A Scalable Synthesis of Meso-Substituted Dipyrromethanes

Article abstract of DOI:10.1021/op034083q

A one-flask synthesis of meso-substituted dipyrromethanes has been refined. The procedure entails reaction of an aldehyde in 100 equiv of pyrrole as the solvent containing a mild Lewis acid (e.g., InCl₃) at room temperature. Following removal and recovery of excess pyrrole, the dipyrromethane is obtained by crystallization. The procedure generates minimal waste and does not require aqueous/organic extraction, chromatography, or distillation. The procedure has been scaled linearly to obtain >100 g of 5-phenyldipyrromethane. The utility of various analytical methods for characterizing dipyrromethanes has been investigated.

Full text of DOI:10.1021/op034083q

Organic Process Research & Development 2003, 7, 799−812
A Scalable Synthesis of Meso-Substituted Dipyrromethanes
Joydev K. Laha, Savithri Dhanalekshmi, Masahiko Taniguchi, Arounaguiry Ambroise, and Jonathan S. Lindsey*
Department of Chemistry, North Carolina State UniVersity, Raleigh, North Carolina 27695-8204, U.S.A.
Chart 1
Abstract:
A one-flask synthesis of meso-substituted dipyrromethanes has
been refined. The procedure entails reaction of an aldehyde in
100 equiv of pyrrole as the solvent containing a mild Lewis acid
(e.g., InCl3) at room temperature. Following removal and
recovery of excess pyrrole, the dipyrromethane is obtained by
crystallization. The procedure generates minimal waste and
does not require aqueous/organic extraction, chromatography,
or distillation. The procedure has been scaled linearly to obtain
>100 g of 5-phenyldipyrromethane. The utility of various
analytical methods for characterizing dipyrromethanes has been
investigated.
The first report of a one-flask synthesis entailed reaction
of 4-pyridinecarboxaldehyde with 2.1 mol equiv of pyrrole
in methanol acidified with gaseous HCl, whereupon the
corresponding 5-(4-pyridyl)dipyrromethane precipitated as
the hydrochloride salt.2 In most applications such precipita-
tion is not possible, in which case an excess of pyrrole is
employed to suppress the continued reaction leading to linear
and cyclic oligomers. A number of reports in the early-mid
1990s described methods where the aldehyde (0.04-0.5 M)
was treated with excess pyrrole (2.1-40 mol equiv) in an
acidified organic solvent: BF3‚O(Et)2/CH2Cl2,3 acetic acid/
DMF4 or THF,5 SnCl4/CH2Cl2,6 p-toluenesulfonic acid/
MeOH7,8 or toluene,9 or aqueous HCl/THF.10 Workup
typically entailed several steps including column chroma-
tography, although Hammel et al. employed flash chroma-
tography followed by Kugelrohr distillation.3 In 1994, we
reported a method that employed the reaction of the aldehyde
(∼0.3 M) dissolved in neat pyrrole (∼14 M) with no other
solvent, relying on column chromatography for purification
(“1994 solventless synthesis”).11 Catalysis was achieved at
room temperature with TFA or BF3‚O(Et)2 or, in some
cases,12 upon heating without added acid. The reaction
proceeded in a few minutes at room temperature and afforded
the dipyrromethane in yields of ∼40-60%, but the use of
chromatography for purification limited the scale.
Introduction
Dipyrromethanes occupy a central place in porphyrin
chemistry. The dipyrromethane structures employed in the
synthesis of naturally occurring porphyrins typically bear
substituents at the â-positions and lack any substituent at
the meso position (Chart 1). In the past decade, dipyrro-
methanes lacking â-substituents but substituted at the meso-
position (i.e., the 5-position) have come to play a valuable
role in the preparation of synthetic porphyrins and related
compounds (dipyrrins, calixpyrroles, chlorins, corroles). A
number of stepwise syntheses of dipyrromethanes lacking
â-substituents have been developed,1 while more direct routes
have employed one-flask condensations of pyrrole and the
desired aldehyde.2-16
* To whom correspondence should be addressed. E-mail: jlindsey@ncsu.edu.
(1) (a) Treibs, A.; Haberle, N. Liebigs Ann. Chem. 1968, 718, 183. (b) Chong,
R.; Clezy, P. S.; Liepa, A. J.; Nichol, A. W. Aust. J. Chem. 1969, 22, 229.
(c) Clezy, P. S.; Smythe, G. A. Aust. J. Chem. 1969, 22, 239. (d) Wilson,
R. M.; Hengge, A. J. Org. Chem. 1987, 52, 2699. (e) Wallace, D. M.; Leung,
S. H.; Senge, M. O.; Smith, K. M. J. Org. Chem. 1993, 58, 7245. (f) Setsune,
J.; Hashimoto, M. J. Chem. Soc., Chem. Commun. 1994, 657. (g) Setsune,
J.; Hashimoto, M.; Shiozawa, K.; Hayakawa, J.; Ochi, T.; Masuda, R.
Tetrahedron 1998, 54, 1407. (h) Volz, H.; Holzbecher, M. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1442. (i) Lin, V. S.-Y.; Iovine, P. M.; DiMagno, S.
G.; Therien, M. J. Inorg. Synth. 2002, 33, 55.
(2) Nagarkatti, J. P.; Ashley, K. R. Synthesis 1974, 186.
(3) Hammel, D.; Erk, P.; Schuler, B.; Heinze, J.; Mu¨llen, K. AdV. Mater. 1992,
4, 737.
(4) Vigmond, S. J.; Kallury, K. M. R.; Thompson, M. Anal. Chem. 1992, 64,
2763.
Several groups have made modifications to the 1994
solventless synthesis. Both Boyle and we altered the workup
protocol to facilitate preparative-scale synthesis. Boyle
employed flash chromatography to remove higher oligomers,
followed by Kugelrohr distillation, enabling isolation of as
(9) (a) Staab, H. A.; Carell, T.; Do¨hling, A. Chem. Ber. 1994, 127, 223. (b)
Shipps, G., Jr.; Rebek, J., Jr. Tetrahedron Lett. 1994, 35, 6823. (c) Carell,
T. Ph.D. Thesis, Ruprecht-Karls-Universita¨t Heidelberg, 1993.
(10) (a) Wijesekera, T. P. Can. J. Chem. 1996, 74, 1868. (b) Nishino, N.; Wagner,
R. W.; Lindsey, J. S. J. Org. Chem. 1996, 61, 7534.
(5) Vigmond, S. J.; Chang, M. C.; Kallury, K. M. R.; Thompson, M.
Tetrahedron Lett. 1994, 35, 2455.
(6) (a) Casiraghi, G.; Cornia, M.; Zanardi, F.; Rassu, G.; Ragg, E.; Bortolini,
R. J. Org. Chem. 1994, 59, 1801. (b) Cornia, M.; Binacchi, S.; Del Soldato,
T.; Zanardi, F.; Casiraghi, G. J. Org. Chem. 1995, 60, 4964. Also see:
Casiraghi, G.; Cornia, M.; Rassu, G.; Del Sante, C.; Spanu, P. Tetrahedron
1992, 48, 5619.
(7) Mizutani, T.; Ema, T.; Tomita, T.; Kuroda, Y.; Ogoshi, H. J. Am. Chem.
Soc. 1994, 116, 4240.
(8) (a) Boyle, R. W.; Karunaratne, V.; Jasat, A.; Mar, E. K., Dolphin, D. Synlett
1994, 939. (b) Boyle, R. W.; Xie, L. Y.; Dolphin, D. Tetrahedron Lett.
1994, 35, 5377.
(11) Lee, C.-H.; Lindsey, J. S. Tetrahedron 1994, 50, 11427.
(12) Gryko, D.; Lindsey, J. S. J. Org. Chem. 2000, 65, 2249.
(13) Bru¨ckner, C.; Sternberg, E. D.; Boyle, R. W.; Dolphin, D. Chem. Commun.
1997, 1689.
(14) Boyle, R. W.; Bruckner, C.; Posakony, J.; James, B. R.; Dolphin, D. Org.
Synth. 1998, 76, 287.
(15) Bru¨ckner, C.; Posakony, J. J.; Johnson, C. K.; Boyle, R. W.; James, B. R.;
Dolphin, D. J. Porphyrins Phthalocyanines 1998, 2, 455.
(16) Littler, B. J.; Miller, M. A.; Hung, C.-H.; Wagner, R. W.; O’Shea, D. F.;
Boyle, P. D.; Lindsey, J. S. J. Org. Chem. 1999, 64, 1391.
10.1021/op034083q CCC: $25.00 © 2003 American Chemical Society
Published on Web 09/09/2003
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
799

thetic reactions.22 (3) In a few cases examined, a larger
pyrrole:aldehyde ratio (e.g., 400:1) gave dipyrromethanes in
yields of 90-95%.21,23 Accordingly, we began our studies
by examining acids and pyrrole:aldehyde ratios that might
afford less black material, which consumes starting material
and complicates the purification procedure.
Chart 2
During the course of our work, two new methods were
reported for carrying out the aldehyde-pyrrole condensation
leading to dipyrromethanes: (1) the use of ion-exchange
resins as acid catalysts24 and (2) the use of refluxing aqueous
acid as a solvent for the reaction from which the dipyr-
romethane is obtained as a crystalline solid.25 The one-flask
solventless synthesis approach also has found other applica-
tions, including (1) reaction of an aldehyde with excess furan
or thiophene, affording the difurylmethane or dithienyl-
methane, respectively;26 (2) reaction of a ketone with excess
pyrrole, affording a 5,5-dialkyldipyrromethane;27 and (3)
reaction of an ortho ester with excess pyrrole, affording the
corresponding tripyrromethane.28 A free-radical reaction has
been employed to obtain 5-trifluoromethyldipyrromethane.29
Such reports illustrate the ongoing interest in efficient
syntheses of dipyrromethanes and related compounds.
In this contribution we report our studies concerning the
further refinement of the solventless synthesis. The studies
include investigation of acid catalysts that are significantly
milder than TFA or BF3‚O(Et)2, a survey of analytical
methods to characterize reaction byproducts, and develop-
ment of a simplified workup procedure. The resulting one-
flask solventless synthesis employs exceptionally mild
conditions, enables purification by crystallization, and can
be scaled to give at least 100-g quantities of the dipyr-
romethane.
much as ∼9 g of product.13-15 We examined the crude
reaction mixture and found the dominant products to consist
of the dipyrromethane, the R,â-linked “N-confused” dipyr-
romethane, and tripyrrane (Chart 2); other species were
attributed to higher oligomers but not characterized.16 We
developed the following purification protocol: (1) aqueous
base treatment and extraction with ethyl acetate; (2) removal
of ethyl acetate and pyrrole; (3: an optional procedure that
was done in cases where the crude product was especially
darkened) filtration through a pad of silica to remove black
material; (4) Kugelrohr distillation to give the dipyrromethane
and N-confused dipyrromethane; and (5) recrystallization to
remove the N-confused dipyrromethane. This multistep
purification protocol was effective for many aldehydes in
small-scale preparations. However, dipyrromethanes derived
from aldehydes bearing large or sensitive substituents could
not be distilled, in which case chromatography was typically
employed for purification. A few reports have appeared of
the direct crystallization of the dipyrromethane from the
reaction mixture, but given the complexity of the reaction
mixture, direct crystallization appears to be viable only for
selected aldehydes.17
We sought to modify the conditions of the solventless
synthesis such that the dipyrromethane could be isolated by
crystallization from the crude reaction mixture. Our approach
was guided by several observations. (1) Our prior analysis
of the product distribution of the dipyrromethane-forming
reaction employed GC and quantitated only the volatile
products (dipyrromethane, N-confused dipyrromethane, and
tripyrrane; ∼80%, 2-3%, and 15% relative yields, respec-
tively, with TFA catalysis) of the reaction.16 However, TLC
analysis of the crude reaction mixture showed the presence
of black material at the origin. The isolated yield of
dipyrromethane generally fell significantly below the 80%
implied by the GC analysis, as expected if the black ma-
terial is nonvolatile and not analyzed by GC. (2) We recently
found that a wide variety of acid catalysts can be used in
the pyrrole-aldehyde condensation leading to the porphy-
rinogen.18,19 We also found that several mild Lewis acids
[InCl3, Sc(OTf)3, Dy(OTf)3, Yb(OTf)3] provide superior
results compared with TFA in porphyrin syntheses via di-
pyrromethane-carbinols.20,21 Mild Lewis acids of this type
have been found to have beneficial effects in diverse syn-
Results and Discussion
I. Analytical Methods. We first examined the applicabil-
ity of various analytical techniques (TLC, NMR spectros-
copy, GC, and elemental analysis) for identifying and
quantifying the products of the pyrrole-aldehyde condensa-
tion. The model reaction of pyrrole + benzaldehyde was
selected for these studies. The dominant products formed in
this reaction are 5-phenyldipyrromethane (1), N-confused
5-phenyldipyrromethane (2), and 5,10-diphenyltripyrrane (3)
(22) (a) Babu, S. A. Synlett 2002, 531. (b) Kobayashi, S.; Sugiura, M.; Kitagawa,
H.; Lam, W. W.-L. Chem. ReV. 2002, 102, 2227.
(23) (a) Loewe, R. S.; Tomizaki, K.-Y.; Youngblood, W. J.; Bo, Z.; Lindsey, J.
S. J. Mater. Chem. 2002, 12, 3438. (b) Yu, L.; Muthukumaran, K.;
Sazanovich, I. V.; Kirmaier, C.; Hindin, E.; Diers, J. R.; Boyle, P. D.;
Bocian, D. F.; Holten, D.; Lindsey, J. S. Inorg. Chem. 2003, 42. In press.
(24) Naik, R.; Joshi, P.; Kaiwar, S. P.; Deshpande, R. K. Tetrahedron 2003, 59,
2207.
(25) Sobral, A. J. F. N.; Rebanda, N. G. C. L.; da Silva, M.; Lampreia, S. H.;
Silva, M. R.; Beja, A. M.; Paixa˜o, J. A.; Gonsalves, A. M. d’A. R.
Tetrahedron Lett. 2003, 44, 3971.
(26) Cho, W.-S.; Lee, C.-H. Bull. Korean Chem. Soc. 1998, 19, 314.
(27) (a) Dube´, T.; Ganesan, M.; Conoci, S.; Gambarotta, S.; Yap, G. P. A.
Organometallics 2000, 19, 3716. (b) Bucher, C.; Zimmerman, R. S.; Lynch,
V.; Kra´l, V.; Sessler, J. L. J. Am. Chem. Soc. 2001, 123, 2099. (c)
Arumugam, N.; Chung, W.-Y.; Lee, S.-W.; Lee, C.-H. Bull. Korean Chem.
Soc. 2001, 22, 932. (d) Depraetere, S.; Dehaen, W. Tetrahedron Lett. 2003,
44, 345.
(17) Bru¨ckner, C.; Karunaratne, V.; Rettig, S. J.; Dolphin, D. Can. J. Chem.
1996, 74, 2182.
(18) Geier, G. R., III; Ciringh, Y.; Li, F.; Haynes, D. M.; Lindsey, J. S. Org.
Lett. 2000, 2, 1745.
(19) Geier, G. R., III; Lindsey, J. S. J. Porphyrins Phthalocyanines 2002, 6,
159.
(20) Geier, G. R., III; Callinan, J. B.; Rao, P. D.; Lindsey, J. S. J. Porphyrins
Phthalocyanines 2001, 5, 810.
(28) Beer, P. D.; Cheetham, A. G.; Drew, M. G. B.; Fox, O. D.; Hayes, E. J.;
Rolls, T. D. Dalton Trans. 2003, 603.
(29) Dmowski, W.; Piasecka-Maciejewska, K.; Urbanczyk-Lipkowska, Z. Syn-
thesis 2003, 841.
(21) Speckbacher, M.; Yu, L.; Lindsey, J. S. Inorg. Chem. 2003, 42, 4322.
800
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

Scheme 1
Table 1. Molecular formula and limitations of elemental
analysis
molecular formula
meso
maximum
substituent dipyrromethane
tripyrrane
tripyrranea (%)
phenylb C15H14N2
C26H23N3
15.6
46.4
17.1
[C1.000H0.933N0.133] [C1.000H0.885N0.115
C9H10N2 C14H15N3
[C1.000H1.111N0.222] [C1.000H1.071N0.214
mesityld C18H20N2
C32H35N3
[C1.000H1.111N0.111] [C1.000H1.094N0.094
]
]
]
Hc
a Calculated maximum percentage of a tripyrrane contaminant in a dipyr-
romethane sample that gives a satisfactory elemental analysis (<0.4% deviation
in any element). b 5-Phenyldipyrromethane and 5,10-diphenyltripyrrane. c Dipyr-
romethane and tripyrrane. d 5-Mesityldipyrromethane and 5,10-dimesityltripyr-
rane.
example, in the case of 5-phenyldipyrromethane (1), 5,10-
diphenyltripyrrane (3) can be present as a contaminant at a
level of up to 15.6%, yet the mixture will still fall within
the 0.4% error limit upon elemental analysis. For other
dipyrromethanes, the amount of contaminating tripyr-
romethane depends on the nature of the meso substituent as
shown in Table 1 (also see Supporting Information). In short,
an isolated dipyrromethane may give a satisfactory elemental
analysis despite the presence of substantial amounts of
N-confused dipyrromethane and tripyrrane contaminants.
(c) NMR Spectroscopy. The 1H NMR spectra of 1, 2, 3
are shown in Figure 1. Each peak was assigned by 2D-NMR
(HH COSY and NOESY) measurements. The spectrum of
a crude reaction mixture (following neutralization and
removal of pyrrole) is also shown. While 1 is the dominant
species, the presence of 2 and 3 can be readily detected by
the characteristic (and nonoverlapping) peaks of the 5-phen-
yldipyrromethane meso-H5 (∼5.4 ppm), N-confused 5-phen-
yldipyrromethane H7 (flanking the â-substituent, ∼6.5 ppm),
and the 5,10-diphenyltripyrrane H7 (two protons at the
â-positions of the central pyrrole ring, ∼5.7 ppm). The
relative amounts of the three species in the crude reaction
mixture can be calculated by integration of these character-
istic peaks, taking into account the 1:1:2 ratio of the number
of hydrogens: 1/2/3 ) 88:9:3. This method is satisfactory
for identifying the presence of the N-confused 5-phenyl-
dipyrromethane and 5,10-diphenyltripyrrane at the few
percent level or greater.
(d) Gas Chromatography. We previously reported that
the GC trace of the crude reaction mixture, obtained with
TFA catalysis, shows three major peaks.16 The GC trace of
the crude reaction mixture obtained with 0.1 equiv of InCl3
and a pyrrole:benzaldehyde ratio of 100:1 for 1 h also shows
three major peaks (Figure 2A) that are assigned as 1 (tR )
16.4 min), 2 (tR ) 17.0 min), and 3 (tR ) 21.8 min) on the
basis of comparison with authentic samples. Closer scrutiny
of the same GC trace upon amplification of the scale reveals
a number of additional minor peaks (Figure 2B). The
appearance and intensity of these minor peaks depend on
the reaction conditions (catalyst, reaction time, and pyrrole:
benzaldehyde ratio). For example, some of the minor peaks
seen in InCl3-catalyzed reactions are not observed in TFA-
catalyzed reactions.
(Scheme 1). We have used pure authentic samples of 1,16
2,16 and 330 for direct comparison during the analysis. The
following results were obtained upon reaction of a 100:1 ratio
of pyrrole:benzaldehyde at room temperature with catalysis
by InCl3 (identical results were obtained with TFA catalysis).
(a) TLC Analysis. Compounds 1, 2, and 3 possess similar
Rf values on silica [hexanes/ethyl acetate (4/1)]: 1, 0.53; 3,
0.48; 2, 0.38. The order of elution changed on alumina in
the same solvent system: 1, 0.52; 2, 0.39; 3, 0.32. Owing
to the similar retention factors on either medium, careful
chromatography is required to distinguish 1, 2, and 3 by TLC
analysis (see Supporting Information). Benzaldehyde is
readily distinguished on silica, however (Rf ) 0.64). TLC
can be used to quantitate the amount of unreacted benzal-
dehyde through the use of a series of standards of known
concentration (see Experimental Section).31 The limit of
detection corresponds to 0.1% unreacted benzaldehyde for
a reaction of pyrrole:benzaldehyde in a 100:1 ratio.
(b) Elemental Analysis. The synthesis of dipyrromethanes
is such that a satisfactory elemental analysis (<0.4%
deviation from calculated composition) for an isolated
dipyrromethane may not necessarily imply the desired level
of purity. Elemental analysis is obviously blind to the
presence of an N-confused dipyrromethane, which is a
dipyrromethane isomer. Perhaps less obvious is the lack of
sensitivity to the presence of contamination by a tripyrrane.
A tripyrrane is an oligomer of the dipyrromethane and has
an elemental ratio similar to that of the dipyrromethane. For
(30) Ka, J.-W.; Lee, C.-H. Tetrahedron Lett. 2000, 41, 4609.
(31) Li, F.; Yang, K.; Tyhonas, J. S.; MacCrum, K. A.; Lindsey, J. S. Tetrahedron
1997, 53, 12339.
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
801

Figure 2. Typical GC trace of the crude mixture (following
neutralization) obtained from the condensation of pyrrole and
benzaldehyde (100:1, 0.1 equiv of InCl3, 3 h) using a GC (FID)
with temperature gradient (temp 1, 35 °C (5 min); temp 2, 315
°C (12 min); rate 20 °C/min, total runtime 31 min). The peaks
were confirmed with standard samples of 1, 2, and 3 or assigned
in separate runs by GC-MS analysis. The peaks labeled as 2′
and 3′ are isomers of N-confused 5-phenyldipyrromethane (2)
and 5,10-diphenyltripyrrane (3), respectively. The trace in (B)
is an expansion of the trace in (A).
(iv) The peaks due to 1 (tR ) 16.4 min), 2 (tR ) 17.0
min), and 3 (tR ) 21.8 min) were readily assigned on the
basis of the use of authentic standards.
(v) GC-MS analysis showed the peak at 17.6 min to have
m/z ) 222, the same as 2, implying this peak stems from an
isomer of 2. We label this peak 2′.
(vi) The peak at 18.8 min (denoted “B”) has m/z ) 413,
which is attributed to a chloro-substituted tripyrrane of
unknown structure. Peak B virtually disappeared when the
reaction was performed with TFA catalysis, suggesting InCl3
as the source of the chloride.
(vii) Most of the remaining peaks fall into two categories,
depending on the observed m/z. A number of the peaks
have m/z ) 377, the same as 3, and are labeled 3′, indicating
they are isomers of 3. Four such peaks are observed in the
GC-FID trace (Figure 3A), while six such peaks are
observed in the GC-MS trace (Figure 3B).
(viii) A number of other peaks have m/z ) 310, consistent
with benzyldipyrrin isomers or tautomers (C22H18N2), and
are labeled “C” (vide infra).
The apparent presence of additional isomers of the
dipyrromethane and tripyrrane was at first surprising. How-
ever, there are three 5-phenyldipyrromethanes given the
possibility of R or â substitution. Thus, 1 is the RR isomer,
2 is the Râ isomer (which exists as a pair of enantiomers),
and 2′ likely stems from the ââ isomer. There are 10 possible
regioisomers of a 5,10-diphenyltripyrrane (3 and 3′), depend-
Figure 1. 1H NMR spectra (in CDCl3, 400 MHz) of 5-phenyl-
dipyrromethane (1); N-confused 5-phenyldipyrromethane (2);
5,10-diphenyltripyrrane (3); and a sample from the crude
reaction mixture (following neutralization and removal of
pyrrole) obtained from the condensation of pyrrole and ben-
zaldehyde (100:1) using 0.1 equiv of InCl3.
The region between 17 and 24 min was further expanded
as shown in Figure 3. Some differences in the chromatogram
are observed on the basis of the choice of detector [flame-
ionization detector (FID) or electron capture (EC)]. We
employed GC-MS to gain information concerning the
identity of the various minor peaks. The peaks in the
chromatograms of Figures 2 and 3 can be assigned as
follows:
(i) The peak at tR ) 8.2 min stems from unreacted
benzaldehyde, which was observed at early times but not
after 1 h of reaction (not shown).
(ii) Broad, low-intensity peaks around 9-11 min were
invariably observed, even upon injection of a sample of
pyrrole alone and, to a lesser degree, upon injection of a
sample of benzaldehyde alone. Such peaks are considered
to be GC artifacts.
(iii) The peak at 12.9 min (denoted “A”) shows m/z )
157 by GC-MS, which is consistent with a benzylpyrrole
(C11H11N) structure.
802
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

Chart 3. Structures of possible tripyrromethane isomers (3′)
Figure 3. Expanded region of the GC traces of the crude
mixture (following neutralization) from the condensation of
pyrrole and benzaldehyde (100:1, 0.1 equiv of InCl3, 3 h). (A)
Further expansion of the traces in Figure 2 (using FID). (B)
The same reaction sample upon analysis by GC-MS with the
same temperature gradient and a nearly identical column.
Numerous peaks from tR ) 20.5 to 21.5 min are attributed to
tripyrrane isomers (3′, m/z ) 377) and unknown species “C”
(m/z ) 310, consistent with a benzyl dipyrrin).
ing on whether the four substitutions occur at the R- or
â-positions. Moreover, the two meso carbons in the tripyrrane
are both chiral, leading to additional stereoisomers. The
tripyrrane regioisomers are shown in Chart 3.
The structures of the components that give rise to peaks
“C” are not known, but the observed m/z ) 310 is consistent
with benzyldipyrrin isomers or tautomers (C22H18N2). An
example of a benzyldipyrin is shown in Chart 4. Eight
benzyldipyrrin isomers with various R/â substitution patterns
are possible (and additional azafulvene tautomers are con-
ceivable). It is noteworthy that a benzyldipyrrin analogous
to 4 has been isolated as a side product in a porphyrin-
forming reaction.32 A mechanism for the formation of 4 in
pyrrole-aldehyde condensations has been proposed.33
In the following studies, the relative yields of 1, 2, and 3
have been identified by GC analysis. The remaining peaks
(except benzaldehyde and any known GC artifacts) in the
12-31 min region have been grouped as “other volatile
components”. While the magnitude of the latter species are
typically in the 1-5% range, these are components that must
be removed during purification and which are not easily
Chart 4
methods in the following survey to identify improved
reaction conditions and workup procedures.
None of the above analytical methods was applicable for
assessing the amount of darkened material formed in the
reactions. TLC analysis provided a means of separating black
materials from the dipyrromethane and other mobile com-
ponents, but was not reliable for gauging the extent of
darkening of the reaction mixture. Accordingly, we employed
visual inspection to gauge the extent of darkening in
comparisons of the effects of different reaction conditions.
1
identified by H NMR spectroscopy or elemental analysis.
We have employed GC in conjunction with other analytical
(32) Hill, C. L.; Williamson, M. M. J. Chem. Soc., Chem. Commun. 1985, 1228.
(33) Lindsey, J. S.; Wagner, R. W. J. Org. Chem. 1989, 54, 828.
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
803

Table 2. Effects of different acids on the pyrrole-benzaldehyde condensationa
relative amounts (%) by GCe
acid (pKa)
solubilityb
darknessc
unreacted benzaldehyded (%)
1
2
3
otherf
Brønsted Acids
0.1-0.5
0.1-0.5
0.1-0.5
>50
trifluoroacetic (0.3)
trichloroacetic (0.7)
oxalic (1.2)
taurine (1.5)
malonic (2.8)
formic (3.8)
soluble
soluble
partially
insoluble
partially
soluble
++++
+++
+++
-
84.7
86.6
84.1
<1
1.5
1.5
12.1
9.9
11.6
n.o.
14.8
2.0
1.7
2.0
2.9
2.8
2.6
3.3
n.o.
n.o.
1.4
n.o.g
1.2
+++
+
0.1-0.5
>50
80.4
13.1
<2
n.o.
n.o.
n.o.
acetic (4.8)
NH4Cl (9.3)
soluble
insoluble
-
-
>50
>50
n.o.
n.o.
<1
Lewis Acids
0.1-0.5
0.1-0.5
0.1-0.5
0.5-1
Yb(OTf)3
InCl3
Sc(OTf)3
MgBr2
CeCl3
partially
insoluble
insoluble
insoluble
insoluble
+
75.4
88.0
71.2
82.7
<2
20.1
1.1
3.1
1.3
5.4
n.o.
3.4
1.4
4.3
1.8
n.o.
+
7.5
23.2
4.0
++
+
-
>50
n.o.
a Condensations were performed with a 100:1 ratio of pyrrole and benzaldehyde at room temperature for 1 h. b Acid solubility was assessed visually. c Relative
darkening of the reaction mixture: ++++, dark brown; +++, light brown; ++, yellow; +, pale yellow; -, colorless. d Unreacted benzaldehyde (%) was obtained
from TLC using standards of known concentration (lower limit of detection is 0.1%; upper range is 50%). e Relative percentage obtained from GC analysis of the
reaction mixture. The amount of benzaldehyde is omitted; hence the values may not sum to 100%. f Refers to “other nonvolatile components” as described in the text.
g Not observed.
While hard to quantify, darkening of the reaction medium
signals formation of materials that diminish the yield and
complicate the purification.
amounts of N-confused dipyrromethane 2. The general trends
observed in the Lewis or Brønsted acid-catalyzed reactions
mirror those observed in our more limited studies with TFA
and BF3‚O(Et)2.16
Among all the acids examined, InCl3 and MgBr2 gave
the best results, considering both the product distribution and
relative absence of darkening of the reaction mixture. MgBr2
gave fewer byproducts compared to other Lewis acids
(although the reaction rate was slower than that of InCl3),
and unreacted benzaldehyde was observed by GC and TLC
analysis. We decided to focus on the use of InCl3 as a catalyst
for the pyrrole-aldehyde condensation.
B. Concentration Effects on Reaction Rate and Product
Distribution. We examined the effects of the pyrrole:benz-
aldehyde ratio (from 25:1 to 400:1) using InCl3 as catalyst
in reactions at room temperature. GC analysis was carried
out to monitor the progress of the reaction and determine
the relative yields of 1, 2, 3, and other components.
Altering the ratio of pyrrole:benzaldehyde caused two
effects: a change in product composition and a change in
reaction rate. The effect of the pyrrole:benzaldehyde ratio
on the product distribution is shown in the reaction time-
courses in Figure 4A-D. The reaction with 25 equiv of
pyrrole was complete within 10 min (not shown). The
reaction with 50 equiv of pyrrole was complete within 30
min. With 200 equiv of pyrrole, the reaction required ∼2 h
to level off, while the reaction with 400 equiv of pyrrole
was still proceeding at 3 h, and the benzaldehyde was
completely consumed only after ∼24 h. Thus, an increase
in the excess of pyrrole caused a decrease in the reaction
rate.
II. Survey of Acids. A. Effects of Acid. A series of acids
was examined in the one-flask synthesis of 5-phenyldipyr-
romethane. The acids include eight Brønsted acids with a
range of pKa values from ∼0.3 (TFA) to 9.3 (NH4Cl), and
five Lewis acids [Yb(OTf)3, InCl3, Sc(OTf)3, MgBr2, and
CeCl3] that span a range of hard/soft acids. Each condensa-
tion was performed with a 50:1 or 100:1 ratio of pyrrole:
benzaldehyde at room temperature for 30 min or 1 h. Samples
were analyzed by TLC to quantitate the amount of unreacted
benzaldehyde, by GC to assess the relative amount of 1, 2,
3, and other species, and by visual inspection to gauge the
amount of darkening of the reaction medium. The ideal acid
would give negligible unreacted benzaldehyde, a high yield
of 1, and no darkening of the reaction medium.
The acids examined and the observations for the 100:1
reaction are listed in Table 2. Similar trends were obtained
with the 50:1 pyrrole:benzaldehyde reaction, but the yield
was lower, and isolation of the dipyrromethane by crystal-
lization was more difficult (see Supporting Information).
Several Brønsted acids (TFA, trichloroacetic acid, oxalic
acid, and malonic acid) readily catalyzed formation of the
dipyrromethane (>80% by GC) with near complete con-
sumption of benzaldehyde. Weaker Brønsted acids showed
lower reactivity as a catalyst. Brønsted acids that gave good
yields of dipyrromethanes generally also gave a substantial
amount (>10%) of tripyrrane 3, negligible (<1.5%) N-
confused dipyrromethane 2, and darkened reaction mixtures.
The direct crystallization of the dipyrromethane from dark-
ened reaction mixtures was typically quite problematic.
On the other hand, the Lewis acids examined herein that
gave good yields of dipyrromethanes generally gave little
tripyrrane 3, substantial amounts of N-confused dipyr-
romethane 2, and little darkening of the reaction mixture. It
is noteworthy that Yb(OTf)3 and Sc(OTf)3 give significant
We attribute the decline in rate with increasing excess
pyrrole to the dilution of the acid catalyst. The amount of
InCl3 was kept constant (0.1 equiv relative to the benzalde-
hyde) across the varying pyrrole:benzaldehyde ratios (Figure
4, A-D). Because pyrrole is the solvent, the acid concentra-
tion declined from 54 to 3.6 mM (and the benzaldehyde
804
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

Figure 4. Effect of pyrrole:benzaldehyde ratio on the reaction rate and product distribution. The data were obtained by GC
analysis. The pyrrole:benzaldehyde ratios and the amount of InCl3 are specified in panels A-F. The early timepoints were at 1, 3,
and 10 min. Legend: 5-phenyldipyrromethane (1, 0); N-confused 5-phenyldipyrromethane (2, 9); 5,10-diphenyltripyrrane (3, O);
benzaldehyde (b); and the sum of all other volatile components (×).
concentration declined from 0.54 to 0.036 M) as the pyrrole:
benzaldehyde ratio increased from 25:1 to 400:1.
InCl3 (0.1 equiv) is shown in Figure 5. Upon increasing the
pyrrole from 25 to 400 equiv, the relative yield of 1 increased
steadily from 82.5 to 92.9%, 3 decreased from 7.3 to 1.4%,
and 2 declined from 5.6 to 1.9%.
The apparent clarity of the data masks underlying
complexity. Close examination of the reaction timecourses
obtained by GC analysis indicated that at least two species
exhibit unusual dynamics. The data for tripyrrane 3 from
Figure 4A are plotted at expanded scale in Figure 6, along
with those for component A (tR ) 12.9 min, m/z ) 157).
The tripyrrane 3 appears in a burst and then declines during
the course of the reaction, while component A forms initially
and then increases late in the reaction.
Given the decreased rate of reaction with increasing
pyrrole:benzaldehyde ratio, it became of interest to examine
an increased amount of acid to compensate for the diminished
reaction rate. Upon increasing the amount of the acid, the
reaction time was shortened: with 0.316 or 1 equiv of InCl3
and a 400:1 ratio of pyrrole:benzaldehyde, benzaldehyde was
consumed within 2 h or 30 min, respectively (Figure 4, E
and F). There was no significant change in the ultimate
product distribution (for a given pyrrole:benzaldehyde ratio)
(of 1, 2, and 3 with alteration of the acid concentration,
thereby indicating that approximately the same final product
distribution can be achieved at a desired rate by controlling
the acid concentration.
While the mechanistic origins of the dynamics observed
for 3 and component A are not known, interpretation of the
changes in product distribution over time must take into
consideration the changing reaction composition during the
The ultimate product distribution obtained as a function
of pyrrole:benzaldehyde ratio and a fixed concentration of
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
805

powdered NaOH (or KOH, LiOH‚H2O, or K2CO3) is added,
and the mixture is stirred for 30 min. (2) Filter. Filtration
affords recovery of most if not all of the acid catalyst and
base. (3) RemoVe pyrrole. The filtrate is placed under
vacuum, enabling removal and recovery of the excess pyrrole
(bp 131 °C). The resulting crude dipyrromethane is often
obtained as an oil. The latter is treated with a small volume
of hexanes, and the hexanes/residual pyrrole mixture is
removed under vacuum, which affords the solid dipyr-
romethane; this procedure may be repeated 2-3 times. (4)
Recrystallize. Recrystallization (aqueous ethanol) affords the
dipyrromethane as white crystals. This four-step procedure
was applied in the subsequent studies described herein that
employ Lewis acid catalysts. For reactions using a Brønsted
acid, the same general procedure was employed, although
quenching was performed by addition of TEA and the
filtration step was omitted.
IV. Comparison of Catalysis by InCl3 or TFA. The
solventless synthesis has previously employed TFA as the
reaction catalyst.11,16 For comparison with catalysis by InCl3,
a number of studies that parallel those reported in Figures
4-6 for InCl3 also were carried out with TFA. The data are
provided in the Supporting Information, and the major results
are summarized as follows. The TFA-catalyzed condensa-
tions were faster than those with InCl3. All TFA-catalyzed
reactions (0.1 equiv) gave complete consumption of benzal-
dehyde within 30 min, and the reaction composition remained
unchanged thereafter. Regardless, the two dominant effects
observed with InCl3 also were observed with TFA cataly-
sis: (1) An increase in the pyrrole:benzaldehyde ratio caused
the reaction rate to decrease. (2) Upon increasing the pyrrole:
benzaldehyde ratio from 25 to 400, the yield of 1 increased
from 65 to 93%, the yield of 3 decreased, and the yield of
2 remained constant (∼2%).
Figure 5. Expanded display showing the effect of pyrrole:
benzaldehyde ratio on the product distribution. The data were
obtained by GC analysis for the condensation with InCl3 (0.1
equiv) when the respective reactions leveled off: pyrrole:
benzaldehyde ) 25:1 or 50:1 (30 min); 100:1 (2 h); 200:1 (3 h);
or 400:1 (24 h). Legend: 5-phenyldipyrromethane (1, 0);
N-confused 5-phenyldipyrromethane (2, 9); 5,10-diphenyl-
tripyrrane (3, O); and the sum of all other volatile components
(×).
Figure 6. Relative yield versus time for two byproducts (3,
component “A”). The data were obtained by GC analysis from
the condensation of pyrrole and benzaldehyde (50:1, 0.1 equiv
of InCl3). The early timepoints were at 1, 3, and 10 min. Legend;
5,10-diphenyltripyrrane (3, O); unknown peak “A” at tR ) 12.9
min upon GC analysis ([).
One distinction between the two catalysts was observed
in the purification process. The reaction mixture obtained
upon TFA catalysis with a 200:1 or 400:1 ratio of pyrrole:
benzaldehyde was worked up as described above. After
neutralization of the acid and removal of the pyrrole,
crystallization of the crude product gave a dark brown solid.
The darkening of the product could not be remedied by
repeated crystallization, requiring instead treatment with
charcoal in hot ethanol (or, on a small scale, silica pad
filtration) followed by recrystallization. By contrast, the
reaction with InCl3 typically gave the dipyrromethane as a
colorless product.
V. Scope of Application. General Method. The refined
synthesis conditions using InCl3 were applied to the series
of aldehydes shown in Scheme 2. We chose a pyrrole:alde-
hyde ratio of 100:1 as a compromise among three factors
affected by an increasing pyrrole:aldehyde ratio (as observed
for benzaldehyde): (i) the increased yield of dipyrromethane,
(ii) the decreased rate of reaction, and (iii) the necessity to
handle a large amount of pyrrole relative to the amount of
product. Each reaction was performed at room temperature
with 50 mmol of aldehyde and 0.1 equiv of InCl3. The
progress of the reactions was monitored by TLC and GC
analysis. The four-step procedure outlined above was em-
course of the condensation. The reaction with a 100:1 ratio
of pyrrole:aldehyde has initial concentrations of 0.14 M for
the aldehyde, 14 M for pyrrole, and 0.014 M for the acid.
At the end of the reaction the pyrrole concentration is
essentially unchanged, while the concentration of water is
equal to that of the initial aldehyde (0.14 M), assuming
complete reaction. Thus, as the reaction progresses, the
increasing concentration of water is likely to impact the
nature of the acid catalysis. It has been shown that ca-
talysis by lanthanides and related metals entails both
Lewis acid catalysis and Brønsted acid catalysis, where the
latter is derived by interaction of water with the Lewis
acid.21,34
III. Workup Method and Purification. We developed
a four-step workup procedure that is simple, scalable, and
has minimal environmental impact. The four steps are as
follows: (1) Quench. To terminate the condensation, excess
(34) Barrett, A. G. M.; Braddock, D. C.; Henschke, J. P.; Walker, E. R. J. Chem.
Soc., Perkin Trans. 1 1999, 873.
806
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

Scheme 2
very slowly. A survey of acid catalysts and acid concen-
tration (0.1 to 1.0 equiv) was performed for the reaction
with a 100:1 ratio of pyrrole:mesitaldehyde. The conditions
that gave the best result for each acid are summarized
in Table 4. The GC traces generally mirrored those ob-
tained for benzaldehyde, including the appearance of
5-mesityldipyrromethane (10) and peaks assigned to N-
confused 5-mesityldipyrromethane (11), 5,10-dimesityltripyr-
rane (12), and isomeric tripyrranes. The structures are
shown in Chart 5. A representative GC trace and complete
results of the acid survey are shown in the Supporting
Information.
GC examination showed generally high relative yields of
5-mesityldipyrromethane (10) with each of the acids. How-
ever, the Brønsted acid catalysts gave dark brown-black
reaction mixtures, while those with Lewis acid catalysts were
only slightly darkened. In general, the extent of darkening
of the reaction mixtureswhile difficult to quantify and
stemming from products that are difficult to analyzesis the
likely cause of the yield of isolated dipyrromethane falling
short of that expected on the basis of GC analysis. A case
in point is provided by the reaction obtained with TFA: the
relative percentage of 10 is 77% (by GC analysis), while
the isolated yield is 27%.16 From these studies, we focused
on the Lewis acid catalysts. The product distribution obtained
with Sc(OTf)3 contained a significant amount of the N-
confused dipyrromethane 11. The reaction with InCl3 at 60
°C afforded a relatively good yield of the dipyrromethane
10, but crystallization to high purity proved difficult. The
best results overall were observed with MgBr2, which also
gave the least darkened reaction mixture. The isolated yield
was 53% (93.0% purity), nearly double that obtained
previously.
VI. Scale-up. A large-scale synthesis of 5-phenyldipyr-
romethane (1) was carried out using benzaldehyde (79.8 g)
and 100 equiv of pyrrole (5.20 L), catalysis by InCl3 (16.7
g), and mechanical stirring for 1.5 h at room temperature.
The reaction mixture remained essentially at room temper-
ature without application of external temperature control. The
reaction was quenched by adding NaOH, and the mixture
was filtered and washed with pyrrole (0.35 L). The mixture
was concentrated, and pyrrole was recovered (5.30 L). The
crude dipyrromethane solidified upon treatment with hexanes
and was recrystallized from ethanol/water (4:1), affording
four crops of crystals. The third and fourth crop were
combined and recrystallized. The three samples each gave
satisfactory elemental analyses, yet the GC analyses indicated
purities of 94.8, 92.3, and 94.0%. The typical GC data (crop
2) indicated the relative amounts of 92.3% (1), 6.3% (2),
ployed for purification. For comparison, the same reactions
were carried out with 20 mmol of the aldehyde, and
purification was carried out by column chromatography.
The results for dipyrromethanes 1 and 5-9 are sum-
marized in Table 3. The general reaction conditions and
workup procedure generally gave good results with each of
the aldehydes. Each dipyrromethane gave a relative GC
purity of >90%. Minor modifications were made for
dipyrromethanes 6 and 8. In the case of 6 a small amount
(∼7%) of transesterification product was obtained when
aqueous ethanol was employed for recrystallization. Hence,
aqueous methanol (10:1) was used for the recrystallization.
Dipyrromethane 8, an oil, was obtained by column chroma-
tography. More extensive changes were made for the
reactions with paraformaldehyde and mesitaldehyde.
Special Cases. A. Paraformaldehyde. Paraformaldehyde
was relatively insoluble in neat pyrrole. The heterogeneous
reaction mixture was heated at 50-55 °C for 2.5 h. Pyrrole
was removed; dipyrromethane (9) was then extracted with
hexanes/ethyl acetate (4:1). Evaporation of the solvent
followed by crystallization [methanol/water (4:1)] afforded
9 in 45% yield (94.6% purity by GC).
1
0.6% (3), and 0.8% (other volatile components). H NMR
analysis showed a corresponding ratio of 1:2:3, while other
components were not observed. Altogether, 115.5 g of 1
(69% yield, 94.2% purity) was obtained. Distillation of the
recovered pyrrole afforded 4.92 L of pyrrole (90% final
recovery).
VII. Material Recovery and Reuse. Filtration of the
reaction mixture affords recovery of the indium employed
B. Mesitaldehyde. The reaction of mesitaldehyde with
pyrrole in the presence of 0.1 equiv of InCl3 proceeded
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
807

Table 3. Scope of application of the refined synthesis methoda
a Condensations were performed with a 100:1 ratio of pyrrole:aldehyde with 0.1 equiv of InCl3 at room temperature for 1.5 h unless specified. b Relative purity (%)
based on GC peak areas is shown in parentheses. The workup procedure involves quenching with powdered NaOH, filtration, removal of pyrrole, and entrainment with
hexanes unless specified, followed by crystallization or chromatography. c May include a first and second crop of crystals. d Obtained as an oil. e Not applicable.
f After reaction at 55 °C for 2.5 h, the reaction mixture was extracted with hexanes/ethyl acetate prior to crystallization.
Table 4. Refined acid catalysis conditions for the synthesis of 5-mesityldipyrromethane (10)a
relative amounts (%) by GC
yield (purity) (%)b
crystallization chromatography
27e
catalyst
TFA
malonic acid
InCl3
conditions
MsCHOc
10
11
12
otherd
0.1 equiv, rt, 1.5 h
0.1 equiv, rt, 3 h
0.3 equiv, 60 °C, 5 h
0.3 equiv, rt, 4 h
0.5 equiv, rt, 1.5 h
0.2
37.5
7.6
0.8
0.9
77.1
52.7
67.9
52.7
82.9
2.1
1.1
18.3
30.2
3.5
12.5
4.2
3.3
1.6
4.8
8.1
4.5
2.9
14.7
7.9
25 (90.0)
n.a.f
n.a.
56 (82.0)
n.a.
49 (85.7)g
n.a.
Sc(OTf)3
MgBr2
53 (93.0)
65 (92.7)
a Condensations were performed with a 100:1 ratio of pyrrole:mesitaldehyde. b Relative purity (%) based on GC peak areas is shown in parentheses. c Mesitaldehyde.
d Refers to “other nonvolatile components” as described in the text. e Obtained by Kugelrohr distillation followed by crystallization using the reported method.16 f Not
applicable. g Yield of 32% (88.0%) in the first crop, 17% (81.2%) in the second crop.
as catalyst in conjunction with the excess base for neutraliza-
tion. We have not attempted to regenerate the indium catalyst.
However, the pyrrole has been recycled numerous times. One
caveat is that an aldehyde (e.g., hexanal, bp 128 °C) may
have a boiling point nearly identical with that of pyrrole (bp
131 °C). In such cases, the recovered pyrrole should be used
exclusively for subsequent reaction with the same aldehyde
to avoid any possibility of contamination by residual unre-
acted aldehyde.
were obtained with InCl3 (for benzaldehyde and other aryl
aldehydes) and MgBr2 (for mesitaldehyde). The workup
procedure entails quenching with NaOH, filtration, removal
of pyrrole, and crystallization. The simplicity of this proce-
dure (no aqueous/organic extraction, distillation, or chroma-
tography) minimizes waste and enables scalability. The
absence of any reaction solvent other than pyrrole enables
facile recovery of the excess pyrrole from the crude reaction
mixture. The recovered pyrrole can be reused. Indeed, the
reaction of 0.75 mol of benzaldehyde in 5.2 L of pyrrole
gave 115.5 g of 5-phenyldipyrromethane and recovery of
90% of the excess pyrrole in a form suitable for reuse.
Characterization of the purity of dipyrromethanes is best
accomplished using a combination of analytical techniques,
including gas chromatography. The N-confused dipyr-
romethane and the tripyrrane species are dominant byprod-
ucts of the reaction. Because the former is an isomer while
Conclusions
The one-flask solventless reaction of an aldehyde with
excess pyrrole provides a simple means of preparing the
corresponding dipyrromethane. Catalysis can be achieved
with a wide variety of Brønsted or Lewis acids. The Brønsted
acids that afford product also give a dark reaction mixture,
while the Lewis acids give less darkening. Excellent results
808
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

Chart 5
GC yield was 91.0%, and the yield upon recrystallization
was 68% (96.7% purity by GC). Given these differences,
we employed distilled pyrrole throughout our experiments.
Pyrrole (reagent grade) was distilled over calcium hydride
before use.
Quantitative TLC. Unreacted benzaldehyde was exam-
ined by TLC analysis [silica, hexanes/ethyl acetate (4:1)]
using known concentrations of benzaldehyde in ethyl acetate
as standards, a method that enables quantitative analysis.31
The standard samples corresponded to yields of unreacted
benzaldehyde of 0.1-50%. Each TLC plate was spotted with
a reaction sample and at least two standards for visual
comparison. In this manner, the yield of benzaldehyde could
be bracketed. The lower limit of detection of unreacted
benzaldehyde is 0.1% for a reaction of 100:1 pyrrole:
benzaldehyde. We made no effort to assess quantities in the
range of 50-100% of unreacted benzaldehyde; such amounts
are designated as >50% unreacted benzaldehyde.
Gas Chromatography. Gas chromatography was carried
out using an HP 6890 gas chromatograph equipped with an
FID detector and a fused silica capillary column HP-5 (30
m × 0.32 mm × 2.5 µm). All samples were examined under
the following temperature gradient: temp 1, 35 °C (5 min);
temp 2, 315 °C (12 min); rate 20 °C/min, total runtime 31
min. Relative yields were determined by direct integration
of the peak areas of the gas chromatogram rather than by
constructing calibration curves using standard solutions of
each component. We demonstrated earlier that the detector
had an approximately linear response to the amount of each
component present.16 GC-MS data were acquired using an
HP 5890 GCD gas chromatograph using the same temper-
ature gradient as described for GC analysis.
Variation in the workup process has little impact on the
product distribution. Workup of the InCl3-catalyzed conden-
sation (100:1 ratio of pyrrole:benzaldehyde at room temper-
ature for 1 h) by quenching with TEA or NaOH (or using
ethyl acetate or CH2Cl2 as the GC injection solvent) gave
nearly identical product distribution profiles. The only
differences were the appearance of several minor peaks in
the 19-24 min region. In general, the GC analyses were
performed following neutralization of the reaction mixture
(but without removal of excess pyrrole).
A control experiment where a sample of pyrrole and
benzaldehyde (100:1) was injected into the GC instrument
resulted in broad, low intensity peaks (9-11 min) as well
as a peak due to 5-phenyldipyrromethane (at ∼1% level).
The former is a GC artifact as described earlier. The latter
is not due to carryover from the prior sample. We have
previously shown that benzaldehyde and pyrrole react, albeit
slowly and in low yield, upon heating in the absence of an
acid to give the dipyrromethane.16 Thus, samples from
reaction mixtures at room temperature where no reaction
occurs can still give small amounts of the dipyrromethane
as a GC artifact.
the latter is an oligomer of the dipyrromethane, a satisfactory
elemental analysis is a necessary but insufficient criterion
of purity.
Experimental Section
1
General. H NMR (400 MHz) spectra were recorded in
CDCl3 unless noted otherwise. CH2Cl2 (ACS grade), hexanes
(ACS grade), ethyl acetate (ACS grade), and powdered
NaOH (20-40 mesh beads) were used as received. The acids
employed in this study were of the following grades (and
were used in quantities without correction for less than 100%
purity): trifluoroacetic acid (99%), trichloroacetic acid
(98%), oxalic acid (97%), taurine (99%), malonic acid (99%),
formic acid (98%), acetic acid (99.99%), NH4Cl (99.5%),
MgBr2 (98.5%), Yb(OTf)3 (99.99%), InCl3 (98%), Sc(OTf)3
(99%), and CeCl3 (99.9%). All other chemicals including
aldehydes were of reagent grade and were used as obtained.
The first-reported characterization data for dipyrromethanes
prepared by one-flask syntheses are as follows: 1,7 5,16 6,35
7,16,36 8,3 9,37 and 10.11 Each compound prepared herein was
characterized by GC, elemental analysis, mp, and 1H NMR
spectroscopy. All reported yields are based on the weight of
the isolated dipyrromethane and are not corrected for the
purity established by GC analysis.
Quality of Pyrrole. Pyrrole slowly darkens on standing.
To determine how the purity of the dipyrromethane depends
on the quality of the pyrrole employed, parallel reactions
were performed using undistilled pyrrole or distilled pyrrole
with a 100:1 ratio of pyrrole:benzaldehyde and InCl3 at room
temperature for 1 h. With undistilled pyrrole, the mixture
was pale green, the GC yield was 89.8%, and the yield upon
recrystallization was 61% (97.3% purity by GC). With
distilled pyrrole, the reaction mixture was pale yellow, the
GC-MS data were obtained with high loading of
samples to detect trace components. We were concerned that
bleeding from earlier eluting species in a given run, espe-
cially those that were (1) present in an overloaded amount
(35) Bru¨ckner, C.; Zhang, Y.; Rettig, S. J.; Dolphin, D. Inorg. Chim. Acta 1997,
263, 279.
(36) Wagenknecht, H.-A.; Claude, C.; Woggon, W.-D. HelV. Chim. Acta 1998,
81, 1506.
(37) Wang, Q. M.; Bruce, D. W. Synlett 1995, 1267.
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
809

and (2) isomeric with later-eluting species could cause
misleading assignment of the later-eluting species. Assign-
ments were made by comparing the intensities of given mass
spectral peaks as a function of time during which a com-
ponent eluted to ensure that the observed spectral signature
originated from the component rather than from earlier-
eluting species.
Survey of Acid Catalysts. Analytical-scale reactions of
benzaldehyde (1.50 mmol) and pyrrole (150 mmol) in the
presence of each catalyst (0.15 mmol) were performed at
room temperature. Reactions were quenched after 1 h.
Quantitative TLC analysis was performed as described above.
GC analysis was done by removing 200-µL aliquots of the
reaction mixture after quenching and then diluting with 800
µL of CH2Cl2.
Effect of Pyrrole:Benzaldehyde Ratio and Concentra-
tion of Acids. Studies on the effect of the pyrrole:benzal-
dehyde ratio and the amount of added acid (TFA or InCl3)
on the product distribution were performed as illustrated
for the following typical example (pyrrole:benzaldehyde
100:1): A mixture of pyrrole (26.8 g, 400 mmol) and
benzaldehyde (424 mg, 4.00 mmol) was treated with 0.1
equiv of InCl3. After 1, 2, 3, and 24 h, 250-µL aliquots from
the reaction mixture were removed and immediately quenched
by injection into a 500-µL solution of ethyl acetate/TEA
(25:1) and then filtered (in the case of InCl3). These samples
were examined by GC analysis as described above.
Found: C, 81.04; H, 6.41; N, 12.67]. The filtrate from the
first crop was concentrated, yielding a second crop of pale
brown crystals (1.22 g, 11% yield; 86.3% purity by GC).
The latter was recrystallized [18 mL of ethanol/water (4:1)]
affording yellow crystals (1.05 g, 10% yield; 92.0% pur-
ity by GC), which were recrystallized [16 mL, 2-propanol/
water (15:1)], affording pale yellow crystals (0.66 g, 6%
yield; mp 99-100 °C; 96.8% purity by GC; Anal. Calcd
for C15H14N2: C, 81.05; H, 6.35; N, 12.60. Found: C, 80.90;
H, 6.36; N, 12.57). The first crop and the final crop were
combined, affording pale yellow crystals (7.88 g, 71% yield;
calcd 96.7% purity by GC).
Notes:
(i) It is essential to quench the reaction with a strong base
(e.g., NaOH) and to stir the reaction mixture for a period
thereafter (45 min). When the reaction mixture was quenched
with an organic base (e.g., TEA) or not quenched at all,
filtration still removed much of the indium material. How-
ever, the filtrate contained the crude dipyrromethane as well
as a blue contaminant. The blue material proved very difficult
to remove from the dipyrromethane by crystallization or
charcoal but could be removed by flash chromatography.
(ii) It is essential that the crude dipyrromethane be
obtained as a solid largely free of pyrrole prior to initiating
crystallization.
(iii) Attempts to crystallize the crude product in other
solvents [e.g., CH2Cl2/hexanes, ethyl acetate/hexanes, ether/
hexanes, THF/water] were not successful.
General Procedure Illustrated for 5-phenyldipyr-
romethane (1). A. Crystallization Method. Pyrrole (347 mL,
5.00 mol) and benzaldehyde (5.31 g, 50.0 mmol) were added
to a 500-mL single-neck round-bottomed flask containing a
magnetic stir bar. The solution was degassed with a stream
of argon for 10 min. InCl3 (1.11 g, 5.00 mmol) was added,
and the mixture was stirred under argon at room temperature
for 1.5 h. The mixture turned yellow during the course of
the reaction. NaOH (6.0 g, 0.15 mol, 20-40 mesh beads)
was added to quench the reaction (see note i below). Stirring
for 45 min afforded a pale yellow mixture. The mixture was
filtered (Fisher brand qualitative filter paper) using a Buchner
funnel. The contents of the flask and the filtered material
were washed with a small amount of pyrrole. The filtrate
was concentrated using a rotary evaporator under vacuum
(0.2 mmHg); the evaporation flask was warmed at 40-50
°C, and the collection trap was cooled with a dry ice-acetone
bath. The collected pyrrole was set aside. Traces of pyrrole
were removed by entrainment with hexanes by performing
the following procedure three times: the crude viscous
residue in the evaporation flask was triturated with hexanes
(50 mL), and the volatile components were evaporated (see
note ii below). The resulting yellow solid was dissolved in
60 mL of ethanol/water (4:1) (see note iii below) with heating
by a water bath. The resulting yellow solution containing a
very small amount of white powder was filtered (Fisher brand
qualitative filter paper), and the filtrate was set aside
overnight at room temperature (see note iv below), affording
a first crop of pale yellow crystals [7.22 g, 65% yield; mp
99-100 °C (lit.7 mp 102-103 °C); 96.7% purity by GC;
Anal. Calcd for C15H14N2: C, 81.05; H, 6.35; N, 12.60.
(iv) Attempts to perform the crystallization at lower
temperature gave less pure product.
B. Column Chromatography Method. Pyrrole (139 mL,
2.00 mol) and benzaldehyde (2.13 g, 20.1 mmol) were
condensed in the presence of 0.1 equiv of InCl3 in a man-
ner similar to that described above. The crude product
obtained after removal of pyrrole was purified by column
chromatography [silica, 6 cm dia. × 14 cm, hexanes/
CH2Cl2/ethyl acetate (7:2:1)]. The crude mixture was dis-
solved in CH2Cl2, silica gel was then added, and the mixture
was evaporated to dryness. The resulting powder was loaded
on the top of the column. Elution (400-1400 mL of eluant)
followed by concentration of the eluted product gave a white
solid (3.64 g, 82% yield; mp 100 °C; >99% purity by GC;
Anal. Calcd for C15H14N2: C, 81.05; H, 6.35; N, 12.60.
Found: C, 80.94; H, 6.39; N, 12.55). Note that chromato-
graphic purification of the crude product using common
solvents such as CH2Cl2 or hexanes/CH2Cl2 (1:1) was
unsuccessful. Even careful chromatography on silica with
hexanes/ethyl acetate (4:1) gave mixtures of the three
components although this solvent system happened to be a
good solvent for TLC analysis.
Applications. 5-(4-Methoxyphenyl)dipyrromethane (5).
(A) The reaction of p-anisaldehyde (6.81 g, 50.0 mmol) with
crystallization [ethanol/water (4:1)] afforded two crops of
crystals: (crop 1) light brown crystals [5.17 g, 41% yield;
mp 95-96 °C (lit.16 mp 99 °C); 93.7% purity by GC; Anal.
Calcd for C16H16N2O: C, 76.16; H, 6.39; N, 11.10. Found:
C, 75.75; H, 6.40; N, 10.98]; (crop 2) brown crystals (2.82
g, 22% yield; 91.0% purity by GC). Combination and
810
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

recrystallization afforded light brown crystals (5.42 g, 43%
yield; mp 98-99 °C; 97% purity by GC; Anal. Calcd for
C16H16N2O: C, 76.16; H, 6.39; N, 11.10. Found: C, 76.03;
H, 6.36; N, 11.08).
(B) Reaction at the 20.0-mmol scale with chromatography
[silica, hexanes/CH2Cl2/ethyl acetate (16:3:1)] afforded a
yellow solid (3.12 g, 62% yield; mp 99 °C; >99% purity by
GC; Anal. Calcd for C16H16N2O: C, 76.16; H, 6.39; N, 11.10.
Found: C, 76.06; H, 6.43; N, 10.99).
stirred for 1 h and then filtered. The filtrate was concentrated,
and the pyrrole was recovered. The crude solid obtained after
removing pyrrole was extracted with 20% ethyl acetate/
hexanes (5 × 50 mL). The solvent was evaporated. Crystal-
lization [methanol/water (4:1)] afforded pale white crystals
[3.29 g, 45% yield; mp 70-71 °C (lit.37 mp 74 °C); 94.6%
purity by GC; Anal. Calcd for C9H10N2: C, 73.94; H, 6.89;
N, 19.16. Found: C, 74.00; H, 6.81; N, 19.15].
(B) The same procedure at the 20.0-mmol scale with chro-
matography [silica, hexanes/CH2Cl2/ethyl acetate (7:2:1)]
afforded a white solid (1.84 g, 63% yield; mp 75 °C; >99%
purity by GC; Anal. Calcd for C9H10N2: C, 73.94; H, 6.89;
N, 19.16. Found: C, 74.18; H, 6.92; N, 19.14).
5-Mesityldipyrromethane (10). (A) A mixture of mesi-
taldehyde (7.42 g, 50.0 mmol) and pyrrole (347 mL, 5.00
mol) in a 500-mL single-neck round-bottomed flask was
degassed with a stream of argon for 10 min. MgBr2 (4.60 g,
25.0 mmol) was added, and the mixture was stirred for 1.5
h at room temperature. The tan mixture was treated with
powdered NaOH (10.0 g, 0.25 mol). The mixture was stirred
for 1 h and then filtered. The filtrate was concentrated, and
the pyrrole was recovered. The crude solid obtained upon
removal of pyrrole was extracted with 20% ethyl acetate/
hexanes (7 × 100 mL). The extract was gravity-filtered
through a pad of silica (80 g). The eluted solution was
concentrated to obtain a yellow solid. Crystallization [ethanol/
water (4:1)] afforded pale yellow crystals [7.00 g, 53% yield;
93% purity by GC; mp 166-167 °C (lit.11 mp 166-167 °C);
Anal. Calcd for C18H20N2: C, 81.78; H, 7.63; N, 10.60.
Found: C, 81.16; H, 7.67; N, 10.19].
5-(4-Carbomethoxyphenyl)dipyrromethane (6). (A) The
reaction of methyl 4-formylbenzoate (8.21 g, 50.0 mmol)
with crystallization [methanol/water (10:1)] afforded two
crops of crystals: (crop 1) pale yellow crystals [7.84 g, 56%
yield; mp 162-163 °C (lit.35 mp 158 °C); 97.2% purity by
GC; Anal. Calcd for C17H16N2O2: C, 72.84; H, 5.75; N, 9.99.
Found: C, 72.59; H, 5.75; N, 9.79]; (crop 2) pale yellow
crystals (2.66 g, 19% yield; mp 161-162 °C; 96.5% purity
by GC; Anal. Calcd for C17H16N2O2: C, 72.84; H, 5.75; N,
9.99. Found: C, 72.34; H, 5.83; N, 9.90). Altogether, 10.5
g (75% yield; calcd 97.0% purity by GC) was obtained.
(B) Reaction at the 20.0-mmol scale with chromatography
[silica, hexanes/CH2Cl2/ethyl acetate (7:2:1)] afforded a white
solid (4.20 g, 75% yield; mp 159-160 °C; >99% purity by
GC; Anal. Calcd for C17H16N2O2: C, 72.84; H, 5.75; N, 9.99.
Found: C, 72.48; H, 5.68; N, 9.74).
5-(Pentafluorophenyl)dipyrromethane (7). (A) The reac-
tion of pentafluorobenzaldehyde (9.80 g, 50.0 mmol) with
crystallization [ethanol/water (4:1)] afforded two crops of
crystals: (crop 1) pale white crystals [8.27 g, 53% yield;
mp 130-131 °C (lit.16 mp 131-132 °C); >99% purity
by GC; Anal. Calcd for C15H9F5N2: C, 57.70; H, 2.91; N,
8.97. Found: C, 57.92; H, 2.76; N, 9.04]; (crop 2) pale
white crystals (4.05 g, 26% yield; mp 129-130 °C;
>99% purity by GC; Anal. Calcd for C15H9F5N2: C, 57.70;
H, 2.91; N, 8.97. Found: C, 57.66; H, 2.82; N, 8.84).
Altogether, 12.3 g (79% yield; calcd >99% purity by GC)
was obtained.
(B) The same procedure was performed at the 20.0-mmol
scale. Following silica-pad filtration, chromatography [silica,
toluene/CH2Cl2 (1:1)] afforded a yellow solid (3.43 g, 65%
yield; mp 165-166 °C; 92.7% purity by GC; Anal. Calcd
for C18H20N2: C, 81.78; H, 7.63; N, 10.60. Found: C, 81.89;
H, 7.66; N, 10.36).
Large-Scale Synthesis of 5-Phenyldipyrromethane (1).
Pyrrole (5.20 L, 75.0 mol; recycled) and benzaldehyde (79.8
g, 0.752 mol) were added to a 12-L two-neck round-bottomed
flask equipped with a mechanical stirrer. The solution was
degassed with a stream of argon for 1 h. InCl3 (16.7 g, 75.5
mmol) was added, and the mixture was stirred under argon
at room temperature for 1.5 h. The mixture was yellow during
the course of the reaction. The reaction was quenched by
the addition of NaOH (90.1 g, 2.25 mol; 20-40 mesh beads).
The mixture was stirred for 1.5 h. The stirrer was stopped,
and the mixture was allowed to settle for 1 h. The mixture
was filtered (Fisher brand qualitative filter paper) using a
Bu¨chner funnel. The flask and filtered material on the funnel
was washed with pyrrole (350 mL). The filtered material
after drying under vacuum with warming for 4 h constituted
118 g. The filtrate was concentrated using a rotary evaporator
under vacuum (0.2 mmHg) at 40-60 °C (the trap was cooled
in a dry ice-acetone bath). The amount of recovered pyrrole
was 5.30 L. A crude viscous residue was left in the
evaporation flask. The crude viscous residue was treated with
hexanes to remove traces of pyrrole in the following manner
four times: hexanes (450 mL) was added, and the volatile
(B) Reaction at the 20.0-mmol scale with chromatography
twice [silica, hexanes/CH2Cl2/ethyl acetate (10:2:1)] afforded
a white solid (5.00 g, 80% yield; mp 131 °C; 98.0% GC
purity; Anal. Calcd for C15H9F5N2: C, 57.70; H, 2.91; N,
8.97. Found: C, 57.64; H, 3.02; N, 8.95).
5-(Pentyl)dipyrromethane (8). The reaction of hexanal
(5.01 g, 50.0 mmol) was performed in the standard way.
The viscous liquid obtained upon removal of pyrrole was
chromatographed twice [silica, hexanes/CH2Cl2/ethyl acetate
(16:3:1); hexanes/ethyl acetate (9:1)], affording a yellow oil
(6.70 g, 62% yield; 94.3% purity by GC; Anal. Calcd for
C14H20N2: C, 77.73; H, 9.32; N, 12.95. Found: C, 77.63;
H, 9.48; N, 12.71).
Dipyrromethane (9). (A) A mixture of paraformaldehyde
(1.50 g, 50.0 mmol) and pyrrole (347 mL, 5.00 mol) in a
500-mL flask was degassed with a stream of argon for 10
min at room temperature. The mixture was heated at 55 °C
for about 10 min under argon to obtain a clear solution. InCl3
(1.11 g, 5.00 mmol) was then added, and the mixture was
stirred at 55 °C for 2.5 h. The heat source was removed,
and NaOH (6.0 g, 0.15 mol) was added. The mixture was
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
811

components were removed under vacuum. The crude product
was obtained as a pale yellow solid (87.0% GC purity;
containing 7.9% of N-confused dipyrromethane 2 and 1.0%
of tripyrrane 3). Recrystallization [ethanol/water (4:1), 1 L]
afforded a first crop of pale yellow crystals [98.9 g; mp
100-101 °C (lit.7 mp 102-103 °C); 94.8% purity by GC;
Anal. Calcd for C15H14N2: C, 81.05; H, 6.35; N, 12.60.
Found: C, 80.84; H, 6.40; N, 12.62]. The filtrate from the
first crop was concentrated to dryness. The residue was
recrystallized, affording a second crop of yellow crystals
(6.00 g; mp 98-99 °C; 92.3% purity by GC; Anal. Calcd
for C15H14N2: C, 81.05; H, 6.35; N, 12.60. Found: C, 80.84;
H, 6.37; N, 12.57). The filtrate from the second crop was
concentrated to dryness, and the residue was recrystallized,
affording a third crop (14.1 g; 88.7% purity by GC). The
filtrate from the third crop was concentrated to dryness, and
the residue was recrystallized, affording a fourth crop (3.33
g; 83.1% purity by GC) of crystals. Crops three and four
were combined and recrystallized [ethanol/water (4:1)],
affording pale brown crystals (10.6 g; mp 99-100 °C; 94.0%
purity by GC; Anal. Calcd for C15H14N2: C, 81.05; H, 6.35;
N, 12.60. Found: C, 80.88; H, 6.39; N, 12.63). Altogether,
115.5 g of the title compound was obtained (69% yield; calcd
94.2% purity by GC).
The filtered material (118 g) from the crude reaction
mixture exceeds the mass of the InCl3 (16.7 g) and NaOH
(90.1 g); the difference is attributed to water of hydration
(maximally ca. 13.5 g originating from the condensation).
The recovered liquid (5.30 L), which contains predominantly
pyrrole and residual water, was distilled. The recovered
pyrrole (4.92 L, 90% recovery based on unreacted pyrrole)
was suitable for reuse.
Acknowledgment
This research was supported by a grant from the NIH
(GM36238). Mass spectra were obtained at the Mass Spec-
trometry Laboratory for Biotechnology. Partial funding for
the Facility was obtained from the North Carolina Biotech-
nology Center and the National Science Foundation.
Supporting Information Available
Description of analytical issues including TLC and
elemental analysis; data for the pyrrole-mesitaldehyde
reaction, and data for TFA catalysis. This material is available
free of charge via the Internet at http://pubs.acs.org.
Received for review June 25, 2003.
OP034083Q
812
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development