The simplest form of glass is single component fused silica (SiO2). This material is, however, very difficult to process, and highly expensive. To minimize these difficulties, some other oxides are added offering specific properties to the resultant glass. Most glasses comprise about 70% silica, which is a glass former, soda as a flux in the form of carbonate and sulfate (about 14%), and lime as a stabilizer in the form of limestone (about 10%). Other types of oxides, like alumina or magnesia, enhance the physical characteristics of glass, especially the resistance to atmospheric conditions. In-depth coloring is obtained by incorporation of different metallic oxides, such a chromium, iron, manganese or copper oxides.
Instrumentation
In this article, an ARL OPTIM’X XRF spectrometer from Thermo Electron (Fig. 1) has been used to derive limits of detection and precision for glasses analysis. The ARL OPTIM’X is a wavelength dispersive system which offers superior resolution and light elements capability. It is fitted with an air-cooled Rh End-Window Tube with thin Be window (0.075 mm), and has a maximum power of 50 W. Due to close coupling between the X-ray tube anode and the sample, the performance of the ARL OPTIM’X is similar to a 200 W conventional WD-XRF instrument. The instrument can be incorporated with the unique SmartGonio, a series of Multichromators, or both.
.jpg)
Figure 1. ARL OPTIM’X XRF Spectrometer.
Calibration and Limits of Detection
A series of pressed glass samples have been measured on an ARL OPTIM’X. Calibration curves have been derived by relating intensities for each oxide (or element) to concentrations in standard samples. X-ray fluorescence measures pure elements, but the results can be related directly to the oxide forms of these elements when only one single form is present in the sample. Using the calibration curves, limits of detection have been derived using the SmartGonio for most common oxides/elements found in soda-lime glasses (Fig. 2). Table 1 shows limits of detection for various elements in soda-lime glasses prepared as pressed powders.
.jpg)
Figure 2. Soda-lime glasses.
Table 1. Analytical parameters and limits of detection for various oxides/element in soda-lime glass (100 sec. counting time).
OXIDE/ELEMENT |
LINE |
CRYSTAL |
DETECTOR |
LOD[PPM] |
Na2O |
Kα1,2 |
AX-06 |
FPC |
100 |
MgO |
Kα1,2 |
AX-06 |
FPC |
60 |
Al2O3 |
Kα1,2 |
PET |
FPC |
47 |
SiO2 |
Kα1,2 |
PET |
FPC |
N.R. |
P2O5 |
Kα1,2 |
PET |
FPC |
48 |
SO3 |
Kα1,2 |
PET |
FPC |
23 |
Cl |
Kα1,2 |
PET |
FPC |
24 |
K2O |
Kα1,2 |
LiF 200 |
FPC |
14 |
CaO |
Kα1,2 |
LiF 200 |
FPC |
12 |
TiO2 |
Kα1,2 |
LiF 200 |
FPC |
12 |
Cr2O3 |
Kα1,2 |
LiF 200 |
FPC |
9 |
MnO |
Kα1,2 |
LiF 200 |
FPC |
9 |
Fe2O3 |
Kα1,2 |
LiF 200 |
FPC |
9 |
ZnO |
Kα1,2 |
LiF 200 |
SC |
3.6 |
SrO |
Kα1,2 |
LiF 200 |
SC |
2.4 |
ZrO2 |
Kα1,2 |
LiF 200 |
SC |
1.8 |
BaO |
Lβ1 |
LiF 200 |
FPC |
51 |
PbO |
Lβ1 |
LiF 200 |
SC |
9 |
N.R. = LoD is not relevant for major elements.
FPC = flow proportional counter.
SC = scintillation counter.
Excitation conditions: 40 kV/1.25 mA.
Collimator: 0.29°.
Precision Tests
Precision tests have been done by repeatedly analyzing the same pressed pellet sample for eleven consecutive analyses. Eighteen elements/oxides are determined using a counting time of 36 seconds per analytical line. The results are summarized below for two different glass samples shown in Tables 2 and 3. For higher precision this counting time can be increased. Doubling the counting time will improve the precision by a factor of about 1.4 (square root of 2)
Table 2. Repeatability for the analysis of the major and minor oxides in sample A.
RUN |
Na2O % |
MgO % |
Al2O3 % |
SiO2 % |
K2O % |
CaO % |
Fe2O3 % |
SO3 PPM |
TiO2 PPM |
P2O5 PPM |
Cl PPM |
Cr2O3 PPM |
MnO PPM |
As2O3 PPM |
SrO PPM |
ZrO2 PPM |
BaO PPM |
PbO PPM |
Time (s) |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
Run 1 |
13.98 |
0.185 |
1.79 |
72.59 |
0.588 |
10.85 |
0.330 |
582 |
579 |
166.9 |
113.4 |
93.6 |
48.4 |
101.6 |
127.4 |
209.1 |
454.3 |
228.2 |
Run 2 |
13.93 |
0.193 |
1.81 |
72.60 |
0.582 |
10.82 |
0.333 |
640 |
563 |
146.3 |
129.5 |
91.4 |
44.7 |
101.5 |
124.8 |
204.6 |
392.6 |
218.9 |
Run 3 |
13.97 |
0.177 |
1.80 |
72.64 |
0.588 |
10.82 |
0.330 |
608 |
563 |
193.3 |
111.2 |
91.1 |
43.8 |
95.9 |
127.1 |
207.0 |
361.8 |
197.8 |
Run 4 |
14.01 |
0.178 |
1.80 |
72.64 |
0.582 |
10.87 |
0.330 |
645 |
581 |
199.2 |
104.6 |
96.2 |
29.9 |
103.8 |
127.0 |
205.4 |
375.7 |
234.6 |
Run 5 |
13.95 |
0.182 |
1.80 |
72.60 |
0.588 |
10.83 |
0.329 |
576 |
564 |
158.1 |
111.8 |
94.6 |
41.8 |
103.7 |
122.7 |
204.7 |
385.2 |
228.5 |
Run 6 |
13.94 |
0.177 |
1.81 |
72.61 |
0.589 |
10.82 |
0.329 |
573 |
569 |
171.3 |
107.9 |
85.2 |
49.5 |
95.0 |
126.3 |
203.8 |
355.4 |
194.8 |
Run 7 |
13.86 |
0.185 |
1.80 |
72.64 |
0.588 |
10.83 |
0.330 |
658 |
569 |
203.6 |
113.4 |
88.9 |
40.3 |
96.0 |
125.0 |
205.4 |
434.1 |
234.4 |
Run 8 |
13.92 |
0.186 |
1.81 |
72.59 |
0.585 |
10.84 |
0.331 |
652 |
566 |
190.4 |
135.6 |
94.5 |
44.7 |
96.8 |
125.9 |
203.4 |
315.1 |
207.3 |
Run 9 |
13.94 |
0.184 |
1.81 |
72.63 |
0.591 |
10.82 |
0.334 |
651 |
579 |
150.7 |
110.1 |
88.6 |
43.1 |
114.9 |
127.1 |
206.6 |
401.2 |
220.8 |
Run 10 |
13.98 |
0.183 |
1.80 |
72.63 |
0.586 |
10.87 |
0.332 |
617 |
526 |
255.0 |
104.0 |
83.6 |
41.9 |
99.2 |
125.8 |
206.2 |
402.2 |
214.6 |
Run 11 |
13.95 |
0.188 |
1.78 |
72.62 |
0.588 |
10.83 |
0.330 |
619 |
561 |
218.3 |
97.9 |
80.6 |
38.9 |
97.7 |
126.8 |
203.1 |
429.8 |
197.8 |
Avg. |
13.95 |
0.183 |
1.80 |
72.62 |
0.587 |
10.84 |
0.331 |
620 |
565 |
186.6 |
112.7 |
89.9 |
42.4 |
100.5 |
126.0 |
205.4 |
391.6 |
216.1 |
Std.Dev. |
0.04 |
0.005 |
0.01 |
0.02 |
0.003 |
0.02 |
0.0015 |
32 |
15 |
32.5 |
11 |
5 |
5.2 |
5.7 |
1.4 |
1.8 |
39.7 |
15 |
Table 3. Table 3: Repeatability for the analysis of the major and minor oxides in sample B.
RUN |
Na2O % |
MgO % |
Al2O3 % |
SiO2 % |
K2O % |
CaO % |
Fe2O3 % |
SO3 PPM |
TiO2 PPM |
P2O5 PPM |
Cl PPM |
Cr2O3 PPM |
MnO PPM |
As2O3 PPM |
SrO PPM |
ZrO2 PPM |
BaO PPM |
PbO PPM |
Time (s) |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
36 |
Run 1 |
13.35 |
0.180 |
1.67 |
73.07 |
0.556 |
10.77 |
773.3 |
0.177 |
556.0 |
200.6 |
100.7 |
63.4 |
9.3 |
118.2 |
122.4 |
227.2 |
883.9 |
895.6 |
Run 2 |
13.33 |
0.180 |
1.68 |
73.08 |
0.564 |
10.76 |
757.9 |
0.181 |
568.0 |
159.5 |
111.2 |
65.9 |
3.9 |
112.0 |
122.2 |
226.1 |
960.6 |
914.5 |
Run 3 |
13.28 |
0.186 |
1.67 |
73.08 |
0.554 |
10.81 |
789.6 |
0.180 |
555.1 |
193.3 |
115.7 |
64.7 |
18.2 |
115.6 |
119.3 |
225.3 |
925.4 |
911.5 |
Run 4 |
13.28 |
0.185 |
1.66 |
73.11 |
0.559 |
10.83 |
768.2 |
0.186 |
587.2 |
156.6 |
103.5 |
74.7 |
8.7 |
105.6 |
126.5 |
225.1 |
891.3 |
900.0 |
Run 5 |
13.35 |
0.181 |
1.67 |
73.05 |
0.554 |
10.79 |
763.9 |
0.181 |
594.7 |
187.4 |
97.9 |
63.6 |
12.3 |
106.8 |
126.3 |
226.2 |
948.8 |
904.8 |
Run 6 |
13.32 |
0.172 |
1.67 |
73.11 |
0.566 |
10.80 |
767.3 |
0.186 |
541.4 |
183.0 |
114.6 |
59.5 |
14.7 |
115.2 |
125.5 |
226.3 |
960.5 |
904.6 |
Run 7 |
13.33 |
0.185 |
1.67 |
73.06 |
0.554 |
10.79 |
758.9 |
0.180 |
570.3 |
193.3 |
101.8 |
67.9 |
6.1 |
113.5 |
124.9 |
226.9 |
960.6 |
910.1 |
Run 8 |
13.26 |
0.185 |
1.69 |
73.04 |
0.555 |
10.78 |
771.7 |
0.185 |
565.2 |
191.8 |
95.2 |
66.9 |
9.3 |
101.6 |
124.7 |
227.4 |
918.0 |
919.7 |
Run 9 |
13.33 |
0.180 |
1.64 |
73.11 |
0.561 |
10.82 |
775.7 |
0.183 |
553.7 |
219.7 |
113.4 |
61.5 |
9.3 |
109.6 |
124.6 |
222.9 |
916.9 |
913.4 |
Run 10 |
13.30 |
0.193 |
1.68 |
73.08 |
0.556 |
10.80 |
764.1 |
0.188 |
572.6 |
243.2 |
101.3 |
64.4 |
1.0 |
103.4 |
123.0 |
226.0 |
980.8 |
912.7 |
Run 11 |
13.31 |
0.184 |
1.66 |
73.06 |
0.561 |
10.78 |
785.3 |
0.186 |
566.0 |
191.8 |
109.0 |
70.8 |
13.8 |
106.9 |
122.1 |
227.7 |
950.9 |
875.6 |
Avg. |
13.31 |
0.183 |
1.67 |
73.08 |
0.558 |
10.80 |
770.5 |
0.183 |
566.4 |
192.8 |
105.8 |
65.8 |
9.7 |
109.9 |
123.8 |
226.1 |
936.2 |
905.7 |
Std.Dev. |
0.03 |
0.01 |
0.01 |
0.03 |
0.004 |
0.02 |
10 |
0.003 |
15 |
24.2 |
7.2 |
4.3 |
4.9 |
5.4 |
2.2 |
1.4 |
31.2 |
12 |
Conclusion
All limits of detection obtained show that the ARL OPTIM’X can offer adequate analysis results, notably for bottle glass application. The repeatability of the analysis is high for major and minor elements, even for Na2O and MgO. Longer counting time may be utilized in case elements present below 100 ppm need to be controlled precisely. These results show that the ARL OPTIM’X spectrometer is well suited to produce precision results for the determination of the main oxides and the coloring agents in glasses.

This information has been sourced, reviewed and adapted from materials provided by Thermo Fisher Scientific - Elemental Analyzers.
For more information on this source, please visit Thermo Fisher Scientific - Elemental Analyzers.