pajaformation(colombia)bylossonignitionand chns analysis

8
島根大学地球資源環境学研究報告 21,9~16ページ(2002年12月) Geoscience Rept.Shimane Univ.,21,p.9~16(2002) Organiccarbonand carbonatecontentsofb且ackshales PajaFormation(Colombia)bylossonignitionand CHNS N.O.Campos*,B.P.Roser*and Y.Sampei* Abstract Organic matter(and hence organic carbon)and calcium carbonate contents of soils an simplelossonignition(LOI)deteminations,attemperaturesranging丘om375to Organic matter(OM)andcarbonate contents of24carbomceous black shale samples methods to compare these two techniques.Two LOI tests were c飢Tied out to dete㎜ine t 550℃and1000℃to ensure complete combustion ofthe OM and carbonate compone exposure at550℃were required to ignite the OM,and3ho皿s further ignition at lOOO carbonate.A complete set of sample were then ignited under these conditions,and the r CHNS analysis.Although the results correlated well,LOI550values(OM)were consistent COrreCtiOn Of the LOI data uSing a generiC OM tO C。,、COnverSiOn faCtOr COmmOnly appli improve the results.A third experimental mn with LOI detemlinations at375℃,5 correlation between LOI375and C。,、(CHNS),but relatively constant loss of weight irrespective ofC。,、content.This suggests most ofthe weight loss below375℃was due that between375℃and550℃was struct皿al water lost ffom clays.Nevertheless,ign giVe reaSOnable eStimateS Of C。,、Ontent,prOVided CalibratiOn data fOr indiVidUal Samp independent technique such as CHNS analysis。The results support proposals that use factors is inappropdate.Test ofthe viability ofcarbonate content dete㎜ination wa nat皿e ofmost ofthe samples.At low levels,neither LOIlooo nor CHNS seem to give good re continued loss of stmctural water and other volatile phases,coupled with weight gains as Fe2+andMn2+. Key words:Black shales,LOI,CHNS,organic content,carbonate content. lntroαuction Loss on ignition method(LOI)is often used as a simple means of determining approximate organic carbon(C。,g), organic matter(OM),and carbonate contents of soils, sediments and sedimentary rocks,particularly when they are Ca1CareOUS and Clay-pOOr (Ball, 1964; Dean, 1974; Sutherland,1998).The method includes a first heating at 375to500-550℃to ignite the OM,followed by a second ignition at950-1000℃to release carbon dioxide ffom carbonate minerals.Contents・are derived丘om gravimetric data.However,many caveats s皿round the method.A number of authors(Sutherland,1998;Heiriαα1.,2001, among others)point out that losses of other compounds may also occ皿 (e.9.volatile salts,stmctural water, inorganic carbon, pyrite), depending on the ignition temperat皿e.Such losses can thus influence the accuracy of the C。,g and carbonate detenninations.Heiri8∫α1.(2001) further noted that LOI is also influenced by heating time and sample size. Chromatographic CHNS elemental analyses can also be USed tO aCCUrately meaSUre Ct。t,C。,g and CarbOna and to detemine the composition of the o (Minoruαα1.,1999;Pimmel&Claypoo1,200 Matsumoto,2001).In this method samples a in a reactor at1000℃in a temporarily enr atmosphere.Temperature reaches1800 combustion.The reaction products are pas glass column packed with a tungsten triox reducer to form CO2,N2,and SO2.These produ analyzed chromatographically,by calibra organic standard. This paper examines the OM,C。,g and carbona of a suite of black shales from the P勾a F Colombia,as detemined by both the LO methods.The primary p皿pose is to compare obtained by the two techniques,and detem technique is apPropriate for rocks of this type. Geologic Setting *DepartmentofGeoscience,ShimaneUniversity,Matsue690-8504 The Eastem Cordillera of Colombia is trending fold belt composed of Precambria basement covered by a thick sequence of 9

Upload: others

Post on 03-Oct-2021

3 views

Category:

Documents


0 download

TRANSCRIPT

島根大学地球資源環境学研究報告 21,9~16ページ(2002年12月)

Geoscience Rept.Shimane Univ.,21,p.9~16(2002)

囮      Organiccarbonand carbonatecontentsofb且ackshalesfromthelower Cretaceous

  PajaFormation(Colombia)bylossonignitionand CHNS analysis:comparisonofmethods

N.O.Campos*,B.P.Roser*and Y.Sampei*

                                     Abstract

 Organic matter(and hence organic carbon)and calcium carbonate contents of soils and sediment can be estimated by

simplelossonignition(LOI)deteminations,attemperaturesranging丘om375to550℃and850-1000℃,respectively.

Organic matter(OM)andcarbonate contents of24carbomceous black shale samples were dete㎜inedby LOI and CHNS

methods to compare these two techniques.Two LOI tests were c飢Tied out to dete㎜ine the ignition times required at

550℃and1000℃to ensure complete combustion ofthe OM and carbonate components.Results showed4to6ho皿s of

exposure at550℃were required to ignite the OM,and3ho皿s further ignition at lOOO℃ensured decomposition of the

carbonate.A complete set of sample were then ignited under these conditions,and the results compared with those ffom

CHNS analysis.Although the results correlated well,LOI550values(OM)were consistently greater than C。,,by CHNS.

COrreCtiOn Of the LOI data uSing a generiC OM tO C。,、COnverSiOn faCtOr COmmOnly applied in the literatUre did nOt

improve the results.A third experimental mn with LOI detemlinations at375℃,550℃,and1000℃showed a strong

correlation between LOI375and C。,、(CHNS),but relatively constant loss of weight(2-4wt%)between375℃and550℃

irrespective ofC。,、content.This suggests most ofthe weight loss below375℃was due to combustion of OM,whereas

that between375℃and550℃was struct皿al water lost ffom clays.Nevertheless,ignition at either375℃or550℃can

giVe reaSOnable eStimateS Of C。,、Ontent,prOVided CalibratiOn data fOr indiVidUal Sample SUiteS iS available frOm an

independent technique such as CHNS analysis。The results support proposals that use of generic OM to C。,,onversion

factors is inappropdate.Test ofthe viability ofcarbonate content dete㎜ination was hampered by the non-calcareous

nat皿e ofmost ofthe samples.At low levels,neither LOIlooo nor CHNS seem to give good results.This is probably due to

continued loss of stmctural water and other volatile phases,coupled with weight gains through oxidation of species such

as Fe2+andMn2+.

Key words:Black shales,LOI,CHNS,organic content,carbonate content.

lntroαuction

  Loss on ignition method(LOI)is often used as a simple

means of determining approximate organic carbon(C。,g),

organic matter(OM),and carbonate contents of soils,

sediments and sedimentary rocks,particularly when they are

Ca1CareOUS and Clay-pOOr (Ball, 1964; Dean, 1974;

Sutherland,1998).The method includes a first heating at

375to500-550℃to ignite the OM,followed by a second

ignition at950-1000℃to release carbon dioxide ffom

carbonate minerals.Contents・are derived丘om gravimetric

data.However,many caveats s皿round the method.A

number of authors(Sutherland,1998;Heiriαα1.,2001,

among others)point out that losses of other compounds

may also occ皿 (e.9.volatile salts,stmctural water,

inorganic carbon, pyrite), depending  on  the ignition

temperat皿e.Such losses can thus influence the accuracy of

the C。,g and carbonate detenninations.Heiri8∫α1.(2001)

further noted that LOI is also influenced by heating time

and sample size.

  Chromatographic CHNS elemental analyses can also be

USed tO aCCUrately meaSUre Ct。t,C。,g and CarbOnate COntentS,

and to detemine the composition of the organic matter

(Minoruαα1.,1999;Pimmel&Claypoo1,20011Sampei&

Matsumoto,2001).In this method samples are combusted

in a reactor at1000℃in a temporarily enriched oxygen

atmosphere.Temperature reaches1800℃d皿ing flash

combustion.The reaction products are passed through a

glass column packed with a tungsten trioxide and copper

reducer to form CO2,N2,and SO2.These products are then

analyzed chromatographically,by calibration against an

organic standard.

  This paper examines the OM,C。,g and carbonate contents

of a suite of black shales from the P勾a Formation of

Colombia,as detemined by both the LOI and CHNS

methods.The primary p皿pose is to compare the results

obtained by the two techniques,and detemine ifthe LOI

technique is apPropriate for rocks of this type.

Geologic Setting

*DepartmentofGeoscience,ShimaneUniversity,Matsue690-8504

  The Eastem Cordillera of Colombia is a NNE-SSW

trending fold belt composed of Precambrian to Paleozoic

basement covered by a thick sequence of Mesozoic to

9

10

   Organic carbon and carbonate contents ofblack shales from the lower Cretaceous

P句aFomation(Colombia)bylossonignitionandCHNSanalysis:comparisonofmethods

740W 720W

60N

40N

Caribbean Plate

1卜

隙婁

South American

Plate

+∵+議+

    ヨ ナ  チ     ナ    ぐ

∵亭+ + o  +

ナ  せ ナ

轡・+♂∴

++1試+

諺.御

十  十  十  十  十

離離雛鑛

               ナ  ナ  ナ  ++      薗+++     N

l12魏、1撒.番

         鯵霧多 1〆

      襲爵饗ノ

謬        む    瓢3                  し        =          □Te由ary メAnticlines    琶

    十十          國8器e・us〆Thrusts

100Km @囲ヒ器e・US≒syncline、          目Triassic-Jurassic              Basement /Extensi・na          回              &Plutons       Faults

嘉、}灘:ン.・…,鰯…灘

繕1、難罪

50

■StudyArea

Fig.1.Geology of the northem segment of the Eastem Cordillera

 and location of the study area(modified from Cooperαα1.,

  1995).

Tertiary sedimentary rocks.The Eastem Cordillera formed

during Miocene-Pliocene times by inversion of two Late

J皿assic-Early Cretaceous basins,the Tablazo-Magdalena

basin to the west and the Cocuy basin to the east.These

former basins are separated by the Santander High,which is

characterized by the occuITence of faults running Parallel to

the general trend of the cordillera itself (Colletta 8∫ α1.,

1990,Dengo&Covey,1993,Cooperαα1.,1995).

  The study area is located in the Velez-Arcabuco

anticlinorium in the westem flank of the East Cordillera

(Fig.1).Micritic limestones of the Valanginian-10wer

Hauterivian Rosablanca Formation are conformablyoverlain by siliciclastic rocks ofthe Hauterivian-Aptian P司a

Fonnation  (Royero  &  Clav麺。, 2000; and  references

therein).The present study focuses on surface samples of

the Paja Fo㎜ation.This sequence consists predominantly

of organic-rich black shales,and minor intercalated silty,

very fine-grained sandstones,massive organic-poor beige

shales, and occasional massive organic-poor green

mudstones.The black shales are generally massive,but

occasionally show very thin lamination and foliation

parallel tO the OVerall Orienta.tiOn Of the Strata.TheSe

laminations usually consist of fine-grained quartz or other

tenrigenous detritus.

ground using a ROCKLABS(Model RC)ring millequipped with a100g capacity tungsten carbide head.The

method used followed that described in Roserαα1.(1998).

Chips showing any sign of oxidation,alteration or

weathering were discarded.Maximum mill times were

30seconds.Subsamples ofthe powders were stored in glass

vials and held at110。C for at least24hours prior to LOI

determinations.

  Powders for the CHNS analysis were prepared by

manually crushing10-20g subsamples of the original hand

specimens to<1mm,followed by grinding in an automatic

agate pestle and moltar for10minutes.The powderedsamples were stored in glass vials and oven-dried at70℃

for24hours prior to analysis.

Methoαology

Cr配sh∫ng

  Samples for LOI deteminations were chipped,washed in

distilled water,and dried at110℃for24h.They were then

Lossonなni∫ionD8∫εrn¢inα∫ions‘LO1フ

  The parameters used to determine organic matter and

carbonate contents by the loss on ignition method vary

widely between studies.For determination of OM,ignition

temperatures used vary between375and550℃,with

ignition times ranging between one and seventeen hours

(Ba11,1964;Dean,1974;Sutherland,1998;Heiriαα1.,

2001; and references therein). For carbonate contents,

temperat皿es used range between850and1000℃,and

times between30minutes and two hours.The method

suggested by Heiriαα1.(2001)for assessing organic matter

and carbonate contents re(luires four hours of exposure at

550℃to ignite the organic matter,and two ho皿s further

ignition at950℃to complete combustion of the carbonate

component.We used the550℃temperat皿e for assessing

organic matter following Heiriαα1.(2001),but carTied out

thecarbonateignitionat1000℃ratherthan950℃.

  Wefirstmadetwo tests to determinereliable combustion

times for each component.In the first test,five samples

were selected at random and processed individually.After

ovemight oven drying at1100C the samples were ignited at

550℃to determine OM content.Weights of the crucibles

were taken every hour until the ninth hour,with the

crucibles retumed to the fumace after each weighing.Each

ho皿一long ignition was timed from when the fumace

regained550℃.The samples werethen ignitedfor afurther

12hours at550℃,and the weights again meas皿ed.The

fumace temperat皿e was then raised to1000℃.After

constant temperature was reached,the samples were ignited

for another two hours,cooled in a desiccator,and

reweighed.The crucibles were then retumed to the furnace

for another three ho皿s.Weights were remeas皿ed every

hour,asinthe550℃experiments.

  In the second test,five duplicates of sample PES-31

varying in net weight from2.7to3.4g were ignited using

the same conditions as above,to test the influence of sample

Slze.

  Following the above experiments,a complete set of

24samples was ignited using optimal ignition times as

determined from the trial samples.Percentage loss on

N.0.Campos,B.P.Roser and Y.Sampei 11

ignitionat550and1000℃aregivenbythefomlulae:

         (CSH・一CS55・)一(C-Ci)

%LO155。=         *100                 Sllo

         (CS55・一CS1。。。)一(C-Ci)

%LOll〔)。。=         *100                  SIlo

  where CSニweight of crucible+sample;Cニcrucible

weight before ignition;Ci=crucible weight after ignition;

Sニsample weight(after Heiri8∫α1.,2001).LOI550thus

corresponds approximately to the amount of OM in the

sample,whereas the corresponding value at1000℃ is

related to the carbonate content.

  Following the above,the CHNS analysis,and preliminary

interpretation,we ca皿ied out a further experiment(test

three)on a selected subset of samples.In this test,the

samples were first ignited at375℃for16hours(after Ba11,

1964;Dean,1974),followed by six hours at550℃,and

three hours at1000℃.Weight losses were dete㎜ined at

each step.

C研VSαnαlysis

  Total CarbOn(Ct。t),OrganiC CarbOn(C。,g),H,N,and S

contents ofthe suite of24samples were determined using a

CHNS automatic elemental analyzer(FISSON,EA llO8).

Total carbon contents were first

powdered samples apProximately

Accurately weighed samples were

capsules,which were crimp-sealed

dete㎝ined, using

10mg in weight.

placed in tin foil

prior to  analysis.

Organic carbon contents were detemined in a separate run.

For these analyses lO mg ofrock powder were placed in Ag

foil containers,and the samples treated by addition of l N

HCl to dissolve the carbonate fraction.If samples reacted

strongly due to high carbonate content,HCl was added

repeatedly until reaction ceased.The Ag capsules were then

dried on a hot plate for45 minutes at110℃,and

subsequently sealed.They were then placed within tin

capsules,which were sealed in tum.

  FOr bOth the Ct。t and C。,g analySeS,prOdUCtS Of the

combustion were calibrated against BBOT standard[2.5-Bis

一(5-tert.butyl-benzoxazol-2-YI-thiopen)].The Eager 200

software was used to run the equipment,and to store and

manipulate the data.TOtal CarbOn(Ct。t)and OrganiC CarbOn

(C。,、)COntentS Were Obtained direCtly frOm the eqUipment,

whereas carbonate carbon(Ci。。,g)was calculated ffom the

difference between C、。t and C。,g.The error inherent in this

type of analysis is about±3%relative.

Results

Lossonlgni∫∫on

  Results for the first test show that weights of the two

samples with lowest LOI(PES-9,PES-14)stabilized after

only three hours ho皿of combustion at550℃(Fig.2).No

significant additional weight loss was observed after

12hours further exposure at that temperat皿e.In contrast,

for samples PES-1,PES-25and PES-35weights did not

stabilize until six hours had elapsed(Fig.2).In the second

test,weight losses converge after six hours ofcombustion at

550℃,and are identical after nine hours(Fig.3).No

significant weight loss was caused by a further12h

expos皿e at550℃.In both tests,constant weight was

achieved in vi血ally all samples after six ho皿s at550℃.

However,weight losses vary considerably with less than six

hours combustion.This variability may be attributed to

differences in OM content(test l)and sample weight(test

25

 0          5          nU          5

            コ          ヨ

(ぷ筆)ω8=8一ΦタΦ>蚕コ∈δ

0

550。C

一十PES-31+PES25十PES-1十PES-14+PES-9

     051015202530                1gn而on time(hours)

Fig.2.Cumulative LOI(wt%)vs time for five black shale samples

ignitedat550℃(h・urslt・21),and1000℃(h・urs22t・27).

25

  0          5          0

  2          1          1

(ぷ琶ω8==9睾Φ>穂3∈δ

5

550℃

hitial weights

+2.73g+2.80g+2.939+3.28g+3.43g PES-31

     0      5      10     15     20     25     30

                1gnition time(hours)

Fig.3.Cumulative LOI vs time for five duplicates of a single shale

 sample(PES-31),with differing initial weights.Ignition at550℃

 (hours l to21),and lOOO。C(hours22to27)as in Fig.2.

12

   Organic carbon and carbonate contents ofblack shales ffom the lower Cretaceous

P司aFomation(Colombia)bylossonignitionandCHNSanalysis:comparisonofmethods

Table1.Loss on ignition and CHNS analyses ofblack shales from the P司a Fo㎜ation,Colombia。

SaNr

LossOnlgnition(wt%)

Tota1  550gC  1000gC Ctot

CHNS Elemental Analyser(wt%)

 CorgCinorgH N S

PES-1

PES-2

PES-3

PES-4PES-5

PES-8

PES-9

PES-10PES-11

PES-12PES-14PES-15PES-16

PES-18PES-20PES-21

PES-23

PES-24PES-25

PES-27PES-28

PES-29PES-30PES-31

816739717058806353109434

079879058580087104299267

902913564585544486943060

112 41           1  1 2

405022607485546598347997

249404624945349741056889

101913464484443376843956

14ーハ∠  41」1                  」-      』-

312617120673369965862557

830475434535737352133377

701000000000000000100003

    3

560189943964854941877404

795583100457362780431629

156417121040000002300404

』i  」一  ハ∠                   」1      詞i

050196369674583963153694

495563198327241769120909

156437110040000001300304

イー  

イー  イー                        イー        -

500093684390371088734710

300010001120111010311610

ααααaααααααααααααααααααα

016897693217638404969311

458536455766756406875008

000000000000000010000110

646376191805402028306794

237224222121020131630415

000000000000000000000000

910054000000000000400000

114040000000000000300000

ααααtααααααααααααααααααα

2).As could be expected,the larger the sample and the

higher the organic content,the more time needed for

combustion.

  In both tests,increasing the temperature to lOOO℃

caused additional weight loss(Figs.2&3).Most of the

additional loss occurred within two ho皿s at1000℃,

implying that carbonate ignition was largely complete.

Weight loss afterthree ho皿s f皿ther ignition was minima1,

and a pla.tea.U WaS reaChed.

  The above tests suggest that six hours ignition at550℃

was sufficient to bum off the OM,and three hours at

1000℃ completed decomposition of the carbomte

component.These times were adopted for the LOI

dete㎜inations of the fhll suite of24samples;results show

thatLOI550was>10wt%in seven samples,with a maximum

value of21.95%(Table1).Among those with LOI550<10%,

the majority yielded values in the range4-6%.In contrast,

LOIlooo values were generally much lower,with only

samples PES-1,一5and-31having significant values(7.83,

30.71,and3.77wt%,respectively).These three samples

also have the greatest CaO contents (5.89, 35.83 and

1.61wt%;Campos,unpubl.data),suggesting the presence

of some calcium carbonate.All other samples have low

LOIlooo(0.25-1.01wt%)and CaO(0.03-1.49wt%;Campos,

unpubl. data), suggesting their carbonate contents are

minima1.

  Results of the third test(sequential ignition at375,550

and1000℃)showedthatinhalfthe samples the bulkofthe

cumulative weight loss at550℃actually occ皿red after

ignition at375℃(Fig.4).In three samples the rate of loss

was almost constant through to550℃,and in one sample

(PES-20),the greatest loss in weight occurred during

ignition at550℃.The only significant weight loss at

1000℃was recordedinthe singlehighly calcareous sample

(PES-5),as expected.

40

   ハU          nU          O

   3         

∩∠         -

(ぷ峯)ω8=昼睾Φ>量コ∈δ

0

375。C 550℃

ゆ一一PES-3一一汕鼈黷oES-8

十PES-20十PES-25十PES-29一一Z一一PES-1

一一「・一一PES-5

一一m二}一一PES-31

a’

1000℃

  1’  .’口

“.・禔轣h一一

      一._’一〇一’”’一が 一””

0 5    10      15      20

1gnition time(hours)

25

Fig.4.Cumulative LOI vs time for eight black shale samples of

 equal weight(3.00±0.02g)ignited at375℃(hours l to16),

 550℃(hours l6-22)and lOOO℃(hours22to25).Solid symbols

 -carbonate-poor s&mples;open circles-carbonate-rich.

N.O.Campos,B.P.Roser and Y.Sampei 13

25

0       5       0       5

ハノ        で        り

(」

ウω≧O⊆OのZ工Q)ぷ峯◎、。Q

0● ●

PES-5

 0

PES-1

PES-31

0●

1:1

y=一3,72+0、89x Rニ0.98

     0510152025           0rganic matter(LOI550wt%)

Fig.5.Comparison of OM content based on LOI550with C。,g

 content determined by CHNS method.Open symbols(carbonate-

 rich samples)are excluded from the regression.

25

0       5       0       5

         コ        て

ぷ一タ(のZ工Q芯、。Q℃Φ」コωOΦΣ

0

y=一3.18+1.54x R=0.98

○    ●

1:1

     0      5      10      15      20      25

        Estimated Corg(LO1550/1.724)wt%

Fig.6.Comparison of estimated C。rg from LOI550with measured

 C。rg(CHNS amlysis),using the generic OM to C。rg conversion

 factor(1.724)commonly applied in the literature(Sutherland,

  1998).

CH/VSα加n4αnc8s

  CHNS abundances in the suite of24samples show

considerablevariation.Organiccarbon(C。,、)contentsrange

from O.03to16.5wt%,and the greatest C。,、values are all

associated with high LOI550values.Although Ci,。,、varies

from O.00to8.19wt%(Table1),in only one sample(PES-

5;8.19wt%)is it significant;all others are<0.67wt%.

Hydrogen(0.44-LO3wt%)and nitrogen(0.02-0.76wt%)

were detected in all samples(Table1).Nitrogen and C。,、

contents are moderately correlated (R=0.77); this

correlation improves when three sample with both high C。,,

and LOIlooo(PES-1,一5,and31)are omitted from the

regression (Rニ0.89).However,no such relation exists

between hydrogen and C。,、(R=:0.26).Sulphur was deteCted

in only six samples,and with the exception of PES-5

(1.45wt%),even in these samples abundances are low(0.04

-0.40wt%).This suggests that sulphide minerals are not

significant components ofthis suite.

Discussion

  As expected from the LOI tests,ignition time and sample

size seem to be the main factors influencing the results,

especially the determinations at550℃ (Figs.2&3).

Samples must be ignited at550℃for sufficient time to

ensure total ignition of the OM component,and hence

preclude any effects on the carbonate determinations at

lOOO℃(Heiriαα1.,2001).This is particularly true for

samples with high organic content.Heiriαα1.(2001)

suggested thatfourhours exposure at550℃was sufficient

to complete ignition of most OM,and that two hours of

ignitionat950℃weresufficienttocompletedecomposition

of carbonate.From our results,we found that six ho皿s

exposure at550℃are required to ens皿e complete weight

loss due to combustion of the OM in samples weighing

<3.3g,and three hours f皿ther exposure at1000℃

guaranteed total loss of the carbonate fヤaction.

  COmpariSOn Of LOI550and C。,、(CHNS analySiS)ShOwS

that a strong linear correlation exists between OM content

(LOI550)and measured C。,g when three SampleS(PES-1,一5

and-31)are excluded from the regression(Fig.5).The

compositions of these three samples differ ffom the others,

as noted above.Although the observed correlation is strong,

the intercept is non-zero.This situation is not unusual,as

evident f士om the summary of regression models between

C。,g(dry CombuStiOn)and LOI given by Sutherland(1998);

slopes may also vary significantly between suites.However,

given the strength of the co∬elation,LOI550values in this

Suite Seem tO be reaSOnable prediCtOrS Of C。,g uSing the

equation derived.

  COnVentiOnally,C。,,iS eStimated ffOm10SS On ignitiOn

data by dividing LOI550by a conversion factor of1.724,

based on the assumptions that LOI550is related to loss of

OM,and thatC。,g compriSeS58%OfOM(SUtherland,1998;

and references therein).Applying this factor to the P勾a

samples produces overestimates at low values(<5%C。,g by

CHNS)and overestimates at higher values(Fig.6).

Sutherland(1998)highlighted the danger of assuming a

fixed conversion ratio of L724 by summarizing an

extensive body of literature data,and noted that factors

actually ranged from l to as high as l6,depending on the

nature of the OM In his study of Hawaiian fluvial

sediments,he found the factors ranges from4.2to33.9.

Conversion factors in given suites are approximated by the

reciprocal ofthe slope of regressions between OM(LOI550)

and meas皿ed C。,g(Sutherland,1998),yielding a conversion

14

   Organic carbon and carbonate contents ofblack shales from the lowerCretaceous

P司aFormation(Colombia)by loss on ignition and CHNS analysis:comparison ofmethods

25

20

5       0

婦『            』-,

器一夕のト。り一〇」

5

0

a.

111

y=1、57+1」25x Rニ0、99

5     10     15

   Co凹wt%(CHNS)

20

25

20

  5       0

  1             1

訳一夕Reり,8』O」

5

0

b

0 25

25

1:1

25

0

20

 5   0

 1       」-

(ま峯)㌍.粁\霞。っ6」

5

             て 

   Corgwt%(CHNS)

20 25

C.

1:1

y二1.41+1.01x R二〇、99

20

  5        0

  イー              -

O創工H工雷OΦ,。昭一〇」

5

d

● ●

y=0.40+0、94x  Rニ0.99

O PES31

1:1

         0        5        10        15       20       25         0        5        10        15       20        25

                   Corgwt%(CHNS)                    Co四+N+40%H+S

Fig.7.Plots of(&)measured C。,g(CHNS)vs weight loss after16h ignition at375℃(LOI375)in the samples from Fig.4,excluding the

 highly calcareous sample PES-5;(b)Weight loss between375。C and550℃in the same samples compared to measured C。,g;(c)

 estimated C。,g(LOI375/L12)vs measured C。,g(CHNS);(d)Modelled OM content from the LOI550data in the entire sample suite,

 assuming60%Of measured hydrOgen iS lost as water,COmpared to measured C。rg+N+H,esidual+S(CHNS).PES-5and PES-31were

 excluded from the regression.

        0                                                0

fa.CtOr Of l.12 in thiS CaSe (Fig。 5). ThiS SUppOrtS

Sutherland’s contention that indiscriminant application of

conversion factors can lead to serious e1Tors in estimation of

C。,g from LOI data.

  The non-zero intercept of the relation between OM

(LOI550)and measured C。,g(Fig.5)is most likely due to loss

of water.Although temperatures of500to550。C are used

by many workers to estimateOMcontents(e.g.Dean,1974;

Heiriαα1.,2001),10ss ofwater from clay minerals at lower

temperatures may be significant.Ba11(1964)stated that

most water loss f士om oven-dried clays occ皿s between450

and600℃,and cited work by Keeling(1962)that found

prolonged ignition(16h)at375℃removed90%ofcarbonaceous matter without affecting structural water in

clays.Clay minerals are a significant component of the

shales analyzed here.XRD analysis has shown the clay

mineralogy consists of primarily of illite,chlorite,and

smectite,with lesser montmorillonite and pyrophyllite in

some samples(Campos,unpubl.data).The samples are also

highly aluminous,with up to27.3wt%A12030n a hydrous

basis(Campos,unpub1.data),suggesting that the clays are a

m句or component.

  The results of the third test(successive ignitions at375,

550and1000℃)also suggest loss of struct皿al water from

clay minerals during ignition between375and550℃.A

StrOng COrrelatiOn eXiStS betWeen LOI375and meaSUred C、)rg,

and the slope is near unity(Fig.7a).In contrast,weight loss

between375and550℃(LOI,,。.,7,)is more or less constant

(2-4 wt%)irreSpeCtive Of C。,g Content(Fig.7b).This

suggests that the bulk of the loss during Prolonged ignition

at375℃is due to combustion of OM,and a more direct

eStimate Of C、)rg Can be Obtained by ignitiOn at the IOWer

temperature.The correlation between LOI375and C。rg can be

imprOVed by applying the LOI5500M tO C。,g COnVerSiOn

factor(1.12)derived from Fig.5.A plot of the converted

data shows the slope is virtually unity and the intercept is

near zero(Fig.7c).

  The laCk Of CO∬elatiOn betWeen hydrOgen and C、,rg

contents noted above also supports loss of water during

ignition.Crude modelling gives some estimate of the

amount of water that may have been lost.If an arbitrary

60%ofthe hydrogen in the CHNS analyses is recast as H20

N.O.Campos,BP.Roser and Y.Sampei 15

8

7    6    5    4    3    2    1

ぶぎ7(ooo』O」)のOQ邸QコΦ一応∈眉ω山

0

勘 ●. ●

1:1

     012345678           Estimated CaCO3(CHNS)wt%

Fig.8.Estimated CaCO3contents in carbonate-poor samples by

 LOllooo and CHNS methods.

and deducted from LOI550,and the residual H summed with

C。,、,N,and S tO better refleCt the COmpositiOn Ofthe organiC

matter,LOI550and CHNS results become comparable(Fig.

7d).In the entire sample suite,only PES-5and PES-31

diverge from a l:1relationship to any extent.The latter

sample has the highest H content(L27wt%)as well as high

C。,、(1353wt%),and a differing prOpOrtiOn of the hydrOgen

may have been lost.

  The comparison ofresults at375and550℃suggests that

ignition at either temperature can provide an estimate of

OM COntentandhenCe C。,g,prOvided ignitiOn iS OfSUffiCient

duration to oxidize all organic matter.Structural water

appears to be a significant component of the LOI550weight

losses,but the strong correlation with CHNS C。,g still allows

reasonable estimates to be made.Although the results at

375℃giveamoredirectcomparison,thelongerignition

time required limits sample throughput.In either case,

determinatiOn Of C。,g in repreSentatiVe SampleS by an

independent method such as CHNS analysis is essentia1,so

that calibration for individual suites can be made.This is

necessary  given  the  variation  in  conversion  facto「s

highlighted by Sutherland (1998). Generic conversion

factors are clearly inapPropriate.

  Comparison of the LOIloooand CHNS methods of

determining carbonate contents was limited here by the

range in carbonate contents of the suite analysed.Only three

samples(PES-1,一5,一31)have significant LOIlooo values;

they were also the only samples which effervesced during

acid addition in the CHNS method.Results for PES-5,a

highly calcareous shale,are good.Assuming all weight loss

at1000℃isCO,,andallC、。.,、iscontainedincalcite,

calculated CaCO3contents are69.84wt%(LOIlooo)and

68.29wt%(CHNS).The results for PES-1and PES-31are

very poor.PES-l has an indicated CaCO3content of

17.81wt%by the LOI!ooo method,but only2.94wt%from

reCa1CUlatiOn Of Ci。。,g frOm the CHNS analySiS.Neither

seems correct.Assuming all CaO in this sample(5.89wt%,

hydrous basis)is present as calcite,then maximum CaCO3

content should be10.5wt.%.PES-31retums an estimate of

8.56wt%from LOIlooo,but Ci.。,g was not detected in the

CHNS analysis.The LOIlooo estimate is unrealistic,because

the CaO content ofthis sample is only1.61wt%,(Campos,

unpub1.data),equivalent to2.87wt%CaCO3.The cause of

the poor results for these two samples is unknown.Both

samples had lower a(Uusted LOI550than expected(Fig.5).

They may contain a carbonaceous phase(e.g.graphite)that

was not lost until the1000℃ignition,or contain other

hydrous phases that s皿vived the lower temperature

ignition.However,their whole-rock elemental compositions

and XRD mineralogy (Campos,unpub1.data) are

unexceptional.

  Indicated CaCO3contents of the remaining samples by

both methods are<6%,and most are less than2%(Fig.8).

PES-14and-29have much greater CaCO3contents by

CHNS than by the LOI method.Their low CaO contents

(0.69and O.38wt%respectively;Campos,unpub1.data)

suggest that the CHNS data are incorrect,possibly due to

dissolution of non-carbonate minerals by the acid addition

USed tO determine C。,、,and COnSeqUently Ci。。,、.AlthOUgh

some of the remaining samples yield comparable values by

the two methods,a number have greater estimates from

LOIlooo.These include five samples in which Ci,。,g was not

detected,but which have LOLooo calcium carbonate values

ranging from O.6to2.25wt%.

  The poor estimates of carbonate contents by both

methods at low contents may be partly due to loss of other

volatiles between550℃and1000℃,as suggested by Heiri

αα1.(2001).This could include more tightly-bound

stmct皿al water from hydrous silicate minerals other than

clays.Other volatile species may include SO2fromoxidation

of sulphide minerals,and F and Cl.Moreover,at these

temperatures,weight gain through oxidation of Fe2+to Fe3+,

and ofMn2+to higher oxidation states may also occur.The

combination of these effects could obscure the results and

contribute to the variability observed.

Conclusions

  Simple loss on ignition procedure is a popular method for

estimating OM and carbonate contents in sediments and

soils。Tests of the method used by Heiriαα1.(2001)and

others suggest that the OM in P句a Fomlation black shales

waslostafterthreeto sixhours heating at550℃,andthree

further hours of ignition at1000℃ were sufficient to

completely decompose the carbonate fraction.OMand

eStimated C。,g COnCentratiOnS determined by the LOI methOd

were greaterthan those determined using a CHNS elemental

analyzer,even when modified conversion factors were used

assuggestedbySutherland(1998).Igniti・ntestsat375℃

16

  Organic carbon and carbonate contents ofblack shales from the lower Cretaceous

Paja Formation(Colombia)by loss on ignition and CHNS analysis:comparison ofmethods

give a better COrrelatiOn betWeen OM and C。,,,and SUggeSt

that much of the weight loss between375and550℃is

structural water from clays,which are an important

component of the samples analysed.However,ignition at

either temperature iS viable,prOVided C。,g data frOm an

independent method such as CHNS analysis is available for

CalibratiOn,and SpeCifiC OM tO C。,g COnVerSiOn faCtOrS Can

be derived for the sample suite concemed.The viability of

estimating calcium carbonate contents from LOIlooo data

could not be tested in our sample suite due to its generally

carbonate-poor nature.Neither the ignition method nor

CHNS analysis produced reliable results at the low levels of

CaCO3present in most samples.Continued loss of volatile

species and water from other hydrous phases may have

contnbuted,and oxidation reactions causing weight gain

may have also obscured the results.

Acknowleαgements

 This work forms part of an M.Sc.study by N.O.C.at

Shimane University,financially supported by a scholarship

from the Inter-American Development Bank(IDB).The

laboratory workwas supportedby aMonbukagakusho grant

-in-aid to B.P.Roser.

References

Ball,D.F.1964.Loss-On-Ignition as an Estimate of Organic Matter and

 Organic Carbon in Non-Calcareous Soils.Joumal ofSoil Science,15,84-

 92.

Colleta,B.,Hebrard,F.,Letouzey,J.,Wemer,P.,and Rudkiewicz,J-L.

 1990.Tectonic style and crustal stmcture of the Eastem Cordillera

 (Colombia)丘om a balanced cross section:1n J.Letouzey(Editor),

 Pαro18況〃1αn4T8αon∫cs∫n〃。わ∫18β81‘s.Edition Technic,Paris,81-100.

Cooper,M.,A.,Addison,F.,T。,Alvarez,R,Coral,M.,Graham,R.,H.,

 Hayward,A.,B.,Howe,S.,Martinez,J.,Naar,J.,Pe且ochas R.,Pulham,

 A.,J.,and Taborda,A.1995.Basin development and tectonic history of

 the Llanos Basin,Eastem Cordillera and Middle Magdalena Valley,

 Colombia.A耀r’cαn Asso6妨’onげ、P8∫ro18配1n G80108’s∫s、Bμl18∫∫n,79,

 1421-1443

Dengo,C.A.&Cove,M.C.1993.Structure of the Eastem Cordillera of

 Colombia:Implications for trap styles and regional tectonics.Aln顔。αn

 Assoo’απon6ゾP8かo18況〃10召olo9∫s孟s Bμ〃8πn,77,1315-1337

Dean Jr,W.E.1974.Detemination of carbonate and organic matter in

 calcareous sediments and sedimentary rocks by loss on ignition:

 Comparison with other methods.10μrnα1犀Sθ4枷8n∫硯y P8孟rology,44,

 242-248.

Heiri,0.,Lotter,A.F.,&Lemcke,G.200L Loss on ignition as a method for

 estimating organic and carbonate content in sediments:reproducibility

 and comparability ofresults./0μrnα1犀Pα1801〃nnolog』y,25,101-l lO.

Keeling,P.S.1962。Some Experiments on the Low-Temperature Removal

 of Carbonaceous Material from Clays.Clαy〃∫n8rαls Bμllα’n,28,155-

 158(cited in Ball.1964).

Minoru.,1.,Kimitaka.,K.,Naohiko,0.,and Asahiko,T.1999.Organic

 geochemistry of greenish clay and organic-rich sediments since the early

 Miocene from hole985a,Norway basin.1nr Raymo,M.E.,Jansen,E.,

 Blum,P.,and Herbert,T.D.(Eds.),Proc8α11ngs犀‘h80cθαn Dr∫ll’ng

 Progr伽,Sc’8n峨’cR8sμ1‘s,162,209-216.

Pimmel,A.,and Claypoo1,G,200L Introduction to shipboard organic

 geochemistry on the JOIDES Resolution.Oc召αn Dr∫ll∫ng Progrα珊丁86h.

 ハゐ。‘8,30[Online],2g PP.

Roser,B.,Sawada,Y.,and Kabeto,K.1998.Crushing performance and

 contamination trials of a tungsten carbide ring mill compared to agate

 grinding.σ80sci8nc8R8poπsρプSh枷αn8Un’v8rs妙,17,1-9.

Royero,J.M,and Clavijo,TJ.2000.Mapa geo160chgico generalizado del

 Departamento de Santander,memoria explicativa.INGEOMINAS.

 Bucaramanga,Colombia、

Sampei,Y.,and Matsumoto,E.,2001.αN ratios in a sediment core from

 Nakaumi Lagoon,southwest Japan-usefulness as an organic source

 indicator.G80ch8規’cα1/0μrnα1,35,189-205.

Sutherland,R.A。1998.Loss-on-ignition estimates of organic matter and

 relationships to organic carbon in fluvial bed sediments.H』y4roわ∫olog’α,

 389,153-167.

(Received:210ct.2002,Accepted:310ct.2002)

(要 旨)

N.O.Campos・B.P.Roser・三瓶良和,2002,コロンビア下部白亜系パジャ層黒色頁岩の有機炭素濃

度と炭酸塩濃度の測定=イグニッションロス法分析結果のCHNS元素分析法による検討,島根大学地

球資源環境学教室、21、9-16.

  堆積岩や土壌の有機炭素濃度と炭酸塩炭素濃度は,一般に,簡便なイグニッションロス法(LOI法)

 を用いて,それぞれ375-550℃と850-1000℃下で加熱して求めることができる.本研究では,24個

 の有機物に富んだ黒色頁岩試料を用いて,有機炭素濃度と炭酸塩炭素濃度を2つの方法(LOI法と

 CHNS元素分析法)で分析し,簡便なLOI法で得られる値の正確な意味について議論した.検討に

 先立ち,有機炭素濃度と炭酸塩炭素濃度を求めるLOI法の最適加熱時間を決めるため,550℃と

 1000℃下で時間を変えて予備加熱実験を行った.その結果,有機炭素の燃焼には550℃下で4-6時

 間(LOI55。),炭酸塩の加熱には1000℃下で3時間(LOI1㎜)がそれぞれ必要であることが分かった.こ

 れらの条件下で黒色頁岩試料のLOI分析が行われCHNS元素分析結果と比較したところ,両者には

 強い相関関係が見られるものの,値には差が生じた.有機炭素濃度は,LOI,,。の方がCHNS元素分

 析値よりも一貫して高くなった.この理由については,LOI,,。では約375℃で有機物が燃焼し,375

 -550℃間でほぼ一定量(2-4wt%)の粘土鉱物の構造水が放出されることが原因と推察された.従っ

 てこれらを回帰計算で補正することにより,LOI,,。でも有機炭素濃度を正確に求めることができる.

 一方,炭酸塩炭素に関して,特に低濃度の場合にLOll㎜とCHNS元素分析値はバラついて一致しな

 かった.その理由には,高温下でも粘土鉱物構造水の放出が継続していることや他の揮発性元素の

 逸散,また,Fe2+やMn2+が酸化して重量が変化することなどが示唆された.