Analysis and geochemistry of chert.

Bejnar, Tor John Stark.

Two aspects of geochemistry were studied: rock analysis tenchique by f 1 u01-escence spect.rometr·y I and geocherni stry of New Zealand cherts and x-t-aY the ir met amorphic equivalants. A. Analysis. The analyses were done using a manually.operated Phillips PW1540 spectrometer and PW1130 generator. The effects the various parts of the system had on analysis were studied. Appropd.ate adjustment, selection, and/or setting of the generator, x-ray tube, sample, sample holder, sample chamber , collimator, analysing crystal, counter, goniometer, and electronics \Jere selected for each of 19 elements (major oxides: Si02 1 Al203 1 Fe203, HgO I cao, K20, Ti02, P205 and Hn02; trace elements: As, Ba, Cr, Cu, Ga, Ni, Rb, Sr, Y, and Zn) . The amount of 1 eeway avai lab! e on many parts z.hov1ed some to require fine tuning, while others needed only quick adjustment to obtain optimum operating conditions. The variation2! when using the pulse height selection radically changed after some modifications were done on the preamplifier by the department technicians. Analyses of pressed powder briquette samples were compared with synthetic standards made from pure chemicalz.l intern:ttional reference material and other Previously anc.lysed rocks. Nearly all briquettes were prepared using a mowiol solution binder. The quality of bt-ibuettes pressed against srlass was superior to those Pressed against poli2.hed steel. No consistent variation of analyses clue to texture quality I however, t-7as recognised from a study of several duplicate briquettes. Two sets of f~e ssed powder briauettes were analysed. One set was of Whole rock cru2.hed pm¥ders. The other was of hydrofluoric acid treated xiii. t hese va lues were modifi ed t o r efl ect t he composition of the whole po~.Jd e r s ; -- rock. Although the t wo sets of data obtained shm-led a good statistical correl ation, many i ndividual el ement an3.lyses from the two set~. of data ~.; e re di fferent by more than a factor of 2. Such discrepencies may be clue to differ ences in grain size <the acid treated samples being much finer), mica flake ori entati on , differences in mass absorbtion eff ects (especially for Si02 and ]1,1 203) and/or inhom·ogeneity of acid h -eated sampl es (differ ent coloured fact ions of the residue migrated to different parts of the teflon basins; collection of all the residue was impossible, and mi xing what was collected may not have been adequate). Utilization of concentration techniques is seen t o be productive for elements whose abundances are near or below their detection limits (P205, As and Ga). A new method of background determination was utilized for most tr·ace · elements using one background determination. Several synthetic .~.tandards devoid of a trace element were used to determine a curve which \-lould predict the background count rate at the peal~ position using a background position count rate. For most trace element analyses, all background count rates were modified before subtracting background from the peak count rate. For most elements, this provided better analysis curves than st:cai.ght peal< minus background analyEds curves. The precision of analysis on eight bdquettes of one powder sample was better than 5 percent (relative) for most major element mddes, and better than 50 percent for most trace elements. I el ement analyses were employed,) <More stringent controls on, major %. I ., llccuracy of r esults was studied by several means. Mn02 concentration va lues for this project were higher than some independant laboratory ~hecl~s, f. ~ and s imilar to others . The concentrat ions , as det &rminetl, are well above typica l cher t anal ys es , as det erm ined by a compari son with va lues in the ~- ""'· xiv. literature. Cu, Ni and Y concentration value:::; determined for this project -·"'vL:- t-1ere usua lly lm.;er ti)an independant l aboratory checks. Cr concentrations · determined for this project were much higher than tho.'::'.e for chert and pelagic sediments and rocks reported by most other authors. Independ~nt checks usually indicated that the Cr concentration2. determined for this project were high, but sti 11 \.Jell above typical chert concentrations. Cu analyses seem to have been adversely affected by Mo l< a lpha Compton Scatter from the Mo x-ray tube. B. Geochernistry . Sixty samples of chert and quartz schist (or 'metachert') were collected £1.-om seventeen areas acro.ss the South Island and southernmost North 1.::\land. Some of the r~e samples were obtained from the GF:ology Department collection or from other field workers. Another 30 quartz schist analyses from one outcrup were a.lso studied (Dowz.e , et al., in prep.) l1o.st samples \-1ere from within the Torlesse or Haast Schist assemblage2.. Field information r egarding textural gr.:tcle , 'country rock' association, and colour of each specimen was obtained. Rigorous sample selection criteria were ~ mployed to avoid weathering, l eaching, ve ining and other contaminating phc:-nomena. A number of chert samples whict! were obvious ly, or possibly, mi xed with adjacent rock material, and several non·-chert samples, ~.Jere Emalysed for comparison purposes. , Chert ,<Y:Jmp l es are a mi xture of .dlica and other material. Although the si lica is almost always of biogenic origin, the evidence for this is · frequently l os t due to l ocalized mobilization and complex recrystallization of I lhe s ilica. The portion of biogenic s ilica in a sample necessary for the sample to appear cbert-·1 H'e E:'.eerru:; to be as 1 01-.' a:::. ten percent. (This means t en percent biogenic silica added to a c~ay whi ch would typically have about EiO per cent non-biogenic s .ll ica. ) The ari thrnE:'ti c aver·age of biogenic s i 1 i ca determined by geochemi ca l wethods was . 54 percent. The moda l averag~ was 84 PE>rc e::nt . The cday faction \~ithin chect most closely re.::~E~mbles typical deep ~ ea clay. XV, The elements in nearly all of the samples of chel..-t and quart: schist foll0\·1 n- ~-~ina l e r_~eochemica l pattern . :;J The ratios between all 'clay' faction l ements except Cu and Cr are fairly cons istant at ctll abundance l evels . e usuallY strong corre lation indicates that these elements are depos ited independantly of the deposit i on of si li ca. Cr has a distinct negative corre l ation with the clay facti on el ements and a positive correlation with Si02. Cu analy;:.es, as deter rn ined , 2.how no strtmg correlation with other elements . This pattern is fairly distinct from non-chert samples. The chert samples are readily separat ed from terri qeneous samples (siltstones and The quartzofc ldspathic schists) using Mn02-Al203 and Mn02-Ti02 variation diagrams. Some trace element abundances r e l ate to crystal chemical behavioral similarities: Ga to Al203, Rb to K20 and Sr to CaO. The ' clay' f act ion e l ements within chert come from several sources : deep sea clay, continental material, volcanic ash, ·and el ements mobilized in hydrotherma l emanations . Within limits, ttae relative contribution from these sources vary. Thr;:y 11i 11 vary lar~r el y due to the environment of cleposi ti on . Sampl es depos i t ed close to co~t1ne ntal 1nargins will have a greater ~ortion of terrigeneous detritus t han sa~ples deposited in isol ated bas ins . Samples deposited cl ose to oceanic spreading centres would be relatively enriched with hydrogeneous metals. A number of geochernicaJ pctrameters. \-;er·e u2.ecl to :::.eparate the various ! field loc.:d i ties from each CJther with r e.spi:·ct to the contributi on of deep ~\ea clay ard ·:·-.··t p -.r·· · r: ; .•. R' ,,o '>··r·~· · ·'- (t:'f'I!J j) ) }' ]· ·t· " - ·- 1 Th . 1 1 d F Al 1 b:t::::. · c, .• l ,_-' L.. 1~'·- L·t:·1.li.J,-:-. .lt., r:. l\ .• - .1 · ,e rnct .E.l. J. ct • ese 1 ne uc e · eralio and ~\l op:=! , Hn··Ti ratio ~1 n•l. slope, f\::/Ti r at i o, Jl.li <Al +Fe+Hn ) ratio, <Fe/Ti ) /( Al/ C Al+Fc~Hn )) ratio , Fe/(Mg~Kl r a tio, Mn/CMg+K) r at io and biogenic ~rt· - · lcn C ;; l 02- '. ?. l ?CJ3+Fe 203 ·1 !+JCl+Cc.~i+ i<20 +T .i. Cli~+P2o~;+l·ln02) :~:2 . 011. 00). .. ·. • I, xvi. The aver age· for each parameter of the sampl es f r om field at-eas \.Jere tabulated. The relative rank for each field area tended to be consistent for all par ameters. Chert 2.ampl es associated with gt-eenschists tended to indicate a closer depositi onal proximity to EPRD-like .sediments than other samples. Individual samples which did not conform to the inter-element b.-ends also indicat ed a closer EPRD-like associati on than the more normal samples. The stratigcaphically lower sampl e~. from Jl.rrow Junction (Dowse, et al., in pr ep.) indicate a much stronger affinity to EPRD-like sediments than the overlying quartz schist samples. Several possible theories for this ate presented. The Fe/Ti versus Al/CAl+Fe+Mn) variation diagram indicates that the clay fact ion in chert samples contain 0 - 25 percent EPRD-like material and 75 - lOO percent typical deep sea clay. Clear geochemical indicators for othel.- 2.ource sedirnents were not ~:. tudied. The elements most strongly associated with terrigeneous influenced chert deposits are Al203, K20, Ti02, Ga and Rb. The elements most strongly associ ated with hydrothermal eminations are Mn, Fe, Ni and Zn. Ne compelling t r ends were noted due to increa.sed textural or metamorphic grade . Geochemical variation correlated with colour of specimens, but the s i gni.ficance of this was not determined. Red chert samples tend to be dch in Fe203 I.Jhile black chert samples tend to be en:dchecl (relative to Fe203) in Mno2. Host chert from the "Nm-1 Zealand geosyncline" was formed in a t er-r i 9eneous 2.ecl iment. starved environment. by the deposition of biogenic Si lie.:-,. The other maj or component in the . chert is deep sea clay rich in hyd ro0eneous metal s . Rat es of s ili ca , cl av. an.d hydrogeneous metal depos ition

137, 90 leaves : ill. (some col., some folded) ; 30 cm.

1983Bejnar