Tuning PtII‐Based Donor–Acceptor Systems through Ligand Design: Effects on Frontier Orbitals, Redox Potentials, UV/Vis/NIR Absorptions, Electrochromism, and Photocatalysis

Abstract Asymmetric platinum donor–acceptor complexes [(pimp)Pt(Q2−)] are presented in this work, in which pimp=[(2,4,6‐trimethylphenylimino)methyl]pyridine and Q2−=catecholate‐type donor ligands. The properties of the complexes are evaluated as a function of the donor ligands, and correlations are drawn among electrochemical, optical, and theoretical data. Special focus has been put on the spectroelectrochemical investigation of the complexes featuring sulfonyl‐substituted phenylendiamide ligands, which show redox‐induced linkage isomerism upon oxidation. Time‐dependent density functional theory (TD‐DFT) as well as electron flux density analysis have been employed to rationalize the optical spectra of the complexes and their reactivity. Compound 1 ([(pimp)Pt(Q2−)] with Q2−=3,5‐di‐tert‐butylcatecholate) was shown to be an efficient photosensitizer for molecular oxygen and was subsequently employed in photochemical cross‐dehydrogenative coupling (CDC) reactions. The results thus display new avenues for donor–acceptor systems, including their role as photocatalysts for organic transformations, and the possibility to introduce redox‐induced linkage isomerism in these compounds through the use of sulfonamide substituents on the donor ligands.


Introduction
Group 10 metalsi nt heir d 8 electronic configuration have served to synthesize ar ange of donor-acceptor metalc omplexes. [1,[2][3][4][5] Thesec ompounds are usually characterized by intense ligand-to-ligand charge transfers (LL'CTs), which impart unique photochemical and photophysical properties on the resulting metal complexes. [2,6] Applications range from dye-sensitized solar cells to small molecule activation and catalysis. [4,7,8] To obtain ac ompound with as trongL L 'CT transition, it is desirable to use as trong p-acceptor and as trong s-a nd/or p-donorl igand, with favorable orbitale nergies. Figure 1s hows prototypical ligandst hat have been successfully employed in this regard. Well-established acceptor ligands are 2,2'-bipyridine (bpy), phenylazopyridine( pap), and also phenyliminomethylpyridine (pimp), with the latter only sparingly used in the construction of such systems. [9][10][11] Another interestinga pplication for compounds with optoelectronically switchable properties are electrochromic devices. [6c, 12] As donor ligands, catecholate/semiquinone/quinone ligands have been well-established due to their strong donor properties and well-defined redox behavior. [7,13] Ac hange in oxidation state for this ligand system results in drastically alteredd onor and acceptorp roperties with the catecholate ligand being a strong p-a nd s-donor and poor acceptor ligand, whereas the fully oxidized quinone is as trong p-acceptor andav ery weak p-donor ligand. If the oxygen donora tom is replaced by an isolobal [N-R] residue, the quinones can be sterically and electronically tuned rather easily. [14] All of the depicted ligandsa re redox-active and thus can act in ap otentially non-innocent manner, when coordinated to am etal center.T he stabilization of additional charges or charges eparationb ecomes important, if one wishes to harvest solar energy.O ur previous work revealed that such compounds show interesting andd iverse reactivity,ifdifferent donor ligands were employed. [3,4] Special emphasis is put on the redox-induced reactivity of the systems featuring o-bis(sulfonamide) ligandsi nt his work. Although this ligand class was described for the first time more than half ac enturya go, [15] the applicationo ft his highly tunable ligand class is stillr ather limited [16,17] and only one platinum complex has been reported. [18] For the mostp art, these reports discuss fundamentals tructurala spects of the complexes.R ecently,w er eported mono-and dinuclearc obalt(II) complexes with chelating and bridging bis(sulfonamido)benzene ligandsr esulting in air-stable single molecular magnets with high switching barriers,h ighlighting the potential of this ligand class. [19] Perutz and co-workers studied rhodium(III)c omplexes with symmetrically and asymmetrically sulfonylated bis(amido)benzenes for transfer hydrogenation,s howingt he catalytic applications of these ligands. [20][21][22] Interestingly,t hey also observed the dimerization of the aforementioned rhodium compounds, in which the oxygen atoms of the sulfonyl group bridge two rhodiumc enters. [20] Kavallieratosa nd co-workerso bserved the formation of coordination polymers using lead(II)s alts, emphasizing the versatile and dynamic coordinationc hemistry these ligandsmay engage in. [17] Results and Discussion Synthesis and structural characterization Complexes 1-5 were prepared by following ap reviously established route (Scheme1). [3,5] The ligandsw ere prepared by reportedr eactions. [23,24] The reactiono fp henyliminomethylpyridine (6)w ith (dmso) 2 PtCl 2 yieldedt he platinum dichloride complex 7 in good yield. In the presence of triethylamine, the respective quinoid ligandsH 2 Q x were deprotonated under inert conditions in acetonitrile to give the title complexes in low to acceptable yields, as shown in Scheme 1. The compounds were initially characterized by meanso f 1 H-and 13 CNMR spectroscopy,m asss pectrometry,a nd elementala nalysis. All complexes are stable towards air and moisture in the solid state and in solution and can be stored for several months withoutdetectable decomposition.
For complexes 1 and 5,t wo regioisomers can be formed; however,o nly one isomer was isolated in contrastt op revious studies. [3] For complex 5,e ven four possible isomersa re conceivable, ift he position of the methylsulfonyl group relative to the plane spanned by the two binding pockets and the platinum center is taken into account.H owever,t he barrier for the rotationi sp robablys ol ow that the isomersi nterconvert too quickly at room temperature. As ystematic screening of various reactionc onditions (time, temperature, solvent, and type of base) did not result in the formation of the other regioisomer or am ixture of both isomers. This is not observed for the strongly relatedp ap (phenylazopyridine) complexes of platinum and palladium, whichwehave reported on earlier. [3] This result is quite surprising; however,i tm ay be partly explained, if the trans-influence of the pyridine and imino fragment are compared. From the crystal structure of (pimp)PtCl 2 (7), we can see that the platinum-chloride bond in the transposition to the imino functioni ss lightly elongated compared with the platinum-chloride bond in the trans-position to the pyridine. Assumingthat the imino functionisanoverall stronger donor than the pyridyl fragment, the first substitution is fa- Scheme1.Syntheses of the precursor( pimp)PtCl 2 (7)a nd the complexes 1-5 (p-Tol = para-tosyl). Chem. Eur.J.2020, 26,1314-1327 www.chemeurj.org 2019 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim vored here. The aminosulfonyl function is more acidic and thus will likely coordinate first, whichw ould explain the stereochemistry,w ith the amidosulfonyl function trans to the imino group, resulting in complex 5.T he same argument can be used for complex 1,j ust that the difference in acidity for the hydroxy groups is less pronounced. This finding is reproduced by DFTcalculations, which predict the experimentally isolatedi somer to be 0.2 eV lower in energy for pimp and thus to be thermodynamically more stable. This is in contrast to an energy differenceo f0 .02 eV for the pap derivatives. [3] This suggestst hat as mallc hange from the azo group to the imino group can have as ignificant influence on the stereochemistry of the resulting compound.
Additionally,a ll complexes and precursors were characterized by means of single-crystal X-ray diffraction (see Figure 2). This was especially useful for the determination of the stereochemistry of 1 and 5.T he single crystals wereo btained by either vapor diffusion or evaporation of the solvent (for details, see the Experimental Sectioni nt he Supporting Information). All complexes display ad istorted squarep lanar geometry,a s expectedf or diamagnetic platinum(II) metal centers. Compounds 1, 3,a nd 6 crystallize in the monoclinic P2 1 /c space group, whereas 2 crystallizes in the monoclinic P2 1 /n space group and 4 and 5 both crystallize in the triclinic space group P1 .
Almost perfectly coplanar aromatic p-systems for the donor and acceptorl igands are observed for 1 and 2,f eaturing two oxygen donor atoms (measured between the planes spanned by the two chelates). For compounds 3 and 4 featuring a NSO 2 Rd onor group, we observe ad eviation from coplanarity of up to 228 (designated by angle q), and for compound 5 featuring an oxygen and nitrogen donor,w eo bserve an intermediate deviation (see Figure 3). Ta ble 1s hows selected bond lengths and angles fort he discussed complexes 1-5,t he precursor,a nd the free pimp ligand. Ac lose inspection of the bond lengths shows that the aromaticity of the catecholato ligands is retained, with bond lengths around 1.40 .T he C1À O1 and C2ÀO2 bond lengths of around1 .35 and C1ÀN1 and C2ÀN2 of around 1.44 are normally observed for CÀO/CÀN single bonds, thus pointing to af ully reduced catecholate Q 2À form in all cases. [3] The bond lengths for complexes 3 and 4 with diamidosulfonyll igandsa lso show bond lengths that are in good agreement with the bond lengths of the free ligand. [23] An inspection of the bond lengths for the free pimp acceptor ligand [25] suggests an unreduced ligand because the N1ÀC7, C7ÀC8, and C8ÀN2 bond lengths do not drastically change (only around0 .03 )u pon coordination with the [PtCl 2 ]f ragment or subsequentc atecholate coordination.T his also applies to the N1ÀPt and the N2ÀPt bonds, which get slightly shorter when an OO-donor ligand is employed as compared to the NN-donors. The torsion angle f in the pimp ligand ranges from 708 to almost 1008.The observedvalues agree with previous literature reports for relatedd onor-acceptor systems [5,26] and also for related platinum chloride complexes. [9,10,27] The nitrogen donors of the diamidosulfonyl ligandsi nc omplexes 3 and 4 show ad ifferent coordination geometry.A trigonal planar coordination is expected for sp 2 -hybridized nitrogen atoms, which is characterizedb yasum of 3608 for all surrounding angles. The nitrogen atom in the trans-position to the pyridine shows at rigonal planar coordination with angles of 355.08 (   when compared to unsubstituted diamidobenzenes,b ecause similar deviations are observedf or platinum(II) and rhodium(III) complexes with symmetric co-ligands. [18] As teric effect can most likely be ruled out, given that the less bulky mesyl (methane sulfonyl;M s) group in 4 shows almost the same deviations as the tosyl groups in 3.
The "electrochemical" HOMO-LUMO gaps of the complexes correlate almost perfectly with the calculated HOMO-LUMO gaps (B3LYP/def2-TZVP), owing to the strong localization of the HOMO on the donor ligand and the LUMO on the acceptor ligand (see Figure 5). However,i ts hould be mentioned that there is as light offset in values if they are directly compared. The magnitude of the HOMO-LUMO gap is determined by the donor ligand, considering that the acceptor ligand is not changed.T he more strongly electron-withdrawing the sub-stituents of the donor ligand are, the lower is the HOMO energy and thus the biggert he HOMO-LUMO gap. With this information,aseries can be established on how electron-withdrawingt he ligandsa re, with the least electron-withdrawing (or strongestd onor) to the most electron-withdrawing( or weakest donor):

UV/Vis/NIR spectroelectrochemistry
To probe the interplay of the opticala nd electrochemical properties,U V/Vis/NIR spectroelectrochemistry using an optically transparent thin layer electrochemical (OTTLE) cell was employed. Because the CV measurements already showed that the higher oxidation states of the compounds likely show follow-upr eactions complicating the electrochemical response, we concentrated on the first oxidation and the first reduction. Early in the investigation we observed ad ecay in the absorption bands when measuring the spectrum in the OTTLE cell without applying ac urrent. We witnessed ar ather quick reaction/decomposition with electromagnetic radiation in the mid-UV range. Upon removing this part of the spectrum (200-300 nm) by using an appropriate filter,w ed id not observe any change of the spectrum during as imple absorption measurement in the OTTLEc ell. The quantitative UV/Vis/NIR measurementsi nas tandard cuvetted id not parallel these observations. If the mid-UV range is used during the spectroelectrochemicalm easurement, we observe only irreversible processes, indicating ad ecomposition for the simultaneousa pplication of mid-UV radiation and ac ertain redox potential. The UV/Vis/NIR spectra for complexes 1-5 are shown in Figure6.I ti se videntt hat variation of the donor ligand has two distinct effects on the absorption spectrum. First, there is the strong decrease in the extinction coefficient for the long wavelength bands, startingf rom 1 with as trong donorl igand to 2 with the electron-deficient tetrachlorocatecholato ligand. This decrease is matched by complexes 3 and 4 for which the donor functions change from "oxido" to amidosulfonyl. The "mixture" of both ligands (i.e.,the complex employing the amidosulfonylphenolatol igand 5)n eatly lies in the middle between the two extremes. Second, we see as hift to highere nergies (or shorter wavelengths)f or complexes with electronwithdrawing ligands. This nicely agrees with the intuitive assumption that we have to excite electrons from al igand with an energeticallyl ower-lying HOMO and hence requirem ore energy to do so. Again, complex 5 is in between the two ex-tremesof1 and 4.These broad and intense bands from roughly 540 to 700 nm can be ascribed to an LL'CT process, whichi s Scheme2.Redox processes for complex 1.  Figure 7s hows the results of UV/Vis/NIR spectroelectrochemistry for complex 1 upon oxidation andr eduction, and comparison of the spectra for 1, 1 À , 1 + ,a nd 1 2 + ;f or the different densities of the respective transitions, see pages S32-S37 (Supporting Information). Upon oxidation, the LL'CT band at 717 nm loses intensity and shifts to higher energies because less electron density is available on the donorl igand.S imultaneously we observe the increaseo faweak band at around 1000 nm, which can be attributed to the intra-ligand charge transfer (ILCT) process of the semiquinonato ligand.T he new bands in the visible region for 1 + can be attributed to metalto-ligandc harge transfers (MLCTs) and p-p*t ransitions from the mesityl group to the donor ligand. Further oxidation to 1 2 + resultsi nt he loss of the NIR band around1 000 nm, and the bands in the visible region shift slightly but maintain their shape. These bands of 1 2 + can be assigned to MLCT processes for the now fully oxidized donorligand.
The reduction takes place on the acceptor ligand:t he LL'CT band againl oses intensity,a nd ac ouple of new andr elatively sharp bands arise in the visible/near UV region of the spectrum. These can be assignedt oI LCT processes taking place on the pimp ligand. There are also minor d-orbital contributions. This is observed for all the complexes,s howing again that the reduction is taking place on the ligand,w hich is also reproduced by our TD-DFT calculations (see pages S32-S45, Supporting Information).
In contrast to the absorptions pectrao ft he native compounds, whichlook qualitatively similar,the spectra for the oxidized and reducedc omplexes differ substantially in some regards (see page S31, SupportingI nformation). If the spectra of all the singly oxidized compounds are compared, we observe rather intense NIR bands for complexes 3 + , 4 + ,a nd 5 + ,w hich all feature an itrogen donor on the semiquinonato ligand, whereas complexes 1 + and 2 + with oxygen-only donors show no absorption in this region ( Figure S78, Supporting Information).I nc ontrast, complexes 1 + and 2 + show strong absorption in the visible region of the spectrum.  The spectra for the reduced species 1 À , 2 À , 3 À , 4 À ,a nd 5 À look relatively similar( Figure S77, Supporting Information), which again indicatest hat the reductions teps are primarily pimp-centered. We observe ar elativelys harp double peak between 400 and 500 nm for all compounds, including the precursor 6.A ll compounds exhibit NIR bandst hat also look similar,e xcept for 1 À ,f or which the band is shifteds ignificantly to higher energy.F or 2 À ,t he band is extremely weak. This indicates av arying degree of influence of the donorl igand on the reduced acceptor ligand, with the influence for 1 À and 2 À (with the oxygen-bearing donor ligands) being the most pronounced.
As discussed above, in these complexes,t he LUMO is located on the pimp ligand and the HOMO is located on the donor ligand,a nd we have additionally confirmed that the low energy band can be assigned to aH OMO-LUMO transition. Thus, we can use UV/Vis/NIR spectroscopy to directly measure the HOMO-LUMO gap. The UV/Vis/NIR data can now be correlated analogously to the CV data with the DFT calculations (B3LYP/def2-TZVP). As shown in Figure 8, the maximumo ft he LL'CT correlates almost perfectly with the theoretical calculations andt he electrochemical data. These correlationsi nde-pendently validate the trend for the donor strength that we have already concluded from the CV measurements in order of decreasing donor strength:

Rearrangements in complexes 3and 4
Complex 3 showed ap articularb ehavior during the UV/Vis/NIR spectroelectrochemicali nvestigation. Upon oxidation of 3,a broad and very intenseb and at 950 nm emerged with a shoulder at around 750 nm (see Figure 9a). This process can be ascribed to MLCT and ILCT processes, whichh ave been observed for diiminiosemiquinonato ligands coordinated to a metal center. [28] Reversal of the scan direction resulted in loss of intensity of the newly emerged bands as expected. However,n ew bands at even lower energies around1 550nm emerged, which had not been observed at the start of the electrolysis (see Figure 9b). These intermediate bands vanish once the initial potential has been reached, and the complex is fully re-reduced (see Figure 9c). The starting spectrum is regainedw ith roughly8 0% of its initial intensity for the LL'CT transition (see Figure9d), indicating ac hemical reaction or substantial rearrangement of the investigatedc omplex. Ar earrangementi sa lready evidentf rom Figure 9a,i nw hich ab and is observed around 1300 nm, whichv anishes again in the course of the constant potential electrolysis (olive trace). Such time-and potential-dependent NIR absorptions are not observed for the oxidative spectroelectrochemistryo f1, 2,a nd 5.I ti sr easonable to assume that these low energy bands are caused by an ILCT [28] of the Q To l C À ligand and MLCT,w hich is also corroborated by TD-DFT calculations. These rearrangementsp ossibly involve ac hange in the position of the tosyl groupsa nd the coordinationg eometry of the nitrogen donor atoms( see discussiona bove). However,i ti sq uite unlikely that simple rotational movements exert such ah eavy influence on the absorption spectrum.T he two independenti sosbestic points at around 1200 and 530 nm also point to two different speciest hat are present during the measurement.A na lternative explanation would be the cleavage of one of the platinum-nitrogen bonds and coordination of the oxygen atoms of the sulfonyl groups. Such rearrangements have been proposed for rhodium complexes employing the same ligands by Perutz and co-workers. [20,21] Also, ap ossible dimerizationc annot be ruled out, as discussed recently by Chang and co-workers. [29] Upon oxidation of compound 3 to 3 + ,t he electron density on the already electron-poor bis(amidosulfonyl) ligand is further reduced, resulting in aw eaker platinum-nitrogen bond. This possibly resultsi nalinkagei somerization at the platinum center with one of the oxygen atoms of the sulfonyl group coordinating to it instead of the nitrogen atom (Scheme 3). A seven-memberedo rafour-membered chelate ring is possible with the oxygen atom coordinating in the trans-position to either the pyridine or imino donor function of the pimp ligand (Scheme 3). Given that the initial UV/Vis spectrumc annotb e fully recovered after re-reduction,i ti sr easonable to assume that the resulting compound formed after oxidationi sn ot very stable. Also, the DFT geometry optimizationf or 3 + starting from the XRD geometry of 3 did not indicate significant rearrangements. However,i ft he intermediates showni nS cheme 3 are used as starting points for the geometry optimization, the calculation quickly converges for all intermediates. The calculations have been performed exemplarily for complex 4 only. The relative energies of all investigated rearranged complexes predict 3 + with its originalf ive-membered chelate ring to be the most stable and the most unstablet etracyclic trans-pyri-dine 3 + to be energetically disfavored by around1 eV (Scheme 3). Bearing in mind that an energy of 1eVc orrespondst oaw avelength of around 1240 nm, the light source could easily trigger this isomerization. Given the fact that the compounds decompose if aw avelength lesst han 300 nm is used, this photoisomerization likely occurs.F or all five isomers, the opticals pectrah aveb een computed (see page S23, Supporting Information). Although somec hanges become appar-Scheme3.Possiblei somerization for compounds 3 + and 4 + (Mes = mesityl). Figure 9. UV/Vis/NIR spectroelectrochemistry for complex 3 in CH 2 Cl 2 /NBu 4 PF 6 measured with ag old working electrode. a) Oxidationof3to 3 + ,b )re-reduction of 3 + to rearranged3 *, c) further re-reductiona nd appearance of an ew band of 3 + to rearranged 3*, d) comparison of 3, 3 + ,3+ *and re-reduced3 . For adetailed discussion see text. Chem.E ur.J.2020Chem.E ur.J. , 26,1314Chem.E ur.J. -1327 www.chemeurj.org 2019 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim ent from the calculations, they are not sufficient to make ad efinitive assignment.
The structurally similar complex 4 also showeds ome similarities to 3 during oxidative spectroelectrochemistry. For the oxidation of 4 to 4 + ,w eo bserve two bands around 900 and 1500 nm of medium intensity.T he spectrum for 4 + (see Figure 10 a) looks very similar to the spectrum that is obtained during the re-reduction of 3 + to 3 (see Figure 9c)a nd can be interpreted analogously (see above). Upon further oxidation, from 4 + to 4 2 + ,t he long wavelength band at 1500 nm vanishes (see Figure 10 b), whereas the band at around 900 nm strongly increases, this time resembling the spectrum of 3 + (Figure 9a). Interestingly,t he starting spectrum is almostp erfectly recovered upon re-reduction;h owever,t he LL'CT band at around5 50 nm does not increases imultaneously with the decrease in the NIRb ands of 4 + at 900 and 1500 nm (see Figure 10 d).
Given the different stericd emands of the tosyl and mesyl substituents, it makes sense to assume that the rearrangementsf or complex 4 will be less sterically hindered and thus faster than the ones for 3.T he above results thus strongly point to the operation of redox-induced linkage isomerism in the sulfonamide-substituted ligands, which is linked with intriguingc hanges in the NIR region of their absorption spec- Figure 10. UV/Vis/NIR spectroelectrochemistryfor complex 4 in CH 2 Cl 2 /NBu 4 PF 6 measured with ag old electrode. a) Oxidationof4t o4+ ,b)oxidation of 4 + to 42 + ,c )re-reductionof4 2+ to rearranged 4 + *, d) furtherr e-reductiono f4+ to 4*, e) furtherr e-reductionand re-emergence of the LL'CT band, f) com-parisonofU V/Vis/NIR spectrao f4,4+ ,42+ ,re-reduceda nd rearranged 4 + *a nd re-reduceda nd rearranged 4*. For adetailed discussion see text. Chem. Eur.J.2020, 26,1314-1327 www.chemeurj.org 2019 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim trum. To the best of our knowledge,s uch electron-transferdriven linkage isomerism hasn ever been observed before for metal complexes of diamidobenzenes. Unfortunately,i tw as not possible to probe these rearrangementsb yo ther means (e.g.,i nfrared spectroscopy) on isolated complexes in the absence of light because the complexes were not stable towards chemicalo xidizing agents ([NO]BF 4 or AgPF 6 )u nder an inert atmosphere (see Figure S81, Supporting Information).
After these observations, the questiona rises as to why complex 5 does not display this typeoflinkage isomerism. At entative explanation mayb et he q angle between the p-systems of the donora nd acceptor ligands. This is slightly closer to 1808 for 5 (see Ta ble 1) and thus resultsi nab etter orbital interaction, whichl eads to as trongers tabilization of 5 + through back-bonding of the Pt II d-orbitals, in comparison to 3 + and 4 + .A dditionally,o ne can argue that the transition state (which likely involves at ilting of the donor ligand with respectt ot he acceptorl igand) is already preformed in complexes 3 and 4, and thus facilitates the rearrangement.

EPR spectroelectrochemistry
To gain further insights into the electronic structure of the redox intermediates, EPR spectroelectrochemistry was applied. Electrolysis inside the EPR cavity led to the observation of signals for the one-electron-reduced and one-electron-oxidized forms of all complexes 1-5.A ll spectra couldb es imulated (see pages S24 and S25, Supporting Information) and show ac oupling to 195 Pt (abundance of 33.3 %a nd nuclear spin I = 1 = 2 ) with no other resolved hyperfine splitting.
The nature of the donor atoms on the donor ligand has a certaini nfluence on the g-value of the radical cationic compounds, with the g-value being slightly higher for the phenylendiamide ligands than that of the catecholates or the amidophenolate ligands. The hyperfine coupling constants (hfcc) vary rather broadly from 2.73 mT for 1 to 11.90 mT for 3,i ndicating varying degrees of contribution from the platinum center (see page S23, Supporting Information).
The g-values of the radicala nionic formsa re almost the same with the only exception being complex 2,m ost likely owing to the tetrachloro substitution. The same applies for the hfcc, which is on average around 9mT, except for 2.T hese values further substantiate that the first reduction is based on the acceptor ligand.T he experimental observations are well-reproduced by as pin population analysis, which correctly predicts the spin to be localized mostly on the pimp acceptor ligand for the reduced species and mostly centered on the quinoid donor ligand for the oxidized species( see page S25, Supporting Information).

DFT calculations and electron flux
To furtherr ationalize the electronic structure of the complexes, DFT calculations on the B3LYP/def2-TZVP level were employed. The optimized structures (BP86/def2-TZVP)a re in good agreement with the crystallographic data. Figure 11 shows the molecular orbital energies from thesec alculations along with the HOMO-LUMO gaps of complexes 1-5,w hich we have already employed for various correlations( see above). The frontier orbitals for all of the complexes show ac ertain metal character but are neverthelessm ostly ligand-based and display mostly  -character.T he discussed deviationo ft he two nitrogen donor atoms for 3, 4,a nd 5 is also apparent in the orbitalp icture. TheHOMOs for 1 and 2 exhibit amirror plane (perpendicular to the p-system and dissecting the OCCO chelate), whereas we observed a" distorted symmetry" for 3, 4,and 5.
To characterizet he ligand-to-ligand charge transfer mechanism for complexes 1-5,i ti si nsightful to analyze the electronic flux densities for the optically active transitions from the ground state to the first absorption band. The electronic flux density yields space-resolved information about the flow of electrons during the excitation process. Because the first absorptionb and at the TD-DFT/B3-LYPl evel is dominated by a HOMO-LUMO transition, the electronic flux densities are calculated in the single-active electron approximation.A ll quantities were computed using ORBKIT [30] and depicted using ZIBAmira, [31] as shown in Figure 12.
As discussed above, the complexes all reveala ni ntense ligand-to-ligand charge transfer for the lowest energy transition, from the various ligand donors to the iminopyridine acceptor.O nt he acceptorl igand of all five complexes, the electronic flux exhibits al arge degree of delocalization, with a pincer-type electron flow incoming through the platinum center,o ver both nitrogen atoms and to the neighboring carbon atoms through the conjugated p-system. The delocalized flow is particularly laminarf or complexes 1, 2,a nd 5, which correlatesw ith al aminar,p incer-type electron flow on the donorl igand as well. The planarity between the donor and acceptorl igands leads to as imple x-shapedf low of the electrons from the coordination atomso ft he donor ligand over the platinum atom to the nitrogen atoms of the acceptor ligand.D espite their similarities, as tronger,m ore localized electron flow is observed on complex 2.T his is possibly due to the symmetry of the substituents on the phenylr ing of the donor.I nc ontrast, the electron flow pattern att he donor li-gands of complexes 3 and 4 is more intricatedue to their nonplanar structure.T his indicates that more electrons are available in the space between the donor ligand and the metal center,w hich in turns favors at hrough-space mechanism for the electron transfer.T he electron flow on the acceptor ligand is perturbed and reduced in intensity in both complexes but, surprisingly,t he spatial extent of the electron transfer remains similar.F or am ore quantitative measure, the charge transfer (CT) number can be computed as the product of the donor population and the acceptorpopulation, as in Equation (1): where P i are the respectiveM ulliken projectors on the donor and acceptor.D espite the marked differences in the electron flow mechanisms observeda bove,t he charget ransfer numbers were found to be similari nc omplexes 2-4,r anging from 0.757 for complex 2 to 0.784 for complex 3.O ft he three planars tructures, only complex 1 was found to have as lightly smallerC Tn umber (0.711). This correlates well with the smaller electron flow observed on the donorl igand. By looking at the three largest CT numbers (complexes [3][4][5], it appearst hat choosing al igand which increases through-space electron flow can increase the degree of charge transfer.T his can come at the expense of al ess laminar flow,a si nc omplexes 3-4,w hich we ratherattribute to as tructural effect.

Application in cross-dehydrogenative coupling reactions
Cross-dehydrogenative coupling (CDC) has gained considerable popularity among organic chemists as an atom-efficient methodf or CÀCb ond formation. [32] Inspired by previous works, whichu tilized platinum for this reaction, we wanted to test the synthesized complexes and investigate their photocatalytic potential. [33,34] They were employedi nt he cross-dehydrogenative coupling of N-phenyltetrahydroisoquinoline (ISQ) with acetone and nitromethane. Special focus was put on the role of the donor atoms, whichi sw hy complexes 1, 4,a nd 5 were used for ap reliminarys tudy with an [OO], [NN], and [ON] donorl igand, respectively.W eo ptimized the reaction conditions in terms of oxygen saturation and overall irradiationt ime (t,e xcitation at 360 nm). Pure oxygen was bubbled through the solutionf or 2minutes, 30 minutes,o rn ot at all (equal to atmosphericc onditions). Interestingly,w eo bserved ah igh yield for complex 1 after only 2minutes of bubblinga nd a long irradiation time of 15 hours in nitromethane, and ac onsiderably lower yield for the same conditions in acetone (see Ta ble 3). If atmosphericc onditions or shorter reaction times are used, the yield diminishes drastically, such that only traces of the product are isolated. Complex 4 also showed only traces under similarr eaction conditions. Complex 5 interestingly showedc onsiderably lower conversions;h owever, as tilla cceptable isolated yield of 40 %i nn itromethane and only 18 % in acetone wereobserved. As acontrol experiment,the precursor 7 wasa lso tested in the CDC, and it displayed considerable activity for the coupling of ISQ and acetone, whereas complex 1 was still superior for the coupling of ISQ with nitromethane. This may have to do with possible side reactions caused by the substantially labile chloride ligands. These results further highlight the high potentialo f the underdeveloped (pimp)PtX 2 systemf or photocatalysis and other applications. To further substantiatet he involvement of dioxygen as an oxidizinga gent, we irradiated as olution of complex 1 in DMF (bubbled for 30 minutes)f or 3minutes with an excess of a-phenyl-N-tert-butylnitrone (PBN). The EPR spectrum shows at riplet (Figure 13), indicating the generation of ar eactive oxygen species (ROS). Given that PBN is an onspecific ROS scavenger, complex 1 could serve as as ensitizer for either singlet oxygen or the superoxide radical, as depicted in Scheme4.T os elec-tively probe the involvement of singlet oxygen ( 1 O 2 )i nt he catalytic cycle, complex 1 was tested in the oxidationo f1 ,5-dihydroxynaphthalene to 5-hydroxy-1,4-naphthoquinone (or juglone). The observation of new bands in the visible region( see Figures S14a nd S15, Supporting Information) that correspond to juglonea re indicative of singlet oxygen being involved in the reaction. [35] Complex 1 will most likely have ar elativelyl ow excited-state oxidation potential, whichw ill not suffice to oxidize ISQ. [36] Thus, we repeated the catalysis at al ower excitation wavelength (around7 00 nm) and observed no conversion for the reaction of ISQ with nitromethane or acetone. This supports the fact that singlet oxygen is the active speciesg enerated with this sensitizer.H owever,b yu sing an irradiation wavelength of 360 nm, highere nergy excited states may be generated, which have as ufficiently high potential to oxidize ISQ. More detailed studies on thep hotophysicso ft he complexes will be necessary to answer thesequestions.
This leads us to the conclusion that as imilarc atalytic cycle as already described in various literature reports is operating here (see Scheme 4). [34,37] The platinum complex is photochemically exciteda nd quenched by the isoquinonline, resulting in the monoanionicp latinum complex and ac ationic tertiary amine. The molecular oxygen oxidizes the reduced platinum complex, recovering the sensitizer and redox-mediator and generating superoxide radical, which abstracts ah ydrogen atom from the oxidized isoquinoline. The hydroperoxyl radical servesa sabase for the nucleophile (acetone or nitromethane), generating am onoanionic nucleophile,w hich combines with the cationic isoquinoline, resulting in the product. Although we have only shown an electron transfer pathwayi nvolving 1 O 2 in Scheme 4, the operation of an alternative energy transfer pathway cannot be completely ruled out.
Further studies will be targeted towardsadeeper understandingo ft he mechanism and the fine-tuning of the catalyst.

Conclusion
We have presented as eries of new platinum(II) donor-acceptor systemsw ith the lesser used phenyliminomethylpyridine ligand and af ocus on the influence of the donor ligand. The title compounds were extensively characterizedb yc yclic voltammetry and UV/Vis/NIR-and EPR spectroelectrochemistry.A ll complexes displayed from two up to four redox events of varying reversibility.D ensityf unctional theory reproduced the ex-  Scheme4.Te ntative reaction mechanism for the CDC. Adapted from Chen, Fu, andcoworkers. [34] (SET = single electron transfer). perimental absorption spectra nicely,a nd the dynamic electron fluxes in such systems were investigated for the first time. UV/ Vis/NIR spectroelectrochemistry revealed interesting redoxdriven linkage isomerism during the oxidation of complexes 3 and 4;a no bservation that, to the best of our knowledge, has been made for the first time in metal complexes of phenylenediamines. The isomerism leads to intriguing changes in the NIR region of the spectrum of the isomers. The stabilization of these isomersr emains ac hallenge with potential applications in photocatalystso rt he development of new materials. The frontier orbitals are strongly localized on the respective ligand, which resulted in an excellent correlation of the "different" HOMO-LUMO gaps. The calculation of dynamic electron flux densities provided an insighti nto the electron dynamics of these systems for the first time and will help in the design of better photocatalysts and opticalm aterials. Lastly,w es howed that these systemsa re also interesting for catalysis. Complex 1 showedh igh yield in the cross-dehydrogenativec oupling of nucleophiles to N-phenyltetrahydroisoquinoline. Future investigationsw ill be dedicated to the exploitation of the different redox states of complexes 1-5,e xploitingt heir switching (redox-induced isomerism) and catalytic potentialt ot he full, and synthesizing the nickel and palladium analoguesf or their potentialuse in catalysis ands witching.