Competition between cyclisation and bisimine formation in the reaction of 1,3-diaminopropanes with aromatic aldehydes

Article abstract of DOI:10.1016/j.tet.2009.10.060

Condensation of 1,3-diamines with aldehydes or ketones gives rise to two major products, the hexahydropyrimidine and the bisimine. Experimental studies of the reaction between a range of aromatic aldehydes and 1,3-diaminopropane or 1,3-diamino-2-propanol establish that the hexahydropyrimidine is favoured by the less nucleophilic amine and by the presence of electron withdrawing groups on the aryl ring of the aldehyde. Calculations indicate that the electronic nature of this aryl ring substituent influences both the relative thermodynamic stability of the final products and the reactivity of the aldehyde as an electrophile.

Full text of DOI:10.1016/j.tet.2009.10.060

Tetrahedron 65 (2009) 10685–10692
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Competition between cyclisation and bisimine formation in the reaction
of 1,3-diaminopropanes with aromatic aldehydes
,
,
,
Julie M. Locke a y, Renate Griffith b z, Trevor D. Bailey a, Robyn L. Crumbie a
*
a Department of Chemistry, University of Western Sydney, Campbelltown Campus, Penrith, NSW 1795, Australia
b School of Chemistry, University of Wollongong, Northfields Avenue, Wollongong, NSW 2522, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
Condensation of 1,3-diamines with aldehydes or ketones gives rise to two major products, the
hexahydropyrimidine and the bisimine. Experimental studies of the reaction between a range of aromatic
aldehydes and 1,3-diaminopropane or 1,3-diamino-2-propanol establish that the hexahydropyrimidine is
favoured by the less nucleophilic amine and by the presence of electron withdrawing groups on the aryl
ring of the aldehyde. Calculations indicate that the electronic nature of this aryl ring substituent influences
both the relative thermodynamic stability of the final products and the reactivity of the aldehyde as an
electrophile.
Received 2 June 2009
Received in revised form
29 September 2009
Accepted 15 October 2009
Available online 21 October 2009
Keywords:
Hexahydropyrimidine
Imine
Ó 2009 Elsevier Ltd. All rights reserved.
Hammett constant
Frontier molecular orbital
Structure–reactivity relationship
LUMO coefficient
1. Introduction
Condensation of 1,3-diamines with aldehydes and ketones is
a well documented reaction. In particular, syntheses of multi-
dentate ligands for metal complexes frequently employ the re-
action of 1,3-diaminopropane (1) with benzaldehydes (2).1–4
However, two potential products may result from this reaction
(Scheme 1), the bisimine (3) and the hexahydropyrimidine (4).
Initially the monoimine (5) or analogous monoiminium ion (5Hþ)
forms and this intermediate can subsequently react with another
molecule of aldehyde to furnish the bisimine 3 as the final product.
Alternatively, the free amino group of the monoimine may
attack the imine carbon in a 6-endo-trig ring closure5 to yield the
hexahydropyrimidine 4.
Historically, aromatic carbonyl compounds have been used to
make bisimines whilst aliphatic carbonyl compounds have been
employed in the synthesis of hexahydropyrimidines.6,7 Interest in
2-arylhexahydropyrimidines often only develops after their initial
Scheme 1. Possible products from the condensation reaction of 1,3-diaminopropane
(1) with benzaldehydes (2).
* Corresponding author. Tel.: þ61 2 46203201; fax: þ61 2 46203025.
E-mail address: r.crumbie@uws.edu.au (R.L. Crumbie).
Present address: Intelligent Polymer Research Institute, University of Wollon-
y
appearance as unwanted side-products in the course of imine
synthesis,8–10 and most modern studies of hexahydropyrimidines
are devoted to characterising the ring-chain tautomerism in these
gong, Northfields Avenue, Wollongong, NSW 2522, Australia.
z
Present address: School of Medical Sciences, University of New South Wales,
Sydney, NSW 2052, Australia.
0040-4020/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2009.10.060

10686
J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
and related 1,3-diheterocyclic systems.11–17 Factors influencing the
product distribution from the reaction of 1,3-diaminopropane (1)
with benzaldehydes have been elucidated to some extent. It has
been shown that an equimolar reaction of the diamine 1 with
various aryl aldehydes gives a mixture of hexahydropyrimidine (4)
and monoimine (5) and that the ratio of products formed depends
upon both electronic and steric factors.9,11
We were interested in probing the influence of intramolecular
hydrogen bonding in determining the conformations of a range of
5-hydroxyhexahydropyrimidines.18 Our early synthetic efforts to-
wards 2-arylhexahydropyrimidines frequently gave complex and
inseparable product mixtures. We therefore turned our attention to
investigating the reactions of 1,3-diamino-2-propanol (6, Scheme 2)
and 1,3-diaminopropane (1) with a series of benzaldehydes (2a–j,
Table 1) to establish the type of products formed and examine the
factors that influence the product distribution in these reactions.
2. Results and discussion
2.1. Experimental product ratios
The diamines 1 and 6 were reacted with a series of variously
substituted benzaldehydes (2a–j) in 1:1 and 1:2 (amine/aldehyde)
molar ratios (Schemes 1 and 2) and the distribution of products
from their condensation reactions was determined. Heating the
diamines and aldehydes in a methanolic solution at reflux provided
crude products from the condensation reaction. Attempts to isolate
products from the reaction mixtures were unsuccessful, and con-
sequently preliminary characterisation of the bisimines and
hexahydropyrimidines in the crude reaction mixtures was accom-
plished through comparison with the 1H NMR spectra of the
bisimines derived from o-hydroxybenzaldehyde (3i and 8i),19,20 as
well as hexahydropyrimidine6,18 and 5-hydroxyhexahydropyr-
imidine.18,21 Characterisation of the products derived from
1,3-diamino-2-propanol (6, Scheme 2) was also facilitated by
comparison of their 1H NMR spectra with those obtained for the
crude products from the reaction of 6 and 2j, where the sole
product obtained was a mixture of cis- and trans-5-hydroxy-2-
(20-pyridinyl)hexahydropyrimidine (9j). Product ratios in the crude
reaction mixtures were determined from the 1H NMR spectra in
CDCl3 via integration values of the well-separated singlets arising
from the proton/s on C2 of the hexahydropyrimidines (4–5 ppm)
and the azomethine proton (8.0–8.5 ppm, see Supplementary
data). The 1H NMR spectra of the crude reaction products contained
no signals from unreacted diamine. In all cases, examination of the
EI mass spectra of the crude reaction products showed no peaks
with greater m/z than the expected parent ion for the bisimine.
Reaction of the diamine 1 with two molar equivalents of various
aldehydes (2a–j) almost exclusively yielded the bisimines 3a–j. The
1H and 13C NMR spectra of bisimines derived from 2a–i indicated
that the compounds were pure and data obtained for known
compounds compared well to available literature values. Conden-
sation of 1 with the pyridyl derivative 2j in a 1:2 (amine/aldehyde)
molar ratio yielded a crude reaction mixture containing a signifi-
cant proportion (34%) of hexahydropyrimidine 4j as well as the
aldehyde 2j and bisimine 3j. The production of a substantial
amount of hexahydropyrimidine in this reaction is most likely
a consequence of the electron withdrawing nature of the pyridyl
ring (see below).
Scheme 2. Possible products from the condensation reaction of 1,3-diamino-2-prop-
anol (6) with benzaldehydes (2).
Condensation of the diaminoalcohol 6 with two molar equiva-
lents of aldehydes 2a–j also gave the bisimine as the major product.
However, in the 1H NMR spectra of the crude reaction mixtures
from the reaction of 6 with 2a, 2b and 2e, traces of the cis and trans
hydroxyhexahydropyrimidines (9) and the associated aldehydes
were also observed. None of the 1H NMR spectra of the crude
reaction mixtures from 6 contained signals attributable to the
aminomethyloxazolidines (10a–j). The amino group is more nu-
cleophilic than the hydroxyl group and a 6-endo-trig ring closure is
more favourable than the 5-endo-trig route therefore this result
was predictable. Again, the aldehyde 2j gave an anomalous result
and the 1H NMR spectrum of this crude reaction product showed
a largely uncharacterisable mixture of compounds, including the
bisimine 8j and the diastereomeric pair of 5-hydroxyhexahy-
dropyrimidines, cis- and trans-9j. The 1H NMR spectra of the crude
reaction products in CDCl3 also contained a number of minor sig-
nals unrelated to the starting materials (6, 2), bisimines (8) or 5-
hydroxyhexahydropyrimidines (9). These extraneous signals,
which consistently included two prominent singlets between 5 and
6 ppm, occurred at varying intensities in crude reaction mixtures
for 8a–8g but were absent from those derived from the hydroxy
benzaldehydes 2h and 2i.
Table 1
Compositiona of the crude products derived from the equimolar reactions of the
diamines 1 and 6 with the aromatic aldehydes 2a–j
Aldehyde
Reaction with 1
Reaction with 6b
%9 %8
80
%4
%3
30
2a p-Nitrobenzaldehyde
2b p-Cyanobenzaldehyde
2c p-Bromobenzaldehyde
2d p-Chlorobenzaldehyde
2e Benzaldehyde
2f p-Methylbenzaldehyde
2g p-Methoxybenzaldehyde
2h p-Hydroxybenzaldehyde
2i o-Hydroxybenzaldehyde
2j 2-Pyridine carboxaldehyde
70c
81c
68c
0
7
19
32
75
69
57
70
60
38
46
0
8
10
21
10
20
32
8d
65
0
100
48
52
11
0
48
100
43
100
18
57
0
82c
100
a
Determined from the relative integral values of singlets from the protons
attached to C2 of the hexahydropyrimidines and the benzylidene protons of the
imines in the 1H NMR spectrum acquired in CDCl3 (300 MHz, 298 K).
b
The percentage of monoimine present in these mixtures makes sum of the
stated values up to 100%.
c
The 1H NMR spectrum of these compounds in CDCl3 (300 MHz, 298 K) showed
a trace of monoimine present.
d
The 1H NMR spectrum of this mixture was obtained in DMSO-d6 (300 MHz,
Additional NMR investigations of N,N0-bis(p-nitrobenzylidene)-
1,3-diamino-2-propanol (8a) and N,N0-bis(p-chlorobenzylidene)-1,3-
298 K) and showed that the crude reaction mixture also contained unreacted
aldehyde.

J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
10687
diamino-2-propanol (8d) revealed these that extra signals are not
present in the 1H NMR spectrum obtained in deuterated DMSO
(400 MHz, 303 K); the only signals observed in this solvent were
those expected for the bisimine (see Supplementary data). This result
suggests the extra signals observed in CDCl3 (400 MHz, 303 K) orig-
diaminopropane 1 with p-chloro-, p-methoxy- and o-hydroxy
benzaldehyde (2d, 2g and 2i) gave the bisimine as the sole product,
whilst the reaction with p-nitrobenzaldehyde (2a) and p-cyano-
benzaldehyde (2b) formed the hexahydropyrimidine in yields of 70
and 81%, respectively. The same trend was observed in the re-
actions with the diaminoalcohol 6, although, often substantially
more heterocycle was produced in reactions with 6. Both cis and
trans hydroxyhexahydropyrimidine 9 were present in the crude
product mixtures isolated from the reaction of 6 with 2a–h and 2j.
In all cases, the cis isomer was the major hydroxyhexahydropyr-
imidine observed and constituted 60–80% of the hexahydropyr-
imidine yield, however no correlation was observed between cis/
trans ratios and the electronic nature of the aryl substituent. The 1H
NMR spectra of the crude products from the reactions of the
diaminoalcohol 6 with all benzaldehydes except 2j also showed
evidence of the monoimine, particularly when an electron donating
group was present in the benzaldehyde. An anomalous result
obtained in the reaction of p-hydroxybenzaldehyde 2h with 6 is
possibly a result of a competing acid–base reaction with the phenol.
The 1H NMR spectrum of this crude reaction mixture showed that
a large amount of aldehyde was still present. As expected from
results obtained in the corresponding reactions of 6 with 2 mol of
aldehyde, none of the 1H NMR spectra of the crude reaction mix-
tures from the equimolar reactions contained signals attributable
to the aminomethyloxazolidines (10a–j).
inate from products formed as the b-hydroxybisimine 8 undergoes
ring-chain tautomerism. Formation of the oxazolidine is a dis-
favoured 5-endo-trig process according to Baldwin’s rules,5 however
breakdown of these rules in the b-hydroxyimine-oxazolidine system
has been reported.22 Intramolecular attack by the hydroxyl group
at either imine moiety would result in formation of a mixture the
cis- and trans-5-[N-(p-chlorobenzylidene)aminomethyl]-2-(p-chlor-
ophenyl)oxazolidine (Scheme 3, 11d). Inhibition of the tautomeric
interconversion in deuterated DMSO, as seen for compounds 8a and
8d, occurs through hydrogen bonding between the solvent and the
hydroxyl group, anobservationpreviously noted in related systems.23
The downfield shift of the phenolic proton (13.05 ppm), in CDCl3 in-
dicates that the imine groups of the o-hydroxy bisimine 8i are en-
gaged in resonance-assisted intramolecular hydrogen bonds, typical
of these compounds.24 Competition from this intramolecular hy-
drogen bonding most likely hampers attack at the imine thereby
preventing ring-chain tautomerism in this molecule.
2.2. Heats of formation
Whilst the results described above displayed obvious trends, it
was apparent from the various anomalies that there were
a number of factors influencing product distribution. Molecular
modelling techniques permitted further examination of these
factors. The thermodynamic stabilities of the observed products
were explored using the calculated heats of formation (DHf) of the
possible products at the semi-empirical (AM1) level (Tables 2 and
3). Table 2 presents calculated results for products formed in the
reaction of 1,3-diaminopropane (1) with the benzaldehydes 2a–j.
Some general trends are apparent in these results. In most cases,
the bisimine (3) is higher in energy than the corresponding
monoimine (5) or hexahydropyrimidine (4), whereas 4 and 5 are
Scheme 3. Possible ring-chain tautomeric equilibrium of compound 8d.
of similar energies. The
extent upon the molecular weight, which possibly contributes to
the higher Hf of the bisimines. However, variations in molecular
weights do not explain why the differences in the heat of for-
mation between the bisimines and the monoimines (DDHf 3–5)
show such a wide variation.
DHf of a compound depends to some
1H NMR spectra of 8d obtained immediately following disso-
lution of the sample in CDCl3 showed that the formation of the
oxazolidines 11d was quite rapid (<5 min). No measurable change
was observed in the ratio of 8d, cis-11d and trans-11d after 24 h,
although a set of minor signals unrelated to these products be-
came apparent in the 1H NMR spectrum acquired in CDCl3, these
were also not observed in the spectrum obtained in DMSO-d6. A
number of key features and relationships in the 1H and 13C NMR
spectra of 8d provide evidence for the presence of the dia
stereomeric pair of oxazolidines cis-11d and trans-11d as sum-
marised in Table S1 (see Supplementary data). Significantly, the
singlets observed at 5.38 and 5.55 ppm arise from protons bonded
to carbons with shifts of 93.0 and 92.1 ppm, respectively, typical
shifts observed for carbons in an O–C–N fragment.25 Un-
fortunately, the presence of numerous overlapping signals, par-
ticularly in the aromatic region of the spectra, precluded definitive
identification of the minor products.
Table 1 shows the compositions of the crude products from the
equimolar reactions of selected benzaldehydes (2a–j) with 1,3-
diaminopropan-2-ol (6) and 1,3-diaminopropane (1). Without ex-
ception, the experimental results showed that the reaction of
benzaldehydes with electron donating substituents favoured
bisimine products and benzaldehydes with electron withdrawing
substituents favoured the formation of the hexahydropyrimidines.
For example, the crude mixtures from the equimolar reactions of
D
Table 2
The calculated heats of formation (
and hexahydropyrimidines 4a–j and the calculated differences in heats of formation
DDHf)b between the intermediate 5a–j and the potential products 3a–j and 4a–j
D
Hf)a for the monoimines 5a–j, bisimines 3a–j
(
Derivative
D
Hf (kJ molꢀ1
)
DDHfb (kJ molꢀ1
4–5 3–5
)
5c
4
3c
a
b
c
d
e
f
g
h
i
154
274
158
107
142
105
ꢀ23
ꢀ45
ꢀ44
194
155
250
162
89
118
110
ꢀ17
ꢀ67
ꢀ69
165
384
613
392
292
342
287
32
ꢀ27
ꢀ25
444
1
ꢀ25
4
ꢀ19
ꢀ23
6
230
338
234
185
200
182
55
19
19
250
6
ꢀ22
ꢀ25
ꢀ29
j
a
Modelled at the semi-empirical level (AM1).
b
c
DDHf¼
D
Hf (product)ꢀ
DHf (monoimine).
All imine bonds were modelled in the anti configuration.

10688
J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
Table 3
The calculated heats of formation (
D
Hf)a for the monoimines 7a–j, bisimines 8a–j, cis and trans-5-hydroxyhexahydropyrimidines 9a–j, the cis- and trans-5-amino-
methyloxazolidines 10a–j and the calculated differences in heats of formation (DDHf)b between the intermediate 7a–j and the potential products 8a–j, cis-9a–j and trans-10a–j
Derivative
D
Hf (kJ molꢀ1
7c,d
ꢀ43
)
DDHfb (kJ molꢀ1
)
8c
192
423
199
98
157
cis-9
trans-9
cis-10
trans-10
8–7
cis-9–7
trans-10–7
a
b
c
d
e
f
g
h
i
ꢀ41
57
ꢀ32
62
ꢀ31
85
ꢀ36
81
235
340
239
174
210
186
58
2
ꢀ26
4
ꢀ24
ꢀ22
6
7
ꢀ2
10
ꢀ5
ꢀ7
11
11
2
83
ꢀ40
ꢀ76
ꢀ53
ꢀ94
ꢀ221
ꢀ239
ꢀ238
2
ꢀ35
ꢀ27
ꢀ26
ꢀ77
ꢀ46
ꢀ79
ꢀ206
ꢀ233
ꢀ233
ꢀ4
ꢀ30
ꢀ81
ꢀ60
ꢀ83
ꢀ210
ꢀ236
ꢀ234
ꢀ7
ꢀ100
ꢀ75
ꢀ105
ꢀ70
93
ꢀ88
ꢀ79
ꢀ163
ꢀ216
ꢀ197
257
ꢀ215
ꢀ261
ꢀ260
ꢀ26
ꢀ206
ꢀ256
ꢀ256
ꢀ23
6
23
41
255
ꢀ23
ꢀ23
ꢀ29
3
ꢀ9
j
a
Modelled at the semi-empirical level (AM1).
b
c
DDHf¼
D
Hf (product)ꢀ
DHf (monoimine).
All imine bonds were modelled in the anti configuration.
Monoimines were modelled as the R stereoisomer.
d
Table 3 shows the differences in
D
Hf values (DDHf) for the
constant corrected for resonance effects, showed this same trend
but with improved correlation (R2¼0.86). However, the same plot
for product 4 from 1,3-diaminopropane (1) showed no quantitative
correlation.
monoimines (7) and the corresponding hydroxyhexahydropyr-
imidines (9), aminomethyloxazolidines (10) and bisimines (8) for
the reactions with the diaminoalcohol 6. The most obvious trend is
the increase in the difference in the
D
Hf of the bisimines (DDHf 8–7)
Whether the imine or iminium ion intermediate undergoes an
intramolecular cyclisation reaction to form the hexahydropyr-
imidine, or reacts with another molecule of benzaldehyde to give
the bisimine, may be related to the comparative reactivities of the
electrophilic carbon atoms in the intermediate and the benzalde-
hyde. To investigate this further, the energies of the frontier mo-
with an increase in the electron withdrawing nature of the alde-
hyde ring substituent. The hexahydropyrimidines cis-9a–j and
trans-9a–j do not show this trend; conversely, they are of compa-
rable energy to the monoimine 7. In the reactions of p-nitro-
benzaldehyde (2a) with 6, for example, the difference in energy
between the hexahydropyrimidine cis-9a and the monoimine 7a is
lecular orbitals (FMOs) of benzaldehydes with
substituents (2a–j) were calculated at ab initio level using the 6-
311G basis set (Table 4). The energy of the LUMO (ELUMO) as well as
a range of
2 kJ molꢀ1, but the difference in
bisimine 8a and the monoimine 7a is 235 kJ molꢀ1. By comparison,
the differences found in the reaction of with o-hydrox-
ybenzaldehyde (2i) are significantly smaller; the difference in Hf
DHf between the associated
*
6
the normalised ELUMO (ELUMOꢁLUMO coefficient at the carbonyl
carbon), and the partial charge on the carbonyl carbon all decrease
in a linear relationship with increasing electron withdrawing na-
ture of the substituent as measured using sp (R2¼0.97, 0.98 and
0.98, respectively). Comparison of these values to sþp gave poorer
correlations (R2¼0.88, 0.92 and 0.93, respectively).
D
of the hexahydropyrimidine cis-9i and the monoimine 7i is
ꢀ23 kJ molꢀ1 and that of the bisimine 8i is 41 kJ molꢀ1
.
These results indicate that the DHf of the product can influence
the product distribution seen in the crude mixtures. Qualitatively,
we observe that the higher the difference in the heat of formation
A normalised ELUMO was calculated to account for the difference
in ELUMO for the individual molecules as it is not strictly correct to
directly compare coefficients of atoms in different molecules.28,29
This measure is related to the density of the ELUMO on the carbonyl
carbon. The same general trend, but no linear correlation, was
observed between the calculated ELUMO and normalised ELUMO
values and the percentage of heterocycle (4a–j and 9a–j) formed.
The reactions of benzaldehydes with lower ELUMO values gave more
2-arylhexahydropyrimidine product, in accord with the fact that
decreased ELUMO and increased substituent electron withdrawing
effect are associated. Likewise, the same general trend, but no linear
correlation, was observed when relating the partial charge on the
carbonyl carbon to the percentage heterocycle formed in the re-
actions with 1,3-diaminopropane. Interestingly, a linear relation-
ship was observed (R2¼0.92) between the partial charge on the
carbonyl carbon of the aldehyde (2a–j), including 2-pyridine car-
boxaldehyde (2j) and o-hydroxybenzaldehyde (2i), and the yield of
the heterocycle (9a–j) in the reactions with 1,3-diamino-2-prop-
anol (6). The same plot for the percentage of hexahydropyrimidine
4 formed in the reaction with 1,3-diaminopropane (1) showed no
quantitative correlation.
(
DDHf) between the monoimine/hexahydropyrimidine and the
bisimine, the less favourable it is for the bisimine to form. Con-
versely, the closer in energy the heats of formation are, the more
likely it is that the bisimine will form. However, results from ex-
perimental product distributions (Table 1) and the relative
DHf of
the products (Tables 2 and 3) shows that the formation of the
hexahydropyrimidine product depends mostly upon the nature of
the substituent attached to the benzaldehyde ring.
2.3. The nature of the aldehyde
The presence of electron donating groups, such as the hydroxy
group, on the aryl ring appears to facilitate formation of the imine
(both bis and mono). When the benzaldehyde has an electron
withdrawing substituent, the formation of the hexahydropyr-
imidine is favoured, as observed for the reactions with the p-nitro-
and p-cyanobenzaldehyde (2a and 2b). To gain further insight into
this phenomenon, we next compared the percentage of hexahy-
dropyrimidine (4a–h and 9a–h) formed in reactions of the para
substituted benzaldehydes (2a–h) with a more quantitative mea-
sure of the electronic effect of the benzaldehyde substituents,
Hammett substituent constants26,27 (see Supplementary data). For
each diamine, plots of the percentage of heterocycle formed against
Hammett constant (sp) showed the same clear trends but no
quantitative correlation; increasing the electron withdrawing na-
ture of the aryl ring substituent increased the amount of
The reactions of (2j) and o-hydroxybenzaldehyde (2i) gave ex-
treme results. The reaction of o-hydroxybenzaldehyde with the
diaminoalcohol 6 gave only bis- or monoimine (8i, 7i) in the
equimolar reaction, whilst the 2-pyridine carboxaldehyde yielded
only the hexahydropyrimidine (9j). Additional effects due to the
ortho substituent/endocyclic heteroatom are likely to influence the
product outcomes in the reactions of these aldehydes. For example,
imines derived from o-hydroxybenzaldehyde are known to gain
hexahydropyrimidine formed.
A plot of the percentage of
5-hydroxyhexahydropyrimidine (9) formed against sþp , a Hammett

J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
10689
Table 4
Computed total energies,a partial charges on the aldehyde carbonyl carbon,b LUMO energies (ELUMO), LUMO coefficients,c calculated values for the normalised ELUMO (ELUMO
norm)d and HOMO–LUMO band gapse for the aldehydes 2a–j after ab initio geometry optimization
Aldehyde
E (104 kJ molꢀ1
)
Charge on the
carbonyl carbon
ELUMO
LUMO
coefficient
ELUMO norm
HOMO–LUMO band
(kJ molꢀ1
)
(kJ molꢀ1
)
gap (103 kJ molꢀ1
)
2a
2b
2c
2d
2e
2f
2g
2h
2if
2j
ꢀ144
ꢀ114
ꢀ765
ꢀ211
ꢀ90
0.542
0.543
0.547
0.548
0.549
0.550
0.552
0.552
0.564
0.533
62
100
169
173
206
217
238
235
192
185
0.353
0.431
0.527
0.532
0.555
0.543
0.607
0.607
0.595
0.491
22
43
89
1.07
1.08
1.09
1.10
1.13
1.11
1.09
1.10
1.05
1.14
92
114
118
144
143
114
92
ꢀ100
ꢀ120
ꢀ110
ꢀ110
ꢀ94
a
Calculated at the HF/6-311G
*
level.
Natural atomic charge of the carbonyl carbon.
LUMO coefficient at the carbonyl carbon; the absolute value of the sum of all coefficients; dominated by pz.
LUMO coefficient multiplied by ELUMO
b
c
d
e
f
.
Energy difference between the HOMO and the LUMO.
Values are for the conformation with an internal hydrogen bond.
extra stabilisation from a strong resonance-assisted hydrogen
bond.24,30 The extra stability gained by the intramolecular hydro-
gen bond may be a factor in why no hexahydropyrimidine forms in
the reaction with o-hydroxybenzaldehyde when hexahydropyr-
imidine is observed in the reaction with p-hydroxybenzaldehyde.
Two different conformations of o-hydroxybenzaldehyde were
modelled here. In one conformation, the substituents were
arranged in such a way that an internal hydrogen bond could be
formed, in the other conformation the carbonyl oxygen and the
hydroxyl hydrogen were turned away from one another. The former
conformation had the lower total energy, lower ELUMO and band
gap, as well as larger LUMO coefficient, normalised ELUMO, and
a more positive charge on the carbonyl carbon than the latter
conformation. For all these reasons, the conformation with the
intramolecular hydrogen bond was used here as the representative
conformation.
Lower values for LUMO coefficients and normalised ELUMO on
benzaldehydes with electron withdrawing substituents are
counter-intuitive and result from delocalisation of the LUMO. The
pictorial representations of the LUMOs for p-nitrobenzaldehyde
(2j) and p-hydroxybenzaldehyde (2h; see Supplementary data)
illustrate this. The density of the LUMO on the carbonyl group of
p-hydroxybenzaldehyde is much greater than that on the carbonyl
of the p-nitrobenzaldehyde. The p-nitrobenzaldehyde molecule
also has a more evenly distributed LUMO. A consequence of this
delocalisation is enhanced stability and hence a lower ELUMO for the
LUMO of benzaldehydes with electron withdrawing substituents.
Similar results were obtained when calculations were performed at
were built to investigate their FMOs. These structures were sub-
jected to a conformation search and the geometry of three con-
formers (chosen to be as different as possible) was optimised at ab
initio (HF/6-31G ) level. In this case, the computed values gave no
*
insight into the reactivity of the monoimines, as the energy dif-
ferences calculated for the different conformations were some-
times large enough to be greater than the energy differences
between the individual molecules. Also, the calculated values for
the ELUMO, EHOMO and the LUMO coefficients32 depended upon the
conformation adopted by the aliphatic chain of the monoimines/
monoiminium ions. In contrast, the benzaldehydes, which are
rather rigid molecules, did not present this problem.
The effect of the amine structure on the imine/heterocycle ratio
was not investigated. The propanediamines are highly flexible and
conformationally complex,33 resulting in difficulties when model-
ling the FMOs. The low energy barriers between the conformations
mean that the LUMO coefficients would be unreliable. Hydrogen
bonding further complicates the matter as both the terminal amino
and the hydroxyl group have the potential to hydrogen bond.
3. Conclusion
The investigations discussed above, both experimental and
computational, demonstrate an identical trend and indicate that
two factors influence the product ratios formed in the reaction of
the diaminopropanes 1 and 6 with benzaldehydes, namely the
electronic nature of the aryl ring substituent and the substituent
group on the carbon 2 of the amine nucleophile. Experimental
evidence shows that hexahydropyrimidine formation is favoured
by the less nucleophilic amine and by the presence of electron
withdrawing groups on the aryl ring of the aldehyde. Calculations
show that the electronic nature of this aryl ring substituent in-
fluences both the relative thermodynamic stability of the final
products and the reactivity of the aldehyde as an electrophile.
a lower level of theory (HF/6-31G , data not shown).
*
The energy difference between the LUMO and HOMO (HOMO–
LUMO band gap) was calculated because it is related to the stability
and therefore the reactivity of a compound. However, the HOMO–
LUMO band gaps showed little variation in this work and no ap-
parent correlation with Hammett constants or with the relative
amounts of heterocycle formed.
The differences between calculated DHf values for the bisimine
Altered reactivity of the benzaldehyde would change the ki-
netics for the formation of the intermediate imine/iminium ion.
Benzaldehydes with electron withdrawing substituents are more
amenable to attack by a nucleophile because they have a more
stable LUMO. This should result in an increased rate of imine for-
mation, as is seen in other imine forming reactions. For example, it
has been shown that the presence of an electron donating group in
the para position of benzaldehydes decreases the rate of imine
formation in the reaction with anilines to form the Schiff base.31 A
number of proposed intermediate monoimine/monoiminium ions
and the hexahydropyrimidine are related to the nature of the
benzaldehyde substituent and can explain experimental observa-
tions. The thermodynamic stability of the hexahydropyrimidines
formed in reactions with electron donating substituents is similar
to that of the analogous bisimines. However, the difference in heat
of formation (DDHf) between the bisimine and the hexahydropyr-
imidine increases with increasing electron withdrawing nature of
the substituent. The relative stability of the bisimine decreases with
increasing electron withdrawing ability of the substituent making
the formation of the hexahydropyrimidine more favourable with

10690
J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
increasing electron withdrawing nature of the substituent. This
trend is mirrored by experimental observations.
CDCl3. Product ratios for the crude reaction mixtures were de-
termined from the 1H NMR integrals of the imine CH protons for the
mono- and bisimines, the methine proton (H2) for the hexahy-
dropyrimidine and when necessary, the CHO proton of the alde-
hyde (see Supplementary data).
Increased hexahydropyrimidine formation with an increase in
electron withdrawing nature of the benzaldehyde substituent can
also be attributed to perturbation of the FMOs of the benzaldehyde.
Electron withdrawing groups in the para position of the benzal-
dehyde lead to stabilisation of the LUMO; the more stable the
LUMO, the higher the percentage of hexahydropyrimidine. The
increased stability of the LUMO is due to it becoming more delo-
calised and hence lower in energy. These results do not demon-
strate why a more stable LUMO in the aldehyde encourages the
formation of a hexahydropyrimidine rather than a bisimine. It is
possible that the stabilising effect of an electron withdrawing
substituent on the LUMO is more pronounced for the monoimines,
making them more susceptible to intramolecular nucleophilic at-
tack, however we have no evidence to support this because of the
conformational flexibility of the monoimines.
4.3. Characterisation data for the isolated
hexahydropyrimidines
4.3.1. 5-Hydroxy-2-(40-nitrophenyl)hexahydropyrimidine (9a). The
crude product was crystallised from methanol to yield a pale tan
solid (0.19 g, 77%). The 1H NMR spectrum of 9a in CDCl3 showed
that the reaction product was a 3:1 mixture of cis (major) and trans
(minor) isomers. 1H NMR (CDCl3, 300 MHz)
d 2.70 (br s, 3H, OH and
NH), 2.78 (dd, J¼11, 9 Hz, 2H, trans axial H4 and H6), 3.16 (dd, J¼11,
2 Hz, 2H, cis equatorial H4 and H6), 3.20 (dd, J¼11, 2 Hz, 2H, cis axial
H4 and H6), 3.36 (dd, J¼11, 4 Hz, 2H, trans equatorial H4 and H6),
3.65 (m, 1H, trans H5), 3.68 (m, 1H, cis H5), 4.63 (s, 1H, trans H2),
4.69 (s, 1H, cis H2), 7.72 (overlapping m, 2Hþ2H, cis and trans ArH),
8.22 (overlapping m, 2Hþ2H, cis and trans ArH). 13C NMR (DMSO-
4. Experimental
4.1. General
d6, 101 MHz)
d 39.5, 39.7, 50.5, 52.7, 62.2, 70.4, 71.6, 122.9, 123.0,
128.0, 128.1, 146.6, 146.7, 150.1, 150.2. LREI-MS m/z 222 ([MꢀH]þ);
Benzaldehyde, 1,3-diaminopropane and 2-pyridine carbox-
aldehyde were purified and/or distilled before use.34 All other al-
dehydes were used as received from freshly opened containers.
CDCl3 was pre-dried over NaSO4 and stored over 4 Å molecular
sieves. DMSO-d6 was dried by vacuum distillation from CaH2. NMR
spectra were acquired at 300 (1H NMR) and 75 MHz (13C NMR) on
a Varian Mercury 300 spectrometer or at 500 (1H NMR) and
126 MHz (13C NMR) on a Varian Inova 500 spectrometer at a probe
temperature of 298 K. NMR spectra obtained at 400 MHz (1H NMR)
were acquired on a Bruker Avance III spectrometer at a probe
temperature of 303 K. NMR spectra are referenced to TMS or the
residual solvent peak. Hydrogen and carbon assignments were
made using standard NOE, APT, DEPT, gCOSY, gHSQC and gHMBC
spectroscopic techniques. Low resolution EI mass spectra (LREI-MS)
between 100 and 700 amu were determined on a Shimadzu
QP5050 spectrometer. High resolution EI mass spectra (HREI-MS)
were determined on VG Autospec spectrometer operating at 70 eV
with a source temperature of 250 ꢂC and were referenced with
perfluorokerosene. Melting points were determined on a Gallen-
kamp capillary or Reichert hot-stage melting point apparatus and
are uncorrected. Infrared spectra were measured on a Nicolet Av-
atar 360 FTIR spectrometer; samples were deposited on a Ge crystal
for IR spectrum acquisition. Characterisation data is reported only
for compounds that gave well-resolved, unambiguous signals in
their 1H NMR spectrum (compounds 3a–i, 8a, 8c–g, 8i, 9a and 9j).
Geminal protons that give well-separated signals are denoted as HA
and HB; the proton with the most downfield signal is designated as
HA. The 1H NMR spectra of compounds 8a, 8c–g and 8i exhibited
a small allylic coupling (w1 Hz) between the azomethine proton
(N]CH) and the protons on C1/C3, which resulted in the N]CH
signal appearing as a broad singlet/unresolved triplet (br s/t).
HREI-MS m/z calcd for [MꢀH]þ C10H12N3O3: 222.0879, found:
222.0884. IR (n
, cmꢀ1) 3242 (N–H), 3073 (OH), 1510 and 1345
(N]O).
4.3.2. 5-Hydroxy-2-(20-pyridinyl)hexahydropyrimidine (9j). The crude
product was a colourless oil that solidified on standing in a desiccator
1
over P2O5. The H NMR spectrum showed that the reaction product
was essentially a 3:2 mixture of cis (major) and trans (minor) isomers
(0.19 g, quant.). 1H NMR (CDCl3, 500 MHz)
d 2.20 (br s, 3H, OH and
NH), 2.76 (dd, J¼10.3, 9.9 Hz, 2H, trans axial H4 and H6), 3.17 (dd,
J¼14, 1 Hz, 2H, cis H4 and H6), 3.22 (dd, J¼14, 2 Hz, 2H, cis H4 and
H6), 3.37 (dd, J¼10.3, 4.5 Hz, 2H, trans equatorial H4 and H6), 3.62
(overlapping m, 1Hþ1H, cis and trans H5), 4.55 (s, 1H, trans H2), 4.67
(s, 1H, cis H2), 7.20–7.24 (overlapping m, 1Hþ1H, cis and trans H40),
7.42 (d, J¼8 Hz, 1H, trans H60), 7.45 (d, J¼8 Hz, 1H, cis H60), 7.66–7.71
(overlapping m, 1Hþ1H, cis and trans H50), 8.54 (d, J¼9 Hz, 1H, cis
H30), 8.55 (d, J¼9 Hz, 1H, trans H30). 13C NMR (CDCl3 126 MHz) cis
isomer d
51.6 (C4 and C6), 63.0 (C5), 73.9 (C2),121.4 (C60),123.33 (C40),
137.2 (C50), 149.3 (C30), 159.5 (C10), trans isomer
d 52.5 (C4 and C6),
65.8 (C5), 73.0 (C2), 121.8 (C60), 123.28 (C40), 137.1 (C50), 149.4 (C30)þ,
159.2 (C10). LREI-MS m/z 179 ([M]þ); HREI-MS m/z calcd for [MꢀH]
C9H12N3O: 178.0975, found: 178.0980. IR (
NH).
n
, cmꢀ1) 3278 (br, OH and
4.4. Characterisation data for the isolated bisimines
4.4.1. N,N0-Bis(p-nitrobenzylidene)-1,3-diaminopropane (3a). The cru-
de product was crystallised from acetonitrile to afford pale yellow
crystals (0.43 g, 93%), mp 202–204 ꢂC (lit.35 197–201 ꢂC). 1H NMR
(CDCl3, 300 MHz)
H3), 7.88 (d, J¼9 Hz, 4H, ArH), 8.25 (d, J¼9 Hz, 4H, ArH), 8.39 (s, 2H,
N]CH). 13C NMR (CDCl3, 75 MHz)
31.9, 59.3, 123.8, 128.6, 141.5,
148.9, 158.9. LREI-MS m/z 340 ([M]þ); HREI-MS m/z calcd for [M]þ
d
2.16 (m, 2H, H2), 3.80 (dt, J¼7, 1 Hz, 4H, H1 and
4.2. Competitive reactions
d
One molar equivalent of the aldehyde (2a–j) in methanol
(15 mL) was added dropwise to a stirred solution of 1 (0.1 g,
1.35 mmol) or 6 (0.1 g, 1.10 mmol) in methanol (15 mL). The mix-
ture was heated at reflux for 2 h and the solvent removed under
vacuum. Residues were lyophilised to remove residual water from
the crude product mixtures before analysis. Synthesis of the imines
was accomplished under the same conditions using 2 M equiv of
the aldehyde. 1H NMR spectra of the crude mixtures were acquired
in CDCl3 with the exception of the crude products derived from 2h,
which were obtained in DMSO-d6 because of their low solubility in
C17H16N4O4: 340.1172, found: 340.1166. IR (n
, cmꢀ1) 1644 (C]N), 1511
and 1342 (N]O).
4.4.2. N,N0-Bis(p-cyanobenzylidene)-1,3-diaminopropane (3b). The
crude product was crystallised from petroleum spirit to give col-
ourless crystals (0.38 g, 95%), mp 163–164 ꢂC. 1H NMR (CDCl3,
300 MHz)
d
2.12 (m, 2H, H2), 3.76 (dt, J¼7, 1 Hz, 4H, H1 and H3),
7.68 (d, J¼8 Hz, 4H, ArH), 7.81 (d, J¼8 Hz, 4H, ArH), 8.32 (br s, 2H,
N¼CH). 13C NMR (CDCl3, 75 MHz)
d 31.8, 59.1, 113.7, 118.3, 128.3,
132.2, 139.8, 159.2. LREI-MS m/z 300 ([M]þ); HREI-MS m/z calcd for

J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
10691
[M]þ C19H16N4: 300.1374, found: 300.1372. IR (
2223 (C^N).
n
, cmꢀ1) 1642 (C]N),
132.1, 160.9, 165.2. LREI-MS m/z 282 ([M]þ). IR (
(C]N).
n
, cmꢀ1) 1632
4.4.3. N,N0-Bis(p-bromobenzylidene)-1,3-diaminopropane (3c). The
crude product was crystallised from petroleum spirit to give
colourless crystals (0.52 g, 95%), mp 73.7–74.5 ꢂC (lit.11 45 ꢂC). 1H
4.4.10. N,N0-Bis(p-nitrobenzylidene)-1,3-diamino-2-propanol(8a). The
crude product was crystallised from ethyl acetate to afford 8a as pale
tan crystals (0.38 g, 98%), mp 137–138 ꢂC. 1H NMR (CDCl3, 300 MHz)
NMR (CDCl3, 300 MHz)
d
2.01 (m, 2H, H2), 3.70 (dt, J¼7, 1 Hz,
d
2.67 (d, J¼4 Hz, 1H, OH), 3.84 (ddd, J¼11, 7, 1 Hz, 2H, H1B and H3B),
4H, H1 and H3), 7.52 (d, J¼9 Hz, 4H, ArH), 7.57 (d, J¼9 Hz, 4H,
3.94 (ddd, J¼11, 5, 1 Hz, 2H, H1A and H3A), 4.29 (m, 1H, H2), 7.90 (d,
ArH), 8.22 (br s, 2H, N¼CH). 13C NMR (CDCl3, 126 MHz)
d 31.8,
J¼9 Hz, 4H, ArH), 8.25 (d, J¼9 Hz, 4H, ArH), 8.44 (br s/t, 2H, N]CH).
59.1, 124.9, 129.4, 131.8, 135.1, 160.0. LREI-MS m/z 406/408/410
1H NMR (DMSO-d6, 400 MHz)
H
d
3.69 (ddd, J¼12, 6, 1 Hz, 2H, H1B and
([M]þ/[Mþ2]þ/[Mþ4]þ); HREI-MS m/z calcd for [M]þ
3B), 3.88 (ddd, J¼12, 5,1 Hz, 2H, H1A and H3A), 4.12 (m, 1H, H2), 4.95
C17H16N279Br2: 405.9680; found: 405.9697. IR (
n
, cmꢀ1) 1644
(d, J¼6 Hz, 1H, OH), 8.03 (d, J¼9 Hz, 4H, ArH), 8.30 (d, J¼9 Hz, 4H,
(C]N).
ArH), 8.51 (s, 2H, N]CH). 13C NMR (CDCl3, 75 MHz)
d 64.9, 70.8,123.8,
128.8, 141.1, 149.1, 160.7. LREI-MS m/z 356 ([M]þ). HREI-MS m/z calcd
4.4.4. N,N0-Bis(p-chlorobenzylidene)-1,3-diaminopropane (3d). The
crude product was crystallised from petroleum spirit to give leafy
white crystals (0.40 g, 95%), mp 65–66 ꢂC (lit.36,37 66.0–67.0 ꢂC). 1H
for [M]þ C17H16N4O5: 356.1121, found: 356.1120. IR ( , cmꢀ1) 1649
n
(C]N), 1525 and 1320 (N]O).
NMR (CDCl3, 300 MHz)
d
2.09 (m, 2H, H2), 3.70 (dt, J¼7,1 Hz, 4H, H1
4.4.11. N,N0-Bis(p-bromobenzylidene)-1,3-diamino-2-propanol
and H3), 7.35 (d, J¼8 Hz, 4H, ArH), 7.63 (d, J¼8 Hz, 4H, ArH), 8.22 (br
(8c). The crude product was a colourless solid (0.46 g, quant.), mp
s, 2H, N]CH). 13C NMR (CDCl3, 75 MHz)
d
31.9, 59.1, 128.7, 129.1,
137–139 ꢂC. 1H NMR (CDCl3, 300 MHz)
d 1.80 (br s, residual H2O and
134.6, 136.3, 159.7. LREI-MS m/z 318 ([M]þ). IR (
(C]N).
n
, cmꢀ1) 1644
OH), 3.73 (ddd, J¼11, 7, 1 Hz, 2H, H1B and H3B), 3.83 (ddd, J¼11, 5,
1 Hz, 2H, H1A and H3A), 4.21 (m, 1H, H2), 7.53 (d, J¼8 Hz, 4H, ArH),
7.59 (d, J¼8 Hz, 4H, ArH), 8.28 (br s/t, 2H, N]CH). 13C NMR (CDCl3,
4.4.5. N,N0-Bisbenzylidene-1,3-diaminopropane (3e)38,1. The crude
75 MHz) d 64.8, 70.9, 125.2, 129.5, 131.8, 134.7, 161.7. LREI-MS m/z
product was isolated as colourless oil that did not require further
421/423/425 ([MꢀH]þ/[(Mþ2)ꢀH]þ/[(Mþ4)ꢀH]þ); HREI-MS m/z
purification (0.33 g, quant.). 1H NMR (CDCl3, 300 MHz)
d
2.12 (m,
calcd for [MꢀH]þ C17H15N2O79Br2: 420.9551, found: 420.9553. IR (
cmꢀ1) 1648 (C]N), 3268 (OH).
n,
2H, H2), 3.72 (dt, J¼7, 1 Hz, 4H, H1 and H3), 7.37–7.42 (m, 6H, ArH),
7.68–7.74 (m, 4H, ArH), 8.28 (br s, 2H, N]CH). 13C NMR (CDCl3,
75 MHz)
d
32.0, 59.2, 127.9, 128.5, 130.4, 136.1, 161.2. LREI-MS m/z
, cmꢀ1) 1643 (C]N).
4.4.12. N,N0-Bis(p-chlorobenzylidene)-1,3-diamino-2-propanol
(8d). The crude product was crystallised from ethanol to give
a colourless solid (0.35 g, 95%), mp 132–135 ꢂC. 1H NMR (CDCl3,
250 ([M]þ). IR (
n
4.4.6. N,N0-Bis(p-methylbenzylidene)-1,3-diaminopropane (3f). The
crude product was a white solid that was characterised without
further purification (0.37 g, quant.) mp 66.5–68.0 ꢂC (lit.37 66 ꢂC).
300 MHz)
d
2.80 (d, J¼4 Hz, 1H, OH), 3.73 (ddd, J¼12, 7, 1 Hz, 2H,
H1B and H3B), 3.83 (ddd, J¼12, 5, 1 Hz, 2H, H1A and H3A), 4.22 (m,
1H, H2), 7.36 (d, J¼8 Hz, 4H, ArH), 7.66 (d, J¼8 Hz, 4H, ArH), 8.29 (br
1H NMR (CDCl3, 300 MHz)
d
2.03 (m, 2H, H2), 2.31 (s, 6H, CH3), 3.61
s/t, 2H, N]CH). 1H NMR (DMSO-d6, 400 MHz)
d
3.58 (dd, J¼12,
(dt, J¼7, 1 Hz, 4H, H1 and H3), 6.90 (d, J¼8 Hz, 4H, ArH), 7.64 (d,
6 Hz, 2H, H1B and H3B), 3.77 (dd, J¼12, 4 Hz, 2H, H1A and H3A), 4.04
(m, 1H, H2), 4.83 (d, J¼5 Hz, 1H, OH), 7.48 (d, J¼8 Hz, 4H, ArH), 7.76
(d, J¼8 Hz, 4H, ArH), 8.32 (s, 2H, N]CH). 13C NMR (CDCl3, 75 MHz)
J¼8 Hz, 4H, ArH), 8.16 (br s, 2H, N]CH). 13C NMR (CDCl3, 75 MHz)
d
21.6, 32.1, 59.2, 127.9, 129.2, 133.6, 140.6, 161.0. LREI-MS m/z 278
([M]þ). IR (
n
, cmꢀ1) 1646 (C]N).
d 64.8 (C1, C3), 70.9 (C2), 128.8, 129.3, 134.3, 136.7, 161.5 (C]N).
LREI-MS m/z 333/335/337 ([MꢀH]þ/[(Mþ2)ꢀH]þ/[(Mþ4)ꢀH]þ);
4.4.7. N,N0-Bis(p-methoxybenzylidene)-1,3-diaminopropane
(3g)1,11,37. Thecrude product was crystallised from petroleum spirit
to give colourless crystals (0.34 g, 90%), mp 81–82 ꢂC (lit.11 81 ꢂC). 1H
HREI-MS m/z calcd for [MꢀH]þ C17H15N2O35Cl2: 333.0561, found:
333.0556. IR (n
, cmꢀ1) 1648 (C]N), 3272 (OH).
NMR (CDCl3, 300 MHz)
d
2.08 (m, 2H, H2), 3.66 (dt, J¼7, 1 Hz, 4H, H1
4.4.13. N,N0-Bisbenzylidene-1,3-diamino-2-propanol (8e)39,40. The
crude product was a viscous oil that solidified on standing (0.29 g,
and H3), 3.82 (s, 6H, OCH3), 6.90 (d, J¼9 Hz, 4H, ArH), 7.64 (d,
J¼9 Hz, 4H, ArH), 8.20 (br s, 2H, N]CH). 13C NMR (CDCl3, 75 MHz)
quant.).1H NMR (CDCl3, 300 MHz)
d
2.10 (br s,1H, OH), 3.75 (ddd, J¼11,
d
32.2, 55.3, 59.2, 113.9, 129.2, 129.4, 160.4, 161.3. LREI-MS m/z 310
7, 1 Hz, 2H, H1B and H3B), 3.85 (ddd, J¼11, 5, 1 Hz, 2H, H1A and H3A),
([M]þ). IR (
n
, cmꢀ1) 1639 (C]N), 1249 (C–O).
4.24 (m, 1H, H2), 7.37–7.43 (m, 6H, ArH), 7.70–7.76 (m, 4H, ArH), 8.33
(br s/t, 2H, N]CH). 13C NMR (CDCl3, 75 MHz)
d
64.9, 70.8,128.0,128.7,
4.4.8. N,N0-Bis(p-hydroxybenzylidene)-1,3-diaminopropane
(3h). The crude product was a brown solid that was not purified
further (0.37 g, quant.) mp 65–68 ꢂC. 1H NMR (DMSO-d6, 500 MHz)
130.4, 136.2, 161.3. LREI-MS m/z 266 ([M]þ). IR ( , cmꢀ1) 1644 (C]N).
n
4.4.14. N,N0-Bis(p-methylbenzylidene)-1,3-diamino-2-propanol
d
1.87 (m, 2H, H2), 3.53 (t, J¼7 Hz, 4H, H1 and H3), 6.78 (d, J¼8 Hz,
4H, ArH), 7.53 (d, J¼8 Hz, 4H, ArH), 8.17 (s, 2H, N]CH), 9.65 (br s,
2H, OH). 13C NMR (DMSO-d6, 126 MHz)
32.2, 58.2, 115.4, 127.4,
(8f). The crude product was a colourless solid (0.32 g, quant.), mp
>250 ꢂC dec 1H NMR (CDCl3, 300 MHz)
d
2.38 (s, 6H, 2ꢁCH3), 2.85
d
(br s, 1H, OH), 3.72 (ddd, J¼11, 7, 1 Hz, 2H, H1B and H3B), 3.82 (ddd,
J¼11, 5, 1 Hz, 2H, H1A and H3A), 4.21 (m, 1H, H2), 7.19 (d, J¼8 Hz, 4H,
ArH), 7.61 (d, J¼8 Hz, 4H, ArH), 8.28 (br s/t, 2H, N]CH). 13C NMR
129.5 159.9, 160.2. LREI-MS m/z 282 ([M]þ). HREI-MS m/z calcd for
[M]þ C17H18N2O2: 282.1368; found: 282.1373. IR ( , cmꢀ1) 1644
n
(C]N).
(CDCl3, 75 MHz) d 21.6, 64.8, 71.1, 128.1, 129.2, 133.3, 141.0, 162.7.
LREI-MS m/z 294 ([M]þ), 293 ([MꢀH]þ); HREI-MS m/z calcd for
4.4.9. N,N0-Bis(o-hydroxybenzylidene)-1,3-diaminopropane (3i). The
[M]þ C19H22N2O: 294.1732; found: 294.1728. IR ( , cmꢀ1) 1646
n
crude product was a bright yellow solid (0.38 g, quant.), mp 59.8–
(C]N), 3242 (OH).
60.5 ꢂC (lit.20 59 ꢂC). 1H NMR (CDCl3, 300 MHz)
d 2.11 (m, 2H, H2),
3.71 (dt, J¼7, 1 Hz, 4H, H1 and H3), 6.86 (dt, J¼7, 2 Hz, 2H, ArH),
6.95 (dd, J¼8, 2 Hz, 2H, ArH), 7.22 (dd, J¼8, 2 Hz, 2H, ArH), 7.29
(ddd, J¼8, 7, 2 Hz, 2H, ArH), 8.35 (s, 2H, N]CH), 13.39 (br s, 2H,
4.4.15. N,N0-Bis(p-methoxybenzylidene)-1,3-diamino-2-propanol
(8g). The crude product was a colourless solid that was not purified
further (0.32 g, quant.), mp 117–119 ꢂC. 1H NMR (CDCl3, 300 MHz)
OH). 13C NMR (CDCl3, 75 MHz)
d
31.7, 56.8, 116.9, 118.5, 118.6, 131.1,
d
2.83 (d, J¼4 Hz, 1H, OH), 3.71 (ddd, J¼11, 7, 1 Hz, 2H, H1B and H3B),

10692
J.M. Locke et al. / Tetrahedron 65 (2009) 10685–10692
3.81 (ddd, J¼11, 5, 1 Hz, 2H, H1A and H3A), 4.20 (br m, 1H, H2), 3.84
(s, 6H, 2ꢁCH3), 6.93 (d, J¼9 Hz, 4H), 7.72 (d, J¼9 Hz, 4H), 8.26 (br s/t,
Supplementary data
2H, N]CH). 13C NMR (CDCl3, 75 MHz)
d
55.7, 65.1, 71.4, 114.2, 129.2,
Supplementary data associated with this article can be found in
online version at doi:10.1016/j.tet.2009.10.060.
129.9, 161.9, 162.4. LREI-MS m/z 327 ([M]þ); HREI-MS m/z calcd for
[M]þ C19H22N2O3: 326.1630; found: 326.1632. IR ( , cmꢀ1) 1649
n
(C]N), 3140 (OH).
References and notes
4.4.16. N,N0-Bis(o-hydroxybenzylidene)-1,3-diamino-2-propanol
(8i). The crude product was a bright yellow solid (0.32 g, quanti-
tative), mp 106–107 ꢂC (lit.41 105 ꢂC). 1H NMR (CDCl3, 300 MHz)
1. Simion, A.; Simion, C.; Kanda, T.; Nagashima, S.; Mitoma, Y.; Yamada, T.;
Mimura, K.; Tashiro, M. J. Chem. Soc., Perkin Trans. 1 2001, 2071–2078.
2. Atwood, D. A.; Harvey, M. J. Chem. Rev. 2001, 101, 37–52.
3. Calligaris, G. M.; Randaccio, L. In Comprehensive Coordination Chemistry; Wil-
kinson, G., Ed.; Pergamon: Oxford, 1987; p 715.
4. Fenton, D. E.; Vigato, P. A. Chem. Soc. Rev. 1988, 17, 69–90.
5. Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976, 734–736.
6. Evans, R. F. Aust. J. Chem. 1967, 20, 1643–1661.
7. Kra¨ssig, H. Makromol. Chem. 1956, 17, 77–130.
8. Fenton, D. E.; Moody, R.; Casellato, U.; Vigato, P. A.; Graziani, R. J. Heterocycl.
Chem. 1985, 22, 1639–1642.
9. Smith, K. M.; Ciccone, J. P.; Ramana, N. V. U.S. Patent 5,410,052, 1995.
10. Farrell, J. R.; Niconchuk, J.; Higham, C. S.; Bergeron, B. W. Tetrahedron Lett. 2007,
48, 8034–8036.
d
2.05 (br s, 1H, C2-OH), 3.75 (ddd, J¼12, 7, 1 Hz, 2H, H1B and H3B),
3.90 (ddd, J¼12, 5, 1 Hz, 2H, H1A and H3A), 4.28 (m, 1H, H2), 6.87
(ddd, J¼8, 7, 1 Hz, 2H, ArH), 6.95 (ddd, J¼8, 7, 0.5 Hz, 2H, ArH), 7.25
(dd, J¼8, 2 Hz, 2H, ArH), 7.31(ddd, J¼8, 7, 2 Hz, 2H, ArH), 8.40 (br s/t,
2H, N]CH), 13.05 (br s, 2H, phenyl OH). 13C NMR (CDCl3, 75 MHz)
d
63.2, 70.5, 116.9, 118.6, 118.7, 131.5, 132.5, 160.9, 167.3. LREI-MS m/z
298 ([M]þ). IR ( , cmꢀ1) 1633 (C]N), 3401 (OH). The spectroscopic
n
data for this compound is consistent with published data.19
11. Zelenin, K. N.; Alekseyev, V. V.; Ukraintsev, I. V.; Tselinsky, I. V. Org. Prep. Proced.
Int. 1998, 30, 53–61.
12. Moodie, R. B.; Moustras, M. Z.; Read, G.; Sandall, J. P. B. J. Chem. Soc., Perkin
Trans. 2 1997, 169–171.
4.5. Computational methods
´
´
¨ ¨
13. Go¨blyo¨s, A.; Lazar, L.; Fulop, F. Tetrahedron 2002, 58, 1011–1016.
14. La´za´r, L.; Fu¨lo¨p, F. Eur. J. Org. Chem. 2003, 2003, 3025–3042.
15. Sinkkonen, J.; Zelenin, K. N.; Potapov, A.-K. A.; Lagoda, I. V.; Alekseyev, V. V.;
Pihlaja, K. Tetrahedron 2003, 59, 1939–1950.
4.5.1. Calculation of heats of formation (DHf). All calculations of DHf
were performed on an SG Indy workstation using Spartan ’02 suite
of programs from Wavefunction Inc. (www.wavefun.com).42 Hex-
ahydropyrimidines 4a–j, the cis and trans forms of 5-hydroxyhexa
hydropyrimidines 9a–j and the cis and trans forms of the amino-
methyloxazolidines 10a–j were built in the chair conformation with
the N–H bonds and phenyl rings in the equatorial position. The
geometry of each compound was optimised at the semi-empirical
level using AM1 theory. Monoimines 5a–j and 7a–j and the bisi-
mines 3a–j and 8a–j were built with the imine bond in the anti
conformation. Monoimines 7a–j were built in the R configuration at
C2. A single derivative of each product was built and subjected to
a systematic conformational search at the molecular mechanics
level using the MMFF94 force field. The three lowest energy con-
formers of each compound were extracted and geometry optimi-
sation of each model was performed at the semi-empirical level
using AM1 theory. The lowest energy conformer after semi-em-
pirical optimisation was then used to build the remaining de-
rivatives and each molecule was then subjected to semi-empirical
geometry optimization (AM1 theory).
16. Maloshitskaya, O.; Sinkkonen, J.; Ovcharenko, V. V.; Zelenin, K. N.; Pihlaja, K.
Tetrahedron 2004, 60, 6913–6921.
17. Maloshitskaya, O. A.; Alekseyev, V. V.; Sinkkonen, J.; Zelenin, K. N.; Pihlaja, K.
Tetrahedron 2006, 62, 9456–9466.
18. Locke, J. M.; Crumbie, R. L.; Griffith, R.; Bailey, T. D.; Boyd, S.; Roberts, J. D. J. Org.
Chem. 2007, 72, 4156–4162.
19. Cameron, J. H.; McEwan, J. A. Inorg. Chim. Acta 1993, 211, 111–116.
20. Atakol, O.; Nazir, H.; Arici, C.; Durmus, S.; Svoboda, I.; Fuess, H. Inorg. Chim. Acta
2003, 342, 295–300.
21. Axenrod, T.; Sun, J.; Das, K. K.; Dave, P. R.; Forohar, F.; Kaselj, M.; Trivedi, N. J.;
Gilardi, R. D.; Flippen-Anderson, J. L. J. Org. Chem. 2000, 65, 1200–1206.
22. Astudillo, M. E. A.; Chokotho, N. C. J.; Jarvis, T. C.; Johnson, C. D.; Lewis, C. C.;
McDonnell, P. D. Tetrahedron 1985, 41, 5919–5928.
23. Paukstelis, J. V.; Hammaker, R. M. Tetrahedron Lett. 1968, 9, 3557–3560.
24. Dziembowska, T. Pol. J. Chem. 1998, 72, 193–209.
25. Hetenyi, A.; Szakonyi, Z.; Klika, K. D.; Pihlaja, K.; Fulop, F. J. Org. Chem. 2003, 68,
2175–2182.
26. Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96–103.
27. Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165–195.
28. Fleming, I. Frontier Orbitals and Organic Chemical Reactions; Wiley: London,
1976.
29. Karelson, M.; Lobanov, V. S.; Katritzky, A. R. Chem. Rev. 1996, 96, 1027–1044.
30. Yıldız, M.; Kılıç, Z.; Ho¨kelek, T. J. Mol. Struct. 1998, 441, 1–10.
31. Layer, R. W. Chem. Rev. 1963, 63, 489–510.
32. Pires, J.; Floriano, W. B.; Gaudio, A. C. J. Mol. Struct.: Theochem 1997, 389,
159–167.
33. Bultinck, P.; Goeminne, A.; Van de Vondel, D. J. Mol. Struct.: Theochem 1995, 339,
1–14.
34. Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory Chemicals, 4th ed.;
Butterworth Heinemann: Boston, 1996.
4.5.2. Calculation of the FMO energies and LUMO coefficients. These
calculations were executed using Spartan ’06 from Wavefunction
Inc.42 on an IBM X Series Server. The aldehyde was built and its
geometry optimised at the molecular mechanics level using the
MMFF as implemented in Spartan. A further geometry optimization
35. Bomfim, J. A. S.; Wardell, J. L.; Low, J. N.; Skakle, J. M.; Glidewell, C. Acta Crys-
tallogr. 2005, C61, o53–o56.
was then performed at the HF/6-311G level.
*
36. Billman, J. H.; Meisenheimer, J. L. J. Med. Chem. 1963, 6, 682–683.
37. Khan, M. S. Y.; Gupta, M. Pharmazie 2002, 57, 377–383.
38. Jentgens, C.; Bienz, S.; Hesse, M. Helv. Chim. Acta 1997, 80, 1133–1143.
39. Medou, M.; Priem, G.; Rocheblave, L.; Pepe, G.; Meyer, M.; Chermann, J.-C.;
Kraus, J.-L. Eur. J. Med. Chem. 1999, 34, 625–638.
40. Enders, D.; Wortmann, L.; Du¨cker, B.; Raabe, G. Helv. Chim. Acta 1999, 82,
1195–1201.
41. Lukov, V. V.; Kogan, V. A.; Novotortsev, V. M.; Tupolova, Y. P.; Gevorkyan, I. E.
Russ. J. Coord. Chem. 2005, 31, 353–356.
Acknowledgements
The authors gratefully acknowledge the Australian Government
for the receipt of an APA scholarship, as well as the Centre for Bio-
structural and Biomolecular Research (CBBR) for funding parts of this
research. We also thank Mrs.. Roza Dimeska for her very able tech-
nical assistance and the Intelligent Polymer Research Institute at the
University of Wollongong for the donation NMR spectroscopy time.
42. SPARTAN ’02 and SPARTAN ’06; Wavefunction: 18401 Von Karman Ave., Irvine,
CA 92715, USA. copyright 1991–2006, www.wavefun.com.