Understanding solid/solid organic reactions

Article abstract of DOI:10.1021/ja0034388

The concept of an organic reaction between two macroscopic solid particles is investigated. Thus, we study several reactions that have been recently reported to proceed "in the solid phase" and clearly show that, in most cases, grinding the two solid reactants together results in the formation of a liquid phase. This is true both for catalytic transformations (e.g., aldol condensations and oligomerization of benzylic compounds) and for noncatalytic reactions (Baeyer - Villiger oxidations, oxidative coupling of naphthols using iron chloride, condensation of amines and aldehydes to form azomethines, homo-etherification of benzylic alcohols using p-toluenesulfonic acid, and nuclear aromatic bromination with NBS). This liquefaction implies the existence of a eutectic mixture with T_fusion below ambient temperature (although both reagents have higher than ambient melting points). In cases where heating is required, it is again clear that a phase change (from solid to liquid) occurs, explaining the observed reaction kinetics. On the basis of 19 experimental examples, we discuss the possibility of solid-phase organic reactions and the implications of these findings to the reaction between two solid reagents. A general description of such reactive systems is proposed, based on a consideration of the potential for eutectic (or peritectic) formation between the constituents of the liquid phases that arise during the process of mechanical mixing of the solid reagents and products.

Full text of DOI:10.1021/ja0034388

J. Am. Chem. Soc. 2001, 123, 8701-8708
8701
Understanding Solid/Solid Organic Reactions
†
‡
§
,‡,⊥
Gadi Rothenberg, Andrew P. Downie, Colin L. Raston, and Janet L. Scott*
Contribution from the York Green Chemistry Group, Chemistry Department, UniVersity of York,
Heslington, York YO10 5DD, U.K., Centre for Green Chemistry, School of Chemistry, Monash UniVersity,
Clayton, Melbourne, 3800 Victoria, Australia, and School of Chemistry, UniVersity of Leeds,
Leeds, LS2 9JT, U.K.
ReceiVed September 19, 2000. ReVised Manuscript ReceiVed March 7, 2001
Abstract: The concept of an organic reaction between two macroscopic solid particles is investigated. Thus,
we study several reactions that have been recently reported to proceed “in the solid phase” and clearly show
that, in most cases, grinding the two solid reactants together results in the formation of a liquid phase. This is
true both for catalytic transformations (e.g., aldol condensations and oligomerization of benzylic compounds)
and for noncatalytic reactions (Baeyer-Villiger oxidations, oxidative coupling of naphthols using iron chloride,
condensation of amines and aldehydes to form azomethines, homo-etherification of benzylic alcohols using
p-toluenesulfonic acid, and nuclear aromatic bromination with NBS). This liquefaction implies the existence
of a eutectic mixture with Tfusion below ambient temperature (although both reagents have higher than ambient
melting points). In cases where heating is required, it is again clear that a phase change (from solid to liquid)
occurs, explaining the observed reaction kinetics. On the basis of 19 experimental examples, we discuss the
possibility of solid-phase organic reactions and the implications of these findings to the reaction between two
solid reagents. A general description of such reactive systems is proposed, based on a consideration of the
potential for eutectic (or peritectic) formation between the constituents of the liquid phases that arise during
the process of mechanical mixing of the solid reagents and products.
Introduction
Solid-state chemistry is a fast-developing science, enhanced
1
by its numerous applications in the high-technology industries.
Often depending on specific crystal modifications and high-
temperature interactions, it is associated primarily with inorganic
crystalline compounds, where reactions are thought to occur
2
via ion displacement and crystal deformation mechanisms. It
seems unreasonable that a physical class of reactions should be
confined solely to “inorganic” materials, so it is not surprising
that several studies on organic reactions in the solid phase have
also been published.3 What is surprising is the fact that many
of these organic transformations were reported to afford high
conversions and yields after short reaction times, at moderate
and even ambient temperatures, yet were reported to occur
between two macroscopic solids.
,4
The concept of a chemical reaction between two solids is a
difficult one. Moreover, various name tags are frequently
employed in this context. In the interests of clarity, we
distinguish here between solid-phase synthesis (the reaction of
molecules from a fluid phase with a solid substrate, e.g., the
polymer-supported peptide syntheses), solVent-free synthesis
Figure 1. Cartoon of solid-phase reaction (top), solvent-free reaction
(middle), and solid-state reaction (bottom).
(any system in which neat reagents react together, in the absence
of a solvent), and solid-state synthesis or solid-solid reactions,
in which two macroscopic solids interact directly and form a
third, solid, product without intervention of a liquid or vapor
phase. A cartoon of these three processes is shown in Figure 1.
In several studies, it has been proposed that when two solid
organic compounds capable of a chemical reaction are ground
in the absence of a solvent (therefore, “solvent-free”), a chemical
reaction occurs in the solid phase. Such solvent-free systems
†
University of York.
Monash University.
University of Leeds.
E-mail: janet.scott@sci.monash.edu.au.
‡
§
⊥
(
1) Smart, L.; Moore, E. Solid State Chemistry, an Introduction, 2nd.
ed.; Chapman and Hall: London, 1996.
2) For a discussion, see: Corbett, J. D. In Solid State Chemistry
(
5
6
Techniques; Cheetham, A. K., Day, P., Eds.; Oxford University Press:
Oxford, 1987; pp 1-35.
have been recently reviewed and variously termed solid-solid
(
3) For reviews, see: (a) Toda, F. Acc. Chem. Res. 1995, 28, 480-486.
(5) Toda, F.; Tanaka, K. Chem. ReV. 2000, 100, 1025-1074.
(
b) Toda, F. Synlett 1993, 303-312.
4) Rastogi, R. P.; Singh, N. B.; Singh, R. P. J. Solid State Chem. 1977,
0, 191-200.
(6) (a) Schmeyers, J.; Toda, F.; Boy, J.; Kaupp, G. J. Chem. Soc., Perkin
Trans. 2 1998, 989-993. (b) Toda, F.; Yagi, M.; Kiyoshige, K. J. Chem.
Soc., Chem. Commun. 1988, 958-959.
(
2
1
0.1021/ja0034388 CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/16/2001

8
702 J. Am. Chem. Soc., Vol. 123, No. 36, 2001
Rothenberg et al.
Table 1. Binary Mixtures of Ketones and Aldehydes Used in
Aldol Condensation Reactions. Shaded Cells Indicate Reactive
Systems (on Addition of NaOH)
7
or solid-state reactions. However, in many cases, this descrip-
tion is oversimplified, as will be demonstrated herewith. It
should be emphasized, though, that our study pertains to the
reaction between discrete solid compounds to produce a
chemically different product by the formation of new covalent
bonds. It does not pertain to reactions resulting in the formation
8
of complex structures such as a charge-transfer or molecular
complex or to reactions in a single crystalline phase such as
photochemically induced solid-state transformations (e.g., 2 +
2
dimerizations) which can, and do, occur within the organic
9
crystal itself. In addition, the reaction of two components that
crystallize together to form a discrete stoichiometric cocrystal,
or inclusion compound, may be considered to occur in the solid
phase, although this is clearly not an interaction between two
discrete solids but rather a reaction between two components
of a single crystalline phase.
1
0
In this paper, we examine some of these so-called “solid-
solid” organic transformations and produce clear evidence that
many are actually reactions in a liquid melt. On the basis of
the experimental evidence of 19 examples, we discuss the
implications of the phase change to a melt as a precondition
for a rapid chemical reaction between solid organic compounds
under solvent-free conditions.
a
“Sticky” powders: xrd analysis reveals peaks due to one component
only, i.e., melt incomplete.
Results
It is important to emphasize that the following results
represent a random selection of both previously reported and
new reactions and not a post-experimental selection of systems
that exhibit solid to liquid phase changes. The systems described
herewith encompass a diverse range of reagents and products
and are entirely representative of the reactions tested.
by grinding the reagents together with a solid base in either
stoichiometric or catalytic amounts.
12
Aldol Condensation. The Aldol condensation in its simplest
form involves the synthesis of R,â-unsaturated ketones from
two carbonyl compounds (aldehyde and ketone, or ketone and
ketone) in the presence of a strong base.11 A common carbon-
carbon homologation protocol, the Aldol reaction has been
reported to proceed with alacrity under solvent-free conditions
Grinding together the solid aldehydes or ketones without
addition of the base catalyst reveals an interesting phenom-
enon: in some cases a liquid melt is observed while in others
the solid reagents remain present as discrete crystalline phases
(
7) (a) Goud, B. S.; Desiraju, G. R. J. Chem. Res., Synop. 1995, 244-
(as evidenced by powder X-ray diffraction analysis). More
2
3
45. (b) Toda, F.; Tanaka, K.; Iwata, S. J. Org. Chem. 1989, 54, 3007-
009. (c) Toda, F.; Takumi, H.; Masafumi, A. J. Chem. Soc., Chem.
importantly, upon addition of the solid base catalyst, reaction
is only observed in those systems that exhibit a phase change
to a melt. Thus, the existence of a liquid phase is a prerequisite
for reaction in these systems. Correlation between appearance
of a liquid phase and reactivity for the aldehydes and ketones
Commun. 1990, 1270-1271. (d) Kaupp, G.; Schmeyers, J.; Kuse, A.; Atfeh,
A. Angew. Chem., Int. Ed. 1999, 38, 2896-2899. (e) Toda, F.; Tatsuya,
S.; J. Chem. Soc., Perkin Trans. 1 1989, 209-211. (f) Im, J.; Kim, J.; Kim,
S.; Hahn, B.; Toda, F. Tetrahedron Lett. 1996, 38, 451-452.
(
8) (a) Rastogi, R. P.; Singh, N. B.; Singh, R. P. J. Solid State Chem.
1
977, 20, 191-200. (b) Rastogi, R. P.; Singh, N. B.; Singh, R. P. Indian J.
1
-6 is illustrated in Table 1 (note that the converse is not true,
Chem., Sect. A 1977, 15A, 941-946. (c) Rastogi, R. P.; Singh, N. B.; Singh,
viz. that the existence of a liquid phase does not necessarily
augur reactivity at ambient temperature).
R. P. Indian J. Chem., Sect. B 1995, 34B, 764-767.
(9) For a very readable treatment of these reactions, see: (a) West, A.
R. Solid State Chemistry and it’s Applications; Wiley: Chichester, 1987;
pp 666-679. See also: (b) Tanaka, K.; Toda, F.; Mochizuki, E.; Yasui,
N.; Kai, Y.; Miyahara, I.; Hirotsu, K. Angew. Chem., Int. Ed. 1999, 38,
In all of the cases where reaction is observed, the mixture
solidifies as the solid dehydration product separates from the
melt. This removal of the dehydration product from the (melt)
reaction mixture as it is formed means that, unlike similar
reactions in solution, the Aldol condensation is inherently
irreversible under these conditions.
3
523-3525. This reactivity in the solid phase pertains chiefly to photo-
chemical processes (e.g., dimerizations, polymerizations) but also for some
thermochemical rearrangements and isomerizations and should not be
confused with the subject of the present study. For examples, see: (c) Cohen,
M. D. In ReactiVity of Solids; Anderson, J. S., Roberts, M. W., Stone, F.
S., Eds.; Chapman and Hall: London, 1972; pp 456-471. (d) Alder, G.
Organic Solid State Chemistry; Gordon and Breach: New York, 1969. For
X-ray diffraction studies of such solid-state photochemical transformations,
see: (e) Gougoutas, J. Z. Pure Appl. Chem. 1971, 27, 305-325.
The phase diagram for the mixture of 1-indanone 1 and
4
-phenylcyclohexanone 2 has been constructed from measure-
ments of thaw and melt points and is presented in Figure 2.
The eutectic temperature of 19 °C is below the ambient
temperature at which the experiments were conducted. However,
it is interesting to note that at a 1:1 molar ratio of 1:2 the liquidus
temperature is above ambient temperature, implying that some
unmelted solid 2 should be present. This is not observed in the
(10) (a) Popovitz-Biro, R.; Tang, C. P.; Chang, H. C.; Lahav, M.;
Leiserowitz, L. J. Am. Chem. Soc. 1985, 107, 4043-4058. (b) Weisenger-
Lewin, Y.; Vaida, M.; Popovitz-Biro, R.; Chang, H. C.; Mannig, F.; Frolow,
F.; Lahav, M.; Leiserowitz, L. Tetrahedron 1987, 43, 1449-1475. (c) Etter,
M. C.; Frankenbach, G. M.; Bernstein, J. Tetrahedron Lett. 1989, 30 (28),
3
617-3620
(
11) (a) For a detailed monograph on “classical” Aldol type reactions,
see: Nielsen, A. T.; Houlihan, W. J. Org. React. 1968, 16, 1-438. (b)
Heathcock, C. H. In ComprehensiVe Organic Synthesis; Trost, B. M., Eds.;
Pergammon Press: London, 1991; Vol. 2, pp 133-179.
(12) (a) Toda, F.; Tanaka, K.; Hamai, K. J. Chem. Soc., Perkin Trans.
1 1990, 3207-3209. (b) Raston, C. L.; Scott, J. L. Green Chem. 2000, 2,
49-52.

Understanding Solid/Solid Organic Reactions
J. Am. Chem. Soc., Vol. 123, No. 36, 2001 8703
phase” at 50 °C in the presence of 2:1 stoichiometric amounts
7b
of FeCl3‚6H2O (eq 3). We have studied this process and found
that no reaction occurs when 2-naphthol (white-cream powder)
and FeCl3‚6H2O (yellow-orange powder) are ground separately,
nor is any reaction or phase change observed when the two
powders are ground together (strong, light yellow powder) at
2
5 °C. However, when the combined powdered mixture is
heated to 50 °C, the sample begins to melt after ca. 1 min,
producing a bright yellow liquid. Reaction progress in the melt
is easily observable, starting from the edge of the Pyrex tube
and proceeding inward. The product solidifies and the reaction
is nearly complete after 45 min. Significantly, no reaction is
observed for an identical sample kept at 25 °C for 24 h.
The previous report states that at 50 °C the reaction took 2
h to complete, while at room temperature (no value is given
but it is reasonable to assume that room temperature is
somewhere between 20 and 30 °C) the reaction took 144 h.
This enormous difference is significant, as both samples are
claimed to react “in the solid state”. However, if both samples
were to react in the solid state, it is difficult to explain the large
differences in reaction rates (k50 °C/k25 °C ∼ 80). Temperature
Figure 2. Phase diagram of 1-indanone 1 and 4-phenylcyclohexanone
2
at constant (ambient) pressure. Filled circles represent liquidus
temperatures, filled squares thaw points, and open squares the onset
temperature of the first endotherm measured by DSC analysis.
bulk samples and may indicate the tendency of the system to
supercool.13 In addition, this implies that some heat is released
upon the grinding of the two components which leads to
complete melting of the mixture. Such heat may be generated
14
by the occurrence of “hot spots” during initial grinding of
the solids, and this phenomenon should be carefully considered
when high-intensity grinding techniques, such as ball milling,
are employed (even in cases where temperature-controlled
apparatus is used). Coexistent solid and liquid phases are also
evidenced in other systems. For example, gentle grinding of
18
sensitivity is clearly not enough to account for this. In fact,
6
-methoxytetralone 4 and veratraldehyde 5 in the absence of a
the reaction proceeds much faster at 50 °C simply because it is
not in the solid phase but rather in a liquid or melt phase.
Condensation of Amines and Aldehydes to Azomethines.
Interest in azomethines relates to their function as primary
reagents for cycloadditions, cyclizations, and enantioselective
oxidations. Their production via the condensation of various
anilines with aromatic aldehydes was recently reported to
base yields a sticky paste which contains solid 4 dispersed in a
viscous liquid phase. The presence of unmelted/undissolved 4
is verified by powder X-ray diffraction. Note that the powder
pattern due to crystalline 5 is not detected.
Baeyer-Villiger Oxidation. The reaction between various
ketones and m-chloroperbenzoic acid (m-CPBA) is a well-known
general method for lactone production.15 It has been recently
reported to occur more rapidly at ambient temperature “in the
solid state” than in solution.6b However, upon repeating the
reaction between 4-tert-butylcyclohexanone and m-CPBA (eq
6a
proceed efficiently in the solid state. We reexamined a number
of such examples and note the formation of a distinctly liquid
or molten phase in a majority of cases. For example, in the
reaction of 4-aminotoluene with 4-bromobenzaldehyde, grinding
2
), we observe that when each reagent is ground separately no
changes occur, yet when the powders are ground together a
liquid melt forms. This melt rapidly solidifies to afford the
product as a glassy material. Previously, the reaction was
reported to be complete in 30 min, but we have observed that
the bulk of the reaction is over in a matter of seconds.
each reactant separately does not lead to any changes in bulk
physical characteristics except particle size (both are white
powders). However, grinding equivalent molar quantities of the
Oxidative Coupling of Naphthols. Simple thermal oxidative
coupling of naphthols to binaphthols is of paramount interest
to the agrochemical and pharmaceutical industries, as biaryls
are key building blocks for pesticides, pharmaceuticals, and
(16) For reviews on biaryl preparation methods and applications, see:
(
1
a) Bringmann, G.; Walter, R.; Weirich, R. Angew. Chem., Int. Ed. Engl.
990, 29, 977. (b) Sainsbury, M. Tetrahedron 1980, 36, 3327-3359.
17) For a review, see: Rosini, C.; Franzini, L.; Rafaelli, A.; Salvadori,
16
natural products. The synthesis of 2,2′-dihydroxybinaphthalene
BINOL) is especially important in the production of various
(
(
P. Synthesis 1992, 503-517.
17
chiral ligands. This reaction was reported to occur “in the solid
(18) If at both temperatures the reactants and the products are solid (i.e.,
there is no other phase present during the reaction), it is to be expected
‡
‡
‡
(
13) Sangster, J. J. Phys. Chem. Ref. Data 1997, 26, 351-502.
that ∆S 25 °C ∼ ∆S 50 °C. Assuming that ∆H is independent of temperature
at this range, the ratio k50 °C:k25 °C would be roughly 6:1. Since this is
obviously not the case here, additional activation must be taking place at
50 °C. See: Isaacs, N. Physical Organic Chemistry, 2nd ed.; Longman:
Essex, 1995; pp 105-118.
(14) Bowden, F. P.; Yoffe, A. D. Fast Reactions in Solids; Butterworth
Scientific Publications: London, 1958; pp 57-77.
15) (a) Baeyer, A.; Villiger, V. Ber. 1899, 32, 3625. For a recent review,
see: (b) Renz, M.; Meunier, B. Eur. J. Org. Chem. 1999, 4, 737-750.
(

8704 J. Am. Chem. Soc., Vol. 123, No. 36, 2001
Rothenberg et al.
Figure 3. 4-Aminotoluene and 2-hydroxy-3-methoxybenzaldehyde ground independently and mixed at room temperature. (i) Immediately after
the powders contact one other an orange, liquid phase forms. (ii) After merely mixing with a glass rod the entire mass is a homogeneous liquid.
Time elapsed is approximately 2 min and the Petri dish has been shifted from a dark to a light background immediately post mixing of the powders.
(iii) The orange, liquid phase is so fluid that it may be drawn into a pasteur pipet. Solid, orange azomethine product rapidly crystallizes from the
liquid phase as the reaction proceeds and the entire mass is a wet crystalline solid in less than 10 min.
Table 2. Phase Changes and Reactivity in the Reaction between
an Aromatic Amine and a Benzaldehyde Derivative To Form an
Azomethine
Ar-NH
-CH
2
Ar-CHO
product
4
3
C
6
H
4
2-OH-3-MeOC
H
6 3
melt
melt
melt
no melt
melt
orange solid
yellow solid
yellow solid
yellow solidb
white solid
d
2
3
4
4
4
-NO
-NO
-NO
-BrC
-OHC
2
2
2
C
C
C
6
H
6
6
6
6
H
H
H
4
4
4
4
a
c
H
4
no melt
4
2
-NO C
6
H
4
4-NO
2
C
H
6 4
melts at 83-85 °C yellow solide
a
Ground mixture changes to bright yellow powder with time.
b
Incomplete reaction after 7 days. c No change in ground mixture after
8 h at ambient temperature (minimal product detectable). Reaction
incomplete after 20 days! Product formed post melt.
d
4
e
two powders together results in the formation of a liquid melt,
of yellow color, which transforms in a matter of seconds into a
white solid product. Similar behavior was noted for a number
of analogous substrates (Table 2) and is particularly striking in
the reaction between 4-aminotoluene and 2-hydroxy-3-meth-
oxybenzaldehyde (o-vanillin) pictured in Figure 3. The solid,
powdered reagents liquefy almost immediately upon contacting
each other without intervention of grinding or even extensive
mixing. The liquefaction is rapid and complete and yields a
liquid phase so fluid that it is easily drawn into a pasteur pipet.
The solid azomethine product begins to separate in a matter of
minutes, and the entire reaction mixture solidifies to a wet
crystalline mass within 10 min. Note that similar behavior was
hinted at in a previous report of the analogous reaction of
Figure 4. Comparison of powder X-ray diffraction traces for the solid
reagents (bottom traces), the product (as calculated from single-crystal
structure data - middle), and the ground mixture post grinding and
storage at 22 °C for 11 days (top). Peaks due to unreacted starting
materials are highlighted with arrows.
reaction post melting of the reagents (or reaction at an elevated
temperature). Indeed, heating the reagents together at 100 °C
for 2 min, followed by cooling to ambient temperature and DSC
analysis, leads to observation of only one significant endotherm
due to product melt, as would be expected after 100% substrate
conversion. This is further borne out by the fact that this system
loses 1 mol of water at elevated temperature in thermal
gravimetric analysis (TGA). The crystal structure of the product
compound, in the polymorphic modification crystallized from
wet dichloromethane, is not a hydrate as evidenced by single-
aromatic amines with vanillin and 4-chlorobenzaldehyde with
a,19
4
-aminotoluene.6
Remarkably, 4-aminotoluene and 4-hydroxybenzaldehyde do
not yield more than traces of azomethine product post grinding
of a 1:1 mixture, even after storage at 22 °C for 5 days. The
reaction mixture remains solid upon grinding, and X-ray powder
diffraction analysis of this mixture after storage at 22 °C for 11
days indicates that solid unreacted starting material is still
present in the reaction mixture as is illustrated in Figure 4.
These results differ from previous reports where the reaction
20
crystal structure analysis (see Figure 5). It is probable that
the water loss measured on TGA analysis arises from a reaction
occurring at a temperature significantly above ambient and post
reagent fusion.
As noted in the Aldol reaction examples, a phase change to
a binary melt appears to precede reaction in all systems tested
except that comprising 4-nitrobenzaldehyde and 4-aminotoluene.
6a
was claimed to be complete after 6 h. Closer examination of
the data presented by Schmeyers et al. gives an indication of
the reduced reactivity of this system as, upon DSC analysis,
endotherms due to fusion of each reagent are measured followed
by an endotherm corresponding to product melt. This may imply
(20) Crystal data for the azomethine product formed from 4-aminotoluene
and 4-hydroxybenzaldehyde: C14H13NO, Mr ) 211.25, orthorhombic, space
group Pbca, a ) 11.1939(3) Å, b ) 9.2537(3) Å, c ) 21.1621(8) Å, V )
(19) “Softening” was reported in the reactions with vanillin. SNOM
3
-1
measurements show that, in the reaction of single crystals of 4-nitroaniline
and 4-chlorobenzaldehyde, the entire surface of the crystal is of a single
chemical type and that there is no differentiation between the chemical
identity of the surface features ascribed to product formation and the
substrate (reagent) between these. The authors ascribe this to “short distance
sublimation” of the aldehyde (ref 6).
2192.1(1) Å , Z ) 8, µ(Mo KR) ) 0.081 mm . Of 7773 reflections
measured, 2588 were unique with 1415 I >2σ(I), R indices [I > 2σ(I)] R1
2
) 0.0496, wR2 ) 0.0923, GOF on F ) 0.959 for 146 refined parameters
and 1 restraint (OH bond length). O-H‚‚‚N hydrogen bonds yield helical
chains of molecules. Double helices composed of chains of opposite
handedness propagate down [010].

Understanding Solid/Solid Organic Reactions
J. Am. Chem. Soc., Vol. 123, No. 36, 2001 8705
Oligomerization of Benzylic Alcohols. We have previously
23
reported that the bowl-shaped supramolecular host compounds
2
4
such as cyclotriveratrylene (CTV) and the analogous tris-(O-
allyl)-CTV may be synthesized in the absence of a solvent (eq
7
).
Grinding of 4-allyloxy-3-methoxybenzyl alcohol with 1 equiv
of TsOH yields a viscous melt which solidifies over a period
of days as tris-(O-allyl)CTV is formed and separates as a
microcrystalline solid. Once again, this solvent-free reaction is
not a solid-solid reaction despite the relatively high melting
points of both reagents (86 and 103-105 °C, respectively). One
mole of water is produced for each mole of benzyl alcohol
condensed, yet we do not believe the presence of free water
accounts for the apparent liquefaction of the reaction mixture.
Although the viscosity of the monomer/acid mixture was lowest
directly after the grinding of the two components, TLC analysis
reveals <5% product at this point. Thus, the water present is
derived chiefly from the monohydrate catalyst and would be
insufficient to account for the degree of liquefaction noted
(particularly given the low aqueous solubility of the benzyl
Figure 5. Packing diagram of (4-hydroxybenzylidene)-p-tolylamine
viewed down a. Oxygen and nitrogen atoms are depicted as filled
circles, and only the OH atom is included for clarity. No water of
crystallization occurs in the structure.
We are currently investigating this specific system further, but,
in general, it appears that solid-solid reactions are not the norm
in azomethine formation.
Nuclear Bromination of Aromatics. Ring bromination of
aromatic compounds with N-bromosuccinimide (NBS) has been
described as a “genuine solid-state reaction” that has “prepara-
tive and industrial value”.7 However, these compounds will
undergo electrophilic aromatic bromination regardless of the
bromine source used,21 and it is unlikely that protocols employ-
ing NBS could compete with syntheses based on molecular
bromine.22 More importantly, upon repeating this reaction we
note that it is not a solid-state reaction at all, but rather occurs
in (again) a liquid melt of the two reagents. Thus, when a 1:3
molar ratio of either 3,5- or 3,4-dimethylphenol (reddish-
transparent crystals and white powder, respectively) and NBS
2
5
alcohol monomer ).
a
Discussion
The above results show clearly that many of the so-called
“solid-solid” reactions are not reactions in the solid state. With
only one exception, all of the reactive systems investigated
exhibited a clear change to a liquid (melt) phase. Even without
this ample experimental evidence, the occurrence of a rapid (i.e.,
faster than in solution) chemical reaction in the solid state is
hard to come to terms with on the basis of chemical intuition.
True, for a reaction in solution one would expect that the
concentration of substrates A and B would be lower than for
the pure compounds, and even when two molecules collide the
reaction may not take place (depending on reactive cross-section
(white powder) are ground together (eq 5), a yellow liquid melt
is observed which solidifies to give the brominated product as
a mixture of di- and tribrominated products, as reported
previously.
2
6
and orientation). However, in the case of two solid particles,
although substrate “concentration” may be high, the actual
number of active substrate molecules would be low because
only those molecules on the particle surface would be able to
27
react. Furthermore, the orientation of the molecules in the solid
is fixed, making for a lower cross-section, and the energy
required to disrupt the crystal lattice to enable the individual
28
molecules to react is often considerable. Solid-solid reactions
(
(
21) Rothenberg, G.; Clark, J. H. Green Chem. 2000, 2, 248-251.
22) Rothenberg, G.; Clark, J. H. Org. Process Res. DeV. 2000, 4, 270-
Etherification of Alcohols. Treatment of alcohols with
equimolar amounts of p-toluenesulfonic acid (TsOH) has been
reported as a method to synthesize the corresponding ethers in
the solid state.7 We find, however, that equimolar amounts of
diphenylmethanol (benzhydrol) and TsOH ground together (eq
274.
(23) Scott, J. L.; Raston, C. L.; MacFarlane, D. R.; Teoh, C. M. Green
Chem. 2000, 2, 123-126.
(
c
24) Systematic name: 2,3,7,8,12,13-hexamethoxy-10,15-dihydro-5H-
tribenzo[a,d,g]cyclononene.
25) BzOH (1 g) dissolves in 25 mL of water: The Merck Index, 11th
ed.; Merck & Co.: Rahway, NJ, p 1137.
26) Levine, R. D.; Bernstein, R. B. Molecular Reaction Dynamics and
Chemical ReactiVity; Oxford University Press: New York, 1987; pp 26-
6.
(
6
) form a sticky white liquid, which solidifies after 10-15 min
to give the reported ether.
(
6
(
27) Mukhopadhyay, S.; Rothenberg, G.; Wiener, H.; Sasson, Y. New
J. Chem. 2000, 24, 305-308.
28) The typical energy barrier for molecular reorientation in a crystalline
(
-
1
organic compound is 10-20 kcal mol . See: Gavezzotti, A.; Simonetta,
M. Chem. ReV. 1982, 82, 1-13.

8706 J. Am. Chem. Soc., Vol. 123, No. 36, 2001
Rothenberg et al.
occurring between two discrete crystalline solids, without
intervention of a mobile phase and which allows a large number
of productive molecular collisions, would be expected to exhibit
diffusion-controlled kinetics.29 Thus, the rapid rates exhibited
by most of the reactions studied (τ1/2,25 °C < 60 s in some cases)
do not support the theory of two solids reacting together without
intervention of a new (liquid) phase which would enable higher
substrate mobility.
Another issue which the solid-solid interpretation is incom-
plete in addressing is the phenomenon of substrate deactivation
via the formation of a solid product layer on the substrate particle
surface. If there is no solvent and no liquid phase, a chemical
reaction of A with a molecule of B on the crystal surface of B
would form the solid product C on, or in close proximity to,
the crystal surface of B (see Figure 1, bottom). Since no melting
is supposed to occur in the solid-solid reaction, in most cases
this would result ultimately in the formation of an microcrys-
talline layer, coating B with C.30 This could arrest the reaction
before full conversion is achieved (in fact, we have observed a
similar phenomenon even in the case of a solid/liquid system,
viz. the coating of CaF2 particles with CaSO4 in the in situ
3
1
production of HF from liquid H2SO4 and solid CaF2).
We propose the following description of such systems: upon
mixing of the reagents, a melt of mutually miscible A and B
exists, so that these may be considered to be mutually soluble.
At this point, the phase diagram is that of two largely immiscible
solids, which do not exhibit the formation of a molecular
complex, as shown in Figure 6i.3
2,33
In the reaction, A + B w C, the overall phase equilibria may
be represented by a triangular prism as shown in Figure 6ii. In
this prism each rectangular face represents one binary diagram
Figure 6. (i) Simple eutectic diagram (at constant P) for two
components, A and B, that are completely miscible as liquids but
practically immiscible as solids. (ii) Triangular prism basis for a three-
component phase diagram at constant P with vertical axis representing
increasing temperature. Consideration of cross-sections such as T1 and
T2 allow analysis of the phases occurring at various component
concentrations at a specific, constant temperature. (iii) and (iv) represent
cross-sections of this prism at particular temperatures. (iii) Three-
component phase diagram at temperatures greater than the AB, BC,
AC, or ABC eutectic temperature. Note liquid phase at C ) 0. (iv)
Three-component phase diagram at temperatures greater than the AC
and ABC eutectic temperature but below AB and BC eutectic. Note a
liquid phase coexists with solid A and solid B at any point where C *
0 (i.e., even a reaction which begins as a solid-solid reaction could
liquefy upon formation of product). (v) Diagram representing constant
A and B composition of 1:1 molar ratio with increasing temperature.
Moving along the dotted lines marked T1 and T2 indicates the phase
changes occurring as C increases.
3
4
(AB, AC, etc.). Since the reactions are not thermally isolated,
these approximate to systems at constant temperature, and the
chemical and phase composition may be represented by
triangular cross-sections of the prism [cf. Figures 6iii and (6iv,
which represent cross-sections at T1 and T2, respectively].
Figure 6iii shows the situation where the temperature is above
that of the binary eutectic formed by A and B (and above the
binary eutectics formed by A and C, B and C, and the ternary
eutectic formed by A, B, and C). Thus as A and B react in 1:1
stoichiometry (as in the azomethine reactions) to form product
C, the composition changes along the line qC. The liquid phase,
initially composed of 1:1 A:B, becomes enriched in C until the
liquidus line is crossed and C begins to crystallize out of the
melt. At complete conversion, the only phase present will be a
solid, crystalline C. As the quantities of A and B, relative to
each other, remain constant throughout the reaction, a vertical
section of the triangular prism [such as the one shown in Figure
fourth chemical component in the phase diagram. The same is
true of any reaction intermediates, such as the â-hydroxy ketones
formed in the Aldol type reactions, and of the catalysts where
appropriate (e.g. NaOH). However, the above description is
adequate for those cases where the quantities of catalysts are
small, the reaction intermediates are short-lived, and the
components A, B, and C have very low aqueous solubility.
Indeed, the inclusion of further miscible components only serves
to increase the proliferation (and likelihood of occurrence) of
possible liquid phases, as was shown by Epple et al. in the case
6
v] can be used to depict the phase equilibria occurring.
In several of the above examples, water is produced as the
condensation reaction proceeds and should be considered as a
(
29) Rastogi, R. P., Singh, A. K., Shukla, C. S., J. Solid State Chem.
982, 42, 136-148.
30) Such passivation would not be expected only in the special cases
1
(
where the newly formed phase would incur a larger molar volume, e.g.,
the oxidation of Fe by O2 to Fe2O3. The issue is further complicated by the
fact that mechanical stress, such as grinding, may disrupt this passivating
surface layer, exposing fresh unreacted reagent.
35
of the reductions of benzophenones with sodium borohydride.
It is interesting to consider the situation where the temperature
at which the reaction is conducted is above the eutectic
temperature of the reagents A and B. If A and B react in a
solid-solid reaction, whether by grain boundary diffusion or
(
31) Rothenberg, G.; Royz, M.; Arrad, O.; Sasson, Y. J. Chem. Soc.,
Perkin Trans. 1 1999, 1491-1494.
32) Moore, W. J. Physical Chemistry; Prentice Hall: Englewood Cliffs,
NJ, 1968; pp 146-149.
33) There are numerous examples of reactivity in systems that do exhibit
(
(
4
surface migration, product C begins to form and, should there
molecular complex formation (cf. ref 10c), but these describe reactions in
a single solid phase containing both reagents and are thus not reactions
between discrete solid phases.
exist an AC, BC, or ABC eutectic which occurs below the
(
34) Masing, G. Ternary Systems; introduction to the theory of three
(35) Epple, M.; Ebbinghaus, S.; Reller, A.; Gloistein, U.; Cammenga,
H. K. Thermochim. Acta 1995. 269, 433-441.
component systems; Dover Publications Inc., New York, 1944; pp 9-31.

Understanding Solid/Solid Organic Reactions
J. Am. Chem. Soc., Vol. 123, No. 36, 2001 8707
reaction temperature, a liquid phase may again intervene. Thus,
a reaction which begins as a solid-solid reaction may proceed
much more rapidly by intervention of a liquid phase arising
due to the existence of a lower melting eutectic formed by the
product and the reagent or reagents. This too could account for
the appearance of liquid phases and for the rapid kinetics
observed in some of the examples studied.
We have noted that, in some binary reactions which require
a catalyst for reaction, e.g., 6-methoxytetralone 4 and veratral-
dehyde 5, incomplete conversion to a liquid phase is observed
prior to addition of the catalyst. This implies that while above
the eutectic temperature, these mixtures do not exhibit the
eutectic composition and one component is present in the
mixture as a solid as would occur in region A + liquid (or B +
liquid) in Figure 6i.
and was kept refrigerated prior to reaction; no data are given regarding
the type of m-CPBA used in the previous study, and we assumed that
the highest commercial grade solid, i.e., similar, was used). p-
Toluenesulfonic acid (monohydrate) was used without further drying
while NaOH was dried to remove water absorbed on storage. Satisfac-
tory spectral data were obtained for all products.
Extreme care was taken to repeat all procedures exactly as reported.
Thermal reactions were carried out in Pyrex tubes immersed in a
thermoregulated silicon-oil bath (Heidolf MR 3003). Temperature was
monitored throughout the reactions using both thermocouple and
mercury thermometers, and the heated oil was stirred continuously to
maintain an even temperature. Temperatures obtained using this setup
were accurate to (1.0 °C. Grinding experiments were performed with
agate or porcelain mortar and pestles, which were acid washed and
dried prior to use. In all cases grinding was quite gentle and in many
cases a similar effect, albeit slower, could be achieved by pregrinding
of the reagents (to achieve particle size reduction) and subsequent
mixing.
Conclusion
Aldol Condensation. A total of 0.67 g (5.1 mmol) of 1-indanone
and 0.84 g (5.1 mmol) of 3,4-dimethoxybenzaldehyde were gently
ground together at 22 °C. A clear light brown liquid formed which,
upon addition of solid NaOH (approximately 1 molar equiv) and further
grinding, solidified to yield a brownish solid. This solid was reground,
slurried in approximately 2 mL of 0.02 M HCl, filtered, washed with
We have demonstrated that few covalent bond forming
bimolecular organic transformations that proceed rapidly and
to a high degree of completion between two solid reactants
actually occur in the solid state. Instead, a liquid or melt phase,
which imbues the individual molecules with the required
mobility for productive (or reactive) collision, intervenes,
allowing rapid reaction between the two solid reagents. These
reactions should therefore be classified together with classical
liquid/liquid and liquid/solid systems that react in the absence
of an added solvent.
1
mL of water, and dried to yield 1.33 g of 2-(3,4-dimethoxyben-
41
zylidene)indan-1-one (94%), mp 183-185 °C (lit. mp 175-177 °C).
Other condensation reactions with the ketones and aldehydes listed in
Table 1 were carried out similarly. The products were identified by H
1
NMR analysis and melting point.
Ketone Oxidation. A total of 1.12 g (3.24 mmol) of m-CPBA and
0.25 g (1.62 mmol) of 4-tert-butylcyclohexanone were ground together
Experimental Section
at 25 °C. A glassy liquid formed which quickly solidified into the
product 3-tert-butylcaprolactone (preparative TLC; hexane:EtOAc 2:1),
mp 49 °C (lit.42 mp 51-54 °C).
Powder XRD analysis was performed on a Syntag PAD5 diffrac-
tometer at 294 K using Cu KR radiation (λ ) 1.54059 Å) in the range
Naphthol Coupling. A total of 1.00 g (7.0 mmol) of 2-naphthol
6
-36° 2θ. DSC analysis was performed on a Perkin-Elmer DSC-7
3 2
and 3.80 g (14.0 mmol) of FeCl ‚6H O were ground separately and
system in sealed 50 µL aluminum pans on ground samples of 4-10
mg at a heating rate of 10 °C/min. The system was calibrated using
then ground together at 25 °C and heated to 50 °C. The mixture melted
and then solidified. Recrystallizing (from EtOH/water) afforded 1.70
g of 2,2′-dihydroxybinaphthalene (85% based on 2-naphthol), mp 212
1
two standards with appropriate melting points. H NMR spectra were
recorded on a Varian Mercury AM 300 MHz spectrometer in CDCl
with TMS reference. Melting points were measured using an Electro-
thermal digital melting point apparatus in open capillary tubes or with
a Reichert hotstage microscope and temperature controller and are
uncorrected.
3
C (lit.43 mp 215-217 °C).
°
Azomethines. A total of 300 mg (2.8 mmol) of 4-aminotoluene and
5
10 mg (2.8 mmol) of 4-bromobenzaldehyde were ground together to
give a yellow liquid melt which rapidly solidified to give (4-
bromobenzylidene)-p-tolylamine quantitatively as a white solid, mp
Binary phase diagrams were constructed from data obtained using
1
44-147 °C. Reactions with the benzaldehyde and aniline derivatives
3
6
the thaw-melt method of Rastogi and measurement of the liquidus
listed in Table 2 were performed in similar manner to yield the
appropriate azomethines. In the case of 4-nitroaniline and 4-nitroben-
zaldehyde, the ground mixture was heated until melted and maintained
at this temperature for 10 min during which time it solidified to the
yellow product.
temperatures using the visual-polythermal method.1
3
Single-crystal X-ray diffraction data were collected on an Enraf
Nonius CCD diffractometer at 123 K using graphite-monochromated
Mo KR radiation (λ ) 0.71073 Å). The data were corrected for Lorentz
polarization effects. Structures were solved by direct methods using
Phenol Bromination. A total of 61.1 mg (0.50 mmol) of 3,4-
dimethylphenol and 266 mg (1.50 mmol) of NBS (white powder) were
ground together to yield a yellow liquid melt which solidified to give
a mixture of 2,6-dibromo-3,4-dimethylphenol and 2,3,6-tribromo-4,5-
dimethylphenol. The reaction of 3,5-dimethylphenol was performed
similarly.
Benzhydrol Etherification. A total of 92.1 mg (0.50 mmol) of
benzhydrol and 95.1 mg (0.50 mmol) of TsOH were ground together
at 25 °C to form a sticky white liquid, which solidified after 10-15
min to give 1,1,1′,1′-tetraphenyldimethyl ether.
Oligomerization of Benzylic Alcohols. A total of 0.33 g (1.68
mmol) of 4-allyloxy-3-methoxybenzyl alcohol and 0.32 g (1.69 mmol)
of p-toluenesulfonic acid monohydrate were ground together to produce
a sticky, pale brown syrup which solidified over a period of 2 days to
yield tris-(O-allyl)CTV. The solid material was dissolved in dichlo-
romethane, insoluble material was removed by filtration, and the pure
crystalline product was obtained by slow evaporation of the solvent.
3
7
the SHELXS-97 program and refined by full matrix least squares
2
38
39
refinement on F using the SHELXL-97 and XSEED programs. Non-
hydrogen atoms were refined anisotropically. All hydrogen atoms,
except that of the OH group, were inserted at geometrically determined
positions with temperature factors fixed at 1.2 times (methine and
methylene) and 1.5 times (methyl) that of the parent atom. The OH
hydrogen atom position was located from electron density difference
maps and refined with a simple bond length restraint. Calculated X-ray
40
powder patterns were generated using the LAZYPULVERIX program
as modified and included in the XSEED program.
All reagents were of 98% purity or greater and were used as
purchased from the supplier unless noted otherwise. Reagent grade
m-CPBA (Aldrich) was assessed prior to reaction to contain 50%
m-CPBA, 15% m-CBA, and ca. 35% water (this compound is a solid
3
9
(
(
(
(
36) Rastogi, R. P.; Rama Verma, K. T.; J. Chem. Soc. 1956, 2097.
37) Sheldrick, G. M. SHELXS-97, University of Gottingen, 1990.
38) Sheldrick, G. M. SHELXL-97, University of Gottingen, 1997.
39) Barbour, L. J. 1999, X-Seed - a graphical interface to the SHELX
(41) Poirier, Y.; Lozac’h, N. Bull. Soc. Chim. Fr. 1966, 1052-1068.
(42) Allinger, N. L.; Greenberg, S. J. Am. Chem. Soc. 1962, 84, 2394-
2402.
program suite, University of Missouri.
40) Yvon, K.; Jetschko, W.; Parth e´ , E. J. Appl. Crystallogr. 1977, 10,
(
7
3.
(43) Akimoto, H.; Yamada, K. Tetrahedron 1971, 27, 5999-6009.

8708 J. Am. Chem. Soc., Vol. 123, No. 36, 2001
Rothenberg et al.
The white crystalline product (0.16 g, 54%) obtained proved to be pure
Supporting Information Available: DSC of the 4-amino-
toluene and 4-hydroxybenzaldehyde system showing changes
effected by heating to melt. Phase diagrams of mixtures of
tris-(O-allyl)CTV when compared with authentic samples by melting
1
point, TLC, and H NMR analysis.
2
-hydroxy-3-methoxybenzaldehyde (o-vanillin)/(2-hydroxy-3-
methoxybenzylidene)-p-tolylamine (azomethine product), and
-aminotoluene/(2-hydroxy-3-methoxybenzylidene)-p-tolyl-
Acknowledgment. We thank the referees for their perspica-
cious comments, which greatly aided us in the revision of this
paper. Part of this work was supported by a Centre for Green
Chemistry research grant and a Special Monash Research Fund
4
amine. Data pertaining to the crystal structure analysis of the
condensation product of p-aminotoluene and p-hydroxybenzal-
dehyde [(4-hydroxybenzylidene)-p-tolylamine] as an X-ray
crystallographic file (CIF). This material is available free of
charge via the Internet at http://pubs.acs.org.
(
SMURF) grant for Green Chemistry, and we gratefully
acknowledge this support. G.R. thanks the European Commis-
sion for a Marie Curie Individual Research Fellowship and
gratefully acknowledges the support of the Leo Baeck (London)
lodge.
JA0034388