Concerning the mechanism of the Friedlaender quinoline synthesis

461

Concerning the mechanism of the Friedländer

quinoline synthesis

Joseph M. Muchowski and Michael L. Maddox

Abstract: Detailed experiments regarding the mechanism of the Friedländer synthesis of quinolines from o-amino-

benzaldehydes and simple aldehydes or ketones are described. Under the basic or acidic conditions commonly used in

this reaction, it is concluded that the first step involves a slow intermolecular aldol condensation of the aldehyde or

ketone with the o-aminobenzaldehyde. The aldol adduct 5 generated in this manner then undergoes very rapid cycli-

zation to 4, which subsequently loses water to produce the quinoline derivative 8. Both 5 and 4 are too short lived to

be detectable (TLC), even when deliberately generated by other means. It is also shown that E-enones corresponding to

6, i.e., the aldol dehydration product, are converted into quinolines (e.g., 21a and 21b from 17a and 17b) under basic

or acidic conditions. Such enones are not detected as intermediates in the base-induced Friedländer synthesis, even

though certain congeners (17b) would be easily observable. Under acidic conditions these enones are too short lived to

be detectable. Schiff bases derived from 2-aminobenzaldehyde (18a) and aldehydes or ketones can be generated under

special conditions, but they show reactivity patterns different from those seen in the usual Friedländer condensations.

Thus, the ytterbium-triflate-catalyzed reaction of aldehydes with 18a at room temperature in toluene generates the E-

Schiff bases (33, R¹= H), from which isomeric mixtures of tetrahydroquinoline derivatives 26 are formed exclusively.

At higher temperatures, the E-Schiff bases 33 are isomerized to the Z-Schiff bases 34, from which the 3-substituted

quinoline derivatives 24 are formed as the major products under appropriate conditions. Also, the ytterbium-triflate-

catalyzed reaction of 18a with the pyrrolidine enamines of the methyl-n-alkylketones 38a,b produces mixtures in which

the 2-monosubstituted kinetic products 37b,d predominate over the 2,3-disubstituted thermodynamic products 21c,e by

a factor of 4:1 to 5:1. These results are opposite to those observed under the usual basic or acidic Friedländer reactions

with methyl-n-alkylketones, where the thermodynamic products are usually strongly favored. The unusual kinetic:ther-

modynamic product ratios observed with 38a,b are ascribed to the generation and rapid cyclization of mixtures of the

Schiff bases 35 and 36, in which the kinetic isomer 35 is highly predominant.

Key words: mechanism, Friedländer synthesis, quinolines, intramolecular aldol reactions, Schiff bases, tetrahydro-

quinolines, high kinetic:thermodynamic product ratios. 478

Résumé : On rapporte les détails d’expériences réalisées dans le but d’élucider le mécanisme de la synthèse des quino-

léines de Friedländer à partir d’o-aminobenzaldéhydes et d’aldéhydes ou des cétones simples. On en déduit que, dans

les conditions légèrement acides ou basiques utilisées pour cette réaction, la première étape implique une condensation

aldolique intermoléculaire lente de l’aldéhyde ou de la cétone avec l’o-aminobenzaldéhyde. L’adduit aldolique ainsi gé-

néré, 5, subit alors une cyclisation très rapide en 4 qui perd subséquemment de l’eau pour conduire au dérivé 8 de la

quinoléine. Les temps de vie des intermédiaires 5 et 4 sont tous les deux beaucoup trop courts pour qu’on puisse les

détecter (CCM), même lorsqu’on tente de les générer par d’autres méthodes. On a aussi montré que les E-énones cor-

respondant à 6, c’est-à-dire le produit de déshydratation de l’aldol, peuvent être transformées en quinoléines (par

exemple, 21a et 21b à partir de 17a et 17b) dans des conditions acides ou basiques. Ce type d’énone n’est pas détecté

comme intermédiaire dans la synthèse de Friedländer induite par les bases même si certains congénères (17b) seraient

facilement observables. Dans des conditions acides, les temps de vie de ces énones sont trop courts pour qu’on puisse

les détecter. Les bases de Schiff du 2-aminobenzaldéhyde (18a) avec les aldéhydes ou les cétones peuvent être géné-

rées dans des conditions spéciales, mais leurs patrons de réactivité sont différents de ceux observés dans les condensa-

tions de Friedländer habituelles. Ainsi, la réaction catalysée par le triflate d’ytterbium des aldéhydes avec 18a, à la

température ambiante et dans le toluène, conduit à la formation de bases de Schiff de stéréochimie E- (33, R¹= H) à

partir desquelles il se forme uniquement des mélanges isomères de dérivés tétrahydroquinoléines 26. À des températu-

res plus élevées, les bases de Schiff de stéréochimie E- (33) s’isomérisent en bases de stéréochimie Z- (34) à partir

desquelles, dans des conditions appropriées, les produits qui se forment d’une façon prépondérante sont des dérivés de

la quinoléine substitués en position 3 (24). De plus, la réaction catalysée par le triflate d’ytterbium du produit 18a avec

Received 3 June 2003. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 27 February 2004.

Dedicated to Professor Victor Snieckus in recognition of his numerous fundamental contributions to heteroatom directed

metalations.

J.M. Muchowski¹and M.L. Maddox. Chemistry, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94304, U.S.A.

¹Corresponding author (e-mail: joseph.muchowski@roche.com).

Can. J. Chem. 82: 461–478 (2004)

doi: 10.1139/V03-211

462

Can. J. Chem. Vol. 82, 2004

des énamines formées par réaction de la pyrrolidine sur les méthylalkylcétones 38a,b conduit à la formation de mélan-

ges dans lesquels les produits cinétiques portant un substituant en position 2 (37b,d) prédominent par rapport aux pro-

duits thermodynamiques disubstitués en 2,3 (21c,e) par un facteur allant de 4:1 à 5:1. Ces résultats sont opposés à

ceux observés dans les conditions usuelles des réactions de Friedländer acide ou basique des méthylalkylcétones dans

lesquelles les produits thermodynamiques sont généralement fortement favorisés. Les rapports inhabituels de produits

cinétiques et thermodynamiques observés avec les produits 38a,b sont attribués à la génération et à la cyclisation ra-

pide de mélanges des bases de Schiff 35 et 36 dans lesquels l’isomère cinétique 35 est fortement prédominant.

Mots clés : mécanisme, synthèse de Friedländer, quinoléines, réactions aldoliques intramoléculaires, bases de Schiff, té-

trahydroquinoléines, rapport élevé des rapports de produits cinétiques et thermodynamiques.

[Traduit par la Rédaction] Muchowski and Maddox

low). These entities, which must exist in the vinylogously

Introduction

stabilized forms 11, have been cyclized under both basic (8,

10) and acidic (9) conditions. No Schiff base derived from a

simple aldehyde or ketone and an aromatic o-aminocarbonyl

compound has ever been isolated.

The formation of a quinoline derivative by the condensa-

tion of an aromatic o-aminoaldehyde or o-aminoketone with

an aldehyde or a ketone containing a methylene group alpha

to the carbonyl moiety, under basic or acidic conditions, is

known as the Friedländer synthesis (1–5). The development

of the heteroatom-directed o-metalation of aromatic systems

(6), which has provided facile access to a variety of o-

aminoaldehydes and o-aminoketones, has made the

Friedländer reaction one of the most reliable of all quinoline

syntheses. Basic reaction conditions are particularly efficient

for the generation of quinoline derivatives from ketones and

o-aminoaldehydes, whereas the condensation of ketones

with o-aminoketones is best effected in acidic media. The

most notable advantage of the Friedländer synthesis is that

the positions of the substituents in the homocyclic ring are

fixed by the structure of the o-aminocarbonyl component,

and the regiochemistry of the substituents in the heterocyclic

ring is predictable because nonsymmetrical ketones usually

give rise to the quinoline derived from the thermodynamic

enolate as the major or exclusive product.

Although the general features of the Friedländer synthesis

are well understood, there is still a considerable degree of

uncertainty regarding the mechanistic details of the process.

Two distinctly different reaction pathways have been

suggested (4, 5) to account for the facts. One involves the

rate-determining formation of the Schiff base 3 (Scheme 1),

followed by an intramolecular aldol-type condensation to the

hydroxyimine 4 and subsequent loss of water therefrom to

produce the quinoline 8. In the other proposal, the rate-

limiting step is the generation of the intermolecular aldol in-

termediate 5, which proceeds to the quinoline via the above

mentioned hydroxyimine 4.

Evidence in support of each mechanistic proposal exists,

but that for the intermolecular aldol process is very rare.

Thus, a unique example of the isolation of the purported

aldol adduct 9 (Scheme 2) in the piperidine-catalyzed, room

temperature reaction of 2-aminonicotinaldehyde with 1,4-

cyclohexanedione has been described (7). Although an NMR

spectrum of 9 could not be obtained, analytical (CHN) and

IR data and the nearly quantitative generation of the ex-

pected quinoline derivative on mere heating in toluene solu-

tion were consistent with the β-ketol structure. On the other

hand, Schiff bases 10 bearing a stabilizing substituent in the

β-position have been generated in several instances (8–10),

although not under the usual Friedländer conditions (see be-

Given the weight of the evidence cited above, the

Friedländer synthesis is usually considered to occur via an

intermediate Schiff base. The intermolecular aldol mecha-

nism has also been considered improbable because of the ex-

pected facile formation of the conjugated unsaturated

carbonyl compound 6 of E-geometry and the unlikely

conversion of this species into the quinoline (via the Z-

stereoisomer 7) (5). Nevertheless, recently devised versions

of the Friedländer synthesis show that neither 5 nor 6 can a

priori be discounted as intermediates (11–14). Indeed, in a

generally overlooked 1883 publication, Drewsen (15) reported

that the nitroenone 12 was converted into 2-methylquinoline

with stannous chloride in hydrochloric acid medium, al-

though neither the exact conditions nor the yield were re-

ported!

The putative generation of aldol intermediates correspond-

ing to 5 (11, 12) and the isolation of enones (14) and the

enals related to 6 (R¹, R³= H) (11) as N-acylated congeners,

as well as the conversion of these species and 12 into

quinolines, encouraged us to take another look at the mecha-

nism(s) of the Friedländer synthesis. Herein are described

our attempts to generate each of the intermediates shown in

Scheme 1 and to establish the presence or absence thereof

under conditions typically used in this synthesis.

Results and discussions

The starting materials required for this study were pre-

pared as shown in Scheme 3. Thus, the sodium-hydroxide-

promoted condensation of the aldehydes 13 with acetone or

2-butanone under conditions similar to or slightly modified

from those of Baeyer and Drewsen (16) gave the β-ketols 14

as the major products. These were accompanied by minor

amounts of the kinetic product 15, from 2-butanone and 13a,

and the enone 16c, from 13b and acetone. The nitroenones

16a and 16b, obtained from the β-ketols 14a and 14b by de-

hydration with hot acetic anhydride, were converted into the

amines 17a and 17b by reduction with iron powder and am-

monium chloride in a two-phase system at room temperature

(17, 18). The very unstable enal 17c was prepared by this

technique as well. This mild, chemospecific reducing system is

also the preferred one for preparing 2-aminobenzaldehydes

Muchowski and Maddox

463

Scheme 1. Possible mechanisms of the Friedländer quinoline synthesis.

Scheme 2.

(18) from 2-nitrobenzaldehydes (13a and 13c). Indeed, for

the synthesis of 2-aminobenzaldehyde itself, it is far superior

in terms of yield (86%–93%) and ease of workup than is the

commonly used method (19) or its recent modification (20).

The generation of an aldol intermediate analogous to 5

(Scheme 1) was examined first. Reduction of the β-ketol 14a

with the iron – ammonium chloride system gave 2-

methylquinoline-N-oxide (20, Scheme 4, Table 1) in 76%

yield and a trace of 2-methylquinoline (21a). The N-oxide

20 is obviously produced by the cyclization of the very

nucleophilic intermediate hydroxylamine 19a. The nearly

exclusive generation of 20 under these conditions is ex-

pected, since during the reduction of the nitro compounds

13a, 16a, and 16b at least, the rapid formation of the inter-

mediate hydroxylamines and the much slower appearance of

the amines is easily seen by thin layer chromatography

(TLC). In contrast, catalytic reduction of 14a over Pd–C cat-

alyst at atmospheric pressure gave a 4:1 mixture favoring the

N-oxide. Somewhat similar results have been reported for

the catalytic reduction of certain Baylis–Hillman (21) prod-

ucts derived from 2-nitrobenzaldehyde (13). Reduction of

the azido compound 14c, where the amine 19b must be an

intermediate, using the iron – ammonium chloride method

gave 21a exclusively in high yield. No intermediates were

detectable by TLC in these reductions; therefore, the lifetime

of both the aldol intermediate 19b and the cyclization prod-

uct 4 (Scheme 1) enroute to the quinoline must be quite

short. This assumption was verified for the latter intermedi-

ate by reaction of the azido compound 14c with triphenyl-

phosphine. No intermediate, not even the iminophosphorane

22, was detectable by TLC during the course of the reaction,

even at room temperature.

The E-enones 17a,b were next studied as potential inter-

mediates in the Friedländer synthesis. Under thermodynamic

conditions, the base-catalyzed aldol condensation is revers-

ible in all steps (22). For this reason, as well as because of

464

Can. J. Chem. Vol. 82, 2004

Scheme 3. Synthesis of the starting materials.

Scheme 4.

microscopic reversibility (23), it should be possible to con-

vert the E-enones into the corresponding quinolines, possi-

bly via the Z-enones 7 (see also ref. 24) under basic

conditions. Indeed, when dissolved in aqueous ethanol con-

taining 20 mol% sodium hydroxide, 17a was cleanly con-

verted into 2-methylquinoline (21a) after ca. 1 day at room

temperature (Table 1). The reaction of 2-aminobenzaldehyde

with excess acetone under very similar conditions produced

21a more slowly but just as efficiently. These are the condi-

tions described by Friedländer himself (2)! Even the alde-

hyde 17c is very slowly transformed into quinoline with less

than an equivalent of base at ambient temperature (entry 7).

In contrast, the enone 17b is stable under these conditions,

but it is converted efficiently into 2,3-dimethylquinoline (21b)

at reflux temperature using 3.3 equiv. of base. Excess so-

dium hydroxide is also required for the generation of 21b

from 2-aminobenzaldehyde and 2-butanone, but the reaction

proceeds at room temperature (entry 9). It is noteworthy that

the enone 17b was not detectable during the course of the

latter reaction, even though it would easily have been ob-

servable by TLC (intense yellow color on silica gel plates)

had it been formed. Thus, although such enones clearly are

viable intermediates, it is unlikely that they play a signifi-

cant role in the alkali-induced version of the Friedländer

quinolone synthesis. Hsaio et al. (14) have recently shown

that the Horner–Emmons reaction of 2-aminonicotinaldehyde

and β-ketophosphonates generated the expected naphthyri-

dines via the E-enones, but they did not examine the corre-

sponding Friedländer synthesis for these intermediates.

The question of a Schiff base intermediate 3 in the alkali-

hydroxide-mediated version of the Friedländer synthesis can

now be addressed with confidence. Under neutral conditions,

2-aminobenzaldehyde (18a) is remarkably resistant to reac-

tion with ketones or aldehydes. For example, it can be recov-

Muchowski and Maddox

Table 1. Formation of quinoline derivatives from putative Friedländer intermediates.

465

Entry

Starting materials

Conditions

Time (h)

Products (%)^a

1

0.25

12

2

1

2

3

4

5

6

7

8

9

14a

Fe-NH₄Cl

20 (76), 21a (trace)^b

20 (80)^c, 21a (20)^c

21a (85)

14a

H₂, Pd–C

14c

Fe-NH₄Cl

14c

Ph₃P, Tol, ∆

21a (73)

17a

0.2 NaOH, EtOH-H₂O(3:1), RT

0.2 NaOH, EtOH-H₂O(4:3), RT

0.8 NaOH, EtOH-H₂O(3:1), RT

3.3 NaOH, EtOH-H₂O(1:1), ∆

3.3 NaOH, EtOH-H₂O(3:2), RT

27

47

75

16

21a (83)

18a + Me₂CO

17c

21a (80)

Quinoline (63)

21b (83)

17b

18a + MEK

7

d

21b (77)

10

11

17c

17a

TFA, THF, RT

Tar

21a (82)

1.5 mol L^–1HCl, Diox-H₂O(1:1), ∆

1

12

13

14

15

16

17b

1.5 mol L^–1HCl, Diox-H₂O(1:1), ∆

cat. H₂SO₄-HOAc, ∆

21b (87)

21a (0)

18a + Me₂CO

17a

2

18

2

21a (26)

21a (35)

21a (75)

cat. H₂SO₄-HOAc, ∆

hν, EtOH

17a

16

Note: Tol = toluene; RT = room temperature; Diox = dioxan; cat. = catalytic H₂SO₄in acetic acid..

^aYield of picrate except where indicated.

^bTLC.

^cProduct ratio measured by NMR.

^dFormation took place within seconds rather than hours.

ered unaltered after many hours at reflux temperature in

acetone solution. Likewise, toluene solutions of 18a, con-

taining various aldehydes in excess, are stable at room tem-

perature for long periods of time. Given the diminished

reactivity of 18a and its congeners toward aldehydes and ke-

tones and the aqueous- and (or) alcohol-based media in

which the alkaline hydroxide-mediated condensations are

usually effected, a Schiff base intermediate in this version of

the Friedländer synthesis is quite improbable.

The formation of quinoline derivatives from N-acylated

congeners of 17c in acidic media (11) and the relatively fac-

ile acid-induced isomerization of Z-enones to E-enones (24–

28) augured well for the acid-promoted generation of quino-

lines from 17a-c. The free enal 17c was, however, rapidly

destroyed, even under the mildly acidic conditions previ-

ously used (11) to cyclize the N-BOC congeners of 17c (en-

try 10). By contrast, enones 17a,b were converted into the

quinolines 21a,b in over 80% yields with 1.5 mol L^–1hydro-

chloric acid in 50% aqueous dioxane after only 1 h at reflux

temperature. An attempt to prepare 21a from 2-aminobenz-

aldehyde and acetone under these conditions resulted in the

destruction of the starting materials. Quinoline 21a was nev-

ertheless slowly formed from these reactants, although only

in 26% yield, when the reaction was carried out in acetic

acid solution containing a catalytic amount of sulfuric acid

(29) at reflux temperature. 2-Methylquinoline (21a) was also

produced (35% yield) from the enone 17a upon heating in

the same reaction medium for 2 h. Although such enones

could well be intermediates under those often-used acidic

conditions, their short half-lives ensure that they would not

be readily detectable, at least by TLC.

enones must indeed be very short lived was obtained by

photolysis of 17a in ethanolic solution. 2-Methylquinoline

was produced as the sole product, but the presumed interme-

diate Z-enone 23 was undetectable by TLC.

Up to this point no convincing evidence for an intermedi-

ate Schiff base had been found under conditions commonly

used in the Friedländer synthesis. It seemed reasonable,

therefore, to attempt to generate such species from alde-

hydes and 2-aminobenzaldehyde (18a) under more classical

conditions. Heating solutions of 18a and 1.2 equiv. of n-

heptanal, 3-phenylpropionaldehyde, or phenylacetaldehye,

freshly generated from the corresponding bisulfite adducts,

in toluene containing a catalytic amount (5 mol%) of p-

toluenesulfonic acid monohydrate with separation of water,

gave modest yields of 3-substituted quinolines 24a-c (Ta-

ble 2). These quinolines were formed extremely slowly, or

not at all, at room temperature. The quinolines were also ob-

tained under similar conditions and in comparable yields us-

ing ytterbium triflate (10 mol%) as the catalyst, except that

the reaction periods were greatly reduced, and water separa-

tion was not necessary (Table 2, entries 2, 4, 6). In every

case, independent of the catalyst used, TLC examination

showed that several much less polar, yellow-colored, neutral

substances were also formed, none of which was a Schiff

base. n-Heptanal gave rise to one such material in relatively

good yield (25%–43%, ytterbium triflate catalyst) as an oil

with a mass spectral molecular weight of 416. This corre-

sponds to the condensation of two molecules of heptanal and

two molecules of 18a with the loss of three molecules of

water (C₂₈H₃₀N₂O). The IR spectrum showed absorptions at

3334, 2733, and 1659 cm^–1, consistent with the presence of

aldehyde and NH groups strongly bonded to each other, and

It was tacitly assumed that Z-enones were involved as ten-

uous intermediates in the conversion of the E-enones 17a,b

into the quinolines 21a,b. Corroborative evidence that the Z-

1

this was supported by the H NMR spectrum (DMSO-d₆),

with peaks at δ 9.80 and 8.91, respectively. In addition to

466

Can. J. Chem. Vol. 82, 2004

Table 2. Generation of quinolines from Schiff bases derived from RCH₂CHO (A) or MeCOCH₂R (B) and 18a or

18b.

Entry

Starting material (R)

Catalyst

Time (h)

Products (%)^a

1

2

18a; A(n-C₅H₁₁)

18a; A(PhCH₂)

18a; A(Ph)

TsOH

20

24a (34–36)^b

Yb(OTf)₃

TsOH

3 to 4

22

24a (29–47)^b

3

24b (28)^b

4

Yb(OTf)₃

TsOH

4

24b (28)^b

5

22

24c (31)^b

6

18a; A(Ph)

Yb(OTf)₃

NaOH

4

24c (44–53)^b

7

18a; A(n-C₅H₁₁)

18b; A(Ph)

8(4)^c

5(3)^c

6(3.75)^c

5(3)^c

58(7)^d

48(7)^d

24

24a (47)^b

8

24c (80)

9

18b; A(n-C₅H₁₁)

18b; A(Ph)

24d (42)

10

11

12

13

14

15

16

17

18

19

24e (78)

18a; B(Me)

18a; B(Et)

21b (48), 37a (28)

21c (34.5), 37b (41)

21c (66), 37b (20)

21d (4), 37c (61)

21d (5), 37c (78)

21c (13), 37b (62.5)

21c (16), 37b (77)

21e (14.5), 37d (59)

21e (3.2), 37d (0.8)

18a; B(Et)

18a; B(i-Pr)

18a; 38a

Yb(OTf)₃

NaOH

49(7)^d

4

Yb(OTf)₃

HgCl₂– Yb(OTf)₃

4.5^e

1

18a; 38a

18a; 38b

6^e

18a; 1-heptyne

72

^aYield of free base unless otherwise indicated.

^bYield of picrate.

^cTime of addition of reagents to catalyst; see Experimental.

^dParentheses, reflux time in toluene; no parentheses, time at 95 °C; see Experimental.

^eRoom temperature reaction.

other absorptions, the NMR spectrum possessed peaks for

eight aromatic hydrogens and three methylene groups at δ

2.78, 2.98, and 5.04 (coupled to the NH by 6.2 Hz). Four of

the aromatic hydrogens had a splitting pattern identical to

and chemical shifts (δ 6.68 (t), 6.90 (d), 7.35 (dt), 7.58 (dd))

similar to those found in 2-aminobenzaldehyde (20). The

chemical shifts of the remaining aromatic hydrogens (δ 7.43

(t), 7.63 (d), 7.78 (d), 8.05 (s)) were essentially identical to

those reported for H-6, H-7, H-5, and H-4, respectively, of

quinoline (30), and the multiplicities were concordant with

those expected for a 2,3,8-trisubstituted derivative. This sub-

stitution pattern was strongly supported by a ROESY experi-

ment, which showed that the δ 2.78 methylene group was

vicinal to the hydrogen at δ 8.05, that the two upfield methy-

lene groups were also vicinal, and that the low field (δ 5.04)

methylene moiety was vicinal to the δ 7.63 hydrogen. The

above information is fully accommodated by the quinoline

derivative with structure 25a (Scheme 5). In particular, the

unusual chemical shift of the methylene group at C-8 stems

from the strong deshielding effect of the peri lone electron

pair of the quinoline nitrogen atom. The formation of this

compound clearly requires the involvement of one or more

Schiff base intermediates.

trace of a material with an Rf expected for 25c. It was as-

sumed that the failure to generate significant amounts of

25b,c was connected with the volatility of the aldehydes;

therefore, the reactions were carried out at room tempera-

ture. After 2 to 3 days, acetaldehyde (4 equiv.) gave a major,

new, yellow-colored oil (ca. 65% yield), but no quinoline.

This material had a mass spectral molecular weight of 294,

i.e., derived from the condensation of two molecules of

acetaldehyde with two molecules of 18a minus two mole-

cules of water (C₁₈H₁₈N₂O₂). The NMR spectrum of this ap-

parently homogeneous material (TLC) indicated, however,

that it was a 4:1 mixture of isomers. The major component

showed signals for two aldehyde moieties (δ 9.86 and 9.81),

two NH groups (δ 8.67 and 8.40), seven aromatic hydrogens,

multiplets corresponding to four mutually coupled aliphatic

protons (δ 4.86, 3.38, 2.31, and 1.71), and a methyl group at-

tached to a secondary carbon atom. The hydrogens at δ 4.86

and 3.83 were also coupled to the low-field NH and the

methyl group, respectively. Deuteration of the NH hydro-

gens and decoupling of the methyl group permitted measure-

ment of all the aliphatic hydrogen coupling constants

(Table 3) and assignment of the cis 2,4-disubstituted tetrahy-

droquinoline structure 26a to this substance, with the methyl

and aniline moieties occupying equatorial sites in a half-

chair system. Full decoupling studies were not carried out on

the minor component, but the data in Table 3 indicate that it

is the trans stereoisomer 26b with the anilino group axial.

When propanal was used in the room temperature reaction, a

not-readily-separable 54:34:12 mixture of stereoisomers was

obtained in about 50% yield. The high-temperature reaction

When the above reaction was attempted with propanal un-

der similar conditions (sealed vessel), only a miniscule

amount (ca. 1.5%) of 25b was formed (crystalline). The ma-

jor product in this complex reaction mixture was a structur-

ally different yellow-colored oily substance (see below). A

reaction with acetaldehyde (4 equiv.) at reflux temperature

gave a complex mixture containing no major product and no

Muchowski and Maddox

467

Scheme 5.

mentioned above gave mainly the major isomer, and the

NMR spectral data in Table 3 establish that this is the 2,3,4-

trisubstituted tetrahydroquinoline 26c, analogous to 26a but

with a C-3 axial methyl group. The isomers of intermediate

and least abundance have structures 26d and 26e, with the

2,3,4-substituents arranged a,e,e and a,a,e. In addition to the

three tetrahydroquinolines, the acyclic dialdehyde 27 was

isolated as a single isomer in ca. 40% yield. The room-

temperature reaction of 2-aminobenzaldehyde and phenyl-

acetaldehyde was not studied in detail, but a very insoluble,

surprisingly high-melting (236 °C) yellow solid could be

isolated in ca. 15% yield. The NMR spectral data (Table 3)

unequivocally show that this is the all-equatorial isomer 26f.

The formation of tetrahydroquinolines from anilines and

aliphatic aldehydes has been known for some time (31, 32)

and proceeds via an intermediate Schiff base dimer 28,

which sometimes can be isolated (e.g., for X = H, R = Et,

Ar = Ph (31)). The dimer cyclization is acid catalyzed, high

yielding, and takes place at room temperature. It is apparent

that the dialdehyde 27 is the hydrolysis product of the inter-

mediate enroute to the tetrahydroquinolines 26. It is also

worth noting that the tetrahydroquinolines and the dihydro-

quinoline described below are well-recognized intermediates

in the Doebner–Miller quinoline synthesis (see ref. 5,

pp. 102–104).

The high-temperature ytterbium-triflate-promoted reaction

of phenylacetaldehyde with 2-aminobenzaldehyde gave, in

addition to 3-phenylquinoline, a complicated neutral mixture

from which a crystalline, yellow-colored, structurally unique

monoaldehyde could be isolated (13%–29%). That this sub-

stance was the 1,2-dihydroquinoline 29 was evident from its

NMR spectrum, which showed, among other signals, absorp-

tions for an aldehyde hydrogen, an NH moiety, an olefinic

hydrogen (singlet), a methine hydrogen, and a benzylic

methylene group at δ 9.77, 8.38, 6.61, 5.03, and 2.82, re-

spectively. The methine hydrogen was coupled to both the

NH (J = 2.6 Hz) and the methylene hydrogens (J = 6.0 Hz).

Solutions of 29 underwent slow aerial oxidation to the

corresponding quinoline 30, which was obtained much more

efficiently by oxidation with Fremy’s salt (33). The alde-

hyde hydrogen of the latter compound is found at δ 11.46,

nearly δ 2 downfield from that observed for 29. The remark-

ably low-field position of this proton must be a consequence

of an aldehyde conformation in which the hydrogen and ox-

ygen are close to and remote from the lone pair of the

quinoline nitrogen atom, which is a conformation that the al-

dehyde must adopt to avoid an unfavorable dipole–dipole

interaction. The hydrogen of a series of quinoline-8-

carboxaldehydes has been reported to resonate at δ 11.3–

11.5 without comment (34).

The isolation of the dihydroquinoline 29 was of consider-

able significance, because it suggested a plausible mecha-

nism for the formation of the quinoline containing dimers

25. Specifically, fragmentation of the tetrahydroquinoline dimers

26 to 2-aminobenzaldehyde and the 1,2-dihydroquinoline

31, followed by Schiff base formation from these fragments,

would generate 32, which could lead to 25 by an intra-

molecular redox process (Scheme 6). In fact, a boiling solu-

tion of the isomeric tetrahydroquinolines 26c, 26d, and 26e

in toluene containing 10 mol% of ytterbium triflate was

slowly (15–21 h) converted into a mixture from which the

previously described quinoline dimer 25b could be isolated

in ca. 25% yield. The tetrahydroquinoline mixture derived

from acetaldehyde (26a and 26b) was converted even more

rapidly (4 h) and more efficiently (>45% yield) into the 2-

methyl analog 25c.

Why are the Schiff bases derived from aliphatic aldehydes

and 2-aminobenzaldehyde converted exclusively into

tetrahydroquinolines at room temperature and only give the

expected 3-substituted quinolines at higher temperature? The

answer must certainly be associated with the geometry of

the Schiff base. At ambient temperature, the thermodynami-

cally more stable E-isomer 33 (Scheme 7) must be generated

essentially exclusively. This geometry cannot, however, lead

to the 3-substituted quinoline; therefore, intermolecular

aldolization to the Schiff base dimers followed by

cyclization to the tetrahydroquinoline derivatives takes place.

At higher temperature the E-isomer exists in equilibrium

468

Can. J. Chem. Vol. 82, 2004

Muchowski and Maddox

469

Scheme 6. Generation of quinolines 25 from tetrahydroquinolines 26.

Scheme 7. Aldehyde Schiff base cyclization.

Scheme 8. Ketone Schiff base cyclization.

with Z-isomers 34, which by an intramolecular aldolization

can now produce the 3-alkylquinoline. The conversion of the

tetrahydroquinoline derivatives 26 into the dimeric quinoline

derivatives 25 must also involve higher energy processes,

and thus, the temperature must be raised for this transforma-

tion to occur at a useful rate as well.

There are two important consequences of the above re-

sults. Firstly, given that the Z-Schiff base is a high-

temperature intermediate, it should be possible to devise

conditions that would favor its formation and thus increase

the 3-alkylquinoline yield. Indeed, addition of a solution of

the 2-aminobenzaldehydes 18a or 18b and a slight excess of

n-heptanal or phenylacetaldehyde in toluene to a boiling sus-

pension of ytterbium triflate in toluene over a 3 to 4 h period

gave the expected 3-substituted quinolines in 40%–80%

yields (Scheme 7 and Table 2, entries 7–10). Secondly, if

Schiff bases of unsymmetrical ketones (e.g., methyl alkyl

ketones) could be generated, and if E–Z equilibration did oc-

cur (e.g., 35 with 36, Scheme 8), then mixtures of two

quinolines (21 and 37) should be formed, and the 2-

470

Can. J. Chem. Vol. 82, 2004

substituted quinoline 37 might well predominate. This would

be quite different from what occurs under the usual basic or

acidic conditions of the Friedländer synthesis (see above).

Schiff bases unfortunately do not form from ketones and 2-

aminobenzaldehyde under the conditions described above

for aldehydes. It was found, however, that heating a solution

of 18a in 2-butanone containing catalytic ytterbium triflate

at 95 °C resulted in the very slow (>2 days) consumption of

the aldehyde. Removal of the excess 2-butanone, followed

by heating the residue in toluene at reflux temperature for

7 h, gave a 1.7:1 mixture of 2,3-dimethylquinoline (21b) and

2-ethylquinoline (37a, Scheme 8) in over 90% yield (Ta-

ble 2, entry 11). In contrast, under alkaline conditions, 2,3-

dimethylquinoline was the only product obtained (Table 1,

entry 9). When 2-pentanone was used in the above reactions,

2-n-propylquinoline (37b) predominated over 2-methyl-3-

ethylquinoline (21c) by a factor of 1.2 for ytterbium triflate,

whereas the latter quinoline was the major product in the so-

dium-hydroxide-mediated reaction (21c:37b = 3.3:1, Ta-

ble 2, entries 12 and 13). Hawes and Gorecki (35) also

observed a slight predominance (factor of 1.1) of the kinetic

over the thermodynamic product in the piperidine-catalyzed

reaction of 4-aminonicotinaldehyde with neat 2-butanone.

These results suggested that the kinetic product might be fa-

vored to an even greater extent for 4-methyl-2-pentanone,

and a 15.3:1 37c:21d ratio supported this prediction. It was

astonishing to find, however, that essentially the same prod-

uct ratio (15.6:1) was observed for the sodium-hydroxide-

promoted reaction, in which a Schiff base intermediate is

unlikely based on the data described herein. Since the eno-

lization of 4-methyl-2-pentanone occurs rapidly at both sites

(deuteration complete in <5 min in perdeuteriomethanol

containing a catalytic amount of the sodium derivative), it is

probable that the highly selective formation of 37c in the

alkali-mediated condensation results from steric inhibition of

aldolization at the methylene site. It is significant that analo-

gous results have been reported for β-methyl substituted ke-

tones both under acidic (36) and basic (37) conditions.

Except for the case of 4-methyl-2-pentanone, the alterations

in product ratios toward the kinetic one, although signifi-

cant, are not so large that they are convincingly diagnostic

of Schiff base intermediates. Since anhydrous conditions

were not maintained in these reactions, it is quite possible

that substantial amounts of the quinoline were formed by

intermolecular aldol processes in both the piperidine- (35)

and the ytterbium-triflate-catalyzed reactions. The facile

acid-catalyzed transamination of enamines and Schiff bases

(38–40) and the well-known water sensitivity of the latter

prompted an examination of the reactivity of the pyrrolidine

enamines of 2-pentanone (38a) and 2-heptanone (38b) to-

wards 2-aminobenzaldehyde (Scheme 8). For example, a so-

lution of 18a in anhydrous toluene containing 2.4 equiv. of

the enamine 38a (1 equiv. of 38a to ensure anhydrous condi-

tions) and catalytic ytterbium triflate was rapidly and effi-

ciently converted into a mixture of the expected quinolines

21c and 37b, both at reflux (<1 h) and room (4 h) tempera-

ture. The kinetic product 37b was favored by a factor of ca.

5 in both cases (Table 2, entries 16 and 17). The kinetic

product was also highly favored for the reaction of 18a with

the enamine 38b at room temperature (37d:21e = 4:1).

These are reasonable results if both the generation of the

Schiff bases (mainly 35) and the cyclization thereof to the

quinolines are very fast processes.

Lastly, the room-temperature hydroarylamination of 1-

heptyne with 18a in the presence of an equimolar amount of

mercuric chloride (41) and catalytic ytterbium triflate gave a

mixture of 37d and 21e (4:1) in low, but reproducible yield

(4%). In this case, the quinolines must be generated via

Schiff base intermediates. The 4:1 kinetic:thermodynamic

product ratio was identical to that observed for 38b and

strongly supports Schiff base intermediates in the enamine

based reactions.

When this study was completed, Dormer et al. (37) re-

ported that the kinetic product was dramatically favored

when the methyl alkyl ketone was added slowly to a hot

ethanolic solution of the 2-aminoarylaldehyde, a slight ex-

cess of various secondary amines, and a catalytic amount of

sulfuric acid. Kinetic:thermodynamic product ratios as high

as 24 were observed when the secondary amine was 1,3,3-

trimethyl-6-azabicyclo[3.2.1]octane (TABO). The authors

did not offer an explanation of these remarkable results. It is

our contention that the reaction conditions reported by these

investigators are precisely those that should result in the

rapid formation and cyclization of the kinetic Schiff base,

which arises by the fast transamination of an intermediate

enamine.

In conclusion, the results described herein show that the

generation of quinolines from simple aldehydes and 2-

aminobenzaldehydes under the typically used acidic or basic

conditions of the Friedländer synthesis takes place by a pro-

cess in which an intermolecular aldol condensation is the

first step. It is highly probable that the same mechanism is

involved when 2-aminoarylketones are used to prepare

quinolines. When Schiff bases are deliberately generated

from aldehydes at room temperature, the kinetically formed

E-isomer is diverted to nonFriedländer products. 3-

Substituted quinolines are formed only at higher tempera-

tures, which effect isomerization of the E-Schiff base to the

Z-isomer, which then undergoes intramolecular aldolization

to the heterocyclic aromatic product. When Schiff bases are

generated from methyl alkyl ketones, the kinetically favored,

less-hindered Z-Schiff base is formed preferentially and rap-

idly cyclizes to the 2-alkylquinoline derivative. Thus, the

hallmark of such cyclizations is a high kinetic:thermody-

namic product ratio, a result completely opposite to that ob-

served when the usual acidic or basic reaction conditions are

employed in the Friedländer quinoline synthesis. This result

has obvious synthetic consequences, which have been con-

vincingly demonstrated in the recently described study of

Dormer et al. (37).

Experimental

General

The melting points are uncorrected. The NMR spectra

(300 and 600 MHz) were recorded in deuteriochloroform

unless indicated otherwise, and the chemical shifts are re-

ported in parts per million (ppm) (δ) with internal tetra-

methylsilane as zero. The IR spectra were measured as

dispersions in KBr for solids and as films for liquids. The

UV spectra were taken in methanol solution. Reactions were

followed by TLC on silica gel plates unless otherwise indi-

Muchowski and Maddox

471

cated. Column chromatographic separations were effected

using 230–400 mesh silica gel. The columns were packed

dry, successively layered with the crude reaction mixture

(absorbed onto silica gel) and a small amount of sand, and

then eluted with the appropriate solvent system. The mix-

tures absorbed on silica gel were prepared by dissolving the

reaction mixture in dichloromethane, adding this solution to

silica gel, and then removing the solvent in vacuo.

The terms “worked up in the usual manner”, “the usual

workup”, etc., signify that the extract was dried over anhy-

drous sodium sulfate, and the solvent was then removed in

vacuo.

the mixture was heated at reflux temperature for 1 h. Water

was added cautiously to the hot solution to decompose the

excess acetic anhydride, and when the vigorous reaction was

over, the mixture was poured into ice-water and extracted

with dichloromethane. The extract was washed successively

with 10 wt % sodium carbonate solution and water and then

worked up as usual. The crude product (4.07 g) was ab-

sorbed onto silica gel (20 g), loaded onto a dry packed col-

umn (6 × 28 cm) of silica gel (380 g), and eluted with

hexane – ethyl acetate (70:30, 50 mL fractions). F 13–16

contained the desired crystalline enone 16b (2.66 g, 65.6%).

Subsequent fractions contained a mixture of 16b and 5-(2-

nitrophenyl)-(4E)-penten-3-one. An analytical specimen of

the desired enone was obtained by crystallization from hex-

4-Hydroxy-4-(2-azidophenyl)-2-butanone (14c) and 4-(2-

azidophenyl)-(3E)-buten-2-one (16c)

1

ane and had mp 59 to 60 °C after drying in vacuo (4 h). H

Aqueous sodium hydroxide (0.25 mol L^–1, 0.6 mL) was

added dropwise at 15 °C to a stirred solution of 2-

azidobenzaldehyde (13b (42)) (294 mg, 2 mmol) dissolved

in a mixture of acetone (2.6 mL), water (2 mL), and ethanol

(2 mL). One half hour after the addition was completed, the

reaction mixture was poured into aqueous sodium chloride

and extracted with dichloromethane. The extract was washed

with water and then worked up as usual to give an oil

(401 mg), which was taken up in a small amount of di-

chloromethane and added to silica gel (2 g), and the mixture

was evaporated to dryness in vacuo. This material was lay-

ered onto a dry packed column (3.5 × 8.5 cm) of silica gel

(40 g) covered with sand and eluted with hexane – ethyl ace-

tate ((70:30), 20 mL fractions (F)). F 2–4 contained the

enone 16c (34 mg, 9.1%), and F 6–10 contained the β-ketol

14c (324 mg, 78.9%), both of which were crystalline.

After crystallization from cyclohexane and drying in

vacuo at 23 °C for 5 h with protection from light, the β-ketol

NMR δ: 1.83 (d, 3H, J = 1.4 Hz), 2.50 (s, 3H), 7.39 (d, 1H,

J = 7.8 Hz), 7.54 (td, 1H, J = 1.2, 7.8 Hz), 7.69 (td, 1H, J =

1.2, 7.5 Hz), 7.81 (bs, 1H), 8.18 (dd, 1H, J = 1.3, 8.2 Hz).

¹³C NMR δ: 13.25, 26.26, 125.38, 129.58, 131.68, 132.30,

133.83, 136.64, 139.22, 147.88, 200.14. Anal. calcd. for

C₁₁H₁₁NO₃: C 64.38, H 5.40, N 6.83; found: C 64.52, H

5.33, N 6.95.

Reduction of nitro compounds with iron powder and

ammonium chloride

The following is a typical reduction. A mixture of the ni-

tro compound (10 mmol) dissolved in ethyl acetate

(100 mL), water (60 mL), iron powder (2.1 g, 37.5 mmol

(Fisher Scientific, electrolytic, <100 mesh; iron powder from

other suppliers, e.g., Malinckrodt, gave similar results, but

the reductions were slower.)), and ammonium chloride

(3.34 g, 62.5 mmol) was stirred vigorously at room tempera-

ture for the time period shown below. The mixture was fil-

tered through Celite, and the inorganic residue was washed

well with ethyl acetate. The ethyl acetate phase was sepa-

rated and combined with an ethyl acetate extract of the aque-

ous phase; the extract was washed with saturated NaCl

solution and then worked up as usual. The crude product

was then purified by column chromatography on silica gel as

described below.

1

14c had mp 69 to 70 °C. H NMR δ: 2.20 (s, 3H), 2.72 (dd,

1H, J = 9.5, 17.6 Hz), 2.91 (dd, 1H, J = 2.7, 17.6 Hz), 3.47

(d, 1H, J = 3.1 Hz, exchanged with D₂O), 5.34 (m, 1H), 7.13

(dd, 1H, J = 1.0, 8.0 Hz), 7.17 (td, 1H, J = 1.1, 7.7 Hz), 7.32

(td, 1H, J = 1.7, 7.7 Hz), 7.54 (dd, 1H, J = 1.6, 7.7 Hz). ¹³C

NMR δ: 30.62, 50.45, 65.31, 117.87, 125.18, 128.65, 133.83,

136.05, 209.22. Anal. calcd. for C₁₀H₁₁N₃O₂: C 58.94, H

5.40, N 20.48; found: C 58.85, H 5.40, N 20.60.

After crystallization from cyclohexane and drying in

vacuo (17 h), the enone 16c had mp 107 to 108 °C. ¹H NMR

δ: 2.39 (s, 3H), 6.70 (d, 1H, J = 16.5 Hz), 7.15 (d, 1H, J =

7.8 Hz), 7.20 (td, 1H, J = 1.0, 8.1 Hz), 7.43 (td, 1H, J = 1.5,

8.5 Hz), 7.6 (dd, 1H, J = 1.5, 7.8 Hz). ¹³C NMR δ: 27.20,

118.82, 125.05, 126.11, 127.90, 128.71, 131.54, 137.62,

139.32, 198.57. Anal. calcd. for C₁₀H₉N₃O: C 64.16, H 4.85,

N 22.45; found: C 64.54, H 4.89, N 22.33.

4-(2-Aminophenyl)-(3E)-buten-2-one (17a)

Reaction time, 17 h. A 4 mmol scale reaction product (ab-

sorbed on silica gel, 3 g) was purified by column (3.5 ×

14 cm) chromatography on silica gel (67 g, 25 mL frac-

tions). The product (0.516 g, 80.0%), present in F 3–5, was

1

obtained as a yellow oil. H NMR : 2.36 (s, 3H), 4.01 (bs,

δ

2H), 6.53 (d, 1H, J = 16.0 Hz), 6.71 (dd, 1H, J = 1.3,

8.3 Hz), 6.77 (td, 1H, J = 1.0, 7.3 Hz), 7.18 (td, 1H, J = 1.5,

7.3 Hz), 7.39 (dd, 1H, J = 1.5, 8.0 Hz), 7.68 (d, 1H, J =

16.0 Hz). ¹³C NMR δ: 28.47, 117.30, 119.48, 120.29,

127.16, 128.57, 131.96, 139.06, 146.32, 198.63. A sample of

this material was converted into the pivalamide (59.9%) by

the technique described below for 17b. A sample on solution

in hot toluene and dilution with an equal volume of hexane

gave material with mp 119 to 120 °C (lit. (43) value mp 117

to 118 °C).

4-(2-Nitrophenyl)-3-methyl-(3E)-buten-2-one (16b)

Sodium hydroxide (0.25 mol L^–1) was added dropwise at

15 °C to a stirred solution of 2-nitrobenzaldehyde (13a,

3.022 g, 20 mmol) in a mixture of 2-butanone (32 mL,

25.7 g), ethanol (12 mL), and water (20 mL), until the solu-

tion was weakly alkaline (ca. 15 mL). After 1 h the solution

was diluted with water, and the mixture was extracted with

ether, washed with saturated NaCl solution, and then pro-

cessed as usual to give a mixture of the β-ketols 14b and 15

(4.42 g). This mixture was dissolved in acetic anhydride

(10 mL); sodium acetate (1.64 g, 20 mmol) was added, and

4-(2-Aminophenyl)-3-methyl-(3E)-buten-2-one (17b)

Reaction time, 20 h. The crude product from an

8.38 mmol reaction was absorbed onto silica gel (8 g) and

472

Can. J. Chem. Vol. 82, 2004

purified on a column (4 × 23 cm) of silica gel (152 g) using

hexane – ethyl acetate (65:35, 50 mL fractions). The product

was contained in F 10–15 as an oil (1.266 g, 83.8%). IR

Preparation of 2-methylquinoline-N-oxide (20) and 2-

methylquinoline (21a) from the β-ketol 14a

1

Iron – ammonium chloride reduction

(film) (cm^–1): 3452, 3366, 3244, 1659, 1623. H NMR δ:

This reduction was effected in the manner described

above on a 1 mmol scale. Evaporation of the ethyl acetate

phase in vacuo gave 92 mg of a crystalline solid. The aque-

ous phase was brought to pH 9 with solid sodium carbonate,

then saturated with sodium chloride and extracted with ethyl

acetate. The extract was processed as usual to give a further

42 mg of solid. The solid material was identical to authentic

20 (yield 75.6%) and contained only a trace of 21a by TLC.

1.95 (d, 1H, J = 1.4 Hz), 2.46 (s, 3H), 3.74 (bs, 2H), 6.74

(dd, 1H, J = 0.9, 8.0 Hz), 6.80 (dt, 1H, J = 1.0, 7.5 Hz), 7.12

(dd, 1H, J = 1.6, 7.6 Hz), 7.17 (dd, 1H, J = 1.5, 7.7 Hz),

7.45 (bs, 1H). ¹³C NMR δ: 13.64, 26.38, 116.20, 118.77,

121.84, 130.02, 130.11, 136.19, 144.70, 200.55.

The pivalamide of this compound was prepared as fol-

lows. Trimethylacetyl chloride (0.25 mL, 245 mg,

2.03 mmol) was added all at once to a stirred solution of the

above amino compound (269 mg, 1.49 mmol) in toluene

(30 mL) containing triethylamine (0.3 mL, 218 mg,

2.15 mmol) and 4-dimethylaminopyridine (18 mg) at 0 °C.

The cooling bath was removed, and the mixture was stirred

at room temperature for 18 h. The mixture was washed suc-

cessively with water, 1 mol L^–1HCl, 10 wt % sodium car-

bonate, and saturated NaCl solution. After the usual workup,

the brown solid was taken up in hot toluene (3 mL) and di-

luted with hexane (9 mL) to give the pivalamide (127 mg,

mp 111 to 112 °C). A second crop (31 mg, mp 110–

111.5 °C, total yield, 158 mg, 40.9%) of product was ob-

tained from the mother liquor. Recrystallization of the first

crop gave the analytical specimen, which had mp 111.5–

112 °C, after drying in vacuo (9 h at 40 °C). Anal. calcd. for

C₁₆H₂₁NO₂: C 74.10, H 8.16, N 5.40; found: C 74.33, H

8.10, N 5.57.

Catalytic reduction

A solution of 14a (418 mg, 2 mmol) in ethanol (30 mL),

containing 10% Pd–C (40 mg), was hydrogenated at room

temperature and atmospheric pressure. After 25 min, hydro-

gen absorption (89% of theory) ceased, the mixture was fil-

tered through Celite, and the solvent was removed in vacuo

1

to give a residue (321 mg), which by H NMR was an 80:20

mixture of 20 and 21a. Crystallization from hot water gave

the monohydrate of 20 (222 mg, 62.6%), which had mp

72.5–74 °C (lit. (45) value mp 77 to 78 °C).

Preparation of 2-methylquinoline (21a) from the azido

compound 14c

Iron – ammonium chloride reduction

This reduction was carried out using the same reagent and

solvent ratios as described above for the nitro compound re-

ductions. Reaction time, 22 h. The ethyl acetate phase, con-

taining 21a, was extracted with 1 mol L^–1HCl; the acid

extract was brought to pH 9 with solid sodium carbonate;

the mixture was extracted with ethyl acetate, and the extract

was washed with saturated NaCl solution. After the usual

workup, the crude base (138 mg from a 1 mmol reaction)

was converted into the picrate (317 mg, 85.1%) with picric

acid (240 mg) in hot ethanol (10 mL). The picrate was iden-

tical to an authentic specimen by mmp, and the free base re-

covered therefrom was identical by NMR spectroscopy to

authentic material.

2′-Aminocinnamaldehyde (17c)

Reaction time, 4 h. After washing the ethyl acetate solu-

tion of the crude product with saturated NaCl solution and

drying over sodium sulfate, the solvent was removed in

vacuo at room temperature or below and the amino com-

pound (100% yield) was immediately dissolved in toluene.

Toluene solutions of this material, which contains a trace of

quinoline (TLC), can be kept for a few days without signifi-

cant decomposition when stored at 5 °C with protection

from light.

2-Aminobenzaldehyde (18a)

Triphenylphosphine

Reaction time, 5 h. The crude product from a 20 mmol re-

action was purified by normal column (4.2 × 18 cm) chro-

matography on silica gel (125 g), using hexane – ethyl

acetate (70:30, 50 mL fractions). F 5–8 contained the prod-

uct (2.08–2.25 g, 86%–93%), which crystallized spontane-

ously and usually had mp 37 to 38 °C (lit. (19) value mp 38

to 39 °C) after drying in vacuo. It is indefinitely stable when

stored at 5 °C.

A solution of the azido compound 14c (61.5 mg,

0.3 mmol) in toluene (10 mL), containing triphenylphos-

phine (79 mg, 0.3 mmol), was stirred at room temperature.

By 1 h, 21a and triphenylphosphine oxide were clearly visi-

ble by TLC (20:80 ethyl acetate – hexane, UV), but no other

new spots were present. After 2 h, the solution was heated at

reflux temperature for 2 h; it was extracted with 1 mol L^–1

HCl, and the acid extract was worked up exactly as de-

scribed for the iron – ammonium chloride reduction. Crude

21a (41 mg) was converted into the picrate (81.7 mg,

73.3%) with picric acid (70 mg) in hot ethanol (4 mL).

2-Amino-4,5-dimethoxybenzaldehyde (18b)

Reaction time, 20 h. The crude product from a 10 mmol

reaction was purified by normal column (3.5 × 19.5 cm)

chromatography on silica gel (100 g) using hexane – ethyl

acetate (1:1, 50 mL fractions). F 7–13 contained the product

(1.395 g, 77%), which had mp 82–84 °C (lit.(44) value mp

Preparation of the quinolines by isomerization of 17a-c

With sodium hydroxide

1

2-Methylquinoline (21a)

86 °C) after drying in vacuo. H NMR δ: 3.85 (s, 3H), 3.89

(s, 3H), 6.12 (s, 1H), 6.88 (s, 1H), 9.69 (d, 1H, J = 0.4 Hz).

¹³C NMR δ: 56.36, 56.92, 98.82, 111.75, 116.43, 141.42,

147.67, 156.56, 191.78.

A solution of 17a (100 mg, 0.62 mmol) in ethanol (3 mL),

water (1 mL), and 1 mol L^–1NaOH (0.1 mL) was stirred at

room temperature for 22 h. Water was added, and the mix-

Muchowski and Maddox

473

ture was extracted with ethyl acetate; the extract was washed

with saturated NaCl solution and worked up in the usual

way. The crude base (96 mg) was converted into the picrate

(192 mg, 83.4%) in the usual way.

Preparation of the quinolines from 2-aminobenzalde-

hyde (18a) and ketones

Sodium-hydroxide-mediated condensations

Quinoline

2-Methylquinoline (21a)

A solution of the freshly prepared enal 17c (147 mg,

1 mmol) in ethanol (6 mL), water (2 mL), and 1 mol L^–1

NaOH (0.1 mL) was left at room temperature. Additional

NaOH was added at 22 h (1 mol L^–1, 0.2 mL) and at 47 h

(2.5 mol L^–1, 0.2 mL). After 75 h the ethanol was removed

in vacuo, and the residue was made acidic with 1 mol L^–1

HCl and washed with ethyl acetate. The acidic phase was

then worked up, as described for the preparation of 21a from

14c above. The crude base was converted into quinoline

picrate (224 mg, 62.5%).

A solution of 18a (121 mg, 1 mmol) in ethanol (4 mL),

water (4 mL), acetone (0.6 mL, 0.475 g, 8 mmol), and 1 mol

L^–1NaOH (0.2 mL) was left at room temperature for 47 h.

The reaction mixture was worked up exactly as described for

the preparation of 21a from 17a under alkaline conditions to

give the picrate (298 mg, 80.0%).

2,3-Dimethylquinoline (21b)

A solution of 18a (121 mg, 1 mmol) in ethanol (6 mL)

and 1 mol L^–1NaOH (3.5 mL) containing 2-butanone

(108 mg, 1.5 mmol) was left at room temperature for 12 h

and then worked up as described above for 17a. The crystal-

line base was converted into the picrate (296 mg, 76.6%).

2,3-Dimethylquinoline (21b)

A solution of the enone 17b (120 mg, 0.67 mmol) in etha-

nol (3 mL), water (1 mL), and 1 mol L^–1NaOH (2.2 mL)

was heated at reflux temperature for 16 h. The reaction mix-

ture was worked up as described above for the preparation of

21a. The crude base was converted into the picrate (213 mg,

82.8%). A sample of the picrate was stirred with excess 1

mol L^–1NaOH and toluene. The toluene phase was sepa-

rated, washed with saturated NaCl solution, and worked up

in the usual way to give 2,3-dimethylquinoline, mp 67–

68.5 °C (lit. (46) value mp 68.2–69 °C).

Acid-mediated condensation

2-Methylquinoline (21a)

To a solution of 18a (61 mg, 0.5 mmol) and acetone

(87 mg, 1.5 mmol) in acetic acid (4.5 mL) was added a 0.18

mol L^–1solution of sulfuric acid in acetic acid (0.5 mL), and

the solution was heated at reflux temperature for 18 h. The

reaction mixture was worked up exactly as described for the

sulfuric-acid-catalyzed isomerization of 17a. The crude

product (43 mg) was absorbed onto silica gel (1 g) and puri-

fied by column (1.8 × 5.5 cm) chromatography on silica gel

(9 g) using hexane – ethyl acetate (60:40, 10 mL fractions).

F 2–4 contained the product (20 mg), which was converted

into the picrate (49 mg, 26.2%).

Acid-mediated isomerization

2-Methylquinoline (21a): 1.5 mol L^–1HCl

A solution of 17a (100 mg, 0.62 mmol) in dioxan (3 mL)

and 3 mol L^–1HCl (3 mL) was heated at reflux temperature

for 1 h. The cooled solution was diluted with water and

made basic with solid sodium carbonate. The mixture was

extracted with ethyl acetate; the extract was washed with sat-

urated NaCl solution and worked up in the usual way to give

the crude base (86 mg), from which was obtained the picrate

(189 mg, 81.8%).

2-Methylquinoline (21a) by photolysis of 17a

A solution of the enone 17a (20 mg, 0.124 mmol) in etha-

nol (4 mL) in a pyrex test tube was placed next to a Pen-Ray

Model PS-1 pencil light, and both were enveloped in alumi-

num foil. The solution was irradiated until TLC showed that

17a was absent (24 h); the solvent was removed in vacuo,

and the crude base was converted into the picrate (34.4 mg,

74.6%).

2-Methylquinoline (21b): Catalytic sulfuric acid in acetic

acid

To a solution of the enone 17a (315 mg, 1.95 mmol) in

acetic acid (8 mL) was added a 0.18 mol L^–1solution of sul-

furic acid in acetic acid (2 mL), and the blood-red reaction

mixture was heated at reflux temperature for 2.25 h. Toluene

was added, and the mixture was evaporated to a small vol-

ume in vacuo. The residue was diluted with water, made ba-

sic with sodium carbonate, and extracted with ethyl acetate.

The extract was washed with 1 mol L^–1HCl, and the acid

phase was made basic with solid sodium carbonate and ex-

tracted with ethyl acetate. The extract was washed with satu-

rated NaCl solution and worked up as usual to give an oil

(219 mg), which was absorbed onto silica gel (2.2 g) and pu-

rified by column (2.2 × 12 cm) chromatography on silica gel

(23 g) using hexane – ethyl acetate (65:35, 10 mL fractions).

F 6–9 contained the product (168 mg), which was converted

into the picrate (251 mg, 34.6%).

Preparation of 3-substituted quinolines 24a-c from 18a

and aldehydes

p-Toluenesulfonic acid catalyst

A solution of 18a (121 mg, 1 mmol) and the appropriate

aldehyde (1.2 mmol, freshly liberated from the bisulfite

adduct) in toluene (50 mL) containing p-toluenesulfonic acid

monohydrate (10 mg) was heated at reflux temperature with

separation of water in a Dean–Stark apparatus until 18a was

no longer present by TLC (see Table 2). The base was ex-

tracted into 1 mol L^–1HCl; the extract was made basic with

solid sodium carbonate and extracted with ethyl acetate. Af-

ter the usual processing the crude base was converted into

the known picrate (See Table 2 for yields). The bases, liber-

ated from the picrates, were also fully characterized by

NMR spectroscopy.

2,3-Dimethylquinoline (21b): 1.5 mol L^–1HCl

The enone 17b was converted into 21b exactly as de-

scribed above to give the picrate in 86.7% yield.

474

Can. J. Chem. Vol. 82, 2004

Ytterbium triflate as catalyst

2-n-Hexyl-3-n-pentyl-8-(2-formyl)anilinomethylquinoline

(25a)

Normal conditions

A 2 mmol reaction was carried out as described in the

section Normal conditions above, except that the acid extrac-

tion to remove 24a was not carried out. The toluene solution

of the product mixture was washed with saturated NaCl so-

lution and then worked up as usual to give an oil (439 mg),

which was absorbed onto silica gel (2.2 g) and subjected to

column (3.5 × 17 cm) chromatography on silica gel (88 g)

using hexane – ethyl acetate (90:10, 25 mL fractions). F 5–7

contained 25a (104–179 mg, 25%–43%) as a yellow oil. IR:

see Discussion section. UV (nm): 385 (ε 6520), 319.5 (ε

4620), 306.5 (ε 4330), 261.5 (ε 10 100), 229 (ε 57 700), 213

(ε 59 600). ¹H NMR: see Discussion section. ¹³C NMR

(CD₂Cl₂) δ: 14.58, 14.69, 23.34, 23.48, 28.93, 29.25, 30.20,

32.58, 32.71, 33.00, 36.12, 43.39, 112.23, 115.47, 119.49,

125.88, 126.85, 127.00, 127.92, 135.32, 136.27, 136.45,

137.31, 162.00, 194.50. MS m/z (%): 417 ([M + H]⁺, 56),

312 (65), 296 (100), 212 (28), 198 (31), 171 (27). HR-

FAB⁺-MS m/z calcd. for [C₂₈H₃₆N₂O]H⁺: 417.2897; found:

417.2895.

A solution of 18a (121 mg, 1 mmol) and the appropriate

aldehyde (1.2–1.4 mmol, freshly liberated from the bisulfite

adduct) in toluene (10 mL) containing suspended ytterbium

triflate (62 mg, 0.1 mmol) was stirred at reflux temperature

(argon atmosphere) for 4 h. The cooled mixture was vigor-

ously stirred with 10 wt % sodium carbonate solution for

15 min, and the toluene phase was separated. The bases

were extracted into 1 mol L^–1HCl and processed as de-

scribed above. See Table 2 for picrate yields. When the

product was 3-phenylquinoline, HCl extraction often caused

the insoluble salt to crystallize and form an emulsion. In this

case, it was necessary to alternatively extract with HCl and

water, until all of the salt had been removed.

Modified conditions: Preparation of quinolines 24a and 24c-e

A solution of 18a or 18b (2 mmol) and the aldehyde (2.4–

2.5 mmol) in toluene (15 mL) was added dropwise (argon

atmosphere) to a stirred suspension of ytterbium triflate

(124 mg, 0.2 mmol) in toluene (15 mL) at reflux tempera-

ture over the time period indicated in Table 2 in parentheses.

Heating at reflux temperature was then continued until the

total elapsed time indicated in Table 2. The cooled mixture

was vigorously stirred with 10 wt % sodium carbonate solu-

tion for 15 min, and the toluene phase was separated. The

subsequent workup depended on the specific compound. For

24a and 24c, the bases were extracted with 1 mol L^–1HCl

and liberated with sodium carbonate as described above. The

yield of 24a corresponds to that of the picrate, whereas that

of 24c refers to the free base. See below for 24d and 24e.

2-Ethyl-3-methyl-8-(2-formyl)anilinomethylquinoline (25b)

A mixture of the isomeric tetrahydroquinolines 26d-e

(227 mg, 0.704 mmol, see below for synthesis) in toluene

(25 mL) containing suspended ytterbium triflate (48 mg,

0.077 mmol) was stirred at reflux temperature under argon

for 22 h. The reaction mixture was stirred with sodium car-

bonate for 15 min, and then the toluene phase was processed

as described above to give a solid residue (206 mg), which

was absorbed onto silica gel (1.0 g) and purified by column

(3.5 × 14 cm) chromatography on silica gel (69 g) using

hexane – ethyl acetate (85:15, 20 mL fractions). F 3–5 con-

tained the product (76 mg), and F 7 contained the product

admixed with a slightly more polar material (8 mg). The ma-

jor fraction on crystallization from cyclohexane gave the

crystalline product in two crops, mp 145–147 °C (47 mg)

and mp 144–146 °C (2 mg). The minor fraction gave an ad-

ditional 2 mg of product, mp 144.5–145 °C, from cyclohex-

ane. Total yields, 51 mg (23.8%). Recrystallization from

cyclohexane gave an analytical specimen that had mp 147–

148.5 °C after drying in vacuo (45 °C, 16 h). IR (KBr) (cm^–1):

3434, 3334, 2750, 1656. UV (nm): 386 (ε 6250), 318.5 (ε

3750), 304.5 (ε 3530), 263 (ε 9370), 228 (ε 51 600), 211 (ε

3-n-Pentyl-6,7-dimethoxyquinoline (24d)

The acid extract was made basic in the usual way and ex-

tracted with ethyl acetate. The extract was washed with satu-

rated NaCl solution and then processed as usual to give the

crude crystalline base (277 mg). This material was absorbed

onto silica gel (1.5 g) and subjected to column (3.5 ×

11.5 cm) chromatographic purification on silica gel (55.5 g)

using 1:1 hexane – ethyl acetate (25 mL fractions). F 10–13

contained the solid product (217 mg, 41.8%). After two

crystallizations from hexane and drying in vacuo for 24 h, it

1

had mp 63.5–65 °C. H NMR δ: 0.90 (t, 3H, J = 7.0 Hz),

1.29–1.40 (m, 4H), 1.65–1.75 (m, 2H), 2.74 (t, 2H, J =

7.6 Hz), 4.01 (s, 3H), 4.02 (s, 3H), 7.00 (s, 1H) 7.40 (s, 1H),

7.71 (d, 1H, J = 2.1 Hz), 8.57 (d, 1H, J = 2.1 Hz). ¹³C NMR

δ: 14.01, 22.51, 31.01, 31.37, 33.06, 56.00, 56.06, 104.78,

107.81, 123.72, 132.73, 133.68, 143.61, 149.58, 149.79,

151.71. Anal. calcd. for C₁₆H₂₁NO₂: C 74.10, H 8.16, N

5.40; found: C 74.25, H 8.16, N 5.56.

1

55 500). H NMR (CD₂Cl₂) δ: 1.45 (t, 3H, J = 7.4 Hz), 2.48

(d, 3H, J = 0.9 Hz), 3.02 (q, 2H, J = 7.4 Hz), 5.16 (bs, 2H),

6.65 (td, 1H, J = 0.9, 6.7 Hz), 6.83 (d, 1H, J = 8.5 Hz), 7.30

(td, 1H, J = 1.7, 7.8 Hz), 7.37 (d, 1H, J = 7.1 Hz), 7.47 (dd,

1H, J = 1.7, 7.8 Hz), 7.57 (dd, 1H, J = <1, >6 Hz), 7.65 (dd,

1H, J = <1, 8.3 Hz), 7.85 (bs, 1H), 8.98 (bs, 1H, exchanged

with D₂O), 9.83 (d, 1H, J = 0.6 Hz). ¹³C NMR δ: 12.29,

19.40, 29.55, 43.22, 125.67, 126.39, 126.51, 127.59, 130.03,

135.82, 136.06, 137.02, 144.86, 151.47, 162.36, 194.25.

Anal. calcd. for C₂₀H₂₀N₂O: C 78.92, H 6.62, N 9.20; found:

C 79.25, H 6.58, N 9.35.

3-Phenyl-6,7-dimethoxyquinoline (24e)

The toluene solution containing the reaction mixture was

not extracted with acid. Instead, after washing with saturated

NaCl solution and the usual workup, the crude product from

a 1 mmol reaction was absorbed onto silica gel (1.5 g) and

subjected to column (2.2 × 20 cm) chromatographic purifi-

cation on silica gel (35.5 g) using 1:1 hexane – ethyl acetate

(20 mL fractions). F 9–18 contained the product (208 mg,

2-Methyl-8-(2-formyl)anilinomethylquinoline (25c)

A solution of a mixture of 26a and 26b (203 mg,

0.69 mmol, see below for synthesis) in toluene (20 mL) con-

taining suspended ytterbium triflate (86 mg) was stirred at

reflux temperature (argon atmosphere) for 4 h and then

1

78.4%), the H NMR spectrum of which was identical to

that reported (47).

Muchowski and Maddox

475

worked up as described for 25b. The crude product

(177 mg) was absorbed onto silica gel (1 g) and subjected to

column (3.5 × 11.5 cm) chromatographic purification on sil-

ica gel (54 g) using hexane – ethyl acetate (85:15, 20 mL

fractions). F 5–10 contained the product (121 mg), which

was crystallized from cyclohexane to give the desired mate-

rial in two crops, mp 95–95.5 °C (85 mg) and mp 94.5–

95 °C (10 mg). Thus the yield of 25c was 95 mg (46.5%).

Recrystallization of a sample from cyclohexane gave the an-

alytical specimen, which had mp 94.5–95.5 °C after drying

in vacuo (9.5 h). IR (KBr) (cm^–1): 3480 (bw), 3350, 2738,

1666. UV (nm): 383 (ε 6850), 317 (ε 3530), 304 (ε 3500),

267 (ε 9980), 224 (ε 52 900), 208 (ε 51 900), 206.5 (ε

and F 10–12 contained pure 27 (50 mg), all as oils. See Ta-

ble 3 for the ¹H NMR spectra of 26c-e. The aldehyde 27 was

a single isomer. ¹H NMR (CD₂Cl₂) δ: 0.96 (t, 3H, J =

7.4 Hz), 1.19 (d, 3H, J = 7.2 Hz), 1.65 (m, 2H), 2.68 (m,

1H, J = 1.2, 4.6, 7.2 Hz), 3.96 (m, 1H), 6.69 (td, 1H, J = 0.9,

7.4 Hz), 6.79 (d, 1H, J = 8.6 Hz), 7.39 (td, 1H, J = 1.7,

7.9 Hz), 7.47 (dd, 1H, J = 1.7, 7.7 Hz), 8.41 (bd, 1H, J =

9.7 Hz, exchanged with D₂O), 9.69 (d, 1H, J = 1.2 Hz), 9.79

(d, 1H, J = 0.7 Hz). ¹³C NMR (CD₂Cl₂) δ: 9.07, 11.05,

26.55, 50.01, 53.94, 111.51, 115.47, 119.01, 136.19, 137.30,

150.90, 194.46, 203.92. MS m/z (%): 219 (87), 190 (20),

161 (44), 162 (100), 144 (100), 132 (31), 122 (45), 118 (47),

93 (42). HR-FAB⁺-MS m/z calcd. for [C₁₃H₁₇NO₂]H⁺:

220.1337; found: 220.1341.

1

52 500). H NMR δ: 2.78 (s, 3H), 5.17 (d, 2H, J = 5.9 Hz),

6.65 (td, 1H, J = 0.8, 7.4 Hz), 6.76 (d, 1H, J = 8.5 Hz), 7.29

(td, 1H, J = 1.6, 8.6 Hz), 7.38 (t, 3H, J = 7.6 Hz), 7.47 (dd,

1H, J = 1.6, 7.7 Hz), 7.61 (dd, 1H, J = 1.0, 7.1 Hz), 7.66 (d,

1H, J = 8.1 Hz), 8.02 (d, 1H, J = 8.4 Hz), 9.02 (bs, 1H, ex-

changed with D₂O), 9.87 (s, 1H). ¹³C NMR δ: 25.93, 43.21,

111.99, 115.19, 119.08, 122.37, 125.71, 126.82, 127.14,

127.42, 135.95, 136.09, 137.03, 146.30, 151.41, 158.68,

194.32. Anal. calcd. for C₁₈H₁₆N₂O: C 78.23, H 5.84, N

10.14; found C 78.50, H 5.89, N 10.22.

A reaction effected at 110 °C had a very different product

spectrum. Thus, a 2 mmol reaction was carried out in tolu-

ene (40 mL) in a pressure bottle for 6 h. Workup as de-

scribed above gave a mixture, which by TLC (hexane – ethyl

acetate (85:15)) showed four major products with Rfs 0.88,

0.78, 0.72, and 0.67. The most polar material corresponded

to 26c. The mixture was absorbed onto silica gel (1.5 g) and

subjected to column (3.5 × 12 cm) chromatography on silica

gel (60 g) using hexane – ethyl acetate (85:15, 25 mL frac-

tions). F 8–10 (112 mg) partially crystallized, and after

slurrying with 95:5 hexane – ethyl acetate, crystalline 25b

(8.8 mg, 1.4%), mp 144 to 145 °C, was collected by filtra-

tion. F 11–13 (102 mg) contained impure 26c. F 14,15

Synthesis of the tetrahydroquinolines 26

A solution of 18a (242 mg, 2 mmol) and the appropriate

aldehyde (8 mmol) in toluene (20 mL) containing suspended

ytterbium triflate (124 mg, 0.2 mmol) was stirred at room

temperature under argon for the time period indicated below.

The reaction mixture was then vigorously stirred with excess

10 wt % sodium carbonate solution for 15 min; the toluene

phase was separated, washed with saturated NaCl solution,

and worked up as usual. The bases were purified as de-

scribed below.

1

(11 mg) was pure 26c. H NMR: see Table 3. ¹³C NMR

(CD₂Cl₂) δ: 6.63, 10.96, 26.46, 54.21, 111.84, 115.36,

116.17, 117.53, 119.66, 122.58, 133.08, 135.78, 136.80,

135.78, 136.80, 137.70, 148.26, 151.17, 194.17, 194.94. MS

m/z (%): 322 (32), 307 (13), 293 (34), 202 (100), 186 (55),

172 (29), 149 (47). HR-EI-MS m/z calcd. for C₂₀H₂₂N₂O₂:

322.1681; found: 322.1667.

trans,trans-2-Benzyl-3-phenyl-4-(2-formyl)anilino-1,2,3,4-

tetrahydroquinoline-8-carboxaldehyde (26f)

cis- And trans-2-methyl-4-(2-formyl)anilino-1,2,3,4-

tetrahydroquinoline-8-carboxaldehydes 26a and 26b

The crude product from a 2 mmol reaction was absorbed

onto silica gel (1.5 g) and purified by column (3.5 ×

12.5 cm) chromatography on silica gel (63.5 g) using hexane –

ethyl acetate (80:20, 25 mL fractions). F 4–7 contained the

product as a TLC homogeneous oil (191–210 mg, 65%–

Reaction was carried out in the usual manner, except that

the ratio of phenylacetaldehyde to 18a was 2 and the reac-

tion time was 41 h. Removal of the toluene from a 2 mmol

reaction gave a residue, which partially crystallized. The

mixture was slurried with toluene; the insoluble solid was

collected by filtration and dried to give 26f (130.5 mg,

14.6%), mp 235–238 °C. A sample was dissolved in hot 1,2-

dichloroethane, diluted with four volumes of hexane, and

seeded to give material with mp 236–238 °C, after drying in

vacuo (75 °C, 6 h). IR (KBr) (cm^–1): 3428, 3292, 2751,

1

72%), which by H NMR spectroscopy was a 4:1 mixture of

26a and 26b. IR (film) (cm^–1): 3317, 2744, 1653. UV (nm):

1

388 (ε 12 000), 262.5 (ε 11 400) 237 (ε 33 800). H NMR:

see Table 3. ¹³C NMR (e,e-isomer 26a) δ: 22.51, 36.34,

46.93, 49.57, 111.39, 115.05, 115.84, 117.59, 118.99,

123.35, 132.40, 135.84, 136.48, 137.41, 147.93, 150.75,

193.96, 194.53. MS m/z (%): 294 (27), 279 (28), 174 (100).

HR-FAB⁺-MS m/z calcd. for C₁₈H₁₈N₂O₂: 294.1368; found:

294.1362.

1

2733, 1652. H NMR: see Table 3. ¹³C NMR (CD₂Cl₂) δ:

40.64, 49.56, 55.97, 58.06, 111.71, 115.04, 115.66, 118.07,

118.67, 124.25, 127.25, 127.67, 128.57, 128.79, 129.05,

129.17, 129.38, 129.88, 133.01, 135.77, 136.95, 137.86,

140.14, 146.85, 151.11, 193.68, 194.14. Anal. calcd. for

C₃₀H₂₆N₂O₂: C 80.69, H 5.87, N 6.27; found: C 80.71, H

5.81, N 6.41.

2-Ethyl-3-methyl-4-(2-formyl)anilino-1,2,3,4-

tetrahydroquinoline-8-carboxaldehydes (26c-e) and 2-

methyl-3-(2-formyl)anilinopentanal (27)

A reaction effected in the usual manner was complete in

2 days. The mixture from a 2 mmol reaction was absorbed

onto silica gel (2 g) and purified by column (4 × 21.5 cm)

chromatography on silica gel (143 g) using hexane – ethyl

acetate (85:15, 50 mL fractions). F 5–7 contained 26c-e

(135 mg); F 8,9 contained a 1:1 mixture of 26c-e and 27;

1,2-Dihydro-2-benzyl-3-phenylquinoline-8-

carboxaldehyde (29)

The reaction was carried out as described for the synthesis

of 25a except that the reaction time was 6 h, and the 3-

phenylquinoline was removed by extraction with 1 mol L^–1

HCl. The toluene phase containing 29 was then washed suc-

476

Can. J. Chem. Vol. 82, 2004

cessively with 10 wt % sodium carbonate and saturated

NaCl solution and then worked up as usual. The crude prod-

uct from a 2 mmol reaction was absorbed onto silica gel

(2 g) and purified by column (4 × 17 cm) chromatography

on silica gel (110 g) using hexane – 1,2-dichloroethane

(35:65, 25 mL fractions). F 9–13 contained the product (52–

115 mg, 13%–29% yield based on phenylacetaldehyde),

which was crystalline. After two crystallizations from cyclo-

hexane and drying in vacuo (7 h), it had mp 119.5–121.5 °C.

IR (CCl₄) (cm^–1): 3313, 2737, 1663. UV (nm): 420.5 (ε

were extracted into ethyl acetate. The extract was washed

with saturated NaCl solution and then worked up as usual.

2,3-Dimethylquinoline (21b) and 2-ethylquinoline (37a)

The crude product was absorbed onto silica gel (1 g) and

purified by column (3.5 × 10.5 cm) chromatography on sil-

ica gel (49 g) using hexane – ethyl acetate (80:20, 20 mL

fractions). F 7–9 contained 37a (54 mg, 28.2%, converted

into the picrate, mp 147–149 °C, lit.(48) value mp 148 °C);

F 11–19 contained crystalline 21b (54 mg, 47.7%), mp 67–

78 °C.

1

9330), 346.5 (ε 3940), 269 (ε 23 800), 234 (ε 18 600). H

NMR: see Discussion. ¹³C NMR (CD₂Cl₂) δ: 43.28, 55.99,

115.99, 118.25, 122.20, 122.67, 125.75, 126.88, 128.36,

128.65, 130.08, 132.61, 135.22, 135.28, 137.66, 138.40,

146.29, 193.44. Anal. calcd. for C₂₃H₁₉NO: C 84.89, H 5.89,

N 4.30; found: C 84.66, H 5.75, N 4.49.

2-Methyl-3-ethylquinoline (21c) and 2-n-propylquinoline

(37b)

Purified in exactly the same manner as described for 21b

and 37a. F 2–5 contained 37b (70 mg, 40.9%, identified by

NMR), and F 8–12 contained 21c (59 mg, 34.5%, mp 69.5–

71.5 °C, lit.(49) value mp 74–74.5 °C).

2-Benzyl-3-phenylquinoline-8-carboxaldehyde (30)

2-Methyl-3-isopropylquinoline (21d) and 2-isobutylquinoline

(37c)

Purified as described above. F 3–5 contained 37c

(112 mg, 60.5%, picrate mp 163–166 °C, lit.(50) value mp

163 to 164 °C), and F 8–12 contained 21d (8 mg, 4.3%). See

below for characterization.

A solution of potassium nitrosodisulfonate (Fremy’s salt,

90 mg, 0.335 mmol) in 4 wt % aqueous sodium carbonate

(6 mL) was added to a stirred solution of 29 (87 mg,

0.267 mmol) in acetonitrile (9 mL). This caused precipita-

tion of some 29, but after 1 h a clear solution was obtained.

After 6 h, the solution was diluted with a large volume of

water and extracted with ethyl acetate. The extract was

washed with saturated NaCl solution and then worked up as

usual. The crude product was absorbed onto silica gel (0.5 g)

and subjected to column (2.2 × 15 cm) chromatographic pu-

rification on silica gel (27.5 g) using hexane – 1,2-

dichloroethane (35:65, 15 mL fractions). F 4,5 contained the

starting material (19 mg), and F 7–21 contained the product

(46.5 mg, 66% yield based on starting material consumed).

On crystallization from cyclohexane 29 was obtained in two

crops, mp 131–132.5 °C (37.5 mg) and mp 126.5–129.5 °C

(3.5 mg). Recrystallization of the first crop from cyclohex-

ane gave pure 29, which had mp 132–132.5 °C after drying

in vacuo (6 h). IR (CCl₄) (cm^–1): 1692. ¹H NMR (CD₂Cl₂) δ:

4.35 (s, 3H), 7.04–7.07 (m, 2H), 7.12–7.20 (m, 3H), 7.31–

7.34 (m, 2H), 7.43–7.48 (m, 3H), 7.65 (td, 1H, J = 0.7,

7.4 Hz), 8.06 (s, 1H), 8.09 (dd, 1H, J = 1.5, 8.1 Hz), 8.26

(dd, 1H, J = 1.5, 7.2 Hz), 11.46 (d, 1H, J = 0.7 Hz). ¹³C

NMR (CD₂Cl₂) δ: 43.15, 126.32, 126.48, 127.44, 128.28,

128.49, 128.82, 128.88, 129.58, 129.88, 131.69, 134.27,

136.94, 137.19, 139.51, 139.63, 146.84, 160.91, 193.11.

Anal. calcd. for C₂₃H₁₇NO: C 85.42, H 5.30, N 4.33; found:

C 85.62, H 5.28, N 4.50.

From ketones under alkaline conditions

2-Methyl-3-ethylquinoline (21c) and 2-n-propylquinoline

(37b)

A solution of 18a (61 mg, 0.5 mmol) and 2-pentanone

(71 mg, 0.82 mmol) in ethanol (3 mL) and 1 mol L^–1NaOH

(1.75 mL) was left at room temperature for 24 h, diluted

with water, and made acidic with 1 mol L^–1HCl. The acid

solution was extracted with toluene, and then the acid phase

was made basic with solid sodium carbonate. The products

were extracted with toluene; the extract was washed with

saturated NaCl solution and worked up as usual to give the

1

mixture of bases (73 mg, 85.2%), which by H NMR was

shown to be a 77:23 mixture of 21c and 37b.

2-Methyl-3-isopropylquinoline (21d) and 2-isobutylquinoline

(37c)

A solution of 18a (121 mg, 1 mmol) and 4-methyl-2-

pentanone (150 mg, 1.5 mmol) in ethanol (6 mL) and

1 mol L^–1NaOH (3.5 mL) was heated at reflux temperature

under argon for 4 h. The reaction was worked up exactly as

described above to give 37c (144 mg, 77.7%) and 21d

1

(10 mg, 5.4%). Compound 21d was an oil. H NMR δ: 1.34

(d, 6H, J = 6.8 Hz), 2.78 (s, 3H), 3.26 (sept., 1H, J =

6.8 Hz), 7.45 (td, 1H, J = 1.1, 8.0 Hz), 7.61 (td, 1H, J = 1.5,

7.7 Hz), 7.74 (dd, 1H, J = 1.2, 81. Hz), 7.92 (s, 1H), 7.99 (d,

1H, J = 8.4 Hz). ¹³C NMR δ: 23.52, 23.54, 29.75, 125.98,

127.49, 127.92, 128.63, 128.86, 131.56, 140.93, 146.51,

158.44. The base was converted into the picrate, mp 216–

218 °C decomposition (dec), in hot ethanol. The mp was un-

changed on recrystallization and drying in vacuo (60 °C,

14.5 h). Anal. calcd. for C₁₉H₁₈N₄O₇: C 55.07, H 4.38, N

13.52; found: C 55.12, H 4.43, N 13.49.

Synthesis of quinolines 21 and 37 from 2-

aminobenzaldehyde (18a) and methyl alkyl ketones

Neat ketones

A solution of 18a (121 mg, 1 mmol) in the neat ketone

(10 mL) containing ytterbium triflate (62 mg, 0.1 mmol) was

placed in an oil bath at 95 °C and stirred in an argon atmo-

sphere for the time given in Table 2, and then the excess

ketone was removed in vacuo. Toluene (20 mL) was added

to the residue, and the mixture was stirred at reflux tempera-

ture in an argon atmosphere for 7 h. After the usual treat-

ment with 10 wt % sodium carbonate solution, the toluene

phase was extracted with 1 mol L^–1HCl; the extract was

made basic with solid sodium carbonate, and the products

From pyrrolidine enamines 38

A solution of 18a (121 mg, 1 mmol) and 38a (51) or 38b

((52), 2.4 mmol) in anhydrous toluene (10 mL) containing

Muchowski and Maddox

477

10. Y. Tamura, T. Tsugoshi, S. Mohri, and Y. Kita. J. Org. Chem.

50, 1542 (1985).

11. I-S. Cho, L. Gong, and J.M. Muchowski. J. Org. Chem. 56,

7288 (1991).

12. J.I. Úbeda, M. Villacampa, and C. Avendaño. Synthesis, 1176

(1998).

13. O.B. Familoni, P.T. Kaye, and P.H. Klaas. Chem. Commun.

2563 (1998).

ytterbium triflate (62 mg, 0.1 mmol) was either left at room

temperature or heated at reflux temperature (see Table 2) in

an argon atmosphere. HCl (1 mol L^–1, 10 mL) was added to

the reaction mixture, which was stirred vigorously at room

temperature for 15 min. The acidic aqueous phase was

worked up as described above to give the crude mixture of

bases. This mixture was spiked with hexamethylbenzene (in-

1

ternal standard) and analyzed by H NMR.

14. Y. Hsiao, N.R. Rivera, N. Yasuda, D.L. Hughes, and P.J.

Reider. Org. Lett. 3, 1101 (2001).

15. V.B. Drewsen. Ber. 16, 1953 (1883).

By arylhydroamination of 1-heptyne

A solution of 18a (121 mg, 1 mmol), mercuric chloride

(472 mg, 1 mmol), and 1-hepytne (0.7 mL, 513 mg,

5.33 mmol) in anhydrous THF (15 mL) was stirred at room

temperature for 18 h in an argon atmosphere. Ytterbium

triflate (62 mg, 0.1 mmol) was added, and stirring was con-

tinued for a total time of 3 days. Even though 18a was still

present by TLC, the solvent was removed in vacuo, water

and ethyl acetate were added, and the mixture was filtered

through Celite. The organic phase was separated and ex-

tracted with 1 mol L^–1HCl. The acidic extract was worked

up as in the section From pyrrolidine enamines 38 above;

the crude mixture was spiked with hexamethylbenzene and

16. A. Baeyer and V. Drewsen. Ber. 15, 2854 (1882).

17. S-D. Cho, W-Y. Choi, S-G. Lee, Y-J. Yoon, and S.C. Shin. Tet-

rahedron Lett. 37, 7059 (1996).

18. K. Ramadas and N. Srinivasan. Synth. Commun. 22, 3189

(1992).

19. I.L. Smith and J.W. Opie. Org. Synth. Coll. Vol. 3, 56 (1955).

20. B.D. Foy, R.A. Smudde, and W.F. Wood. J. Chem. Educ.

70 322 (1993).

21. G.P. Black, F. Dimon, S. Fratucello, P.J. Murphy, M. Nielson,

and H.L. Williams. Tetrahedron, 52, 8001 (1996).

22. A.T. Nielsen and W.J. Houlihan. Org. Syn. 16, 1 (1968).

23. M.B. Smith and J. March. In March’s advanced organic chem-

istry. Reactions, mechanisms, structure. 5th ed. John Wiley &

Sons, Inc. New York. 2002. pp. 285–286.

1

analyzed by H NMR.

24. H.O. House and R.S. Ro. J. Am. Chem. Soc. 80, 2428

(1958).

Note added in proof

25. D.S. Noyce and M.J. Jorgensen. J. Am. Chem. Soc. 83, 2525

(1961).

26. J.E. Dubois and M. Dubois. Compt. Rend. 256, 715 (1963).

27. T. Vondrák and M. Procházka. Coll. Czech. Chem. Commun.

48, 82 (1983).

28. D.M. Sammond and T. Sammakia. Tetrahedron Lett. 37, 6065

(1996).

29. E.A. Fehnel. J. Org. Chem. 31, 2899 (1966).

30. G. Barbieri, P. Lazzaretti, L. Schenetti and F. Taddei. Org.

Magn. Res. 7, 451 (1975).

The Friedländer synthesis of quinolines from o-

aminoarylketones and ketones also proceeds by a rate-

determining aldol mechanism. This is strongly supported by

the observation that heating a toluene solution (reflux, 17 h)

of 2-hexanone (4 equiv.), o-aminoacetophenone (1 equiv.),

and pyrrolidine (2 equiv.), containing suspended anhydrous

magnesium sulfate (10 equiv., in situ generation of the

pyrrolidine enamine of 2-hexanone (53)), and ytterbium

triflate (0.1 equiv.), gave a 5.5:1 mixture (ca. 25% yield) of

2-n-butyl-4-methylquinoline (kinetic) and 2,4-dimethyl-3-n-

propylquinoline (thermodynamic). In contrast, effecting the

condensation of 2-hexanone and o-aminoacetophenone in

acetic acid (reflux), containing catalytic sulfuric acid (29),

gave a 1:1.7 kinetic to thermodynamic product ratio. Com-

parable results were obtained starting with 2-butanone and

o-aminoacetophenone.²

31. M. Funabashi, M. Iwakawa, and J. Yoshimura. Bull. Chem.

Soc. Jpn. 42, 2885 (1969).

32. S. Talukdar, C-T. Chen, and J-M. Fang. J. Org. Chem. 65,

3148 (2000).

33. P.A. Wehrli and B. Schaer. Synthesis, 288 (1974).

34. J.B. Wommack, T.G. Barbee, D.J. Thoennes, M.A. McDonald,

and D.E. Pearson. J. Heterocycl. Chem. 6, 243 (1969).

35. E.M. Hawes and D.K.J. Gorecki. J. Heterocycl. Chem. 11, 151

(1974).

36. G. Kempter, P. Andratschke, D. Heilmann, H. Krausmann, and

References

M. Mietasch. Chem. Ber. 97, 16 (1964).

37. P.G. Dormer, K.K. Eng, R.N. Farr, G.R. Humphrey, J.C.

McWilliams, P.J. Reider, J.W. Sager, and R.P. Volante. J. Org.

Chem. 68, 467 (2003).

38. P. Dalla Croce, P. Ferruti, and R. Stradi. Gazz. Chim. Ital. 97,

589 (1967).

39. D. Pocar, G. Bianchetti, and P. Ferruti. Gazz. Chim. Ital. 97,

597 (1967).

40. G.W. Fischer. J. Prakt. Chem. 330, 981 (1988).

41. J. Barluenga, F. Aznar, R. Liz, and R. Rodes. J. Chem. Soc.

Perkin Trans. 1, 2732 (1980).

1. P. Friedländer. Ber. 15, 2572 (1882).

2. P. Friedländer and C.F. Gohring. Ber. 16, 1833 (1883).

3. E. Reimann. In Houben-weyl. Methoden der organischen

chemie. Hetarene II. Part 1. Vol. E7a. G. Thieme Verlag,

Stuttgart. 1991. pp. 323–337.

4. C-C. Cheung and S-J. Yan. Org. React. 28, 37 (1982).

5. G. Jones. In The chemistry of heterocylic compounds. Vol. 32.

Quinolines. Part 1. Edited by G. Jones. John Wiley & Sons,

New York. 1977. pp. 181–191.

6. V. Snieckus. Chem. Rev. 90, 879 (1990).

42. A. Capperucci, A. Degl’Innocenti, M. Funicello, G. Mauriello,

P. Scafato, and P. Spagnolo. J. Org. Chem. 60, 2254 (1995).

43. H. Horina and N. Inoue. J. Org. Chem. 46, 4416 (1981).

44. A. Rilliet. Helv. Chim. Acta, 5, 547 (1922).

7. T.G. Majewicz and P. Caluwe. J. Org. Chem. 40, 3407 (1975).

8. K. Schofield and R.S. Theobald. J. Chem. Soc. 395 (1950).

9. E.A. Fehnel, J.A. Deyrup, and M.B. Davidson. J. Org. Chem.

23, 1996 (1958).

²J.M. Muchowski. Unpublished results.

478

Can. J. Chem. Vol. 82, 2004

45. J. Buckingham (Editor). Dictionary of organic compounds.

Vol. 4. 5th ed. Chapman & Hall, London. 1982. p. 4053.

46. J.H. Markgraf, R.J. Katt, W.L. Scott, and R.N. Sherfin. J. Org.

Chem. 34, 4131 (1969).

50. T. Higashino, K. Suzuki, and E. Hayashi. Chem. Pharm. Bull.

23, 746 (1975).

51. R. Jacquier, C. Petrus, and F. Petrus. Bull. Soc. Chim. Fr. 2845

(1966).

47. M.P. Maguire, K.R. Sheets, K. McVety, A.P. Spada, and A.

52. J. Koyama, I. Toyokuni, and K. Tagahara. Chem. Pharm. Bull.

46, 332 (1998).

Ziberstein. J. Med. Chem. 37, 2129 (1994).

48. R. Delaby and J. Hiron. Bull. Soc. Chim. Fr. 47, 1395 (1930).

49. N. Satoshi. Yakugaku Zasshi. 81, 479 (1961); Chem. Abstr.

55, 105 853 (1961).

53. P.A. Zoretic, F. Barcelos, and B. Branchaud. Org. Prep.

Proced. Int. 8, 211 (1976).

Article Doi

Concerning the mechanism of the Friedlaender quinoline synthesis

DOI: 10.1139/v03-211

Source and publish data:

Authors:

Article abstract of DOI:10.1139/v03-211

Full text of DOI:10.1139/v03-211

Products guided by the article

R&D Labs maybe for 57044-36-7

Relevant to this article

Hot Product