Sample Preparation

In filament pyrolysis techniques, ideally the sample is coated on the filament by applying one or more drops of a solution followed by drying in air, in an inert gas or in a vacuum chamber. Although readily soluble samples may be found among both synthetic and natural polymers, more often than not the sample is insoluble in most chemically inert solvents. Generally, insoluble samples can be handled by mechanical grinding or milling to form a fine powder (with admixture of quartz sand to aid homogenisation), followed by ultrasonic suspension in a suitable solvent. Among the refractory samples which have been successfully processed in this manner are cells, tissues, whole soils, coals and shale rocks. Sampling and homogenisation of sticky or elastic materials, e.g. certain microorganisms and natural rubber, can be aided by grinding with cooling in liquid nitrogen. Some samples, e.g. bacterial colonies, sludges or urine* can be applied directly to the filament. Suitable solvents or suspending media are methanol, carbon disulphide and water. Carbon disulphide is most easily dried and produces relatively stable suspensions by virtue of its high specific density. However, it is toxic and stored samples tend to dry out rapidly. Water, although often the most natural choice for biomaterials, dries very slowly and therefore requires vacuum drying if time is an important factor. Methanol has been the preferred solvent or suspending medium in most of our recent studies.

Further important factors in sample preparation are the pH and ionic strength of the sample solution or suspension, as these factors may strongly influence the pyrolytic fragmentation of (especially polar) organic compounds. This is demonstrated in Figure 11, where pyrolysis mass spectra of polygalacturonic acid sampled from different suspension or solution media are compared. Such "matrix effects" may lead to artifacts on comparative analysis of samples by Py-MS. A suitable means of preventing such artifacts is to "standardise" solutions or suspensions throughout the series of samples to be compared by using buffered saline, e.g. aqueous PBS buffer (0.01 M phosphate, 0.145 M CI", 0.17 M Na+; pH 7.2). Even when disregarding any signals directly derived from residual solvent, a change of solvent, e.g. from methanol to water or carbon disulphide may appreciably influence the pyrolysis pattern (refs. 100, 101).

Careful attention should be paid to the cleanliness of the pyrolysis wire and of the glass or quartz reaction tube surrounding the filament. The use of ferromagnetic wires made of a fast rusting alloy and in particular pure iron wires (Tc = 770°C),

Figure 11. Pyrolysis mass spectra of polygalacturonic acid, using different sample solvents, (a) Suspension in methanol;(b) suspension/solution in water, pH 3.8;(c) solution in aqueous sodium hydroxide, pH 9.5;(d) solution in phosphate buffered saline (PBS), pH 6.5. Conditions:sample concentrations 1 mg/ml; samples 10 yg; Tc 510°C; Eei 14 eV. Note that the acidic aqueous and methanolic suspensions (a, b) give similar spectra. Use of the alkaline medium (c) results in marked intensity changes, e.g. at m/z 31, 32, 60, 68, 74, 85, 96, 102 and 114. The formation of furanoic components seems to be reduced under the influence of alkali. The neutral buffered solution gives an intermediate spectrum;the high concentration of chloride ions gives rise to the intense peaks at m/z 36 and 38 (HC1+').

in combination with aqueous solutions or suspensions should be avoided. As Curiepoint wires are very inexpensive, they may be discarded after use. Used glass or quartz reaction tubes, however, can be cleaned by boiling in acids (e.g. chromic acid), rinsing in de-ionised water and subsequent oven drying. The choice of the filament cleaning method, e.g. prolonged heating in a water-saturated hydrogen atmosphere, ultrasonic cleaning in organic solvents or pre-pyrolysing in a vacuum environment, also noticeably influences the pyrolysis pattern (refs. 100, 101). We most frequently use the reductive hydrogen cleaning technique because it removes oxide layers which may cause catalytic reactions and/or change the emissivity of the filament surface. However, it should be noted that reductive cleaning techniques can cause severe hydrogen absorption by the metal. This may influence the pyrolytic reactions, as demonstrated by Kutter et at. (ref. 102) using small ferromagnetic cylinders for the pyrolysis of nitro and azo compounds. Hydrogen can also be formed pyrolytically from residual solvents (methanol, water).

To apply the sample from a solution or suspension, a micropipette is used to deposit a 5-microlitre drop close to the tip of the ferromagnetic wire. Figure 12 shows a batch of 12 wires being coated while protruding from the glass reaction tubes and slowly rotating to ensure uniform distribution of the sample. If necessary, the whole assembly can be pumped off in a vacuum enclosure for fast drying. After drying, the wires are retracted to the proper position in the glass reaction tube, which also serves as a protective cover and even allows shipping of coated wires by mail (see Figure 13). An alternative coating technique may be used with slurries, sludges or pastes which are too thick for micropipetting. Such samples can be simply smeared on the wire using a platinum sample loop or by sticking the wire directly into the sample. With bacterial colonies, this technique, although providing less uniform sample coatings than the drop technique, has considerable advantages over sequential washing, freeze-drying and re-suspending procedures which may cause irreproducible changes in the chemical composition of the sample and loss of characteristic metabolites (ref. 103). Preferred amounts of sample in Curiepoint Py-MS vary between 1 and 20 micrograms. Below 1 microgram, background signals may significantly contribute to the patterns. As reported by lieuzelaar (ref. 100) and Windig et al. (ref. 101) changing the amounts of sample from 2 to 20 micrograms causes minimal changes in the relative peak heights of the patterns of glycogen and

Figure 12. Apparatus for the batch coating and drying of Curie-point wires. Note that the wires are revolved by magnetic attraction to the base plate assembly (right) which makes a slow circular movement. All wires can be withdrawn simultaneously into the reaction tubes by sliding the manifold away from the base-plate.

Figure 12. Apparatus for the batch coating and drying of Curie-point wires. Note that the wires are revolved by magnetic attraction to the base plate assembly (right) which makes a slow circular movement. All wires can be withdrawn simultaneously into the reaction tubes by sliding the manifold away from the base-plate.

albumin. Above 20 mi-crograms, however, the increase in overall peak height is no longer proportional to sample weight, indicating that parts of the material are blown off the wire before being pyrolysed. Occasionally, the use of amounts of samples up to 60 - 70 yg cannot be avoided, e.g. when pyrolysing whole soil samples containing less than \% of organic material (ref. 51). In this case, however, sample blow-off is not as severe, as the inorganic part of the sample usually does not produce large amounts of gaseous pyrolysis products.

Figure 13. Coated Curie-point wires and reaction tubes in shipping container which can hold up to 96 tubes.

Typical concentrations of solutions or suspensions used for sample coating are 1 - 2 mg/ml. When applying 5 pi drops, this amounts to 5 - 10 micrograms per drop. The normal repeatability of sample amounts applied to the filaments is approximately ±10% when using the drop technique and ±50% when using the smear technique.

0 0

Post a comment