Determination of the Lewis acidity of amide–AlCl3 based ionic liquid analogues by combined in situ IR titration and NMR methods

A combinatorial method to determine both acidic strength and acidic amount of each Lewis acid site in amide–AlCl3 based ionic liquid (IL) analogues was developed by the combination of in situ IR titration and NMR analysis. 31P NMR was used to distinguish effectively the acidic strength of each Lewis acid site in the amide–AlCl3 based IL analogues. Nitrobenzene was used as a molecular probe to measure the total Lewis acidic amount of the amide–AlCl3 based IL analogues by in situ IR titration. The acidic amount of each Lewis acid site in the amide–AlCl3 based IL analogues was calculated with the assistance of 27Al NMR analysis.

Acidic amide-AlCl 3 based ionic liquid (IL) analogues have attracted signicant attention as good alternatives to traditional imidazolium and pyridinium based halometallate ILs due to their broad acidity-adjusting range, high catalytic activity, low toxicity and cost, and easy preparation. 1-3 Amide-AlCl 3 based IL analogues exhibit a mixture of neutral molecular Al species, and cationic and anionic Al species in equilibrium, which contribute to the incomplete asymmetric splitting of Al 2 Cl 6 under the induction of amide. 4,5 Therefore, multiple Lewis acidic species with catalytic activity exist in these ILs analogues. 6 The Lewis acidity of the amide-AlCl 3 based IL analogues, including acidic strength and amount, is correlated with their catalytic activity and selectivity. [7][8][9] Hence, it is necessary to establish a suitable method to determine the acidic strength and amount of each Lewis acid in these IL analogues, which can guide corresponding acid-catalyzed reactions.
For traditional ILs, the spectral measurement methods of the acidity are mainly UV-vis, NMR and IR spectroscopies. The UVvis spectroscopy method determines semi-quantitatively the acidic strength of total Brønsted acid in ILs according to the Hammett function, 10-12 but it could not be applied in the analysis of Lewis acid in ILs, such as [Al 2 Cl 7 ] À in chloroaluminate ILs. The Lewis acidic strength can be quantied by the Gutmann acceptor number, which is directly proportional to the 31 P NMR chemical shi of triethylphosphineoxide (TEPO) dissolved in ILs. 13,14 The 31 P NMR method can distinguish effectively the acidic strength of each Lewis acid in ILs with multiple Lewis acids, but it could not measure the acidic amount of each Lewis acid. [15][16][17] The traditional KBr tabletting IR uses nitrogencontaining compounds as molecular probes, such as pyridine and ethanenitrile. The change in the IR frequencies of the molecular probes is correlated to the acidic strength of the acid species in ILs. The tabletting IR method can distinguish evidently the Brønsted and Lewis acid according to the wavenumber of the characteristic peaks. 18,19 For example, two peak at 1450 cm À1 and 1540 cm À1 were the indication of pyridine coordinated to Lewis and Brønsted acid, respectively. 20 But this method neither distinguishes easily the acidic strength of each Lewis acid in ILs with multiple Lewis acids because of the overlap of characteristic peaks, nor can it measure the acidic amount of each Lewis acid. In addition, infrared studies of ammonia adsorption and microcalorimetry were also used by Dupont Company to investigate the acidity of zeolite. 21 In this communication, we rst establish a combinatorial method to determine the acidity of amide-AlCl 3 based IL analogues with multiple Lewis acids by combining in situ IR titration with NMR analysis. This method not only distinguished effectively the acidic strength of each Lewis acid in amide-AlCl 3 based IL analogues, but also measured the acidic amount of each Lewis acid.
Firstly, 31 P NMR was used to identify the acidic strength of each Lewis acid in amide-AlCl 3 based IL analogues, as shown in Fig. 1. A single peak at 83.48 ppm was observed in the 31 P NMR spectra of molecular probe (TEPO) dissolved in neat Et 3 NHCl-AlCl 3 IL (molar ratio of Et 3 NHCl to AlCl 3 was 0.65), which was assigned to the coordination of TEPO to Lewis acid. This result indicated that neat Et 3 NHCl-AlCl 3 IL only contained single Lewis acid, namely [Al 2 Cl 7 ] À . However, two peaks at 83.48 and 84.92 ppm were observed in neat NMA-AlCl 3 IL analogue (molar ratio of N-methylacetamide to AlCl 3 was 0.65, marked as 0.65NMA-1.0AlCl 3 ) with the addition of TEPO. This phenomenon indicated that another Lewis acid in addition to [Al 2 Cl 7 ] À existed in 0.65NMA-1.0AlCl 3 . The peak at 84.92 ppm was assigned to the cationic Al species because the molecule Al species was neutral. 5 Meanwhile, the acidic strength of cationic Al species located in low eld was stronger than that of [Al 2 Cl 7 ] À .
Subsequently, in situ IR titration method was used to measure the acidic mount of two Lewis acids in neat 0.65NMA-1.0AlCl 3 . The principle of this method is based on the online monitoring of the variation in the characteristic peaks formed by the coordination of indicator (nitrobenzene) with 0.65NMA-1.0AlCl 3 . 22,23 A quantitative measurement of the acidic amount of 0.65NMA-1.0AlCl 3 was made based on the typical procedure. 0.65NMA-1.0AlCl 3 (10 g) was placed into a 25 mL two-necked ask equipped with a stirrer. The silicon probe of the in situ IR apparatus was inserted into the 0.65NMA-1.0AlCl 3 , and then the data on the IR spectra were collected. Next, nitrobenzene (0.25 g) was added dropwise to the ask and IR spectra were collected continuously until the absorbance of the characteristic peaks remained constant, meanwhile, the peaks of nitrobenzene itself were observed. The aforementioned steps were repeated until the absorbance of the characteristic peaks did not change with the addition of nitrobenzene. This point was marked as the terminal point of titration, the total mass of nitrobenzene added into 0.65NMA-1.0AlCl 3 was collected.
As a premise of the in situ IR titration method, the characteristic peak formed by the coordination of nitrobenzene with 0.65NMA-1.0AlCl 3 and the peak of nitrobenzene itself needed to be marked. Fig. 2 shows the IR spectra of neat nitrobenzene, neat 0.65NMA-1.0AlCl 3 , and the mixture of 0.65NMA-1.0AlCl 3 with nitrobenzene. Two peaks at 1520 and 1346 cm À1 were observed in neat nitrobenzene, which were assigned to the y as (O-N-O) and y s (O-N-O) stretching vibration of -NO 2 group, respectively. 24,25 A new peak at 1260 cm À1 was observed in the mixture of 0.65NMA-1.0AlCl 3 with nitrobenzene, which should be assigned to the coordination of nitrobenzene with Lewis acids. Meanwhile, the y as (O-N-O) stretching vibration at 1520 cm À1 shied to higher wavenumber 1537 cm À1 . The y s (O-N-O) stretching vibration at 1346 cm À1 appeared only in the case that excess nitrobenzene were added into 0.65NMA-1.0AlCl 3 . Therefore, the peaks at 1260 and 1346 cm À1 were chosen as the characteristic peaks to observe in the following in situ IR titration method.    This journal is © The Royal Society of Chemistry 2018 Fig. 3 shows the variation of the characteristic peaks at 1260 cm À1 and 1346 cm À1 from the coordination of nitrobenzene with 0.65NMA-1.0AlCl 3 and y s (O-N-O) stretching vibration of nitrobenzene, respectively. A surface plot was generated during the continuous addition of nitrobenzene into 0.65NMA-1.0AlCl 3 (Fig. 4). The absorbance of the peak at 1260 cm À1 increased with the increasing addition of nitrobenzene, while the absorbance of the peak at 1346 cm À1 remained almost constant before the terminal point, which attributed that the Lewis acidic amount of 0.65NMA-1.0AlCl 3 was continuously consumed by nitrobenzene. When the Lewis acidic amount of 0.65NMA-1.0AlCl 3 was used up, the absorbance of the peak at 1346 cm À1 had a signicantly increase with the addition of nitrobenzene. The total mass of nitrobenzene from start to terminal point was recorded, and "the molar consumption of nitrobenzene per 1000 g IL analogue" was dened as "activity index" to evaluate the acidic amount of 0.65NMA-1.0AlCl 3 . 26,27 The Lewis acidic amount of several amide-AlCl 3 based IL analogues with different amide structures and amide/AlCl 3 molar ratios were measured by in situ IR titration method, as shown in Fig. 5. The amide structure affected the Lewis acidic amount of amide-AlCl 3 based IL analogues, 28 for example, the Lewis acidic amount of amide-AlCl 3 based IL analogues (NMA-AlCl 3 and DMA-AlCl 3 ) with bidentate coordination was higher than that of amide-AlCl 3 based IL analogues (AA-AlCl 3 and Ur-AlCl 3 ) with monodentate coordination under the same amide/ AlCl 3 molar ratio. 29 This phenomenon was attributed to the fact that the bidentate coordination was more favorable to the asymmetric splitting of AlCl 3 than the monodentate coordination with the same amide/AlCl 3 molar ratio, resulting in the more active Lewis species (anionic Al species and cationic Al species). On the other hand, the amide/AlCl 3 molar ratio also affected the Lewis acidic amount of amide-AlCl 3 based IL analogues. The Lewis acidic amount of amide-AlCl 3 based IL analogues increased with the decreasing amide/AlCl 3 molar ratio. For amide-AlCl 3 based IL analogues, the balance between neutral molecular Al species and ionic Al species was readily broken with the change of amide/AlCl 3 molar ratio. The asymmetric splitting degree of Al 2 Cl 6 increased and the molecular species transformed into ionic species as the amide/AlCl 3 molar ratio decreased, so the Lewis acidic amount of amide-AlCl 3 based IL analogue also increased.
The total Lewis acidic amount of amide-AlCl 3 based IL analogues could be measured by in situ IR titration method, but the acidic amount of anionic Al species and cationic Al species needed to be further determined. 27 Al NMR is a good tool to distinguish these Al species, and the peaks at 102.75, 89.30 and 77.05 ppm should be assigned to the anionic Al species ([Al 2 Cl 7 ] À and [AlCl 4 ] À ), molecular Al species [AlCl 3 L n ], and cationic Al species [AlCl 2 L n ] + , respectively. 29 The integral area ratio of anionic Al species ([Al 2 Cl 7 ] À , [AlCl 4 ] À ) to cationic Al species ([AlCl 2 L n ] + ) was obtained by the normalization method of the peak areas, as shown in Fig. 6. The integral area

Conclusions
An efficient method to determine the acidic strength and acidic amount of each Lewis acid in amide-AlCl 3 based IL analogue was proposed in this study. The 31 P NMR using triethylphosphineoxide as a molecular probe showed that two active Lewis acids ([AlCl 2 L n ] + , [Al 2 Cl 7 ] À ) existed in amide-AlCl 3 based IL analogues, and the acidic strength of [AlCl 2 L n ] + was stronger than that of [Al 2 Cl 7 ] À . The principle of the in situ IR titration method was described in detail, and the total Lewis acidic amount of these IL analogues was measured with nitrobenzene as indicator. The results indicated that the total Lewis acidic amount of amide-AlCl 3 based IL analogues was related with both the amide structure and the amide/AlCl 3 molar ratio. The amide-AlCl 3 based IL analogues with bidentate coordination structure and low amide/AlCl 3 molar ratio had a high Lewis acidic amount. The Lewis acidic amount of each Lewis acid was calculated further by combining 27 Al NMR analysis with in situ IR titration.

Experiments
The general route for the synthesis of 0.65amide-1.0AlCl 3 IL analogue was as follows: anhydrous AlCl 3 (0.2 mol) was placed in 250 mL two-necked ask; then, amide (0.13 mol; N-methylacetamide, NMA; N,N-dimethylacetamide, DMA; acetamide, AA; urea, Ur) was added slowly while stirring for 30 min. The mixture was then heated to 80 C and maintained at that temperature until all solids "dissolved" (approximately 4 h). 30 IR spectra over the 4000 cm À1 to 650 cm À1 frequency range were obtained at room temperature and at 8 cm À1 resolution using an in situ IR spectrometer (Mettler-Toledo) equipped with an attenuated total reectance based silicon probe and a liquid nitrogen-cooled mercury-cadmium-tellurium (MCT) detector. IL analogue (10 g) was placed into a 25 mL two-necked ask equipped with a stirring bar at room temperature. The silicon probe was then inserted into the IL analogue, aer which date on the IR spectra were collected. Next, the indicator (0.05 g) was added dropwise to the ask and IR spectra were collected continuously until the characteristic peaks remained constant. The aforementioned steps were repeated. During the measurement, the optical path of the spectrometer was continuously purged with dry N 2 at a ow rate of 2 mL min À1 to eliminate moisture and CO 2 . 27 Al and 31 P NMR spectra were obtained using a Bruker Avance spectrometer.
The samples were placed into a 10 mm standard tube by inserting a well-centered capillary. Thereaer, the NMR tube was capped and sealed with paralm. The aqueous solutions of Al(NO 3 ) 3 (1.0 mol L À1 ) and H 3 PO 5 (85 wt%) in the capillary was used as an external reference for the 27 Al NMR and 31 P NMR chemical shi, respectively. Peak intensities and areas were carefully measured using the Brucker-NMR soware package.

Conflicts of interest
There are no conicts to declare.