Alcohols and di-tert-butyl dicarbonate: How the nature of the Lewis acid catalyst may address the reaction to the synthesis of tert-butyl ethers

Article abstract of DOI:10.1021/jo061402d

The reaction between alcohols and Boc₂O leads to the formation of ferf-butyl ethers and/or Boc-alcohols, depending on the nature of the Lewis acid catalyst. Product distribution is mainly tuned by the anionic part of the salt. Perchlorates and Inflates, anions with highly delocalized negative charge, give prevalent or exclusive ether formation. On the other hand, Boc alcohols are the main or exclusive products with un-delocalized isopropoxide or low-delocalized acetate ions. The metal ion influences only the reaction rate, roughly following standard parameters for calculating Lewis acidity. A reaction mechanism is supposed, and a series of experimental evidences is reported to support it. These studies allowed us to conclude that, to synthesize tert-butyl ethers, in reactions involving aliphatic alcohols, Mg(ClO₄) ₂ or Al(ClO₄)₃ represents the best compromise between costs and efficiency of the reaction, while, in reactions involving phenols, Sc(OTf)₃ is the best choice, since aromatic tert-butyl ethers are not stable in the presence of perchlorates.

Full text of DOI:10.1021/jo061402d

Alcohols and Di-tert-butyl Dicarbonate: How the Nature of the
Lewis Acid Catalyst May Address the Reaction to the Synthesis of
tert-Butyl Ethers
,
†
†
†
‡
Giuseppe Bartoli,* Marcella Bosco, Armando Carlone, Renato Dalpozzo,
†
†
†
Manuela Locatelli, Paolo Melchiorre, and Letizia Sambri
Dipartimento di Chimica Organica “A. Mangini”, UniVersit a` di Bologna, Viale Risorgimento 4, I-40136
Bologna, Italy, and Dipartimento di Chimica, UniVersit a` della Calabria, Ponte Bucci, cubo 12C, I-87036
ArcaVacata di Rende (Cs), Italy
bartoli@ms.fci.unibo.it
ReceiVed July 6, 2006
2
The reaction between alcohols and Boc O leads to the formation of tert-butyl ethers and/or Boc-alcohols,
depending on the nature of the Lewis acid catalyst. Product distribution is mainly tuned by the anionic
part of the salt. Perchlorates and triflates, anions with highly delocalized negative charge, give prevalent
or exclusive ether formation. On the other hand, Boc alcohols are the main or exclusive products with
un-delocalized isopropoxide or low-delocalized acetate ions. The metal ion influences only the reaction
rate, roughly following standard parameters for calculating Lewis acidity. A reaction mechanism is
supposed, and a series of experimental evidences is reported to support it. These studies allowed us to
conclude that, to synthesize tert-butyl ethers, in reactions involving aliphatic alcohols, Mg(ClO
4 2
) or
Al(ClO represents the best compromise between costs and efficiency of the reaction, while, in reactions
4 3
)
involving phenols, Sc(OTf)
presence of perchlorates.
3
is the best choice, since aromatic tert-butyl ethers are not stable in the
Introduction
SCHEME 1
1
In a preliminary communication, we recently reported that,
in the presence of magnesium perchlorate, the reaction of
alcohols and tert-butyl dicarbonate (Boc2O) leads to the
formation of tert-butyl ethers in high yields, instead of the
expected tert-butyl carbonate (Scheme 1). This methodology
represents a completely new and efficient entry to tert-butyl
ethers, in particular to aromatic ones.
preventing its application to aromatic substrates, which under
The classical route to tert-butyl ethers is, in fact, the addition
these conditions preferentially give Friedel-Crafts ring alky-
2
of alcohols to isobutene, catalyzed by strong acids, which
3
lation.
involves the formation of carbocationic intermediates, thus
Besides, tert-butyl ethers are very important in the chemistry
of protecting groups, since they are stable under strongly basic
conditions (e.g., organolithium and magnesium compounds), but
they are generally underused owing to the harsh conditions
*
To whom correspondence should be addressed. Fax: +39-051-209-3654.
Tel: +39-051-209-3624.
†
Universit a` di Bologna.
Universit a` della Calabria.
‡
(
1) Bartoli, G.; Bosco, M.; Locatelli, M.; Marcantoni, E.; Melchiorre,
P.; Sambri, L. Org. Lett. 2005, 7, 427-430.
2) (a) Greene, T. W.; Wuts P. G. M. In Protecting Groups in Organic
(3) (a) Zhang, K.; Zhang, H.; Xu, G.; Xiang, S.; Xu, D.; Liu, S.; Li, H.
Appl. Catal., A 2001, 207, 183-190. (b) Skthivel, A.; Badamali, S. K.;
Selvam, P. Microporous Mesoporous Mater. 2000, 39, 457-463. (c) Zhang,
K.; Huang, C.; Zhang, H.; Xiang, S.; Liu, S.; Xu, D.; Li, H. Appl. Catal.,
A 1998, 166, 89-95.
(
Syntheis, 3rd ed.; Wiley: New York, 1999; pp 65-67. (b) Kocienski P. J.
In Protecting Groups, 3rd ed.; Thieme: Stuttgart, Germany, 2004; pp 59-
6
1.
10.1021/jo061402d CCC: $33.50 © 2006 American Chemical Society
9
580
J. Org. Chem. 2006, 71, 9580-9588
Published on Web 12/01/2006

Lewis Acid Role in Synthesis of tert-Butyl Ethers
TABLE 1. Reaction between 1-Octanol (1a) (1 equiv) and Boc2O (2) (2.3 equiv) in Dichloromethane at Reflux in the Presence of Various
Lewis Acids (10% mol)
entry
catal
conv (%)a
t (h)
Perchlorates
3a/4a ratioa
2 recovered (equiv)a
1
2
3
4
5
6
7
8
9
Al(ClO4)3‚9H2O
Ce(ClO4)3‚6H2O
Zn(ClO4)2
Mn(ClO4)2‚6H2O
Mg(ClO4)2 dry
Mg(ClO4)2‚6H2O
Ca(ClO4)2‚H2O
Sr(ClO4)2‚3H2O
Ba(ClO4)2‚H2O
HClO4
89
82
84
84
85
86
75
68
45
70
2.3
4
100/0
98/2
98/2
99/1
97/3
99/1
97/3
99/1
100/0
100/0
0
0
0
0
0
0
0.1
0.9
1.8
0.1
4.5
4.5
4.5
4.5
7
7
7
7
1
0
Triflates
11
12
13
14
15
16
In(OTf)3
Sc(OTf)3
Al(OTf)3
Ce(OTf)3‚xH2O
Zn(OTf)2
Mg(OTf)2
81
80
84
73
67
64
4.5
4
4.5
3.5
7
99/1
100/0
100/0
97/3
94/6
94/6
0
0
0
0
0.13
0.7
7
Miscellaneous
17
18
19
20
21
22
23
24
25
26
27
ZnBr2
ZnI2
ZnCl2
CeCl3 dry
MgBr2
Mg(NO3)2‚6H2O
MgCl2
InCl3
27
39
26
65
48
18
47
100
100
100
100
7
7
7
7
7
7
7
33/67
28/72
19/81
18/82
15/85
11/89
9/91
>2
>2
>2
1.3
1.4
>2
>2
0.64
1.2
1.2
0.7
b
4.5
4/96
Ti(iPrO)4
Zn(OAc)2
Zn(OAc)2‚2H2O
7
7
7
0/100
0/100
0/100
a
Determined by NMR. b Dioctyl carbonate (3%) was also detected.
2
required for their formation; thus, to place a mild reaction at
the chemists’ disposition appears to be very important.
Moreover, an evaluation of the whole set of parameters
involved in this new synthesis of tert-butyl ethers is needed. In
this paper, we report an accurate and detailed study both of the
Lewis acid catalysis and of the other factors such as solvent
and temperature, which could control the reaction course and
the product distribution, to understand the reaction mechanism
and to optimize the reaction conditions.
SCHEME 2
by NMR integration of the most typical signal of each compound
after usual workup.
At the end of the reaction, the NMR analysis of the crude
always showed only starting materials, ether 3a and/or carbonate
4
a.
Since Lewis acids can catalyze tert-butyl ethers cleavage
Results
5
under certain experimental conditions, stability of pure 3a was
tested and it was found that the employed Lewis acids did not
decompose 3a under these experimental conditions.
As reported in Table 1, only perchlorates and triflates are
able to give high yields and selectivity toward ether formation,
and among them, Mg(ClO4)2, Al(ClO4)3, and Sc(OTf)3 appeared
to be the most efficient catalysts, because 3a was obtained in
high yields and reasonably short times. Therefore, we carried
out screenings on the influence of other parameters on product
distribution and on other substrates considering only these
catalysts.
Aliphatic Primary Alcohols. First the influence of the Lewis
acid catalyst on the product distribution was tested by modifying
the nature of both the metal and the counterion. The reaction
between 1-octanol (1a) and Boc2O (2) was chosen as the model
reaction (Table 1). All the reactions were carried out in
dichloromethane under reflux adding a 10% molar amount of
catalyst, with a 2.3 molar excess of 2. An excess of Boc2O was
4
necessary owing to its instability under heating or in acid
conditions (Scheme 2).1
All reactions were monitored by TLC and GC-MS analyses
and were quenched after the complete disappearance of one of
the reagents or, anyway, after 7 h. Product ratio was obtained
In nonpolar solvents such as chloroform or toluene, the
limiting parameter was the solubility of the catalyst. Solubility
(
4) (a) Pope, B. M.; Sheu, S.-J.; Stanley, R. L.; Tarbell, D. S.; Yamamoto,
(5) Bartoli, G.; Bosco, M.; Carlone, A.; Locatelli, M.; Marcantoni, E.;
Melchiorre, P.; Sambri, L. AdV. Synth. Catal. 2006, 348, 905-910 and
literatures cited therein.
Y. J. Org. Chem. 1978, 43, 2410-2414. (b) Dean, C. S.; Tarbell, D. S.;
Friederang, A. W. J. Org. Chem. 1970, 35, 3393-3397.
J. Org. Chem, Vol. 71, No. 26, 2006 9581

Bartoli et al.
TABLE 2. Influence of Temperature and Solvent on the Reaction between 1-Octanol (1a) (1 equiv) and Boc2O (2) (2.3 equiv) in the Presence
of the Most Efficient Catalysts (10% mol), Except Otherwise Mentioned
entry
catal
solvent
T (°C)
t (h)b
conv (%)a
3a/4a ratioa
1
2
3
4
5
6
7
8
9
Al(ClO4)3‚9H2O
Al(ClO4)3‚9H2O
Al(ClO4)3‚9H2O
Al(ClO4)3‚9H2O
Mg(ClO4)2
CH2Cl2
CH2Cl2
neat
neat
CH2Cl2
neat
neat
neat
CH2Cl2
CH2Cl2
neat
40
25
40
25
40
40
25
40
40
25
40
25
2.3
10
8
24
4.5
7
18
2
4
89
69
87
82
86
82
66
98
80
85
89
92
100/0
100/0
100/0
100/0
99/1
99/1
100/0
99/1
100/0
98/2
100/0
100/0
c
Mg(ClO4)2
Mg(ClO4)2
Mg(ClO4)2
d
Sc(OTf)3
Sc(OTf)3
Sc(OTf)3
Sc(OTf)3
1
1
1
0
1
2
4.5
1
2
neat
a
Determined by NMR. b Times refer to the disappearance of Boc2O in the reaction mixture. c Data from ref 1. d 3.5 equiv of 2 is used.
TABLE 3. Influence of the Temperature on the Reaction between 2-Octanol (1b) (1 equiv) and Boc2O (2) (2.3 equiv) in the Presence of the
Most Efficient Catalysts (10% mol), Except Otherwise Mentioned
entry
catal
solvent
T (°C)
t (h)b
conv (%)a
3b/4b ratioa
1
2
3
4
5
6
7
Mg(ClO4)2c
neat
CH2Cl2
neat
CH2Cl2
neat
CH2Cl2
neat
50
40
40
40
25
25
25
30
4
6
22
67
5.5
3.5
85
77
99
77
73
79
86
100/0
100/0
100/0
100/0
100/0
100/0
100/0
d
Mg(ClO4)2
Mg(ClO4)2
d
Al(ClO4)3‚9H2O
Al(ClO4)3‚9H2O
Sc(OTf)3
Sc(OTf)3
a
Determined by NMR. Times refer to the disappearance of Boc2O in the reaction mixture. c Data from ref 1. d 3.5 equiv of 2 is used.
being equal, the reaction proceeded in comparable rate and
yields as in dichloromethane, but solvent removal was more
difficult because of their higher boiling points. If the solubility
was lower, the reaction rate slowed down. In more polar
solvents, such as diethyl ether and THF, conversions generally
decreased, although the catalyst solubility increased. When the
catalyst was soluble in the mixture and its viscosity was not
too high, neat conditions gave better or similar results. As a
consequence, the best choice appeared to be dichloromethane
or neat conditions.
Then we evaluated the influence of decreasing temperature
on product distribution especially to depress the thermal acid-
catalyzed decomposition of Boc2O (Table 2). Reactions were
carried out at 40 and 25 °C in a thermostatic bath in neat or in
dichloromethane solutions. Results reported in Table 2 showed
that a temperature decrease did not significantly influence either
conversion into ether 3a or Boc2O decomposition. In fact, the
reaction never went to completion before complete decomposi-
tion of the excess of Boc2O and times reported in Table 2 refer
to its disappearance.
corresponding adamantyl tert-butyl ether (3c) after 22 h reflux-
ing in dichloromethane.
Phenols. The reaction of Boc2O with phenol (1d) was then
examined. A screening among various Lewis acids allowed
concluding that also in this case Mg(ClO4)2, Al(ClO4)3, and Sc-
(OTf)3 are the most efficient catalysts for the tert-butoxybenzene
(3d) synthesis.
Conversely from aliphatic alcohols, the reaction appeared to
be more complex. In fact, we found that prolonged reaction
times decreased yields in 3d and ring-alkylated products started
to be detected among the reaction products under Al(ClO4)3
and Mg(ClO4)2 catalysis. On the other hand, the Sc(OTf)3-
catalyzed reaction did not show any variation in yields even
after 17 h.
To understand this anomaly, reactions were performed in a
NMR tube and monitored at 20 °C. (Figure 1). Since CO2 and
isobutene gases are evolved, the NMR tube was not sealed. The
Al(ClO4)3 amount was increased to a 20% mol amount, to
complete the reaction in conveniently short times. However,
solvent evaporation could not be avoided in such an extent to
allow the consequent concentration variation to be neglected.
Under these conditions, a rigorous rate law could not be
deduced. Nevertheless, some interesting considerations on the
reaction course can be obtained.
To increase yields up to almost quantitative, we had to
increase the excess of Boc2O up to 3.5 equiv (Table 2, entry
8
).
Secondary and Tertiary Aliphatic Alcohols. The reaction
of 2-octanol (1b) was tested as representative of a secondary
alcohol under the reported experimental conditions (Table 3).
The observed trend was almost superimposable with that of
The profile of the Sc(OTf)3-catalyzed reaction (Figure 1a)
shows, as expected, an increase of product amount within about
3 h and then a plateau at about 89% of conversion. At the same
time Boc2O amount decreases to zero and tert-BuOH amount
increases until about 2.3 equiv. Also isobutene was detected
among the reaction products, but its amount could not be
measured, since it rapidly moved away from the mixture.
Some interesting differences can be observed for the Al-
(ClO4)3-catalyzed reaction (Figure 1b). First, a rate enhancement
is observed only when some tert-butanol is formed (about 1.3
equiv, after 100 min). It acts as the solvent for the catalyst,
1
-octanol, with the only difference that the reaction time was
1
generally longer and conversion was lower.
Once more, an increase of Boc2O up to 3.5 equiv allowed an
almost quantitative yield to be reached after 6 h under neat
conditions (Table 3, entry 3).
Finally, tertiary alcohols failed to give the desired ether and
only starting material was recovered after prolonged reaction
1
times. Only 1-adamantol (1c) led to a 17% yield of the
9582 J. Org. Chem., Vol. 71, No. 26, 2006

Lewis Acid Role in Synthesis of tert-Butyl Ethers
2
FIGURE 1. Profiles of tert-butoxybenzene (3d, green), tert-butanol (6, magenta), and Boc O (2, blue) amounts in the reaction of phenol (1d) with
2
3 4 3 3
in the presence of Sc(OTf) (5% mol) (a) or Al(ClO ) (20% mol) (b) at 20 °C in CDCl .
SCHEME 3. Product Distribution after Decomposition of
Ether 3d
which is not completely dissolved in the reaction mixture at
the beginning of the reaction, thus confirming that the reaction
rate is favorably influenced by homogeneous catalysis. In fact,
when we carried out an experiment using a dichloromethane/
tert-butanol (6/1) mixture, in which Al(ClO4)3 is completely
dissolved, we observed an increase in reaction rate (3 h versus
4
.5 h to reach the maximum yield).
tert-Butoxybenzene (3d) increases as long as Boc2O is present
into the reaction mixture. The top is reached after 4.5 h (55%),
when 0.1 equiv of 2 is still present. Then the amount of 3d
starts to slowly decrease and the formation of ring-alkylated
phenols is observed. The appearance of ring-alkylated products
only after 2 disappearance in the reaction mixture means that,
These results evidence that ether decomposition starts only
when insignificant amounts of Boc2O are present in the reaction
mixture, very likely because it strongly chelates the Lewis acid.
Therefore, the acid is not available for O-tert-Bu cleavage. Only
at the end of the reaction, some metal ions are free to coordinate
to the ether oxygen causing the cleavage of the ethereal bond
of 3d.
5
conversely from 3a, 3d is cleaved by the Lewis acid.
Then, we allowed 3d to stand at room temperature for 24 h
in the presence of 20% Al(ClO4)3 and actually we recovered
phenol (69%) and tert-butylphenols (31%) (Scheme 3).
Finally, we wanted to compare the previous reported condi-
1
tions [i.e. Mg(ClO4)2 (10% mol), 40 °C] with the best
Mg(ClO4)2 itself is able to decompose pure tert-butoxyben-
zene (3d) when refluxed in dichloromethane for 5 h, giving
unaffected 3d (16%), 1d (79%), 7 (1%), and 8 (4%). Neverthe-
less, the synthesis of 3d catalyzed by Mg(ClO4)2 was already
reported in our preliminary communication and the product was
recovered in 90% yields, when the reaction was carried out for
conditions set now [i.e. Sc(OTf)3 (5% mol) at room temperature]
in reactions of Boc2O with some more phenols (Table 4). Sc-
(OTf)3 appeared to be generally more effective.
In some cases, Boc2O amount was raised to 3.5 equiv to
overcome the lack of reactant due to self-decomposition.
However, electron-withdrawing substituents slowed down ether
formation rate, so Boc2O self-decomposition occurred at a
comparable rate and an increase of its amount did not result in
1
3 h, in refluxing dichloromethane, in the presence of 2.3 equiv
1
of Boc2O/mol of 1d.
J. Org. Chem, Vol. 71, No. 26, 2006 9583

Bartoli et al.
TABLE 4. Comparison between Mg(ClO4)2 and Sc(OTf)3 Catalysis in the Reaction between Phenols (1d-i) and Boc2O (2) [(d) Ar ) Ph, (e)
Ar ) 4-OMePh, (f) Ar ) 3-ClPh, (g) Ar ) 4-NO2Ph, (h) Ar ) 4-CHOPh, (i) Ar ) 1-Naphthyl]
entry
substrate
catal
solvent
2 (equiv)
T (°C)
t (h)a
yield (%)
b
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
1d
1d
1d
1e
1e
1e
1f
Mg(ClO4)2 (10% mol)
Mg(ClO4)2 (10% mol)
Sc(OTf)3 (5% mol)
Mg(ClO4)2 (10% mol)
Sc(OTf)3 (5% mol)
Sc(OTf)3 (5% mol)
Mg(ClO4)2 (10% mol)
Sc(OTf)3 (5% mol)
Sc(OTf)3 (5% mol)
Mg(ClO4)2 (10% mol)
Sc(OTf)3 (5% mol)
Mg(ClO4)2 (10% mol)
Sc(OTf)3 (5% mol)
Sc(OTf)3 (5% mol)
Mg(ClO4)2 (10% mol)
Sc(OTf)3 (5% mol)
CH2Cl2
Neat
CH2Cl2
Neat
CH2Cl2
CH2Cl2
Neat
CH2Cl2
CH2Cl2
Neat
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
2.3
2.3
2.3
2.3
2.3
3.5
2.3
2.3
3.5
2.3
2.3
2.3
2.3
3.5
2.3
2.3
40
40
20
40
20
20
40
20
20
40
20
40
20
20
40
20
13
21
5
24
5
16
32
5
23
32
5
13
5
18
54
5
90 (3d)
93 (3d)
84 (3d)
84 (3e)
b
b
87 (3e)
94 (3e)
82 (3f)
72 (3f)
82 (3f)
30 (3g)
35 (3g)
30 (3h)
48 (3h)
48 (3h)
b
1f
1f
1
1
1
1
1
1
1
1g
1g
1h
1h
1h
1i
b
65 (3i)
65 (3i)
1i
a
b
Times refer to the disappearance of Boc2O in the reaction mixture. Data from ref 1.
a yield improvement. In these cases, the cheapest practice
becomes the use of only a low excess of Boc2O, followed by
separation of ether from the unreacted starting alcohol, a very
easy process, and finally alcohol can be recycled.
and softness of anions are generally correlated to their nucleo-
philicity or to their capability to delocalize the negative charge.
Perchlorate is a very stable anion; in other words it can be
classified as “soft”. On the other hand, bivalent and trivalent
metal cations are generally considered “hard”. As a consequence,
they do not match with perchlorate anion to form a tight ion
pair.
Triflates are a little more nucleophilic than perchlorates, since
the nucleophilicity of an anion is related both to the electrone-
gativity of the core atom (sulfur instead of chlorine) and to the
Discussion
Influence of the Lewis Acid on the Product Distribution.
From data reported in Table 1, we noticed a good correlation
6
between the reaction rate and the pKh (hydrolysis constant) of
the metal ion in both perchlorate and triflate series, although
the reactions were carried out in a nonpolar solvent such as
dichloromethane. The lower the pKh, the faster is the reaction.
In the alkaline-earth element series, for example, reactivity
follows the order Mg (pKh ) 11.44) > Ca (pKh ) 12.85) > Ba
number of oxygen atoms involved in the delocalized system
10
(
three instead of four). As a consequence triflates mostly give
ether 3a, but significant amounts of carbonate 4a are also
recovered. For the same reasons, nitrate ion is harder and, in
fact, yields of 4a are increased (Table 1, entry 22).
(
(
pKh ) 13.47) (Table 1, entries 6, 7, 9). The trivalent aluminum
pKh ) 4.97) should be much more reactive and, in fact, gave
Halide anions give a mixture between ether 3a and carbonate
4
a, since they are still harder. Finally, a highly nucleophilic
the highest conversion (89%) in the least time (Table 1, entry
).
The efficiency of triflates is lower than that of perchlorates,
anion such as acetate leads to the exclusive formation of
carbonate 4. This trend has been observed into a series of both
zinc (Table 1, entries 3, 15, 17, 18, 19, 26, 27) and magnesium
salts (Table 1, entries 6, 16, 21, 22, 23). Moreover, other metal
halides or Ti(OPr)4, a highly nucleophilic anion, confirm our
statement.
1
and the reaction is less selective toward ether 3a formation, as
shown by the comparison between perchlorates and triflates of
Al, Ce, Zn, and Mg (Table 1, entries 1/13, 2/14, 3/15, and 5/16,
respectively). The higher efficiency of the metal ion of per-
chlorate vs triflate to act as Lewis acid has been already observed
In conclusion, dissociated salts gave tert-butyl ethers, whereas
intimate ion pairs led to Boc alcohols.
7
8
in the acetylation of alcohols and in the Fischer esterification.
Reaction Mechanism. The synthesis of ethers is peculiar to
In conclusion, the nature of the metal ion only influences
the reaction rate but not the product distribution.
Boc2O, which has a characteristic decomposition (Scheme 4).4
Other dicarbonates, which follow different decomposition
On the other hand, the influence of the anion in product
distribution was clearly observed. In fact, with increase of the
nucleophilicity of the anion, the yield of carbonate 4a increased
despite formation of 3a. The role of the counterion can be
4a
pathways or do not decompose at all, are unable to give the
same reaction. Therefore, a mechanism involving an interaction
between some electrophilic species arising from Boc2O self-
decomposition and alcohol (1) appeared to be a reasonable
hypothesis.
9
explained in term of HSAB theory. In this theory, hardness
Actually Boc2O self-decomposition is supposed to occur by
initial scission of an internal carbonyl oxygen bond and
subsequent loss of carbon dioxide followed by proton abstraction
and decarboxylation again (Scheme 4, path a) or alternatively
by formation of the tert-butyl cation (path c), which decomposes
(
6) Kobayashi, S.; Nagayama, S.; Busujima T. J. Am. Chem. Soc. 1998,
1
2
20, 8287-8288. Kobayashi, S. Manabe K. Acc. Chem. Res. 2002, 35, 209-
17.
9
a
(
7) Bartoli, G.; Bosco, M.; Dalpozzo, R.; Marcantoni, E.; Massaccesi,
M.; Sambri, L. Eur. J. Org. Chem. 2003, 4611-4617.
8) Bartoli, G.; Boeglin, J.; Bosco, M.; Locatelli, M.; Massaccesi, M.;
Melchiorre, P.; Sambri, L. AdV. Synth. Catal. 2005, 347, 33-38.
9) (a) Hati, S.; Datta D. J. Chem. Soc., Dalton Trans. 1994, 2177-
180. (b) Pearson, R. G. J. Am. Chem. Soc. 1963, 85, 3533-3539.
(
(
(10) Wulfsberg G. Principles of descriptiVe inorganic chemistry; Wad-
sworth: Monterey, CA, 1987; Chapter 2.
2
9584 J. Org. Chem., Vol. 71, No. 26, 2006

Lewis Acid Role in Synthesis of tert-Butyl Ethers
SCHEME 4. Decomposition of Boc
Dataa
2
O from Literature
SCHEME 5. Proposed Mechanism
N
SCHEME 6. A Possible S 2 Alternative Mechanism
N
SCHEME 7. A Possible S 1 Alternative Mechanism
with potential electron donor atoms such as diethyl ether and
THF can compete for the Lewis acid, so lowering the reaction
rate.
a
See ref 4a for paths a,c and ref 4b for path b.
to isobutene (5) by loss of a proton. Coupling to di-tert-butyl
ether has been never observed.
Then, 10 can undergo nucleophilic addition of the alcohol to
give 16, whose efficiency depends on the Lewis acidity of the
metal ion. Among parameters used to estimate the Lewis acidity
of ions, we found a good correlation with hydrolysis constants:
However, a cyclic transition state was also proposed (path
b).4b
2
The Ba(ClO4)2-catalyzed reaction (Table 1, entry 9) provides
the first useful suggestions to exclude this hypothesis. After 7
h, the ether 3a was recovered only in 45% yield, without
extensive decomposition of the excess of 2.
cations with great pKh values have small (charge) /(ionic radii)
6
values and by consequence low Lewis acidity, so the lower
pKh, the more tightened is the chelate complex 10.
However, dependence of the efficiency of the reaction on
pKh could suggest the generation of a Brønsted acid from the
hydrolysis of the crystallization water by the metal ion.12
However this hypothesis was easily excluded, since perchloric
acid, a classical Brønsted acid, is less efficient than the best
Lewis acid catalysts (Table 1, entry 10).
Ba(ClO4)2 and Boc2O were allowed to react under the same
experimental conditions in a blank run, in the presence of 1,4-
di-tert-butylbenzene as the internal standard. After 7 h, the NMR
signal ratio showed a decrease of Boc2O signal within 9-10%.
In addition, if free tert-butyl cations were involved in the
reaction mechanism, we should observe ring-alkylated products
in the reaction with aromatic systems. It is well-known, in fact,
that phenols undergo electrophilic substitution at the aromatic
Phenols with electron-withdrawing substituents are worse
nucleophiles, and consequently their attack onto 10 is delayed.
Tertiary alcohols are unable to give this reaction, since a
crowded carbonyl function like that of 10 offers steric resistance
to the nucleophile approaching. As a consequence, we never
observed formation of di-tert-butyl ether from reaction of the
11
ring in the presence of alkyl cations. To corroborate this
statement, we carried out an independent test, in which 2 was
decomposed in the presence of Mg(ClO4)2 and anisole as the
carbocation scavenger. We did not observe the formation of
tert-butyl anisole, not even in trace amounts.1
1
byproduct tert-butanol and Boc2O. Only 1-adamantol, less
hindered than other tertiary alcohols owing to its shape, gave
partial attack.
Results of these trials contrast with any hypothesis of ether
formation from reaction between alcohol 1 and an electrophilic
tert-butyl species arising from Boc2O self-decomposition and
allowed us to confidently affirm that the ether synthesis and
the Boc2O self-decomposition are reactions which run inde-
pendent and parallel.
Then ether 3 could be originated together with activated
Boc2O (10), by the intervention of adventitious tert-butanol from
Boc2O self-decomposition on activated alcohol 16 via a SN2
pathway (Scheme 6). However, this option can be excluded,
because aromatic substrates cannot undergo any nucleophilic
substitution, but tert-butyl aryl ethers 3d-i are obtained in high
yields (Table 4).
The stability of intermediate 16 depends, in our opinion, on
the complexation of the metal ion, in other words, on the
measure in which the charged oxygen in 16 or the catalyst anion
matches with the metal ion.
As a rationalization of the whole of the reaction, we propose
the independent mechanism, depicted in Scheme 5, which
involves the formation of a mixed dicarbonate. This mechanism
1
was already hypothesized in our previous communication, and
it is now supported by experimental evidence.
The metal ion activates Boc2O, forming the chelate complex
1
0, already responsible of Boc2O decomposition (Scheme 4).
Highly nucleophilic anions such as isopropoxide and acetate
do not dissociate from their metal counter part, and as a
The solvents medium is an important factor, because solvent
(
11) Harris, T.; Campbell, C.; Tang, Y. J. Phys. Chem. A 2006, 110,
(12) Wabnitz, T. C.; Yu, J.-Q.; Spencer, J. B. Chem.sEur. J. 2004, 10,
484-493.
2
246-2252.
J. Org. Chem, Vol. 71, No. 26, 2006 9585

Bartoli et al.
SCHEME 8. Decomposition of a Mixed Pyrocarbonate to tert-Butyl Ether
SCHEME 9. Possible Heterolytic Breakdowns
ate (19) and dicyclopentyl carbonate (20). Treating this mixture
with Mg(ClO4)2 (Scheme 8), under our experimental conditions,
1
4
we observed formation of ether 3j from 21, while 19 and 20
remained unaffected.
Unfortunately, the lack of efficient syntheses, which allow
significant amounts of pure mixed dicarbonate to be obtained,
actually prevents us from accurately studying this pivotal
decomposition stage. Therefore, every mechanistic hypothesis
could appear speculative. Nevertheless, some hypotheses are
very unlike on the basis of both the present results and literature
data.
SCHEME 10. Decomposition of 18 through a Six-Ring
Mechanism
Homolytic breakdowns, for example, can be ruled out because
it has been demonstrated that coupling of phenoxy and tert-
butyl radicals leads to ring-alkylated products even if it occurs
1
5
inside the solvent cage.
A heterolytic breakdown like that supposed in Boc2O
decomposition (Scheme 4, path a) has to occur leading to cation
1
2 and alcoholate 22 (Scheme 9), but according to Tarbell’s
4
a
mechanistic hypothesis, large amounts of isobutene should
arise by proton abstraction of alcoholate on 12. Although
isobutene amount cannot be measured, since it evolves rapidly
from the mixture, such a mechanism should lead to large
amounts of unreacted alcohol, but we had demonstrated that
starting materials are completely consumed if a large excess of
Boc2O is used.
consequence, 16 is a “naked” anion. So it is more prone to
restore the carbon-oxygen double bond, so favoring elimination
of the good leaving group carbonate monoester and leading to
the exclusive formation of carbonate 4 (path b). Once formed
under the experimental conditions, 4 is stable, so it remains
unaltered and neither can be responsible for further transforma-
tions nor its formation can be involved into an equilibrium
reaction, since carbon dioxide evolves from the reaction mixture.
On the other hand “naked” metal ions, such as those of
perchlorates and triflates, tightly bind to the oxygen anion of
Alternatively, we can suppose that 12 decomposes to tert-
butyl cation (15) before interacting with the alcoholate. How-
ever, in aromatic systems under homogeneous catalysis, ionic
mechanisms lead to mixtures of C- and O-alkylated products.16
In the Sc(OTf)3-catalyzed reaction we never observed ring-
alkylated products, while in the case of Al(ClO4)3 they are
evidently formed by decomposition of 1d. Therefore, coupling
of 15 with 1 or its salt 22 can be excluded.
16 and allow an equilibrium between 16 and 17 to be established
(path a).
Ether 3 could arise from intermediate 17, if it acts as an
intimate ion pair between a tert-butyl cation and a mixed
dicarbonate hydrate, via an intramolecular SN1-like attack of
alcoholic oxygen on incipient tert-butyl cation (Scheme 7). Once
more the absence of any ring addition product even in traces in
the reaction of highly activated 4-methoxyphenol (1e) demon-
strates that recombination of the two partners must be much
faster than any rotation into the solvent cage, which is very
unlikely in our opinion.
The opposite breakdown, which can give the cationic species
2
3 and tert-butylate (14), can be also ruled out. In fact, Tarbell
affirmed that these cations are unable to give alkyl cations, but
4a
they couple with alcoholate to give carbonate. Adamantyl
4a
dicarbonate itself decomposes to carbonate, but we never
observed tert-butyl adamantyl carbonate among the reaction
products, so indicating that such a breakdown cannot happen
in our system.
Then intermediate 17 is able to lose tert-butanol leading to
the metalated mixed dicarbonate 18.
The most conceivable hypothesis remains that the decomposi-
tion of 18 occurs through a synchronous mechanism in a “six-
We were unable to demonstrate the formation of mixed
dicarbonates like 18 along the reaction course, by following
the reaction by NMR, since we never observed signals assign-
able to this compound in a free region of the spectrum.
However, we could independently demonstrate that 18
(14) Decomposition of 20 is fast, as expected, since asymmetric
dicarbonates must be much more reactive than the symmetric ones, as
already demonstrated by the Ba(ClO4)2-catalyzed reaction, where ether
formation (from decomposition of asymmetric dicarbonate) largely over-
whelms Boc2O (a symmetric dicarbonate) decomposition.
13
decomposes to ether. In fact, following literature suggestions,
tert-butyl cyclopentyl dicarbonate (21) was obtained as an
inseparable mixture together with tert-butyl cyclopentyl carbon-
(
15) Galindo, F.; Miranda, M. A.; Tormos, R. J. Photochem. Photobiol.,
A 1998, 117, 17-19.
16) Wilson, K.; Adams, D. J.; Rothenberg, G.; Clark, J. H. J. Mol. Catal.,
A 2000, 159, 309-314.
(
(13) Basel, Y.; Hassner, A. J. Org. Chem. 2000, 65, 6368-6380.
9586 J. Org. Chem., Vol. 71, No. 26, 2006

Lewis Acid Role in Synthesis of tert-Butyl Ethers
1
ring” transition state (TS, Scheme 10) according to the
The crude mixture was then submitted to H-NMR analysis, and
alternative hypothesis of Boc2O decomposition (Scheme 4, path
relative amounts of remaining 1a and 2 or products 3a and 4a were
4
b
determined by integration of signals at δ ) 3.61 (CH O of 1a),
b).
2
1
.53 (Me
Temperature Tests. In a two-necked flask equipped with a
magnetic stirring bar and a condenser coil, Lewis acid (0.10 mmol)
and 1a (1.0 mmol) were dissolved in 1.5 mL of CH Cl or mixed
3 2 2
C of 2), 3.31 (CH O of 3a), or 4.05 (CH O of 4a).
The substantial independence from temperature of the reaction
supports the idea of a concerted mechanism. Moreover, recent
DFT calculation on the alkylation of phenols demonstrated that
tert-butyl phenyl ether arises from a neutral six-ring concerted
mechanism even under acidic conditions, while C-alkylated
products arise from the ionic rearrangement of the early formed
ether.11
2
2
together. Then 2 (2.3 mmol) was added. The mixture was stirred
at 40 or at 25 °C with a thermostatic bath until the TLC analysis
revealed the presence of Boc O. The crude reaction mixture was
2
1
worked up as usually and was then submitted to H-NMR analysis.
Relative amounts of remaining 1a and products 3a and 4a were
determined by integration of signals at δ ) 3.61 (CH
.31 (CH O of 3a), or 4.05 (CH O of 4a).
Reaction of 1b,c and 2. In a two-necked flask equipped with a
magnetic stirring bar and a condenser coil, Mg(ClO (0.10 mmol)
and 1b,c (1.0 mmol) were dissolved in 1.5 mL of CH Cl . Then 2
3.5 mmol) was added. The mixture was stirred at reflux until the
2
O of 1a),
Conclusions
3
2
2
The study of the reaction between alcohols and tert-butyl
dicarbonate allowed optimizing the choice of the catalyst for
addressing the reaction toward tert-butyl ethers or Boc-protected
alcohols. We found that the metal ion of the catalyst influences
only the reaction rate, while product distribution is tuned by
counterions. Perchlorates and triflates, anions with highly
delocalized negative charge, give prevalent or exclusive ether
formation. On the other hand, Boc alcohols are the main or
exclusive products with undelocalized isopropoxide or low-
delocalized acetate ions.
4 2
)
2
2
(
TLC analysis revealed the presence of 2. The crude reaction mixture
1
was worked up as usually and then submitted to H-NMR analysis;
relative amounts of 1b (23%) and 3b (77%) were estimated. The
crude mixture from reaction of 1c was worked up as usual and
filtered on a short silica gel column with recovery of 3c (17% yield
and 99% purity) and 1c (82% yield and 99% purity). 3c was
1
submitted to H-NMR analysis, which gave a spectrum superim-
posable to that of literature data.17
The results of this study allowed concluding that, to obtain
tert-butyl ethers, the reaction course must be addressed to the
formation of a mixed dicarbonate intermediate, which very likely
decomposes through a concerted mechanism.
In a two-necked flask equipped with a magnetic stirring bar and
a condenser coil, Mg(ClO ) (0.10 mmol) and 1b (1.0 mmol) were
4 2
mixed together. Then 2 (3.5 mmol) was added. The mixture was
stirred at 40 °C in a thermostatic bath until the TLC analysis
revealed the presence of 2 (6 h). The crude reaction mixture was
worked up as usually and filtered on a short silica gel column, and
Mg(ClO4)2 represents the best compromise between costs and
efficiency of the reaction in aliphatic systems. Al(ClO4)3, which
is a little more expensive and efficient, can be used if low
temperatures are necessary. We discarded the use of Sc(OTf)3
from an economical point of view, because it did not signifi-
cantly improve the reaction efficiency and its cost is 100 times
higher than that of perchlorates. With a large excess of Boc2O,
the reaction is almost quantitative in the aliphatic series.
Finally, both Mg(ClO4)2 and Sc(OTf)3 behave as comparably
efficient catalysts in the synthesis of aromatic tert-butyl ethers.
The former Lewis acid is cheaper, but the reaction course must
be carefully followed, since prolonged reaction times cause
decomposition of the product. On the other hand, the more
expensive Sc(OTf)3 offers both warranty that no decomposition
of the product occurs and chance of low reaction temperatures,
which allows survival of sensitive functions. Overall yields are
not very different for both Lewis acids; therefore, the choice
can be made at each time.
1
recovered 3b (99% yield and purity) was submitted to H-NMR
analysis, which gave a spectrum superimposable on that for the
literature data.1
General Procedure for the Synthesis of Aromatic tert-Butyl
Ethers. All the reactions were carried out in CH
solvent. In a two-necked flask equipped with a magnetic stirring
bar and a condenser coil, Mg(ClO (0.10 mmol) or Sc(OTf) (0.05
mol) and the alcohol 1d-i (1.0 mmol) were dissolved in 1.5 mL
of CH Cl . Then 2 (2.3 mmol, or 3.5 mmol) was added, and a
2 2
Cl or without
)
4 2
3
2
2
bubbling was immediately observed. The mixture was stirred at
reflux until the TLC analysis revealed the presence of 2. The crude
reaction mixture was diluted with water and extracted with CH
2
-
Cl . The organic layer was separated, dried over MgSO , and
2
4
filtered, and the solvent was removed by rotary evaporation. The
tert-butyl ethers were separated from the residual alcohol by flash
chromatography on silica gel with a mixture of 9:1 petroleum ether/
1
Et
2
O. H-NMR analysis gave a spectrum superimposable on that
1
for the literature data.
When the reaction was carried out without solvent, the alcohol
and the catalyst were mixed together; then 2 was added, and the
mixture was heated at 40 °C until the TLC analysis revealed the
presence of 2.
The reactions with Sc(OTf) were carried out as above-described
3
but at 20 °C in a thermostatic bath.
Experimental Section
1
General Methods. The H NMR spectra were recorded at 400
MHz or 600 MHz, while 13C NMR spectra were recorded at 100
MHz. Purification of reaction products was carried out by flash
chromatography on silica gel (230-400 mesh). Commercial grade
reagents and solvents were used without further purification. All
Kinetic Experiment. 1d (0.32 mmol) and Al(ClO
)
4 3
or Sc(OTf)
3
starting alcohols, salts, and Boc
2
O were purchased and used as
(0.06 mmol) were dissolved in 0.6 mL of CDCl and submitted to
3
1
received. All reaction products were already fully characterized
H-NMR analysis at 600 MHz. After the instrument was set, 2 (0.73
mmol) was added. A spectrum was recorded immediately after and
then every 15 min during 6 h. Then spectra were recorded after 7,
previously.1
General Procedure for the Reactions of Table 1. In a two-
necked flask equipped with a magnetic stirring bar and a condenser
coil, Lewis acid (0.10 mmol) and 1a (1.0 mmol) were dissolved in
9
, 11, 13, and 15 h. Amounts of 3d, 6, and 2 were measured by
integration of signals at δ ) 1.35 (CMe of 3d), 1.23 (CMe of 6),
and 1.53 (CMe of 2) with respect to the whole aromatic part of
3
3
1
.5 mL of solvent. Then 2 (2.3 mmol) was added and a bubbling
was immediately observed. The mixture was stirred at reflux until
the TLC analysis revealed the presence of Boc O or, anyway, for
h. The crude reaction mixture was diluted with water and extracted
with CH Cl . The organic layer was separated, dried over MgSO
and filtered, and the solvent was removed by rotary evaporation.
3
the spectrum and related to starting ratio. Among the signals
recorded during these experiments, we never observed peaks
assignable to mixed pyrocarbonate in a free region of the spectrum.
2
7
2
2
4
,
(17) Miller, J. B.; Salvador, J. R. J. Org. Chem. 2002, 67, 435-442.
J. Org. Chem, Vol. 71, No. 26, 2006 9587

Bartoli et al.
On the other hand, signals assignable to 5 [δ ) 1.82 (Me), 4.75
NMR analysis. A multiplet at 6.8-6.9 typical of 3d was less than
(
CH
d (0.32 mmol) and Al(ClO
.6 mL of CDCl and 0.1 mL of tert-butanol. The mixture became
2
d)] were detected but they were impossible to be quantified.
5% with respect to singlet at 1.35 (CMe
were observed.
3
of 3d). No other signals
1
4 3
)
(0.06 mmol) were dissolved in
0
3
Decomposition Test of 2. 2 (0.5 mmol) and 1,4-di-tert-
butylbenzene (0.5 mmol) were dissolved in 0.6 mL of CD Cl and
clear, and 2 (0.73 mmol) was added. A spectrum was recorded
immediately after and then every 15 min. In the first time of the
reaction (within 2 h), the kinetic profile was very similar to that of
2
2
1
submitted to H-NMR. The solution was poured into a two-necked
flask equipped with a magnetic stirring bar and a condenser coil
Sc(OTf)
3
.
4 2
containing Ba(ClO ) (0.022 mmol). The mixture was stirred at
1
Ether Decomposition Experiments. In a two-necked flask
equipped with a magnetic stirring bar and a condenser coil, ether
reflux for 7 h and then resubmitted to H-NMR. The intensity of
signal at δ ) 1.53 decreased of about 9-10%.
In a two-necked flask equipped with a magnetic stirring bar and
3a (0.5 mmol) was dissolved in 1 mL of dichloromethane and Mg-
(
ClO or (0.05 mol) was added. The mixture was stirred at 40 °C
4
)
2
a condenser coil, Mg(ClO
were dissolved in 1 mL of CH
mixture was stirred at reflux until the TLC analysis revealed the
4
)
2
(0.043 mmol) and anisole (1 mmol)
in a thermostatic bath for 5 h. The crude reaction mixture was
worked up as usually and submitted to H-NMR analysis. Signal
at δ ) 3.61 (CH
δ ) 3.31 (CH O of 3a).
2
2
Cl . 2 (1 mmol) was added. The
1
1
2
O of 1a) was less than 3% with respect to that at
2
presence of Boc O and then after usual workup submitted to H-
2
NMR. Neither significant modification of the aromatic part of the
spectrum nor signals assignable to o- and p-tert-butyl frameworks
were observed.
In a two-necked flask equipped with a magnetic stirring bar and
a condenser coil, ether 3d (0.5 mmol) was dissolved in 1 mL of
dichloromethane and Al(ClO
was stirred at room temperature for 24 h. The crude reaction mixture
was worked up as usually and submitted to H-NMR analysis. The
4
)
3
(0.1 mmol) was added. The mixture
Reaction of 21. Mixed dicarbonate 21 was prepared according
to the reported procedure11 from 1j and 2 in the presence of DMAP.
The crude mixture of mixed dicarbonate (21), mixed carbonate (19),
and symmetric carbonate (20) was dissolved in 1.5 mL of CH Cl ,
1
crude reaction mixture was worked up as usually and submitted to
2
2
1
H-NMR analysis. The following product ratio was obtained using
4 2
and then Mg(ClO ) (0.1 mmol/mmol of 21) was added and the
the whole aromatic region as standard (5H): 1d [69% based on
multiplet at 6.8-6.9 (3H)]; 7 [11% based on singlet at 1.28 (9H)];
mixture was stirred at room temperature. After 18 h (instead of 30
h required by the direct reaction of 1j and 2), 21 disappeared and
1
8
[10% based on singlet at 1.41 (9H)]; 9 [10% based on singlets at
the crude reaction mixture was diluted with water and extracted
with CH Cl . The organic layer was separated, dried over MgSO ,
1.30 and 1.42 (9H)].
2
2
4
In a two-necked flask equipped with a magnetic stirring bar and
and filtered, and the solvent was removed by rotary evaporation.
The crude mixture was then submitted to H-NMR analysis, which
revealed the presence of 3j together with unreacted mixed carbonate
1
a condenser coil, ether 3d (0.5 mmol) was dissolved in 1 mL of
dichloromethane and Mg(ClO (0.05 mmol) was added. The
4
)
2
mixture was stirred at 40 °C in a thermostatic bath for 5 h. The
and symmetric carbonate.
crude reaction mixture was worked up as usually and submitted to
1
H-NMR analysis. The following product ratio was obtained using
Acknowledgment. Work was carried out in the framework
of the National Project “Stereoselezione in Sintesi Organica.
Metodologie e Applicazioni” supported by MIUR, Rome, by
the University of Bologna, in the framework of “Progetto di
Finanziamento Pluriennale, Ateneo di Bologna”, and by National
Project FIRB “Progettazione, preparazione e valutazione bio-
logica e farmacologica di nuove molecole organiche quali
potenziali farmaci”.
the whole aromatic region as standard (5H): 1d [79% based on
multiplet at 6.8-6.9 (3H)]; 3d [16% based on singlet at 1.35 (9H)];
7
[1% based on singlet at 1.28 (9H)]; 8 [4% based on singlet at
.41 (9H)].
In a two-necked flask equipped with a magnetic stirring bar and
1
a condenser coil, ether 3d (0.5 mmol) was dissolved in 1 mL of
dichloromethane and Sc(OTf) (0.05 mmol) was added. The mixture
was stirred at 40 °C in a thermostatic bath for 5 h. The crude
3
1
reaction mixture was worked up as usually and submitted to H-
JO061402D
9588 J. Org. Chem., Vol. 71, No. 26, 2006