The Synthesis of 2,6,7-Trioxa-1,4-diphosphabicyclo[2.2.2]octane Revisited: The Synthesis of 2,6,7-Triphenyl-2N,6N,7N-triaza-1,4-diphosphabicyclo[2.2.2]octane and the Synthesis of 1λ5-Phosphiranol

Article abstract of DOI:10.1002/hc.8

We report herein a reliable synthesis of P(OCH2)3P in a study aimed at understanding the factors that control its formation. We also report the serendipitous synthesis of the novel phosphiranol (CH2)2PO2H and the synthesis of P(PhNCH2)3P in good yield.

Full text of DOI:10.1002/hc.8

Heteroatom Chemistry
Volume 12, Number 2, 2001
The Synthesis of 2,6,7-Trioxa-1,4-
diphosphabicyclo[2.2.2]octane Revisited: The
Synthesis of 2,6,7-Triphenyl-2N,6N,7N-triaza-
1,4-diphosphabicyclo[2.2.2]octane and the
Synthesis of 1k5-Phosphiranol
Philip Kisanga and John Verkade
Department of Chemistry, Iowa State University, Ames, IA 50011
Received 19 September 2000; revised 10 November 2000
species, 4 analogous to 3 has been reported [6], to
ABSTRACT: We report herein a reliable synthesis of
our knowledge no mention of a stable phosphiranol
has been made in the literature. The only report of a
2N,6N,7N-triaza-1,4-diphosphabicyclo[2.2.2]octane
is that for 5, which was reported from our labora-
P(OCH2)3P in a study aimed at understanding the fac-
tors that control its formation. We also report the ser-
endipitous synthesis of the novel phosphiranol
(CH2)2PO2H and the synthesis of P(PhNCH2)3P in
tories to form in low yield via Reaction 1 [7]. Here
good yield. ᭧ 2001 John Wiley & Sons, Inc. Hetero-
we describe a more efficient route to this ligand.
atom Chem 12:114–117, 2001
INTRODUCTION
The bicyclic ligand 1 was first reported from our lab-
oratories [1–3] and later by others [4]. Because of
anecdotal information gathered from other investi-
gators regarding difficulties in repeating our synthe-
sis of 1 from 2 and indeed our own subsequent prob-
lems with consistently repeating the synthesis of 1,
we decided to examine this reaction more closely.
The desire to pursue a more reliable synthesis was
spurred by the observation that 1 can act as a ligand
that forms interesting metal complexes [5]. During
the course of the present study, we serendipitously
encountered a synthesis for 3. Although a transient
(1)
RESULTS AND DISCUSSION
We previously showed that the synthesis of 2 from
phosphonium salt 6 in methanol with 1 equiv. of
commercial NaOMe is incomplete (Reaction 2) [2].
Here we confirm that this is the case even with the
use of up to 2.0 equiv. of commercial NaOMe. Thus,
addition of 1.05 equiv. of a methanolic suspension of
commercial NaOMe to 6 followed by stirring for 15–
Correspondence to: John Verkade.
Contract Grant Sponsor: Donors of the PRF administered by
the American Chemical Society.
᭧ 2001 John Wiley & Sons, Inc.
114

The Synthesis of 2,6,7-Trioxa-1,4-diphosphabicyclo[2.2.2]octane Revisited 115
20 minutes resulted in a ca. 65% conversion to 2 as
by Ellis and coworkers [8] (i.e., treatment of anhy-
drous 6 with dry triethylamine) led to a 42% yield of
1 when transesterified with P(OMe)3 in refluxing
THF. Compound 2 prepared in this way contained
less than 5% of 7 as the only impurity. When 0.25
equiv. of 6 was deliberately added to a sample of 2
produced by the method of Ellis et al. [8] and the
reaction with P(OMe)3 was attempted, a 46% iso-
lated yield of 3 was obtained. This suggests that the
formation of 3 stems from 2 and not its oxide 7,
which under these circumstances would be expected
to produce only trace amounts of 3 because the con-
centration of 7 is so low after the first step (Ͻ5%).
To study the effect of other potentially damaging
impurities on the formation of 1, water was added
to 2 (freshly made by the method of Ellis et al. [8])
in increasing amounts before the reaction with tri-
methyl phosphite in refluxing THF. In the presence
of two drops of water, the reaction of 0.1 mol of 2
afforded a 21% yield of 1. This yield decreased to 9%
when the amount of water was increased to 0.5 mol
equiv, and only trace amounts of 1 were isolated
when the amount of water was increased to 1.0
equiv. When small amounts (two drops) of dilute HCl
were added to 0.1 mol of 2 followed by the trimethyl
phosphite reaction in refluxing THF for 16 hours, a
1:1 mixture of 3 and 8 was obtained in a total yield
estimated by 1H and 31P NMR integration.
P(CH2OH)4Cl ם NaOMe →
6
(2)
P(CH2OH)3 ם CH2O ם MeOH ם NaCl
2
1H NMR spectroscopic analysis showed the reaction
mixture obtained (after filtration and removal of sol-
vent from the filtrate) to consist of a mixture of 65%
of 2 along with its oxidized form 7, 30% 6, and about
5% paraformaldehyde. However, by adding 1.2
equiv. of freshly prepared NaOMe (made by reacting
sodium with methanol), 2 was produced from 6 in
74% yield, in addition to 18% of 7 and ca. 8% para-
1
formaldehyde as estimated by H NMR integration.
Analysis of this reaction mixture by 31P NMR spec-
troscopy showed it to be free of starting material 6
(d ס 27.0). The presence of 7 is apparently due to
the oxidation of 2 with adventitious oxygen. This ob-
servation had also been reported previously by Ellis
et al [8]. By bubbling air through a CD3OD solution
of 2, we were able to observe a gradual decrease in
the 31P NMR signal corresponding to 2 and a con-
comitant increase in the signal for 7 (d 31P ס 48.7).
Complete conversion to 7 occurred after 72 hours.
of 37%. This experiment indicated a possible role for
ם
H
in the formation of 3. A rationale for the forma-
tion of 8 in this reaction is suggested in Scheme 1,
wherein formaldehyde originates from the prepara-
tion of 2 (Reactions 2 and 3).
When a formaldehyde-free sample of 2 (pre-
pared by a procedure from the literature [8]) was
allowed to react with trimethyl phosphite in the
presence of catalytic amounts of dilute HCl, the only
product obtained upon sublimation was 3. To fur-
ther support the role of the acid in the formation of
3, compound 2 was reacted with trimethyl phosphite
in THF in the presence of catalytic amounts (5%) of
p-toluene sulfonic acid (PTSA). After removal of the
volatiles and sublimation of the residue, a 36% iso-
lated yield of 3 was obtained. We have so far found
no evidence for the formation of 3 in the absence of
acid, 6 or P(OMe3). This suggests that in the presence
of 6, HCl is perhaps produced as shown in Reaction
4. The sequence of reactions in Scheme 2 represents
a highly speculative attempt to rationalize the for-
P(CH2OH)4Cl ם NaOH →
6
P(CH2OH)3 ם CH2O ם H2O ם NaCl
(3)
2
The reaction of an ethanolic solution of 6 with one
equiv. of ethanolic NaOH was complete in 15 min-
utes at room temperature as shown by both H and
1
13C NMR spectroscopic analysis. Even though water
is formed in this reaction as well as paraformalde-
hyde (Reaction 3), we selected this route for the syn-
thesis of 2 because of its simplicity and complete
conversion of 6 to product 2. Thus, after filtration of
the reaction mixture under vacuum, the filtrate was
passed through anhydrous potassium carbonate
˚
onto 4A molecular sieves. The solvent was removed
under vacuum at room temperature to afford 2,
which was then transesterified with P(OMe)3 in re-
fluxing tetrahydrofuran (THF) to give 1 in yields con-
sistently in the range of 29–32%, which is compa-
rable to the 32% yield that we reported originally
[1,2]. A sample of 2 prepared by the method reported
SCHEME 1

116 Kisanga and Verkade
SCHEME 3
SCHEME 2
mation of the novel phosphiranol 3 from 2, which is
facilitated by a source of aqueous acid. The forma-
tion of methanol in this reaction was confirmed by
1H and 13C NMR spectroscopy and by observation of
SCHEME 4
1
the growth in the H NMR resonances assigned to
the methanol OH and CH3 protons upon adding
methanol to the reaction mixture. Although the for-
mation of 3 could in part be attributed to resonance
stabilization of its anion formed on proton dissoci-
ation of 3, it should be noted that water solutions of
3 are neutral to litmus, indicating that 3 is a very
weak acid. This accords with our failure to separate
a mixture of 3 and 8 using aqueous NaOH. On the
other hand, stability for 3 may be achieved via H-
bonding as depicted in dimer (3)2 or perhaps an anal-
ogous trimer.
alogues of 5. The products of such reactions with
methyl or ethylamine, for example, consist mainly of
uncharacterized oligomers.
EXPERIMENTAL
All reactions were carried out under nitrogen. All sol-
vents were dried according to standard procedures.
Preparation of Chloride-Free 2 Containing
20% 7
[P(CH2OH)4Cl ם P(OMe)3 → P(CH2OH)3
6
2
Compound 2 was prepared according to a published
method using ethanolic sodium hydroxide [1]. After
stirring of the solution for 15 minutes, it was allowed
to settle for 1 hour. The clear ethanolic solution was
vacuum transferred by means of a cannula, through
a fritted filter tube packed with anhydrous potas-
sium carbonate into a Schlenk flask containing ac-
ם Me(O)P(OMe)2 ם CH2O ם HCl
(4)
˚
tivated 4A molecular sieves. The solution was stored
over the sieves for 6 hours after which it was trans-
ferred by means of a cannula into a round-bottomed
flask. Upon removal of the solvent by distillation un-
der reduced pressure at 40ЊC, a clear viscous liquid
was obtained that consisted of 2 plus ca. 20% of its
oxidized form 7 and about 7% of paraformaldehyde
as determined by 1H NMR integration.
Because of the poor yield we obtained previously for
5 [7], we sought a more practical synthesis. The re-
quired triamine P(CH2NHPh)3 9 (Scheme 3) was eas-
ily prepared in 83% yield by the method reported by
Frank and Drake [9]. However, the conversion of 9
to 5 proved to be challenging. The attempted trans-
amination of P(NMe2)3 with 9 at 90ЊC followed by
heating for 72 hours afforded only 18% of 5 as a
white solid with a melting point of 194–196ЊC, which
agreed favorably with that previously observed in
our laboratories [7]. We were disappointed by the
low yield, however. A 70% yield of 5 was achieved
using CIP(NMe2)2 as a reagent as shown in Scheme
4. Thus far, we have been unable by the first step in
Scheme 3 to obtain more than very small yields of
P(CH2NHR)3, which is required to form trialkyl an-
Preparation of 2 Containing [P(CH2OH)4]Cl (6)
To a round-bottomed flask containing 1 equiv. of 6
dissolved in 250 mL of methanol was added 1 equiv.
of commercial NaOMe (Aldrich) suspended in 65 mL
of methanol. The reaction mixture was stirred at
room temperature for 15 minutes. The precipitated
NaCl was removed by filtration without protection
from the atmosphere, and the volatiles were re-
moved under reduced pressure to afford 15.2 g of

The Synthesis of 2,6,7-Trioxa-1,4-diphosphabicyclo[2.2.2]octane Revisited 117
1
impure 2 (as shown by H and 31P NMR spectros-
copy) that was used for the preparation of 3 without
further purification.
cool to room temperature. The volatiles were re-
moved under reduced pressure, and the residue was
sublimed under vacuum to afford 0.15 g (36% yield)
of 3 after recrystallization of the sublimate from
benzene.
Preparation of 1 from Chloride-Free 2
This was achieved using the sample of 2 described
previously using a procedure detailed elsewhere
[1,2].
Preparation of 5
To ca. 8.6 mmol of CIP(NMe2)2 prepared in situ in
the equilibrium reaction of 2.8 mmol of PCl3 with
5.73 mmol of P(NMe2)3 in 5 mL of acetonitrile at 0ЊC
was syringed 3.0 g (8.6 mmol) of P(CH2NHPh)3 [9]
dissolved in 10 mL of dry acetonitrile, whereupon a
white precipitate formed immediately. The reaction
mixture was allowed to stir for 2 hours at room tem-
perature, after which it was filtered and the precip-
itate was washed twice with 5 mL portions of cold
dry acetonitrile. Drying of the residue under vacuum
Preparation of 3 from Chloride-Free 2
To a round-bottomed flask connected to a water con-
denser was added 1.5 g of 6 (7.8 mmol) under nitro-
gen. Using a syringe, 4.5 g (37.5 mmol) of chloride-
free 2 was added, followed by 30 mL of dry THF. The
flask was then warmed to 50ЊC, 6.0 g (0.05 mol) of
trimethyl phosphite was added dropwise, and the re-
action mixture was refluxed for 16 hours. Removal
of the volatiles under reduced pressure followed by
sublimation afforded a white solid that was purified
by recrystallization from benzene to afford 1.4 g
1
afforded 2.30 g (71%) of 5. H NMR (CDCl3): d 3.63
(dd, 6H), 6.88 (t, 3H), 7.08 (m, 6H), 7.24 (m, 6H). 13
C
NMR (CDCl3): d 148.6 (d, J ס 21 Hz), 148.5, 129.3,
120.4, 116.0, 115.9 (d, J ס 14 Hz), 40.4 (d, J ס 13
Hz). 31P NMR (CDCl3): d 49.7 (d), מ45.32 (d). HRMS
Calcd. for C21H21N3P2 377.1211, observed m/e (M )
377.1210.
1
(34%) of 3. H NMR (CDCl3): d 1.47 (dd, 4H), 8.96
ם
(bs, 1H). 13C NMR (CDCl3): d 18.01 (d, J ס 69 Hz).
31P NMR (CDCl3): d 39.76. LRMS Calcd. for C2H6O2P,
ם
93.04; observed m/e (M ם H ) 93.02. Phosphiranol
3 is a white crystalline solid that melts at 142–143ЊC.
It is soluble in most organic solvents but not in pen-
tane. When left in air, 3 slowly absorbed water to
form a clear colorless solution that showed no
ACKNOWLEDGMENT
The authors are grateful to Professor Robson
Matos for experimental assistance of the PRF ad-
ministered by the American Chemical Society for
grant support.
1
change in either its H or 31P NMR spectrum. At-
tempts to grow crystals of 3 suitable for X-ray dif-
fraction failed.
REFERENCES
Preparation of 3 Using 2 Containing 6
[1] Coskran, K. J.; Verkade, J. G. Inorg Chem 1965, 4,
1655.
[2] Rathke, J. W.; Guyer, J. W.; Verkade, J. G. J Org Chem
1970, 35, 2310.
[3] Volcko, E. J.; Verkade, J. G. Phosphorus Sulfur 1984,
21, 111.
[4] Koslov, E. S.; Tovstenko, V. I. Zh Obschch Khim 1980,
50, 1210.
To a round-bottomed flask connected to a condenser
was added 6.2 g of impure 2 containing 6. Dry THF
was then added, and the flask was warmed to 50ЊC
after which 6.0 g (0.05 mol) of trimethyl phosphite
was added. The reaction was continued as detailed
in the previous paragraph to afford 1.2 g (26%) of 3.
[5] (a) Allison, D. A.; Clardy, J.; Verkade, J. G. Inorg Chem
1972, 11, 2804; (b) Stricklen, P. M.; Volcko, E. J.; Ver-
kade, J. G. J Am Chem Soc 1983, 105, 2494.
[6] Goeller, A.; Heydt, H.; Clark, T. J Org Chem 1996, 61,
5840.
[7] Stricklen, P. Ph.D. Dissertation, Iowa State University,
Ames, IA, 1970.
[8] Ellis, J. W.; Harrison, K. N.; Hoye, P. A.; Orpen, A. G.;
Pringle, P. G.; Smith, M. B. J Inorg Chem 1992, 31,
3026.
[9] Frank, A. W.; Drake, G. L., Jr. J Org Chem 1972, 37,
2752.
Preparation of 3 Using PTSA
To 0.55 g (4.4 mmol) of 2 (prepared by the method
reported by Ellis and co-workers) [8] in a round-bot-
tomed flask was added 30 mL of dry THF under ni-
trogen. The flask was warmed to 50ЊC with constant
stirring followed by the addition of 5.0 mg PTSA dis-
solved in 2 mL of THF. The reaction mixture was
refluxed for 48 hours, after which it was allowed to