Current research note Fresenius J Anal Chem (1994) 350:642-644 - © Springer-Verlag 1994

First analytical results using a multi-ion counting system of a spark source mass spectrometer

Max-Planck-Institut ffir Chemic, Postfach 3060, D-55020 Mainz, Germany

switch to other mass ranges. To realize first experiments, we have tested the system with 5 channeltrons. Each channeltron is connected to its individual high voltage power source (voltage used = 2100 V), a separate preamplifier (deadtime = 25 ns, input threshold = 20mV) and a counter (Philips). They are mounted at an angle of deflection of 45 ° with respect to the ion optical axis in order to increase the compactness of the channeltrons at a given mass dispersion and to incorporate a conversion plate in front of each channeltron.

Received: 10 May 1994/Revised: 5 August 1994/ Accepted: 5 August 1994

Samples and analytical techniques

K.P. Jochum, H.-J. Lane* H.M. Seufert, A.W. Hofmann

Abstract. Isotope dilution measurements of geological reference materials are described using a newly developed multi-ion counting (MIC) system for a spark source mass spectrometer. Compared with the conventional photoplate detection system, there are considerably improved analytical features of the MIC system, especially the high sensitivity, which leads to short measuring times of I rain to 1 h for trace element analysis on the gg/g - ng/g level.

Introduction Spark source mass spectrometry (SSMS) is a powerful method for multi-element analysis of geological samples [1,2]. The main advantages are simple sample preparation (with no chemical treatment of the samples), simultaneous determination of 3 0 - 4 0 elements and low detection limits ( 1 - 1 0 ng/g). However, there are some disadvantages with photoplates which are the standard detection system, such as long measuring times (exposure and densitometer times of 1 4 - 2 0 h for a 3 photoplate analysis), complex evaluation techniques and limited precision of 3 - 1 5 % . The appearance of small channeltrons opened a new possibility to compensate the drawback of the photoplate detection (PhD) system and to detect many ions simultaneously along the straight focusing plane of the Mattauch-Herzog instrument. Laue et al. [3] have therefore equipped a commercial spark source mass spectrometer with a multi-ion counting (MIC) system designed to consist of 25 separate channeltrons for ion counting measurements. In this paper first analytical data are presented using the MIC system and the results are compared with those obtained by the conventional PhD system.

Multi-ion counting system (MIC) A full description of the MIC system will be published elsewhere [4]: A short account follows: The MIC system consists of a 8 cm long box which can be equipped with up to 25 channeltrons each having a width of 1.7 mm (Dr. Sjuts Optotechnik). The channeltron box is situated at the high mass end of the image plane of the mass spectrometer. This set-up allows the simultaneous determination of up to 25 isotopes (from 171yb- 238U). However, changing the magnetic field enables to

* Present address: Spectromat GmbH, Blankenburger Strasse

17, D-28205 Bremen, Germany Correspondence to: K.P. Jochum

Geological standard reference materials from the United States Geological Survey (USGS) and the Geological Survey of Japan (GSJ) were used for this work. For a precise and accurate determination of the element content, a multi-element isotope dilution technique [2] has been applied. About 5 0 - 1 0 0 mg of the rock powder was mixed with a spiked graphite containing 235U and other spike isotopes and then briquetted to rod-shaped electrodes. Ions were produced in a spark plasma (voltage = 25 kV, pulse repetition rate = 300 Hz, pulse length = 100 gs) and analyzed in an AEI-MS 702R mass spectrometer. The magnetic field strength was set at suitable values to measure simultaneously 23~U and 238U isotopes with the MIC system. The total ion charge determined by a monitor before mass separation was used as a measure of the ions entering the magnetic field. To compare the analytical results with the conventional PhD system, the same electrodes were also analyzed using Ilford Qplates for ion detection.

Results and discussion Table I shows our results for different reference samples. The data represent the mean of 1 0 - 2 0 measurements using total ion charges of 3 - 1 0 n C (corresponding to measuring times of about 1 0 - 3 0 s). The measured values for U and Th isotopes are overlapped by a background (about 1 - 10 counts/nC) which is mainly caused by charge exchange between ions and neutral molecules. Recently, the background has been reduced to a level of about 0.1 counts/nC by improving the instrumental vacuum with turbomolecular and ion pumps. Reproducibility of count measurements is relatively poor (10-25°7o) because the position of the electrodes in the ion source (possibly causing different fractions of the total ion beam entering the magnetic field) influences the measurements. However, the ion optics influences the measurements of different isotopes in the same way and therefore ratios of different isotopes are nearly uniform. The 238 235 U/ U ratios and U concentrations of Table 1 were calculated by using background corrected values. They are compared with the results using the conventional PhD system. In the following, important analytical features of the new MIC technique are discussed: Precision. The precision mainly depends on ion statistics. How-

ever, because of the very high sensitivity of the MIC system (only 6 - 2 0 ~tg sample is consumed for each measurement), sample heterogeneities in the electrodes may also influence the reproducibility. Nevertheless, we achieved a precision of 1-3o70 for most coarse grained (e.g. granite NIM-G) and fine grained (e.g. basalt BCR-1) samples having U concentrations of about 1 - 10 Ixg/g. The very low U concentration of 15 ng/g in JP-1 could be determined with a precision of about 10°70. Compared

643 Table 1. Uranium measurements in geological reference materials using the mult-ion counting (MIC) system. Samples are spiked with 235U. Most measured isotope ratios and isotope dilution values agree within 1 (lm error with data obtained by the conventional photoplate detection (PhD) system. Concentrations are also compared with reference values

Sample (rock type)

23sU counts (total ion charge used for one measurement)

U [gg/g]

238U/235U

MIC

PhD

Reference values

MIC

PhD

BCR-1 (basalt)

1200

(3nC)

2.55_+0.04

2.44_+0.07

1.75 _+0.02

1.67 _+0.05

1.75 [5]

BE-N (basalt)

800

(3nC)

2.76_+0.06

2.81+0.10

2.70 _+0.07

2.75 -+0.13

2.4 [5]

STM-1 (syenite)

800 a (3nC)

2.88_+0.10

2.99-+0.12

9.20 _+0.30

9.56 _+0.40

9.06 [5]

NIM-G (granite)

8000

(3nC)

12.3 _+0.2

13.9 -+0.6

JR-I (rhyolite)

3500

(3nC)

16.3 _+0.5

14.1 _+1.8

JP-1 (peridotite)

72 (10nC)

0.94_+_0.11

0.93_+0.28

14.1

-+0.3

16.2

_+0.9

15 [5]

8.10 _+0.28

8.0

+1.0

9 [5]

0.015_+0.002

0.013_+0.005

0.013 [7], 0.011 [8], 0.05 [5]

a Sample was diluted with 75°70 pure quartz 1 0 - 2 0 repeated measurements and reproducibilities of better than 3°70, it is between 1 min and 1 h for trace element analysis on the ~tg/g - ng/g level. Compared to the conventional PhD system this means an improvement by more than a factor of 20.

to the PhD system, the precision is by a factor of 2 - 3 better. This is especially obvious for samples which are underspiked 238 235 (i.e. having ratios of U / U of _> 10) because the dynamic 4 5 range of the M1C system (about 10 - 10 ) is much larger than that of the PhD system (about 50).

Accuracy. The U concentration in BCR-1 (which is probably the best-known geochemical reference sample in the world) is identical with the compiled value [5]. JP-1 has a very low U content; however, the value agrees with recently published data by modern sensitive analytical methods, such as ICP-MS [71 or HPLC [8]. Accuracy of the new MIC-SSMS technique is estimated to be similar to the attainable precision, i.e. better than about 3% for concentrations > 1 ~tg/g.

Sample amount. Because of the high sensitivity, the sample consumption is very low. It is about 6 - 2 0 ~tg for each measurement using total ion charges of 3 - 10 nC. As mentioned, this can be some disadvantage for the analysis of coarse grained rocks, such as granitoids, where careful homogenization of the samples is therfore necessary. On the other hand, the MIC system permits the analysis of small (typically 1 mg) samples. Therefore sensitive multi-element analysis of mineral grains may be an important application.

Measuring time. One of the main advantages of the MIC system is the very short measuring time which is only dependent on the element concentration and the theoretical precision. Assuming

Detection limits. The detection limits (3~ background noise) for Th and U are determined using the depleted sample JP-1. Their values of about 4 ng/g for measuring times of 100 s are similar

12Cll 132Ba

2500

Fig. 1. Mass spectra of Cs, Ba and the rare earth elements of the spiked standard rock sample JR-2 (spike isotopes: 136Ba, ~43Nd, 147Sm, 151Eu, 161Dy, ~71yb). Each data point represents a measurement of 1 nC total ion charge. A mass resolution M / A M of 600 is obtained by using an 0.4 mm wide exit slit. The inset shows a highly resolved mass spectrum ( M / A M = 3000) at mass number 132 for the BCR-1 standard; the low abundant ~32Ba isotope is separated from a carbon cluster

2000

1500

10oo

500

0 Mass number

Lit_ o....*....o.., 130 Cs

Ba

e e O e

140

150

ee

O

me

•

eOI

o

e e O e e o e Q I o e

160

Rare earth elements (La - Lu)

170

eO

644 to those obtained by the PhD system. However, we expect a factor of 10 improvement of the detection limits by the recently obtained low background using the new vacuum pumps. Further improvements can be achieved by increasing the measuring times ( > 100 s) and/or by applying element enrichment techniques, such as the spark source mass spectrometric tiptop-technique [6].

Mass resolution. The PhD system is characterized by a high mass resolution M / A M of about 5000. Mass resolution of the MIC system will be presumably lower and dependent on the widths of the exit slits. The first results of Table 1 were obtained without the use of exit slits. Therefore the mass resolution was only about 200, which is however sufficient to resolve the spectral lines of the isotopes of interest, i.e. 232, 235 and 238. To improve the mass resolution and to separate mass lines close to each other, we recently performed experiments with different exit slits (widths < 0.5 ram). As an example, Fig. 1 shows the highly resolved mass spectra of Cs, Ba and the rare earth elements ( M / A M = 6 0 0 - 3000).

Acknowledgement. We thank C. Dienemann and J. Diefenbach for technical assistance with the analyses.

References

1. Taylor SR, Gorton MP (1977) Geochim Cosmochim Acta 41:1375-1380 2. Jochum KP, Seufert HM, Midinet-Best S, Rettmann E, SchOnberger K, Zimmer M (1988) Fresenius Z Anal Chem 331:104-110 3. Laue H-J, Seufert HM, Jochum KP (1994) Verh D P G 4, 523 (abstract) 4. Laue H-J, Seufert HM, Jochum KP (1994) In preparation 5. Govindaraju (1989) Geostandards Newsletter 13:1-113 6. Rocholl A, Jochum KP, Seufert HM, Midinet-Best S (1988) Fresenius Z Anal Chem 331:140-144 7. Ionov DA, Savoyant L, Dupuy C (1992) Geostandards Newsletter 16:311 - 315 8. Rehk~imper M (1994) Chem Geol (in press)

Short communications Fresenius J Anal Chem (1994) 350:644-646 - © Springer-Verlag 1994

NMR spectroscopic determination of enantiomers via their tripheny|tin derivatives: A new application of chiral solvating agents J. Klein, R. Borsdorf

Department of Chemistry, University of Leipzig, Linn6strasse 3, D-04103 Leipzig, Germany Received: 21 June 1994/Revised: 5 August 1994/ Accepted: 10 August 1994 Abstract. A method is described for the NMR determination of the enantiomeric composition of chiral carboxylic acids and alcohols using their triphenyltin derivatives together with the chiral solvating agents quinine hydrochloride and phenylethylamine.

Introduction

As well as chromatographic methods NMR investigations of enantiomers require the presence of a chiral auxiliary that converts the enantiomeric mixture into diastereomers. While chiral derivatising agents (CDA) force the formation of diastereo-

Correspondence to: R. Borsdorf

meric compounds prior to the NMR analysis, chiral solvating agents (CSA) can be used directly and no isolation of the diastereoisomeric solvation complexes is necessary. The peak areas of the separated resonances permit a direct calculation of the enantiomeric composition. Unfortunately, the application of the latter method is limited to compounds strongly interacting with the CSA. Chemical shift anisochrony is caused by the relative size of the diastereomeric complexation constants of the enantiomers and by magnetically anisotropic groups (e.g. phenyl groups). However, signal splitting is often rather small or not observable at all [l]. Compared with a large number of methods for NMR analysis of chiral alcohols, only relatively few CSA are known for investigations of chiral carboxylic acids [2-41. Therefore we present here a method that enables us to determine the enantiomeric composition of chiral compounds forming only weak solvation complexes with the CSA. Results and discussion

Starting from standard methods, many triorganotin derivatives of compounds containing acid hydrogen atoms, are available by azeotropic dehydration with bis-(triphenyltin)-oxide using toluene as solvent [5, 6]. R-X-H+(Ph3Sn)20 X= -C(O)O-

,

~ 2Ph3Sn-X-R+H20

-0-

This reaction converts a functional group, that is only capable of interacting weakly with the CSA, into a strong coordinating group. Triorganotin derivatives act as Lewis acids due to the coordinatively unsaturated tin. For this reason they form