Polyampholytic Graft Copolymers as Matrix for TiO2/Eosin Y/[Mo3S13]2− Hybrid Materials and Light‐Driven Catalysis

Abstract An effective strategy to enhance the performance of inorganic semiconductors is moving towards organic‐inorganic hybrid materials. Here, we report the design of core–shell hybrid materials based on a TiO2 core functionalized with a polyampholytic (poly(dehydroalanine)‐graft‐(n‐propyl phosphonic acid acrylamide) shell (PDha‐g‐PAA@TiO2). The PDha‐g‐PAA shell facilitates the efficient immobilization of the photosensitizer Eosin Y (EY) and enables electronic interactions between EY and the TiO2 core. This resulted in high visible‐light‐driven H2 generation. The enhanced light‐driven catalytic activity is attributed to the unique core–shell design with the graft copolymer acting as bridge and facilitating electron and proton transfer, thereby also preventing the degradation of EY. Further catalytic enhancement of PDha‐g‐PAA@TiO2 was possible by introducing [Mo3S13]2− cluster anions as hydrogen‐evolution cocatalyst. This novel design approach is an example for a multi‐component system in which reactivity can in future be independently tuned by selection of the desired molecular or polymeric species.


Introduction
Global challenges such as climatec hange and overconsumption urgently requiret he supply of sustainable cleane nergy.
This couldb et ackledb yt he utilizationo fh ydrogen as as econdary,c arbon-free energy carrier. Therefore, tremendouse fforts have been invested in the direction of production,s torage, and delivery of hydrogen. [1] Light-driven catalytic hydrogen evolution is af avourable carbon-neutral method, exploiting solar energy. [2] The first example of light-induced water splitting was described in 1972 by utilisationo faTiO 2 electrode, thoughi tw as limited to irradiation with UV light. [3] Althoughb eing chemically stable, non-toxic, and al ow cost material, light harvesting using abroader solar energy spectrumi s desirable. [4] Recent approaches towardsh ydrogen evolution are therefore targeting visible-light harvesting throughb and gap engineering, [5] utilization of photosensitizers, [6] introduction of co-catalysts, [7] electron relays such as polyoxometalates, [8] or the use of novel semi-conducting materials. [9] In this context,T iO 2 as potential candidate can be tuned for visiblelight-driven hydrogen evolution by the addition of different cocatalysts and sensitizers. Recently,c ombining TiO 2 and molybdenum sulfide was identified as ap romising composite system. [4c, 10] Molybdenumsulfides (MoS x )a s( co)catalystsare gaining considerable interestd ue to low cost, long-term stability,a nd earth-abundance. In that regard both, 2D structured MoS 2 + x as well as molecular molybdenum sulfides (thiomolybdates), for example, [Mo 3 S 13 ] 2À or [Mo 2 S 12 ] 2À have been investigated in combination with different sensitizers and (co)catalysts. [4b, 11] Thiomolybdates typically carry ah ighn umber of active sites and allow homogenous hydrogene volution with high turnover numbers (> 41 000). [11b, c, 12] With respect to multi-componenth ybrid materials, immobilization of different compounds within suitable matrices plays a key role in connecting individual elements of light-driven catalytic systems and has been realized on solid substrates such as carbon nanomaterials, [13] metal oxides [14] or semiconductors such as p-Si. [15] Although being well soluble in aqueous environmenta nd exhibiting potential binding sites for both different catalysts and sensitizers, only few examples focus on soft matrices based on polyelectrolytes.Romanenko et al. described the preparation of block copolymer membranes, where the molecular catalyst[ Mo 3 S 13 ] 2À and photosensitizer [Ru(bpy) 3 ] 2 + were immobilized using positively charged groups alongt he poly(N,N-dimethylaminoethyl methacrylate) (PDMAEMA) segment. [16] Besides, polyelectrolyte-based hydrogels are promising scaffolds and in this regard Weingarten et al. [17] and H. Sai et al. [18] attached perylene monoimide as wella ss uitable catalysts for efficient hydrogen production. Also 'free' polymers could molecularly interactw ith catalysts as it has been shown by Hu et al.,e xploiting conjugated polyelectrolytes for the interaction with Pt catalysts in hydrogen evolution reactions, [19] or double-hydrophilic block copolymers as templates for CdS nanoparticles. [20] Polyampholytic polydehydroalanine (PDha) is as uitable template featuring ah igh density of functional( charged) groups and strongly interacting with metal oxides, metal nanoparticles and dyes in water,w hich led us to the assumption that this is ap romising matrix for light-driven catalysis. [21] PDha exhibits both positively charged amino groups as well as negatively chargedc arboxylic acid moieties for the specific interaction with various compounds. Furthermore, it was found to be a platform for modification reactionst oo btain materials with tailoreds olubility,f or example, as sensors, smart dispersants, or templates. [21c, 22] Besides this, PDha-g-PEG was already successfully used as dispersant for aw ater-insoluble perylenebased photosensitizer. [23] In this regard, we prepared tailormade PDha graft copolymers for the application in visible-light driven hydrogen evolution. We herein introduce an ovel catalytic system consisting of promising and low cost catalysts TiO 2 and [Mo 3 S 13 ] 2À ,a sw ell as Eosin Y( EY) as dyes ensitizer in the presenceo ft he sacrificiala gent triethanol amine (TEOA), revealing > 800 fold increased activity compared to bare EY/TiO 2 hybrids. This is realized by the attachment of phosphonic acid side-chains for increased solution stability againsts edimentation and as strong anchor groups for TiO 2 . [15] The overall combination not only leads to visible light-driven catalysis, but also enablest he physical combination of materials, which has not been possible before.T he mere combination of TiO 2 and EY has already been reported, but efficient hydrogen evolution strongly depends on the way of interaction due to insulating, quenching and stabilityi ssues. [1c, 24] On the other hand, while EY was successfully used as ap hotosensitizer together with MoS 2, [25] it failed in case of [Mo 3 S 13 ] 2À presumably due to rather weak interactions with the catalyst. [11b] Results and Discussion

Synthesis of phosphonicacid modified graft copolymers
In our earlier work we have shownt hat post-polymerization modification of PDhab yg raftingi sap owerful synthetic route to fine-tune solution properties and functionality. [21c, 22a] To develop the applicability of PDha as ac oating for TiO 2 NPs, both increasing its water solubility at pH values < 7a nd the attachment of an additional, strong anchor group was desired. Therefore, PDha with an averageo f6 0r epeat units was used as ar eactive backbone, obtained from deprotection of PtBAMA (M n = 13 200, = 2.55), and n-propylp hosphonica cid acrylamide (PAA) was successfully graftedv ia an aza-Michael addition (Figure 1 A).
The modifier was synthesized in accordance with the protocol of Hu et al. [26] Hereby, 66 %o ft he monomer units were functionalized when 5equivalents of the acryl amide were used. The degree of functionalization was determined from 1 HNMR spectroscopy in accordance with our previous work. [22a] As expected, the obtained graft copolymer was soluble over the entire pH-range as well as in methanol. NMR ( 31 P, 1 Ha nd 13 C) spectroscopy( Figure 1A,B ;F igure S2) showst he presence of both the PDha backbone as well as phosphonic acid side groups and SEC traces reveal an arrowing of the elution traces in comparison to the pristine polymer,w hile still maintaining am onomodal distribution. Regarding the multiple peaks in the 31 PNMR spectrum being in close proximity,w e assume ar andom distribution of the side-chains alongt he backbonel eadingt od ifferent chemical environments of the phosphonic acid groups. Besides that, protonation might also play ar ole.

Formation of TiO 2 based core-shell hybrids
The amino and carboxylic acid moieties of PDha are exceptionally strong ligands for inorganic NPs as already proven for stable dispersionsw ith iron oxide, gold, silver,o rA g/Au nano- alloy particles. [21a, c, 27] However,t he readily grafted phosphonic acid side-chains are additional strongl igandsf or TiO 2 surfaces and besides that they implement pH solution stability as well as additional negative charges, the latter possibly facilitating proton transfer processes. [28] Therefore, we utilizedP Dha-g-PAA as coating for TiO 2 NPs using simple ultra-sonication ( Figure 2A)a nd the successful formation of core-shell hybrid materials was provenb yt hermogravimetric analysis (TGA),t ransmission electron microscopy (TEM), X-ray photoelectron spectroscopy( XPS) and dynamic light scattering (DLS) (FiguresS3-S6).
To determine the amount of organic shell material, TGA was measured under air for the pristineT iO 2 ,P Dha-g-PAA and the hybrid materials with different polymer to TiO 2 ratios (5:1 and 15:1 (w/w) polymer/TiO 2 ,F igure S3). Any free polymer was removed by dialysis and three washing steps after centrifugation of the solidm aterial. For the TiO 2 NPs aw eightl oss of around 17 wt %w as observed between 30 8Ca nd 450 8C, which is mainly assigned to the loss of water.I nc ase of PDha-g-PAA three weight-loss steps wereo bserved, the first step between rt and 140 8Ca sar esult of loss of residual water,a nd two further steps at 290 and 470 8Ca saresult of polymer decomposition. No furtherw eight loss is observed at T > 600 8C( residual mass:2 0wt%). Regarding the hybrid material, similard ecompositions teps were observed between 290 and 600 8C, providing evidencef or the presenceo ft he polymer shell.H igh polymer contents of 29 wt %a nd 66 wt %w ere calculated, respectively.T oc onfirmt hese results, TEM images of PDha-g-PAA@TiO 2 (Figure 2a nd Figures S4 and S5) show individual TiO 2 NPs with an average radius of around 11 nm and an organic layer with at hickness of several nm is clearlyv isible. Additionally, in DLS an increase of the hydrodynamic radius (R H ) from 11 to 38 nm waso bserved after shell formation (Figure S3). We attribute this significant increase in size to multilayer formation and chain expansion of the grafted polymers.
[Mo 3 S 13 ] 2À À as cocatalyst modification To furthero vercome the recombination of photoexcited carriers, cocatalyst modification was explored for the PDha-g-PAA@TiO 2 core-shell hybrid system. Thereby,a saco-catalyst we used [Mo 3 S 13 ] 2À clusters in accordance with our earlier work. [16] Hereby, PDha-g-PAA on one hand solubilizes negatively charged molybdenum sulfide, ando nt he other hand is anchored on the surface of the TiO 2 nanoparticles (Figure 3A), therefore bringingb oth building blocks in close proximity while at the same time mediating solubility.
The resulting[ Mo 3 S 13 ] 2À @PDha-g-PAA and [Mo 3 S 13 ] 2À @PDhag-PAA@TiO 2 hybrid materials were characterized via DLS and TEM (Figure 3a nd FigureS3, S4). TEM data indicates small aggregates with the size of only af ew nm for [Mo 3 S 13 ] 2À @PDhag-PAA ( Figure 3B). After successful solubilisationo f[ Mo 3 S 13 ] 2À , TiO 2 NPs were added before furthers onication and the resulting three-component hybrid material ([Mo 3 S 13 ] 2À @PDha-g-PAA 15 @TiO 2 )w as subjected to TEM investigations ( Figure 3C,D and Figure S9), where in addition to the above-described coreshell nanoobjectsa dditional dark spots with different contrast could be found. We ascribe these likely to the presence of the co-catalyst as schematically shown in Figure 3A and Figure S8. However, DLS reveals as ignificant increase of the R H of the three-component hybrid material compared to PDha-g-PAA@TiO 2 (15 w/w) from 38 to 82 nm, which could be the result of multilayer formation, chain expansion, and some secondary aggregation taking place.
In addition, z-potential measurements were carriedo ut and while the overall charge of the TiO 2 NP was found to be slightly negative with À8 AE 2mV, it decreased to À20 AE 1a fter addition of the graft copolymer.W ea scribe this to the negatively chargedc arboxylic acid and phosphonic acid groups. After decoration of the core-shell materialw ith anionic [Mo 3 S 13 ] 2À the z-potential became À32 AE andf inally À37 AE 1a fter attachment of EY,i ndicating successive incorporation of the different compounds. This is noteworthy as electrostatic repulsion might alsob ee xpected-nevertheless, additional (attractive) interactions, for example, with exposed -NH 3 + functional groups from the PDha backbone or of hydrophobic nature seem to favor binding of the individual components. Indeed, the amine groups of the polyampholytic backbone shouldb e protonated at the corresponding pH of 7-8, although an overall negative charge is observed. [21c, 22a, 29] In order to furtheri nvestigate the interactions between PDha-g-PAA with [Mo 3 S 13 ] 2À and TiO 2 ,X -ray photoelectron spectroscopy (XPS) was used. The C1s, P2p, Mo 3d, Ti 2p, N1s, and O1ss pectra obtainedb yX PS for differents amples are shown in Figure 3E-G and Figures S5,S 6. In the Ti 2p spectra, the doublet was assigned to TiO 2 ,w ith Ti 2p 3/2 at ab inding energy of % 459 eV and Ti 2p 1/2 at ab inding energy of % 464.5 eV.T he binding energy difference of 5.5 eV between those two peaks of TiO 2 corresponds well to literature values (DE from 5.5 to 5.8 eV). [30] Following grafting of PDha-g-PAA, no significant changes areo bserved in the Ti 2p spectrum. After grafting of the polymer,aP 2p signal was detected, which was not observed for the TiO 2 reference particles. This confirms the presence of the phosphonic acids on the modified TiO 2 particles. Interestingly,t he binding energy of phosphonic acid is shifted towards lower binding energies upon graftingo nto TiO 2 from 133.8 eV forf ree PDha-g-PAA to 132.9 and 133.4 eV for PDha-g-PAA@TiO 2 and [Mo 3 S 13 ] 2À @PDha-g-PAA@TiO 2 ,r espectively ( Figure 3E-G). The observed downshift is another indication for successful grafting using phosphonic acid as anchoring group. [30a, 31] Visible-light driven hydrogen evolution Light-driven catalytic H 2 evolution performances of the samples were evaluated in the presence of TEOA (0.5 m)a ss acrificial reagentu nder LED light irradiation (530 nm (with AE 50 nm; 281 mW; 330 mA;3 .1 V) in a3 D-printed irradiationr eactor with af an ( Figure S1). To investigate the effect of each compound of our system different testsw ere made. Figure 4A shows the time courses of H 2 production forT iO 2 ,T iO 2 /EY,a nd EY/PDha-g-PAA/TiO 2 with differentw eight percent( wt %) of polymer.T iO 2 withoutE Ya nd physicalm ixtureso fT iO 2 with EY show no or trace H 2 evolution activity under these conditions. Wec ould explain this by inactivity of TiO 2 under visible light irradiation and the lack of firmly bound EY to the TiO 2 surface. [32] Besides low loading of TiO 2 via aw eak ester-like linkage, physically adsorbedd yes tend to desorbi nto the solution during irradiation, leadingt oq uick degradation due to the formation of unstable anion radicals (EY À C), and decreasing efficiency in H 2 evolution catalysis. In sharp contrast, coating TiO 2 with PDha-g-PAAs ignificantly increased the H 2 production under visible light. Overall, during irradiation, EY/PDha-g-PAA/TiO 2 with 5:1 and 15:1 of PDha-g-PAA to TiO 2 (w/w) exhibits high stability for hydrogen production with average rates of 0.301 and 0.276 mmol g À1 h À1 ,respectively.
The enhanced activity of EY/PDha-g-PAA/TiO 2 is attributed to an effective EY loading by PDha-g-PAA and stable fixation of EY within the polyampholyte shell. The maximum loading amount of EY by PDha-g-PAA and TiO 2 NPs was evaluated by a UV/Vis study ( Figure S7 and Ta ble S1). Our data revealed capacity values of 151 mgmg À1 for TiO 2 ,a nd 102 and 76 mgmg À1 for EY/PDha-g-PAA/TiO 2 with 5:1a nd 15:1 (w/w) of graftc opo- lymer,r espectively.T he maximum loading capacity for the EY/ PDha-g-PAA/TiO 2 coating materialr esembled the capacity of PDha-g-PAA as well, whereas it was slightly lower comparedt o pristine TiO 2 ,p robably due to the rather high surface area of TiO 2 .N evertheless, this might also enhance the fast decomposition of dye during irradiation, leading to lower hydrogen evaluation rates. On the other hand, although PDha-g-PAA/TiO 2 shows lower capacity for EY adsorption,i ti ss eemingly enough adsorbed and stable to inject the electron thought he graft copolymer shell to the TiO 2 core.
Finally, previous studies by Streb et al. [11c] indicated that [Mo 3 S 13 ] 2À shows higher catalytic activity in methanol/water mixtures compared to pure water.T herefore, we also investigated our EY/PDha-g-PAA/(TiO 2 /[Mo 3 S 13 ] 2À system in am ethanol/water (1:1) mixture under otherwise unchanged conditions. As shown in Figure 4B,w eo bserve increased H 2 evolution rates in methanol-water as well as highera pparent stability (i.e. prolonged reactivity compared to the system in water as exclusive solvent). This is in line with previouss tudies, which suggested that ligand exchange on [Mo 3 S 13 ] 2À in water is a major deactivation pathway. [11c] Indeed, our data indicates that [Mo 3 S 13 ] 2À in water shows promising activity but fast catalyst deactivation because of complete exchange of the terminal disulfides resulting in decreased catalytic activity. In contrast, in methanol-water mixtures, significantly higher reactivityi sa chieved by stabilizing highly active catalytic species. We could attributet his enhancement to partial exchange of one or two terminald isulfidesw ith aqua ligands which leads to the formation of more active species. Comparedt oe arlier studies where EY was covalently or electrostatically graftedt ot he surface of TiO 2 ,t his work demonstrates that PDha-based polyampholytic graft copolymers are as imple, tunable, and effective method to achievestable EY sensitization on TiO 2 with enhanced activity and stability for light-driven H 2 evolution.

Conclusion
In summary,w er eport the successful grafting of at ailor-made polyampholytic graft copolymer to TiO 2 nanoparticles, which enabledt he binding of EY photosensitizer and [Mo 3 S 13 ] 2À hydrogen evolution cocatalysts. Our results revealt hat this is a straightforward approachf or the preparation of at unable and versatile soft matter matrix which can effectively co-integrate severalm olecular componentsr elevant for light-driven catalysis. The main role of the graft copolymer is to provide close proximity and the potentialt oi nteractf or all individual com-ponents.S pecifically,w ef ind that the obviousi mprovement of the light-driven catalytic activity for hydrogen production was found by immobilizing [Mo 3 S 13 ] 2À clusters, reaching TONs > 500. Ours trategy indicates that we can use polyampholytic graft copolymers to improve and regulate different molecular catalytic systems by immobilization on TiO 2 .T his strategy introduces as ystem, which could be valuable for other molecular catalysts, dyes or semiconducting sensitizers.