into the ice-cooled gas bomb, the coil was heated to about 60 °C
using a heat gun. The quantity of transferred DME was controlled
by weight. The filling was repeated until the necessary amount
of DME was transferred into the bomb. The corresponding
amount of ethane was transferred using the same procedure.
Research grade sulfur hexafluoride, laughing gas, and carbon
dioxide were obtained from Scott Specialty Gases (Plumsteadville,
PA). Xenon (>99.99%) was obtained from SL Gas, Lenzburg,
Switzerland. Xenon was transferred from the gas reservoir into
the pump as described in the literature.16
SFC/ MS. A Perkin-Elmer Sciex API III+ triple-quadrupole
mass spectrometer with an atmospheric pressure chemical ioniza-
tion device (Perkin-Elmer Sciex, Toronto, Canada) and a custom-
built interface (see Figure 2) was used for SFC/ MS experiments.
Nebulization gas (nitrogen >99.999%) pressure was held at 140
kPa (20 psi, resulting in a flow of 0.6 L/ min). If not mentioned
otherwise, MS conditions were as follows. For positive chemical
ionization, the corona discharge needle supplied a discharge
current of ∼3 µA at 8000 V. The interface plate was held at 650
V, and the orifice, at 75 V. The curtain gas flow (nitrogen
>99.999%) was maintained at 0.8 L/ min. The first quadrupole of
the mass spectrometer was scanned with a step rate of 0.4 between
m/ z 140 and 1000, a dwell time of 0.97 ms, and a pause time of
0.02 ms, resulting in 0.47 scan/ s.
Reagents and Samples. Tributylamine, N-ethylcyclohexyl-
amine, and triethylenetetramine (technical grade) were received
from Fluka, Buchs, Switzerland. If not stated otherwise, 1%
solutions in a 1:1 mixture of methanol and dichloromethane (both
from Fluka) were prepared and injected as such. The Mannich
bases characterized were prepared as follows. Mannich base 1
(MB-1) was prepared by the Mannich reaction of phenol, N,N-
dimethyl-1,3-propanediamine, and formaldehyde (36% solution in
water) (all from Fluka) (see Figure 3). Mannich base 2 (MB-2)
was prepared by transamination of 2,4,6-tris(dimethylaminometh-
yl)phenol (Ciba Specialty Chemicals, Basel, Switzerland) with N,N-
dimethyl-1,3-propanediamine (see Figure 4). Solutions of MB-1
and MB-2, 1% (w/ w), were prepared in methanol puriss. (Fluka)
and injected as such.
Figure 8. Oxazine derivative of the compound eluting at 8.2 min
(see Table 1).
such as decylamine and dodecylamine, can be eluted with CO2,
supporting the theory that only multifunctional primary amines
are not eluted by CO2.14,15 Despite their inability to form carba-
mates, N2O and xenon yield similar results, eluting the tertiary
and secondary amines of the test mixture, as well as decyl- and
dodecylamine, but not eluting TETA. The failure to elute TETA
and poor chromatographic resolution eliminate SF6 as a mobile
phase for phenolic Mannich bases. And ethane, as an inert alkane
with suitable critical values, does elute the tertiary and secondary
amines but again fails to elute TETA.
Trials to increase the elution power of CO2 through the addition
of polar modifiers such as methanol or 2-propanol did not yield
any useful results, as they quickly destroyed the column, resulting
in clogging of the restrictor and irreproducible results. As a
consequence, inert but less polar modifiers such as tetrahydro-
furan (THF) and dimethyl ether (DME) were tested for the
separation of the amine mixture. The high critical temperature of
THF (267.3 °C) and probably related but not further investigated
problems with UV detection resulted in the use of DME as the
modifier. For reasons of miscibility, and to avoid the formation of
carbamates, a mixture of DME with ethane was preferred to a
mixture of DME with CO2. Several mixtures were prepared with
different mixing ratios, and although 11% (mol/ mol) DME in
ethane elutes TETA, a mixture with 24% DME performed best
with respect to elution power and separation for the compounds
described below. As can be seen in Figure 6, it gives reasonable
separation of the three test compounds as well.
Although the use of mobile phases with higher amounts of
DME may be necessary for the elution of amines with higher
polarity or higher molecular weight compounds and a lower
concentration of DME may yield better separations for less polar
products, the mixture with 24% gave excellent results in most cases
and was used throughout this work. A 100% dimethylsiloxane
stationary phase was preferred to higher polar stationary phases
such as DB-17 (50% diphenylsiloxane) due to its higher resistance
against the supercritical mobile phase.
The application of this SFC method to the separation of the
constituents of technical phenolic Mannich bases reveals a
multitude of different products and byproducts (see Figure 7) and,
even more importantly, huge differences between so-called identi-
cal Mannich bases (Mannich bases consisting of the same phenol
and the same amine). Although differences in application-related
properties, such as viscosity and reactivity, can be detected
between these Mannich bases, investigations regarding differ-
ences in their structural composition have not previously been
carried out.
RESULTS AND DISCUSSION
The initial goal was to develop a capillary SFC method for the
separation of polar phenolic Mannich bases. An amine mixture
consisting of tributylamine [102-82-9] as tertiary amine, N-
ethylcyclohexylamine [5459-93-8] as secondary amine and trieth-
ylenetetramine (TETA [112-24-3]) as highly polar polyamine was
used to evaluate the elution power of different mobile phases with
respect to amines (see Figure 5).
Due to their different functionalities, these amines allow a good
assessment of each mobile phase for the elution of amines.
Additionally, TETA shows amine functionalities similar to those
of a multifunctional Mannich base, making it a good indicator for
the applicability of the mobile phase for their separation. As can
be seen in Figure 6 and in agreement with the literature, CO2
only elutes the secondary and tertiary amines of the test mixure;
TETA as a multifunctional amine is not eluted. Additional
investigations have shown that monofunctional primary amines,
In this case, SFC-UV coupled to MS is the method of choice
for the identification of unknowns. An SFC/ MS interface for
atmospheric pressure chemical ionization (APCI) similar to the
(16) Kirschner, C. H.; Taylor, L. T. J. High Resolut. Chromatogr. 1 9 9 4 , 17, 61-
67.
2328 Analytical Chemistry, Vol. 71, No. 13, July 1, 1999
Fuchslueger, Ulf
