a Laboratoire d’Étude et de Recherche sur l’Hygiène et la Qualité des Aliments de l’Agence Française de Sécurité Sanitaire des Aliments, AFSSA/LERHQA/CEN-RAD, 10 rue Pierre Curie,
94704 Maisons Alfort, France b Service de Protection contre les Rayonnements du Commissariat à l’Énergie Atomique,
CEA/SACLAY/UGSP/SPR/SRSE, 91191 Gif sur Yvette Cedex, France
(Received 20 December 2000; accepted 24 May 2002)
Abstract – This paper presents a review on all available analytical methods for 90Sr and 89Srdetermination in milk. Special attention was focused on the parameter of the performing time inorder to assess the rapid measurement of milk contamination, especially in the case of emergencysituations such as a nuclear plant accident. Preliminarily, a short description of the milk matrix andradiostrontium transfer from soil to milk and its derivatives is given. Then analytical procedures areexamined step by step. 90Sr and 89Sr can be separately determined in less than 2 days by a rapidsample pretreatment and the use of a liquid scintillation detector.
Résumé – Revue des méthodes de dosage du radiostrontium dans le lait. Cet article présentel’ensemble des méthodes analytiques disponibles pour le dosage du 90Sr et du 89Sr dans le lait. Cetteétude attache une importance toute particulière à la durée d’analyse dans le but d’une estimationrapide de la contamination du lait telle qu’elle est rendue nécessaire dans le cas de situationsd’urgences comme l’accident d’une centrale nucléaire. En préliminaires, une rapide description dela matrice lait et du transfert du radiostrontium du sol au lait et ses dérivés est donnée. Puis, lesprocédures analytiques sont examinées étape par étape. Les isotopes 90 et 89 du strontium peuventêtre déterminés séparément en moins de 2 jours par un traitement rapide de l’échantillon etl’utilisation d’un détecteur à scintillation liquide.
Owing to its chemical and biochemicalsimilarities to calcium, more than 99% ofstrontium is efficiently incorporated intobone tissue and teeth. Characterized by along physical and biological half-life1,90Sr may cause damage to bone marrowbecause of its high-energy � particle(E = 546 keV).
The Chernobyl nuclear reactor accidentrevealed the need for a rapid 90Sr analyti-cal method. The analytical time is consi-dered as an important economic factor inconnection with the closing down time ofdairy farms suspected of having been incontact with clouds of radioactive dust.
In such an emergency situation, thedetermination of radiostrontium in a matrixsuch as milk is essential for several rea-sons; (a) milk ingestion is a substantial2
path of radiostrontium incorporation in thehuman body especially that of infants;(b) the milk is usually consumed as a freshproduct, therefore industrial storage ofmilk is not practicable; (c) the strontiumtransfer from soil-plant to cow-milk is effi-cient; and (d) the milk contamination levelgives an indication of the radiostrontiumdeposition over a wide area [17].
The EURATOM regulation publishedin the Official European Community Jour-nal of July 22, 1989 [14] stipulates upperlimits of radiostrontium activity in infantfood and dairy products at 75 and125 Bq·kg–1, respectively.
This regulation, which should beapplied in emergency situations such as anuclear accident, reveals that a low detec-tion limit is not of first importance in com-parison with analytical time.
A comparison of published methods isdifficult for several reasons; some essentialinformation is missing in the literature(yield or yield calculation, performance,
procedure time, uncertainty and detectionlimit calculations); the aim of each methodcan be different (determination of: (a) totalradiostrontium, and (b) 90Sr and 89Sr,elimination of some interfering radionu-clides such as 140Ba). In 1991, such a studywas done by Wilken and Joshi [52]. Thepurpose of this paper is to examine theavailable methods for 90Sr determinationin milk, taking into account time and safetyas priority factors.
2. RADIONUCLIDES TRANSFER
2.1. Transfer of radiostrontium from soil to milk
Cows ingest radionuclides with conta-minated water and forage, and inhale themwith the air. During fallout radionuclidesare directly deposited on plants and par-tially absorbed into plant cells. During andfor many years after the fallout, plantsincorporate radionuclides through theirroots. Comar and Wasserman [11] andTwardock et al. [48] have estimated thatabout 1% of the strontium present in forageis transferred into milk. Other studies esti-mated that 4–6% of strontium intake froma contaminated diet appear in cow milk[53]. Nevertheless, owing to its efficientroot uptake and long lifetime, strontiumcan contaminate milk for decades afterdeposition.
The transfer of radionuclides in the foodchain occurs in two major steps: the firstfrom soil to plant, and the second shorterstep from cow to milk. The plant/soil andmilk/cow concentration ratios3 are used forpredicting radionuclide concentration infoodstuffs and dose impact to man. Theirdetermination is rather difficult and thevalues are strongly influenced by many
1 Physical half-life: 28.5 years / biological half-life: about 7 years.2 In the literature, the relative contribution of milk to 90Sr incorporation in the human organism varies. Most often, that contribution does not exceed 10%.3 The concentration ratio is calculated as follows: Conc ratio
Bq per kg of end MatrixBq per kg of first Matrix----------------------------------------------------------- .=
Determination of radiostrontium in milk 3
physical, chemical and biological factors.For instance, investigations in Finland afterthe Chernobyl reactor accident have shownthat the transfer of 137Cs and 90Sr from soilto plant differs according to the geographicsite and plant species considered. In thesouthern area, the plant/soil concentrationratio of 90Sr was about 9 times greater thanthat of the 137Cs ratio [34].
The 90Sr transfer from cow to milk issomewhat less efficient than the soil/planttransfer. After a single uptake by the cow,the maximum strontium concentration isreached after 1–2 d [34].
2.2. Milk composition
Milk is a complex fluid containingmany components in several states ofdispersion. A rough classification of theprincipal constituents of milk with appro-ximate percentages of contents is given inTable I [50].
Milk and its derivatives are the mainsources of calcium and phosphorus in gen-eral nutrition. The mineral composition ofmilk varies according to species (Tab. II)[9] and also to race; the calcium and phos-phorus content in a “Normande” cow arelarger than in a “Frisonne”, “Pie Rouge” or“Holstein” cow. In theory, mineral concen-tration is not affected by nutrition. How-ever, while this statement may be true for
essential minerals that are under homeo-static control, it probably does not apply tonon-essential elements such as strontium.Moreover, important variations are observedduring the lactation period.
Speciation of 90Sr in contaminated milkis of interest with respect to the bioavaila-bility to the human organism, processingand decontamination of milk and radio-chemical analysis.
The minerals in milk are mainly inor-ganic salts, partly ionized and partly presentas complex salts. The mineral distributionbetween the soluble and colloidal phases ofmilk is heterogeneous. Temperature, pHand citrate content are the principal physi-cal and chemical factors likely to changemineral distribution: saline concentration inthe soluble phase increases when: (a) pHand temperature decrease, and (b) citrate
Table I. Approximate composition of milk [50].
Component Average content percentage (w/w)
Range percentage (w/w)
Average percentage of dry matter
Water 87.3 85.8–88.7
Solids (not fat) 8.8 7.9–10.0 69
Fat in dry matter 31 21–38
Lactose 4.6 3.8–5.3 36
Fat 3.9 2.4–5.5 31
Protein 3.25 2.3–4.4 26
Casein 2.6 1.7–3.5 20
Mineral substances 0.65 0.53–0.80 5.1
Organic acids 0.18 0.13–0.22 1.4
Miscellaneous 0.14 1.1
Table II. Mineral composition of milk forseveral species (g·L–1) [9].
Cow Goat Sheep Woman
Ca 1.20 1.30 1.90 0.31
P 0.90 0.95 1.50 0.15
Mg 0.12 0.12 0.16 0.037
K 1.50 1.60 1.25 0.52
Na 0.45 0.40 0.45 0.15
Cl 1.15 1.40 0.70 0.50
4 S. Brun et al.
concentration increases. Strontium isbound similarly to calcium, for example, tothe phosphate groups of caseins, and con-sequently is located in about 80% of themilk’s casein micelles [27, 29, 50]. At pH6.6–6.8, less than 50% of the strontiumcontent is exchangeable. By lowering thepH to 5.3, a major part of the strontium istransferred to the milk soluble phase [49].
2.3. Transfer of radiostrontium from milk to its products
The milk product:milk concentrationratios are given in Table III [27, 50]. Thehigher ratio is obtained for casein whencoagulation is carried out with rennet.Regarding cheese, the ratio between 3.8and 6.6 is quite substantial, whereas thebutter content is 11 times lower than themilk content.
3. METHODS FOR DETERMINING RADIOSTRONTIUM IN MILK
89Sr, 90Sr and its daughter 90Y are prac-tically pure � emitters. Their disintegrationschemes are given below.
Figure 1 shows the time required toreach the secular equilibrium of the activi-ties of 90Sr and 90Y. The three radionu-clides (90Sr, 89Sr, 90Y) must be separatedfrom a large amount of inactive matrixconstituents (fats, proteins, minerals) andfrom a number of interfering radionuclides(3H, 14C, 40K, 137Cs), prior to measure-ment. In addition, when the milk is con-taminated with fresh fallout, 140Ba and itsdaughter 140La must be removed. Thestandard methods for determination of 90Srin milk are often laborious4 and time-con-suming. The necessity of separating 90Srfrom 90Y contributes to this. Three proce-dure examples are given in Figures 2, 3and 4. The first procedure (AOAC [3])
Table III. Derivative:milk concentration ratio.
Fat rate Fermentation type
Fresh cheese fat rennet 3.91
lactic ferment 0.66
skim rennet 3.83
lactic ferment 0.97
Fermented cheese 4.7–6.6
Skim milk 0.92
Cream 24% 0.6
Butter 0.09
Casein rennet 20.5
acid 9.97
1 Milk contamination by metabolic way.
4 The determination of radiostrontium based on the nitrate strontium precipitation with fuming nitricacid requires nine different precipitation steps.
90Sr
Half life 28.5 y
�max 0.546 MeV
90Y
Half life 64.0 h
�max 2.28 MeV
90Zr (stable)
89Sr
Half life 50.4 d
�max 1.46 MeV
89Y (stable)
85Sr derivative85Sr milk 1
Determination of radiostrontium in milk 5
determines 89Sr and 90Sr activities sepa-rately. The analysis requires at least 15 d.Whereas, in an emergency situation, a sim-
ple and rapid radiostrontium determinationcan be accomplished in 24 h as shown inFigure 3 (Vaney et al. [49]). Tait et al. [45]
Figure 1. Secular equilibrium of 90Sr and its daughter 90Y.
Figure 2. Determination of 90Sr and 89Sr in milk (from Baratta [3]).
6 S. Brun et al.
Figure 4. Rapid determination of 90Sr and 89Sr in milk (from Tait et al. [45]).
Figure 3. Rapid determination of radiostrontium 90 and 89 in milk (from Vaney et al. [49]).
Determination of radiostrontium in milk 7
quantified 90Sr and 89Sr separately in 24 h(Fig. 4).
The separation of strontium is based onfour major steps: the carrier addition; thesample preparation; the radiostrontiumextraction by chemical treatment, and theactivity determination. The second, thirdand fourth steps are examined in succes-sion, and discussed under the aspect oftime saving.
3.1. Carrier addition
Two isotopes are commonly used ascarriers to evaluate the chemical yield of90Sr separation. Strontium 85, a � emittercharacterized by a short half-life (T =64.9 d), is simply measured by gammaspectrometry (E��= 514 keV) [20]. The sta-ble strontium 87 is also well adapted to ageneral analytical method. The chemicalyield is measured by gravimetry or atomicabsorption spectrometry (AAS). However,the gravimetric method is less precise.
The chemical yield of 90Y is frequentlydetermined by means of stable 89Y as car-rier. Numerous methods use yttrium oxalateprecipitation for the gravimetric evaluationof chemical yield. Nevertheless, at roomtemperature, the amount of hydrated waterconsidered for precipitate theoreticalweight differs between 7 [38] and 9 [3, 31].
Stable barium carrier is sometimes [3]used for BaCrO4 precipitation. Tait andWiechen [42] used stable cesium carrier toreduce 137Cs contamination.
3.2. Sample preparation
In most cases, the time required for thepretreatment operation depends on theamount of the sample. Pretreatment ofsmaller samples is faster, but a longercounting time is required.
Fresh milk samples are often preservedwith formaldehyde (HCHO) or sodiumazide (NaN3) and stored to obtain 90Y inthe cases where 90Sr and 89Sr must bedetermined [3].
The classical procedure of milk samplepretreatment is carried out in three steps[5, 18, 19, 21, 24, 26, 38]; the sample isdried at about 105–110 °C to a constantweight and incinerated in a muffle furnacebetween 400 and 600 °C; then the ashes aremost often dissolved in concentrated nitricacid (between 5 and 14 mol·L–1) except byBouquiaux and Gillard-Baruh, who dissolvedthe ashes in concentrated hydrochloric acid[6, 7]. Actually, this pretreatment is dif-ferent for each method with respect to, thetemperature chosen and sometimes thetemperature gradient used, and the nitricacid concentration and temperature usedfor dissolution. Moreover, the evaporationtime is strongly dependent on samplevolume.
Although calcination pretreatment isefficient and commonly used, there is nodoubt that it is time-consuming (>1 d)[4, 16, 25, 26, 31].
In a recent paper, Abbadi et al. [1]incinerated the milk sample directly with-out any time-consuming drying stages inan electrically-heated fluidized bed reactormade of quartz at 650 °C using com-pressed air. The milk is conveyed with aflow rate of 5 mL·min–1 so that less than2 h are required to ash 0.5 L of milk.
An alternative “leaching procedure” [19]requires boiling and evaporating milk sam-ples with 50 mL of 65% nitric acid(3 times) to dryness, taking up in 2 mol·L–1
nitric acid and warming for 30 min for dis-solution, then cooling and centrifuging.The supernatant is collected for the separa-tion step. Except for the evaporation step,which still requires a subsequent amount oftime, the time saving is accomplished byreplacing the calcination step with theshort leaching acid procedure (about 1 h).
Mikulaj and Svec [31] proposed, when90Sr/90Y radioactive equilibrium is reached,adding concentrated nitric acid, to precipi-tate proteins and the major part of the fats.The precipitate and supernatant are sepa-rated by centrifugation and filtration afterbeing heated at 90 °C and left at room
8 S. Brun et al.
temperature for 2 h. The disadvantage ofsuch a procedure is that a substantial partof the fats remains in the analytical solution.
The strontium extraction from milk isfrequently carried out with exchange orchelating resins which have a strong affi-nity for alkaline earth cations. Initial inves-tigations were carried out in order toreduce milk contamination. In 1954,Nervik et al. reported by Kirchman [27]studied milk decontamination from stron-tium by means of cation exchange resins.In 1960, Migikowsky [30] also developeda milk decontamination process whichrequired the cation exchange resin Dowex50-X12 (sulfonate groups) for strontiumextraction. Successively, mass of resin:volume of milk ratio, number of treat-ments, contact time, temperature and resinform (Ca, Na, K, CaNaK) are treated. Thebest extraction yield reached was 97%.These extraction techniques were appliedin sample preparation of radiostrontiumanalysis in order to save time.
Ovarec and Navarcik [33] chose to pre-concentrate radiostrontium on a strongacid cation exchanger OSTION KS-0807(amine groups) under static conditions.The authors estimate that the procedure is8 times faster than the drying and calcina-tion method.
To determine 89Sr and 90Sr, the Asso-ciation of Analytical Chemists (AOAC)[3] suggests the use of a double columnwhich contains at the top anion exchangeresin (Dowex 1-X8 trimethyl ammoniumgroup, for yttrium/citrate complex extrac-tion) and at the bottom cation exchangeresin (Dowex 50W-X8, sulphonate group,for Sr, Ca and Ba extraction).
Tait and Wiechen [42] treated the milkin a batch procedure with a Chelite P resin5
(aminomethylphosphonate group) in Na+
form. The contact time is between 30 and40 min at 65–70 °C in a water bath. Thestrontium is eluted with diluted nitric acid.
To reduce the milk/resin contact time,Vaney et al. [49] adjusted the pH of milk to5.3 with citric acid and mixed it withDowex 50W-X8 resin for 15 min, thenafter decantation, the supernatant was firsttransferred to a chromatographic columncontaining 5 mL of fresh resin, followedby the resin itself. The procedure can beachieved in less than 1 h and the yttriumseparation yield reaches 74%.
In 1995, Tait et al. [43] bound a bicyclicpolyether cryptand C222 to several resinsto improve their affinity for strontium. Forthe Dowex 50W-X8 system termed “D/222”, the sorption of strontium from milkis about 30% greater than sorption by thecorresponding untreated resin. The columnis washed with 10 mg·mL–1 solution ofstable cesium to remove milk traces andminimize contamination of radiostrontiumsorbed with radiocesium. Two years later[44, 45], the same authors extracted morethan 95% of strontium with D/222 underthe following conditions: the volume ratioof resin to milk is 1:50; the pH of the milkis adjusted to between 4.8–5.4; the contacttime is about 4 h at 22 °C.
There is no doubt that the ion exchangemethod is faster than the calcinationmethod (except for the use of a fluidizedbed reactor). However, in cases where themilk contains more cream or curd, Jeterand Grob suggest the second method [25].
For goat milk Comar and Wasserman[11] and Twardock et al. [48] observed asubstantial difference between spikedstrontium (66% for extraction yield) andmetabolic strontium (86% for extractionyield) by using the ion exchange method.That difference may be explained by agreater concentration of calcium in goatmilk than in bovine milk. Ion exchangerswith a relatively small amount of Sr-spe-cific binding sites probably become satu-rated with either strontium or interferingions such as calcium. Thomas [47] demon-strated that strontium metabolized and
5 100 mL of milk for 15 g of resin.
Determination of radiostrontium in milk 9
secreted into milk by the cow could beremoved in the same manner as strontiumdirectly spiked in milk.
It is obvious that the sample preparationtime can be significantly reduced byion exchange techniques under preciseconditions.
3.3. Strontium extraction
Since the sixties, a large variety of pro-cedures have been described in order toseparate and purify radiostrontium andradioyttrium. The different procedures areselective precipitation, liquid-liquid extrac-tion and solid-liquid extraction. All thetechniques available are successivelyexamined and discussed, as far as possible.
3.3.1. Precipitation
This technique is based upon differentsolubilities of cations in a given solution.For this purpose fuming nitric acid is oftenused [15, 20, 24] to separate strontiumfrom calcium, although these conditionsare difficult and dangerous for health (con-centration acid > 90%). The strontium/calcium separation is also feasible by usingdifferent solubilities of strontium and cal-cium oxalates (in presence of a large excessof calcium [10]).
Other interfering cations such as 140Baor 90Y require specific precipitation stepsto be extracted. In most procedures whichconsider a fresh nuclear reactor accident,the probable presence of 140Ba and itsdaughter 140La in milk requires theirextraction by means of chromate precipita-tion [3, 20, 24–26, 29, 33, 49]. The chro-mate precipitation is also well suited to theselective extraction of 210Pb. The stron-tium/yttrium separation is carried out bycoprecipitation with iron hydroxide [4, 24,26, 49], or selective precipitation of yttriumhydroxide [25, 33, 39].
Other precipitation reactions (phosphate,carbonate, oxalate, sulfate) are often usedto eliminate, among other things, traces of137Cs and 40K.
Though precipitation steps are fre-quently used, they are tedious [16] andmust often be repeated several times [29]to obtain adequate purity of strontium.They consequently lead to large losses ofstrontium.
3.3.2. Liquid – liquid extraction
In the case where 90Y is in secular equi-librium with 90Sr, numerous methods useTBP (tributyl phosphate) [3, 4, 31, 38] orHDEHP (Bis(2-ethylhexyl) phosphoricacid) [16] extraction of 90Y from aqueousacid. The technique enables us to deter-mine activity due to 90Sr and 89Sr sepa-rately by quantifying simultaneously 90Yand total radiostrontium activity. How-ever, these efficient methods require morethan 15 d to reach secular equilibrium.
Kimura et al. [26], among others [49],have proposed the extraction of strontiumfrom a large amount of calcium by liquid-liquid extraction with a macrocyclic ether(dicyclohexyl-18-crown-6, DC18C6) intochloroform. These methods give a suffi-cient decontamination factor, except for140Ba which follows strontium in thechemical operation. Consequently, chro-mate precipitation has to be carried out toremove barium.
Tait and Wiechen [42] have used thecrown ether DC18C6 into chloroform fol-lowed by extraction of barium into dichlo-romethane containing 21 crown 7 (21C7).The limitation of the procedure is the gen-eration of organic solvent wastes. Thedisposal of such wastes is difficult andexpensive.
3.3.3. Chromatographic methods
3.3.3.1. Ion exchange chromatography
Stella et al. [39, 40] have used two cat-ion exchangers; tin dioxide and copperchromate, which have simultaneously astrong affinity, respectively, for alkalineearth and barium, and no affinity for alkali
10 S. Brun et al.
metal ions such as cesium, sodium andpotassium. The use of both exchangersallows direct extraction of 75% of stron-tium from milk previously skimmed bycentrifugation.
Grahek et al. [21] carried out the stron-tium/calcium separation from milk ash bymeans of the anion exchanger amberliteCG-400 (Quaternary ammonium group)and 0.25 mol·L–1 HNO3 in methanol aseluent for calcium. However, the methodrequired 3 precipitation steps to removeyttrium and barium.
In 1967, Noshkin and Mott [32] usedcomplexing agent to separate strontiumand calcium. After sorption on ion exchangecolumns (Dowex 50-X12), strontium andcalcium can be eluted separately by com-plex formation with cyclo-hexanediamine-tetra-acetic acid (CyDTA) and carefulcontrol of pH. In the same way, Bouquiauxand Gillard-Baruh [6, 7] substituted ethyl-enediaminetetra-acetic acid (EDTA) forCyDTA.
Abbadi et al. [1] used successively 2different ion chromatography columns forselective extraction of strontium. Firstly,coarse-grained AMP (molybdophosphoricacid) filled in a column allowed them toseparate strontium from alkaline ions.Secondly, an acidic cation exchanger wasused to separate strontium from transitionmetal ions (Fe3+, Cu2+, Zn2+, Ru3+, Co3+),alkaline earth ions (Ca2+, Ba2+) and otherelements such as lead (Pb2+) by means ofan EDTA pH gradient. Both columns areintegrated into a HPLC system.
Heilgeist [22] used ion chromatographyas an ideal technique for a radiochemicalfine purification. It was mainly intended toseparate Sr from Ba when the separationby extraction chromatography was notcomplete.
3.3.3.2. Extraction chromatography
Horwitz et al. [23] have developed anovel extraction chromatographic resin forstrontium, “Sr-SPEC”, consisting of anoctanol solution of 4,4'(5')-bis(t-butyl-
cyclohexano)-18-crown-6 sorbed on aninert polymeric support. The resin hasshown excellent selectivity for strontiumover a number of alkali, alkaline earth andother metal cations. Most often, percola-tion, rinsing and elution can be carried outin about 3 h.
Because of its simplicity, rapidity andefficiency, numerous teams [2, 18–20, 25,42–45] have used the “Sr-SPEC” resin forstrontium analysis in milk. However, theresin cannot be used directly with liquidmilk but only with a fairly “clean” extract.Furthermore, the resin inefficiently sepa-rates strontium from barium. However,Tait et al. [44, 45] proposed an elution pro-cedure which enables us to separate bothelements. The second limitation of such amaterial is a low regeneration power(� 4 times) which incurs a substantial costfor routine analysis. They have obtainedabout 10% decrease in chemical yield forthird and fourth usages [45].
Recently, Tait et al. [46] succeeded inseparating strontium from calcium on theD/222 system in order to eliminate the“Sr SPEC” separation step. Previously toacid elution of strontium, calcium isremoved by adding 300 mL of an aqueoussolution containing 0.1 mol·L–1 penta-sodium tripolyphophate Na5P3O10 and stirring for 40 min. Then, 140Ba and its daugh-ter 140La can be separated from Sr byadjusting the acid eluate from the D/222 toa pH value of 3 and passing it through acolumn containing 0.5 g of manganesedioxide. However, that Sr/Ba separationrequires further optimization.
From the above considerations regardingchemical separation procedures, it is obviousthat no significant time saving can beobtained in the strontium separation step.
3.4. Activity determination
The activity determination by all availa-ble techniques is most often well separatedfrom radiochemical separation. However,Abbadi et al. [1] detected 90Sr and 89Sr
Determination of radiostrontium in milk 11
online with the help of an installed HPLCdetector. Lamb et al. [28] have automatedan entire separation system under compu-ter control.
The pure � emissions from 90Sr, 89Srand 90Y can be measured with a propor-tional counter, a liquid scintillation counterand a surface barrier detector. In the caseof 89Sr and especially 90Y, the high parti-cle energies enable us to use a fourth detec-tion system: the Cerenkov counter. Thedetermination of 90Sr and 89Sr can also becarried out with a non-radiometric techniquebased on mass spectrometry. The limit ofdetection is not systematically given tocompare counting techniques because thegiven values are not determined in thesame conditions (sample volume, chemicalrecovery, counting time…).
3.4.1. Proportional counter
Largely used [3, 20, 21, 26, 31, 33, 39,40, 49] for its convenience, low back-ground (0.5 dpm), significant efficiency(about 50%) and relative immunity from �interferences, the proportional counter isthe oldest technique. Nevertheless, thiscounting system is unable to discriminatethe � particles of each of the elements 90Sr,89Sr and 90Y. Consequently, in the case of90Y elimination prior to radiostrontiumcounting, the content of 90Sr and 89Sr mustbe evaluated by the degree of 90Yingrowth at some interval of time after theSr/Y separation. Broadway and Guy [8]show how a generalized linear system ofequations may be used to solve this prob-lem when considering measurements takenafter additional time intervals. The mathe-matical model can be carried out by usinga simple linear regression. Sutherland [41]compares count data against a spreadsheet-calculated matrix of ingrowth/decaycurves from 0 to 100% 90Sr and at the ana-lytical count intervals between about 24 to650 h after the Sr/Y separation, to deter-mine the best fit of the data set. Therefore,the fit allows him to calculate 90Sr and 89Srpercentages in the sample.
Both methods gave satisfactory resultsand can be carried out with a simple Excelspreadsheet. However, they require about10 measurements, so that 10 d are neededto obtain acceptable standard of deviations(about 3%), leading to a total analyticaltime of over 11 d.
Most often, the limit of detection isabout 5–100 mBq·L–1 for a sample volumeof 1 L, with a counting time of between1–2 h and a chemical recovery of over 50%.
3.4.2. Liquid scintillation detector
Although the lower background obtainedwith ultra low level liquid scintillation(> 3 dpm) is greater than for the pro-portional counter, liquid scintillation is fre-quently used [2, 5, 18, 19, 23, 28, 36, 42–45]because it gives a � spectral distributionfor the three radionuclides. Furthermore,liquid scintillation is characterized by a highcounting efficiency (> 85% for 90Sr and90Y) and avoids counting problems such asself absorption, poor geometry or non-uniformity of sample mounting.
Among methods which attempted to assayfor 90Sr in non-equilibrium, Piltingsrudand Stencel [36] used Cab-O-Sil gellingagent to suspend an insoluble strontiumprecipitate of a sample containing 90Y,90Sr and 89Sr, and a spectrum-unfoldingcomputer program to determine the quan-tity of each of the three isotopes.
As reported by Passo and Cook [35],Heilgeist [22] uses a calculation programthat defines three energy windows on theliquid scintillation analyzer. The calibrationof the spectrometer and the evaluation ofthe sample spectrum require spectra of thebackground and of the reference prepara-tions, which contain only one of the radio-nuclides 89Sr, 90Sr and 90Y. After eliminationof the background, 90Y is calculated on thebasis of its undisturbed share in window�816–950 keV�, 89Sr is determined afterdeduction of 90Y in window �690–815 keV�, and finally, after deduction of89Sr and 90Y, 90Sr is calculated in channels
12 S. Brun et al.
�< 690 keV� , according to the spectra sub-traction method. However, in the casewhere the concentration ratio of 90Sr/89Sris over or under the unit, the deviation ofthe minor component increases as well.
The limit of detection is about 10–100 mBq·L–1 for a sample volume between1–4 L, with a counting time of between1–2 h and a chemical recovery of over 80%[13, 35].
3.4.3. The Cerenkov detector
Cerenkov counting was first employedby Randolph [37]. Owing to the high Cer-enkov counting efficiency for 89Sr and 90Y(> 40%) and the low efficiency for 90Sr(< 1.4%), the activity determined from asample containing a fresh mixture of 90Srand 89Sr is entirely due to 89Sr [12].
The lowest background acquired in lowlevel count mode using glass scintillationvials is about 6 dpm.
Wilken and Joshi [52] mention that theuse of a 85Sr � emitter as a yield monitorshould be avoided if Cerenkov counting isto be performed at very low levels, and anon-radiometric technique is proposed.
Passo and Cook [35] reported thatRucker obtained a limit of detection of0.35 Bq·L–1 for 89Sr and 0.29 Bq·L–1 for90Sr, for 1 L samples, with a 20 min counttime and 80% chemical recovery.
3.4.4. Surface barrier detector
Wilken and Joshi [52] report that Minkused a silicon surface barrier detector sys-tem to obtain � particle spectra. Thesystem is able to resolve the differentpeaks arising in the conversion electronspectra for radionuclides such as 207Bi. Inthe case of � particle emissions from radio-nuclides such as 90Sr, 89Sr and 90Y, theparticle energy varies continuously fromzero to maximum energy. Consequently,the surface barrier detector is unable to dis-criminate these � particle emissions.
The background is of a few counts aday, so the limit of detection can be signi-
ficantly lower than that of the proportionalcounter. The major advantage of this systemis to be able to solve the lead’s problem. Infact, the possible presence of 210Pb (�emitter) in a sample will cause an interfer-ence because of its chemical similaritieswith strontium. By using the surface bar-rier detector the 210Bi daughter of 210Pbcan be detected .
3.4.5. Mass spectrometry
Wendt et al. [51] used a non-radiometrictechnique for ultra traces analysis of 90Srand 89Sr by combining conventional massspectrometry with collinear laser reso-nance ionization spectroscopy. The limit ofdetection for 90Sr is like that of the radio-metric techniques previously described(2 mBq); for 89Sr the limit of detection isstill too high (15 Bq). The techniquerequires a separation of strontium from cal-cium and barium, which would interfereduring the subsequent atomization and ion-ization process.
In the case where the sample containslarge quantities of stable strontium, theselectivity reached is not high enough toenable ultra traces determination of 90Srand 89Sr.
4. DISCUSSION
Following a nuclear reactor accident,milk produced in a contaminated areausually undergoes regular checking forradiostrontium. A rapid analytical method(performed in less than 2 d) is required forthe application of the EURATOM regula-tion [14], in order to screen a large numberof milk samples in a short period of timeand to minimize the contamination of allthe dairy food chain. In the presence ofinterferents such as 140Ba, 134Cs, 137Cs…,existing detectors are unable to adequatelyresolve the spectra of the three radionu-clides 90Sr, 89Sr and 90Y. In addition tocounting technology delay, an importantradiochemical separation is required. Twodifferent approaches are possible for radio-strontium determination. The first is a
Determination of radiostrontium in milk 13
radiochemical procedure based on 90Yanalysis with the intent of determining 90Srand 89Sr separately. It is the best methodfor a specific determination of 90Sr. How-ever, it needs more than 15 d to reachsecular equilibrium, and so is inappropri-ate for managing an emergency situation.
The second consists of measuring radio-strontium activity after several radiochemicalseparations. Many rapid methods are avai-lable according to the nature of the inter-ferences and the level of the selectivityrequired. Among them, though the nitrateprecipitation with fuming nitric acid fol-lowed by several other steps is consideredto be one of the most selective methods, itis tedious and time-consuming. Moreover,the fuming nitric acid manipulation is dan-gerous. On the other hand, Tait et al. [45]have developed a simple and efficient pro-cedure for a crisis situation. The samplepreparation is strongly accelerated byusing cation exchange resin, then the stron-tium separation is carried out by usingcrown ether extraction chromatography.89Sr and 90Sr are determined with a liquidscintillation analyzer and a simple calcula-tion program.
Among all the available detection sys-tems, liquid scintillation is particularly welladapted for the quantification of 90Sr, 89Srand 90Y. Its high background (about3 dpm) is well under the minimum valuerequired for the application of the EURATOMregulation. Furthermore, the ability of thesystem to screen the 3 radionuclides allowsone to save a significant amount of time.
In the case where the proportional coun-ter is used, the 89Sr and 90Sr activities canbe discriminated by following the activitydecay. The mathematical treatment iseffective under one major condition: the
ratio must have a certain min-imal value so that deviation in the observedactivity induced by 89Sr decay is greaterthan the counting uncertainty. The timerequired to determine activity due to 89Srand to 90Sr shortens with the increasingvalue of the ratio.
Detection systems based on mass spec-trometry have the great advantage of signi-ficantly reducing chemical operationscompared with traditional radioactivitycounting. However, these techniques stillrequire the separation of 89Sr and 90Srfrom a large amount of stable strontium.Furthermore, the limit of detection givenfor 89Sr is still too high.
5. CONCLUSION
The comparison of different methods isquite difficult because the objectives arenot always the same. In an emergencysituation related to a nuclear reactor acci-dent, a rapid estimation of the radio-stron-tium present in milk will prevail. In thisway, available rapid methods are able togive a result in less than two days. A signi-ficant amount of time can be saved in thesample preparation steps by using ionexchange chromatography. Furthermore,simple mathematical treatments of theliquid scintillation spectrum allow a rapid89Sr and 90Sr discrimination.
Finally, detection systems based on massspectrometry allow new prospects relatingto 90Sr determination. The radiochemicalseparation can be strongly minimized.
REFERENCES[1] Abbadi S., Diercks H., Knöchel A., Sen
Gupta R., Tödter K., Rapid procedures forthe determination of radioactive strontiumisotopes in food and environmental samples,Kerntechnik 62 (1997) 91–95.
[2] Alvarez A., Navarro N., Salvador S., New method for 90Sr determination in liquid
samples, J. Radioanal. Nucl. Chem. 191(1995) 315–322.
[3] Baratta E.J., Strontium-89 and strontium-90in milk, in: Horwitz W. (Ed.), OfficialMethods of Analysis of AOAC Internatio-nal, 17th edn. AOAC International, Gaithers-burg, Maryland, USA, chapter 13, 2000,pp. 3–6.
[4] Bem H., Bakir Y.Y., Shukerb S.M., A rapidmethod for the determination of 90Sr inpowdered milk, J. Radioanal. Nucl. Chem.147 (1991) 263–268.
A89Sr:A90Sr
A89Sr:A90Sr
14 S. Brun et al.
[5] Benzi P., Operti L., Volpe P., On the reliabi-lity of a rapid method for the determinationof strontium-90 in natural samples, J. Ra-dioanal. Nucl. Chem. 126 (1988) 245–256.
[6] Bouquiaux J., Gillard-Baruh J.C.H., Déter-mination du strontium radioactif après sépa-ration par l’EDTA sur résine cationique,Radiochim. Acta 9 (1968) 153–157.
[7] Bouquiaux J., Gillard-Baruh J.C.H., Sépara-tion du strontium sur résine cationique pourla détermination du strontium radioactif,Radiochim. Acta 14 (1970) 39–43.
[8] Broadway J.A., Guy G.Y., Calculation of 90Sr and 89Sr concentrations in environmen-tal samples using a generalized linear sys-tem, Health Phys. 47 (1984) 458–462.
[9] Brulé G., Les minéraux, in: CEPIL-INRA,(Éd.), Le lait matière première de l’industrielaitière, CEPIL, Paris/INRA, Versailles,France, 1987, pp. 87–98.
[10] Bunzl K., Kracke W., A simple radiochemi-cal determination of 90Sr in environmentalsamples, J. Radioanal. Nucl. Chem. 148(1991) 115–119.
[11] Comar C.L., Wasserman R.H., Strontium,in: Comar C.L., Bronner F. (Eds.), Mineralmetabolism an advance treatise 2, part A,Academic Press, New York, USA, 1964,pp. 523–566.
[12] Cook G.T., Passo Jr C.J., Carter B., Stron-tium-89 and-99, in: L’Annunziata M.F.(Ed.), Handbook of radioactivity analysis,Academic Press, San Diego, USA, 1998,pp. 369–370.
[13] De Regge P., Radecki Z., Moreno J., BurnsK., Kis-Benedek G., Bojanowski R., TheIAEA proficiency test on evaluation ofmethods for 90Sr measurement in a mineralmatrix, J. Radioanal. Nucl. Chem. 246(2000) 511–519.
[14] EURATOM, Niveaux maximaux admissi-bles pour les denrées alimentaires et les ali-ments pour bétail après un accident nucléaireou dans toute autre situation d’urgence ra-diologique, J. Off. Commun. Eur., règlementN° 2218/89 du 22/07/89.
[15] Florou H., Savidou A., Chaloulou C., Stron-tium-90 activity in monthly milk samplesfrom Greece, J. Dairy Sci. 79 (1996)1679–1682.
[16] Fourie H.O., Ghijsels J.P., Radiostrontiumin biological material: a precipitation stepand extraction procedure eliminating the useof fuming nitric acid, Health Phys. 17(1969) 685–689.
[17] Galle P., Toxiques nucléaires, 2nd edn.Masson, Paris, France, 1998.
[18] Gastberger M., Steinhäusler F., GerzabekM.H., Hubmer A., Lettner H., 90Sr and 137Csin environmental samples from Dolon nearthe Semipalatinsk nuclear test site, HealthPhys. 79 (2000) 257–265.
[19] Gjeçi E., Analysis in environmental and bio-logical samples by extraction chromatogra-phy using a crown ether, J. Radioanal. Nucl.Chem. 213 (1996) 165–174.
[20] Goutelard F., Nazard R., Bocquet C.,Coquenlorge N., Letessier P., Calmet D.,Improvement in 90Sr measurements at verylow levels in environmental samples, Appl.Radiat. Isotopes 53 (2000) 145–151.
[21] Grahek Z., Lulic S., Kosutic K., Eskinja I.,Cerjan S., Kvastek K., Separation ofradioactive strontium from natural samplesby means of mixed-solvent anion exchange,J. Radioanal. Nucl. Chem. 189 (1995)141–146.
[22] Heilgeist M., Use of extraction chromato-graphy, ion chromatography and liquidscintillation spectrometry for rapid determi-nation of strontium-89 and strontium-90 infood in cases of increased release of radio-nuclides, J. Radioanal. Nucl. Chem. 245(2000) 249–254.
[23] Horwitz E.P., Chiarizia R., Dietz M.L., Anovel strontium-selective extraction chro-matographic resin, Solvent Extr. Ion Exch.10 (1992) 313–336.
[24] International Atomic Energy Agency,Measurement of radionuclides in food andthe environment, a Guidebook, Tech. Rep.Ser. 295 (1986) 70–91.
[25] Jeter W.H., Grob B., Determination of ra-diostrontium in milk using an extractionchromatography column, Radioact. Radio-chem. 5 (1994) 8–17.
[26] Kimura T., Iwashima K., Ishimori T.,Hamada T., Separation of strontium-89 and90 from calcium in milk with a macrocyclicether, Anal. Chem. 51 (1979) 1013–1016.
[27] Kirchmann R., Rétrospective des études surla décontamination du lait et la répartitiondes produits de fission majeurs dans les déri-vés, Seminary, Radioactivity transfer duringfood processing and culinary preparation,Gembloux, Belgique, 1989, pp. 231–249.
[28] Lamb J.D., Nordmeyer F.R., Drake P.A.,Elder M.P., Miles R.W., Lash R.P., Ion chro-matographic separation for analysis of radios-trontium in nuclear reprocessing solutions ofhigh ionic strength, J. Radioanal. Nucl.Chem. 134 (1989) 317–331.
[29] Macasek F., Gerhart P., Distribution andspeciation of cesium, strontium and euro-pium in the aqueous two phase system milk-
Determination of radiostrontium in milk 15
pectin, J. Radioanal. Nucl. Chem. 186(1994) 87–95.
[30] Migikowsky B.B., Radioisotopes and envi-ronmental circumstances reduction of thehazard in the food chain, in: Caldecott R.S.,Snyder L.A. (Eds.), Symposium on Radioi-sotope in the biosphere, University of Min-nesota, USA, 1960, pp. 293–305.
[31] Mikulaj V., Svec V., Radiochemical analy-sis of strontium-90 in milk, soil and plantsby solvent extraction, J. Radioanal. Nucl.Chem. 175 (1993) 317–324.
[32] Noshkin V.E., Mott N.S., Separation ofstrontium from large amounts of calcium,with application to radiostrontium analysis,Talanta 14 (1967) 45–51.
[33] Ovarec J., Navarcik I., Determination of137Cs and 90Sr in milk by the static methodwith the ion exchanger, J. Radioanal. Nucl.Chem. 121 (1988) 331–335.
[34] Paasikallio A., Rantavaara A., Sippola J.,The transfer of cesium-137 and strontium-90from soil to food crops after the Chernobylaccident, Sci. Total Environ. 155 (1994)109–124.
[35] Passo Jr C.J., Cook G.T., Handbook of envi-ronmental liquid scintillation spectrometry,Packard a Canberra Company, Meriden,USA, section IV (1994) 9-1–9-10.
[36] Piltingsrud H.V., Stencel J.R., Determina-tion of strontium 90Y, 90Sr and 89Sr in sam-ple by use of liquid scintillation betaspectroscopy, Health Phys. 23 (1972)121–122.
[37] Randolph R.R., Determination of strontium90 and strontium 89 by Cerenkov and liquidscintillation counting, Appl. Radiat. Isoto-pes 26 (1975) 9–16.
[38] Shabana E.I., Al-Hussan K.A., Al-JaseemQ.K., Adaptation of radioanalytical methodfor strontium 90 to different kinds of envi-ronmental samples and its application for theinvestigation of some samples in the Riyadhregion, J. Radioanal. Nucl. Chem. 212 (1995)229–240.
[39] Stella R., Ganzerli Valentini M.T., MaggiL., Ion exchange chromatography on a newform of tin dioxide for the isolation of stron-tium radioisotopes from fission products: anapplication to milk samples, Appl. Radiat.Isotopes 41 (1990) 905–908.
[40] Stella R., Ganzerli Valentini M.T., MaggiL., Determination of 90Sr in milk by usingtwo inorganic exchangers, Appl. Radiat. Iso-topes 44 (1993) 1093–1096.
[41] Sutherland J.K., A spreadsheet method for90Sr and 89Sr interpretation, Health Phys. 54(1988) 69–72.
[42] Tait D., Wiechen A., Improvements to arapid method for separating strontium fromliquid milk by treatment with a chelatingresin and crown ethers, J. Radioanal. Nucl.Chem. 159 (1992) 239–247.
[43] Tait D., Wiechen A., Haase G., Binding ofSr from milk by solid phase extraction withcryptand C222 sorbed on silica gel, cationexchange, chelating or adsorbent resins forsimplified 90Sr analysis, Sci. Total Environ.173/174 (1995) 159–167.
[44] Tait D., Haase G., Wiechen A., Rapid andefficient separation of strontium from liquidmilk with a cation exchange resin Dowex50WX8 treated with cryptand 222, J. Ra-dioanal. Nucl. Chem. 226 (1997) 225–228.
[45] Tait D., Haase G., Wiechen A., A fastmethod for the determination of Sr-90 inliquid milk by solid phase extraction withcryptand 222 on cation exchange resin,Kerntechnik 62 (1997) 96–98.
[46] Tait D., Haase G., Wiechen A., Use of a sin-gle strontium binding resin in a batch pro-cess followed by elution and precipitation ofSrCO3 for the efficient analysis of 89Sr and90Sr in liquid milk, Environmental radioche-mical Analysis, Special Publication, NewtonG.W.A. (Ed.), Royal Society of Chemistry,234 (1999) 192–200.
[47] Thomas C.W., Determination of strontium-90 in milk, Health Phys. 8 (1962) 407–415.
[48] Twardock A.R., Prinz W.H., Comar C.L.,The state of calcium and strontium in goat’smilk, Arch. Biochem. Biophys. 89 (1960)309–312.
[49] Vaney B., Friedli C., Geering J.J., Lerch P.,Rapid trace determination of radiostrontiumin milk and drinking water, J. Radioanal.Nucl. Chem. 134 (1989) 87–95.
[50] Walstra P., Jenness R., Dairy Chemistry andPhysics, Wiley-Interscience Publication,John Wiley & Sons, New York, USA, 1984.
[51] Wendt K., Kratz J.V., Lantzsch J., Müller P.,Nörtershäuser W., Seibert A., Trautman N.,Waldek A., Zimmer K., Rapid ultratrace de-termination of 89,90Sr in environmental sam-ples by collinear laser resonance ionisationspectrometry, Kerntechnik 62 (1997) 81–84.
[52] Wilken R.D., Joshi S.R., Rapid methods fordetermining 90Sr, 89Sr and 90Y in environ-mental samples: a survey, Radioact. Radio-chem. 2 (1991) 14–27.
[53] Wong N.P., Jenness R., Keeney M., MarthE.H., Fundamentals of Dairy Chemistry, 3rdedn., Van Nostrand Reinhold, New York,USA, 1988.
