Probing Homogeneous Catalysts and Precatalysts in Solution by Exchange-Mediated Overhauser Dynamic Nuclear Polarization NMR

Triphenylphosphine (PPh3) is a ubiquitous ligand in organometallic chemistry that has been shown to give enhanced 31P NMR signals at high magnetic field via a scalar-dominated Overhauser effect dynamic nuclear polarization (OE DNP). However, PPh3 can only be polarized via DNP in the free form, while the coordinated form is DNP-inactive. Here, we demonstrate the possibility of enhancing the 31P NMR signals of coordinated PPh3 in metal complexes in solution at room temperature by combining Overhauser effect DNP and chemical exchange between the free and coordinated PPh3 forms. With this method, we successfully obtain 31P DNP enhancements of up to 2 orders of magnitude for the PPh3 ligands in Rh(I), Ru(II), Pd(II), and Pt(II) complexes, and we show that the DNP enhancements can be used to determine the activation energy of the ligand exchange reaction.


S1 Temperature control
Under microwave heating and cooling gas, the temperature read out by the sensor in the sample stator is not close to the real sample temperature with DNP.Therefore, we use the chemical shift of free PPh3 as an internal thermometer since it is temperature-dependent.We first measured a sample containing 10 mM BDPA and 40 mM PPh3 in benzene-d6 on a liquidstate NMR machine (9.4 T) and found the chemical shift change (relative to 298 K, labelled as Δ) almost linearly with the sample temperature as shown in figure S1 and Table S1.We also tested this dependence using 100 mM PPh3, 10 mM BDPA and with [Rh(PPh3)3Cl] and found almost the same relation as listed in Table S1.Therefore, we established a linear relation below to calculate the sample temperature by the relative chemical shift to the experiment at 298 K.

S2 DNP OFF experiments
To rule out the temperature effect which is always coupled with the microwave, we applied a series of microwave experiments but the magnetic field is adjusted so that the electronic spin of BDPA is not on resonance and the DNP effect is therefore quenched.As shown in Figure S2, by comparing these experiments with their corresponding DNP on experiments, it is clear that the coordinated PPh3 peaks were enhanced because of DNP.Since only the magnetic field was changed by sweeping, we calibrated the free PPh3 peaks to the same chemical shift as the DNP in the experiment.

S3 DNP of PPh3 without metal complexes
The DNP experiment of free PPh3 without any metal complexes has been intensively studied before.Here we perform this experiment to get a reference sample measured with the same setup for comparison.The resulting DNP enhancement and measured T1 are plotted in Figure 2 and 5 of the main text with the related conditions and results in Table S2.
Table S2 For [Rh(PPh3)3Cl], we also performed the experiment without microwave irradiation at different temperatures and the resulting spectra are shown in Figure S3.It is clear that as the temperature increases, the T1 of free PPh3 decreases due to the exchange.In addition, the signal intensity was reduced when acquired with a fixed recycle delay of 1s.
Figure S3.The kinetics in the system with metal complexes could be considered as two parts as shown in Scheme S2.The left part is the DNP process, where the free PPh3 binds with the BDPA radical in the solution to form a transient complex and the PPh3 is hyperpolarized during this process.
The right part represents the ligand exchange process, where the hyperpolarized free PPh3 exchanges with a coordinated PPh3.The DNP process is complex and the BDPA-PPh3 complex is not observable in the NMR spectrum.To simplify the model, we assume the ligand exchange is relatively slow compared to the elementary DNP process.With this assumption, we can treat the DNP separately by assuming that the magnetization of the free PPh3 is building up to the DNP enhancement with the intrinsic T1 (T1F for the free PPh3).Meanwhile we ignore the weak dipolar-dominated negative DNP effect on the coordinated PPh3, and then the coordinated PPh3 should tend to the thermal polarization at the rate 1/ T1C.
Now, we can consider the exchange reaction as presented by Scheme S3.Here, we assume that the exchange reaction happens instantaneously and the polarization states are preserved after the exchange.Thus, for n mol/L of exchange reaction, it will cause  2 3 of magnetization transfer, which is negative for free PPh3 and positive for coordinated PPh3.Therefore, there is no net magnetization change in the system due to the exchange reaction at equilibrium.We know that the overall exchange rate r is defined as  = 12 13 .In the exchange reaction in Scheme S3, no chemical change is involved and therefore gives an equilibrium constant of 1, and the forward and back transformation rate equals to each other, r+=r-=r.Combining this exchange effect with the formula above, we have the magnetization change with exchange rate: As the concentration of the two species are constant in our experiments, we can use the polarization level to replace the magnetization in the equations above by  =  and d =  .Then divide the equations by corresponding concentration, and we get The polarization level is proportional to the NMR signal integrals and the scaling factor is the for a given experimental series.Since the enhancement is measured using a microwave off experiment under quantitative conditions, we can assume the polarization at microwave off is the thermal polarization level, and the thermal level is the same for both species.
Finally, we can use the enhancement to replace the polarization level by dividing the equations by Pthermal, and we have: Because we apply saturation pulses before the recycle delay, we have the initial conditions that  !,7 =  ',7 = 0.The term  #$% presents the enhancement that the DNP process can achieve in our conditions, since we assume the exchange process does not interfere the DNP process, we use the DNP enhancement measured under similar conditions but without the metal complexes.
For [Rh(PPh3)3Cl], we have ~2 mM complexes that have three coordinated PPh3 molecules in two different environments with the T1 of 0.8 s and 0.6 s at room temperature as shown in S8.
In principle, we need to treat these two different coordinated PPh3 separately and need to consider the exchange between these two sites.To simplify, we here treat these two sites as one and the  ' = 0.006 .The T1 is in principle temperature dependent, but above room temperature it is relatively stable, so we use  &! = 10  and  &' = 0.8  . #$% is also temperature dependent, but stabilizes at about 130 above room temperature when there is no metal complex, so we use  #$% =130 here.[Rh(PPh3)3Cl] because the T1C here is not significantly shorter than the free PPh3.Consequently, as compare to Figure S9, the coordinated PPh3 also reaches higher DNP enhancement at the same exchange rate, which is also consistent with our observations.

Determination of activation energy
As shown in section S9, the enhancement of the coordinated PPh3 in a certain regime strongly depends on the exchange rate, which allows us to extract the exchange rate from the DNP experiments.However, the absolute enhancement is related to the ideal DNP enhancement  #$% , which fluctuates between different experimental sessions due to the hardware performance.The ratio of enhancements, : " : ! is less sensitive to  #$% especially when the exchange rate is high.Here, we use  for the data in Table S5, and the results are shown in Table S8.(The data measured at 271.1 K and 330.3K are not used because their enhancement ratio is too close to 0 and 1 where the ratio is not sensitive to the ligand exchange rate.)We then plot the calculated ratio as a function of exchange rate r using the measured parameters from Figure S10, according to the model of section S9, but with  #$% = 90 and 130 according to the data in Table S8.As Figure S12A shows,  #$% = 90 and 130 give almost identical curves, with only a small difference observed when the enhancement ratio is close to 0. Therefore, we used the curve with  #$% = 90 to extract the exchange rates as shown in Figure S12A, and the resulting values are shown in Table S8.Finally, as we assume the reaction rate can be written according to the formula below, the concentrations are constant in this case and can be treated together with the pre-exponential factor.S8), the best fit line is shown in red.)) 1/T (K -1 ) Fitting ln(r)=-14110 1/T + 41 (A) (B) Figure S1.A series of 31 P NMR spectra of a benzene-d6 solution contains 10 mM BDPA and 40 mM PPh3 measured at different sample temperatures on a 9.4 T liquid NMR spectrometer.

Figure S2 .
Figure S2. 31P NMR spectra of (A) [Rh(PPh3)3Cl], (B) [Ru(PPh3)3Cl2], (C) [Pd(PPh3)2Cl2] and (D) [Pt(PPh3)2Cl2] in benzene-d6 with the addition of excess PPh3 (100 mM), 10 mM BDPA obtained at 9.4 T with continuous-wave microwave irradiation at (red) and out of the DNP field (blue).The spinning sidebands are labelled with asterisks.For (A), (B) and (D), spectra in each group were obtained with the same number of scans.While in (C), the DNP off spectrum (blue) is obtained by twice of number of scans of the DNP on spectrum (red), the spectrum intensity is normalized by the number of scans.

FigureFigure S8 .
Figure S7. 31P NMR spectra of [Pt(PPh3)2Cl2] in benzene-d6 with the addition of excess PPh3 Scheme S3.The ligand exchange and the magnetization change after n (mol/L) of

Figure S9 .
Figure S9.DNP enhancement a function of polarization time calculated for different enhancement as a function of polarization time calculated for different exchange rates r(mol L -1 s -1 ), parameters are used according to the experiment with [Pd(PPh3)2Cl2] shown in Figure 4.  ' = 0.006  ,  != 0.1  ,  &! = 10  ,  &' = 8  and  #$% = 130.To demonstrate the behaviour of [Pd(PPh3)2Cl2] shown in Figure 4, we calculated the enhancements using  &' = 8  and found the results shown in Figure S10.As discussed in the main text, the enhancement of the free PPh3 does not decrease as dramatically as for activation energy of the ligand exchange reaction in [Pd(PPh3)2Cl2] (chosen because of the high signal-to-noise ratio we obtained with this sample and since there is only one environment of coordinated PPh3).
Figure S12B.The fitted slope is 1.4×10 4 (± 5×10 3 ) which corresponds to an activation energy Ea of 28 (± 10) kcal/mol.(If we use only the two points in the linear regime of FigureS12Athat potentially have less error on the exchange rates, then the activation energy obtained is 18 kcal/mol.)We note that these activation energies are similar to reported activation energies of the PPh3 ligand dissociation in palladium or other transition metal complexes (ranging from 15 to 40 kcal/mol)2-5 .This could suggest that the ligand exchange process is dominated by the dissociation of the PPh3 ligand.

Figure
Figure S12.(A) Calculated ratio of enhancements : " : ! at 50 s as a function of exchange rate r

Table S3 .
Summary of the conditions and results of DNP experiment of free PPh3 with 31P NMR spectra of [Rh(PPh3)3Cl] mixed with excess PPh3 (100 mM). 10 mM BDPA in benzene-d6, obtained at 9.4 T with (red and green) and without (blue) continuouswave microwave. 31 spectra of [Ru(PPh3)3Cl2] mixed with excess PPh3 (100 mM). 10 mM BDPA in benzene-d6, obtained at 9.4 T with (red) and without (blue) continuous-wave microwave.TableS4.Summary of the conditions and results of DNP experiment of free PPh3 with [Ru(PPh3)3Cl2] with microwave irradiations measured at different temperatures (spectra in Figure3).

Table S5 .
Summary of the conditions and results of DNP experiment of free PPh3 with [Pd(PPh3)2Cl2] with microwave irradiations measured at different temperatures (spectra in Figure4).

Table S6 .
Summary of the conditions and results of DNP experiment of free PPh3 with The phosphorus atoms with chlorine or another phosphorus at the opposite site are labelled in blue or red, respectively.

Table S8 .
The parameters used to extract the activation energy for the ligand exchange of ": !