UNDP/GEF Danube Regional Project

Preparation of Reference Materials for
Analytical Quality Control in the Water
Laboratories

Report on Homogeneity Tests of the First
Sediment Reference Material

August, 2003

VITUKI Plc.,
WATER RESOURCES RESEARCH CENTRE Plc.
H-1095 Budapest, Kvassay Jen út 1.
Hungary
Hungary

LIST OF CONTENT

A.
PROJECT OBJECTIVES..................................................................................1
B.
APPROACH OF WORK IN LINE WITH THE REQUIRED SERVICE..............1
Sediment RMs:.................................................................................................1
C.
HOMOGENIZATION .........................................................................................2
D.
HOMOGENEITY TESTING...............................................................................2
1. Bulk homogeneity test................................................................................2
2. Specific homogeneity test..........................................................................4
E.
EVALUATION OF HOMOGENEITY .................................................................8
F.
PROGRESS OF WORK..................................................................................13

A. PROJECT
OBJECTIVES

Objective: Ensure the reliability and comparability of monitoring results during the
implementation of the transnational water quality monitoring (TNMN) in the Danube
river basin (DRB) by providing appropriate reference materials (RM) for analytical
quality Control (AQC).

Outputs: Ensure availability of homogenous reference samples (RM), as AQC
sample, for analyzing specified pol utant characteristics in water and sediment. The
RMs shall be available for performance testing (intercalibration) and intra-laboratory
quality control.

This project output will assist DRB-countries to control the water quality monitoring
results by reference samples in their laboratories and to ensure sustainable quality
work as well as to improve their working quality as needed.

Implementation of the project will ensure the continuity of the quality assurance
activities in the DRB which have been developed and maintained since 1995 in the
frame of different projects supported from different financial sources, e.g. individual
countries, PHARE programme and the ICPDR.

B.
APPROACH OF WORK IN LINE WITH THE REQUIRED SERVICE

In line with the Work Program of the MLIM Expert Groups, there is a need to ensure
and maintain the analytical quality control measures in the water laboratories in the
DRB as a basic requirement of the quality assurance in the trans-national
monitoring. This Project provides significant quantities of water and sediment RMs:

Sediment RMs:
based on earlier experience concerning the problem of determination of the different
river quality characteristics the target determinands include nutrients, i.e. nitrogen
and phosphorus forms, heavy metals, as well as selected determinands to
characterize organic contamination, e.g. total organic carbon (TOC), chlorinated
hydrocarbons, ((DDT, HCB, HCH (-, -, -), PAHs (fluoranthene,
benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, indenopyrene,
benzo(g,h,i)perylene (Borneff PAHs)).

The compounds of interest in the sediment RMs will be in the range of
concentrations found in Danube sediments.

In the case of sediment reference materials all the preparation steps and tests are
specified. The procedure for producing powdered RMs is the following:

1

a) sampling
b) cleaning
c) air drying
d) grinding
e) sieving

fraction > 63 ľm

discarded
f) fraction < 63 ľm
g) homogenization
h) bulk homogeneity testing
i) storage
j) testing
k) transport

C.
HOMOGENIZATION

Homogenization of sediment (fraction < 63 ľm) was carried out by a special
equipment in VITUKI. The time of homogenization at a constant rotating speed was
five days.

D.
HOMOGENEITY TESTING

1. Bulk homogeneity test

During production, sub-samples were taken from the bulk material to control the
homogeneity on three heavy metals: copper, lead, nickel, using a relatively fast and
accurate method (atomic absorption spectrometric method). The test results were
used as indicators of successful bulk homogenization.

2

Randomly selected four sub-samples were tested. Each sub-sample was analyzed in
five replicates. The results of bulk homogenization tests can be seen in Table 1.

Table 1. Results of the first homogenization

Sample-1 Sample
-2
Replicate
sample
Cu
Ni
Pb
Cu
Ni
Pb
(mg/kg) (mg/kg) (mg/kg) (mg/kg) (mg/kg) (mg/kg)
/1 45,20
33,00
32,00
45,80
34,40
32,50
/2 45,40
35,40
31,90
46,20
35,20
32,90
/3 46,10
38,60
33,00
45,40
37,30
31,90
/4 45,40
37,50
31,30
46,20
37,20
32,40
/5 45,10
37,60
31,30
45,80
37,30
31,30
Minimum
45,10 33,00 31,30 45,40 34,40 31,30
Maximum
46,10 38,60 33,00 46,20 37,30 32,90
Median
45,40 37,50 31,90 45,80 37,20 32,40
Average
45,44 36,42 31,90 45,88 36,28 32,20
SD
0,39 2,24 0,70 0,33 1,38 0,62
Replicate
Sample-3 Sample
-4
sample
/1 45,80
35,30
31,30
45,20
35,40
32,80
/2 45,70
36,80
31,30
45,40
38,00
31,30
/3 45,10
35,30
31,90
46,20
35,60
32,00
/4 45,70
37,80
32,00
45,80
37,20
31,90
/5 46,10
37,30
31,30
45,30
37,50
31,40
Minimum
45,10 35,30 31,30 45,20 35,40 31,30
Maximum
46,10 37,80 32,00 46,20 38,00 32,80
Median
45,70 36,80 31,30 45,40 37,20 31,90
Average
45,68 36,50 31,56 45,58 36,74 31,88
SD
0,36 1,15 0,36 0,41 1,17 0,60

As the analytical results indicated inhomogeneity, a repeated homogenization was
necessary.

3

2. Specific homogeneity test

After the repeated homogenization specific homogeneity both for between-units and
within-unit was tested.

Between-units homogeneity

It was necessary to check using a reasonable number of samples if the
differences between-units stay within acceptable limits.

The selected analytical method was atomic absorption spectrometric method, which
was applied under the best repeatable conditions: all samples were measured on the
same day, with the same instrument, by the same operator, against the same
calibrates.

Five representative replicate samples were selected from each bottle in order to
verify their homogeneity. All the representative samples were analyzed by VUVH
and VITUKI as well.

Within-unit homogeneity

Within-unit homogeneity was calculated from the five replicates per bottle (unit).

The analytical results are presented in Table 2-7. The analytical data summarized in
Table 2, 4, 6 originate from VUVH, while analytical results in Table 3, 5, 7 originate
from VITUKI.

4

Table 2. Results of homogeneity test

Bottle-1 Bottle-2 Bottle-3 Bottle -4

Replicate
sample
Cu
Cu
Cu
Cu

(mg/kg) mg/kg) (mg/kg) (mg/kg)
/1
46,85
46,45
45,49
45,89

/2
45,13
46,04
45,86
46,94

/3
45,93
46,16
45,32
46,65

/4
45,63
45,95
45,95
45,46

/5
45,89
46,66
46,05
45,99

Minimum
45,13 45,95 45,32 45,46 45,13Minimum
Maximum
46,85 46,66 46,05 46,94 46,94Maximum
Median
45,89 46,16 45,86 45,99 45,94Median
Average
45,89 46,25 45,73 46,19 46,01Average
SD
0,63 0,30 0,31 0,60 0,46SD
RSD (CVw)
1,36 0,64 0,69 1,30 1,00RSD (CVb)
Uw
0,43 0,20 0,22 0,41 0,32Ub
CVw+Uw
1,80 0,84 0,90 1,71 1,31CVb+Ub
CVw-Uw
0,93 0,44 0,47 0,89 0,68CVb-Ub

Table 3. Results of homogeneity test

Bottle-1 Bottle-2 Bottle-3 Bottle -4

Replicate
sample
Cu
Cu
Cu
Cu

(mg/kg) (mg/kg) (mg/kg) (mg/kg)
/1
45,20
45,80
45,80
45,20

/2
45,40
46,20
45,70
45,40

/3
46,10
45,40
45,10
46,20

/4
45,40
46,20
45,70
45,80

/5
45,10
45,80
46,10
45,30

Minimum
45,10 45,40 45,10 45,20 45,10Minimum
Maximum 46,10 46,20 46,10 46,20 46,20Maximum
Median
45,40 45,80 45,70 45,40 45,55Median
Average
45,44 45,88 45,68 45,58 45,65Average
SD
0,39 0,33 0,36 0,41 0,38SD
RSD (CVw) 0,86 0,73 0,80 0,91 0,82RSD (CVb)
Uw
0,27 0,23 0,25 0,29 0,26Ub
CVw+Uw
1,13 0,96 1,05 1,20 1,08CVb+Ub
CVw-Uw
0,59 0,50 0,54 0,62 0,56CVb-Ub

5

Table 4. Results of homogeneity test

Bottle-1 Bottle-2 Bottle-3 Bottle -4

Replicate
sample
Ni
Ni
Ni
Ni

(mg/kg) (mg/kg) (mg/kg) (mg/kg)
/1
34,83 36,97 34,74 34,84

/2
35,90 35,72 35,02 35,23

/3
36,19 36,19 35,15 35,43

/4
36,37 36,39 35,12 35,04

/5
36,00 36,15 35,32 34,54

Minimum
34,83 35,72 34,74 34,54 34,54Minimum
Maximum
36,37 36,97 35,32 35,43 36,97Maximum
Median
36,00 36,19 35,12 35,04 35,56Median
Average
35,86 36,28 35,07 35,02 35,56Average
SD
0,60 0,45 0,21 0,34 0,40SD
RSD (CVw)
1,68 1,25 0,61 0,98 1,13RSD (CVb)
Uw
0,53 0,40 0,19 0,31 0,36Ub
CVw+Uw
2,21 1,65 0,80 1,30 1,49CVb+Ub
CVw-Uw
1,15 0,86 0,42 0,67 0,77CVb-Ub

Table 5. Results of homogeneity test

Bottle-1 Bottle-2 Bottle-3 Bottle -4

Replicate
sample
Ni
Ni
Ni
Ni

(mg/kg) (mg/kg) (mg/kg) (mg/kg)
/1
33,00 34,40 35,30 35,40

/2
35,40 35,20 36,80 38,00

/3
38,60 37,30 35,30 35,60

/4
37,50 37,20 37,80 37,20

/5
37,60 37,30 37,30 37,50

Minimum
33,00 34,40 35,30 35,40 33,00Minimum
Maximum
38,60 37,30 37,80 38,00 38,60Maximum
Median
37,50 37,20 36,80 37,20 37,20Median
Average
36,42 36,28 36,50 36,74 36,49Average
SD
2,24 1,38 1,15 1,17 1,49SD
RSD (CVw)
6,15 3,81 3,15 3,18 4,07RSD (CVb)
Uw
1,94 1,20 1,00 1,01 1,29Ub
CVw+Uw
8,09 5,01 4,15 4,19 5,36CVb+Ub
CVw-Uw
4,20 2,60 2,16 2,18 2,78CVb-Ub

6

Table 6. Results of homogeneity test

Bottle-1 Bottle-2 Bottle-3 Bottle -4

Replicate
sample
Pb
Pb
Pb
Pb

(mg/kg) mg/kg) (mg/kg) (mg/kg)
/1
31,18
31,10
30,67
30,95

/2
30,51
30,95
30,72
31,44

/3
30,46
30,89
31,00
31,68

/4
30,37
30,99
30,62
31,65

/5
30,79
31,62
30,95
31,35

Minimum
30,37 30,89 30,62 30,95 30,37Minimum
Maximum
31,18 31,62 31,00 31,68 31,68Maximum
Median
30,51 30,99 30,72 31,44 30,86Median
Average
30,66 31,11 30,79 31,41 30,99Average
SD
0,33 0,30 0,17 0,29 0,27SD
RSD (CVw)
1,07 0,95 0,56 0,94 0,88RSD (CVb)
Uw
0,34 0,30 0,18 0,30 0,28Ub
CVw+Uw
1,41 1,25 0,73 1,23 1,16CVb+Ub
CVw-Uw
0,73 0,65 0,38 0,64 0,60CVb-Ub

Table 7. Results of homogeneity test

Bottle-1 Bottle-2 Bottle-3 Bottle -4

Replicate
sample
Pb
Pb
Pb
Pb

(mg/kg) (mg/kg) (mg/kg) (mg/kg)
/1
32,00 32,50 31,30 32,80

/2
31,90 32,90 31,30 31,30

/3
33,00 31,90 31,90 32,00

/4
31,30 32,40 32,00 31,90

/5
31,30 31,30 31,30 31,40

Minimum
31,30 31,30 31,30 31,30 31,30Minimum
Maximum
33,00 32,90 32,00 32,80 33,00Maximum
Median
31,90 32,40 31,30 31,90 31,90Median
Average
31,90 32,20 31,56 31,88 31,89Average
SD
0,70 0,62 0,36 0,60 0,57SD
RSD (CVw)
2,18 1,91 1,13 1,87 1,78RSD (CVb)
Uw
0,69 0,61 0,36 0,59 0,56Ub
CVw+Uw
2,87 2,52 1,49 2,47 2,34CVb+Ub
CVw-Uw
1,49 1,31 0,78 1,28 1,21CVb-Ub

7

E.
EVALUATION OF HOMOGENEITY

The coefficients of variation (relative standard deviation) were obtained from the
between-unit and within-unit homogeneity tests (CVB and CVW respectively).

Their respective uncertainties UCV, are defined as follows:

UCV CV/ 2n , where n = number of replicates.

No inhomogeneity is detected when the uncertainty ranges of the two coefficients of
variation overlap, or when
CVB ą UCV, and CVW ą UCV show overlap.

Variation coefficients and uncertainties are also presented in Table 2-7 based on
copper, nickel and lead analytical results.

The evaluated values from between-unit and within-unit homogeneity test are
summarized in Table 8.

The bolded values inside the range (CVB ą UCV) indicate the homogeneity of
sediment RM and it is plotted in Fig. 1-3.

The requirement of homogeneity namely the uncertainty ranges of the two
coefficients of variation shall be overlapped is achieved.

8

Table 8. Evaluation of homogeneity test results

Copper
VUVH VITUKI
CVB ą UCV CVW ą UCV
CVB ą UCV
CVW ą UCV
w1 0,93-1,80 w1 0,59-1,13
w2 0,44-0,84
w2 0,50-0,96
0,68-1,31
0,56-1,08
w3 0,47-0,90
w3 0,54-1,05
w4 0,89-1,71
w4 0,62-1,20

Nickel
VUVH VITUKI
CVB ą UCV CVW ą UCV
CVB ą UCV
CVW ą UCV
w1 1,15-2,21 w1 4,20-8,09
w2 0,86-1,65 w2 2,60-5,01
0,77-1,49
2,78-5,36
w3 0,42-0,80
w3 2,16-4,15
w4 0,67-1,30
w4 2,18-4,19

Lead
VUVH VITUKI
CVB ą UCV CVW ą UCV
CVB ą UCV
CVW ą UCV
w1 0,73-1,41 w1 1,49-2,87
w2 0,65-1,25 w2 1,31-2,52
0,60-1,16
1,21-2,34
w3 0,38-0,73
w3 0,78-1,49
w4 0,64-1,23
w4 1,28-2,47

9

Cu (VUVH)
1,9
1,8
1,7
1,6
1,5
1,4
1,3
1,2
CV
1,11
ą U B
0,9
between-unit
CV
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,10
S-1
S-2
S-3
S-4
CVw-Ucv
within-unit
CVw+Ucv

Cu (VITUKI)
1,3
1,2
1,1
1
V
0,9
C
0,8
ą U B
0,7
between-unit CV
0,6
0,5
0,4
0,3
0,2
0,1
0
S-1
S-2
S-3
S-4
CVw-Ucv
within-unit
CVw+Ucv

Figure 1. Graphical representation of homogeneity of sediment RM

10

Ni (VUVH)
2,5
2
1,5
CV
1
ą UB
between-unit
CV
0,5
0
S-1
S-2
S-3
S-4
CVw-Ucv
within-unit
CVw+Ucv
Ni (VITUKI)
9
8
7
6
5
-unit
CV
n
4
e
ą U B
twe
3
be
CV
2
1
0
S-1
S-2
S-3
S-4
CVw-Ucv
within-unit
CVw+Ucv

Figure 2. Graphical representation of homogeneity of sediment RM

11

Pb (VUVH)
1,5
1,4
1,3
1,2
1,1
1
CV
unit
0,9
een- ą U
0,8
B
w
0,7
CV
bet
0,6
0,5
0,4
0,3
0,2
0,1
0
S-1
S-2
S-3
S-4
within-unit
CVw-Ucv
CVw+Ucv

Pb (VITUKI)
3,5
3
2,5
CV
2
ą U B
1,5
CV
between-unit
1
0,5
0
S-1
S-2
S-3
S-4
CVw-Ucv
within-unit
CVw+Ucv

Figure 3. Graphical representation of homogeneity of sediment RM

12

F.
PROGRESS OF WORK

The Project goes forward according to the work plan (Annex 2: paragraph E). The
first set of the water RMs prepared for nutrients (ammonium-N, nitrate-N, phosphate-
P, total P) and heavy metals (cadmium, copper, chromium, lead, mercury, nickel,
aluminum arsenic) were delivered in August.

To prepare the first set of sediment RMs bulk sediment was collected in the Danube.
The sediment was processed according to the flow chart of sediment RM preparation
(cleaning, drying, grinding, sieving and homogenization). After homogenization bulk
homogeneity test was carried out. As the analytical results did not show
homogeneity this step was repeated again. After the second homogenization the
sediment was bottled and between- and within-units homogeneity tests were carried
out. As the evaluated results of homogeneity tests showed homogeneity, the
sediment RMs were delivered together with water RMs.

The reference material samples have been prepared and tested according to the
"Practical Manual for Production of Laboratory Reference Materials" as well as ISO
Guide 34, 1996. Quality system guidelines for the production of reference materials.
Ibid.

Budapest, 29/08/2003

Dr. Jolan Schneider

team leader

13