Tuning the H‐Atom Transfer Reactivity of Iron(IV)‐Oxo Complexes as Probed by Infrared Photodissociation Spectroscopy

Abstract Reactivities of non‐heme iron(IV)‐oxo complexes are mostly controlled by the ligands. Complexes with tetradentate ligands such as [(TPA)FeO]2+ (TPA=tris(2‐pyridylmethyl)amine) belong to the most reactive ones. Here, we show a fine‐tuning of the reactivity of [(TPA)FeO]2+ by an additional ligand X (X=CH3CN, CF3SO3 −, ArI, and ArIO; ArI=2‐(tBuSO2)C6H4I) attached in solution and reveal a thus far unknown role of the ArIO oxidant. The HAT reactivity of [(TPA)FeO(X)]+/2+ decreases in the order of X: ArIO > MeCN > ArI ≈ TfO−. Hence, ArIO is not just a mere oxidant of the iron(II) complex, but it can also increase the reactivity of the iron(IV)‐oxo complex as a labile ligand. The detected HAT reactivities of the [(TPA)FeO(X)]+/2+ complexes correlate with the Fe=O and FeO−H stretching vibrations of the reactants and the respective products as determined by infrared photodissociation spectroscopy. Hence, the most reactive [(TPA)FeO(ArIO)]2+ adduct in the series has the weakest Fe=O bond and forms the strongest FeO−H bond in the HAT reaction.


Introduction
Thed evelopment of non-heme iron(IV)-oxo complexes has been inspired by enzymatic oxidations. [1] Enzymes working with non-heme iron(IV)-oxo reactive forms can oxidize inert C À Hb onds in as elective fashion. [2] Thee fforts on synthetic non-heme iron complexes have focused on understanding the ligand effect on the properties and reactivity of the generated reactive iron(IV)-oxo species. [3] Ap rototypical complex in this field is one supported by the tetradentate tripodal ligand TPA( tris(2-pyridylmethyl)amine). [4] It has been shown that [(TPA)Fe II (TfO) 2 ]( TfO À = trifluoromethylsulfonate or triflate) can be oxidized by peracids or iodosobenzene (or its derivatives) in acetonitrile to form af airly reactive [(TPA)Fe IV O(MeCN)] 2+ complex (Scheme 1). [5] Subsequent modifications of the TPAl igand have led to even more reactive complexes. [6] Thec ommon feature of the [(L)Fe IV (O)] 2+ complexes (L = tetradentate tripodal ligand) is that asolvent molecule or another 2e À donor ligand is often attached as the sixth ligand (unless the supporting ligand is sterically hindered). [7] This labile ligand offers another opportunity to tune the reactivity of this type of iron(IV)-oxo complexes,a lthough this aspect has thus far not been investigated in that much detail. [7] In solution, we expect an equilibrium among complexes with different labile ligands.The observed bulk reactivity then reflects the average of the reactivities of these different complexes weighted by their abundance.H ere,w es how an approach that allows different forms of reactive complexes to be detected and their reactivity to be compared. Fort his study,w ef ocus on af amily of iron(IV)-oxo complexes [(TPA)Fe IV O(X)] 2+ (X = labile ligand, Scheme 1). In addition, this family of analogous complexes offer an ideal platform to evaluate how the properties of the Fe=Ob ond and the newly formed FeO À Hb ond affect the reactivities of these complexes in HATreactions.

Results and Discussion
Mass spectrometry has been an established method in the field for detection of reactive iron(IV)-oxo complexes. [8] Most often, researchers use the cryospray method at À40 8 8Ct o transfer the reactive complexes to the gas phase. [9] In this study,w ep repare the reactive complexes in af low reactor from as ilica capillary coupled to the electrospray ionization chamber,t hereby shortening the time for the oxidation reaction ( Figure 1a). Thew hole setup is cooled to À40 8 8C and the reaction time (t r )i nt he flow is typically about 10 seconds.  Figure 1b). [5] In addition, the low temperature and short reaction time allow the detection of [(TPA)Fe IV O(ArIO)] 2+ and [(TPA)Fe IV O(ArI)] 2+ ions,c orresponding to signals of iron(IV)-oxo complexes,r espectively bound to the oxidant itself as well as the oxidation byproduct ArI.

Hydrogen Atom Transfer (HAT)
Thereaction of the iron(IV)-oxo complexes that is of most interest is hydrogen atom transfer with hydrocarbons,w hich we have studied with 1-methylcyclohexene (che,F igure 1c). After adding the hydrocarbon to the vial C, we can monitor the HATr eaction proceeding in solution before ESI-MS detection. Thereaction conversion can be tuned by changes in the concentration of the hydrocarbon (vial C). Thep roducts are iron(III)-hydroxo species (Figure 1c)a nd iron(III) alkoxide complexes bearing the deprotonated hydroxylated product 3-methylcyclohex-2-en-1-ol (cho,s ee Figure S8 for fragmentation patterns). Note that the HATr eaction proceeds in solution in the flow reactor as evidenced by the conversion changes depending on the reaction time (i.e., the capillary length, see Figure S5).
In  10.4 kcal mol À1 ) % TfO À (DBDE = 12.6 kcal mol À1 ) ! ArIO (DBDE = 25.0 kcal mol À1 ). Forc omparison, we have studied the trend of the binding energies experimentally in the gas phase,w hich is consistent with these theoretical values ( Figure S9). Thee ntropic effects in solution disfavor the stability of [(TPA)Fe IV O(ArI)] 2+ leading to the predicted trend in the equilibrium reactions shown in Table 1 (Table 1a nd  Table S3). The[ (TPA)Fe IV O(X)] 2+ complexes with X = TfO À or ArI and their role for the reactivity in solution can be neglected.

UV/Vis Spectroscopy
We have tested the hypothesis that [(TPA)Fe IV O-(ArIO)] 2+ is preferentially formed in solution by UV/Vis spectroscopy.T he parent S = 1c omplex [(TPA)Fe IV O-(MeCN)] 2+ exhibits am aximum absorbance (l max )a t7 22 nm when generated by oxidation with peracetic acid. [4] The maximum slightly blue-shifts when the iron-oxo species is generated by oxidation with an excess of ArIO ( Figure 2). [5] Previous studies showed that l max is sensitive to the nature of the cis-ligand.
Mayer [17] and Shaik [18] have shown that the barrier for HATreactions can be estimated based on the bond dissociation energy of the newly formed O À Hb ond (BDE OH ). [17e] However,i ti sd ifficult to assess the BDEso fF eO À Hb onds experimentally,therefore the studies rely on theoretical data. To date,t he experimental BDE OH values of only two nonheme Fe III OH complexes have been obtained, but these values have been determined only by indirect means. [19] In principle,t he Fe = Oa nd FeOÀHb onds can be studied [a] The calculations were performed at the B3LYP [11] -D3 [12] /def2tzvp [13] level using SMD [14] to account for the solvation by acetonitrile.  directly by vibrational spectroscopy.Stretching frequencies of the bonds correlate with their strength but, so far, it was impossible to obtain vibrational information on the FeO À H bonds of the product complexes.I nt he following,w e demonstrate that infrared photodissociation spectroscopy can easily carry out direct experimental measurements of the bonds of interest and shed light on ac ritical question in this field. [20] Infrared Characteristics of [(TPA)Fe IV O(X)] +/2+ and [(TPA)Fe III (OH)(X)] +/2+ Helium-tagging infrared photodissociation (IRPD) spectroscopy allows the measurement of IR spectra of massselected ions. [21] TheF e = Os tretching frequencies of all four detected [(TPA)Fe IV O(X)] +/2+ complexes can readily be assigned based on their 16 O/ 18 Os hifts.W eo bserve that the Fe=Ostretching frequencies of these four [(TPA)Fe IV O(X)] +/ 2+ complexes decrease in the following order:8 49 cm À1 for X = TfO À ,843 cm À1 for X = ArI, 839 cm À1 for X = ACN, and 834 cm À1 for X = ArIO, (Figure 3, Table 2). For[(TPA)Fe IV O-(ArIO)] 2+ ,t here is also af eature at about 640 cm À1 that is sensitive to 18  Them easured stretching frequencies of the reacting and the newly formed bonds predict that the HATreactivities of [(TPA)Fe IV O(X)] +/2+ should follow the order of X: ArIO > MeCN > ArI % TfO À .T his trend is the same as the trend observed experimentally.I ts hows that it is valid to approximately evaluate the reactivities of iron(IV) oxo complexes based on the properties of the Fe IV =Oa nd the Fe III OÀH bonds.
However,t he trends of the Fe=Oa nd FeO-H stretching frequencies are not reproduced by the DFT calculations (Tables 2, S1 and S2, Figure S12   [a] The error refers to the accuracy of the measurement and evaluation of the data. The relative error is much smaller.
[c] The B3LYP-D3/def2tzvp frequenciesscaled by 0.96. The unscaled frequenciesare in the brackets. The bands with asterisks * represent OÀHbands hydrogen bonded to the respective ligands X (see Figure S11).
These contradicting trends point to the limits of DFT in describing this type of complexes.T here is an ample of examples that DFT describes correctly reactivity and properties of various hypervalent metal complexes.However,there might be al imitation when going towards description of subtle differences such as the effect of labile cis-ligands as described here.H ence,w eb elieve that our results show the need for experimental data that describe properties and reactivity trends of these complexes to benchmark the theoretical methods.

Conclusion
We have investigated the formation and hydrogen-atom transfer (HAT) reactivity of [(TPA)Fe IV (O)(X)] +/2+ (X = MeCN,T fO À ,A rI, ArIO) complexes in af low setup with electrospray ionization mass spectrometry (ESI-MS) detection. We show that we can investigate solution reaction kinetics and employ sensitive ESI-MS detection at the same time.T his approach has allowed us to show that the HAT reactivity of [(TPA)Fe IV (O)(X)] +/2+ with 1-methylcyclohexene decreases in the order of X: ArIO > MeCN > ArI % TfO À .The theory predicts that the binding energy of ArIO to the iron core is about 16 kcal mol À1 larger than that of MeCN and that the formation of [(TPA)Fe IV (O)(ArIO)] 2+ in solution should be largely preferred if an excess of ArIO is used in the experiment. Theimportance of [(TPA)Fe IV (O)(ArIO)] 2+ , as pecies which to date has not been considered to be aparticipant in this chemistry,isfurther supported by UV/Vis spectroscopy.
Furthermore,wehave evaluated the effect of the ligand X on the Fe IV =Oa nd the Fe III OÀHb onds in the [(TPA)Fe IV -(O)(X)] +/2+ and [(TPA)Fe III (OH)(X)] +/2+ complexes,r espectively.T he Fe = Os tretching frequency increases in the order of X: ArIO < MeCN < ArI < TfO À ,while that of the FeO-H unit decreases in the same order.Hence aweaker Fe=Obond gives rise to astronger FeOÀHbond, which in turn facilitates the HATr eaction. This trend directly correlates with the measured HATa ctivities of these complexes and demonstrates in particular the role of the O-H formation in serving as adriving force for HATreactions,aspostulated by Shaik [18] and Mayer. [17]