Patent Application
20160054241
The invention relates to preparation of samples by pressing, for example for X-ray fluorescence analysis (XRF), including for example the preparation of dairy samples.
Taking measurements of samples frequently includes preparing the sample in some way for measurement. For samples that are not fully homogenous, some form of breaking up and mixing may be necessary to avoid variation across the sample. Further, the sample may need to be formed into a sample shape of the correct shape.
In particular, samples may be pressed into pellets for analysis.
Milk contains mineral elements in different fractions (Rinaldoni et al, “Analytic determinations of mineral content by XRF, ICP and EEA in ultrafiltered milk and yoghurt”, Latin American Applied Research, Volume 39, pages 113 to 118, 2009). These components may be considered to be proteins, fat and sugar related phases.
If samples with relatively high fat content (for example greater than 10%) are processed into pellets, then the application of high pressure during a pressing step, for example, larger than 3 to 5 ton/cm2 (3 to 5×1012 Pa) may cause the fat to extrude from the sample. This makes precise measurement impossible. However, samples with lower fat content may need larger pressures to form stable pellets with sufficient integrity to survive the pellet manufacturing process. These issues are particularly relevant for dairy products which may have fat levels either above or below 10% and hence which may require a range of different pressures based on the composition of the sample.
Similar effects may be observed for samples with other components that may mobilise under pressure. Such components may include oils, moisture, protein, or other biological material. Mobilisation of components within mixed samples can lead to separation of the mobile component or simply inhomogeneity that gives poor or non-repeatable results when carrying out measurements on the sample.
Pressed samples may be used, for example, for X-ray fluorescence measurements.
A prior approach to XRF measurement of dairy samples, containing fat, is provided in Pashkova, “X-ray fluorescence determination of element contents in milk and dairy products”, Food Anal. Methods (2009) 2.303-310. This describes preparing milk powder pellets weighing 4 g and with 40 mm diameter using pressures from 2 to 8 tons with a hydraulic press. Lower pressures were used for samples with high fat content and higher pressures for samples with lower fat content. Additionally it is well documented that dried milk samples with fat content >10%, when pressed under pressure >2 tons, will extrude fat.
However, using different processes for different samples leads to difficulties since it is then difficult to directly compare results in the case that different methods are used to obtain the results. Comparing results obtained using different methods can lead to unpredictable results. This is particularly the case for low density samples which are highly compacted during the preparation of the pellets. Different pressures can lead to different results.
There is therefore a need for an improved method of sample preparation which can be applied to samples in particular to dairy samples, especially in powder form. Similar issues can arise in other products, for example other types of human food, animal feed, and dietary supplements.
A good degree of homogeneity is required, since inhomogeneity will introduce variations into the X-ray fluorescence results which will deliver poor reproducibility if the samples are not mixed.
According to a first aspect of the invention there is provided a method of preparing samples according to claim 1.
The component that may mobilise may be fat, moisture, protein, oils etc. . . . any phase which is disturbed by pressure and becomes mobile, and the additive may be a binding additive for binding that component. The additive may comprise activated carbon, or alternative material such as activated alumina.
In a particular embodiment, the additive may in particular be a fat-binding additive for binding fat. A fat binding additive that binds to the fat in the immediate vicinity of its segregation when the sample and binder are pressed is preferred. The fat-binding additive may be activated carbon or similar acting compounds for example activated alumina.
The aforementioned binder may further comprise a wax, for example a micronized wax. The amount of wax may vary according to the physical requirements of the desired sample pellet.
The step of pressing into a pellet may be carried out at a pressure of over 5 ton/cm2. A particular benefit of embodiments of the invention is that the fat-binding additive means that the same process may be carried out for varied amounts of fat.
The step of milling and mixing may be carried out in a mixer-mill. Alternatively, separate milling and mixing steps may be carried out to achieve homogenisation and thorough mixing of the additives.
The amount of binder varies according to sample characteristics and as such can represent a weight fraction of the total weight of the sample and binder, preferably the amount of binder is 6% to 15% by weight of the total weight of the sample and binder.
The sample may be a dry powder dairy sample but also any sample type which may benefit from the additive previously described. In particular, as well as a dry powder dairy sample, the sample may be an animal feed, a dietary supplement, or processed human food.
The invention also relates to a method of making XRF measurements including preparing a sample as discussed above.
For a better understanding of the invention, embodiments will now be described, purely by way of example, with reference to the accompanying drawing, in which
Dairy powder samples and other similar samples were used for testing. A sample preparation method was targeted which takes into account the small-scale inhomogeneity of some samples and which was applicable for all fat content, while providing best possible repeatability results.
Although the method is described for use with samples containing fat, the method is also applicable to samples including other similar content which can segregate or otherwise become heterogeneous. Such samples may include animal feed, dietary supplements or processed human food.
To homogenise the samples in a homogenisation step was carried out using a mill to homogenise, mix and grind the sample. In alternative arrangements, a simple mixer may be used or alternatively a multi-step process may be used for example with separate sample preparation and mixing apparatus.
In order to bind the fat, the method according to the specific embodiment uses an additive for binding fat which may segregate when the sample is pressed to form a pellet. Preferably, when the sample is pressed and the fat segregates, the fat should be bound at its immediate vicinity.
Many additives were tested and it was observed from visual inspections and also XRF measurements that activated carbon provided the best fat-binding characteristics. Activated alumina is an alternative. Other substances tested included boric acid, different grades of cellulose, starch, Hoechst wax, and mixtures of those.
Without wishing to be bound by theory, it is believed that the activated carbon, due to its extremely high surface area, adsorption capacity, chemical and physical binding characteristics, allowed for fatty samples to be pressed into a stable pellet at pressures as high as 10 ton/cm2 without significant fat exudation. In particular, good results were obtained with the addition of 5% activated carbon.
By way of comparison, without using activated carbon, fatty infant formula samples are completely wet from exudating fat when pressed at 5 ton/cm2.
In a step to further improve the repeatability of the sample preparation, mixtures of additives were tested and it was observed that the use of activated carbon and wax rendered even better results than the pure activated carbon.
To illustrate this,
The use of pure activated carbon as a fat-binder, although very efficient, proved to be a problem when producing pellets from very low fat products, as those were not mechanically stable. To address this, a mixture of carbon and wax was used as the binder.
Pellets were formed as set out above and tested in a Panalytical XRF spectrometer, of type “E3-XL”. The amounts of a number of elements was tested, namely Ca, Cl, Cu, Fe, K, Mg, Mn, Na, P, S and Zn. Calibration plots of counts per second against parts per million of the variety of elements was obtained. Good calibration was achieved, in other words the known concentration of the elements was highly correlated with the measured count, bearing in mind the wide variety of samples tested.
The calibrated XRF apparatus was then used to measure a variety of samples of skimmed milk powder and high-fat infant formula milk. All were thoroughly mixed to assure maximum homogeneity.
As a comparison, an alternative sample preparation method according to a comparative example was used. The comparative example used direct pressing of the product at 10 ton/cm 2 for skimmed milk and 3 ton/cm2 for fatty infant formula.
The samples according to the invention were carried out and relative standard deviations were calculated for a range of elements.
Table 1 shows the system repeatability values for ten repeats of a comparative example using skimmed milk powder simply pressed into pellets (top ten lines) and for ten repeats of the same skimmed milk powder using a sample prepared using the recipe as set out above. Note that the system repeatability is the repeatability of the measurement using the same sample.
Table 2 shows the method repeatability values for ten repeats of a comparative example using skimmed milk powder simply pressed into pellets (top ten lines) and for ten repeats of the same skimmed milk powder using the invention. Note that method repeatability repeats the whole experiment using the same two procedures.
Table 3 shows the method repeatability values for ten repeats of a comparative example using high fat infant formula sample (26% fat) simply pressed into pellets (top ten lines) and for ten repeats of an example using the invention. Note that method repeatability repeats the whole experiment using the same two procedures.
In brief, comparison of the relative standard deviation, expressed in %, in tables 2 and 3 shows better results for the lower half of the table (the method of the invention) compared with the upper half (the comparative example). This is particularly the case for table 3, the high fat sample, which shows that the method according to the invention is successful at dealing with such samples.
Good repeatability of measurement was obtained for Ca, Cl, K, Mg, Na, P, S and Zn.
For the low fat sample (skimmed milk) illustrated in table 2, relatively poor results were obtained for Cu, though better using the invention than the comparative example, in view of the fact that the concentration of Cu was very low, near to the method detection limit.
Results for Fe and Mn were also poor in table 2, being also below detection limits. In the case of Mn, note for example the negative numbers obtained in table 2 for the comparative examples. No usable results for Fe were obtained using the comparative example of table 2. However, Fe did give repeatable results using the recipe according to the invention in table 2.
As for table 3, it is notable that reasonable values of Fe and Cu were obtained in this case, i.e. for the high fat sample. The use of the recipe according to the invention, in the lower half of the table, increased repeatability, for example for Ca and Cl the repeatability improved by factors of 8.2 and 6.3 respectively.
It thus appears that the use of the fat-binder avoids fat transport during pressing and that the homogeneity of the samples is improved by the mixing process used. The method of preparing samples accordingly improves the measurement of various element in samples containing variable amounts of fat, for example milk powder of different fat concentrations. The same method may accordingly be used for milk powder of widely varying fat concentration, improving repeatability, reliability and the comparison accuracy.
Those skilled in the art will appreciate that variations and additions may be made to the invention.
The embodiments presented above all use activated carbon. However, the inventors have also achieved positive results with finely powdered graphite
We further attach data prepared with five samples of the pressed dairy pellet containing 16% fat. Table 4 relates to five samples simply pressed into a pellet (upper half of table) and five samples mixed with wax and graphite powder and pressed into a pellet (lower half of table).
A marked improvement using the stabilizing binder of wax and graphite is clearly seen for all elements except Mn which is below the method detection limit and Zn.
Note that instead of activated carbon or graphite a mixture of both, i.e. graphite and activated carbon may be used.
To further improve the measurement accuracy various techniques may be used, for example by increasing measurement times. Fine tuning of the apparatus and the calibration may also be used, for example by including additional secondary standards based on additional milk powder samples.
As well as being relevant for binding fat, other materials may also be bound such as moisture, protein or oils.
| Number | Date | Country | Kind |
|---|---|---|---|
| 13162472.8 | Apr 2013 | EP | regional |
| 13165326.3 | Apr 2013 | EP | regional |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2014/056867 | 4/4/2014 | WO | 00 |
