Nuclear Instruments and Methods in Physics Research B47 (1990) 143-147 North-Holland

143

RADIATION DAMAGE IN PIXE ANALYSIS OF MUSEUM PAPER-LIKE OBJECTS *

Xianzhou ZENG, Xiankang WU, Qiyun SHAO, Jiayong TANG and Fujia YANG

Nuclear Science Department, F&m University, Shanghai 200433, PR China

Received 15 June 1989 and in revised form 22 November 1989

Paper discoloration and its tensile strength degradation caused by 2.1 MeV external proton beam bombardment under low beam current density ( ~10 nA/cm’) were measured as a function of the number of incident protons. The target temperature reached

during bombardment was measured using an infrared thermometer. Under low beam current density, paper heating is negligible and the paper deterioration depends only on the total beam fluence. Some of the experimental results are presented and discussed for several types of paper, including one used for traditional Chinese paintings.

1. Introduction

Particle induced X-ray emission (PIXE), as a useful analytical technique, has been widely applied to art and archaeology [l-5]. PIXE analysis can provide informa- tion regarding the techniques employed in ancient times, the origin and the authenticity of museum objects. Trace element analysis of documents, stamps, paintings and other paper-like samples by PIXE has attracted much attention in recent years [6-121. The possible paper deterioration caused by proton bombardment is still one of the important problems to be solved.

The target temperature reached during irradiation is determined by the energy balance between the energy deposited by the incident beam and the heat removal. Since paper is a delicate material and is subject to damage, a temperature rise in paper may cause severe discoloration and sometimes may lead to burning through, even in the external beam case. In order to minimize the temperature rise and the paper damage thus caused, a low beam current density is desired. This is accomplished by a very tight geometry, forced con- vection cooling, and on-demand beam pulsing [l&12, 141.

One may ask what. is to be expected regarding paper damage at low beam current density. There is no guarantee that paper damage will still not be seen, since there are several additional possible damage mecha- nisms [13] (e.g. sputtering, element migration, atomic displacements) besides heating during irradiation. In fact, many paper analysis experiments show that even in

* This work was supported by National Natural Science Foundation of China under grant no. 18 60 122.

the case of low beam current density, where the temper- ature rise was negligible, paper damage did occur: visi- ble color changes, obvious fragility, etc. These effects were seen as long as the total number of incident protons was large enough. This kind of damage, caused by ion beam bombardment, is cumulative in effect and should also be treated with care.

In this study, some results of paper deterioration measured as a function of the beam fluence under low beam current density are reported for several types of paper, including one used for traditional Chinese paint- ings.

2. Experiment

2. I. Proton beam

In the present work, a 2.6 MeV proton beam from the Van de Graaff accelerator at Fudan University was used. The proton beam was deflected by an analyzing magnet into a 90” exit port and passed through two tantalum collimators to the external PIXE system with on-demand beam pulsing [12]. In order to obtain a homogeneous proton beam spot with adjustable size, an alurninium diffusion foil and several. selective collima- tors were used. Therefore, a proton beam size ranging from a diameter of 0.1 mm to 6.0 mm was available. In the present work, a beam size of 6.0 mm was chosen. The proton beam was then extracted through a 7.5 urn Kapton window, which was replaced every 24 h to avoid breakdown due to radiation damage. The proton beam traveled 3 mm in air to reach the paper sample. The proton energy at the paper was about 2.1 MeV. A

0168-583X/90/$03.50 0 Elscvier Science Publishers B.V. (North-Holland)

144 Xianzhmc Zeng et al. / PIXE radiation &n&age in paper

Faraday cup, located at the rear of the paper, was used to monitor the external proton beam.

2.2. Measurement of paper temperature

Since paper is not a thermally conducting material and any elevated temperature vanishes only a few minutes after the irradiation, the method used to mea- sure the paper temperature reached during irradiation should be noncontactive, sensitive, and on-line, i.e. it should be able to indicate the temperature dut-lng irradi- ation.

In the present work, a digital infrared thermometer ND400 was used, manufactured by the Shanghai In- stitute of Technical Physics, Academia Sinica [15]. Its probe contains a collimator (4 mm diameter) which ensures that the detector receives thermal radiation only from the region of interest of the sample irradiated by the proton beam (6 mm diameter). The emissivity of the paper surface should be determined separately in ad- VtIXW3.

2.3. vehement of paper deterioration

Bombarding paper by an energetic proton beam may cause the breakage of molecular chains, leading to color changes and mechanical performance degradation of the paper. Therefore, the change in brightness of white paper reflects the damage caused to the paper and can serve as a measurable macroparameter.

The paper brightness was measured using a digital brightness tester SBD-1.’ In order to avoid the difficul- ties caused by the discrepancies in brightness among the paper samples, a self-comparison method was used, in which each paper sample was measured twice before and after the bombardment at the same spot on the paper specimen. The brightness change between the measurements reflects the paper discoloration pro- duced.

The decrease in paper tensile strength, which serves as another measurable macroparameter of the paper damage, can also be obtained using a quasi-self-com- parison method. In doing so, two measurements of paper tensile strength were performed for two strips cut from a paper specimen in the direction of the fiber near each other. One strip served as a specimen to be bombarded while another as a reference. Each strip was 6 mm wide and 100 mm long, which was pulled in a Ion~tudu~ direction to measure its strength. To calcu- late the tensile strength (in kN/m) a correction factor for strip width was taken into consideration. The ratio of the two measured strengths was called the relative tensile strength (in W) of the paper. A tensile tester manufactured by Lorentzen and Wettre Co., Sweden, was used for these measurements.

2.4. Mann- W&itney test

A statistical method for significance testing, named the Mann-Whitney test [16], was performed in the present work.

Suppose there are two sets of random samples, one from group I and another from group 2, and the null hypothesis is He: the two groups are identical. To compute the observed value of the test statistic, we combine the two set samples and rank all sample ob- servations from smallest to largest. The test statistic is

T= S, - fnl(nl + I),

where St is the sum of the ranks assigned to the sample observations from group 1. Therefore we reject H, if the computed value of T is less than Wa,2 or greater

than K--cr,ZI where a is the significance level, W& is the critical value of T given in ref. [16], and Wl_-01,2 is given by

K-42 = 111112 - K/2*

where at and n2 are the sample size of group 1 and group 2, respectively. Otherwise the null hypothesis is accepted at a significance level of a.

3. Results and discussion

Three types of paper were used in the present work. Xuan paper (- 28 g/m’), composed mainly of pure natural cellulose, has been used for traditional Chinese paintings and calligraphy for some thousands of years. Art paper (90-105 g/m’) contains a large amount of chemicals and is normally used for pictures and mag- azines. White tissue (- 11 g/m2) contains both natural cellulose and a small amount of chemicals.

3. I. Paper temperature rise and heat damage

Under the conditions of switching off both the on- demand beam pulsing and the forced convective cool- ing, the Xuan paper sample was bombarded with a 2.1 MeV external proton beam. The temperature rise AT reached its maximum value AT_ (O C) in 30-40 s, which varied with beam current density J (nA/cm2) as shown in fig. 1. By fitting these experimental data, the following formula was obtained for the range 5 to 30 nA/cm2

AT_ = -0.17 -I- 0.164 J.

It shows that the temperature rise is less than 5 o C when the beam current density is below 30 nA/&.

In order to further ensure that the degradation of paper tensile strength under constant beam fluence was not associated with beam current density as well as the temperature rise, the Mann-Whitney test was per-

Xianzhou Zeng et al. / PIXE radiation damage in paper 145

0 15 30

CURRENT DENSITY C nW=m’ B

Fig. 1. Effect of beam current density on paper temperature (external proton heam energy: 2.1 MeV).

formed for three groups (A, B, C) of Xuan paper bombarded by 3.5 pC/cm’ protons at beam current densities of 0.28 nA/cm2, 1.0 nA/c& and 5.4 nA/cm2, respectively. The results of the Mann-Whitney test between every two out of the three groups are sum- marized in table 1. The average value of the tensile strength of the Xuan paper samples before irradiation was 1.03 f 0.12 kN/m. Table 1 shows that the degrada- tion of paper tensile strength caused at beam current densities from 0.28 nA/cma through to 5.4 nA/cm2 was statistically identical.

To investigate the effect of paper temperature exclu- sively on its tensile strength, an aging test of Xuan paper was also performed. Two groups (D, E) of paper, each containing 10 samples, were or were not kept in an oven at 105OC for 72 h, respectively, without having been irradiated by the proton beam. These paper sam- ples were then measured to obtain their tensile strength. The Mann-Whitney test between the groups D and E was performed and the results are summarized in table 2. The measured tensile strengths of the groups D and E were statistically identical, indicating that a temperature as high as 105“C did not cause an obvious decrease in tensile strength of the Xuan paper.

These experimental data imply that under low beam current density ( < 10 nA/cm2) paper deterioration de- pends only on the beam fluence as the heating damage in this case is negligible.

3.2. Radiation damage under low beam current density

The degradation of paper tensile strength and paper discoloration before and after proton bombardment were measured as a function of the total beam fluence. The results obtained are shown in fig. 2 through to fig. 4. Among the three types of paper studied, Xuan paper could withstand bombardment to the largest proton beam fluence before deterioration and Art paper the least. This is not surprising, since Xuan paper contains the greatest content of natural cellulose while Art paper contains the least, and the artificial chemicals are widely known to be more subject to change than natural materials.

Table 1 Mann-Whitney test of Xuan paper for different beam current densities for a beam fluence of 3.5 PC/C& (external proton beam energy: 2.1 MeV)

Group Beam current Relative tensile strength [%] 6.025 w,-0.025 T Comment density J

bVm21

A 0.28 37 42 36 46 44 I TAB= 8 identical B 1.0 44 43 48 47 36 3 22 Tat = 18 identical C 5.4 32 42 37 35 53 Tc- =16 identical

Table 2 Mann-Whitney test of Xuan paper for aging effect, for control samples (D) and samples kept in an oven at 105OC for 72 h (E)

Group Time [hl

Tensile strength of paper [kN/m] w, - 0.025 T Comment

D 0 0.80 0.91 0.91 1.0 1.11

E 72 0.97 1.1 1.1 1.3 0.90 0.95 0.96 1.1

1.1 1.2 1.2 1.2

1.3 1.1

1.3

24 76 44 identical

146 Xianzhou Zeng et al. / PIXE radiation damage in paper

The results obtained imply that even under low beam current density there is a maximum limitation of beam fluence for each type of paper, above which the proton bombardment may cause significant paper de- terioration. For Xuan paper, the maximum beam fluence is about 0.4-0.6 pC/cm* in the case of 2.1 MeV pro- tons.

In our laboratory some valuable ancient Chinese paintings composed of Xuan paper were analyzed by using external beam PIXE technique. To avoid any

100

75

50

c

, ,L 0 a o,o

1tY2 lo-’

b-

i-

,-

I

10 2

BEAM CHARGE C ~.~C/crn~ 3

b I I I

1 o-2 lo-* 10 o 10 * 10 2

BEAM CHARGE C /.Ucm= 3 Fig. 2. Effect of beam fhnce on Xuan paper deterioration (external proton beam energy: 2.1 MeV). (a) Relative tensile strength of Xuan paper; and (b) relative brightness of Xuan

paper.

1 ci2 id’ 10 O 10 i 10 2

a 0 I I I

, -

I-

I

BEAM CHARGE C j.Xktn2 1

3 I I I

Id2 id’ 10 O 10 1 10 2

BEAM CHARGE C /K/cm= II

Fig. 3. Effect of beam fluence on art paper deterioration (external proton beam energy: 2.1 MeV). (a) Relative tensile strength of Art paper; and (b) relative brightness of Art paper.

damage in the analysis, a beam current density of 0.2-0.6 nA/cm* and a beam fluence of 0.05-0.1 uC/cm* for each painting were used. We have not observed any damage to the paintings since the analysis began two years ago.

4. Conclusions

For PIXE analysis of precious irreplaceable paper- like objects, both low beam current density and low

Xianzhou Zeng et al. / PIXE radiotim damage in paper 147

100

n

x 75 ”

ttf

5 Id 50 t-

‘;!

5 25

i!

0 IO-- 1 ci-’ 10 O 10 = 10 =

3EAM CHARGE C &+_xn2 3

Fig. 4. Effect of beam Buence on the tensile strength of white tissue (externaI proton beam energy: 2.1 MeV).

total beam fluence are important. Under low beam current density (< 10 nA/cm’) the paper deterioration caused by external proton beam bombardment depends only on beam fluence. For each type of paper there is a maximum limitation of beam fluence, above which sig- nificant paper damage may occur. We suspect that the paper damage effects observed experimentally in this study may be due in huge part to atomic displacement damage.

The authors wish to thank Dr. Zhou Zhenzhen of Shanghai Paper Research Institute for her help with

measuring paper samples and Dr. Zhang Caigen of Shanghai Institute of Technical Physics, Academia Sinica, for his help with measuring paper temperature.

References

PI

I21

131

141

151

PI

[71

RI

[91

WI WI

WI

iI31 [I41

WI WI

F. Yang, Atomic Physics (Shanghai Science and Technol- ogy, Shanghai, 1985). JR. Bird, P. Dnerden and D.J. Wilson, Nucl. Sci. Appl. Bl(5) (1983) 357. G. Arnsel, Ch. Heitz and M. Menu, Nucl. Instr. and Meth. B14 (1986) 30. T.A. CahiII, B. Kusko and RN. Schwab, Nucl. Instr. and Meth. 181 (1981) 205. R.A. Ekired, B.H. Kusko and T.A. Cahill, Nucl. Ins&. and Meth. B3 (1984) 579. J. Wisnovsky, Time, 10 March 1986, p. 75. B.H. Kusko and R.N. Schwab, Nucl. Instr. and Meth. B22 (1987) 401. B. Nens, P. TroceIlier, Ch. Engehnann and Ch. Lahanier, Nucl. In&r. and Meth. B14 (1986) 148. E.M. Johansson, S.A.E. Johansson, K.G. Mabnqvist and I.M.B. Wiman, Nucl. Instr. and Meth. B14 (1986) 45. F. Stross, The Vortex, 69 (1988) no. 1, p. 6. T.A. CahiII, D.W. McColm and B.H. Kusko, Nucl. Instr. and Meth. 814 (1986) 38. X. Zing, X. Wu, Y. Zong, D. Wci, Q. Shao and H. Yao, Vacuum 39 (2-4) (1989) 91. J.A. Cookson, NucI. Instr. and Meth. B30 (1988) 324. X. Zeng and X. Li, NucI. In&r. and Meth. B22 (1987) 99. C. Zhang, Acta Physica Sinica 31(9) (1982) 1191. W. W. Daniel, Applied Nonparametric Statistics (Hougb- ton Miffhn, Boston, 1978).