Tiêu chuẩn iso tr 18336 2016

© ISO 2016 Guidelines for good XRF laboratory practice for the iron ore industry Lignes directrices de bonnes pratiques de laboratoire de spectrométrie de fluorescence de rayons X pour l’industrie du[.]

Guidelines for g ood X RF laboratory

Ligne s direc trice s de b n e s pratiques de la oratoire de

spectrométrie de flu re scenc de ra o s X p ur l’industrie du

COPYRIGHT PROTECTED DOCUMENT

© ISO 2016, P blshed in Sw itz rlan

A ll rig hts r eserved Unles otherw ise spe ified, nopar of this p blc tion ma y be r epr od c d or utilz d otherw ise in an form

or b an me ns, ele tr onic or me hanic l, inclu in p oto opying , or postin on the internet or an intranet , w ithout prior

written permis ion Permis ion c n be req esed from either ISO at the ad r es below or ISO’s member bod y in the c u try of

F reword i v

Introduction v

1 Sc ope 1

2 Re g ents 1

3 A pparatus 2

4 F sed glas be ds 5

4.1 General 5

4.2 Storag e 5

4.3 Disc making pr ecision 5

4.4 Bead dis ortion (curv tur e and flatnes ) 6

5 Qual ty c ontrol 6

5.1 S le tion of Q C samples an freq ency of pr ep ration 6

5.2 A nalysis of Q C an analytical samples 7

5.3 Contr ol chart 7

5.4 Particip tion in pr oficiency tes pr og rams 8

A nne x A (informative)Results for flux los on ig nition testing 9

A nne x B (informative)Proc edur e to check dis making precision 10

A nne x C (informative) Meth d to determine relationship betwe n height and c onc entratio 11

A nne x D (informative) Pr oductio of a height adjustable cup

12 A nne x E (informative) Be d me sur ement ap aratus 13

A nne x F (informative) Flow she t for fused be d qual ty 14

A nne x G (informative) Microsof E c el pr og ram for calculating dis pr ecisio 15

A nne x H (informative)Data input s reen for calculating dis precisio 19

A nne x I (informative)L s of ac curacy w ith no los of pr ecision 20

A nne x J (informative) L s of ac curacy w ith los of pr ecision 22

A nne x K (informative) L s of ac curacy an precisio — L s of be d making pr ecision 23

A nne x L (informative) Results for spectr omete precision test 25

A nne x M (informative)Dr if c or r ectio 26

Biblog raphy 27

ISO (he Int ernational Org nization for Stan ardization) is a worldwidefede ation of national s an ards

b dies (ISO membe b dies) The work of pr p ring Int ernational Stan ards is normaly car ied out

through ISO t ech ical committ ees Each membe b dy int er st ed in a subje t for w hich a t ech ical

committ ee has be n es a lshed has the right t o be r pr sent ed on that committ ee Int ernational

org nizations, g ove nmental an non-g ove nmental, in laison with ISO, also take part in the work

ISO cola orat es closely with the Int ernational Ele trot ech ical C mmis ion (IEC) on al matt ers of

ele trot ech ical s an ardization

The proc d r s used t o develo this document an those int en ed for it furthe maint enanc ar

desc ibed in the ISO/IEC Dir ctives, Part 1 In p rticular the dife ent a pro al c it eria ne ded for the

dife ent ty es of ISO document should be not ed This document was draft ed in ac ordanc with the

edit orial rules of the ISO/IEC Dir ctives, Part 2 ( e www.iso.org dir ctives)

A tt ention is drawn t o the p s ibi ity that some of the element of this document ma be the subje t of

p t ent right ISO shal not be held r sponsible for identifying any or al such p t ent right Detais of

any p t ent right identified d ring the develo ment of the document wi be in the Introd ction an / r

on the ISO ls of p t ent de larations r c ived ( e www.iso.org p t ent )

Any trade name used in this document is information given for the convenienc of use s an does not

cons itut e an en orsement

F or an ex lanation on the meaning of ISO spe ific t erms an ex r s ions r lat ed t o conformity

as es ment, as wel as information a out ISO’ s adhe enc t o the WTO principles in the Te h ical

Bar ie s t o Trade (TBT), se thefolowing URL:F or word — Sup lementary information

The committ ee r sp nsible for this document is ISO/TC 1 2, Iro or a d dir c t r duc ed iro ,

Subcommitt ee SC2, Chemic al a alys i s

This Te h ical R eport is int en ed for use in conju ction with othe Int ernational Stan ards for the

chemical analy sis of iron or s Although it was writt en for a high through-put iron or la orat ory,

the proc d r s desc ibed can be modif ied t o suit othe in us ry or la orat ory r q ir ment Some

la orat ories ma f in the r commen ed fr q ency of t es ing r commen ed b this Te h ical R ep rt

t o be ex ces ive for their situation or the pr cision r q ir d b them In this case, the o e at or ma use

their informed disc etion t o ada t her commen ations of the guidel nes t o their situation

Guidelines for g ood X RF laboratory practic e for the iron

This Te h ical R eport spe ifies r commen ed q alty control proc d r s for XRF la orat ories

o e ating within the iron or in us ry

Al r ag ent ( inclu ing fusion fluxes an calbration r ag ent ) should be pur hased from r puta le

supple s an should me t the minimum r q ir ment for purity as lst ed in ISO 9 1 -1 Al r ag ent

should ha e a b t ch n mbe an , w he e a ai a le, a c rtif icat e of analy sis Detai s of pur hased r ag ent

( up le , amou t pur hased, q al ty, an b t ch n mbe ) should be r corded These r cords should

inclu e w hat the r ag ent ar used for F or bat ches of flux, the r cords should in icat e w hich samples

we e analy sed with a particular b t ch

2.1 F sion flux

As the levels of contamination ma v ry from b t ch t o b t ch of flux, the purity of fusion flux es should

be che ked prior t o use This can be achieved b fusing d plcat e beads of pur si ica an iron with

the new flux, an analy sing these along with beads pr par d using a pr viously t est ed (c rtified) flux

Backgrou d conc ntrations should not ex ce d 1 p m t o 2 p m (as comp r d t o a c rtif ied b t ch of

flux) for each of the folowing o ides Mn

3O

4, SnO

2, V

2O

5, Cr

2O

3, Co

3O

4, NiO, CuO, ZnO, As O

3, PbO, BaO,

Na

2

O an Cl an the sum of the p sitive dife enc s should not ex ce d 40 p m t o 5 p m

The conc ntrations of the o ides should not ex ce d 0,0 % for each of the fol owing o ides F e

2O

3, SiO

2,

2

O an P

2O

5, (he a solut e sum of the dife enc s should not ex ce d 0,0 % ),

w hie backgroun s should not dife b mor than 0,0 % Sulfur (r port ed as SO

3) can fr q ently v ry

b 0,0 % Whe e flux does not conform t o spe if ications, a se on d plcat e set of beads (made with old

an new flux)should be pr par d b a dife ent o e at or on the same da , or b the sameo e at or on a

dife ent da If the mat erial fais t o me t the minimum spe if ications, the sup le of non-conforming

flux should be contact ed an a r plac ment b t ch o tained an t est ed

Whe e non-significant deviations ar o se ved for major an trac element between flux bat ches,

these beads can be used t o updat e calbration int er ept In al cases, r cords of calbration prior an

aft er amen ment hould be kept

Prior t o cal bration amen ment, the conc ntration levels of al pr viously analy sed blank beads (prior

t o calbration amen ment) should be plott ed, an tr n s not ed If conse utive set of d plcat e beads

show consist ent p sitive conc ntration inc eases, pr vious beads should be r fused an r -ru , an the

tr n s conf irmed or neg t ed

Whe e la orat ories ele t t o use ad itive flux es (o idizing, r lease ag ent or int ernal s an ard

comp u ds), the homog eneity of the flux should be t est ed, as es ed an compar d a ains the q ot ed

q alty or a ains a flux b t ch that is k own t o be homog enous Tes ing methods inclu e dir ct

measur ment of ad ed analyt es, or in ir ct measur ment of a q alty p ramet er ( ignition los ) An

ex mple of flux t es ing r sult can be fou d in Annex A

As cal brations would ha e be n amen ed, tr n s wi l be se n as neg tive v lues pro r s ing t owards

a mor positive r sult If p s ex e t ed tr n s can ot be r plcat ed, the XRF ins rument (cal bration,

monit or) should be inspe t ed If pr viously se n tr n s ar r peat ed, flux sup le s should be contact ed,

an the pro lem discus ed

2.2 Calbration reag ents

R eag ent used for XRF r calbration should be che ked in a simiar manne t o that used t o che k flux

(b pr p ring b th the old r ag ent an the new r ag ent in the same ty e of flux) He e the level of

contaminant should not ex ce d those r p rt ed on the r ag ent up le ’ s c rtif icat e of analy sis Please

not e that it is common for r ag ent sup le s t o r port r ag ent purity b sed on an “as dife enc ” b sis

Conseq ently, a 9 , 999 % r ag ent ma ha e only be n analy sed for a single contaminant w hose cont ent

is les than 1 p m Howeve , this r ag ent ma contain othe contaminant w hich have not be n analy sed

w hose conc ntration ma be signif icant

Alt ernatively, if no in-house high purity r ag ent is a aia le, smal q antities of analy sed r ag ent ma

be o tained from r puta le la orat ories In ad ition, w he e r ag ent ar suspe t ed of contamination,

they can be ext ernal y analy sed

Al eq ipment used t o pr p r an measur fused glas beads should be che ked on a r gular b sis in

ac ordanc with the sched le set out in Ta le 1 The fr q encies defined in Ta le 1 ar those r q ir d b

a high through-put iron or la orat ory As this Te h ical R ep rt is a guidel ne rathe than a pr sc iptive

s an ard, la orat ories with les e deman s can modify the f igur s ac ordingly

3.1 Be d pr eparatio equipment

Al eq ipment used for the prod ction of fused glas beads ( uch as b lanc s, fusion war , fusion

furnac s an sample drying eq ipment) should be ins aled, maintained an o e at ed in ac ordanc t o

the man factur r’s r commen ations The ext ernal surfac s of al fusion furnac s should be inspe t ed

at the commenc ment of each shif for ex ces ive d s an glas fra ment or glas spi s If any spi lag e

is det ect ed it should be cleaned up befor furthe use On a we kly b sis, or if spi ag e has oc ur ed,

fusion furnac s should be alowed t o co l an the int erior should be inspe t ed for fault (broken or

lo se furnac l nings that ma contaminat e samples) and cleaned (an r pair d if ne es ary) Fusion

furnac t empe atur s should be che ked we kly for ac uracy an u iformity of t empe atur using a

calbrat ed r fe enc the mocouple As a g ene al rule, an in epen ent pe formanc au it of bead making

eq ipment ( inclu ing any environmental fact ors)should be pe formed yearly

Table 1 — Summary of frequency of tests and procedures

a

Monit or data should b checked for drif (3.2.4).If pro lems/is ues

are suspected within the instrument, then inc e sed monitor samples should b ru

throughout the day

A t the commencement of every shif , al la oratory per on el involved in the prod c

-tion of XRF fusion b ads should prepare at le st one q ality control sample (reference

or certif ied reference material) in d plicat e (5.1)

A t the commencement of every shif , al la oratory per on el involved in the prod c

-tion of XRF fusion b ads should prepare at le st one q ality control sample (reference

or certif ied reference material) in d plicat e (5.1) These specimens should b me

s-ured an the results evaluat ed b fore an u k own specimens are validated

Monitor chiled water flow rate an t emperature (3.2.2)

Monitor compres ed air pres ure (3.2.2)

Monitor detect or g s flow rat e an pres ure (3.2.2)

Monit or instrument er or lo s (3.2.2)

Inspect fusion furnaces an cle n if neces ary (3.1)

a

Monitor is used e clusively for drif c r e tion Where the power set in s have not chan ed more than 1 k W fom the

nominal operatin power an the spe t ometer has be n shown to have a hieved s abi ty over a desired time f ame (3.2)

Monitor u dates c n be perormed for the c r espon in time fame

Fr quency Tes

We kly C eck resolution an p lse shif for flow detector (3.2.2)

Perform dis making precision t ests for new operat or Once competent, perform

checks monthly (4.3)

Cle n int erior an e terior of fusion furnaces (3.1)

C eck furnace t emperature an u iformity of t emperature (3.1)

Fortnightly Back u spectrometer data f iles,calibrations an conf igurations (3.2.3)

Me sure flatnes of moulds an casting dishes (4.4)

Monthly Perform dis making precision tests for compet ent operat or (4.3)or automat ed

eq ipment

Cle n monit or (3.2.4)

T re -monthly Prepare sy thetic calibration stan ards ( Sy Cals) in q adru licate an compare

results to calibration data (5.3)

Half-yearly an prior to

calibration or af er

major repair

Perform precision tests in ISO/TR 1 2 1

Perform long -t erm sta ility t ests in ISO/TR 1 2 1

Perform detector line rity tests in ISO/TR 1 2 1

Ye rly C eck al b ad preparation eq ipment b lances, fusion ware, fusion furnaces an

en ironmental con itions)(3.1)

a

Monitor is used e clusively for drif c r e tion Where the power set in s have not chan ed more than 1 k W fom the

nominal operatin power an the spe t ometer has be n shown to have a hieved s abi ty over a desired time f ame (3.2)

Monitor u dates c n be perormed for the c r espon in time fame

3.2 X RF Spectrometer

3.2.1 General

Al XRF ins rument should be t est ed t o ensur conformanc t o ISO/TR 1 2 1 for ins rument pr cision

an det ect or lnearity Pr cision t es ing should be pe formed twic yearly, or w heneve major r p irs

t o an XRF (r plac ment of det ect or, X-ra tube, g ene at or or measur ment ele tronics) or t o the chi led

wat er cir uit has be n pe formed

F or al ins rument t es s, a r lative pr cision of bett er than 0,0 % coeff icient of v riation is r q ir d

for the analy sis of F e in iron or s, and so spe tromet er pr cision an l nearity t es s should be car ied

out using ac umulat ed count rat es of 2 × 1

7

count for each measur ment Whe e separat e cou ting

channels ar used for the dife ent element ( imultaneous ins rument ), or w he e the det ect or is

chang ed, the dead time of each chan el an each det ect or should be det ermined in epen ently, using

an a pro riat e wa elength for the det ect or/c y stal combination

L ng t erm (24 h) XRF s a i ty t es s should be pe formed using v rious kV an mA set ings, det ect or,

col imat or, an c y stal combinations Al t es s should be con uct ed biann al y at a signif icanc level of

0,0 % (2 × 1

7

cou t ), an immediat ely prior t o pe forming an XRF calbration Ins rument should

exhibit short-t erm s a i ity in the orde of 0,0 % o e a 6 h t o 1 h pe iod If an XRF fais t o me t

desir d pr cisions, ins rument man factur rs should be consult ed an the cause of p or ins rument

s a i ty r ctified

A dditional spe tromet er t es s can be fou d in Annex es I, J an L

3.2.2 Each shif/we kl y spectrometer checks

As the o e ation an pe formanc of an XRF spe tromet er is highly depen ent on the q alty of ext ernal

se vic s such as chi led wat er ( flow rat e an t empe atur ), compr s ed air (pr s ur ) an det ect or

g s (flow rat e an pr s ur ), the s atus of these sup l es should be monit or d on a daiy b sis an the

Table 1 (c ontin ed)

r sult r corded in a che k she t In ad ition, e ror lo s should be che ked dai y an cor e tive action

taken w hen fa ult ar r p rt ed As spe tromet ers p s es int ernal se vic s ( uch as chi led wat er flow,

t empe atur an con uctivity, int ernal spe tromet er t empe atur an ins rument v cu m levels), the

s atus of these should be monit or d Not e that as al mode n XRF sof war p s es es spe tromet er

s atus sc e ns, these should be con gur d t o be o en in a minimised win ow so that the s atus of

spe tromet er se vic s can be monit or d

As the pe formanc an con ition of spe tromet er flow det ect ors ar highly dependent on the q alty

of det ect or g s, the r solution an pulse shif should be che ked on a we kly b sis Che ks should be

pe formed with r g rd t o the man factur r’s r commended proc d r Minor chang es in pulse shif

(≥ , ≤ ) should be cor e t ed (not e as spe tromet ers ty icaly ha e dife ent det ect or v ltag e set ings

for each c y stal, al should be che ked w hen t es s an amen ment ar pe formed) Signif icant chang es

in det ect or r solution or pulse shif p sition should be inves ig t ed as these ma be caused b win ow

faiur or g s contamination (leaks)

3.2.3 Hard war e and sofware backup

A b ck p of spe tromet er data files, calbrations an conf igurations should be pe formed eve y two

we ks This b ck p should be st or d on a r mot e comput er or shar d data st orag e ar a XRF sof war

can be st or d in a se ur ar a within thela orat ory

It is also highly r commen ed that XRF sof war an spe tromet er data (calbration f iles an

conf iguration set ings) be lo ded on seve al comput ers (la t ops pr fe r d) w hich pos es the ne es ary

hardwar ( e ial p rt )r q ir d t o conne t an commu icat e with thespe tromet er These comput ers

wi se ve as a hardwar an sof war back p

3.2.4 Instr ument mo itor

To compensat e for drifs in X-ra tube output int ensity, al X-ra measur ment should be made

r lative t o a monit or disc Although dife ent monit or discs could be used for each component, it is

mos convenient t o use a single disc, containing al comp nent t o be measur d The r q ir ment of

the monit or disc is that it be s a le o e a long pe iod Also, the monit or should contain suff icient of

each element t o ensur that the int ensity of each analytical l ne is much highe than the int ensity of

the b ckgroun an can be measur d with the r q ir d pr cision in a r asona le time Suita le s a le

monit or discs made for the analy sis of iron or ar a ai a le comme cialy

The es ential pro e ty of the monit or is it long -t erm s a i ity It should be flat so that it can be plac d

r prod cibly in the sample holde of the XRF, and the analytical surfac should be pol shed for easy

cleaning Monit ors should be cleaned in ac ordanc with man factur r’ s spe if ications or eve y two

months with ethanol or ac t one In-house pr p r d b rat e monit ors ar not r commended as they ar

uns a le an det eriorat e over time

The ac umulat ed count for the v rious element should be such that they ar highe than those from

iron or s, so that pr cision is not l mit ed b the monit or F or the major element , inclu ing F e, the cou t

rat e ma be les than o se ved from iron or s, but it should be high enough that, in a short cou ting

time, the counting e ror is sufficiently smal Monit ors should be run eve y 4 h t o 6 h d ring, an prior

t o thecommenc ment of routine analy ses

Monit or cou t timescan be det ermined at he time of cal bration using the formula below

The monit or count time TM ( in se onds) for each element is given b F ormula (1):

F or F e, the int ensity should be measur d using 1 0 % F e

2O

3

T is the r q ir d co nting time ( in se on s), for each element as de ived in ISO 9 1 -1:2 0 , 7.2.6;

RM is the int ensity, c /s, from the monit or for each analyt e, measur d for 1 s

Whenever XRF spe tromet er p wer set ings ar chang ed mor than 1 k W, ins rument tube ins a i ty

ma oc ur Depen ing on the ty e of spe tromet er an X-ra tube design, this ins a i ty ma pe sis

for 3 min t o 6 min C nseq ently, prior t o measur ment, X-ra eq ipment should be powe ed up for a

suita le time an al owed t o s a i ise

Whe e spe tromet ers ar maintained at a s an b set ing that is les than 0,5 kW of the nominal

o e ating p wer, monit ors ne d not be run prior t o the commenc ment of analy sis Monit or ac eptanc

control chart should be plott ed on Shew hart chart using the lmit spe if ied in the control chart

se tion of this Te h ical R eport

3.2.5 Calbration val datio

Onc a cal bration has be n pe formed in ac ordanc t o ISO 9 1 -1, the method should be v ldat ed

( ruenes t es ing an conformanc t o method pr cision) using a pro riat e r fe enc s an ards The

c rtif ied r fe enc an othe r fe enc mat erials used t o v ldat e a calbration should be chosen using

ISO 1 042 for conc ntration an matrix element

In ad ition t o v l dating a calbration using r fe enc mat erials, si con dio ide an iron o ide, pur

an binary calbration s an ards an synthetic calbration s an ards ( SynC ls) should be analy sed as

un nowns t o v ldat e the calbration Not e, as the calbration an calculation alg orithms ar dife ent,

a slght bias betwe n the r gr s ed and measur d chemical conc ntrations ma oc ur This bias should

be not mor than 1,5 times the counting s atis ical e ror (measur ment pr cision) for 1 % of the

analy sed samples

Al samples, r g rdles if they ar calbration, q al ty control or for routine analy sis should be pr p r d

with eq al car an conside ation Conseq ently, al samples should be pr p r d using the proc d r s

l st ed in ISO 9 1 -1:2 0 , 7.1.2 t o 7.1.7 using compet ent o e at ors To ensur that theq al ty of analy ses

ar not o e at or compromised or biased, al bead pr p rations should be o e at or trac a le

4.2 Storag e

Beads should not be st or d in desic at ors containing any drying ag ent ( uch as si ica g el) or in drying

o ensas e roneousd s p rticles wi l contaminat e the analytical surfac It is r commen ed that beads

be clearly la eled an plac d in clean r -seala le PVC b gs prior t o analy sis an for long t erm st orag e

Whe e beads ar t o be analy sed within 3 min of pr p ration, co led beads should be trans e r d

dir ctly from the cas ing dish/moulda le int o a clean spe tromet er sample tra

4.3 Dis making pr ecision

Al pe sonnel/eq ipment inv lved in the prod ction of XRF samples should be t est ed for disc making

pr cision Initial y, o e at ors should be taught b an o e at or w ho has pro en ex e ienc in the

pr paration of fused beads Disc making should be che ked we kly, unti such a time as beads can be

made consist ently t o conform t o r q ir d la orat ory pr cisions (normaly, thr e conse utive ru s)

The eaft er, disc making pr cision should be che ked monthly A g ene al proc d r for che king disc

making can be foun in An ex B, with a calculation pro ram for measur ment pr cision in Annex G

Whe e aut omat ed bead making faci ities ar used (weigh s ations and fusion l nes), al pr paration

l nes should be t est ed

As a g ene al guide, r ading e rors should not ex ce d s atis ical cou ting e rors, w hie disc making

e rors should not ex ce d r ading e rors b mor than 1,5 % of the theor tical cou ting e ror If

r ading e rors (measur ment pr cision) ex ce d those ex e t ed for cou ting times, this is in icative of

ins rument pr cision e rors (non-s a i ty of det ect ors, tube, col imat ors, lo ding e rors) Alt ernatively,

this could in icat e insufficient ins rument warm up time ( ta i isation time) If conse utive ru s

(pe formed at long er ins rument warm up pe iods) fai t o me t desir d ins rument pr cisions,

ins rument man factur rs should be consult ed an the cause of p or ins rument s a i ty r ctif ied

If disc making e rors (pr cision) ar in exces of r ad e rors b 1,5 %, beads should be broken an

r fused/ re-p ur d using ex ctly the same fusion w ar (moulds an moulda les) used t o pr p r the

beads Ifin ivid al r analy sed bead e ror tr n s cor elat e pr viously analy sed beads, this is in icative

of an in e ent pro lem in the composition of the bead This could be d e t o incomplete drying of flux

or o idant or incomplet e ignition of sample Alternatively, o e ators ma ha e b lanc pro lems, p or

weighing t ech iq e, bead curv atur pro lems or q antitative los of material d ring trans e o e ations

If disc making is foun to impro e folowing r fusion an r analy sis (XRF r ading e rors ar cons ant)

this is in icative that the originaly cas beads we e not homog eneous or we e p or ly pr p r d beads

(c acked, u dis olved mat erial c y stal isation) If this oc urs, the o e ation of the fusion eq ipment

should be che ked to ensur that the t empe atur within the furnac , the a itation an the bead co lng

o e ation ar al u iform Whe e a significant los in the a i ty t o cas beads is encou t er d, inspe tions

an t es s of weighing an cas ing eq ipment should be pe formed ( e An ex K) If these ar fou d t o

be in orde , the fusion p ramet ers (a itation time, spe d et c ) should be adjust ed an the bead making

pr cision r che ked using fr shly weighed samples Not e that w hen man aly o e at ed fusion eq ipment

is used, dife enc s in the p uring time an co lng rat esha e be n fou d t o afe t pr cision

The r sult of al pr cision beads ( inclu ing the o e at or name, dat e, eq ipment used an the r sult of

any cor e tive actions used t o impro e bead pr cision) should be ta ulat ed an st or d

4.4 Bead distortion (curvature and flatnes )

As the X-ra int ensity (an henc conc ntration) g ene at ed from an analytical sample is depen ent on

it r lative g eometry from the det ect or an X-ra ex citation sour e, impe fe tions in the curv tur

an flatnes of XRF beads (dist ortion) wi l dir ctly afe t the ac uracy of analytical det erminations

As the r lationship between bead dist ortion an measur ment ac uracy (measur d conc ntration)

is u iq e for each spe tromet er configuration an bead pr paration t ech iq e (diution an bead

siz ), this r lationship should be det ermined for each spe if ic combination An ex C desc ibes a

method t o det ermine the r lationship betwe n bead dist ortion an conc ntration, w hie An ex D

depict modif ications t o t es eq ipment Onc a r lationship has be n det ermined, bead dist ortion

( impe fe tions in cas ing a p ratus) should be as es ed an r ctified using the eq ipment an

proc d r desc ibed in Annex E an An ex F Fusion war -spe ific r forming t ools can be pur hased

from the a pro riat e ven ors

Alt ernatively, moulds an cas ing dishes should be r formed on a we kly or monthly basis, or w heneve

flatnes is suspe t ed as a sour e of analytical e ror

5 Qual ty control

5.1 S lection of Q C samples and frequency of preparation

A t the commenc ment of eve y shif, al la orat ory pe son el inv lved in the prod ction of XRF fusion

beads should pr par at leas one q alty control sample (r fe enc or c rtif ied r fe enc mat erial)

in d plcat e 5.1 These spe imens should be measur d an the r sult ev luat ed befor any un nown

spe imens ar v ldat ed

Qualty control samples should be chosen in ac ordanc with ISO 1 042 for the r p rted element

Note that any r fe enc mat erial that has be n made t o conform t o ISO Guide 3 is ac epta le for

in-house use Whe e p s ible, the same s an ards that we e used t o v l date the calbration should be

used as Q C s an ards

The use of d pl cat e beads is highly r commen ed as they can be used t o as es o e at or disc making

proficiency an short-t erm pr cision

Pur si icon dio ide (99, 99 % containing les than 3 μg g of each of the othe o ides lst ed in

ISO 9 1 -1:2 0 , Ta le 1) should be used as a q alty control sample t o monit or contamination of

r ag ent an deviations in b ckgrou d eq iv lent conc ntrations Duplcat e beads should be pr p r d

at he commenc ment of each shif an the r sult plott ed Eve y thr e months, q adruplcat e synthetic

calbration s an ards ( SynCals), as desc ibed in ISO 9 1 -1, should be pr p r d and the r sult

comp r d t o those of the cal bration data

R esult for Q C samples should be s atis icaly analy sed using the truenes t es spe if ied in

ISO 9 1 -1:2 0 , 9.6 an the r sult plott ed in an a pro riat e control chart containing Z scor s Fused

bead samples w hich fai t o p s the truenes or pr cision t es s should be r run or r fused If a dife enc

betwe n subseq ent XRF analy ses is gr at er than four times the iron analytical e ror ( y ical y 0,0 %

t o 0,0 % ), the s a i ity or lo ding me hanics of the spe tromet er should be t est ed in ac ordanc with

ISO/TR 1 2 1

R eanaly sed beads w hich s i l fai truenes or pr cision t es s should be r fused an analy sed Samples

w hich fai subseq ent r fusing an analy sis should be r analy sed as an in ependent d plcat e p ir

Whe e conse utive d plcat e beads fai truenes t es s, the spe tromet er an analytical b lanc

pr cision should be as es ed In ad ition, the o e ation ofthe fusion eq ipment (mixing, fusion time

an t empe atur ) an con ition of the fusion war (bead curv tur an contamination) should be

che ked an a pro riat e cor e tive action taken

5.2 A nal ysis of QC and anal ytical samples

The ac uracy an pr cision of routine analy ses ar depen ent on sampl ng an pr paration proc d r s

Conseq ently, analytical samples should be pr p r d in ac ordanc with ISO 3 8

W ithout u de taking multiple sample analy ses it is imp s ible t o asc rtain w hat pe c ntag e of the

o e al pr cision can be at ribut ed t o sample pr paration an w hat can be at ribut ed t o measur ment

pr cision W ith r spe t t o analytical data, although it is desira le t o analy se routine samples in

d plcat e, in many cir ums anc s this is impractical d e t o o e ational l mitations Nevertheles , a

pro ortion of analytical samples (5 % t o 1 % )should be measur d in d plcat e

In an ideal situation, the pr cision of sample pr p ration should be eq al for b th routine samples an

c rtif ied r fe enc mat erials used for q alty control data The efor , w he e a signif icant pr cision

dife enc is o se ved between q alty an routine analy ses, this can be at ribut ed t o p or sample

pr paration In this case, the pro lem should be inves ig t ed an cor e t ed

5.3 Control char ts

Chart should be cons ruct ed in ac ordanc with ISO 7 7 - 2 for each q al ty control s an ard an for

each XRF monit or Chart should be cons ruct ed for eve y la orat ory t echnician an XRF ins rument

Whe e multiple analytical b lanc s an fusion eq ipment (furnac an fusion war ) ar used in a

la orat ory, samples should be pr p r d in a wa that al eq ipment is uti iz d The Shew hart chart

should be updat ed eve y time a Q C sample /CRM is analy sed Significant long t erm tr n s ( hr e

conse utive biased r sult ) should be inves ig t ed an a pro riat e cor e tive measur s taken t o

cor e t this bias Pr fe a ly, a single chart for each XRF ins rument should be cons ruct ed w he e the

t ech ician an pe mutations of fusion furnac s an b lanc s ar clearly identif ied

Whie the truenes tes for each use of the CRM s i l a ples ( e ISO 9 1 -1:2 0 , 9.4 ), the use of the

Shew hart chart is r commen ed t o identify outle s, tr n s and ru s, folowing the principles outlned

in ISO 7 7 - 2, using “t es s for as igna le causes” Statis ical ev luations should inclu e information

pe taining t o short (bead d plcate) an long te m bead making pr cision of o e at ors an bead

making eq ipment In ad ition, the short an long t erm spe tromete pr cision should be as es ed on

a we kly b sis

F or pur si icon dio ide, the mean trac element conc ntrations should be not mor than two times

the measur d backgroun conc ntration (o tained d ring calbration) F or a cor e tly calbrat ed

XRF ins rument, this should be a pro imat ely eq al t o half the q antification l mit de ived from

ISO 9 1 -1:2 0 , Ta le 7 (using S = s andard deviation) Ex ces ive b ckgrou d conc ntrations ( in the

orde of 1 times the measur d XRF v lue) ar in icative of gros contamination from fusion war or

fusion furnac s

Short-t erm monit or drif should be not mor than 0,1 % v riation in count between monit or

measur ment for iron, si ica, calcium, mang nese, aluminium, titanium, ma nesium, phosphorous,

sulfur an p tas ium If this lmit is ex ce ded, al measur ment should be r peat ed Su den e ratic

chang es in monit or cou t should be cor e t ed prior t o analy sis L ng t erm monit or drif should not

chang e mor than 0,5 % pe year for at omic mas es gr at er than p tas ium F or at omic n mbe s les

than p tas ium, long t erm drif should not v ry mor than 2 % pe year Due t o tube ag eing, monit or

v lues ma eithe de r ase (l ght element ) or inc ease over time

If drif ex ce ds these figur s, o e at ors should inves ig t e the ca use of the drif

5.4 Participation in pr oficiency test prog rams

Al la orat ories should p rticip t e in roun ro in proficiency t es ( int er-la orat ory comp rison)

pro rams w hich conform t o the guidelnes s ipulat ed in ISO/IEC 1 043 These proficiency pro rams

should be used t o ev luat e the la orat ory’ s ac uracy and the short-t erm (r plcat e e ror) an long

t erm analy sis pr cision In ivid al la orat ories r sult should be ranked t o the o e al la orat ory

pe formanc b means of Z scor s

The r sult of prof iciency pro rams should be made r adiy a ai a le t o al la orat ory s af

Pr fe ential y, r sult should be r p rt ed gra hical y an should inclu e a clear an concise

int erpr tation of ac uracy an pr cision of data

La oratories should ensur that suff icient r sour es ar commit ed to maintaining r cords,

inclu ing the outcomes of inves ig tions of any unsatis actory r sult an subseq ent cor e tive or

pr ventative actions A la oratory should be a le to draw it own conclusion a out it pe formanc

from prof iciency tes s

A nne x A

Results for flux loss on ig nition testing

Table A.1 — Results for flux los o ignitio testing

R esult of ignition trials in icat e that the r lative s an ard deviation ( RSD ) for 1 ignitions of flux

b t ch 1B is highe than that o se ved for 1A As the ignition los is d e t o the de omp sition of the

ad ed o idizing ag ent, the pe c ntag e RSD an a e ag e ignition los can be used as measur of ad itive

homog eneity an q antity

A nne x B

Make sur that moulds an cas ing dishes have be n as es ed t o ensur flatnes (4.4) Pr p r betwe n

8 an 1 beads of ignit ed spe trosco ical y pur F e

2O

3( 99, 99 % purity) using the normal diution

r cipe (flux:sample:o idant ratio) an fusion t echniq e(fusion times an t empe atur s)

Using an exis ing iron or analytical method as a base, c eat e a new analytical proc d r w he e only

iron is measur d an matrix cor e t ed (al major an trac element ar r mo ed from the calbration)

Measur ment times should be sele t ed so that t otal ac umulat ed cou t of 2 × 1

7

ar o tained for a

pr cision bead

Beads should be measur d twic in suc es ion, r -measuring the f irs bead at the en of each sample

ru Data should be proc s ed ac ording t o the sample calculation spr adshe t in An ex G an Annex H

(a co y of the calculation pro ram is a ai a le from CSIRO, A delaide)

In al cases, R ead e ror should be eq al t o the s atis ical cou ting e ror of the analytical method

(≈ 0,0 % or 2 × 1

7

cou t ) Disc e ror should be les than 1,5 of the theor tical cou ting e ror, but

pr fe a ly eq al t o the measur ment e ror (w hen r p rt ed at 1 0 % F e

2O

3conc ntrations)

A nnex C

c onc entration

The proc d r t o det ermine r lationship betwe n height an conc ntration is as fol ow s:

a) Using bead measur ment a p ratus (An ex E) an an u modified X-ra cup, set depth ga ug e

r fe enc p int t o z ro using datum bead

b) Using height adjus a le cup (as desc ibed in Annex D), set bead height 0,5 mm lowe than the

datum bead height (plus 0,5 mm on depth ga ug e)

c) Using a flat 1 0 % Fe

2O

3fusion bead, measur the iron conc ntration ( in q adruplcat e) uti izing a

cou ting s atis ical e ror of 0,0 5

NOT A flat usion b ad can b prepared from a nominal fusion b ad that has b en dry lap ed using

12 0 grit diamon polishing paper an machinist’ s surface plate

d) R epeat measur ment proc d r (a o e) for datum height at (− ,2 , 0,00, + ,2 an + ,5 ) mm

e) Plot a e ag e F e

2O

3conc ntration a ains dife enc from datum point (pr paration of height/

conc ntration r lationship)