Discussion 84 Determination of volatile fatty acids VFAs in silage

We still await a routine method for the rapid analysis of large batches of silage samples for VFAs. Although initially expensive, the use of NIRS on fresh silages would be ideal, but sufficiently robust equations for components other than lactic, acetic, propionic and butyric acids are not yet widely available, although it is reported that commercial services are offering VFA analysis by NIR. The abstract of the poster by Deaville and Givens (1996) is apparently the only published data in this area (D.I. Givens, ADAS, Stratford-on Avon, 2001, personal communication), which suggests further work is needed to improve the accuracy of prediction, especially for acetic and propionic acids. That leaves gas chromatography (GC) and high performance liquid chromatography as the two methods most commonly employed, although automatic titration methods also exist. The chromatographic methods have traditionally required that the acids undergo a time-consuming derivatization step (Jones and Kay, 1976; MAFF/ADAS 1986, pp. 235-239) converting them to esters, which are more easily separated and detected. For about 20 years, procedures have been developed allowing the direct injection of silage juice, or aqueous extracts. The analysis of synthetic mixtures of the VFAs usually give clearly separated peaks, but when the actual silage extract is injected, there are several problems. For example, in GC, lactic acid tends to give a large broad peak that obscures the following small peaks of any n-valeric or iso- and n-caproic acids. In HPLC, there are so many different organic compounds in the silage extract with each producing a peak on the chromatogram, that it is difficult to identify a peak with any degree of certainty. Even knowing the exact elution time of the pure acid is of no real benefit when analysing the silage extract, because an increase in the column temperature or in the acidity of the eluant has been found to cause the peak previously attributed to just one VFA to be resolved into two. There can thus be two or more peaks eluting simultaneously, which leads to an erroneously high estimate of concentration for that particular VFA. Succinic acid, for example, elutes at only about 18 s before lactic acid in HPLC analysis using a Bio-Rad Aminex HPX 87 H ion-moderated partition column, and may well overlap to varying degrees. It may be advantageous to analyse some components, such as ethanol, by GC, and others, such as lactic acid, by HPLC. There are numerous columns available for both GC (packed and capillary) and HPLC, each with special instructions from the manufacturer.

In this case, therefore, an actual analytical procedure will not be recommended, but references to some published methods and some guidelines will be given.

Gas chromatography of VFAs

A GC method using direct injection of silage juice extract has been proposed by Fussell and McCalley (1987) using a Carbopack B-DA/4% Carbowax 20M (80-120 mesh) column. This improved on the tailing lactic acid peaks of Playne (1985) who used a Chromosorb 101 (80-100 mesh) column. A column capable of derivatizing in situ was suggested by Suzuki and Lund (1980). It consisted of poly(ethylene glycol phthalic acid ester) coated on a solid terephthalic acid support, and sharpened up the lactic acid peak. Galletti and Piccaglia (1983) used a Porapak QS (80-100 mesh) column to separate lactic acid and C2-C6 VFAs in 10 min using acidified portions of aqueous maize silage extract.

We obtained the best results with the Carbopack B-DA/4% Carbowax 20M, 80-120 mesh Supelco column, 2000 x 2 mm. It was initially conditioned for 21 h at 245°C, but the normal running temperature is 175°C. The injector/detector temperature is 200°C and a flame ionization detector is used. A glass sleeve is fitted to the injector and the glass wool plug removed from the column inlet. The carrier gas is nitrogen with a flowrate of 40 ml min-1 at 310 kN m-2. The sample solution (9 ml) is mixed with 1 ml of pivalic acid solution (1.6% m/v) as internal standard. Then 1 ml of this solution is mixed with 1 ml 0.3 M oxalic acid solution and 3 ml deionized water before injecting 1 pl into the septum.

A GC chromatogram of a mixture of known VFAs, lactic acid and pivalic acid (internal standard) is shown in Fig. 8.3, and a chromatogram of a silage juice extract with pivalic acid internal standard is shown in Fig. 8.4.

HPLC of sllage VFAs

Careful work at the Scottish Agricultural College (www.sac.ac.uk) by Rooke et al. (1990) and Salawu et al. (1997) has enabled the successful analysis of silage VFAs using the Bio-Rad Aminex HPX-87H ion-moderated partition column and a refractive index detector that enables ethanol in addition to the VFAs to be detected. They also successfully used an equivalent Supelco Supelcogel C-610H 300 x 7.7 mm ID column. In either case, a guard column is used (Bio-Rad Cation H+ Cat. No. 125-0129). Other workers (Kubadinow, 1982; Canale et al., 1984; Siegfried et al., 1984) have also used the Aminex HPX-87H 300 x 7.8 mm column with an Aminex HPX-85H 40 x 4.6 mm guard column and UV detection at 210 nm.

Minutes

Fig. 8.3. GC chromatogram of mixed silage juice standards using a Carbopack B-DA column. Identity and concentrations (before diluting 4:1 standard:0.3 M oxalic acid): a, ethanol, 1 mg ml-1; b, acetic acid, 1.25 mg ml-1; c, propionic acid, 0.25 mg ml-1; d, isobutyric acid, 0.25 mg ml-1; e, n-butyric acid, 0.25 mg ml-1; f, pivalic acid (internal standard), 0.4 mg ml-1; g, isovaleric acid, 0.25 mg ml-1; h, lactic acid, 10 mg ml-1; i, n-valeric acid, 0.25 mg ml-1; j, isocaproic acid, 0.25 mg ml-1; k, n-caproic acid, 0.25 mg ml-1.

Minutes

Fig. 8.3. GC chromatogram of mixed silage juice standards using a Carbopack B-DA column. Identity and concentrations (before diluting 4:1 standard:0.3 M oxalic acid): a, ethanol, 1 mg ml-1; b, acetic acid, 1.25 mg ml-1; c, propionic acid, 0.25 mg ml-1; d, isobutyric acid, 0.25 mg ml-1; e, n-butyric acid, 0.25 mg ml-1; f, pivalic acid (internal standard), 0.4 mg ml-1; g, isovaleric acid, 0.25 mg ml-1; h, lactic acid, 10 mg ml-1; i, n-valeric acid, 0.25 mg ml-1; j, isocaproic acid, 0.25 mg ml-1; k, n-caproic acid, 0.25 mg ml-1.

Fig. 8.4. GC chromatogram of a third-cut silage juice using a Carbopack B-DA column. Identity: a, ethanol; b, acetic acid; c, propionic acid ; e, n-butyric acid; f, pivalic acid (internal standard), 0.32 mg ml-1 in injected solution; h, lactic acid.

We experimented firstly with a Phenomenex Rezex ROA-Organic Acid 300 x 7.8 mm column with a Rezex Organic Acid 50 x 7.8 mm guard column; the mobile phase was 0.013 N (0.0065 M) H2SO4. UV detection was at 215 nm, the flowrate 0.6 ml min-1, column temperature 35°C and injection volume 20 [jl. The internal standard was 2-ethylbutyric acid. The lactic acid peak was preceded by a partially overlapping, probably succinic acid peak; the to £

Methyglucuronoxylobiose
Fig. 8.5. HPLC chromatogram of a first-cut silage juice using a Rezex ROA-organic acid column. Identity: a, lactic acid; b, formic acid; c, acetic acid; d, propionic acid; e, n-butyric acid; f, 2-ethylbutyric acid (internal standard).

two peaks coincided at 45°C. A typical chromatogram is shown in Fig. 8.5, with elution times in minutes added by the integrator-printer and positively identified acids (acetic, formic, propionic, n-butyric and lactic), indicated by letters over the elution times.

Second, we experimented with a Spherisorb C8 column using 0.2 M H3PO4 as solvent at a flowrate of 0.6 ml min-1. A typical chromatogram is shown in Fig. 8.6.

The peak at 17.34 min is 50 pg ml-1 mesaconic acid which was added as a possible internal standard, but it proved unsuitable because of an unknown peak eluting at 16.78 minutes seen when analysing silage juice without added internal standard; oxalic acid suffered from the same problem. We still need a suitable internal standard for this column, thus further use was suspended. Other failed compounds included: adipic acid, fumaric acid, d-glucuronic acid, glutaric acid, glycolic acid, 3-hydroxybutanone, itaconic acid, malic acid, maleic acid, malonic acid, pimelic acid and succinic acid.

Extraction procedure

We have found that whatever separation/detection method is used, the extraction procedure used for obtaining the silage juice should be standardized. This is because slightly different amounts of the various components will be extracted from the fibrous silage material depending on whether it is homogenized, shaken or compressed. A compression method is given in 8.4. below. Even the strength of torque applied in the compression method can affect the recovery of VFAs. Thus the recommended torque of 8 Nm recovers on

00 Si

Fig. 8.6. HPLC chromatogram of a first-cut silage juice using a Spherisorb C8 column. Identity: a, formic acid; b, lactic acid; c, acetic acid; d, propionic acid; e, n-butyric acid. The peak at 17.34 min is 50 ^g ml-1 mesaconic acid added as a possible internal standard, which proved unsuitable because of an unknown peak eluting at 16.78 min obtained using silage juice without added internal standard.

average 12% more of the various VFAs and lactic acid than a torque of 5.4 Nm, which is easier to apply. This is to be expected if, as is likely, the concentration of VFAs is not homogeneous throughout the sample, and thus any juice held in the deeper interstices will be expressed by the greater torque.

Internal standard

It is good practice to include an internal standard, which should not be degraded in any way, such as by being heated in a gas chromatograph. It should also give a peak that does not overlap any of the peaks of the sample being determined. Acids and ketones should be suitable, but aldehydes are not recommended. Pivalic acid is often used for this purpose in the GC analysis of VFAs. Although variations in the sensitivity of the determination can be corrected by measuring the peak area of a constant amount of pivalic acid included with the sample, it has its limitations. Repetitive injections of the same sample of silage juice may show that any variation in sensitivity between injections is not always equal for all the peaks. However, the use of an internal standard reduces the error and improves the precision. With HPLC of silage extracts, a widely used internal standard is 2-ethylbutyric acid.

Method 8.4. Extraction method for obtaining silage juice for analysis for VFAs

Equipment.

• Extractor - the extractor is shown in Fig. 8.7, and the means of attaching the small torque wrench in Fig. 8.8. Ideally it should be constructed from stainless steel, but we only used a stainless steel threaded rod or studding (10 x 5/8 in, 11 t.p.i., or 254 x 15.9 mm), and the rest was machined from brass. The silage juice should not remain in contact with the brass for long, however, because the brass will be attacked by the acids. The dimensions

Butyl rubber sealing ring

0 0

Post a comment