Journal of Chromatography, 431 (1988) 351-364

Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

CHROM. 20 215

SEPARATION OF OPIUM ALKALOIDS BY CARBON DIOXIDE SUB- AND

SUPERCRITICAL FLUID CHROMATOGRAPHY WITH PACKED COL-
UMNS

APPLICATION TO THE QUANTITATIVE ANALYSIS OF POPPY STRAW

EXTRACTS

J. L. JANICOT, M. CAUDE* and R. ROSSET

Laboratoire de Chimie Analytique de I’.??coleSuperieure de Physique et de Chimie de Paris, 10 rue Vauquelin,
75231 Paris Cedex OS (France)
(First received July 31st, 1987; revised manuscript received November 26th, 1987)

SUMMARY

The optimization of the separation of seven opium alkaloids by sub- and su-
percritical fluid chromatography (SFC) using packed columns and carbon dioxide as
the primary mobile phase was studied. The influence of aminated polar modifiers on
bare and aminopropyl-bonded silica was investigated. It was found that the presence
of an amine in the mobile phase could enhance the retention of alkaloids on ami-
nopropylsilica.
Various separation schemes were possible depending on the type of analysis
needed. An aminopropyl-bonded silica used with a carbon dioxideemethanol-
triethylamine-water mixture (82.95:16.25:0.50:0.30, w/w) gave a very rapid separa-
tion (2 min). A bare silica with a carbon dioxide-methanol-methylamine-water mix-
ture (83.37:16.25:0.15:0.23, w/w) gave longer analysis (10 min) but a higher resolu-
tion. This last procedure was applied to a poppy straw extract and to the determi-
nation of three alkaloids of interest after a peak purity study using a diode-array UV
detector. For these alkaloids, SFC appears to be a promising technique for the
routine analysis of opium alkaloids.

INTRODUCTION

The use of supercritical fluids as mobile phases for chromatography has re-
cently become popular and this method is now an attractive one for difficult sepa-
rationslp5. In comparison with those used in liquid chromatography, these fluids offer
lower viscosities and higher analyte diffusivities and hence higher efficiencies per unit
time and, consequently, faster analysis are obtained6.
It seemed interesting to apply supercritical fluid chromatography (SFC) to
opium alkaloid separations and especially to quantitative analysis of poppy straw
extracts because in this instance high-performance liquid chromatography (HPLC)

0021-9673/88/$03.50 0 1988 Elsevier Science Publishers B.V.

352 J. L. JANICOT, M. CAUDE, R. ROSSET

is time consuming or requires the use of gradients 7- l l. However, the polar character
of opium alkaloids necessitates the addition of polar modifiers to carbon dioxide, the
most frequently used supercritical fluid12-14.
In this paper, we describe the optimization of the separation of seven major
opium alkaloids by SFC using carbon dioxide-methanol-amine-water mixtures as
the mobile phase. The procedures were applied to a poppy straw concentrate, aiming
not at a qualitative screening but at a quantitative determination of three alkaloids
of interest (thebaine, codeine and morphine).

EXPERIMENTAL

The apparatus has been described elsewhere 4. Experiments were carried out
using a Varian 5500 chromatograph (Varian, Palo Alto, CA, U.S.A.) supplied with
carbon dioxide by a cylinder equipped with an eductor tube. The pump efficiency
was improved by maintaining the pump head in close thermal contact with a clamp-
on heat exchanger cooled to 0°C by a recirculating refrigerated ethanol bath. As
it is impossible to use the proportioning valves of the Varian chromatograph under
the carbon dioxide cylinder pressure (55 bar), polar modifiers were added to carbon
dioxide through a Gilson Model 302 pump (Gilson, Villiers-le Bel, France). Tem-
perature control was achieved with a constant-temperature water-bath (Type TX 9,
P. M. Tamson, E.V., Zoetermeer, The Netherlands). A Varian UV 200 spectropho-
tometer was used with a special detection cell modified to permit detection up to 350
bar. A Varian Polychrom 9060 diode-array spectrophotometer did not need any cell
modification to be used within the SFC pressure range. The pressure was controlled
by a back-pressure regulator (Model 26-3220-24004, Tescom, Minneapolis, MN,
U.S.A.).

Chromatographic columns
Stainless-steel columns (12-23 cm x 0.46 cm I.D.) were packed using a con-
ventional technique. Various stationary phases were used: (i) 5 pm LiChrosorb Si 60
silica (Merck, Darmstadt, F.R.G.); (ii) 10 pm aminopropyl-bonded silica (Dr. Her-
bert Knauer, Bad Homburg, F.R.G.).

Chemicals and reagents

Solvents. Carbon dioxide was of technical grade (Air Liquide, France), meth-
anol of HPLC grade (Prolabo, Paris, France) and chloroform of HPLC grade (Rath-
burn Chemicals, Walkerburn, U.K.). Methylamine was used as a 40% aqueous so-
lution (Prolabo) and ethylamine as a 70% aqueous solution (Merck); triethylamine
was of Zur Synthese grade (Merck).
Solutes. Morphine, codeine, ethylmorphine, thebaine and poppy straw con-
centrate were obtained from Francopia (Paris, France), cryptopine, narcotine and
papaverine from Sanofi, Chemical Division (France) and 2,6-dinitrophenol from Pro-
labo.
SUB- AND SUPERCRITICAL FLUID CHROMATOGRAPHY OF OPIUM ALKALOIDS 353

RESULTS AND DISCUSSION

Influence of polar modifiers

Opium alkaloids and related molecules are polar compounds that are highly
retained on bare silica or aminopropylsilica with pure carbon dioxide as the mobile
phase, whatever its density. Hence the use of a polar modifier is necessary. A con-
clusion from our experiments was that a complex mixture of methanol, an aliphatic
amine and water gave the best results with a sufficiently high resolution in a given
analysis time.
The influence of polar modifiers in carbon dioxide SFC on both eluting power
and kinetic parameters has been discussed in various papers’ 2P1‘. Here, we shall exam-
ine the effect on the retention and selectivity of the various polar modifiers we in-
vestigated, namely methanol, amines and water.

Injuence of methanol
Tables I and II give the capacity factors and the selectivities, respectively, ob-
tained for six opium alkaloids and related compounds on an aminopropyl-bonded
silica, the percentage of methanol in the eluent varying from 5 to 25% (w/w). With
such methanol contents and under the temperature and pressure conditions that we
used (40.7”C, 220 bar), we must conclude that the mobile phase was in a subcritical
rather than a supercritical state18.
We worked with two complementary operating modes. On aminopropyl-bond-
ed silica and in order to achieve short separation times, we used about 16% (w/w)
of methanol in carbon dioxide. The retention time of morphine (the last compound
eluted) was nearly 5 min, but narcotine and papaverine were coeluted. On bare silica,
the use of the same methanol content gave a longer analysis (15 min), but also a
partial separation of narcotine and papaverine. However, it appeared that pure meth-
anol was not sufficient to give good separations of alkaloids.

TABLE I

CAPACITY FACTORS OF OPIUM ALKALOIDS ON AN AMINOPROPYL-BONDED SILICA

WITH VARIOUS METHANOL CONTENTS IN CARBON DIOXIDE

Column, 23 x 0.46 cm I.D.; stationary phase, IO-pm aminopropyl-bonded silica (LiChrosorb-NHz);

flow-rate, 4 ml min-‘; mean pressure, 220 bar; temperature, 40.7”C; detection, UV at 280 nm.

Solute Methanol in CO2 (%, w/w)

5 10 17 25

Narcotine 1.5 0.75 0.50 0.30

Papaverine 1.9 0.80 0.50 0.30
Thebaine 5.6 2.8 1.3 0.75
Ethylmorphine 6.0 3.3 1.7 1.0
Codeine 8.9 4.0 1.9 1.2
Morphine 63 17 7.0 3.1
354 J. L. JANICOT, M. CAUDE, R. ROSSET

TABLE II

SELECTIVITY VALUES OBTAINED ON AN AMINOPROPYL-BONDED SILICA WITH VAR-

IOUS METHANOL CONTENTS IN CARBON DIOXIDE

Operating conditions as in Table I.

Solutes pair Methanol in CO2 (%, w/w)

5 10 17 25

Thebaine-ethylmorphme 1.07 1.18 1.31 1.34

Narcotine-papaverine 1.27 1.07 1.00 1.00
Codeine-morphine 7.08 4.25 3.68 2.58

Injhence of amines
Influence on capacity factors and selectivities. Methyl-, ethyl- and triethylamine
were studied with bare silica and aminopropyl silica stationary phases.
On silica the addition of methyl- or ethylamine reduced the capacity factors
of all alkaloids (Fig. 1). The two amines showed very similar behaviour. The decrease
was monotonous up to a molar fraction of amine (xA) in the mobile phase of 8 .
10P3, except for narcotine and papaverine, whose k’ values increased above XA =
4. 10-3.
A similar phenomenon was observed with aminopropylsilica (Fig. 2), but the
capacity factors of morphine, cryptopine, codeine, ethylmorphine and thebaine
showed a minimum when methyl- or ethylamine was added to carbon dioxide. On
the other hand, k’ increased monotonously for narcotine.

Fig. 1. Variation of capacity factors of opium alkaloids on bare silica versus molar fraction (x) of amine
in the mobile phase. C’olumn, 23 x 0.46 cm I.D.; stationary phase, 5-pm LiChrosorb Si 60 silica; mobile
phase, carbon dioxide--methanol&water (83: 16.25:0.75, w/w); added amine, (a) methylamine (MeNH& (b)
ethylamine (EtNH& flow-rate, 8 ml min-‘; mean pressure, 220 bar; temperature, 40.7”C; detection, UV
at 280 nm. Solutes: C, narcotine; & papaverine; 0, thebaine; A, ethylmorphine; n , codeine; 0, cryp-
topine; 0, morphine.
SUB- AND SUPERCRITICAL FLUID CHROMATOGRAPHY OF OPIUM ALKALOIDS 355

0.3
1
I , , .
01234561
wwp’

Fig. 2. Capacity factors of opium alkaloids versus molar fraction (x) of amine in the mobile phase on
aminopropylsilica. Column, 23 x 0.46 cm I.D.; stationary phase, lo-pm LiChrosorb-NHz; mobile phase,
carbon dioxide-methanol-water (83:16.25:0.75, w/w); added amine, (a) methylamine, (b) ethylamine, (c)
triethylamine; flow-rate, 4 ml mini; other conditions as in Fig. 1. Solutes: A, narcotine; 0, thebaine;
W, ethylmorphine; 0, codeine; A, cryptopine; 0, morphine.

With regard to selectivity, on silica the addition of amines led to a decrease in

the lower values of a, i.e., for narcotine-papaverine and codeine<ryptopine pairs
(Fig. 3). On aminopropylsilica the amines did not change significantly the selectivities
obtained with narcotine-papaverine, thebaineeethylmorphine and codeinecrypto-
pine pairs, but triethylamine gave the best values for this last pair (Fig. 4).
Zq%~~ce on hold-up time. Hold-up time values (to) were measured with the
negative peak produced by injection of hexane with detection at 280 nm. 1-Decene,
detected at 205 nm, gave identical results.

a:

2.5
i

0 2 4 6 8
XMeNH,.1a3 XEtNH2.10’

Fig. 3. Influence of molar fraction (x) of amine in the mobile phase on the selectivity obtained on silica
for opium alkaloids. Conditions as in Fig. 1; added amine, (a) methylamine, (b) ethylamine. Solutes pairs:
0, narcotine-papaverine; W, codeine-cryptopine; A, thebaineeethylmorphine; 0, thebaine-codeine;
0, codeine-morphine.
356 J. L. JANICCJT, M. CAUDE, R. ROSSET

Fig. 4. Influence of molar fraction (x) of amine in the mobile phase on the selectivity measured on ami-
nopropylsilica for opium alkaloids. Conditions as in Fig. 2; added amine, (a) methylamine, (b) ethylamine,
(c) triethylamine. Solutes pairs: 0, thebaine-ethylmorphine; W, codeinexryptopine; 0, thebainexodeine;
A, narcotine-thebaine; 0, codeine-morphine.

Fig. 5. Influence of molar fraction (x) of amine in the mobile phase on the hold-up time. (a) Stationary
phase, 5-pm LiChrosorb Si 60 silica, other conditions as in Fig. 1; (b) stationary phase, lo-pm aminopro-
pylsilica LiChrosorb-NHz, other conditions as in Fig. 2. Added amine: l (-), methylamine; A
(---------), ethylamine; W (- -), triethylamine.
SUB- AND SUPERCRITICAL FLUID CHROMATOGRAPHY OF OPIUM ALKALOIDS 357

An increase in amine concentration in the eluent led to a decrease in to due to

impregnation of the pore volume of the support by the amine. Fig. 5 shows results
obtained with the two stationary phases using the same reduced velocity of the eluent.
With both supports, methyl- and ethylamine had similar effects up to XA = 6.5 .
10d3; the dead time was reduced by 16% compared with a mobile phase without any
amine. Triethylamine gave a smaller reduction as the impregnation of the stationary
phase was less efficient, probably for steric reasons.
Influence on the retention time. Fig. 6 shows that, on aminopropylsilica, an
increase in the amine content in the eluent can induce an increase in the retention
times of alkaloids. This effect was smaller if methylamine was replaced by ethylamine
and disappeared when triethylamine was used.

Influence of water
Without amine in the mobile phase. On bare silica, the addition of water led to
a decrease in the retention of alkaloids due to adsorption of water on silanol groups;
for thebaine, k’ was 45 with 0.71% (w/w) and 36 with 1.42% of water in carbon
dioxide.
Using aminopropylsilica, the proportion of water in carbon dioxide was in-
creased from 500 to 7100 ppm (0.05 to 0.71%, w/w), maintaining the methanol con-
tent close to 16.3%. Surprisingly, the retention of alkaloids was hardly altered (Fig.
7a), showing that residual silanol groups, solvated by methanol or water, were not
easily accessible to alkaloids.
In the presence of methylamine. On aminopropylsilica, with methylamine in the
eluent, water enhanced the retention of solutes (Fig. 7b). According to the results

t ,(min)

10
1

981

I.-_ ,
0 ‘i 2 3 --“‘-i
4 5 6 8
x&lo3

Fig. 6. Comparison of the effect of amines (molar fraction xa) on the retention times of codeine and
morphine on aminopropylsilica. Conditions as in Fig. 2. Added amine; 0, 0, methylamine; A, A,
ethylamine; 0, n , triethylamine. Solutes: 0, A, 0, codeine; 0, A, n , morphine.
358 J. L. JANICOT, M. CAUDE, R. ROSSET

10

3-

l-

OS
-1

0.00
‘i
----r-1--
0.20 0.40 0.60 0.86 1.00
i ,
0.00 0.20 0.40 0.60 0.80 1.00
.

%H~Oh/w) %H,O(W/W)

Fig. 7. Variation of capacity factors of opium alkaloids with water content in the mobile phase. Column,
23 x 0.46 cm I.D.; stationary phase, aminopropylsilica, lo-pm LiChrosorb-NHz; mobile phase, (a) carbon
dioxide-methanol (83 75:16.25, w/w), (b) carbon dioxide-methanol-methylamine (83.45:16.25:0.30, w/w);
other conditions as in Fig. 2. Solutes: A, narcotine; 0, thebaine; 0, ethylmorphine; A, codeine; 0,
morphine.

0 a 0
3

2.5

1.5

1 *
0.2 0.4 0.6 0.8
%HzO(w/wl

Fig. 8. Influence of percentage of water in the mobile phase on (a) capacity factors and (b) selectivities.
Mobile phase, carbon dioxide-methanol-methylamine (83.60:16.25:0.15, w/w); other conditions as in Fig.
1. Solutes: (a) 0, narcotine; a, papaverine; 0, thebaine; A, ethylmorphine; W, codeine; 0, cryptopine;
0, morphine; (b) 0, narcotine-papaverine; A, thebaine-thylmorphine; 0, codeine-cryptopine; Cl,
thebainesodeine; 0. codeine-morphine.
SUB- AND SUPERCRITICAL FLUID CHROMATOGRAPHY OF OPIUM ALKALOIDS 359

obtained without any amine in the mobile phase, it is possible to exclude the role of
residual silanol groups; water does enhance alkaloid-amine-aminopropyl chain in-
teractions.
On bare silica, with an amine in the eluent, it appeared that the percentage of
water must be as low as possible; on the one hand, the capacity factor of morphine
(and the analysis time) increased with increasing water content and, on the other
hand, water led to a decrease in selectivity for the codeine-cryptopine pair (Fig. 8).

Choice of operating conditions for alkaloids separations

From the preceding results, it was possible to choose various operating con-
ditions according to the type of alkaloid mixtures to be separated.

On aminopropylsilica
The narcotine-papaverine pair was not well separated on this stationary phase
and it was therefore reserved for rapid separations of narcotine, thebaine, ethylmor-
phine, codeine, cryptopine and morphine.
As the selectivity values obtained for the codeineecryptopine pair were not
strongly influenced by the amount of water in the mobile phase, a low percentage of
water (0.30%, w/w) was chosen in order to obtain a rapid analysis. Triethylamine
was used as it gave the best selectivity for the codeinesryptopine pair (R, = 2.70
for an amine molar fraction of 2 . 10e3).

c 0

l I w
0 2 4 6 8 0 1 2
Time(min) Time(min)

Fig. 9. Separation of opium alkaloids on aminopropylsilica. (a) Mobile phase, carbon dioxide-
methanol-triethylaminewater (82.95: 16.25:0.50:0.30, w/w); other conditions as in Fig. 2. (b) Column, 12
x 0.4 cm I.D.; stationary phase, 3-pm aminopropylsilica Spherisorb-NHZ; mobile phase, carbon
dioxide-methanol-triethylamine-water (87.62: 11.80:0.36:0.22, w/w); flow-rate, 8 ml min- ‘; other condi-
tions as in Fig. 2. Solutes: 1, narcotine; 2, thebaine; 3, codeine: 4, cryptopine; 5, morphine.
 J. L. JANICOT, M. CAUDE, R. ROSSET

‘ 4 I l
0 2 4 6 a 10
Time @in)

Fig. 10. Separations of opium alkaloids on bare silica. Column, 23 x 0.46 cm I.D.; stationary phase,
S-pm LiChrosorb Si 60 silica; mobile phase, carbon dioxide-methanol-methylamine-water (83.37:
16.25:0.15:0.23, w/w); other conditions as in Fig. I. Solutes: 1, narcotine; 2, papaverine; 3, thebaine; 4,
codeine; 5, cryptopine, 6, morphine.

o’i
I 4 6 8 111
.--

Tim ?(mIn)

Fig. 11. Chromatogram of a poppy straw concentrate. Operating conditions as in Fig. 10 except for
detection, UV at 220 nm; injection solvent, methanol. Solutes: 1, papaverine; 2, thebaine; 3, codeine; 4,
morphine.
 SUB- AND SUPERCRITICAL FLUID CHROMATOGRAPHY OF OPIUM ALKALOIDS 361

The chromatograms obtained are shown in Fig. 9a (particle diameter 10 pm)

and 9b (particle diameter 3 pm).

On bare silica
This stationary phase was preferred when the narcotine-papaverine pair was
to be separated. A 0.15% (w/w) concentration of methylamine was added to the
mobile phase; this did not favour rapid analysis but it provided a partial resolution
of narcotineepapaverine (R, z 1.0). To preserve the selectivity obtained for the
codeineecryptopine pair, we used as low as possible a percentage of water with the
40% aqueous methylamine solution used, viz., 0.23% (w/w) (Fig. 10).
Application to a poppy straw concentrate

The procedure described for silica was applied to a poppy straw extract. The
operating conditions were chosen in order to resolve partially narcotine from pa-
paverine. An example of a chromatogram is shown in Fig. 11, where four alkaloids
(papaverine, thebaine, codeine and morphine) are revealed. The minor peaks did not
correspond to the other alkaloids of interest that we studied (narcotine, ethylmor-
phine and cryptopine). When compared with those obtained on poppy straw by
Pettitt and DamonlO, this chromatogram shows the same major peaks in a shorter
analysis time (11 min against 25 min).
Using a Varian Polychrom 9060 diode-array spectrophotometer we then stud-
ied the peak purity. Purity was estimated from the “purity parameter” delivered by
the spectrometer. This parameter, P, is calculated for a wavelength range chosen by
the operator according to absorption bands characteristic of a given solute. The peak
purity can be calculated by comparison of P values on the peak upslope, apex and
downslope. With this technique, for example, 5% cryptopine can be detected in co-
deine when these two alkaloids are coeluted. Of course, the closer the spectra of the

TABLE III

PURITY PARAMETER VALUES OBTAINED FOR A POPPY STRAW CONCENTRATE AND A SYN-
THETIC MIXTURE OF ALKALOIDS

Solute Wavelength range Position Synthetic mixture Extract

for P in peak*
calculations (nm)

Papaverine 210-254 U 230.22 230.16 229.94 230.22 230.12 230.01

A 230.49 230.65 230.35 230.46 230.45 230.57
D 230.12 230.35 230.46 230.34 230.46 230.46
Amp 0.37 0.49 0.41 0.24 0.34 0.56
Thebaine 220-3 11 U 243.36 243.14 242.21 241.92
A 243.02 242.78 241.19 241.89
D 242.99 242.71 241.83 241.62
Amp 0.37 0.43 0.38 0.30
Codeine 220-292 U 228.01 227.17 225.00 225.10
A 227.38 226.95 225.29 225.41
D 227.47 227.04 225.10 225.46
Amp 0.54 0.22 0.29 0.36

* Positionin peak: U = upslope; A = apex; D = downslope; Amp = amplitude of variations.

362 J. L. JANICOT, M. CAUDE, R. ROSSET

APz(P-P,)
A

0.40-
0
0
0.30-

I-----
0
0.20 .
.

i
0.10 0 0
0
0

l
.
0 0 ’ n )

-0.20 - 0.10 ‘“. t-tR(min)

O.iO

. .
-0.10 , 0 o

I .

1
-0.20

-0.30

Fig. 12. Variations of AP = P- Ps for two chromatographic peaks of pure morphine (purity parameter
and retention time plotted after subtraction of their values at the peak apex). Wavelength range used for
P calculations, 258-297 nm. P values at peak apex: 0, first injection, Papel = 278.57; 0, second injection,
P apex = 278.59.

AP=(P-P,)

0.40

0.30

0.20

0
0.19

0 o-o.1o 0
-A--o-. 0.10 0.20 0 0 .

. A t-tR(min)
.
0 x 0 0
- 0.10
00
.
-0.20 0
.
.
.
-030

- 0.40

1
Fig. 13. Variations of AP = P-Ps for two chromatographic peaks of morphine in a poppy straw con-
centrate. Wavelength range for P calculations as in Fig. 12. P values at peak apex: 0, 278.64; A, 278.98.
SUB- AND SUPERCRITICAL FLUID CHROMATOGRAPHY OF OPIUM ALKALOIDS 363

;
Y

-f
0 2 4 6 8 10 12
Time(min)

Fig. 14. Chromatogram of a poppy straw concentrate containing 2,6-dinitrophenol as an internal stan-
dard. Conditions as in Fig. 11 except for the injectionsolvent (methanol+zhloroform, 50:50,v/v). Solutes:
1, chloroform; 2, papaverine; 3, 2,6_dinitrophenol; 4, thebaine; 5, codeine; 6, morphine.

two compounds the higher is the detection limit of one in the other. As we did not
know possible impurities contained in the poppy extract, large wavelength ranges
were used in order to detect any possible impurities (Table III).
Table III compares the results obtained for the concentrate with those obtained
with a synthetic mixture of pure papaverine, thebaine and codeine. As the variations
of P for the extract were similar to those of the synthetic mixture, we considered the
peak purity to be sufficient to permit a quantitative analysis.
With regard to morphine, the chromatographic peaks were large enough to
calculate P more than three times. As shown in Figs. 12 and 13, we plottted P- Ps
versus t - tR, where Ps is the P value at the peak apex, t the retention time where the
purity parameter is equal to P and tR the retention time at the peak apex. Fig. 12
relates to pure morphine and Fig. 13 to poppy straw concentrate and, as the P
variations were comparable in the two instances, the peak of morphine was con-
sidered to be pure.

TABLE IV

RESPONSE FACTORS AND CORRELATION COEFFICIENTS OBTAINED THROUGH CALI-

BRATION BY LINEAR REGRESSION

Solute Responsefactor x IO3 (Img-‘) Correlation coejjicient

Papaverine 6.14 0.9941

Thebaine 5.63 0.9946
Codeine 3.43 0.9986
Morphine 3.56 0.9971
364 J. L. JANICOT, M. CAUDE, R. ROSSET

Quantitative analysis
As papaverine was a minor component of the poppy straw extract, detection
was effected at 220 nm, a compromise between an acceptable baseline stability and
a maximum response for the other three alkaloids. 2,6-Dinitrophenol was chosen as
an internal standard.
The best solvent for dissolution of the extract was methanolchloroform
(50:50, v/v). Chloroform generated an unretained peak but did not perturb mea-
surements of alkaloid concentrations (Fig. 14).
Table IV shows response factors and correlation coefficients obtained with five
standard solutions. Finally, relative standard deviations obtained for nineteen anal-
yses were 2.9% for thebaine, 9.0% for codeine and 8.6% for morphine.

CONCLUSIONS

The results demonstrate that sub- and supercritical fluid chromatography using
carbon dioxide-amine polar modifiers as the mobile phase is an effective technique
for opium alkaloid separations.
Narcotine, thebaine, codeine, cryptopine and morphine can be separated in
less than 2.5 min on aminopropylsilica. Bare silica gives a higher resolution in an
1 I-min analysis and has been successfully applied to the determination of three al-
kaloids in a poppy straw concentrate. Even though it is necessary to increase the
reproducibility further, SFC appears to be a promising technique for the routine
analysis of opium alkaloids. In comparison with HPLC on poppy strawlo, our results
show the same major peaks but the analysis time is halved without the drawbacks
of gradients such as column regeneration.

ACKNOWLEDGEMENTS

This study was supported by the Societe Nationale Elf-Aquitaine, Sanofi-Elf-

Bio-Recherches, Labege and Sanofi-Chimie, Aramon.

REFERENCES

1 E. Clement, I. Onuska, J. Yang, A. Eiceman and H. Hill, Jr., Anal. Chem., 58 (1986) 321R.
2 C. M. White and R. K. Houck, J. High Resolut. Chromatogr. Chromatogr. Commun., 9 (1986) 4.
3 P. Mourier, E. Eliot, M. Caude, A. Tambute and R. Rosset, Anal. Chem., 57 (1985) 2819.
4 P. Mourier, M. Caude and R. Rosset, Anaiusis, 13 (1985) 299.
5 P. Mourier, P. Sassiat, M. Caude and R. Rosset, Analusis, 12 (1984) 229.
6 J. C. Fjelsted and M. L. Lee, Anal. Chem., 56 (1984) 619A.
7 N. R. Ayyanger and S. R. Bhide, J. Chromatogr., 366 (1986) 435.
8 H. A. H. Billiet, R. Wolters, L. de Galan and H. Huizer, J. Chromatogr., 368 (1986) 351.
9 W. R. Sisco, C. T Rittenhouse and W. M. Maggio, Chromatographia, 20 (1985) 280.
10 B. C. Pettitt, Jr., and C. E. Damon, J. Chromatogr., 242 (1982) 189.
11 J. L. Janicot, M. Caude and R. Rosset, Analusis, 14 (1986) 441.
12 P. Mourier, P. Sassiat, M. Caude and R. Rosset, J. Chromatogr., 353 (1986) 61.
13 J. M. Levy and W. M. Ritchey, J. Chromafogr. Sci., 24 (1986) 242.
14 A. L. Blilie and R. Greibrokk, Ana/. Chem., 57 (1985) 2239.
15 B. W. Wright and R. D. Smith, J. Chromatogr., 355 (1986) 367.
16 C. R. Yonker and R. D. Smith, J. Chromatogr., 361 (1986) 25.
17 P. Sassiat, P. Mourier, M. Caude and R. Rosset, Anal. Chem., 59 (1987) 1164.
18 J. B. Crowther and J. D. Henion, Anal. Chem., 57 (1985) 2711.