Vol. 66, No. 3 (2018)
Chem. Pharm. Bull.
305
Table 2. The Screening of the Benzylation of Primary Alcohol 1b Using X-b
Entry
Solvent
Additive
Temp. [°C]
Time [h]
Yield [%]a)
1
DME
DME
MgO (1.2 equiv.)
TfOH (1mol%)
rt
56
54
5
91
81
95
97
89
80
95
96
2
rt
3
DME
MgO (1.2 equiv.)
MgO (1.2 equiv.)
MgO (1.2 equiv.)
MgO (1.2 equiv.)
MgO (1.2 equiv.)
MgO (1.2 equiv.)
40
60
40
40
40
40
4
5b)
DME
1
DME
28
24
10
9.5
6
CH2Cl2
1,4-Dioxane
Diglyme
7
8
a) Yields were calculated by 1H-NMR analysis using an internal standard. b) The reaction was performed in 1.0mol/L DME solution of 1b.
in the entries 3, 4 in Table 2. The results are summarized in respectively (Chart 2[A]). A similar reaction was conducted
Table 3. For comparison, the yields using DPT-BM (X-a) or in the absence of Y-a because the relative nucleophilicities of
Dudley reagent that were reported in the literature are also Y-b, Y-c, and Y-d were ambiguous owing to the low yields of
listed. The benzylation of simple primary alcohols, 1b and 1c, Z-b, Z-c, and Z-d, which result from the high nucleophilicity
proceeded in high yields (entries 1, 2). Bulky secondary and of Y-a. As a result, Z-b, Z-c, and Z-d were formed in a ratio
tertiary alcohols, 1a and 1d–f, were converted into their cor- of 0:77:23, respectively (Chart 2[B]). To conduct the reaction
responding benzyl ethers in >80% yields which are compa- in the presence of one equivalent of Y-b, other benzylating
rable to those using DPT-BM (X-a) (entries 3, 5, 7, 9). When reagents were used instead of X-b. The reaction employing
the reaction temperature was increased to 60°C, the reactions DPT-BM (X-a) in the presence of Y-b, Y-c, and Y-d also
completed within 1h without a decrease of the yields (entries produced a similar ratio of products as in reaction [A] (Chart
4, 6, 8, 10). Base-labile alcohols, such as 1g and 1h, possess- 2[C]). When X-d was employed as the benzylating reagent
ing a chloroalkyl or acetoxy group, provided benzyl ethers, 2g in the presence of Y-b and Y-c, only a slight increase in Z-c
and 2h, in high yields (entries 11, 12). β-Hydroxyester 1i and was observed, and Z-b was not produced (Chart 2[D]). These
2-(trimethylsilyl)ethanol 1j, which are prone to decompose experimental results clearly indicate that Y-b that possesses
under acidic or basic conditions, were converted into products two t-butyl groups on the triazine ring does not react with a
2i and 2j, respectively, in good yields (entries 13, 14). No benzyl cation species at all.
racemization was observed during the benzylation of 1i. The
As we previously reported,11) these results again indicate
O-benzylation of cholesterol 1k also afforded the correspond- that the formation of Z proceeds intermolecularly between
ing ether 2k in good yields (entries 15, 16). Although tertiary BnOTf and Y, and not via an intramolecular rearrangement.
alcohols, such as 1l and 1m, and their ethers (2l and 2m) are As expected, the introduction of a bulky t-butyl group into the
known to decompose under acidic or heating conditions, 2l triazine core caused a decrease in the rate of the N-benzylation
and 2m were obtained without decomposition using approxi- of Y; therefore, the O-benzylation of alcohols produced high
mately stoichiometric amount of X-b (entries 17–20). In par- yields by using only 1.2 equiv. of X-b, X-c, or X-d.
ticular, better yields were observed at 40°C, which is milder
than the temperature used for Dudley reagent.
We reported that DPT-BM (X-a) dissolved in CDCl3 disap-
peared according to first-order kinetics, and the release of the
The formation of N-benzyltriazinium Z-b from X-b was benzyl cation species from DPT-BM (X-a) may occur via a
not observed in all the reactions listed in Table 3. Therefore, nucleophilic attack to the N-benzyl group by the triflate coun-
we synthesized the benzylating reagents possessing a t-butyl ter anion11) (SN2 reaction, Fig. 2). Thus, we conducted a kinetic
and an alkoxy group on the triazine ring (X-c and X-d) to study of the degradation of X-b, X-c, and X-d because the re-
study the effect of t-butyl group on the formation of the corre- lease rate of BnOTf from X should be affected by substituents
sponding N-benzyltriazinium (Z-c and Z-d). The benzylation on the triazine ring. As a result, all the reagents degraded ac-
of 1a using X-c or X-d resulted in the formation of 2a in cording to first-order kinetics. The order of the first-order re-
higher yields compared with that obtained using DPT-BM action rate constants, k of X, were determined to be DPT-BM
(X-a), and it did not produce N-benzyltriazinium Z-c or Z-d (X-a)>X-d>X-c>X-b (Table 4). This result can be accounted
(Table 1, entries 5, 6).
for by the electronic effect of the substituents on the triazine
23)
,
the
To compare the relative nucleophilicities of the morpho- ring. Based on the Hammett’s substituent constant, σm
linotriazines Y,22)
which are co-products produced from X electron-withdrawing ability of the phenoxy group is larger
and could undergo N-benzylation to form N-benzyltriazinium than that of the methoxy group. In contrast, the t-butyl group
Z, crossover experiments were performed. A mixture of the has an electron-donating ability. The observed first-order rate
equimolar amount of morpholinotriazines, Y-a, Y-c, and Y-d, constants of X are assumed to correlate with the leaving abil-
was treated with the same amount of X-b, which is a source ity of Y altered by the substituents on the triazine. It has been
of benzyl cation species. As a result, N-benzyltriazinium, Z-a, reported that the Hammett equation is applicable to triazine
Z-b, Z-c, and Z-d, were produced in a ratio of 94:0:4:2, derivatives using the sum of σm of two substituents on the tri-