____
~~~
On the Determination of Boron Mixed with Polyphenyls SIR: The most suitable known organic compounds used in the organic cooled heavy water moderated nuclear reactors are polyphenyls. Therefore, many attempts have been and will be made by physicists t o understand the dynamics of polyphenyl molecules, and also t o measure the neutron parameters of these compounds. One technique for measuring the diffusion parameters of thermal neutrons is to add boron, at various concentrations, t o the medium and t o measure the relaxation length of the thermal neutrons. Thus, the accuracy of the extracted results is partly dependent upon the accuracy of the boron concentrations in the mixture. Such a measurement was performed in Dowlherm-A (a eutectic mixture of 26.Sz diphenyl and 73.5% diphenyl oxide by weight) at various temperatures (I) (Table I). Trihexylane glycol biborate [tri-(2-methyl-2,4-pentanediol)biborate], commercially known as Borester-7 was used t o mix with Dowtherm-A at various concentrations. Unfortunately, the experimental arrangement did not allow the direct determination of boron concentrations by the addition of the known amount of Borester-7. The neutronic measurements were accurate t o about one-half per cent, but some difficulties exist in determining the boron concentrations to a similar accuracy. This is an example of a common analytical problem that is difficult t o solve. Chemical and physical methods of analysis have failed t o give reliable results, but infrared spectra measurement method was found t o be adequate. The Borester-7 in the mixture was hydrolyzed with 5 % sodium carbonate, the organic materials were removed by chloroform, and both the titration as the manitol complex and spectrophotometric estimation by carmine were performed.
(1) F. Sefidvash, Ph.D. Thesis, University of London, 1967.
The efficiency of the hydrolysis was demonstrated by finding the value of 5.66 f 0.01% for boron in pure Borester-7, compared to 5.64% quoted by its supplier. But in mixtures with Dowtherm-A, low and random results were obtained. The error was found t o increase with increasing quantities of Dowtherm-A, and extrapolation t o zero weight of Dowtherm-A produced inconsistent results. Systematic variation of sample size, increase of hydrolysis time from 20 t o 200 minutes, extraction by chloroform alone, chloroform after mechanical separation, and by other solvents failed t o reveal the precise cause of error. No significant quantities of boron could be found in the separated organic material. The hydrolysates were noticed t o develop a yellow hue after one or two days, indicating the incomplete removal of organic matter, but prolonged extraction and heating did not alleviate this situation. The hydrolysates were made strongly alkaline with sodium hydroxide, and were boiled under reflux up t o three hours. This produced a significant improvement, in that recoveries rose t o a level between 80 and 9Oz, but reproducibility was poor, extrapolation t o zero Dowtherm-A wejght became impossible, and the high salt concentrations gave rise to imprecise end points in titration. The differences in the refractive indices of Dowtherm-A and Borester-7 suggested the refractometry method of analysis. The refractive indices were measured at 22 "C at the sodium wavelength. Samples of Dowtherm-A and Borester-7 showed no change in refractive index after heating either 24 or 148 hours at 180 "C. The results are the mean value of five refractive index measurements using the value of 5.66x by weight of boron in Borester-7. The variation between refractive index and per cent boron for synthetic samples was linear. Infrared spectra measurement method is based upon the
Table I. Boron Content of Borester-7 Dowtherm-A Mixtures
Sample A 80" A 140 A 200 A 200 B 20 B 140 B 200 B 200
c C c
20 140 200
Refractive index, % B by weight 0.055
0.052 0.037
...
0.030
0.052 0.063 0.048 0.063 0.055
0.063
Infrared spectra, B by weight 0.0081 0.00815 0.00782 0.0085 0.0153 0.0155 0.0180 0.0162 0.0214 0.0209 0.0211
Sample E 20 E 80 E 200 F 20 F 200 G 20 G 200
Infrared spectra, % B by weight
Refractive index, B by weight 0.136 0.166 0.147
0.0870
0.0880 0.0860
0.232
0.155
0.276
...
0.226 0.223
H 20 H 200
0.369 0.342
0.292 0.289
20 200
0.652 0.588
0.524 0.514
I I
...
0.155
D 20 0.081 0.0295 D 200 0.085 0.0289 J 20 1.03 1.09 The letter indicates a particular mixture followed by a number giving the temperature ("C)of the medium at which the sample was taken. ~~
~~
VOL. 40, NO. 7, JUNE 1968
1165
measurement of absorption frequencies of the B-0-B bond which is present in the Boresters. Initial experiments, using the double beam technique, gave sharp absorptions at 1412 cm-1 and 1370 cm-', corresponding to the doublet associated with the B-0-B stretching frequency. For each sample a full spectrum was recorded, and then the portion of spectrum between 1600 cm-1 and 1200 cm-' was repeated four times, each time with a different aliquot of sample in the sample cell. The mean of the four values for absorbance was used. It seems unlikely that the chemical analysis of boron in these mixtures would be feasible without much further research. This view is supported by the notorious difficulty of precise boron determination in many materials. The results show the unreliability of the refractive index method. The infrared spectra measurement is the most suitable of all the presently tried methods. There still remains the need for finding relatively cheap
boron compounds of high boron to hydrogen ratios, and stable with a high boiling point, to mix with various polyphenyls in large quantities, and to develop a simple and reliable method of determining their boron content to an accuracy of better than 1%. This then will provide physicists with a tool to make careful and detailed studies of the neutronic properties of polyphenyls. FARHANG SEFIDVASH
Department of Technical Physics Technical University of Helsinki Otaniemi, Finland RECEIVED for review October 31, 1967. Accepted January 12, 1968. Work performed under the auspices of the Imperial College of Science and Technology (University of London).
Solvent and Temperature Effects in the Determination of Pyrrolizidine Alkaloids with 4-Dimethylaminobenzaldeh yde SIR: Recently Mattocks (1) described the application of the Polonovsky reaction (2) t o the development of a reliable colorimetric method for the determination of microgram quantities of unsaturated pyrrolizidine alkaloids. The preliminary steps to the final reaction with 4-dimethylaminobenzaldehyde (Ehrlich reagent) to form a colored product are conversion of the alkaloid base to the N-oxide with methanolic hydrogen peroxide and dehydrogenation of the formed N-oxide to a pyrrole by means of acetic anhydride. For these steps to be reproducible, all excess peroxide must be removed, alkaloid must not be decomposed in the process, and dehydrogenation with acetic anhydride must be moderated by a suitable solvent. Since isoamyl acetate appeared to meet the criteria of a suitable moderating solvent, namely nonreactivity with acetic anhydride and adequate polarity for solution of N-oxides, while lacking the disadvantages of peroxide formation, it has been investigated as an alternative t o diglyme. Reproducible results were obtained with all the alkaloids tested except lasiocarpine which invariably gave erratic results both within a series and from series to series. By omitting the oxidation step and studying lasiocarpine N-oxide alone, reproducible values were obtained. This finding showed that the moderating solvent was not involved and that the effect observed was specific to the oxidation stage. Therefore, this stage was examined to determine whether use of a temperature high enough to volatilize methanol but low enough t o restrain the vigor of oxidation might yield more consistent results. A temperature of 77 f 1 "C was selected after tests had shown that methanol was completely volatilized from the tube in 20 to 30 min at this level. During this study of the Mattocks procedure, it was observed that solutions prepared from heleurine and supinine formed strong colors in the cold following addition of Ehrlich reagent (1) A. R. Mattocks, ANAL.CHEM., 39, 443 (1967). (2) M. Polonovsky and M. Polonovsky, Bull. SOC.Chim. France, 41, 1190 (1927). 1166
ANALYTICAL CHEMISTRY
Table I. Molar Absorptivity (e) of Reaction Product between Treated Pyrrolizidine Alkaloids and 4-Dimethylaminobenzaldehyde in Two Different Solvents E
x
10-
Alkaloid Isoamyl acetate Diglymeb Heliotrine 73.8 =k 0.96(8). 75.2 Heliotrine N-oxide 80.4 i 1 . 9 (7) ... Lasiocarpined 80.0 =t 1 . 3 ( 5 ) 80.5 Lasiocarpine N-oxided 79.0 i 1.8 (6) ... Heliotridine 62.0 55.2 i 1.6 ( 5 ) Heliotridine trachelanthate 7 6 . 2 1 1.1 (5) ... 78.0 Supinine 74.2 12.6 (7) ... Heleurine 63.1 12.6 (5) a Average value with standard deviation. * Data of Mattocks ( I ) . The figure in parentheses is the number of determinations. Temperature at oxidation step, 77 "C.
in contrast to those from other pyrrolizidine alkaloids which required heating before appreciable color developed. Therefore, to determine whether this difference might be used to differentiate between alkaloids, the extent of color development at room temperature, compared with that at the elevated temperature, was studied. EXPERIMENTAL
Apparatus. A Beckman Model B spectrophotometer fitted with Corex cells of 1-cm path length was used for all absorbance measurements. The absorbance maximum was generally in the region of 557-560 mp for all alkaloids examined, tending to 557 mp for supinidine base alkaloids and 559 mp for heliotridine base alkaloids; 559 mp was routinely used. Reagents. All reagents with the exception of isoamyl acetate were prepared essentially by the details given by Mattocks (1).
