CS8:. SURFACE Cleaning of Ancient Glass by Plasma Generated in Water Solution

The application of plasma in combination with liquids for the surface cleaning of ancient glass is a new perspective direction in this material conservation/restoration. Currently, there are only a few studies available in this field (Hlochova et al., 2016). This study shows the first part of the results obtained by our research team using a specially designed electrode system for plasma generation in liquids (Krcma, 2015, Krcma, 2019).

The glass sample was discovered during the emergency excavation in the field of former Vlnena Company in Brno (Czech Republic, 49.18876 N, 16.61840E) in 2017. The whole artifact is shown in Figure 8.58.

Glass sample used for the presented study

FIGURE 8.58 Glass sample used for the presented study.

According to the excavation stratigraphy, this sample was made during the end of the eighteenth or the beginning of the nineteenth century. One-half of the sample was cut into six parts (each of them with an area of about 2 cm2) that were used for the current study; the second half was kept for another experiment.

The plasma treatment was carried out according to the scheme shown in Figure 8.59. The specially designed handy movable electrode system (shown in Figure 8.60. practical realization is demonstrated in Figure 8.61) was immersed into the K1CO3 water solution in the glass vessel with dimensions 16 x 10 x 10.2 cm3. The grounded polished stainless steel electrode was installed at the glass vessel wall. The ground from the specially designed laboratory power supply (Lifetech, 700 W) was connected through the capacitance C = 961 nF that was used together with the

Scheme of the experimental setup

FIGURE 8.59 Scheme of the experimental setup. 1 - audio-frequency power supply. 2 - high voltage transformer, 3 - capacitor, 4 - movable high voltage electrode system (see Figs. 8.60 and 8.61), 5 - grounded stainless steel electrode, 6 - plasma. 7 - treated sample, 8 - glass vessel with water solution.

Scheme of the used electrode system and its realization

FIGURE 8.60 Scheme of the used electrode system and its realization. 1 - head of the electrode system with the pin-hole, 2 - tungsten wire (diameter of 0.5 mm). 3 - Quartz capillary, 4 - open space filled by liquid. 5 - silicone gland. 6 - plasma in the gaseous bubble.

Used electrode system (left) and generated plasma in water solution (right)

FIGURE 8.61 Used electrode system (left) and generated plasma in water solution (right).

high voltage probe (Tektronix P6015A) for the measurement of energy dissipated into the electrode system. The evaluation technique based on Lissajous pictures described for dielectric barrier discharges (Wagner et al, 2003) and later adapted for high- frequency discharges in liquids (Krcma et al. 2015) was used for the energy calculation. The energy calculation is based on the chart of both high voltage and voltage at known ballast capacitance C as it is shown in Figure 8.62. The area is approximated by the parallelogram with area A. Energy P is calculated as

where/is frequency directly measured from the applied voltage record.

In order to check if the distance between the grounded and the movable high voltage electrode system influences the dissipated energy, the set of measurements at distances of 2.5, 7.5, 9, and 13 cm was carried out in NaCl solution with the conductivity of 205 pS/cm; the operating power supply frequency was (16780 ± 30) Hz. As the discharge character is random (see later), each experiment was repeated five times. The calculated energy was 24 W at all distances with the uncertainty around 1 W.

Chart for the dissipated energy calculation. Points represent the measured data

FIGURE 8.62 Chart for the dissipated energy calculation. Points represent the measured data.

The discharge itself is created in microbubbles according to the bubble theory of the discharge ignition (see Chapter 3.4). The bubbles are dynamic as it is demonstrated in Figure 8.63. These images were taken from the i-SPEED ultrafast camera movie (objective NIKON AD, AF MICRO NIKKOR 200 mm with aperture f/16). The system was illuminated by the high illuminance reflector with the lens of 15°. The acquisition rate was 10,000 frames per second, and single image exposure time was 0.01 ms. It is well visible from Figure 8.63 that the discharge is operating inside the bubble that propagates into the capillary as well as into the open solution. This ensures the direct contact of plasma with the treated surface. The bubble cavitation is a very fast process that leads to the generation of mechanic waves (for a better demonstration of shock waves, see Krcma et al., 2018).

The presented experiment shows the surface treatment dependence on the solution conductivity. The surface corrosion layers and other surface contamination should be mainly oxidized and consequently dissolved in the solution. The most common salt used for the solution conductivity adjustment is NaCl. But atomic chlorine that is produced by electrolysis at the anode is strongly oxidative and can damage the glass surface if the anode is close to the surface. Due to this fact, we chose less reactive inorganic salt K1CO3 for the solution conductivity adjustment. Moreover, this compound was frequently used during history as an additive in glass production (Eitel, 1975. Fanderlik, 1991). The solution conductivity was changed in an interval of 400-900 pS/cm. The discharge was operating at the fixed frequency of (16,780 ±30) Hz, dissipated power was (23.0 ± 1.5) W. The area of approximately 10 mm in diameter was treated for 10 min. The high voltage electrode system was hand-moved about 1 mm above the surface to ensure good contact between the plasma and the sample surface.

Series of discharge images showing discharge bubble formation and propagation

FIGURE 8.63 Series of discharge images showing discharge bubble formation and propagation.

Surface analyses of samples were carried out using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The excimer laser ablation system Analyte Exite + (Teledyne, Photon machine) with an ArF* excimer laser operating at 193 nm is equipped with a tunable two-volume ablation cell HelEx II enabling fast transport of an ablated material into the ionization source. Ablation was conducted in the He atmosphere, set to a constant flow of 0.8 L.min'1 in total. The carrier gas was mixed with Ar gas (1 L.min'1) prior to the ICP source of a quadrupole ICP-MS Agilent 7900 (Agilent technologies). ICP ionization source was operating at RF power of 1550 W and argon flow of 15 L.min'1. Optimization of LA-ICP-MS parameters was performed with the glass reference material NIST

SRM 612 with respect to the maximum signal sensitivity, and minimal doubly charged ions and oxides formation. The potential interferences were minimized via a collision cell (He 1 mL.min'1).

Signal of isotopes of interest was investigated by line scanning (see Figure 8.64) using a laser beam spot diameter of 150 pm (square) within the sample area 150 x 150 pm2 (variable related to the area of interest), approximately. The scan speed of 50 pm.s'1, fluence of 3 J.cm'" and frequency of 10 Hz were kept constant.

The NIST 612 and 610 standards were used for quantification. Preselected elements (see later in Table 8.9) were finally recalculated to their oxides and represent 99% of the whole sample, the rest 1% belongs to elements such as hydrogen, barium, cobalt, fluorine, and lanthanides.

Sample 1 was treated in the solution with initial conductivity of 400 pS/cm. Conductivity was increased to 405 pS/cm, only. The samples before and after the treatment are shown in Figure 8.65.

The surface cleaning was visible since the fifth minute, no remarkable changes were visible at the treatment beginning. The treatment was not so effective (see analyses in Table 8.9), this also reflects only a small increase in solution conductivity.

Sample 2 was treated in the solution with initial conductivity of 500 pS/cm. Conductivity was increased to 516 pS/cm after the treatment. The samples before and after the treatment are shown in Figure 8.66.

The surface cleaning was visible since the third minute. Two separate surface analyses were carried out. The first one was completed at the clean glass surface (analysis 2-g), the second one was done at the surface that was covered by a white layer after the treatment (analysis 2-w). The first analyzed part showed better cleaning than in the first sample. The second one showed a completely different surface composition and we suppose that this part of the sample was originally covered by some color decor.

Laser ablation scan over one of the analyzed samples. The distance between the neighbor lines is 15 pm (optical microscope LCD MIKRO 5MP, BRESSER)

FIGURE 8.64 Laser ablation scan over one of the analyzed samples. The distance between the neighbor lines is 15 pm (optical microscope LCD MIKRO 5MP, BRESSER).

TABLE 8.9

Surface composition of samples

mass.%

Li2O-10'3

B2O3

С02 10'3

Na20

MgO

AI2O3

Si02

1

2.11

0.034

0.76

0.219

1.47

4.02

81.82

2-g

2.54

0.045

0.99

0.285

2.04

3.76

78.00

2-w

0.02

0.005

0.75

0.040

0.18

5.37

88.00

3

2.87

0.041

0.70

0.304

2.23

4.77

74.95

4

0.03

0.019

1.75

0.196

0.28

7.07

84.88

5

2.66

0.045

1.41

0.294

2.20

4.38

76.90

reference

bulk

0.09

0.075

1.97

0.800

0.27

5.80

88.84

mass.%

P2O5

SO.,

Cl

k2o

CaOlO-'

ТЮ2

Cr2O3-10‘3

1

0.86

0.31

0.029

9.23

1.13

0.272

3.65

2-g

0.98

0.43

0.016

12.49

1.43

0.178

1.62

2-w

0.08

0.23

0.061

4.37

0.18

0.154

2.01

3

0.92

0.65

0.034

13.96

1.51

0.131

1.52

4

0.10

1.03

0.068

3.90

0.78

0.186

2.74

5

0.89

0.36

0.033

12.93

1.50

0.176

1.63

reference

bulk

0.18

0.58

0.037

1.01

0.21

0.213

4.03

mass.%

Mn02

FeO

NiOlO-'

CuOlO-2

ZnO

M0O3IO-5

1

0.22

0.41

1.59

1.61

0.0800

7.51

2-g

0.29

0.43

1.76

1.18

0.0490

8.43

2-w

0.01

0.44

1.35

2.96

0.0264

1.82

3

0.31

0.63

3.72

2.65

0.0403

7.73

4

0.06

1.04

7.90

9.97

0.0589

3.66

5

0.30

0.43

1.68

1.05

0.0381

7.98

reference

bulk

0.01

1.06

5.46

4.72

0.0741

5.41

Sample 3 was treated in the solution with initial conductivity of 600 pS/cm. Conductivity was increased to 635 pS/cm after the treatment. The samples before and after the treatment are shown in Figure 8.67.

The surface cleaning was visible since the fourth minute. Cleaning efficiency was well visible by the naked eye and this result was also confirmed by the elementary analysis (see Table 8.9). Solution after the treatment was also exposed to the additional discharge (using the same settings as during the treatment) and optical emission spectrum was collected (see later).

Glass sample treated in the solution with initial conductivity of 400 pS/cm. The treated area is marked by a circle

FIGURE 8.65 Glass sample treated in the solution with initial conductivity of 400 pS/cm. The treated area is marked by a circle.

Glass sample treated in the solution with initial conductivity of 500 pS/cm. The treated area is marked by a circle

FIGURE 8.66 Glass sample treated in the solution with initial conductivity of 500 pS/cm. The treated area is marked by a circle.

Sample 4 was treated in the solution with initial conductivity of 800 pS/cm. Conductivity was increased to 839 pS/cm after the treatment. The samples before and after the treatment are shown in Figure 8.68.

The surface cleaning was visible after the first minute. The corrosion products were removed from the surface after the first minute of the treatment. This sample was rather inhomogeneous and thus the results of the surface analysis are significantly different from the others.

The last sample, sample 5, was treated in the solution w'ith initial conductivity of 900 pS/cm. Conductivity was increased to 962 pS/cm after the treatment. The samples before and after the treatment are shown in Figure 8.69.

The surface cleaning was visible just after the beginning. The surface cleaning was generally good as it is visible from the picture as well as from the analysis.

Glass sample treated in the solution with initial conductivity of 600 gS/cm. The treated area is marked by a circle

FIGURE 8.67 Glass sample treated in the solution with initial conductivity of 600 gS/cm. The treated area is marked by a circle.

Glass sample treated in the solution with initial conductivity of 800 gS/cm. The treated area is marked by a circle

FIGURE 8.68 Glass sample treated in the solution with initial conductivity of 800 gS/cm. The treated area is marked by a circle.

Glass sample treated in the solution with initial conductivity of 900 gS/cm. The treated area is marked by a circle

FIGURE 8.69 Glass sample treated in the solution with initial conductivity of 900 gS/cm. The treated area is marked by a circle.

The discharge in the solution with initial conductivity of 600 pS/cm was also studied by optical emission spectrometry. The emitted light was led to the multi- mode optical fiber using the quartz lens with a focal length of 100 mm, fiber input was at the focus, i.e., the whole discharge volume was analyzed. The other end of the fiber was mounted to the Jobin Yvon Triax 550 spectrometer with the liquid nitrogen cooled CCD detector (back-illuminated. 1026 x 256 pixels, UV enhanced). Emission spectra were recorded using the 1200 gr/mm ruled diffraction grating with the entrance slit of 0.03 mm and integration time of 30 s that ensured elimination of the discharge instabilities.

The spectra are shown in Figure 8.70. The characteristic of species for the discharges in water solutions such as molecular band of OH radical and hydrogen and oxygen atomic lines is well visible in both spectra. Also, the potassium lines are presented. The presence of sodium lines at 588 and 589 nm is probably due to the reactor contamination from another experiment where NaCl was used for the conductivity adjustment. The spectrum of the solution after the sample treatment also contains a few lines that are probably titanium. Comparing this result with the surface composition shown in Table 8.9, this is very probable because titanium was significantly removed from the sample surface. The presence of other atomic lines related to the surface contaminants was not detected at this pilot experiment. The full used solution examination will be performed by the ICP-OES technique that allows the determination of nearly all periodic table elements with the appropriate sensitivity.

Based on the experimental results, we can conclude that the application of the electrical discharge generated directly in the liquid can be a promising way for the surface cleaning of ancient glass. The equipment is not so complicated (and thus not so expensive, too) and good cleaning results can be achieved without any damage to the fragile ancient object. The surface analyses showed successful removal of elements presented at the surface from the surrounding environment and the increase of compounds related to the bulk material. The best results were achieved at solution conductivity in the range of 600-900 pS/cm. The treatment at lower conductivities is not so efficient because of lower plasma reactivity. On the

Optical emission spectra of the discharge in fresh (left) and used (right) solution with initial conductivity of 600 pS/cm

FIGURE 8.70 Optical emission spectra of the discharge in fresh (left) and used (right) solution with initial conductivity of 600 pS/cm.

contrary, higher solution conductivities can lead to stronger shock wave generation

that can damage the treated object.

The separate studies are prepared for the application of other water solutions as

well as the determination of effects of different applied powers and treatment times.

 
Source
< Prev   CONTENTS   Source   Next >