The reactivity of o-hydroxybenzyl alcohol and derivatives in solution at elevated temperatures

Article abstract of DOI:10.1021/jo981110f

The reactivity of o-hydroxybenzyl alcohol (o-HBA, 1), as a model compound for lignin, has been studied in various solvents between 390 and 560 K. Both in polar and apolar solvents the benzylic cation is the reactive intermediate. In alcoholic solvents, the benzylic cation reacts with the solvent to give the corresponding ethers. Relative reaction rates have been determined for different alcohols; a factor of 14 is encountered between the most (methanol) and least (tert-butyl alcohol) reactive ones. The etherification is reversible, in contrast to the electrophilic aromatic substitution with phenol and anisole, for which k(PhOH) = 1 X 10⁵ M^-1 s^{- 1} and k(anisole) = 1 x 10⁴ M^-1 s^-1, at 424 K. In apolar hydroaromatic solvents, 7H-benz[de]anthracene, 9,10-dihydroanthracene, and 9,10- dihydrophenanthrene, the formation of o-cresol proceeds via hydride transfer from the solvent to the benzylic cation; rate constants at 555 K are 2 x 10⁶, 5 x 10⁴, and 5 x 10³ M^-1 s^-1, respectively.

Full text of DOI:10.1021/jo981110f

3012
J . Org. Chem. 1999, 64, 3012-3018
Th e Rea ctivity of o-Hyd r oxyben zyl Alcoh ol a n d Der iva tives in
Solu tion a t Eleva ted Tem p er a tu r es
Edwin Dorrestijn, Marieke Kranenburg, Maria Victoria Ciriano, and Peter Mulder*
Leiden Institute of Chemistry, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
Received J une 10, 1998
The reactivity of o-hydroxybenzyl alcohol (o-HBA, 1), as a model compound for lignin, has been
studied in various solvents between 390 and 560 K. Both in polar and apolar solvents the benzylic
cation is the reactive intermediate. In alcoholic solvents, the benzylic cation reacts with the solvent
to give the corresponding ethers. Relative reaction rates have been determined for different alcohols;
a factor of 14 is encountered between the most (methanol) and least (tert-butyl alcohol) reactive
ones. The etherification is reversible, in contrast to the electrophilic aromatic substitution with
phenol and anisole, for which kPhOH ) 1 × 105 M-1 s-1 and kanisole ) 1 × 104 M-1 s-1, at 424 K. In
apolar hydroaromatic solvents, 7H-benz[de]anthracene, 9,10-dihydroanthracene, and 9,10-dihy-
drophenanthrene, the formation of o-cresol proceeds via hydride transfer from the solvent to the
benzylic cation; rate constants at 555 K are 2 × 106, 5 × 104, and 5 × 103 M-1 s-1, respectively.
In tr od u ction
cessing o-methoxyphenolic units are partially converted
into o-HBA derivatives.6
During the past decades, the demand for alternative
sources of chemical feedstock has increased due to a
growing awareness of the depletion of crude oil reserves.
The utilization of woody biomass as a renewable source
is therefore an appealing option. Lignin, for example, may
be converted into (substituted) phenols. Liquid-phase
hydrocracking, in the presence of hydroaromatic solvents
such as tetralin, seems to be a promising route for the
degradation of lignin into monoaromatic products.1 Fur-
thermore, alcohols, 1,4-dioxane, and water can be used
in processing lignin or model compounds.2
Quinone methides (QMs) have been suggested as
important intermediates in the photochemical and ther-
mal chemistry of lignin.3 However, only a limited number
of studies detail the dynamic behavior of o-quinone
methide (o-QM). o-Hydroxybenzyl alcohol (o-HBA, 1) has
been reported as a precursor for o-QM4 and can also be
considered as a model compound for structural elements
in lignin: 5% of the aromatic units in lignin contain a
benzylic hydroxyl moiety.5 Moreover, during lignin pro-
In the photosolvolysis of o-HBA, under alkaline and
neutral conditions, Wan et al.7 have proposed the forma-
tion of o-QM via intramolecular proton transfer. In the
presence of methanol, o-QM is converted into the corre-
sponding methyl ether (2). The existence of o-QM has
been further substantiated on the basis of product studies
in the presence of nucleophiles or dienophiles. However,
the observed products can be equally well explained by
an ionic pathway in which the benzylic cation is the
reactive intermediate. A nanosecond laser flash photoly-
sis study of a neutral aqueous solution of o-HBA and
derivatives has revealed a long-lived species that, ac-
cording to its lifetime (>5 s) and the absorption spectrum,
was identified as o-QM. If the intermediate were a
benzylic cation, the lifetime would have been less than 1
µs.8 During photocondensation of o-HBA in alkaline
media the formation of oligomers has been reported,9
which may or may not proceed via o-QM. Similar oligo-
mers have been observed in the acid-catalyzed formation
of Bakelite resins; in that case the benzylic cation acts
as the intermediate.10
* To whom correspondence should be addressed. Fax: (31)-71-
5274492. E-mail: P.Mulder@Chem.LeidenUniv.NL.
(1) For a review: Dorrestijn, E.; Laarhoven, L. J . J .; Arends, I. W.
C. E.; Mulder, P. J . Anal. Appl. Pyrolysis, in press.
(2) E.g.: Sola´r, R.; Kacik, F. Holz Roh. Werkst. 1995, 53, 123-128.
Destine´, J . N.; Wang, J .; Heitner, C.; Manley, R. S. J . J . Pulp Paper
Sci. 1996, 22, J 24-J 30. Pardini, V. L.; Vargas, R. R.; Viertler, H.
Tetrahedron 1992, 48, 7221-7228. Ni, Y.; Hu, Q. J . Appl. Polym. Sci.
1995, 57, 1441-1446. Kleinert, T. Monatsh. Chem. 1952, 82, 623-
628. Wang, M.; Smith, J . M.; McCoy, B. J . Energy Fuels 1993, 7,
78-84. Sakai, K.; Takagi, T.; Imamura, H. Mok. Gak. 1990, 36, 553-
558.
Previously, we have identified o-QM directly with
on-line mass spectrometric analysis in the gas phase
at low pressure by thermal treatment of o-HBA
and chroman.4a At atmospheric pressures, this inter-
mediate appears to be extremely reactive and rapid
oligomerization occurs, even under dilute conditions.4b
In the presence of toluene as a hydrogen source, o-QM
(3) For a review: Shevchenko, S. M.; Apushkinskii, A. G. Russ.
Chem. Rev. 1992, 61, 105-131.
(6) Dorrestijn, E.; Mulder, P. J . Chem. Soc., Perkin 2, 777-780.
(7) Wan, P.; Chak, B. J . Chem. Soc., Perkin Trans. 2 1986, 1751-
1756.
(8) Diao, L.; Yang, C.; Wan, P. J . Am. Chem. Soc. 1995, 117, 5369-
5370.
(9) Wan, P.; Hennig, D. J . Chem. Soc., Chem. Commun. 1987, 939-
941.
(10) E.g.: Sprung, M. M.; Gladstone, M. R. J . Am. Chem. Soc. 1949,
71, 2907-2913. Forrester, A. R.; Wardell, J . L. In Chemistry of Carbon
Compounds, vol. IIIA, 2nd ed.; Elsevier: Amsterdam, 1971; pp 304-
305. Natesan, R.; Yeddanapalli, L. M. Ind. J . Chem. 1974, 12, 691-
694.
(4) (a) Dorrestijn, E.; Epema, O. J .; Van Scheppingen, W. B.; Mulder,
P. J . Chem. Soc., Perkin Trans. 2 1998, 1173-1178. (b) Dorrestijn, E.;
Pugin, R.; Ciriano Nogales, M. V.; Mulder, P. J . Org. Chem. 1997, 62,
4804-4810. (c) Eck, V.; Schweig, A.; Vermeer, H. Tetrahedron Lett.
1978, 27, 2433-2436. (d) McIntosh, C. L.; Chapman, O. L. J . Chem.
Soc., Chem. Commun. 1996, 771. (e) Mao, Y. L.; Boekelheide, V. Proc.
Natl. Acad. Sci. U.S.A. 1980, 77, 1732-1735.
(5) Adler, E.; Hernestam, S. Acta Chem. Scand. 1955, 9, 319-334.
Hon, D. N. S., Ed. Chemical Modification of Lignocellulosic Materials;
Marcel Dekker: New York, 1996.
10.1021/jo981110f CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/03/1999

Reactivity of o-Hydroxybenzyl Alcohol and Derivatives
J . Org. Chem., Vol. 64, No. 9, 1999 3013
Ch a r t 1
can be transformed into o-cresol in two steps: a reverse
radical disproportionation mechanism (RRD, see eq 1)
followed by hydrogen abstraction.4b Hence, the o-cresol
formation can be directly related to the presence of
o-QM.
Exp er im en ta l Section
The liquid-phase experiments were carried out in sealed
Pyrex tubes (ca. 1.5 mL). After the tube was filled approxi-
mately one-third with the reagents and an internal standard
(naphthalene), the air was removed by three freeze-pump-
helium-thaw cycles. Subsequently, the tube was sealed under
vacuum (ca. 400 Pa) and placed in a thermostat-controlled
oven. The temperatures applied were reached within 5 min.
After the thermal treatment the reaction chamber was cooled
to room temperature (in about 5 min), and toluene or acetone
and an external standard (n-dodecane) were added. For calcu-
lation of the initial concentration, the amount of o-HBA weigh-
ed into the tube was corrected for the amount of GC-identifi-
able impurities (5%) consisting of o-cresol and o-hydroxyben-
zaldehyde. However, during the course of the experiments it
was noticed that the results varied depending on the purity
of o-HBA. These differences could not be explained by the pres-
ence of o-cresol and o-hydroxybenzaldehyde but may arise due
to small amounts of an acid, most likely o-hydroxybenzoic acid.
However, proper detection by GC analysis was not possible.11
(1)
This report deals with the behavior of o-HBA, as a
lignin model compound, under thermal liquefaction
conditions. For comparison, the reactivities of some
benzylic alcohol derivatives have been explored as well.
Hydroaromatic and alcoholic solvents have been used, in
a temperature range of 390-560 K, to investigate whether
quinone methide chemistry is relevant under thermal,
solution-phase conditions (Chart 1).
The samples were analyzed in quadruplicate on an HP 5890
GC (column: CPsil 5 CB) equipped with hydrogen as the

3014 J . Org. Chem., Vol. 64, No. 9, 1999
Dorrestijn et al.
Ta ble 1. P r od u ct Distr ibu tion (%)a of Rea ction s w ith
carrier gas and an FID detector. For identification of unknown
products an HP 5972 mass selective detector was used.
Response factors for available compounds were calculated by
injecting mixtures of known composition. For compounds not
available, response factors were estimated on the basis of the
number of carbon and oxygen atoms and related to similar,
available chemicals. 1H NMR spectra of 200 scans in acetone-
d6 were performed on a 300 MHz DPX300 apparatus from
Bruker. IR spectra were measured on a Unicam SP3-200
double beam infrared spectrometer.
Ch em ica ls. o-(Methoxymethyl)phenol (o-MMP, 2) was pre-
pared according to a synthesis reported previously12 and
further purified by aqueous phase extraction (final purity
96%). All other reagents were obtained in the purest grade
and used as received, except for 9,10-dihydroanthracene, which
was crystallized twice from methanol (GC purity >99%).
o-HBA was used as such, and a purified batch (>99%) was
prepared as well by means of sublimation or recrystallization
from a 1/9 toluene/n-hexane mixture.
o-HBA (0.22 M) in Va r iou s Solven ts a t 424 K
solvent
methanol
ethanol
2-propanol
o-HBAb
etherc
o-cresolc
otherc,d
5
21
35
47
29
18
62
33
82
68
40
94
78
65
52
62
82
< 1
0
0
0
< 1
0
0
0
0
0
0
< 1
0
< 1
0
tert-butyl alcohol
cyclohexanol
benzyl alcohol
diphenyl ether
1,4-dioxane
acetonitrile
0
8e
< 1
1
f
AnH2
2
0
AnH2/BzH 10/1f
2
a
Average of four analyses, accuracy ( 1%, reaction time 1 h.
Recovered substrate: (100%{[o-HBA]t/[o-HBA]t)0}). c Detected
b
d
products: (100%{[product]t/[o-HBA]t)0}). High boiling products.
Based on GC-MS analysis, 60-100% of these products was the
dimer 14. No other products could be identified. e Of which 7%
aromatic substitution products (ortho/para ) 1/6). f Initial o-HBA
concentration 0.07 M.
Resu lts
o-HBA in Hyd r oa r om a tic Solven ts. The chemistry
of o-hydroxybenzyl alcohol (o-HBA, 1, 0.07 M) was
studied in mixtures of 9,10-dihydroanthracene (AnH2, 3.3
M) and 7H-benz[de]anthracene (BzH, 0.3 M), where BzH
acts as the hydrogen donor and is regenerated via
hydrogen abstraction from AnH2 by the Bz• radical.1 The
conversion of o-HBA was complete after 1 h at temper-
atures above 475 K. o-Cresol emerged as a product and
the yield gradually increased with temperature, up to a
maximum of 88% at 555 K. In these experiments no
conversion of BzH was observed and about equimolar
amounts of anthracene were formed with respect to
o-cresol. At 555 K, the o-cresol yield was independent of
the reaction time (0.75-2 h); this reaction product was
shown to be stable under the applied conditions. By
diluting the hydrogen-donor mixture with (inert) diphe-
nyl ether (DPE), it was found that the o-cresol yield
increased linearly with the BzH concentration.
in a mixture of BzH (1.3 M) and AnH2 (3.3 M). Thus,
relative rate constants for the o-cresol formation at 555
K can be derived as kPhenH2/kAnH2/kBzH ) 1/10/340.14 With
purified o-HBA similar results were obtained, although
a somewhat lower reactivity at temperatures below 500
K was noticed accompanied by a somewhat higher
selectivity toward o-cresol.
o-HBA in Alcoh olic Solven ts. In alcoholic solvents
(methanol, ethanol, 2-propanol, tert-butyl alcohol, and
benzyl alcohol) at relatively low temperatures, o-HBA
was converted with a high selectivity (e.g., 95% in
methanol after 1 h at 424 K) into the corresponding
ethers (2, 4, 5, 6, and 7, see Table 1), according to eq 2.
(2)
Particularly at low reaction temperatures, an incom-
plete mass balance was obtained. No products other than
o-cresol could be detected besides traces (e2%) of a high-
boiling compound. On the basis of GC and GC-MS
analysis, this material was identified as a (condensation)
dimer of o-HBA (14). Presumably, the dimer is indicative
for the formation of other oligomers (which could not be
detected) and may account for the low mass balance. With
o-HBA heated (for 1 h at 474 K) in inert solvents
(diphenyl ether and 1,4-dioxane) a precipitate could be
isolated and characterized as diphenylmethane deriva-
tives.13 When the oligomer was heated in 9,10-dihydroan-
thracene at 555 K for 1 h, only coalification took place.
The nature of the applied hydroaromatic solvent was
varied: at 555 K, starting with 0.18 M o-HBA, the
o-cresol yield amounted to 5% in 9,10-dihydrophenan-
threne (PhenH2, 4.8 M), 34% in AnH2 (4.7 M), and 84%
Only above 500 K did the total mass balance decrease.
The o-HBA purity did not affect the reaction rates or
selectivities. Two other benzylic alcohols that were ap-
plied under the same conditions, benzyl alcohol and
o-methoxybenzyl alcohol (o-MBA, 3), appeared to be
almost inert.
With o-(methoxymethyl)phenol (o-MMP, 2) dissolved
in 2-propanol or tert-butyl alcohol, an ether/alcohol
interchange was observed. Hence, under the applied
conditions the ether formation is reversible. The conver-
sion of o-MMP, another precursor for o-QM,15 in 2-pro-
panol at 424 K (22%) is lower in comparison with that of
(13) The precipitates from diphenyl ether and 1,4-dioxane were
analyzed by 1H NMR and IR spectroscopy and shown to be identical.
In the IR spectrum absorptions were visible at 3100-3700 (s), 2900
(11) On the basis of the response factor of separately synthesized
o-MMP and after correction for the observed amounts of o-cresol and
o-hydroxybenzaldehyde as impurities in o-HBA, a complete mass
balance was obtained in the experiments with o-HBA in methanol.
Hence, the amount of acid impurities in o-HBA, most likely o-
hydroxybenzoic acid, is less than 1%. Further quantification is not
possible in view of the strong interaction of benzoic acids with the
applied GC column. It should be noted that the presence of small
amounts of acid may well be important for the overall solution’s acidity.
For with 0.22 M o-HBA (together with 1% of o-hydroxybenzoic acid)
dissolved in phenol (10 M) and using the pKa values of around 4 and
10, respectively, the acidity is still dictated by the organic acid.
(12) De J onge, J .; Bibo, B. H. Rec. Trav. Chim. Pays-Bas 1955, 74,
1448-1452.
(w), 1600 (w), 1500 (s), 1450 (m), 1250 (s), 1100 (w), and 750 (w) cm-1
.
The 1H NMR spectrum showed two broad signals at δ ) 3.6-4.0 and
6.6-7.3 ppm, which were assigned to the methylene and aromatic
protons, respectively, of the oligomeric structure. In addition, three
broad phenolic resonances were observed in the region 7.9-8.6 ppm.
Because of the absence of the characteristic carbonyl frequencies (e.g.,
strong band at 1650-1720 cm-1) in the IR spectrum, an oligomer of
o-QM could be excluded. Both the IR and 1H NMR spectra are in good
accordance with those reported by Wan from Bakelite resins produced
in the photocondensation of o-HBA in an alkaline medium.9
(14) By using [o-cresol]t)1h/[oligomer]t)1h equals (kPhenH2 × [PhenH2]t)0)/
(rolig) in PhenH2, (kAnH2 × [AnH2]t)0)/(rolig) in AnH2, and {(kBzH
×
[BzH]t)0) + (kAnH2 × [AnH2]t)0)}/(rolig) in BzH and assuming equal rates
for the oligomerization (rolig).

Reactivity of o-Hydroxybenzyl Alcohol and Derivatives
J . Org. Chem., Vol. 64, No. 9, 1999 3015
Ta ble 2. P r od u ct Distr ibu tion (%)a of Rea ction s w ith
o-HBA (0.22M) in (Mixtu r es of) P h en ol a n d An isole a t
424 K
o-HBA (61%). Apparently, the elimination of methanol
is somewhat slower than that of water.
From Table 1, a distinct difference in the rate of o-HBA
conversion at 424 K can be noticed when applying
different alcoholic solvents: methanol > benzyl alcohol
≈ ethanol > cyclohexanol > 2-propanol > tert-butyl
o-
HBAc ether 2,2′-HMD 2,4′-HMD otherd,e
2,2′-DHDd 2,4′-DHDd
solventb
phenol
anisole
2-propanol/phenol 1/1
2-propanol/anisole 1/1
anisole/phenol 1/1
<1
68
5
56
23
50
3
10
<1
34
1
31
10
7
<1
17
3
4
<1
2
<1
8
alcohol. To determine the rate of decomposition (kdecomp
)
70
42
the decay of o-HBA in 2-propanol was followed in time (t
) 0-90 min) at this temperature, using an initial
substrate concentration of 0.20 M. With a forced mass
balance consisting of the remaining o-HBA and formed
o-(2-propoxymethyl)phenol (o-PMP, 5) ([o-HBA]t)0 ) [o-
HBA]t + [o-PMP]t, which was close to the actual mass
balance) a rate constant of kdecomp ) 4.0 × 10-4 s-1 was
obtained from the linear plot (r2 ) 0.99) of -ln([HBA]t/
[HBA]t)0) vs reaction time. At a lower o-HBA concentra-
tion (0.005 M), kdecomp ) 3.5 × 10-4 s-1 (r2 ) 0.97) was
found, indicating first-order behavior.
a
Average of four analyses, accuracy ( 1%, reaction time 1 h.
Solvent ratio in M/M. c Recovered o-HBA: (100%{[o-HBA]t/[o-
b
d
HBA]t)0}). Detected products: (100%{[product]t/[o-HBA]t)0}).
e High boiling product, mainly 14.
toward EVE can be regarded as upper limits, since the
ether formation is already reversible under these condi-
tions. For comparison, kEVE/kMeOH is around 0.1 for
(diphenylmethyl) cations at 298 K.17
The relative rate constants for ether formation with
the various alcohols (kROH), derived from competition
o-HBA in P h en ol a n d An isole. To explore the option
of the benzylic cation as intermediate, phenol was used
as a solvent (Table 2), being a very potent trap for
carbocations. When o-HBA was heated in neat phenol,
two major products were observed with high selectivi-
ties: 2,2′- and 2,4′-dihydroxydiphenylmethane (2,2′- and
2,4′-DHD, 11 and 15).20 The same products (11, 15) were
obtained with o-MMP, whereas o-methoxybenzyl alcohol
(o-MBA, 3) was almost inert. Ether formation between
o-HBA and phenol did not take place. The products can
only be explained to arise from electrophilic aromatic
substitution of the benzylic cation at the ortho and para
positions of the aromatic ring. The third isomer, stem-
ming from addition to the meta positions, could not be
detected (detection limit: 0.5%). Very low yields of the
meta isomer in the electrophilic substitution with acti-
vated monoaromatics have been reported before.21 In
phenol/alcohol mixtures, the ratio DHD/ether increased
both with time and temperature. The formation of the
ether from o-HBA with the alcohol is kinetically favored
but reversible, in contrast to the aromatic substitution.
The ortho/para ratio varied with temperature and cosol-
vent between 1 and 3. The higher preference for the ortho
substitution has been documented quite well; the differ-
ence in isomer ratio between neat phenol and phenol
mixed with another solvent may be caused by the
variation in the hydrogen bonding of phenol.21,22
experiments, revealed kMeOH/kEtOH/kbenzOH/k2-PrOH/kt-BuOH
)
14/8/5/4/1, independent of reaction time (1-4 h) and
temperature (394-475 K). Ether formation can be envis-
aged by either o-QM or the benzylic cation as the
intermediate reacting with the alcoholic solvent. The
ratios in rate constants as we observed are quite similar
to those found for the addition of the diphenylmethyl
cation.16
In tert-butyl alcohol, the formation of (2,2-dimethyl-
chroman, 9) was observed as an additional product at
475 K. It is known that isobutene can be formed from
tert-butyl alcohol at elevated temperatures in a slightly
acidic environment.18 Indeed, when the tube was opened,
evolution of a volatile compound was visible. The mech-
anism of formation can be rationalized by a Diels-Alder
addition of o-QM or the benzylic cation to isobutene. In
alcohol mixtures including tert-butyl alcohol, the yield
of 2,2-dimethylchroman (9) increased with time,19 at the
expense of the product ethers, although their ratios
remained constant. Clearly, the formation of 9 is kineti-
cally irreversible, in contrast to the ether formation. This
also implies that both processes start with one common
intermediate, and thus, ether formation through a direct
SN2 mechanism can be excluded.
Ethyl vinyl ether (EVE), a commonly used trap for
o-QM8 and benzylic cations,17 was also applied to inter-
cept the intermediate in the o-HBA thermolysis. In a
mixture of EVE (0.3 M), methanol (3.0 M), and benzyl
alcohol (4.8 M) at 424 K the o-HBA conversion was 77%
(see Table 1). Ether derivatives from methanol (2) and
benzyl alcohol (7) as well as 2-ethoxychroman (10) were
found as products in a ratio of 40/28/9. At 475 K o-HBA
is quantitatively converted into 10. On the basis of the
No significant changes were observed when purified
o-HBA was applied, albeit that the ortho/para ratio
increased to about 4.
Rates of aromatic substitution relative to ether forma-
tion were determined with o-HBA in mixtures of phenol
and 2-propanol. By extrapolation to t ) 0 of the plot of
the ratio DHD/ether at 424 K vs time (which gave a linear
relation), the ratio of kPhOH/k2-PrOH could be retrieved.
Using the rate constant for addition of the benzylic cation
to 2-propanol, k2-PrOH ) 5 × 106 M-1 s-1 at 424 K,23 the
product yields at 424 K it can be approximated that kEVE
kMeOH ) 2 and kEVE/kbenzOH ) 5. These relative reactivities
/
(15) (a) Gardner, P. D.; Sarrafizadeh R. H.; Brandon, R. L. J . Am.
Chem. Soc. 1959, 81, 5515. (b) Cavitt, S. B.; Sarrafizadeh R. H.;
Gardner, P. D. J . Org. Chem. 1962, 27, 1211-1216.
(16) Ratios kMeOH/kEtOH/k2-PrOH/kt-BuOH ) 17/12/4/1 were found at 293
K in acetonitrile.17
(17) Bartl, J .; Steenken, S.; Mayr, H. J . Am. Chem. Soc. 1991, 113,
7710-7716.
(18) McMurry, J . Organic Chemistry, 2nd ed.; Brooks/Cole: Pacific
Grove, 1988; p 596. o-HBA or acidic impurities can serve as the acid
catalyst.
(20) In phenol, 2,2′- and 2,4′-DHD (11 and 15) are identified by
comparison with spectra in the NIST GC-MS library. In anisole, 2,2′-
and 2,4′-HMD (12 and 16) are identified based on GC (retention times)
and GC-MS analysis.
(21) Harvey, D. R.; Norman, R. O. C. J . Chem. Soc. 1961, 3604-
3610.
(22) Watson, W. D. J . Org. Chem. 1974, 39, 1160-1164.
(23) The reaction rate of the benzylic cation with an alcohol is about
24
(19) Remarkably, next to 2,2-dimethylchroman (9), another, yet
unknown product, was observed when tert-butyl alcohol was used as
a solvent together with methanol, 2-propanol, and benzyl alcohol.
105 M-1 s-1 at 295 K,8, and by applying a preexponential factor of
109 M-1 s-1, an activational energy of 23 kJ mol-1 is obtained. Thus,
at 424 K, kROH ) 5 × 106 M-1 s-1
.

3016 J . Org. Chem., Vol. 64, No. 9, 1999
Dorrestijn et al.
rate constant for the electrophilic aromatic substitution
products in various solvents: ethers (alcohols), aromatic
substitution compounds (phenol, anisol), chroman deriva-
tives (isobutene, ethyl vinyl ether), and o-cresol (hy-
droaromatic solvents) no real distinction can be made
between the two routes. Moreover, the benzylic cation
may also arise through rapid protonation of o-QM. From
the competition experiments with various types of sol-
vents it is, however, clear that the same intermediate is
present in all cases.
In anisole a strong dependence of the conversion and
product spectrum exists on the solution’s acidity. In the
absence of protons, aromatic substitution does not occur,
indicating that o-QM needs to be protonated first before
such a reaction can take place. The main pathway for
o-QM is now oligomerization. Our results show that
phenol is a strong enough acid to protonate o-QM to
produce the prerequisite intermediate for the aromatic
substitution: the benzylic cation. On the other hand,
o-MBA (due to the methyl group a direct elimination can
be excluded) is almost inert in phenol as the solvent,
indicating that the acidity is still too low to render the
benzylic cation. Thus, the initial step in the thermal
degradation of o-HBA seems to be direct formation of
o-QM.
Conversely, the constant ratio of ether products from
o-HBA in a mixture of alcohols over a range of 80 K may
be an indication for the direct protonation mechanism
for o-HBA decomposition at least in polar solvents. The
variation in addition rates to different alcohols, which
are quite similar to those advanced for benzylic cation
reactivity,17 may be explained by a steric effect:7 with tert-
butyl alcohol the lowest addition rate is found. Since this
is an entropic contribution and probably only small
differences in reaction enthalpies exist for the different
alcohols, the individual rate constants will only slightly
vary with the temperature. Moreover, at our reaction
temperatures the ether formation is reversible. Therefore,
the observation that the product spectrum does not
change with time and temperature indicates that the
ether decomposition also proceeds via an ionic mecha-
nism. Otherwise, the rates for direct, concerted alcohol
elimination from the ethers need to be almost identical
to accommodate our observations, which seems unreal-
istic.
with phenol was obtained: kPhOH ) 1 × 105 M-1 s-1
.
After 1 h at 475 K, o-MBA in a mixture of phenol and
2-propanol appeared to be inert. However, after addition
of an equimolar amount of nonpurified o-HBA, o-MBA
was converted for 23% with a 5% yield of 8 and 15% of
the substitution products 13 and 17 (ortho/para ) 1/1).
Simultaneously, o-HBA was completely converted into 5
(10%), 11 (62%), and 15 (24%). Apparently, the phenol
acidity is not sufficient enough to render a benzylic cation
from o-MBA, which demonstrated that a (stronger)
organic acid is present in nonpurified o-HBA.11
With unpurified o-HBA in anisol, similar substitution
products were found: 2-hydroxy-2′-methoxydiphenyl-
methane and 2-hydroxy-4′-methoxydiphenylmethane (2,2′-
and 2,4′-HMD, 12 and 16).20 The ortho/para ratio of ca.
0.32 in anisole tallies nicely with the ratio found for other
electrophilic substitution reactions21,25 and was indepen-
dent of temperature and (alcoholic) cosolvents. By using
a mixture of phenol and anisole, the former was found
to be 10 times more reactive, and thus, kanisole ) 1 × 104
M-1 s-1 at 424 K. Surprisingly, in anisole the results did
depend on the purity of the starting material. Upon
purification the conversion decreased and only traces of
aromatic substitution products were obtained. Oligomer-
ization appears to be the major reaction pathway.
The relative reactivity of the cation toward either BzH
and anisole was further investigated.26 The product ratio
o-cresol/HMD at 555 K linearly increased with the ratio
BzH/anisole. This clearly indicates that both o-cresol and
the aromatic substitution products are formed from the
same precursor, the benzylic cation. From the slope, the
ratio in rate constants kBzH/kanisole ) 11 (r2 ) 0.997) was
derived.27
Discu ssion
o-HBA Degr a d a tion . Two mechanisms can be ad-
vanced for the conversion of o-HBA in solution at elevated
temperatures (Scheme 1): thermal, concerted elimination
of water to yield o-QM and an acid-catalyzed dehydration
to form a benzylic cation. From the obtained reaction
Sch em e 1
Thus, no decisive conclusion can be drawn from the
current experiments as to the mechanism of the o-HBA
thermolysis in the liquid phase, but it is certain that in
all cases the benzylic cation acts as the reactive inter-
mediate for product formation.
The rate of decomposition of o-HBA depends on the
type of solvents, as can be seen in Table 1. An explanation
can be found in the different degrees of hydrogen bond
accepting capacities of the solvents. To semiquantify the
role of intermolecular hydrogen bonding we used the
solvent parameter â2H, a commonly used parameter to
quantify the contribution of hydrogen bonding.28 In
Figure 1, the plot is displayed of log(kdecomp) as a function
(24) Richard, J . P.; J encks, W. P. J . Am. Chem. Soc. 1984, 106,
1373-1383.
H
of the solvent â2 for various solvents. For the alcoholic
solvents, a fair correlation is obtained, not only for the
neat solvents (including phenol) but also for mixtures
(25) Norman, R. O. C.; Radda, G. K. J . Chem. Soc. 1961, 3610-
3616.
(26) In these experiments, the mass balance amounted to 43-47%
of the initial o-HBA concentration. In all experiments ca. 2% of the
dimer of o-HBA was found and ca. 3% of an unknown, high-boiling
product.
(28) (a) MacFaul, P. A.; Ingold, K. U.; Lusztyk, J . J . Org. Chem.
1996, 61, 1316-1321. (b) Valgimigli, L.; Banks, J . T.; Ingold, K. U.;
Lusztyk, J . J . Am. Chem. Soc. 1995, 117, 9966-9971. (c) Abraham,
M. H.; Lieb, W. R.; Franks, N. P. J . Pharm. Sci. 1991, 80, 719-724.
(d) Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J . J .; Taylor,
P. J . J . Chem. Soc., Perkin Trans. 2 1990, 521-529.
(27) Assuming the same temperature dependence as in ref 23, kanisole
) 2 × 105 M-1 s-1 at 555 K; thus, kBzH ) 2 × 106 M-1 s-1. Consequently,
from kPhenH2/kAnH2/kBzH ) 1/10/340 (see text): kAnH2 ) 5 × 104 and
kPhenH2 ) 5 × 103 M-1 s-1
.

Reactivity of o-Hydroxybenzyl Alcohol and Derivatives
J . Org. Chem., Vol. 64, No. 9, 1999 3017
H
F igu r e 1. Correlation between log(kdecomp) and â2 for o-hydroxybenzyl alcohol conversion at 424 K in neat alcoholic solvents ([),
mixtures of alcohols (O), acetonitrile (9), 1,4-dioxane (b), anisole (0), and diphenyl ether (2). Values for â2H for the alcohols from
ref 28c, for the other compounds from ref 28d.
Sch em e 2
F igu r e 2. Conformations of o-hydroxybenzyl alcohol.
of these solvents, for which we used linearly averaged
â2H values. In the non-hydrogen-bonding solvent, carbon
tetrachloride, three different intramolecular-hydrogen-
bonded conformers of o-HBA have been identified by
infrared spectroscopy at room temperature. The conform-
ers A and B predominate (see Figure 2),29 while type C
only exists in small amounts. Apparently, when o-HBA
is less intermolecularly bonded the elimination of water,
most likely from the intramolecularly bonded conformer
A, proceeds faster.
is not feasible and BzH is recycled by hydride abstraction
from AnH2.
The derived rate constants for hydride abstraction by
the benzylic cation (e.g., kAnH2 ) 5 × 104 M-1 s-1) are in
range with those published before for reaction of carbon
cations with hydride donating compounds.31 The forma-
tion of o-cresol by hydride transfer can be compared with
the hydrogen atom transfer from the same donor solvents
to the benzyl radical. In both cases, the reaction rate
constants increase in the same order, PhenH2 < AnH2 <
BzH, and the derived ratios in rate constants are quite
similar. Surprisingly, the absolute rate constants for
hydride transfer27 are larger than those for hydrogen
atom transfer.32 Thus, on the basis of dynamic studies
only, it can be difficult to distinguish between both
processes. More in general, the analogy between an ionic
and radical pathway needs to be treated with great
caution.35
o-Cr esol F or m a tion . In hydroaromatic solvents, a
quantitative conversion of o-QM into o-cresol can be
expected. The RRD of o-QM with the used hydrogen
donors is sufficiently fast4b relative to other, nonionic
processes. However, the o-cresol yield depends on both
the hydrogen-donating capability of the solvent and on
the hydrogen-donor concentration. Instead of the RRD
mechanism, the formation of o-cresol involves the shut-
tling of a hydride ion from the solvent to the benzylic
cation. Analogous hydride-transfer reactions from AnH2
have been reported before.30 After formation of the
anthracenylic cation, anthracene is formed by a fast
proton elimination30a (Scheme 2). In the case of the more
reactive BzH, proton elimination from the carbocation
(30) (a) Nenitzescu, C. D. In Carbonium Ions; Olah, G. A., Schleyer,
P. v. R., Eds.; Wiley: New York, 1970; Vol. 2, pp 463-520. (b)
Bonthrone, W.; Reid, D. H. J . Chem. Soc. 1959, 2773-2779. (c) Bethell,
D.; Gold, V. Carbonium Ions; Academic Press: London, 1967.
(31) Mayr, H.; Patz, M. Angew. Chem., Int. Ed. Engl. 1994, 33, 938-
957.
(29) Langoor, M. H.; Van der Maas, J . H. J . Mol. Struct. 1997, 403,
213-229. Mori, N.; Morioka, K. Bull. Chem. Soc. J pn. 1975, 48, 2213-
2214.

3018 J . Org. Chem., Vol. 64, No. 9, 1999
Dorrestijn et al.
An analogeous conversion mechanism for benzylic
alcohols may also be operative during the reduction of
benzophenone to diphenylmethane in hydroaromatic
solvents at more elevated temperatures, as has been
performed by Choi and Stock.36 Initially, the ketone is
reduced to the benzhydrol, via radical chemistry (i.e.,
RRD followed by hydrogen abstraction). Subsequently,
due to the sensitivity of the benzylic hydroxyl group
toward acid-catalyzed dehydroxylation, a benzylic cation
is formed as intermediate, Ph2C+H. As we have shown,
such a cation can abstract a hydride from the hydroaro-
matic solvent, resulting in diphenylmethane. This ratio-
nale excludes the formation of an ether from the cation
and benzhydrol as well as numerous other reactions that
have been proposed to explain the observations.
thermal, liquid-phase degradation of o-HBA in the
presence of a catalytic amount of acid. However, many
results (but not all) can also be explained on the basis of
o-QM as the intermediate. Experiments to distinguish
o-QM and the benzylic cation by spectroscopic methods
at these conditions could be helpful in solving this
matter.37
In many approaches for isolating and converting lignin,
polar solvents are applied. On the basis of the results in
this paper, it can be concluded that, in the presence of
small amounts of an acid, benzylic cations can be formed
in the lignin matrix upon heating. A large number of
acidic phenolic groups are present in lignin, and the polar
solvent increases the proton mobility through the matrix.
Next to ionic chemistry,38 radical chemistry (e.g., hydro-
gen shuttling)1 can take place as well, which increases
the complexity of lignin degradation. Electrophilic aro-
matic substitution of the formed cationic moieties may
be regarded as one of the steps in the unwanted coke
formation during the processing of lignin.
F in a l Rem a r k s
On the basis of the results, we are convinced that
the benzylic cation is the reactive intermediate in the
(32) For the reaction PhCH2 + RH f PhCH3 + R•, the reaction
•
enthalpies, ∆rH, can be calculated as -21, -54, and -100 kJ mol-1
,
respectively, for RH ) PhenH2, AnH2, and BzH.33 The Arrhenius
parameters can be derived according to the Evans-Polanyi relation:
34 log A ) 8.0 per H, Ea ) 67 + 0.35 × ∆rH kJ mol-1. At 555 K the rate
constant for hydrogen abstraction by benzyl radicals from PhenH2,
AnH2, and BzH can thus be calculated as 1 × 103, 1 × 104, and 2 ×
105 M-1 s-1, respectively.
Ack n ow led gm en t. The authors wish to thank Da-
nilo Santoro for kindly supplying a sample of 7H-
benz[de]anthracene.
J O981110F
(33) By using differential functional theory (DFT) calculated values
from: Santoro, D.; Korth, H. G.; Mulder, P. To be published.
(34) Savage, P. E. Energy Fuels 1995, 9, 590-598.
(37) The authors wish to thank an anonymous reviewer for insisting
that our data are also consistent with o-QM reactivity.
(38) Britt, P. F.; Buchanan, A. C., III; Thomas, K. B.; Lee, S. K. J .
Anal. Appl. Pyrolysis 1995, 33, 1-19.
(35) Cheng, J . P.; Handoo, K. L.; Parker, V. D. J . Am. Chem. Soc.
1993, 115, 2655-2660.
(36) Choi, C.; Stock, L. M. J . Org. Chem. 1984, 49, 2871-2875.