[PDF] - Measurements on radioactivity of ancient roman lead to be used as PDF Document

SLIDE 1

106

Nuclear Instruments and Methods in Physics Research B61(1991) 106-117

North-Holland

Measurements on radioactivity

f ancient roman lead to be used as

shield in searches for rare events

A. Alessandrello a, C. Cattadori a, G. Fiorentini b, E. Fiorini ‘, G. Gervasio ‘, G. Heusser d,
G. Mezzorani b, E. Pernicka d, P. Quarati b, D. Salvi e, P. Sverzellati ’ and L. Zanotti ’

’ Loboratori Nazionaii de1

Gran Sa~so dell’fNFN, Assergi, L’Aquila, Ita& ’ ~ipariime~to di Fisica de~~u~ivers~t~ di Cagliari e Sezione INFN di CagEor& Cagliari, Ita@ ’ ~ipartime~to di Fisica del~~~iversit~ di ~iiano e Sezione INFN di Milana, M~lono~ Itaiy d Max Plunck Institut jiir Kern.5h~sik, Heidelberg, Germany e Soprintendenta Archeologica per le provincie di Cagliari e Oristano, Cugliari, Italy

Received 14 January 1991 and in revised form 22 March 1991

An ingot from the load of lead carried by a Roman ship sunken near Sardinia has been anaiysed for its composition and purity from radioactive contaminants in view of its possible use to shield low level counting experiments. Measurements canied out with X-ray fluorescence, X-ray diffraction and neutron activation show that the lead was desilvered, a beneficial procedure as far as the elimination of radioactive contaminants is concerned. Measurements of gamma and X-rays and of alpha particles have been carried

ut on a substantial part of the ingot and, for comparison, on samples of modern electrolytic and of specially prepared lead and of

lead from a presumably 500 year old Dutch ship. Co~ta~nations yielding high energy gamma rays have been found to be negligible in all sampIes. The analysis of gamma rays, X-rays, atpha particles and bremsstrahlung spectra on the contrary have consistently shown that the modem lead contains a considerable amount of “‘Pb, which is totally absent in Roman lead. This metal, together with the special low radioactivity lead, is by far the best shielding material among the samples investigated in this study. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

1. Introduction

Recent developments of sub~uclear physics and the so-called particle physics-astrophysics connection fl,2] have stimulated interest in “passive” experiments, car- ried out without accelerators. Most of these are based

n the search for very rare events which produce in

suitable counters signals corresponding to delivered en- ergies ranging from fractions of MeV up to a few MeV. We would like to quote as examples the interactions

f

solar neutrinos [3], or some rare process like double beta decay and the possible decay of the electron [4,5]. In all these experiments it is essential to strongly sup- press the background of spurious counts of environmen- tal origin like charged cosmic rays, neutrons and gamma rays. The background

f charged cosmic rays is strongly

reduced if the laboratory is located deep underground. In the Laboratori Nazionali de1 Gran Sasso (LNGS) [6], located at a depth of about 3500 meters of standard rock, where most of the measu~ments reported here have been carried out, the intensity of charged cosmic rays is reduced by a factor of about 10e6 with respect to the flux on the surface [7]. The reduction of neutrons is also considerable, but not so relevant since only a part of them is produced by cosmic rays, the remaining

nes being generated

by spontaneous fission in the

rocks. In the Gran Sasso Laboratory the radioactivity of

the rocks is rather low f8] and therefore the fluxes of thermal and fast neutrons are three to four orders of magnitude smaller [8,9] than in any low background laboratory

n the surface. In an underground

labora- tory the background

f gamma rays is an the contrary

similar to or even higher than in a surface laboratory due to the radioactivity

f the surrounding

rocks or of the construction materials. Reduction of gamma rays, which is essential in the above mentioned experiments, can only be accom- plished by shielding the detector with suitable materials

f high atomic number and low intrinsic activity [lO,ll].

A s~elding material commonly adopted at least for the internal “core” of the shield is oxygen free high conduc- tivity (OFHC) copper, a metal normally used in elec- tronics where the process to eliminate

xygen also

strongly reduces radioactive contaminants. Copper is, however, rather expensive and has a low atomic number and high cross section for capture of thermal neutrons and for the cosmogenic production of radioactive nuclei 1121. Apart from its rather large thermal neutron cross section, mercury would constitute an excellent shielding

0168-583X/91/$03.50 0 1991 - Elsevier Science Publishers B.V. (North-Holland)

SLIDE 2

A. Ales~andr~i~o et at. / Radi~cfivit~~ of ancient reman lead

107 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

material [13], since it can be repeatedly purified by

distillation. It is, however, rather expensive and requires

suitable containers. Due to its reasonable cost and mechanical properties, low neutron cross section and high atomic number, lead represents an ideal material for shiefding gamma rays. For radiation protection pur- poses its intrinsic ra~oacti~ty is negligible, but for highly sensitive low radioactivity experiments its resid- ual activity may limit its usefulness. The nature and

rigin of this ~onta~nation

has been thoroughly studied by many authors and is partly reviewed 114-191. The

verall consistent result was that only “‘Pb

with its daughter nuclides 210Bi and “‘PO is responsible for the intrinsic radioactivity of lead. Due to the long half-life

f *l’Pb (22 years) the abundance of these nuclides in

the lead could be much higher than ex acted li from secular eq~lib~~ [ZO]. The presence of “Bi and its daughters has been demonstrated by Welier et al. [15]

n the basis of alpha, beta and gamma correlations. The

very soft beta (E,,,,

f 16.5 and 63 kev) and gamma

(46.5 kev) radiation of “‘Pb hardly escapes seif-ab- sorption, but the energetic (1.16 MeV) beta rays of 21*Bi induce bremsstrahlung and characteristic X-rays in lead. The bremsstr~ung-continuum peaks at about 170 keV, while the lead X-rays have energies of 72.8, 75.0, 84.9 and 87.4 keV. It has been recently discovered [11,21,22] that 210Po, the alpha-decaying member of this chain with an energy of 5.30 MeV, diffuses to the surface of freshly scraped lead, resulting in an enhancement of the surface activity. Con~ntrations

f 2*oPb reported in the

literature range from detection limit of the respective detector up to 2500 Bq/ kg [23]. In lead from solder, where “Pb can be also introduced by tin 112,241, this cont~nation can be as high as 50000 Bq/ kg [24]. We would like to point out that, apart from 2’oPb and its daughters, no U or Th cont~nations have been reported so far for commercial lead. Since the upper ~nta~nation limits found [12] for 08Tl and *14Bi are as low as 0.33 mBq/ kg and 0.67 mBq/ kg, respectively, the corresponding concentrations

f Th

and U are expected to be less than 8.2 X zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA lo-” and 5.4 X IO-” g/ g. Probable ~o~t~nation path of “‘Pb is through the uranium-untying minerals accompany- ing or added to the lead ores [16,22]. If they are not completely separated at the beginning of the lead pro- duction, their accumulated “‘Pb is taken up by the lead bullion during smelting. Othe~,~athways are: added scrap lead, already wnt~~ng Pb, or coal added in the reduction process. Coal normally contains Urania in large quantities from which “‘Pb is transferred to the smelting charge. Reduction processes without coal, like electrically heated furnaces, avoid this way of wn-

tamination. Today, certified lead with a specific ‘lOPb

activity of less than 50 Bq/ kg is w~erci~iy available (e.g. from Boliden, Sweden) f19] at a price which is just about twice that for regular lead. Another wmpany (J

hnson

and Matthey) succeeded in producing lead with at least 20 times lower specific radioactivity 1251 by

ptimizing each step of the lead production

process using control me~urements

f alpha counting and,

presumably, also ore selection. This lead, also included in this investigation, is mainly produced for the elec- tronics industry and by far too expensive to be used in large quantities. An alternative is old lead produced several half-lives of 210 Pb ago. The av~lability of such lead is, however, very limited, since “sources of supply” like water pipes older than 200 years, sunken shiploads,

r ballast of sailing ships are not found frequently. The

age and the water protection of the Roman lead dis- cussed here is such to prevent visible conta~nations not only of “‘Pb, but also possibly of other long lived ~osmogenically produced radioactive cont~n~ts. The present search has been stimulated by the recent discovery of a sunken Roman commercial ship (navis

neraria) near Sardinia carrying an ex~ption~ly

large load of lead (261, For more than two thousand years this ship has lain at the bottom of the sea at a depth of about 30 m, an excellent shield against en~ro~ental neutrons and also against radioactive remnants of the Chernobyl

accident. Due to the generous help of the

Soprintendenza Archeologica di Cagliari ed Oristano we were able to obtain a sizable amount of this Iead for

analysis. We have carried out measurements of X-ray

fluorescence and diffraction, neutron activation and alpha, X-ray and gamma ray spectroscopy. Alpha, X-ray and gamma ray spectroscopy was also carried out for comparison

n samples of modem lead, on specially

produced low radioacti~ty lead and on a sample of lead which is about 500 years old. The intrinsic ra~oacti~ty

f the Roman lead was found to be extremely low, thus

showing that this material could be extremely useful in low activity experiments on rare decays and in low level gamma spectroscopy.

2. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Ancient lead and the Roman ship Al~ou~ lead is one of the earliest metals known and used by man 127,281, ores wnt~~ng it were only rarely smelted for their lead contents in pre-Roman

periods. The metal sought was silver, for which lead
res formed

the principal

source. However, not all

argentiferous lead ores contained enough silver that could be extracted economically. Therefore some lead was further produced as a by-product in the search for

silver. This is also indicated by the contemporary

ap- pearance of lead and silver in the archaeological record roughly in the middle of the fourth millenium B.C. Such lead always contains silver, in the order of 400 to 800 ppm, and several other impurities derived from the ore zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

1291.

SLIDE 3

108

A. Alessandrelio et al. / Radioactivity
f ancient reman lead

The threshold for the silver content that could be extracted economically from lead decreased

ver the

course of time and was probably around 100 ppm from the Roman period to the beginning

f the industrial

age [30]. This can be readily explained by the separation technique

f silver from lead called cupellation.

It in- volves the oxidation

f molten

lead to litharge (PbO) and of any other base metals which are dissolved in the

litharge. The noble metals (silver, gold, and the platinum

group metals) are left largely unaltered while the litharge is absorbed by the hearth material

r skimmed
ff.

However, some silver passes into the litharge especially near the end of the process either as Ag,O

r in the

form of small occluded globules of silver-rich lead metal. Therefore even lead produced from litharge by reduc- tion with charcoal usually contains some silver, but rarely more than about 100 ppm. The composition

f the litharge

varies through the process. While at the beginning it contains most of the base metals such as arsenic, antimony, tin and any nonmetallic impurities from the lead bullion, it consists

f rather

pure PbO usually with some CuO in the middle and is enriched in bismuth and silver at the end

f the process.

Since impurities

f base metals have the

effect

f hardening

lead, which was undesirable for many purposes even in Roman times, it is likely that the impure litharge from the beginning

f the cupellation

was collected separately and discarded. The rest was probably reduced to lead and cast into ingots. Many Roman lead ingots have the sign “EX ARG” imprinted

n them indicating

that the lead has been desilvered. From the standpoint

f radioactive

impurity content it is much better, if the lead has been desilvered. Al- though the principal lead mineral galena (PbS) does not contain uranium and thorium in any significant amounts, some siliceous material has to be added to the smelting charge to form a slag and this may well con- tain uranium and thorium in the ppm range. Since both are typical lithophile elements, which on smelting con- centrate in the slag, it is not impossible that some slag particles may be occluded by the lead bullion. Other elements taken up by the lead bullion besides the noble metals are copper, arsenic, antimony, bismuth, thallium and sulfur. Depending

n

the smelting conditions selenium, tellurium, tin, and even cobalt, nickel and zinc may be added to this list. Generally however the purity of the lead bullion

r work lead has copper

and sulfur as principal impurities 90% to 99% of the time zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 1311. The Roman ship wreck [26] is located near the “Ma1 di Ventre” island,

ff

the south western coast

f

Sardinia, to the north

f Sulcis-Iglesiente

where lead mines have been in operation since ancient

times. Due

to the strong wind and to the small protruding rocks this area is known for many ship wrecks. The lead ingots lie 28 m below the water on a flat, sand covered, sea bottom and covers an area measuring 36 m zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

x 12 m.

These figures are indicated by the position

f the

anchors:

ne iron anchor

lies on the prow, three lead anchors

n the stern and two lead counterbalances

are symmetrically disposed

n the stem. Thanks

to the size

f the ship and to the large quantity
f lead ingots

stored in its center it has been possible to classify it as an “oneraria magna”, suitable to carry heavy loads of

metals. The ship’s keel is reinforced

by iron nails, each measuring more than 70 cm. It is estimated that there are about 1500 ingots, each weighing about 33 kg (about 100 Roman pounds). 102 ingots

f

tronco-pyramidal shape and bearing the manufacture’s name have so far been excavated. The shape and weight allows us to date them to the first half

f the first century

B.C. This date is also in agreement

with the pottery found (amphoras Dressel 1) and with the names of the ingot manufacturers, who are familiar to us from previous discoveries. The name that appears most frequently is that of the Pontilieni family (Socie- ties of Caius and Marcus, sons of Marcus, a family of the Fabia tribe, active in the middle of the first century

B.C.).

Other ingots bear the mark Cams Hispallus (or Hispalius)

f the Menenia
tribe. One should note that a

“Cams Russinus”, whose activity is well known throughout Sardinia, also belongs to this tribe. The mining activity

f these

families

f Italian
rigin

in Spain is well known. On the other hand the discoveries

f a considerable

number

f metal ingots in the Medi-

terranean sea [32,33] demonstrate that such metals were widely spread. Lead was used in large quantities for a variety

f purposes

such as water pipes, anchors, net sinkers and for lead clamps used in the construction

f

stone buildings. The large quantity

f this metal found

as a part of the shipload near the Ma1 di Ventre island proves that specific ships were used to carry metal ingots to urban markets where there was a greater demand

f them.

The lead ingots

f the Pontilieni

family found in Agde (France) were analysed in 1973 [33] and revealed the presence

f copper

and silver as well as traces of iron and nickel and, in one case only,

f bismuth.

No evidence was found for arsenic, tin, antimony, gold, zinc and cobalt. The data presently available support the hypothesis that such ingots came from mines in Spain, namely from Cartagena

r Rio Tinto,

where most of the pro- duction took place. However, lead mines were also active in Sardinia in the first century

B.C.. In this region

it is possible to distinguish two major areas Monteponi- San Giovanni and Montevecchio-Ingurtosu. The latter has, since the Nuraghic era [34,35], played an important role in the production

f silver and lead. The lead which
riginated

from the former area does not contain tin or antimony, while the other one contains both metals [36]. As it will be seen later the ingot analysed here contains both antimony and tin. The lead from Montevecchio-

SLIDE 4

A. Alessandrello et al. / Radioactivity
f ancient reman lead

109 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Ingurtosu was normally brought to the mouth of the Rio Piscinas river, where ingots similar in size to those found near Ma1 di Ventre island have recently been

discovered. There the metal was loaded on large ships
r taken to the market of Sulci with smaller ones. The

Piscinas river is located at approximately 50 km from the position of the wreckage. The unaltered position of the ingots, stored in the center of the ship, and that of the anchors, still situated at the prow and at the stern, indicate that the ship sunk without changing the arrangement of the load. It may have sunk because of its excessive weight or because of an accidental collision with the rocks that protrude near the island. The ship probably sailed, perhaps after a stop, from one of the nearby Sardinian ports and then sank during the first part of its journey. It should be of considerable archaeological interest to identify the source region of this ship load. In princi- ple, this should be possible by comparing trace elements and lead isotope ratios of the ingots with ore from possible source regions mainly on Sardinia itself and southern Spain [29,37]. Although these regions are geo- logically rather similar it is still possible that they can be geochemically differentiated. In addition, it would be interesting to see if all ingots have similar composition and therefore derive from a single ore deposit or not. This will hopefully be possible only after the entire ship load has been recuperated.

3. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Preparation and chemical analysis of the sample The ingot sent to us for investigation is shown in fig.

1. It is covered by a corrosion crust. Before taking

samples we determined the radioactivity of the bulk and

f the surface of one of the lead bullets (“glandes

missiles”) transported

n the same ship for defence

purposes, with the same setup to be described later. While no activity was found in the bulk, the surface has shown a Ra-226 activity of about 40 Bq cm-*, clearly due to contamination by the sea water. Therefore the bottom surface of the ingot was first removed in order to eliminate this contamination. In a careful operation carried out using distilled water and specially cleaned tools we were then able to extract from the bottom of the ingot about 4.5 kg of bulk lead, without damaging the upper surfaces. During this operation we were surprised to find a few cavities containing gas and about 20 g of transparent crystals. This material has been analysed qualitatively by energy dispersive X-ray fluorescence and was found to contain only lead and

chlorine. There are two minerals with this composition

that could form as corrosion products of lead in sea water, namely cotunnite (PbCl,) and laurionite (Pb- ClOH). Further measurements with X-ray diffraction enabled us to ascertain that most of the material was indeed cotunnite. The metal samples have also been analysed by X-ray fluorescence and, as expected, contain only lead as a major element. Impurity concentrations in the lead and crystals have been determined by instrumental neutron activation analysis following the technique of Wytten- bath and Schubiger [30]. Samples of about 100 mg were shaped to discs of roughly 8 mm diameter and 0.1 mm

thickness. The crystalline material was gently crushed

and homogenized. An aliquot of 100 mg was taken for

analysis. All samples were packed in polythene con-

tainers and irradiated together with a metallic lead standard of similar shape in a neutron flux of 5 x lOI* ncm-* s-t for 4 hours. After a decay time of 24-h the

Fig. 1. The lead ingot used in this analysis.

Samples of lead extracted from the ingot are also shown.

SLIDE 5

110 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

A. Alessandrello et al, / Radioactivity of ancient reman lead

Table 1 Impurity concentrations in samples from the lead ingot. The laboratory numbers for the various samples refer to: (a) lead from the surface (1397); (b) drill shavings from the interior (1398); (c) a solid chunk from the interior (1399); and (d) crystailine material (1400). Mean values of the concentrations of the same elements in 13 Roman lead ingots found in Great Britain and Switzerland (Wyttenbach and Schubiger, ref. [30]) are also included for comparison. All concentrations are given in ppm Lab no. cu 1397 (a) 1095 1397 (b) 1080 1398 (a) 1910 1398 (b) 1790 1399 (a) 1290 1399 (b) 1160 1400 52 13 ingots 190 As Sb < 0.08 0.11 < 0.08 0.12 < 0.08 0.17

=Z

0.07 0.16 < 0.08 0.61 < 0.04 0.25 < 0.9 < 0.4 35 160 Sn A8 Au c 80 54.2 0.0041 cl00 56.3 < 0.004 c9i.l 165 0.0102 c90 149 0.0067 <120 52.1 0.0124 < 90 35.4 0.0036 < 450 < 32 < 0.021 4 85 0.01 resulting gamma spectrum was measured at several de- cay intervals. The results of replicate analyses are given in table 1. We would like to note that the relatively high

detection limits in sample no. 1400 are due to the presence of bromine, which derives from the sea water and which, on irradiation, produces high background from 82Br. Cast lead is usually rather homogenous. Analysis of a Roman lead sarcophagus from Syria showed no sig- nificant variation in trace element content

ver the

whole cover, which was cast in one piece 1381. Therefore the differences between replicate samples and samples taken from different parts of the ingot are significant and suggest that some redistribution

f impurities,

probably by recrystallization

f the metal, has taken

place during the deposition time. Judging from the low concentrations

f arsenic and antimony

and from the fact that the b&k concentration

f silver in the ingot is

probably below 100 ppm it is very likely that the lead has been desilvered. For the purpose of using this lead as shielding material for low level experiments this is beneficial, because any radioactive impurities in the lead bullion would have been greatly reduced in the cupellation

process. As it will be shown later this lead

can already be used as shielding material without fur- ther purification. It has also to be borne in mind that

r” t g

d)

LOO.

e50.

lam. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

1400.

2ooo. 2400, 2x!KI.

Energy Ii-WI

Fig. 2. Energy spectra of background: modem (a), Dutch (b), low activity (c), Roman (d) lead, and background (OFWC copper) (e).

SLIDE 6

A. Alessandrello et al. / Radioactivity
f ancient roman lead

111

for this purpose the lead would have to be remelted anyway, which would result in further removal of easily

xidizable

impurities and nonmet~~c inclusions. Al- though probably not necessary the lead could then be additionally purified by zone melting.

4. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

M~ ~ ~ ernents

f gamma and X-rays on the various

samples of the lead

Measurements

f gamma and high energy X-rays

were carried out in our low background laboratory located in the Gran Sass0 Tunnel [6] with a GeLi diode

f 168 and 145 cm3 total and active volume, respec-
tively. This detector was specially made with low radio-

activity material: OFHC copper for the cap and gold instead of indium for the gaskets. It was shielded with at least 8 cm of previously tested OFHC copper and with an external layer of at least 20 cm of low radioac- tivity lead. The internal shield was designed in order to allow insertion around the detector of a layer of lead of 1.5 cm average thickness corresponding to a total mass

f about 4.5 kg. Spectra of pulses from the detector

were automatically analysed for peaks due to radioactiv-

ity. We would like to point out that two peaks con-

stantly appear in all spectra and are int~nsic~ly due to the detector: the peak at 1460.4 keV due to the varnish nearest to the detector and the peak at 661.65 keV due to an accidental contamination

f Cs-137. These un-

welcome peaks however allow a continuous monitoring and energy calibration

f the setup.

Measurements have been carried out placing the following materials around the detectors: (1) a standard shield of OFHC copper to provide a “blank spectrum” (280 h of effective running time); zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Table 2 Counting rates of the various samples (in units of lo-’ counts keV-* hh’ and of counts h-r for the peaks)

En. (keV)

Origin Background Modem Roman Low activity Dutch 100-200 200-400 400-600 600-800 800-100 1000-120 1200-1400 140@-1600 1600-18~ 1800-2000 75 86.8 186.2 238.6 241.6 295.2 352 569.7 583.1 604.7 609.3 661.65 803.17 911.2 1063.7 1120.3 1173 1332 1460.8 1764.5 1770.2 2614.6 18.3 f0.3 216 +1 9.1 10.2 112 *I 3.2 iO.1 26.7 +0.3 1.84+0.06 5.6 +O.l 0.74 tt 0.04 1.08 50.06 0.74 *o&t 0.82 rto.05 0.52 t 0.03 0.56 kOo.04 0.52 * 0.03 0.52 +0.04 0.07 f 0.01 0.08 +0.02 0.03 * 0.01 0.037 f 0.008 X-rays X-rays 226Ra 2’2Pb ‘14Pb 2’4Pb zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

21dBi 207Bi 208 Ti ‘“CS 214Bi 13%S

40K 214Bi

207Bi 208 n

0.01* 0.1 0.17 f 0.06 0.1 10.1 0.15 f0.06 0.14+0.05

0.05 * 0.05

0.1 +0.1 0.08 f 0.04 0.98 + 0.06 0.05 j, 0.02 _ 0.04 4 0.02 0.09 + 0.0 0.04* 0.03 0.79 * 0.05 0.04 f 0.02 _ 0.03 * 0.01 14 fl 12 *1 0.27 +0.30 0.2 fO.l 0.2 i-o.1 0.4 rto.2 0.29 f0.20 0.41 *0.02 0.05 f0.09 0.16 +0.09 0.28 kO.08 0.8 iO.1 0.34 *to.06 0.02 io.04 0.15 *0.05 0.09 +0.04 0.1 kO.1 0.1 rto.1 0.76 iO.06 0.02 i-o.01 0.03 io.01 0.03 &to.01 14.8 Ifro.2 14.4 F0.2 8.0 40.1 7.9 kO.1 3.06 If: 0.07 2.76 * 0.07 1.80&0.07 1.69 f 0.05 0.90+0.04 0.74 f 0.03 0.78 to.04 0.71* 0.03 0.48 f 0.03 0.49 f 0.03 0.48 f 0.02 0.42+0.03 0.09 * 0.01 0.07 f 0.03 0.03 jr 0.008 0.03 * 0.03 15.3 +0.3 8.1 ;tO.2 2.9 rf:O.l 1.84+0.0’7 0.88 Ifr 0.05 0.70 f 0.04 0.56 f: 0.04 0.49 * 0.04 0.07 + 0.01 0.02+0.10 0.04+0.08 0.23 rt 0.02 0.02 Ifr 0.10 0.02 _I 0.08 0.26 f 0.02 0.15 rto.11 0.10 f 0.09 0.03 + 0.03 0.10 * 0.09 O.lOrtO.08 0.12rtLO.06 0.1 +0.1 0.10 f 0.08 0.10+0.5 0.1 +0.1 0.16 i 0.08 0.15 io.05 0.21+ 0.05 0.18irO.06 0.53 +0.15 0.05 f 0.05 0.08 * 0.07 0.13 i 0.04 0.07 f 0.03 0.15-+0.04 0.15 rfo.04 0.07 rt 0.03 0.16&0.03 0.19*0.04 1.1 +0.1 0.98 + 0.09 0.77f0.10 0.06 i 0.02 0.02 f 0.02 0.04*0.02 0.02 * 0.02 0.09 rf 0.03 0.10 f 0.03 0.09 f 0.03 0.05 _L 0.03 0.78 rt 0.05 0.65 i 0.06 0.06 rF 0.01 0.04 * 0.01 0.01 rt 0.01

0.02 f 0.01

0.02 + 0.01

0.08 + 0.03

0.08 + 0.02 0.04 f 0.02 0.77 + 0.07 0.03 rf: 0.01

0.02 10.01

SLIDE 7

A. Alessandrello et al. / Radioactivity
f ancient roman lead

112 (2) zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

(3) (4) (5) zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 5 kg of common electrolytic modern lead previously analysed and found free from gamma ray activity, which will be from now on indicated as “modem” (190 h); and 4.5 kg of “Roman” lead obtained as specified be- fore (236 h); 5 kg of modern lead specially produced by J

hnson

and Matthey, later indicated as “low activity” (335 h); and 4.5 kg of lead obtained from 500 year old ingots from a sunk Dutch ship later indicated as “Dutch” (137 h). The five spectra are shown in fig. 2. The correspond- ing counting rates in the various energy regions and in correspondence to the various expected radioactivity peaks are reported in table 2. We would like to note that:

(1) (2)

The counting rates in the peaks at 661.65 and 1460.6 due to the above mentioned contaminations

f the detector are the same in all spectra and are

therefore only due to contamination of the detector. All samples, chosen, we stress, for their low radiac- tive contamination, show the same counting rate in all gamma ray peaks as the blank. The only excep- tions are the peaks at 569.7, 803.17 and 1063.7 in the modern lead sample. The first and the last indicate the presence of *“Bi. This nuclide, of cos- mogenic or contamination origin, with a half-life of 32.2 yr, is absent in the blank and in the other zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

5.0

;

$3.0 :

d

2.0 0.0

40. 80. 1W.120.1~.160.18).2M).Za).XO.260.280.300.320.~0.360.380.410.40.~.~.480.500.

Enaw IKmVI

Fig. 3. Comparison

in the low energy region in modem and Roman

lead. The spectrum

for low radioactivity lead is indistinguishable from the Roman

ne.

samples of lead. The same is true for the peak at 803.17 keV, due to the decay of 210Po into 206Pb, which indicates the presence of 210Pb with an activ- ity of 240 k 40 Bq/ kg. This figure will be compared later with the value obtained with X-rays, alpha particles and bremsstrahlung spectra. (3) Excess counting rates with respect to the blank at the energies of lead X-rays is absent for Roman and low radioactivity lead, weak for Dutch lead, but considerable for the “modern” one. (4) The integral counting rate in the low energy region (< 600 keV) follows approximately the same be- haviour (fig. 3). The last two points will be further discussed later on. We would like, however, to stress already that the counting rate in the low energy region for the Roman and low activity lead are very low, even in comparison with the blank, obtained with OFHC copper, which is normally considered among the best shielding materials. Limits on the natural radioactivity of the various lead samples have been obtained from the difference be- tween the corresponding spectra and the blank and are reported for a few relevant nuclei in table 3.

5. Measurements
f alpha activity

For alpha measurements we have used samples in form of discs of 1 mm thickness and 5.6 cm2 surface

SLIDE 8

A. Alessandreh

et al. / Radioactivity of

ancient

roman lead 113

Table 3 Gamma activities (Bq/kg) and X-ray counting rates (h-‘)

f

the various samples. Limits are at two standard deviations Modern Roman Low activity Dutch 138U < 0.03 < 0.01

c. 0.01

< 0.02

214& c 0.003

< 0.001 < 0.001 c 0.002

208 ll < zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

0.001

z
.ocKl9 < 0.0007

c: 0.001 Pb X-rays 33.5 * 1 < 0.1 < 0.1 0.5 * 03 area, placed at a distance

f 4 mm from a 900 mm’

silicon surface barrier detector. All samples were cleaned in an ultrasonic bath immediately before placing them zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

0.04 Low activity

in front

f the detector.

Before each measurement

f

lead samples, blank runs have been performed (fig. 4a) by exposing the detector to the sample holder (a plate

f OFHC

copper)

nly. The spectra
btained

with the four samples are also reported in fig. 4. Those concern- ing Dutch, Roman and low activity lead show no P ar- tic&r feature, except for some radioactivity

f 22 Rn

(5492 keV), which is often present at the beginning

f

the measurement, but disappears after prolonged pump-

ing. The excellent

upper limit on 2’4Bi obtained from gamma spectroscopy suggests, however, that such con- tamination is not intrinsic

f lead. Comparison

with the background

f the counting

rates in the energy window

0.06 L 0.05 0.04 5000 6000 Background

E ikeV)

1.6 7 14 2 1.2 Electrolytic " 1 0.8 0.6 in

2

”

Monte carlo simulation

Fig. 4. Alpha particle

spectra.

SLIDE 9

114

A. Alessandrello et al. / Radioactivity of ancient reman lead zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Table 4 *“Pb activity from alpha measurements. Limits are at two standard deviations Sample Measurement Counts Activity time [h] (no/ h)

Pq/kl

Background 361.2 0.13+0.02

Modem

45.3 8.8kO.4 186 * 10 Roman 281.5 0.19 f 0.03 < 2.6 Low activity 165.5 0.11* 0.03 < 0.9 Dutch 67.0 0.16 + 0.05 < 2.8 below the previously discussed ***Rn line leads to the results shown in table 4. The spectrum of fig. 4b indi- cates alpha particle emission with maximum energy of 5300 keV, corresponding to the decay of “‘PO into 06Pb. In order to evaluate the corresponding activity a Monte Carlo simulation has been performed: the results are also shown in fig. 4b. Comparison of this simulation with experimental data clearly shows that the shape of the spectra is different in the energy region between 4800 and 5300 keV. A possible explanation has been proposed by Wojcik [ll] and by Zastavni et al. [21] who found a growing concentration of l’Po on the sample surface during very long measurements. To avoid problems due to such an effect, calculation

f the activity have been performed only in the energy

region between 3500 and 4800 keV, which corresponds to a depth ranging from 1.8 and 6.6 pm. Results for all samples after background subtraction, are shown in table 4. The samples of ancient and low activity lead yield a counting rate which is compatible with back- ground within two standard deviations. The *r’Pb con- tamination is in very good agreement with the value

btained by gamma ray spectroscopy.
6. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

X-ray production by ionizing particles Since the sample of modern lead presents evidence both of X-ray peaks and of a bump in the counting rate below 1 MeV (fig. 3), we have tried to account for such effects in terms of radiation produced by ionizing par- ticles propagating through the material. Alpha and beta particles have been considered separately. 6.1. Production by alpha particles From the data available in the literature [39-441 the cross section for ionization of K-shell electrons by alpha particles is very low, e.g about 3.5 mb at 5 MeV. Since these findings are purely theoretical we have checked them by comparing the cross sections computed for the same process induced by protons [45] with experimental results [46]. As shown in table 5 the agreement is more than satisfactory. Another approximate, but direct, estimate of these cross sections can be obtained as follows. Let Pd be the probability that, in the alpha decay of *r’Po, a vacancy is created in the K-shell of the daughter atom. The probability of creating the same vacancy by a central collision of alpha particles of the same energy on Pb is PC = 4P,. In fact the transition amplitude is doubled, because the path is doubled, and consequently the prob- ability is four times greater.

Pd

has a quadratic dependence from the particle velocity because the term of the Hamiltonian which breaks the adiabaticity inducing the transition to an excited electronic state, is proportional to collision

velocity. From experimental data on “‘PO alpha decay

[46] we obtain P,, = 1.6 x 10e6 for an alpha energy of

E. = 5.3 MeV and therefore P,(

E)/k( E/E,) with k = 6.4 X 10m6 (E is the alpha particle energy). An alpha particle, colliding with a Pb atom, will create a vacancy if it crosses the K-shell region, the area

f which is, including Coulomb deflection,

A=7ra$(l-$j, where Z = 82 and a, = 0.6 X 10-r’ cm is the K-shell

radius. As a consequence the cross section is:
=nkga$(l-

$1.

(2)

Values obtained by this equation are greater (i.e.

6 X 10m6 b at 5.3 MeV) than the data quoted in [46].

We can therefore “a fortiori” conclude that the X-ray production by alpha particles is indeed negligible. By considering as an example an alpha particle of 5 MeV, an energy typical in radioactive chain, and by assuming a constant cross section along its path, we evaluate about 4 x lop6 produced K-shell ionizations. As a con- sequence an alpha-decaying contamination of 1 Bq in a kilogram of lead would correspond to a totally negligi- ble X-ray production rate of 0.3 per day. Table 5 Cross section (barn) for K-shell ionization induced in lead by alpha particles and protons Energy lMeV1

‘%lpha

ref. [45]

4

ref. [45]

4

ref. [44]

1

1.5 2 2.5 3 4 5 6 2.43 x lo-’ 1.26~10-~

(1.1 *0.1)x10-4

3.68~10-~ 9.79x104

3.85 x 1O-6

3.13 x10-3 (2.97*0.2)x10-3 1.57x10-4 6.96~10-~

4.11

x10-4

1.48~10-~
_

3.51 x 10-3

_

6.67~10-~

_

SLIDE 10

A. Alessandrello et al. / Radioactivity of ancient reman lead

115

Ge(Li) dhector 5cm

Fig.

5. Scheme
f the detector

used in zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

ur Monte Carlo

simulation. We have performed a direct test of X-ray production by alpha particles by exposing a lead sheet to an 241Am source. Observed Pb X-rays can be completely accounted for by interactions of low energy gamma rays accompanying alpha decay. 6.2. Production by beta emission We have studied both X-ray and bremsstrahlung production by a Monte Carlo simulation of the propa- gation in lead of electrons with initial energy ranging from 200 keV to 3 MeV. The geometry chosen is shown in fig. 5. At each energy the code recorded the number

f generated photons of both kinds and the energy

spectrum from the detector (fig. 6). As a second step, the empirical law connecting initial energy and number

f photons generated and detected was obtained by

linear and quadratic interpolation. The numbers Nxs, N,,, N,,s, Nbr of X-rays and bremsstrahlung photons emitted and detected per gen- erated electron are given by: Nxs = 2.70 x 10-4E - 0.0822, (3) N,, = 1.10 x 10-6E - 1.70 x 10-4, (4) Nbs = 8.38 x lo-‘E

0.0255,

(5)

1000 2000 3000 4000

Electron

Energy (keV)

Fig. 6. Number of bremsstrahlung

photons detected as a func-

tion of the energy of the generated electron. Table 6 Detection yields for X-rays and bremsstrtiung generated by beta rays (photons per bequerel

f the parent

nucleus) Parent activity X-rays Bremsstrahlung 23sU in equilibrium 23sU before 222Rn 238U after 222Rn 2’oPb and daughters 232Th in equilibrium 232Th before 220Rn 232Th after 220Rn 9.65~10-~ 2.85 4.32~10-~ 1.48 5.33 x10-2 1.37 1.79x10-2 0.32 4.76~10-~ 1.51 1.86x10-2 0.66 2.90~10-~ 0.85

N,, = 8.96 x lo-“E*

1.68 x 10-7E - 4.27 x 10-7.

(6) These relations have been folded with the known shape of the beta spectrum, in order to compute a yield, z, defined as the ratio between the rate of photons detected per second and the activity (in Bq/ kg) of the parent nucleus of the decay chain considered in the selected geometry. It should be stressed that such yields depend explicitly on the sequence of beta decays in a specific radioactive chain. They have in fact been com- puted separately for the chains shown in table 6. The bremsstrahlung spectra of all samples are obvi-

usly due to two contributions:

from the sample and from a “background” due to other sources, essentially from the detector itself. Inspection of fig. 3 shows that the low energy spectra for Roman and low activity lead are identical, which suggests they are due to background

nly. This is confirmed by the following considerations:

(1) (2)

Let us assume that the low energy spectrum of the Roman lead sample was completely due to brems- strahlung from beta decay in lead of nuclides of the 232Th and 238 U chains in secular equilibrium. In this hypothesis, and using eq. (6) we can calculate for these chains a total activity A = 2.0 + 0.1 Bq/ kg. This figure is larger by three orders of magnitude than the upper limits obtained by gamma ray spec- troscopy and reported in table 3. The presence in Roman or low activity lead of a contamination due to the 238U or 232Th in secular equilibrium can therefore be excluded. Let us, conversely, drop the previous hypothesis of secular equilibrium and adopt the assumption that all the Roman lead continuous spectrum is due to bremsstrahlung

f beta decays following *l’Pb

in the sample. We find that the main contribution comes from 210Bi. In this case the activity computed from the bremsstrahlung spectrum is A = 15.3 + 0.1, in contraddiction with the limit obtained from X- rays analysis (table 7). We conclude that the observed continuous radiation spectrum for the Roman and low activity lead is

SLIDE 11

116

A. Alessandrello et al. / Radioactivity of

ancient

rornan lead Table 1 Activities (Bq,/kg) computed from X-rays and bremsstrahlung

under different hypothesis for their origin Origin 238U

232Th

238 U

232Th *“Pb 2’oPb “‘Pb *“Pb

Method X-rays

X-ray bremss. bremss. X-rays bremss. gamma spectr. alpha spectr.

Modem 24.2* 0.5

49.1* 1.0 21.7* 0.2 41.0* 0.4 147 +30 194 + 2 240 *40 186 +lO

Roman i 0.17

< 0.35 < 0.03 < 0.06 cl.0 < 0.3 ‘z 40 < 2.6

Low activity < 0.15

i 0.30 < 0.4 < 0.09 < 0.9 < 0.4 < 40 < 0.8

Dutch 0.37*0.12

0.15 f 0.24 < 0.08 < 0.15 2.2 +0.7 < 0.7 i 60 < 0.9

due to sources outside the sample, namely on the surface or inside the detector and can therefore be considered as background. The limit on the activity

f the Roman lead has therefore been computed by

taking a two standard deviation statistical fluctua- tion over this background. The activities of the

ther samples have then been evaluted by subtract-

ing from the corresponding spectra the spectrum of the Roman lead taken as background. The results are reported in table 7. We would like to point out that:

(1) (2)

(3) (4)

The activity of the modem lead obtained from X-ray and bremsstrahlung measurements is fully inconsistent with those from gamma-ray spec- troscopy if attributed to contaminations of 238U and 232Th in secular equilibrium. This activity can there- fore be attributed only to the chain following *“Pb; No net measurable activity is present in Roman and low activity lead; There is some evidence from X-rays for a weak activity due to the 2’o Pb chain in the Dutch lead. This evidence is not confirmed by bremsstrahlung. The possible activity of Dutch lead cannot obvi-

usly be attributed to original 2’0Pb, but only to

contaminations after the extraction of the material; The activities due to 210Pb in modem lead obtained from X-rays and bremsstrahlung are in good agree- ment between themselves and with the values ob- tained with alpha and gamma ray spectroscopy, which are also reported in table 7 for comparison.

7. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Conclusions From the previous measurements and analysis we would like to conclude that:

(1) (2)

The material from the Roman ingot is 99% pure lead, containing only traces of copper, antimony and silver. From the high purity and expecially from the low content of this last metal it seems clear that lead has been desilvered with a consequently strong reduction of radioactive contaminations; No gamma ray peak in excess of background is (3) (4) (5) found for any of the samples analysed here, with the exce 4 tion of weak gamma ray signals due to 207Bi and “PO which are present in modem lead. Upper limits for activity due to the components

f the

uranium and thorium chains in secular equilibrium are of the order of 10m2 Bq kg-‘. Alpha and gamma-ray spectroscopy and analysis of the bremsstrahlung background and of the lead X-ray peaks lead, in excellent agreement, to a 210Pb contamination in modem lead of about 190 Bq/ kg; The best samples are Roman and special low radio- activity lead where the *l’Pb contamination is defi- nitely below one Bq/ kg. As a consequence the counting rate in the low energy region is the lowest for these two samples: these materials appear there- fore to be excellent for shielding experiments on rare events; The counting rates in the presence of Roman and low radioactivity lead are identical and lower than for OFHC copper. They can be fully accounted for by intrinsic radioactivity of the detector. Only the use of a larger quantity of lead and of a more efficient and less contaminated detector could allow a definite determination of what the ultimate radio- activity of Roman lead is. Acknowledgements We would like to thank the Soprintendenza Archeo- logica per le Provincie di Cagliari ed Oristano for the generous and understanding collaboration, and the di- rector and staff of the Laboratori Nazionali de1 Gran Sasso for their warm hospitality. The highly profes- sional work of Roberto Mazza and Sergio Parmeggiano and the help of our student Anna Somigliana are also gratefully acknowledged. References

[l] See for instance:

L. Van Hove, CERN-TH.5775/90,

June 1990, preprint.

SLIDE 12

A. Alessandrello et al. / Radioactivity zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
f zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

anc ie nt reman lead

111 [2] E. Fiorini,

Physica B169 (1991) 388. [3] T. K&ten, Proc.

f the

13th Int. Conf.

n Neutrino

Physics and Astrophysics, Boston (World Scientific, 1989) 317. [4] M. Moe, Proc. 1990 Int. Conf. on Neutrino Physics and Astrophysics, CERN (Geneva), to be published in Nucl.

Phys. B.

[S] H. Ejiri, Int. Symp. on Nuclear Physics in 1990’s, Santa Fe, May (1990) to be published in Nucl. Phys. B. 161 E. Bellotti, Europhysics News 21 (1990) 121. [7] E. Giacomelli, report to the Int. Conf.

f the European
Phys. Sot., Madrid,
Sept. 6-14

(1989) to be published in

Nucl. Phys. B.

[S] E. Bellotti, M. Buraschi,

E. Fiorini

and

C. Liguori,

INFN,‘TC-85/19

Oct. 15 (1985).

[9] P. Belli, R. Bemabei,

S. d’Angelo,

M.N. De Pascale, P. Jucci,

L. Paoluzi,
R. Santonico

and G. Villoresi, Nuovo

Cim. 1OlA (1989) 959.

[lo] D.E. Watt and D. Ramsden, in: High Sensitivity Count- ing Techniques (Pergamon, London, 1964) and reference therein. [ll]

M. Wojcik, Nucl. Instr. and Metb. 189 (1981) 645.

(121 R.L. Brodzinski, J.H. Reeves, N.A. Wogman, F.T. Avig- none III and H.S. Miley, Nucl. Instr. and Meth A292 (1990) 337, also for previous references. [13] E. Bellotti, 0. Cremonesi,

E. Fiorini,
C. Liguori,
A. Pullia,
P. Sverzellati

and L. Zanotti, Nuovo

Cim. 95 (1986) 1.

[14] H.L. De Vries and G.W. Bardensen, Physica 19 (1953) 987. [IS] R.I. Weller, E.C. Anderson and J.L. Barker, Jr., Nature 206 (1965) 1211. (161 W.A. Kolb,

Proc. 1st Int. Congr.
Rad. Protection,

Rome (1966), ed. W.S. Snyder (Pergamon, New York, 1968) p. 1385. [17] O.G. Bartels, Health

Phys. 28 (1975) 189.

[18] RI. Weller, Health

Phys. 41 (1981) 15.

[19] W.A. Kolb, Environment. Intern. 14 (1988) 367. [20] M. Eisenbud, in: En~ronment~ Radioacti~ty (Academic Press, New York, 1973). [21] A. Zastavny,

J. Bialan and T. Sosinski,
Appl. Radiat.

Isot. 40 (1989) 19. [22] W.A. Kolb, PTB Braunschweig, private communication. [23] R. de Boeck,

F. Adams

and

J. Hoste,
J. Radioanalyt.
Chem. 1 (1968) 397.

[24] B. Bunzl and W. Kracke, Nucl. Instr. and Meth. A238 (1985) 131. [25] G. Heusser, Proc. XVth Regional Congress

f IRPA,

Visby 1989, ed. W. Feldt, Progress in Radiat. Protection (TUV ~einl~d, Cologne, 1989) p. 468. [26] D. Salvi, Prime considerazioni ml10 scavo di un relitto romano press0 l’isola di Ma1 di Ventre (1989) preprint. [27] Georgius Agricola, in: De Re Metallica, translated by H. Clark (Hoover and L.H. Hoover, New York, 1950). [28] W.A. Kolb, Die Eigenaktivitaet von Blei, Proc 1st Int. Congr. Rad. Protection, Rome (1966)

ed. W.S. Snijder

(Pergamon, Oxford, 1968) pp. 1385-1391. (291 E. Pernicka, Jahrb. Rem.-Germ. Zentralmuseum 34 (1989) 607. [30] A. Wyttenbach and P.A. Schubiger, Archeometry 15 (1973)1 199. [31] E. Pemicka and H.G. Bachmann, Erzmetall36 (1983) 592. 1321 M. Besnier, Rev. D’Arch. VII 211 (1920). 1331 F. Laubenheimer-L~~~~~t, Recherches sur le lingots de cuivre et de plomb d’epoque romaine dan les regions de Languedoc

Roussillon

et de la Provence

Corse, (De

Broccard, Paris, 1973) p. 192; C. Domergue, Le Lingot de Plomb romains du Mu&e Archeologique de Canthagene et du Mu&e Naval de Madrid, Archive Espairol de Arcqueologia, XXXIX (1966) p. 113. [34] G. Rolandi, Notixie sull’ industria de1 piombo e dell0 zinc0 in Italia (Sot. Montevecchio ed., Milan0 1948). [35] G. Ugas and L. Usai, Atti de1 II Convegno di Studi: Un millennio di relazioni tra la Sardegna ed i Paesi de1 Mediterraneo (Provincia di Cagliari, 1986) p. 167. [36] P. Zuffardi, private communication. [37] N.H. Gale and Z.A. Stos-Gale,

Ann. Brit. School Athens

76, (1981) 169. [45] D.D. Cohen and

M. Hanigan,

Atom. Data and Nucl. Data Tables 33 (1985) 255. [38] E. Pernicka, Arbeitsblatter

f. Restauratoren

1 (1989). [39] J.S. Hansen,

Phys. Rev. A8 (1973) 822.

[40] J.D. Garcia,

Phys. Rev. Al (1970) 280.

[41] J.D. Garcia,

Phys. Rev. Al (1970) 1402.

[42] J.S. Hansen,

Phys. Rev. A9 (1974) 40.

(431 W. Brandt and G. Lapicki,

Phys. Rev. A23 (1981) 1717.

[44] H. Paul and J. Sacher, Atom. Data and Nucl. Data Tables 42 (1989) 105. [46] CM. Lederer and VS. Shirley, in: Tables of Isotopes, 7th ed (Wiley, New York, 1987).