classical separated creep recovery and stress relaxation tests?In the stress relaxation step, is it instantaneous loading or ramped loading being used to apply the strain to 300%?What is the strain rate if it is ramped loading?
Response: Thank you for the reviewer's careful comments.It appears there was some confusion caused by our insufficient explanation, for which we apologize.To clarify, we measured the strain recovery and stress relaxation of our OCAs separately.Initially, we determined the stress corresponding to a 300% strain from dynamic mechanical analysis.Subsequently, we used this stress level to measure strain recovery.
We also wish to highlight that our focus was on the measurement of strain recovery rather than creep recovery, given our targeted applications in foldable displays.Creep recovery typically involves long-term deformation under constant stress until the material reaches a viscous deformation region.The reviewer correctly noted that foldable durability of an OCA can be evaluated by both creep recovery and stress relaxation, as outlined in 'Pressure-Sensitive Adhesives for Flexible Display Applications (2019)'.However, considering the real-world use of foldable displays, an OCA must exhibit durability during rapid folding and unfolding processes, and it needs to maintain its integrity when the display is kept in a folded state for extended periods.To simulate the practical use of foldable displays, we conducted the strain recovery tests by applying instantaneous loading at a constant strain, confined to the elastic deformation region.For additional context, applying a 300% strain is a common standard for testing the strain recovery of OCAs used in foldable displays.This level of strain correlates with the curvature that these displays typically experience in practical applications (see Fig. R1; now Supplementary Fig. 20 in the revised SI).For instance, the shear behavior of an OCA can be illustrated by the difference in length between an inner curve (Fig. R1, blue line) and an outer curve (Fig. R1, green line), as represented by Equations 1−3.
In the given context, D1 and D2 represent the outer and inner folding curves, respectively, while D signifies a displacement in the shear strain.T corresponds to the thickness of the OCA, and a is the folding radius.It follows from the relationship in the shear strain that the displacement is likely to increase in proportion to the thickness, multiplied by π.Consequently, under the conditions of 180° folding, we applied a strain of approximately 300% in order to assess the strain recovery and stress relaxation.
2) In the rheology test, the temperature sweep test in the range of -20 °C to 80 °C has been conducted, can the glass transition temperature (Tg) of the material be identified from the data?
Response: Thank you for the reviewer's insightful comment.Because the Tg of the OCA we prepared is beyond the temperature range covered by the original manuscript's temperature sweep test, it was not observed in the initial data.
In response to the reviewer's comments, we expanded the temperature range from -50°C to 90°C and conducted another temperature sweep.The Tg was measured to be approximately -40°C, as illustrated in Fig. R2 below (now Supplementary Fig. 21 in the revised SI).
Generally, the Tg of OCAs used in foldable displays is targeted to be below -20°C to ensure adequate flexibility (SID Symposium Digest of Technical Papers, 48, 198 (2017)).This target also takes into account the use of the display in extremely cold environments (Samsung Display Newsroom.Samsung Foldable OLED: 'Folding Test in Extreme Cold').
Therefore, in synthesizing OCAs, we prepared UV-blocking OCAs using 2-ethylhexyl acrylate (EHA, Tg (poly(EHA)) ~ -58°C) and 4-hydroxybutyl acrylate (HBA, Tg (poly(HBA) ~ -65°C) (BASF, Acrylic and Methacrylic Monomers (2021)).With the combination of these acrylic monomers, we anticipated, based on the Flory-Fox equation (Equation 4), that the Tg of the prepared OCA would theoretically fall within the range of -65 to -58°C; in this equation, wx represents the weight fraction of component x, and Tg,x denotes the glass transition temperature of component x (where x = A or B).
Despite expectations based on a [EHA]:[HBA] ratio of 3:1 (i.e., a weight percentage ratio of EHA:HBA = 3.83:1), the actual Tg of our synthesized OCAs was observed to be -39.7°Cfrom a dynamic temperature sweep ranging from -50°C to 90°C (Fig. R2).We have properly added these results in the revised manuscript as follows: - "The storage modulus (G'), loss modulus (G''), and damping factor (tan δ) were measured by a rheometer through a dynamic temperature sweep ranging from -50 °C to 90 ℃ (Supplementary Table 12).Notably, the measured values at all temperatures met commercial OCA requirements.At low temperatures, the storage modulus value (G' = 150.3kPa at -20 ℃) was found to be slightly higher than that of CEF 3602 (G' = 115.0kPa at -20 ℃), yet it still meets the commercial standard where the storage modulus of commercially available OCAs is within the 100-150 kPa range. 70,71  Furthermore, the Tg of the OCA film, measured at -39.7 °C during a dynamic temperature sweep, is consistent with the observed low G' at -20°C, as detailed in Supplementary Fig. 21  However, in response to comment #3-which hypothesized that failures occurred in the peel strength test-we surmise that adhesive failure would be the most likely.Although the common monomer ratio between EHA and HBA is 4:1 for synthesizing foldable OCAs (Patent US 10640689 B2), we utilized a ratio of [EHA]:[HBA] = 3:1.This was done to enhance the crosslinking in our prepared OCAs, thereby preventing cohesive failure (Int.J. Adhes. Adhes., 74, 137-143 (2017);Prog. Org. Coat., 88, 155-163 (2015); Polym.Int., 52, 347-357 (2003)).Furthermore, to provide a clearer understanding of our process, we have included images of the peeling test in the revised SI (Fig. R3).

Glass substrate
Sample OCA; ~ 50 μm thickness 4) In page 6, the authors wrote "Notably, the measured values at all temperatures met commercial OCA requirements.… , still fulfilling the standard for commercialization ", are there any specific numerical ranges in the commercial OCA requirements?Can the authors provide the references that describe the specific commercial requirements or standards?
Response: Thank you for the reviewer's insightful comments.Unlike standardized methods for measuring physical properties, universally accepted standards for the commercialization of these properties do not exist; they vary among companies and are often closely guarded as trade secrets.For example, the standards for the physical properties of flexible OCA differ between companies such as LG Display and Samsung Display.Consequently, it is difficult to establish precise benchmarks for commercialization; please be aware that disclosing this information in your paper could potentially lead to legal issues.In the case of the OCA developed in this study, it has been internally evaluated by Samsung Display and found to nearly meet their commercialization standards.Additionally, we can draw on the approximate commercial standards based on 3M's CEF 3602, which is already being utilized in foldable devices.
Therefore, to assess the commercial viability of our OCA films in this study, we compared their viscoelastic properties with those of the commercially available foldable OCA, specifically 3M's CEF 3602.Table R2 illustrates that the UVblocking OCA produced under our conditions demonstrates a storage modulus (G'), loss modulus (G''), and damping factor (tan δ) comparable to those of CEF 3602, both at room temperature and at elevated temperatures.At lower temperatures (-20 ℃), our OCA exhibits slightly higher storage modulus values than CEF 3602.However, considering that the storage modulus for commercialized OCAs ranges from 100 to 150 kPa at -20 ℃, as reported in the SID Symposium Digest of Technical Papers, 48, 198 (2017) and Patent US 11502270 B2, our OCA-with a G' of 150.3 kPa-demonstrates strong potential for commercialization.We have included these references in the revised manuscript to further substantiate our claims.

Table. R2
Results for the storage modulus (G'), loss modulus (G''), and damping factor (tan δ) of the OCA films prepared in this study are presented.The viscoelastic properties of these OCAs were evaluated at -20°C, 25°C, 60°C, and 85°C.For comparison, CEF3602 was used as a benchmark in the control experiments.Response: Following the reviewer's comments, we have properly corrected.
Response: Following the reviewer's comments, we have properly corrected.
Reviewer #2 (Remarks to the Author): To address the need for reducing power consumption in foldable smartphones, the development of an advanced optically clear adhesive (OCA) capable of blocking UV light while preserving crucial properties like adhesion and strain relief is imperative.The authors introduce a novel photo-initiating system (PIS) that enables UV-blocking OCA production at a rate 10 times faster than their previous work when exposed to visible-light irradiation.While the overall design is interesting, the manuscript's presentation lacks clarity in conveying the innovations, potentially leading to reader misinterpretation.
My comments are outlined below: 1) In the Abstract and Introduction, the authors assert that they employed a newly designed PC (Photocatalyst), specifically 4DP-IPN.However, it is worth noting that this PC, 4DP-IPN, has been previously explored in their earlier research.Throughout the manuscript, substantial effort is invested in comparing 4DP-IPN with 4Cz-IPN, both in the main text and Figure 1.This extensive comparison may inadvertently obscure the innovative aspects and significance of the current study.It is essential to emphasize that the primary distinction between the photo-initiating systems (PIS) developed in this work and their prior research lies in the choice of the photoinitiator.This pivotal distinction could be highlighted more prominently in the present manuscript.Additionally, Figures 1 and 2 contain a large amount of redundant information from their earlier work, warranting attention for a more concise and focused presentation.

Response:
We thank the reviewer for the encouraging comments and constructive suggestions.We acknowledge the reviewer's point that an excessive focus on the previous work may have detracted from the uniqueness of this study, and the detailed characterizations of the chemicals could have obscured the clarity in Fig. 1 and 2. In response to the referee's remarks, we thus have substantially revised the abstract and introduction to more clearly delineate the differences between our previous and current work (see below).Additionally, Fig. 1 and 2 have been updated to enhance the readers' comprehension of our novel contributions.Based on this, we clarified the issues present in the past studies to better convey the novelty of our current research (see Fig. R4; Fig. 1b in the revised manuscript).Moreover, as suggested by the reviewer, we have incorporated references to previous studies that used 4DP-IPN, offering a more comprehensive understanding of its application.
The abstract of the revised manuscript has been rewritten as follows: "In developing an organic light-emitting diode (OLED) panel for a foldable smartphone (specifically, a color filter on encapsulation) aimed at reducing power consumption, the use of a new optically clear adhesive (OCA) that blocks UV light was crucial.However, the incorporation of a UV-blocking agent within the OCA presented a challenge, as it restricted the traditional UV-curing methods commonly used in the manufacturing process.Although a visible-light curing technique for producing UV-blocking OCA was proposed, its significantly slow curing speed posed a barrier to commercialization.Our study introduces an innovative and efficient photo-initiating system (PIS) for the rapid production of UV-blocking OCAs utilizing visible light.We have re-engineered the photocatalyst (PC) to minimize electron and energy transfer to UV-blocking agents and introduced new co-initiators that allow for faster electron transfer and quicker PC regeneration compared to previously established amine-based co-initiators.This advancement enabled a tenfold increase in the production speed of UV-blocking OCAs, while maintaining their essential protective, transparent, and flexible properties.When applied to OLED devices, this new OCA demonstrated outstanding UV protection, suggesting its potential for broader application in the safeguarding of various smart devices." The following sentences have now been added in the introduction: "Through a mechanistic analysis of existing PIS, we found that electron transfer (ET) and energy transfer (EnT) between the PC and the UVAs predominates over that between the PC and the co-initiator.This greatly reduces the efficiency of PIS and, consequently, impedes the curing rate.To address this issue, we delicately modified the structure of the existing PC to design a new PC, 4DP-IPN, with reduced ET and EnT efficiency with UVAs (Fig. 1b).Additionally, we introduced co-initiators, i.e., HNu 254 and Borate V, that facilitate faster electron transfer and PC regeneration compared to previously used amine-based co-initiators." 2) The molecules are initially excited to the singlet state and subsequently transition into the triplet state in the photoexcitation.The emphasis in this work primarily centers on the design principles pertaining to the triplet state of various species within the photo-initiating system (PIS).However, the question arises: why is the singlet state of these materials not equally critical and deserving of discussion?
Response: We appreciate the reviewer's insightful comments, which have certainly helped to prevent confusion among potential readers.We apologize for any misunderstanding caused by our previous insufficient explanation.As the reviewer has pointed out, cyanoarenes, a class of thermally activated delayed fluorescence (TADF) compounds, can exhibit both excited singlet and triplet states (Chem. Soc. Rev., 50, 7587 (2021)).While both singlet and triplet states in 4DP-IPN and 4Cz-IPN demonstrate similar magnitudes of electron transfer rate constants due to comparable driving forces (−ΔG) with co-initiators (Fig. R5a), it is indeed more accurate to consider the electron transfer process in terms of the actual rate (νPET (M•s −1 ) = kPET [PC * ][Q]), rather than the rate constant (kPET, M −1 •s −1 ).From this standpoint, the concentrations of the PC in the excited state and the quencher are crucial in determining the electron transfer rate.
To this end, we performed a kinetic simulation based on the rate law to ascertain the concentration of PC in the excited state under photostationary conditions, utilizing the reported Jablonski diagrams of the two PCs (Adv. Mater., 35 2204776 (2023); see Fig. R5b).The simulation results indicated that for 4Cz-IPN, the concentration ratio between the singlet and triplet excited states was approximately 100-fold, while for 4DP-IPN, this ratio exceeded 10,000-fold.This suggests that even with low concentrations of co-initiators, the triplet state of both PCs can contribute more significantly to the electron transfer process than the singlet state (Nat. Commun. 14, 92 (2023);Chem. Rev. 122, 1830−1874(2022)).
In our experimental setup, the concentrations of co-initiators, HNu 254 (5.2 mM) and Borate V (3.1 mM), are much lower than those used in photoluminescence quenching experiments.Consequently, the electron transfer from the triplet state is expected to be predominant in the actual polymerization process, in contrast to the singlet state, which would require significantly higher concentrations of co-initiators for quenching.
For clarification, we have added the following sentence in the revised manuscript as follows: "4Cz-IPN, a cyanoarene-based thermally activated delayed fluorescent (TADF) material, generates singlet and triplet excited states with negligible energy differences, 33,34 and its triplet state concentration is approximately 100 times higher, predominantly contributing to ET (Supplementary Fig. 4)." and "in the case of 4DP-IPN, the excited state triplet concentration is nearly 1000 times higher than that of the singlet, making the triplet the principal contributor, similar to 4Cz-IPN (see Supplementary Fig. 4)." 3) It's better to differentiate more for the terms "triplet-triplet energy transfer (EnT)" and "energy transfer (ET)".
Response: Thanks for the reviewer's suggestion.However, we have chosen to retain the abbreviations 'EnT' and 'ET' for energy transfer and electron transfer, respectively, as they are commonly recognized in the field.To avoid confusion for the reader, we refrained from using abbreviations in the abstract and corrected any instances where they were used inaccurately or ambiguously within the text.
4) The authors achieved developing UV-blocking OCAs at a rate approximately 10 times faster than before.And claimed that the technique could be immediately commercialized.However, the manuscript's current presentation may leave readers seeking more compelling evidence to fully substantiate this claim.
Response: This information is proprietary to the company and cannot be fully disclosed.However, according to insights from Samsung Display and its partner companies, an integrated light dose of approximately 1500-3000 mJ cm -2 is generally required to ensure the productivity of the film.This specification is corroborated by 3M product catalogs and various patents (Fig. R6).For instance, the recommended curing energy for 3M's Liquid OCA 2321, as detailed in their guidelines, is 3000 mJ cm -2 with irradiation at wavelengths ranging from 315 to 420 nm.Moreover, the feasibility of our approach is bolstered by a patent that has been filed for an acrylic resin curing system, which operates with a light dose of 2000-5000 mJ cm -2 at 365 nm (Patent JP 2020-46557 A).Therefore, drawing from these precedents, we suggest that our OCA exhibits strong potential for successful commercialization.We have included the patent reference in the revised manuscript to underscore this potential.

Fig. R6
Product catalog for 3M's OCA 2321, illustrating the cured properties of the adhesive.The catalog specifies the required light dosage for curing (3000 mJ cm -2 ), which is highlighted with a red line.
5) What about the UV blocking ability of the layers "BM and CF", "ToE", and "TFE"?In the dome of Fig. 4, it might be better to include those layers for the comparison.
Response: Thank you for the reviewer's comments.We wish to clarify that the UV light passing through the color filter (CF) and reaching the pixel layer presents a challenge due to the color filter's inadequate UV-blocking capability.
Samsung Display has recognized this issue following the establishment of the color filter on encapsulation (CoE) technology and has tasked our lab with developing an OCA with UV-blocking abilities to address it.
To begin, the black matrix ("BM") containing carbon black can absorb UV light (as stated in Patent JP 3508399B2).
However, since the BM's primary functions are to prevent color mixing and to enhance display resolution, it is positioned between RGB pixels (Fig. R7a).As a result, the BM layer does not efficiently protect RGB pixels from the vertical incidence of UV light.Secondly, the CF incorporates additives (such as pigments, dyes, solvents, and dispersants) and is typically produced using a UV-photolithography method (ACS Photonics, 6, 3132 (2019) and Appl.Opt., 59, G137 ( 2020)) (Fig. R7b).Introducing UV-blocking capabilities into the CF layer may impede the UV-photolithography process, underscoring our manuscript's emphasis on the necessity of visible-light curing methods.The touch electrode (ToE) is processed atop the thin organic layer of the thin film encapsulation (TFE) using a patterning process for a metal mesh sensor (such as ITO, Cu, Ag, etc.) with a dielectric layer (Patent US 20170153729A1, Patent US 20060097991A1, and International Journal of Precision Engineering and Manufacturing, 16, 2347 ( 2015)) (Fig. R8a).Although ITO in the ToE layer can absorb UV light in the range of 300 nm to 400 nm (Fig. R8b), its patterned mesh primarily covers electrode parts, not the RGB pixels.Thus, despite ITO's UV-blocking properties, it is likely inefficient in preventing RGB OLED degradation.Lastly, TFE layers serve to both flatten the pixel define layer (PDL) for a uniform surface and to protect RGB pixels from moisture and oxygen.Inorganic thin films in TFE layers, such as SiOx, SiNx, or SiOxNy, exhibit UV-blocking properties primarily for UV-B (290-320 nm) and UV-C (100-290 nm) regions (Fig. R8c).
However, additional protection against UV-A (320-400 nm)-which penetrates the ozone layer, reaches the Earth's surface, and significantly impacts living organisms-is still necessary to enhance the lifespan of OLED devices.
While ideally, all layers would be integrated to create an OLED device for the UV-blocking test, we have prioritized testing the OCA layer in our laboratory due to accessibility and to simplify the experimental process.In order to commercialize the OCA we developed, we are currently planning to conduct a UV-blocking test on an actual device in cooperation with Samsung Display.Response: Thank you for the reviewer's comments.In fact, we performed the peel strength test using an LS1 (AMETEK, USA) machine and adhered to the standard outlined in the Korean Industrial Standards (KS T 1028).A 10 kgf capacity load cell was employed, and the test specimens were secured using clamps designed for tensile testing.The acrylic syrup, cured between PETE film and release films, formed test specimens that were 1 cm in width.After the release films were removed, the adhesive film was bonded to the adherend using a 2 kg roller, passed over the film twice.Following a 24hour bonding period, the peel strength of the tape-type adhesive was measured.This value was calculated as an average of the strength values recorded from 20% to 80% of the working range.We have updated the SI to include these detailed testing conditions as suggested by the reviewer as follow: "The peel strength test was conducted with LS1 (AMETEK, USA) and followed the Korean Industrial Standards (KS T 1028).A load cell with a capacity of 10 kgf was used, and the test specimens were fixed by clamp used for a tensile test.
For the peel test specimens, the acrylic syrup was cured between PETE film and release films, and the tested specimens were prepared as 1 cm wide.After removal of the release films from the prepared specimens, the OCA film was attached to the adherend using a 2 kg roller (2 round trips over the film).After 24 h of attachment time, the peel strength of tapetype adhesive was measured.Peel strength was obtained from the average of measured strength values from 20% to 80% of working range.".
The reviewer appears to have misconceived the adhesion values.Contrary to what was suggested, our data, as summarized in the accompanying Table R2, actually indicate that adhesion to glass was higher, while it was lower on CPI.The OCA developed in this study, intended for use with a cover window (ultrathin glass (UTG) in smartphones, and colorless polyimide (CPI) in larger displays such as tablets-as depicted in the device structure in Fig. 1 of the manuscript, which has not yet been commercialized), may exhibit these variations in adhesion due to alterations in the monomer combination.In this study, we slightly adjusted the monomer mixture to further optimize the rheological properties, aiming to enhance the commercial viability of the OCA under development.Regardless, it's important to note that the adhesion levels typically required are in the range of approximately 1 to 4 N cm -1 .Therefore, the adhesion values we measured do not present a significant concern.Contrast Enhancement Film CEF36XX Series.Accordingly, we prepared acrylic OCAs with a thickness of approximately 50 μm in our laboratory.We acknowledge the challenge of consistently producing OCAs thinner than 50 μm.In response to the reviewer's comment, we evaluated the UV-blocking abilities of OCAs at various thicknesses by measuring UV/Vis transmittance using absorption spectroscopy (Fig. R9).We found that all films, irrespective of thickness, effectively block UV light below 400 nm.However, thicker films demonstrated lower transmittance in the visible light spectrum above 400 nm.Consequently, taking into account the requirements for display application efficiency, we have identified 50 μm as the optimal thickness for our OCAs.These insights into the correlation between the UV-blocking ability of OCAs and their thickness have been added to the revised manuscript and SI as follows:
9) The misspelling of 'TFE' is found in the Fig. 1 legend.

Response:
We have properly addressed in the revised manuscript.The authors responded to my quesfion very diligently, and I am okay with their answers.
Reviewer #2 (Remarks to the Author): Most of my comments and suggesfions are not addressed properly.I can't recommend accepfing this manuscript at the current version.

For my previous Comment 1.
There is sfill major misleading informafion in the manuscript, including the abstract, introducfion, Figs., etc., regarding disfinguishing the innovafion in the current work from their previous work.Again, I would like to point out that the photocatalyst (PC), 4DP-IPN, is not new.However, the authors sfill emphasize that "We have re-engineered the photocatalyst (PC)… " and "we delicately modified the structure of the exisfing PC to design a new PC, 4DP-IPN,…".In addifion, Figures 1 and 2 sfill contain a large amount of redundant informafion from their earlier work.

For my previous Comment 2.
The formafion of excitons during photoexcitafion primarily consists of singlet excitons, a well-established fact within the photophysical community.However, the authors' calculafion results appear to draw the opposite conclusion, suggesfing that triplet excitons, rather than singlet excitons, dominate the photocatalyst.(Note: There seems to be a discrepancy or confusion regarding whether singlet or triplet excitons are dominant in the author's Response) The author adopted the calculafion method used in their previous work (Adv.Mater., 35 2204776 (2023)), which is likely flawed.
a.The authors calculated the concentrafions of [S1] and [T1] based on the equafions S1 and S2 in the Supplementary Informafion.The rate constants are shown in Table S3.However, the kr, T1 and knr, T1 are not available in the Table, how did the authors do the calculafions?
b.The calculated results might be meaningless to esfimate the concentrafions of [S1] and [T1] in the photo-inifiafing system (PIS) if the equafions neglect the electron and energy transfer between the PCs and the acceptors (photoinifiator and UVAs).The authors menfioned that the rate constant of photoinduced ET (PET) between 4Cz-IPN and DMAEAc is kPET = 2.1 × 107 M-1 s-1, which means that, at least in the fime range longer than ~5μs, the concentrafions of [S1] and [T1] are largely influenced by the PET.Not to menfion the singlet energy transfer rate, which might be faster.If the electron and energy transfers between the PCs and the acceptors (photoinifiator and UVAs) are ignored, only the results of t=0 are useful.BTW, why is the rafio between the concentrafion of [S1] and [T1] at t=0 not equal to 1 for some PCs, e.g., 4DP-IPN, and 4-p,p-DCDP-IPN?
3. For my previous Comment 3: It's befter to differenfiate more for the terms "triplet-triplet energy transfer (EnT)" and "energy transfer (ET)".
The authors retained the abbreviafions 'EnT' and 'ET' for energy transfer and electron transfer, respecfively.Why are the abbreviafions the authors retained not the same with their previous version of manuscript?And how did they do the improvement, and where?

For my previous Comment 4
The current work employs visible light to cure the opfically clear adhesive (OCA) and claims that the technique could be immediately commercialized.I asked why, and the authors responded their approach is bolstered by a patent that has been filed for an acrylic resin curing system, which operates with a UV light source, which means the technique can't be commercialized with visible light?I'm sfill not convinced how this technique could be immediately commercialized in the current presentafion of the manuscript.In the author's response to my previous Comment 5, the authors menfioned that "Introducing UV-blocking capabilifies into the CF layer may impede the UV-photolithography process,…".But introducing UV-blocking capabilifies into the OCA is OK for the UV curing process?If so, it is sfill feasible to introduce UV-blocking capabilifies into the CF layer regarding my previous Comment 5 and Comment 7?

For my previous Comment 5
It would be highly valuable to also include the UV blocking ability of the CF layer or at least have a discussion in the manuscript.
Lastly, but certainly not least, it is imperafive to conduct the discussions based on the facts rather than relying on statements or opinions from Samsung Display.
Most of my comments and suggestions are not addressed properly.I can't recommend accepting this manuscript at the current version.

Response:
We regret that our initial revision did not fully meet the expectations of the reviewer.In this current revision, we have made every effort to address the reviewer's feedback comprehensively.Specifically, we have eliminated phrases and sentences that might lead to an overestimation of our results, thereby more clearly highlighting the study's importance and novelty.We have also addressed the reviewers' technical questions and requests in detail.We are deeply appreciative of the reviewers' time and effort dedicated to our manuscript and believe that these contributions have significantly enhanced its quality.
1.For my previous Comment 1, In the Abstract and Introduction, the authors assert that they employed a newly designed PC (Photocatalyst), specifically 4DP-IPN.However, it is worth noting that this PC, 4DP-IPN, has been previously explored in their earlier research.Throughout the manuscript, substantial effort is invested in comparing 4DP-IPN with 4Cz-IPN, both in the main text and Figure 1.This extensive comparison may inadvertently obscure the innovative aspects and significance of the current study.It is essential to emphasize that the primary distinction between the photoinitiating systems (PIS) developed in this work and their prior research lies in the choice of the photoinitiator.This pivotal distinction could be highlighted more prominently in the present manuscript.Additionally, Figures 1 and 2 contain a large amount of redundant information from their earlier work, warranting attention for a more concise and focused presentation.
There is still major misleading information in the manuscript, including the abstract, introduction, Figs., etc., regarding distinguishing the innovation in the current work from their previous work.Again, I would like to point out that the photocatalyst (PC), 4DP-IPN, is not new.However, the authors still emphasize that "We have re-engineered the photocatalyst (PC)… " and "we delicately modified the structure of the existing PC to design a new PC, 4DP-IPN,…".
In addition, Figures 1 and 2 still contain a large amount of redundant information from their earlier work.
Response: Responding to the reviewer's comments, we have revised or removed certain sentences and phrases in the introduction and abstract that could lead to an overvaluation of our results.As the reviewer correctly noted, the catalyst 4DP-IPN has been developed and utilized in previous studies.Therefore, to avoid reader confusion, we have replaced the term 'reengineered' with 'selected in our previously constructed catalyst library.' Additionally, to more clearly convey the novelty of this study, we slightly modified Figure 1b and relocated Figure 1c from the main figure to the supplementary information.Regarding Figure 2a, which depicts a mechanism already proposed in the previous paper, we considered its removal.However, we chose to keep it to aid comprehension for readers who may not be familiar with our prior work, thereby eliminating the need for them to refer to earlier publications.

#Previous version: "We have re-engineered the photocatalyst (PC) to minimize electron and energy transfer to UVblocking agents and introduced new co-initiators ~~"
#Revised version: "We have carefully selected the photocatalyst (PC) to minimize electron and energy transfer to UV- to the triplet state of various species within the photo-initiating system (PIS).However, the question arises: why is the singlet state of these materials not equally critical and deserving of discussion?
The formation of excitons during photoexcitation primarily consists of singlet excitons, a well-established fact within the photophysical community.However, the authors' calculation results appear to draw the opposite conclusion, suggesting that triplet excitons, rather than singlet excitons, dominate the photocatalyst.(Note: There seems to be a discrepancy or confusion regarding whether singlet or triplet excitons are dominant in the author's Response).The author adopted the calculation method used in their previous work (Adv.Mater., 35 2204776 ( 2023)), which is likely flawed.to misunderstandings and thus requires further clarification.As the reviewer notes, organic dyes are typically excited to the singlet excited state manifold (S n ), which then rapidly deactivates to S 1 .This is indeed the primary step occurring in our dyes.However, our photocatalysts (PCs) differ from classic dyes as they are based on thermally activated delayed fluorescence (TADF).In this process, T 1 is efficiently populated from S 1 via intersystem crossing (ISC), and subsequently, S 1 is repopulated from T 1 via reverse ISC (RISC), resulting in a cycling effect.This mechanism leads to a prolonged excited state lifetime, making these compounds particularly effective as photocatalysts (see Chem.Soc. Rev., 50, 7587 (2021)).Our primary strategy focuses on the utilization of the triplet excited state in photocatalysis for preparing OCA films.The longer lifetime of the triplet state enables a higher effective PC * concentration, thereby reducing the required PC loading compared to organic dyes that predominantly generate the S 1 state.To illustrate this, we conducted a kinetic simulation for the exciton population, enabling us to quantitatively evaluate the contributions of each excited state in the electron transfer process.Detailed explanations of the kinetic simulations are provided below.
a.The authors calculated the concentrations of [S1] and [T1] based on the equations S1 and S2 in the Supplementary Information.The rate constants are shown in Table S3.However, the kr, T1 and knr, T1 are not available in the Table , how did the authors do the calculations?

Response:
The rate constants k r,T1 and k nr,T1 were in fact neglected in the calculations; this approximation is justified for so-called "strong TADF emitters", as shown earlier by others and by us (J.Phys.Chem. A 117, 5607−5612 (2013), Chemical Society Reviews, 50, 7587-7680 (2021), The Chemical Record, 20, 831-856 (2020) and Nature Commun.14, 92 ( 2023)).The condition for strong TADF emitters is fulfilled for compounds with (i) a high total fluorescence quantum yield  F and, at the same time, (ii) a high fraction of delayed fluorescence  DF / F .Both conditions are in fact fulfilled for 4DP-IPN (Macromolecules.52, 5538 (2019)).In this case, the triplet deactivation is largely dominated by RISC.
For a comprehensive understanding, we now added a detailed description of the process to obtain each rate constant in the Jablonski diagram of the PC in the revised Supplementary materials with the corresponding references:

■ Evaluation of k r,S1 , k nr,S1 , k ISC and k RISC .
The general TADF kinetics can be obtained from the differential equations for the singlet and triplet excited state (S 1 and T 1 ) deactivation.
where k S = k r,S1 + k nr,S1 + k ISC , k T = k r , T1 + k nr,T1 + k RISC ,  is the intensity of excitation, and  is the absorption coefficient; the solutions are described below and the total intensity is obtained as The exponents A 1,2 (which correspond to the reciprocal values of the prompt/delayed PL lifetime constants, i.e., τ PF -1 and τ DF -1 , respectively) are given by The total PL quantum yield Φ PL is given as the sum of fluorescence and phosphorescence quantum yields (Φ PL = Φ F + Φ PH ), where Φ F consists of a prompt fluorescence (Φ PF ) and delayed fluorescence (Φ DF ) part; the prompt part is defined by In the presence of large number of TADF cycles, the total Φ F under steady state is obtained as, 14 where η ISC = k ISC /k S and η RISC = k RISC /k T are the efficiencies for ISC and RISC, respectively.Similarly, Φ PH is obtained as The condition for strong TADF emitters (i.e.large  F and large  DF / F ) emitters translates to k RISC >> k r,T1 , k nr,T1 , so that η RISC  1; this simplifies equation ( 13) to Furthermore, for TADF compounds with a non-negligible ΔE ST , RISC is much smaller than ISC (i.e., k RISC << k ISC ). 2 Under these conditions, with a Taylor expansion ( = √1 +  ≈ 1 +  2 for x << 1), the solutions for A 1,2 simplify to Finally, the radiative rate constant  , 1 can be estimated from the Strickler-Berg formula, which in its simplified form reads, 15,16    r,  1 ,SB = 0.667( −1  2 ) where  is the TD-DFT calculated oscillator strength of vertical absorption,  is the refractive index of solvent and  is the energy of vertical absorption and emission respectively for the lowest energetic CT transition.
In summary, the photophysical rate constants of PCs in Jablonski diagram were evaluated by experimental (i.e., prompt/delayed fluorescence decays) and computational method (i.e., TD-DFT), which each relation is simplified to where , , ,   ,   , and   have been defined earlier.
b.The calculated results might be meaningless to estimate the concentrations of [S1] and [T1] in the photo-initiating system (PIS) if the equations neglect the electron and energy transfer between the PCs and the acceptors (photoinitiator and UVAs).The authors mentioned that the rate constant of photoinduced ET (PET) between 4Cz-IPN and DMAEAc is kPET = 2.1 × 107 M-1 s-1, which means that, at least in the time range longer than ~5μs, the concentrations of [S1] and [T1] are largely influenced by the PET.Not to mention the singlet energy transfer rate, which might be faster.If the electron and energy transfers between the PCs and the acceptors (photoinitiator and UVAs) are ignored, only the results of t=0 are useful.BTW, why is the ratio between the concentration of [S1] and [T1] at t=0 not equal to 1 for some PCs, e.g., 4DP-IPN, and 4-p,p-DCDP-IPN?
Response: We appreciate the reviewer for the constructive comments.The primary aim of our kinetics simulations is to estimate the kinetics of radical production in each photoinitiation system (PIS), as clearly depicted in Figure 3e in the original manuscript.These calculations take into account all factors that could potentially affect radical generation for photopolymerization.This includes the concentration of the excited state of the photocatalyst (PC), the rates of electron and energy transfer between the PC and the co-initiator, and between the PC and UV absorbers (UVAs).By considering these factors, we were able to determine the efficiency with which each PIS can generate radicals, thereby influencing the speed of curing.
However, this simulation does not intuitively explain why PISs exhibit such radical generation kinetics.We aimed to clarify this through Figure 3d.Specifically, the cyanoarene-based PCs used in our study possess TADF properties, which efficiently generate triplet excited states.The extent to which these triplet excited states are formed significantly influences the overall curing speed.The energy transfer from the PCs' triplet states to the UVAs competes with the electron transfer to the co-initiators, impacting the curing speed.However, as the reviewer pointed out, the energy or electron transfer from PCs to the UVAs or the co-initiator is indeed competitive.Calculating the concentration of the triplet excited state without considering the co-initiators might lead to misunderstandings or confusion among readers.
To address this, we provided data on the amount of triplet excited state generated solely from the PCs, in the absence of UVA and co-initiator (see below).This approach allows readers to more clearly comprehend the differences between 4DP-IPN and 4Cz-IPN.

Fig. R1. Chemical structures and photophysical properties of UVAs including UV/vis absorption and photoluminescence (PL)
emission spectra in ethyl acetate acetate (1.0 × 10 -5 M).E (S 1 ) and E (T 1 ) were extracted from the onset of PL and onset of gated PL in ethyl acetate at 65 K, respectively.The onsets were obtained by tangential method, i.e., the intersection of the tangent, set at the high energy slope of the spectrum, with the x-axis.Because the phosphorescence of UVA-2 was not observed, therefore E (T 1 ) of UVA-2 was referred to the literature, where E (T 1 ) was estimated from the phosphorescence quenching experiments with the quencher having E (T 1 ) = 2.59 eV (Journal of Luminescence 166, 203-208 (2015)).E (T 1 ) of UVAs calculated by TD-DFT were also given.
3. For my previous Comment 3: It's better to differentiate more for the terms "triplet-triplet energy transfer (EnT)" and "energy transfer (ET)".
The authors retained the abbreviations 'EnT' and 'ET' for energy transfer and electron transfer, respectively.Why are the abbreviations the authors retained not the same with their previous version of manuscript?And how did they do the improvement, and where?Normalized intensity (a.u.) 5. For my previous Comment 5, it would be highly valuable to also include the UV blocking ability of the CF layer or at least have a discussion in the manuscript.Lastly, but certainly not least, it is imperative to conduct the discussions based on the facts rather than relying on statements or opinions from Samsung Display.
Reviewer's previous Comment 5) What about the UV blocking ability of the layers "BM and CF", "ToE", and "TFE"?
In the dome of Fig. 4, it might be better to include those layers for the comparison.
Response: Following the reviewer's argument, we added the following sentences to the introduction and results parts: "Additionally, introducing UV-blocking properties to the color filter might appear viable, but it is impractical for two main reasons: ⅰ) these filters are usually produced using a UV curing process, 9,10 incompatible with the addition of UV absorbers (UVAs), and ⅱ) visible-light curing, while a potential alternative, is ineffective because color filters inherently absorb visible light."And "Ideally, for UV blocking testing, all layers, including the color filters, should be integrated to create an OLED device.However, due to practical difficulties in fabricating color filters in the laboratory, we opted for a simplified device structure.This approach is justified, given that color filters block very little UV light."See aftached.
The further revisions addressed most of the questions, making it more ready for publication.But I still have some remaining questions and comments: 1.For the calculation of the concentrations of [S1] and [T1]: a) k r,T1 and k nr,T1 were neglected in the calculations.I agree with the authors that those two neglections are reasonable for strong TADF emitters with two requirements: a high  F and a high  DF / F .However, the  F of 4DP-IPN is only about 26% (Adv.Mater., 35 2204776 (2023)), which is not high at all.Besides, the  RISC of 4DP-IPN is as low as ~ 10 4 s -1 , which is notably lower than the strong TADF emitters like 4CZ-IPN, indicating the triplet excitons are more likely to be quenched through processes like TTA.
There was a different model (Nat.Mater.14, 330, ( 2015)), which neglected the k nr,S1 , for estimating the kinetic processes in TADF emitters.It was suggested that for non-efficient TADF emitters, e.g., 4DP-IPN, the k nr,T1 is much faster than the k RISC , implying that nonradiative decay from T1 dominates over the RISC process back to the singlet state.Therefore, it is not appropriate to neglect the contribution of k nr,T1 to estimate the k T in equation 6.
Since the dominating energy transfer process in the photo-initiating systems is the triplet-triplet energy transfer, it is crucial to include the k nr,T1 in equation 6.Therefore, it would be more rigorous to do the calculations using the models that consider the k nr,T1 (Nat. Mater. 14, 330, (2015)) or both the k nr,S1 and k nr,T1 (Nat.Photonics.6, 253, ( 2012)).
In the meantime, since the k PET between PCs and UVAs is estimated by the authors to be around 2.1 × 107 M-1 s-1, which is around 3 orders higher than either k r,T1 or k nr,T1 .So, in equation 6, calculating T1 concentration without considering k PET is meaningless to explain why 4DP-IPN is a better PC than 4CZ-IPN (as shown in Fig. R1).
b) I noticed an extra term αI in equation 5 compared to the authors' previous calculations (Adv. Mater., 35 2204776 (2023))?It seems equations 7 and 8 are derived as solutions to equations 5 and 6 without considering the αI term?BTW, what is the As in equation 10?
2. As the authors suggested, the OCA is compatible with the UV curing process, which implies that adding UVAs in color filters would not prevent their UV curing process.However, it would be helpful for readers to better appreciate the rationale behind adding UVAs to the OCA if the authors could discuss the UV-blocking ability of the color filters.E 00 were extracted from the a onset of PL and b onset of gated PL in ethyl acetate at 65 K, respectively.The onsets were obtained by tangential method, i.e., the intersection of the tangent, set at the high energy slope of the spectrum, with the x-axis.c Oscillator strengths were obtained by TD-DFT calculation.d Φ F are relatively measured against the coumarin C153. 9e Φ PF are obtained from the following relationship, Φ PF = k r,S1 / (k r,S1 + k nr,S1 + k ISC ).f Φ DF are obtained from the following relationship, Φ DF = Φ F -Φ PF .g Φ ISC are obtained from the following relationship, Φ ISC = k ISC / (k r,S1 + k nr,S1 + k ISC ).h Φ RISC are obtained from the following relationship, Φ RISC = k RISC / (k r,T1 + k nr,T1 + k RISC ).i K C are correlated with the number of TADF cycles determined from the following relationship, K C = Φ F / Φ PF . 10j Values in parentheses of both PCs were determined from the total PLQY (Φ F ) and the proportion of the integrated area of individual components in the TCSPC decay to the total integrated area. 11,12k k r,S1 was estimated via the Strickler-Berg equation. 13,14 Photophysical values were obtained following by the procedure 15,16 under assumption of k nr,S1 ~ 0 and k r,T1 ~ 0. Response: Thanks for the reviewer's careful comments.αI term is related to the extent of photoexcitation to excited states.αI term is not involved to solve ordinary differential equation to obtain the photophysical rate constants.Thus, even some publications of the TADF molecules sometimes omit αI term to describe their differential equations system (e.g., J. Phys. Chem. A 125, 8074−8089 (2021), J. Phys. Chem. C 122, 29173−29179 (2018)).However, to provide comprehensive understanding of photophysical behaviors, we have revised our descriptions including the photoexcitation term, namely, P n , fixing some typos like 'As' in equation 10.

#Revised version: ■ Derivation of kabs
Our LED setups are based on 452 nm (I 0 = 10 mW cm -2 ) LEDs, therefore with consideration of photonflux (m -2 s -1 ), P n , the rate of S n generation from S 0 via photoexcitation (i.e., S 0 → S n ), can be expressed by following equation ( 3), 20,21     =  × where  is quantum yield of the transformation for S 0 → S n ,  is cross-sectional area (100 cm 2 ),  0 is the reaction volume (5 mL), N A is Avogadro number, ℎ is Planck constant,  is frequency of the photon,  is the extinction coefficient of PCs in ethyl acetate (e.g.,  452nm = 7.8 × 10 3 M -1 cm -1 for 4DP-IPN and  452nm = 2.1 × 10 3 M -1 cm -1 for 4Cz-IPN),  is the concentration of PC (i.e., [S 0 ] = 4.75 × 10 -5 M for the 10 ppm in the synthesis of poly(EHA)) and  is the optical path length (50 μm).Because it is tricky to evaluate all the factors affecting photonflux (e.g., refractive index and surface curvature of release film), we excluded them in this kinetic simulation.Furthermore, as the internal conversion (i.e., S n → S 1 ) is highly fast, we would assume the lowest S 1 state are mainly generated.Quantum yield () for S 0 → S 1 is assumed as unity, hence, in accordance with our experimental conditions, equation ( 3) can be converted to equation ( 4).

■ Evaluation of kr,S1, knr,S1, kISC and kRISC.
The general TADF kinetics can be obtained from the differential equations for the singlet and triplet excited state (S 1 and T 1 ) deactivation.The exponents A 1,2 (which correspond to the reciprocal values of the prompt/delayed PL lifetime constants, i.e., τ PF -1 and τ DF -1 , respectively) are given by 2. As the authors suggested, the OCA is compatible with the UV curing process, which implies that adding UVAs in color filters would not prevent their UV curing process.However, it would be helpful for readers to better appreciate the rationale behind adding UVAs to the OCA if the authors could discuss the UV-blocking ability of the color filters.

Response:
We appreciate the reviewer's comment.To address any potential misunderstanding, we would like to clarify that the conventional production of OCAs using UV-photoinitiators is not compatible in the presence of UVAs due to their high absorbance of UV light.Consequently, we opted for visible-light irradiation, successfully obtaining UVblocking OCAs.In our revised introduction, we aim to present our argument more clearly and emphasize the potential of UV-blocking color filters.

Previous version:
"This innovative approach necessitates the use of an adhesive material possessing UV-blocking characteristics to replace the polarizer's function of safeguarding the panel from external UV radiation.However, the preparation of UV-blocking adhesive proves to be a complex task due to the UV-blocking agent's high absorbance of UV light, which impedes the effectiveness of the conventional photocuring method that relies on a UV-photoinitiator.Additionally, introducing UV-blocking properties to the color filter might appear viable, but it is impractical for two main reasons: ⅰ) these filters are usually produced using a UV curing process, 9,10   incompatible with the addition of UV absorbers (UVAs), and ⅱ) visible-light curing, while a potential alternative, is ineffective because color filters inherently absorb visible light." "A simple approach is to utilize visible light instead of UV light for curing, with the aid of a photosensitizer that efficiently absorbs visible light. 11-26However, this method presents a challenge as it substantially reduces the film's optical transparency in the visiblelight range. 27-30"

Revised version:
"This innovative approach necessitates the modification of existing layers to possess UV-blocking characteristics to replace the polarizer's function of safeguarding the panel from external UV radiation.To incorporate a UV-blocking ability into the target layer, its fabrication process need to be compatible with UV absorbers (UVAs) considering the potential hindrances of UV-light curing due to their high UV-light absorbance.One strategy would involve imbedding UVAs into the color filter layer.However, since the fabrication of these filters typically relies on UV-light curing, 9,10 substantial efforts are required to introduce UV-blocking capabilities into the color filters (see below).Optically clear adhesives (OCA) layer could be regarded as an alternative, but the preparation of UV-blocking adhesive still proves to be a complex task in the presence of UVAs, which impedes the effectiveness of the conventional photocuring method that relies on a UV-photoinitiator.
A simple approach is to utilize visible light instead of UV light for curing, with the aid of a photosensitizer that efficiently absorbs visible light; 11-30 while visible-light curing can technically be used to add a UV-absorber to the color filter layer, it is likely inefficient and thus impractical, mainly due to the high absorption of visible light by the pigments. 9,10"

Fig. R1
Fig. R1 Schematic illustration of 180° folded OCA in the foldable display.

." 3 )
For the peeling test, what is the major debonding failure mode, cohesive failure or adhesive failure?Can the authors provide experiment pictures showing the peeling process?Response: We appreciate the referee's thorough comments.Generally, the failure modes in peeling tests are categorized as adhesive failure, cohesive failure, and structural failure (Adhesives Technology Handbook (Second Edition), 1-19 (2009)).Adhesive failure typically occurs where there is insufficient physical bonding strength between the adhesive and the adherend, resulting in sticky-slip.In contrast, cohesive failure originates from a breakdown of the intermolecular bonding strength (i.e., cohesive force) within the adhesive itself.Structural failure refers to the failure of the entire structure, which implies an unbreakably strong bond and adhesion.In light of these failure modes, we observed no residue on the substrate during the peeling tests conducted in this work, indicating an absence of failure.

Fig. R3
Fig. R3Images show a representative peel strength test conducted on a glass substrate.To affix the specimens (OCAs) onto the glass substrate secured in the clamp, we used a commercially available acrylic adhesive, 3M Scotch™ Tape 810.The peel strength of the prepared OCAs was measured after 24 hours of attachment.

Fig. R4 ■
Fig. R4 Schematic illustration of this work.Schematic illustration of this work; here EnT and ET denote energy transfer and electron transfer, respectively.Image of the prepared UV-blocking OCA is shown.

Fig. R5
Fig. R5 Comparison of contributions to electron transfer from the singlet and triplet excited states of 4DP-IPN and 4Cz-IPN.(a) The photoinduced electron transfer rate constants (kPET, M −1 •s −1 ) between the PCs and co-initiators were measured.These constants for both the singlet and triplet excited states of the PCs were determined by monitoring the changes in prompt fluorescence (PF) and delayed fluorescence (DF) using time-correlated single-photon counting (TCSPC) techniques at an excitation wavelength (λex) of 377 nm and a detection wavelength (λdet) of 520 nm.The photoluminescence (PL) decay spectra at room temperature were obtained from degassed solutions of PCs in ethyl acetate (1.0 × 10 −5 M) with varying concentrations of quenchers.(b) The simulated relative populations of the lowest singlet (S1) and triplet (T1) excited states of the PCs in ethyl acetate (4.75 × 10 −5 M) were calculated under the photostationary state during illumination with a 452 nm LED at an intensity of 10 mW•cm −2 .
b ■ Scheme of kinetic simulation of PC concentration in excited state under photostationary state a ■ Measurements of photoinduced electron transfer rate constant (k PET , M -1 •s -1 ) between PC and co-

Fig
Fig. R7 (a) Schematic of a foldable display device structure featuring a UV-blocking optically clear adhesive (OCA) film.The acronyms TFT, BM, CF, ToE, TFE, and PDL stand for thin-film transistor, black matrix, color filter, touchpanel on encapsulation, thin-film encapsulation, and pixel define layer, respectively.(b) Process flow schematic for the UV-photolithography-based fabrication of multispectral filter arrays (MSFA).The contents of Fig. R7b have been reproduced from the cited publication (ACS Photonics, 6, 3132 (2019)).

Fig
Fig. R8 (a) Schematic representation of touch-panel on encapsulation (ToE) and thin-film encapsulation (TFE) layers in an OLED device.(b-c) Optical properties of the components comprising the ToE and TFE layers: (b) UV/visible transmission spectra of materials (e.g., metal mesh and ITO) in the ToE layer, and (c) absorption spectra of SiO2 in the TFE layer.The contents of Fig. R8(b-c) are reproduced from the cited publications (International Journal of Precision Engineering and Manufacturing, 16, 2347 (2015) and IOP Conference Series: Materials Science and Engineering, 310, 042029 (2019)) the synthesis of acrylic syrup, all reaction mixtures were irradiated by 455 nm LEDs at RT and then the synthesized acrylic syrup was casted to a thickness of 50 μm between silicon-treated release films (100 μm-thick), then they were cured under 452 nm LEDs.The data of Entry 1 have been reproduced from the cited publication (Adv.Mater., 35, 2204776 (2023)).7)What advantages does the design of the OCA with UV-blocking ability offer over alternative strategies, like incorporating UVAs into the BM-CF layer?This alternative approach might potentially expedite OCA development significantly, given its potentially reduced demands on the choice of PC and initiators.Response: As outlined in the manuscript, the technology we describe is known as Color Filter on Encapsulation (CoE) ( https://www.oled-info.com/samsung-display-announces-polarizer-free-eco2-oled-technology).Developed to replace traditional polarizers, CoE aims to enhance energy efficiency.It has already been implemented in the Galaxy Z-Fold and is anticipated to be incorporated into future smartphones and other devices.Competitors of Samsung Display, i.e.BOE and Tianma, are also diligently working to refine this technology.Given that the CoE production process has been established and is currently scaled for mass production, the integration of a UV-blocking agent into the color filter may be technically feasible but could incur significant development and process costs.Consequently, this might lead to financial challenges.[REDACTED] 8) How does the UV-blocking ability depend on the thickness of the OCA?Response: Thank you for the reviewer's constructive comments.The thickness of film layers such as cover film, OCA, and OLED in foldable displays typically ranges from 25 to 150 μm to ensure the necessary mechanical properties, including flexibility, as documented in SID Symposium Digest of Technical Papers 48,938-941 (2017)  and the 3M OCA (50 mm) UV-blocking OCA (100 mm) UV-blocking OCA (150 mm) UV-blocking OCA (200 mm)

■■
Comparison of curing rate in acrylic adhesive film without or with UVAs b Comparison of curing rate in acrylic adhesive film without or with UVAs b

■■
Comparison of curing rate in acrylic adhesive film without or with UVAs b Comparison of curing rate in acrylic adhesive film without or with UVAs b Counts of radical generation (RG) d[RG] dt =  PET,1 [T 1 ][HNu254] +  ET,2 [DA][HNu 254] +  disso.[BR] +  amino.[DRC] b) I noticed an extra term αI in equation 5 compared to the authors' previous calculations (Adv.Mater., 35 2204776 (2023))?It seems equations 7 and 8 are derived as solutions to equations 5 and 6 without considering the αI term?BTW, what is the As in equation 10?

Comparison of original and revised figures: 2
"Here, we have developed a highly efficient PIS that operates under visible-light irradiation, allowing for the production of UV-blocking OCAs at a much faster rate compared to the previously reported PIS.31Through a mechanistic analysis of existing PIS, we found that electron transfer (ET) and triplet-triplet energy transfer (EnT) between the PC and the UVAs ~~" .For my previous Comment 2, The molecules are initially excited to the singlet state and subsequently transition into the triplet state in the photoexcitation.The emphasis in this work primarily centers on the design principles pertaining blocking agents and have chosen co-initiators ~~"#Previous version: "Here, we develop a highly efficient PIS that operates under visible-light irradiation, allowing to produce UV-blocking OCAs at a fast enough rate for commercialization.Through a mechanistic analysis of existing PIS, we found that electron transfer (ET) and energy transfer (EnT) between the PC and the UVAs ~~"#Revised version:#Previous version: "In pursuit of resolving the issue, we employed the systematic molecular catalyst design principles proposed in our previous study.34Thismethodologyutilizes~~" #Revised version: "To resolve the issue, we screened the PC candidates established in our previous study, utilizing the systematic molecular catalyst design platform.34Thisplatform enables ~~ #