Counterion effects in the preparation of aldehyde-bisulfite adducts

Article abstract of DOI:10.1016/j.tetlet.2013.09.105

The identification and development of an aldehyde-bisulfite adduct as an isolable starting material in the synthesis of the CETP inhibitor Evacetrapib are described. The physical properties of the sodium and potassium analogs are compared, and the extension of the scope of this study to include an investigation into the solid state properties of a range of sodium and potassium bisulfite adducts of commonly encountered aldehydes is discussed.

Full text of DOI:10.1016/j.tetlet.2013.09.105

Accepted Manuscript
Counterion effects in the preparation of aldehyde-bisulfite adducts
Marie G. Kissane, Scott A. Frank, Gregory A. Rener, Christopher P. Ley,
Charles A. Alt, Paul A. Stroud, Radhe K. Vaid, Sathish K. Boini, Laura A.
McKee, Jeffrey T. Vicenzi, Gregory A. Stephenson
PII:
S0040-4039(13)01670-5
DOI:
http://dx.doi.org/10.1016/j.tetlet.2013.09.105
TETL 43605
Reference:
To appear in:
Tetrahedron Letters
Received Date:
Revised Date:
Accepted Date:
2 September 2013
16 September 2013
21 September 2013
Please cite this article as: Kissane, M.G., Frank, S.A., Rener, G.A., Ley, C.P., Alt, C.A., Stroud, P.A., Vaid, R.K.,
Boini, S.K., McKee, L.A., Vicenzi, J.T., Stephenson, G.A., Counterion effects in the preparation of aldehyde-
bisulfite adducts, Tetrahedron Letters (2013), doi: http://dx.doi.org/10.1016/j.tetlet.2013.09.105
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Counterion effects in the preparation of aldehyde-bisulfite adducts
Marie G. Kissane,* Scott A. Frank,* Gregory A. Rener, Christopher P. Ley, Charles A.
Alt, Paul A. Stroud, Radhe K. Vaid, Sathish K. Boini, Laura A. McKee, Jeffrey T. Vicenzi,
Gregory A. Stephenson
Chemical Product R&D, Lilly Research Laboratories, Eli Lilly and Company, Indianapolis,
Indiana 46285, United States
*
Corresponding authors
Graphical Abstract
Abstract
The identification and development of an aldehyde-bisulfite adduct as an isolable starting
material in the synthesis of the CETP inhibitor Evacetrapib is described. The physical properties
of the sodium and potassium analogs are compared, and the extension of the scope of this study
to include an investigation into the solid state properties of a range of sodium and potassium
bisulfite adducts of commonly encountered aldehydes is discussed.
Keywords: aldehyde-bisulfite adduct, Rosenmund reduction, solid state properties, Evacetrapib
Introduction
Aldehydes are an important class of compounds in organic synthesis due to their functional
reactivity and participation in a range of reactions such as condensations, oxidations, reductions
1
and hydrations. Aldehydes are extensively utilized in the pharmaceutical industry, in
2
3
4
transformations such as organometallic coupling reactions, allylations, Wittig reactions and
5
reductive aminations. However, the long term storage of synthetically useful aldehydes can be
6
problematic as aldehydes often suffer from instability and non-crystallinity issues. These issues
can lead to challenges affecting variability in product purity and supply robustness. Many of
these challenges can be conveniently overcome by the preparation of aldehyde bisulfite adducts,
7
most commonly as the sodium salt by reaction of the aldehyde with sodium bisulfite. Bisulfite
-9
adducts are generally crystalline solids that are easy to handle and can be stored for prolonged

6
, 10-11
periods of time due to their enhanced stability.
In most cases, the bisulfite adducts are
readily cleaved back to their aldehyde precursors.
As part of an ongoing project into the synthesis of the CETP inhibitor Evacetrapib,
multikilogram quantities of aldehyde 1a were required. However, 1a, which is an oil, is rather
unstable due to its ability to epimerize readily and its potential to oxidize in air to the carboxylic
acid, and was thus deemed to be unsuitable for long term storage. It was hoped that the
conversion of 1a to the corresponding sodium bisulfite adduct 2a would lead to a crystalline
solid with the stability and solid state characteristics necessary for prolonged storage of the
material. It was found that 1a was easily converted to 2a by reaction with NaHSO
water. Unfortunately, the solid state characteristics of 2a were not very favourable; in particular
a was found to be quite hygroscopic, with a weight gain as high as 30% at relative humidities of
3
in THF and
2
up to 95%. This would have a major impact on the long term storage of 2a, as controlled storage
conditions would be required. Additionally, a thick slurry is formed during the preparation of 2a
resulting in long filtration times. As a result of these issues with the solid state properties, we
were interested in investigating if changing the counterion of the bisulfite adduct would lead to a
solid with improved physical properties. Although not as common as its sodium counterpart, the
1
2-14
preparation of potassium bisulfite adducts have also been reported.
The potassium bisulfite
adduct 3a was prepared and was found to have much improved solid state characteristics
compared to 2a. Herein is reported the study on the solid state characteristics of the sodium and
potassium bisulfite adducts of 1a. The scope of this study was also extended to include an
investigation into the solid state properties of a range of sodium and potassium bisulfite adducts
of commonly encountered aldehydes.
Results and Discussion
1
5
The aldehyde 1a was prepared as outlined in Scheme 1. Conversion of the monoacid 4 to the
acid chloride 5 is achieved by reaction of 4 with 1.14 eq. of oxalyl chloride and 0.1 eq. of 2,6-
lutidine in toluene at rt for 17 h. The acid chloride then undergoes a Rosenmund reduction by
1
6
reaction with 0.26 eq. of 5% Pd/C, 1.2 eq. of 2,6-lutidine in toluene to afford the aldehdye 1a,
with yields over the 2 steps typically 88-92%. Aldehyde 1a formed in this manner is typically
quite pure, with levels of the undesired cis-isomer consistently less than 0.5%.
4
5
1a

Scheme 1 Preparation of 1a
The sodium and potassium bisulfite adducts of the aldehyde 1a were easily prepared by reaction
of 1 equivalent of sodium or potassium bisulfite with 1 equivalent of 1a in a THF-water mixture
to give the sodium bisulfite adduct 2a and potassium bisulfite adduct 3a in yields of >90%
following stirring at room temperature for 2-16 h and filtration. The purity (determined by
quantitative NMR in DMSO using maleic acid as an internal standard) of 3a is usually up to 10%
greater than that of 2a, with water and suspended inorganic salts presumably accounting for the
remaining mass (by NMR spectroscopy and HPLC there are no other organic impurities present).
Additionally, aldehyde regenerated from either the sodium or potassium adduct is typically
equivalent in purity to the crude material emanating from the Rosenmund reduction. This
procedure is very amenable to scale-up and has been demonstrated successfully for both 2a and
3
a on multikilogram scales.
Table 1 Preparation of aldehyde bisulfite adducts
%
Na
K
% yield
Na
% purity % yield
a
Entry
Aldehyde
purity
a
K
adduct
adduct
Na
K
1
2a
2b
3a
3b
quant.
quant.
85
90
95
93
2
88
84
3
4
2c
3c
79
77
92
72
93
77
84
90
2d
3d

5
6
7
2e
2f
3e
3f
67
quant
91
84
97
91
75
92
89
94
93
O
Ph
H
2g
3g
100
8
2h
3h
91
89
86
98
2
a. Determined by quantitative NMR in DMSO or D O using maleic acid as an internal standard
In addition to an enhanced purity profile, the filterability properties of 3a were found to be much
more favorable than those of 2a. As mentioned earlier in the introduction, a thick slurry is
formed during the preparation of 2a which leads to long filtration times. In contrast, the slurry
formed during the preparation of 3a is much more flowable, leading to filtrations which are up to
6
4
3
times faster. The filtration of 2a and 3a was studied using a 200 mL Pocket Pressure Filter with
.0 cm diameter and Tetco filter medium at pressures of 30, 15 and 5 psi (see Table 2 and Table
2
). For 2a, the filtration flux ranged from 280-832 L/hm and the filtration rate from 0.09-0.26
2
g/s, while the corresponding values for 3a were 678-1700 L/hm and 0.22-0.54 g/s respectively.
Further optimisation of the preparation of 3a to include a co-addition of aqueous potassium
bisulfite and a THF solution of 1a to a solution of THF and water containing a seedbed of 3a at
4
1
5 °C have led to further enhancement of the filtration properties, with a filtration flux of up to
2
1,200 L/hm and a filtration rate of up to 3.6 g/s obtained.
Table 2 Filtration study of 2a
Pressure (psi)
30
15
5
Filtration rate (g/s)
0.26
0.17
551
0.09
280
2
8
32
Filtration flux (L/hm )
Table 3 Filtration study of 3a
Pressure (psi)
30
15
5

Filtration rate (g/s)
0.54
700
0.43
0.22
678
2
1
1341
Filtration flux (L/hm )
With the benefits of enhanced filterability, we desired to further explore the physical property
difference of the sodium and potassium bisulfite adducts. In a variable humidity solid state
experiment, the sodium bisulfite adduct 2a was found to be quite hygroscopic, with a weight
gain of 27% at relative humidity of up to 95%. The corresponding moisture sorption pattern for
the potassium bisulfite adduct 3a showed it to be much less hygroscopic, with a weight gain of
just 2% at relative humidity of up to 95% (entry 1, Table 5 ). Presumably, the sodium adduct
exists as a variable hydrate, while the potassium adduct exists in the solid state as a hemihydrate,
which was further confirmed by single crystal analysis on a sample that was crystallized from
2
EtOH-H O. The X-ray also verified the sulfonate structure (Figure 1 and Figure 2). Despite
repeated attempts, we were unable to grow a crystal of 2a suitable for single crystal analysis.
Figure 1 X-Ray structure of 3a (unit cell viewed down the y-axis)

Figure 2 Hydrogen bonded ribbons propagate along the b-direction
The thermal profile of both 2a and 3a was recorded using TGA-DSC analysis. As shown in
Figure 1 and Figure 2, the potassium bisulfite adduct 3a is usually isolated as the hemihydrate,
and TGA analysis confirms this with a 4.1% loss in mass observed between 80-130 °C, although
this is a little higher than the calculated wt% water for a hemihydrate (2.9%). Meanwhile, the
sodium bisulfite adduct 2a showed a 9.1% mass loss between 50-100 °C in the TGA profile (the
calculated wt% water for a monohydrate is 6.2%). Interestingly, decomposition rather than a
phase transition from solid to liquid was observed for both salts, with a significant mass loss
accompanying each major thermal event. The onset temperature of decomposition for 2a was
1
52 °C, while the corresponding temperature for 3a was 168 °C (entry 1, Table 6).
The amount of water present was further verified by oven Karl Fisher coulometric titrations
standard volumetric Karl Fisher titrations were not possible due to interactions between the
(
bisulfite ion and the Karl Fisher titrant). The average water content in samples of the sodium
bisulfite adduct 2a is usually 8% (the theoretical amount for the monohydrate is 6%), while the
amount of water present in samples of the potassium bisulfite adduct 3a is typically 3%, which
corresponds to the theoretical amount for the hemihydrate (see Table 6).

Table 4 pH Measurements of bisulfite adducts
a
a
Entry
Na adduct
K adduct
pH Na
6.5
pH K
5.2
5.1
5.0
4.9
6.1
4.9
3.7
4.6
1
2
3
4
5
6
7
8
2a
2b
2c
2d
2e
2f
3a
3b
3c
3d
3e
3f
5.7
5.8
6.4
7.3
5.8
2g
2h
3g
3h
6.0
5.3
a. pH of a 1% aqueous solution
Following the successful preparation of 2a and 3a, the scope of this study was extended to
include an investigation into the solid state properties of a range of sodium and potassium
bisulfite adducts of commonly encountered aldehydes. To that end, the sodium and potassium
1
1
bisulfite adducts of cyclohexane carboxaldehyde, butyraldehyde, 2-methylbutyraldehyde,
1
1
10
hexanal, benzaldehyde, phenylacetaldehyde and diphenylacetaldehyde were prepared (Table
, entries 2-9) in yields ranging from 67-100% and purities of 72-100%. In general, the
1
preparation of the potassium bisulfite aldehyde adducts gave material of higher purity than the
sodium counterparts. The pH values of 1% aqueous solutions of each of 2a-2h and 3a-3h were
measured, and in most instances the potassium bisulfite adducts were found to be more acidic
than their sodium analogues (Table 4, entries 2-8).
The sodium and potassium bisulfite adducts of cyclohexane carboxaldehyde 2b and 3b had
similar filterability properties to 2a and 3a, with 2b forming a thick slurry and had longer
filtration times than 3b. For all the other adducts 2c-2h and 3c-3h, the filtration properties were
comparable for the sodium and potassium salts.
We were particularly interested in determining if a similar trend to that observed in the moisture
sorption patterns of 2a and 3a would be seen for 2b-2h and 3b-3h. The sodium bisulfite adducts
of cyclohexane carboxaldehyde 2b, hexanal 2e, phenylacetaldehyde 2g and
diphenylacetaldehyde 2h were found to absorb moisture much more readily than the
corresponding potassium bisulfite adducts (Table 5). The sodium and potassium bisulfite adducts
of butyraldehyde 2c and 3c and 2-methylbutyraldehyde 2d and 3d were both found to be quite

hygroscopic, with a weight gain of up to 63% observed at 95% relative humidity for 2c and 3c
Table 5, entry 3) and a weight gain of up to 37% observed for 2d and 3d (Table 5, entry 4),
(
while a low mass increase (up to 3%) was seen for the bisulfite adducts 2f and 3f at similar
relative humidity (Table 5, entry 6)-see supporting information for moisture sorption patterns.
Table 5 Moisture sorption of aldehyde bisulfite adducts
a
a
Entry
Na
K
Moisture sorption Na
Moisture sorption K
(%wt gain)
(
%wt gain)
1
2
3
4
5
6
7
8
2a
2b
2c
2d
2e
2f
3a
3b
3c
3d
3e
3f
26.7
2.0
7.0
29.6
61.3
63.2
37.4
2.5
32.9
19.5
3.3
3.4
2g
2h
3g
3h
7.0
1.3
17.7
0.2
a. At 95% relative humidity
The thermal profiles of 2b-2h and 3b-3h were also recorded using TGA-DSC analysis. As the
decomposition of 2d had an onset temperature of 71 °C and 2h had an onset temperature of 75
°C, it was not possible to determine how much water was present in these samples from TGA
analysis (Table 6, entries 4 and 8). The cyclohexane carboxaldeyde adducts 2b and 3b behaved
similar to 2a and 3a, with a mass loss of 10.3% in the TGA of 2b indicative of a monohydrate,
while the 3.4% mass loss in the TGA of 3b corresponds to the hemihydrate (Table 6, entry 2).
The potassium bisulfite adduct of 2-methylbutyraldehyde 3d was also isolated as the
hemihydrates (confirmed by the 3.7% loss in mass in the TGA, Table 6, entry 4). All other
adducts had minimal loss in mass around 100 °C. Again, similar to 2a and 3a, decomposition
rather than a phase transition from solid to liquid was observed for each salt, with a significant
mass loss accompanying each major thermal event. The onset temperatures of the decomposition
varied from 75 °C for 2h up to 170 °C for 3g (Table 6, entries 2-9).

The amount of water in each sample was also measured by oven KF, with good correlation in all
instances between the KF data and the TGA data. As the oven KF was measured at 90 °C, which
is above the onset temperatures of degradation for 2d and 2h, the KF values are likely not
representative of the amount of water present in these samples.
Table 6 Thermal analysis and oven KF of aldehyde bisulfite adducts
%
TGA
9.1
10
1.7
-
H
2
O Na
KF
2
% H O K
Na
K
T decomp T decomp
Entry
adduct adduct Na (° C)
K (° C)
a
b
a
b
TGA
KF
1
2
3
4
5
6
7
8
2a
2b
2c
2d
2e
2f
3a
3b
3c
3d
3e
3f
152
127
139
71
168
157
142
113
143
153
170
147
8
4.1
3.4
0.6
3.7
0
3
11.8
1.8
2.0
0.9
8.3
4.0
157
127
187
75
0
<0.4
<0.4
<0.4
11.9
<0.4
<0.4
<0.4
<0.4
0
0
2g
2h
3g
3h
0.4
-
0.7
0.3
a. Estimated from TGA by loss in mass at 100 °C
b. Measured by oven KF at 90-115 °C
As multigram quantities of the free aldehyde 1a were required for subsequent chemistry, the
generation of 1a from 2a and 3a was investigated. The use of NEt led to <25% conversion to
a, while TMSCl led to >95% conversion but was accompanied by 30% isomerisation to the
3
1
1
1
1
0
cis isomer. Aqueous basic hydrolytic cleavage proved to be much more successful, with K
and Na CO giving complete conversion to the desired aldehyde with minimal formation of the
cis isomer. Na CO is now the preferred base as better layer separations are observed. The
optimized conditions involve reaction of the bisulfite adduct with 3 equivalents of Na CO in 5
volumes of water and 5 volumes of toluene or 2-MeTHF at room temperature for 0.5-2 h to give
a in yields of >90% (entries 1 and 2, Table 7). Under these conditions, the level of
2 3
CO
2
3
2
3
2
3
1
epmierization is typically <0.5%. The regeneration of cyclohexanecarboxaldehyde 1b from
bisulfite adducts 2b and 3b (entries 3 and 4, Table 7) and diphenylacetaldehyde 1h from 2h and
3
h (entries 5 and 6, Table 7).

Table 7 Regeneration of aldehyde from bisulfite adduct
Entry
Bisulfite Adduct
Aldehdye
% Yield
90-95
90-95
57
1
2
3
4
5
6
2a
3a
2b
3b
2h
3h
1a
1a
1b
1b
1h
1h
55
89
95
Conclusion
An aldehyde-bisulfite adduct was identified and developed as a stable, isolable starting material
in the synthesis of the CETP inhibitor Evacetrapib. Both the sodium and potassium bisulfite
adducts were prepared in multikilogram scales, and the potassium bisulfite was found to have
more favorable solid state properties in terms of its filterability and hygroscopicity. The scope of
the study was extended to include an investigation into the solid state properties of a range of
sodium and potassium bisulfite adducts of commonly encountered aldehydes. In general, the
solid state properties of the potassium bisulfite adducts were more advantageous.
Author Information
Corresponding authors email: kissane_marie_g@lilly.com, frank_scott_a@lilly.com
Acknowledgements

We wish to thank Steven Bandy for HRMS, Shanna Neely for qNMR advice, Lori Hilden for
single crystal analysis, Luke Lutz for XRPD, thermal and moisture sorption analysis and Steve
Hulbert for KF analysis.
References
1
2
. March, J, Advanced Organic Chemistry, Fourth Edition, Wiley-Interscience, 1992.
. Yu, H.; Richey, R.N.; Stout, J.R.; LaPack, M.A.; Gu, R.; Khau, V.V.; Frank, S.A.; Ott,
J.P.; Miller, R.D.; Carr, M.A., Zhang, T.Y., Org. Proc. Res. Dev. 2008, 12, 218.
3
. Mickel, S.J.; Sedelmeier, G.H.; Niedereer, D.; Schuerch, F.; Seger, M.; Schreiner, K.;
Daeffler, R.; Osmani, A.; Bixel, D.; Loiseleur, O.; Cercus, J.; Stettler, H.; Schaer, K.;
Gamboni, R., Org. Proc. Res. Dev. 2004, 8, 113.
4
5
6
7
8
. Alimardanov, A.; Schmid, J.; Afragola, J.; Khafizova, G., Org. Proc. Res. Dev. 2008, 12,
4
24.
. Cooper, C.G.F.; Lee, E.R.; Silva, R.A.; Bourque, A.J.; Clark, S.; Katti, S.; Nivorozhkin,
V., Org. Proc. Res. Dev. 2012, 16, 1090.
. Ragan, J.A.; Am Ende, D.J.; Brebek, S.J.; Eisenbeis, S.A.; Singer, R.A.; Tickner, D.L.;
Teixeira, J.J.; Vanderplas, B.C.; Weston, N, Org. Proc. Res. Dev. 2003, 7, 155.
. Cong, C.; Wang, H.; Hu, Y.; Liu, C.; Ma, S.; Li, Xin,; Cao, J.; Ma, S., Eur. J. Med.
Chem., 2011, 46, 3105.
. Liu, J.; Yi, W.; Hu, J.; Wu, F.; Zhao, L.; Song, H.; Wang, Z., Chem. Pharm. Bull., 2010,
5
8, 1127.
9
1
1
1
1
. Ozkay, Y.; Tunali, Y.; Karaca, H.; Isikdag, I., Eur. J. Med. Chem., 2010, 45, 3293.
0. Kjell, D.P.; Slattery, B.J.; Semo, M.J., J. Org. Chem. 1999, 64, 5722.
1. Pandit, C.R.; Mani, N.S., Synthesis, 2009, 4032.
2. Haines, A.H.; Hughes, D.L., Carbohydr. Res. 2010, 345, 2705.
3. Cole, E.R.; Craig, D.C.; Fitzpatrick, L.J.; Hibbert, D.B.; Stevens, J.D., Carbohydr. Res.
2
001, 335, 1.
1
1
4. Kuroda, T.; Ashida, T.; Sasada, Y.; Kakudo, M., Bull. Chem. Soc. Jap., 1967, 40, 1377.
5. Sawada, N.; Uehara, R.; Kabuto, S.; Kyotani, Y.; Kawamura, K. Jpn. Kokai Tokkyo
Koho, 1974, JP 49048639 A 19740511.

1
6. Houpis, I. N.; Molina, A.; Reamer, R. A.; Lynch, J. E.; Volante, R. P.; Reider, P. J.
Tetrahedron Lett. 1993, 34, 2593.