Rapid Quantification of the Activating Effects of Hydrogen Bonding Catalysts with a Colorimetric Sensor

Abstract

A sensor to quickly and simply assess the relative reactivity of different hydrogen bonding catalysts has been developed. Specifically, blue shifts seen upon treatment of hydrogen bonding catalysts with the colorimetric compound, 7-methyl-2-phenylimidazo[1,2-a]pyrazin-3(7H)-one, correlate well to the K_eq of binding to the sensor. The blue shifts also show a high degree of correlation with relative rates in Diels-Alder reactions of methyl vinyl ketone and cyclopentadiene employing the hydrogen bonding catalysts. The relevance of the sensor blue shifts to the LUMO lowering abilities of the hydrogen bonding catalysts is discussed.

Electrophile activation by small-molecule hydrogen bond donors has emerged as an important paradigm for enantioselective catalysis.ⁱ Nonetheless, a thorough understanding of the principles and features that govern the reactivity and selectivity of these catalysts remains incomplete. A number of physical organic measurements have provided scales that can be used to estimate the reactivity such as the pK_a table,ⁱⁱ the nucleophilicity and electrophilicity parameters,ⁱⁱⁱ the Irving – Williams order^iv,v etc but no scales have been made for all categories of H-bonding catalysts. Contributing to this problem is the large range of hydrogen bond strengths, from 0.2–40 kcal/mol.^vi While the strength of a hydrogen bonding interaction can be inferred from ΔpK_a,^vii,viii such a measurement gives an incomplete account with respect to catalysis since a water molecule poorly mimics a substrate. As a result, secondary interactions, such as sterics, dual hydrogen bonding,^ix and hydrogen bonding directionality, between a hydrogen bond donor and an electrophilic substrate are not fully incorporated. Here, we present a simple spectroscopic measurement using a colorimetric sensor to determine the effectiveness of hydrogen bonding catalysts in electrophilic activation of a monodentate substrate. The measurement is effective for a range of catalysts encompassing a pK_a window of approximately 7–20.

We assessed a number of methods to judge the ability of different hydrogen bond donors to activate a carbonyl (LUMO lowering), but found that methods effective for strong Lewis acids, such as changes in IR or NMR signals, provided insufficient signal or were technically challenging. In search of a simple, easily applied measurement, we elected to use a colorimetric sensor molecule. 7-Methyl-2-phenylimidazo[1,2-a]pyrazin-3(7H)-one (1), which gives good correlations between λ_max shifts and the Fukuzumi parameters for a small number of Lewis acids^x,xi,xii was discovered to give a readily discernable signal upon coordination (eq 1) with a range of hydrogen bond donors (Chart 1). Figure 1 illustrates the simplicity of the method with changes in color, which are readily visible to the naked eye, upon saturation with different hydrogen bonding catalysts.

Chart 1.

Hydrogen bonding catalysts.

Figure 1.

Change in color upon addition of hydrogen bonding catalysts (see Chart 1) to the pyrazinone sensor 1 in dichloromethane.

(1)

Figure 2 further illustrates the blue shift in the λ_max of the sensor when combined with increasing amounts of a hydrogen bonding catalyst, in this case N,N′-di(3,5-bis(trifluoromethyl)phenyl)thiourea. With this data, K_eq values (Table 1) for the sensor – hydrogen bond donor association ^xiii could be readily obtained from the corresponding titration curves as illustrated for N,N′-di(3,5-bis(trifluoromethyl)-phenyl)thiourea.^xiv The inverse of the λ_max shift obtained upon saturation with 2–6 showed a strong correlation with the K_eq value (Figure 3) indicating that this λ_max shift could be used as a reliable indicator of the association between the sensor and a prospective hydrogen bonding catalyst.

Figure 2.

Response of sensor 1 at 2.22×10⁻⁵ M to increasing amounts of N,N′-di(3,5-bis(trifluoromethyl)-phenyl)thiourea (4) in dichloromethane.

Table 1.

Hydrogen bonding catalyst saturated λ_max values and K_eq values for binding to 1 along with k_cat values for the reaction in eq 2 at 1 mol% catalyst loading in benzene.

Hydrogen bond donor	pKa (DMSO)	λmax (nm)	Keq (M−1)	kcat (s−1)
None		499		–a
2	13.4vii	490	1.67 × 10	1.26 × 10−6
3	17.1xv,b	487	3.23 × 10	1.80 × 10−6
4	8.5vii	477	1.77 × 103	2.09 × 10−5
5	12.8–13.6xvi	473	3.34 × 103	4.90 × 10−5
6	12.8–13.6xvi,c	465	3.47 × 105	1.79 × 10−4

^ak_uncat = 7.50 ×10⁻⁵ s⁻¹.

^bFor 2-naphthol.

^cFirst pK_a may be 1–2 units lower due to dicationic nature of 6.

Figure 3.

Correlation between wavelength shift and K_eq.

Importantly, this sensor coordinates very weakly to water (Δλ_max at saturation = 3.4 nm), which is easily displaced by catalyst. Thus, implementation is simple: sufficient catalyst is added until no further blue shift is seen. At this point, any water has been displaced and the sensor is saturated. The λ_max obtained at this juncture is then used in the correlations to binding (K_eq) and rate (k_rel, see below). For example, a measurement can be made using 10 μg of the sensor and ≤10 mg of the catalyst without special precautions to exclude moisture.

The Diels-Alder reactions of α, β-unsaturated carbonyl dienophiles is well established to undergo rate acceleration with Lewis acids by LUMO lowering of the dienophile ^{xvii,xviii,xix,xx} and a similar activation is believed to operate for hydrogen bonding catalysts.^xxi In order to limit the number of different interactions between the substrates and the hydrogen bonding catalyst, the monodentate substrate methyl vinyl ketone was selected along with a nonbonding diene, cyclopentadiene (eq 2). Rate measurements by NMR^xxii,xxiii showed a range of activities for different hydrogen bonding catalysts (Table 1).

(2)

A plot of ln(k_rel) (k_rel = k_cat/k_uncat) vs the inverse of the λ_max shift (Figure 4) showed a strong correlation indicating that the binding to the sensor provides a reasonable account of the LUMO lowering ability of different hydrogen bonding catalysts. In contrast, the pK_a values do not track well with the reactivity (Table 1, pK_a vs k_cat).

Figure 4.

Correlation of Diels-Alder k_rel values from different hydrogen bonding catalysts with their wavelength shifts of sensor 1.

In conclusion, pyrazinone sensor 1 was found to rapidly provide a read out of the relative reactivity of hydrogen bonding catalysts in the Diels-Alder reaction of methyl vinyl ketone and cyclopentadiene. Namely, catalysts that cause a greater blue shift at saturation of the sensor are more reactive. Thus, it appears that the interaction between hydrogen bond donors and the carbonyl of the sensor provides a good approximation of the LUMO lowering potential available via hydrogen bonding. These preliminary results support the use of sensor 1 as a tool to gauge the relatively reactivity of new hydrogen bonding catalysts and to further the understanding of why different hydrogen bonding catalysts are more effective than others. Exploration of additional hydrogen bonding donors and Lewis acids with the pyrazinone sensor and with other reactions is underway.