Earth-Science Ret,tews, 32 (1992) 2 3 5 - 2 8 3 235

Elsevier Science P u b l i s h e r s B V , A m s t e r d a m

Manual for the geochemical analyses of marine sediments

and suspended particulate matter

D.H. Loring and R.T.T. Rantala

Marine Chemistry DIt lslon, Physical and Chemical Sciences Branch, Dept of Fisheries and Oceans, Bedford Institute of Oceanography,
Dartmouth, N S B2Y 4A2, Canada
(Accepted May 22, 1991)

ABSTRACT

Lorlng, D H and Rantala, R T T , 1992 Manual for the geochemical analyses of marine sediments and suspended
particulate matter Earth Sct Rec, 32 235-283

Accurate and precise sampling and analytical procedures are essential in environmental geochemical studies This report
provides a detailed description of the techniques and analytical procedures for samphng, grain size determinations, and for
precise and accurate AAS determination of the major and trace metals in marine sediments and suspended particulate
matter In addition, it describes the procedures for the chemical partition of the metals, determination of readily oxidizable
organic matter, and calcium carbonate A separate section discusses the normalization of trace metal data

Table of c o n t e n t s

l 0 INTRODUCTION
2 0 SAMPLING ESTUARINE AND COASTAL SEDIMENTS
2 1 Estuarme sediment classification 241
2 2 Samph" collection and fteM obsercatlons . . . . . 241
2 2 1 G r a b s a m p l i n g of surface s e d i m e n t s . . . . . . . . 241
2 2 2 Barrel core s u b - s u r f a c e s e d i m e n t s a m p l i n g . . . 242
2.2 3 Box core s e d i m e n t s a m p h n g . . 243
3 0 SAMPLE PREPARATION PROCEDURES
3 1 Sample preparation . . . . . . . 244
3.1 1 L a b o r a t o r y e q m p m e n t ...... 244
3 1 2 Imtlal p r e p a r a t i o n ......... 244
3.1.3 Drying s e d i m e n t s . . . . . . . . . . . . ..... 244
314 Water content . . . . . . . . . . . 245
3 1 5 D e t e r m i n a t i o n of w a t e r c o n t e n t in wet s a m p l e s (Oven drying) . . . . . . 245
3 1 6 D e t e r m i n a t i o n of w a t e r c o n t e n t m air dried s a m p l e s ( O v e n drying) 245
3 1 7 Salt c o n t e n t correction . . . . 245
3 1 8 R e m o v a l of soluble sea salts .. 246
3 1.9 Storage of s a m p l e s 246
4 0 SEDIMENTOLOGICAL ANALYSES
4 1 Gram size separation . . . . . . . . . . . . . . . . . . . . 246
4 1.1 D e t e r m i n a t i o n of s a n d a n d m u d s~ze m a t e r i a l 246
4 1 2 D e t a i l e d grain size s e p a r a t i o n s for trace metal d e t e r m i n a t i o n s 247
4 1 3 Storage of s a m p l e s . . . . . . . . 248

Correspondence to D . H Loring, M a r i n e C h e m i s t r y Dwlsion Physical a n d C h e m i c a l Sciences B r a n c h D e p t of

F i s h e r i e s a n d O c e a n s B e d f o r d I n s t i t u t e of O c e a n o g r a p h y , D a r t m o u t h , N S B2Y 4A2, C a n a d a

0 0 1 2 - 8 2 5 2 / 9 2 / $ 1 5 . 0 0 ~) 1992 - Elsevier Science P u b h s h e r s B.V All rights reserved

236 D H LOR1NG AND R T T RANTALA

5.0 G E O C H E M I C A L A N A L Y S E S O F S E D I M E N T S
5.1 S e d i m e n t d e c o m p o s i t i o n m e t h o d s . . . . . . . . . . . 249
5 1 1 Total decomposition m e t h o d s . . . . . . . 249
5.1.2 Strong acid ( n o n - H F ) digestzon . . . . . . . . . . . . . 249
5 1.3 M o d e r a t e or weak acid extractions . . . . . . . . . 249
6.0 T O T A L (HF) D E C O M P O S I T I O N O F S E D I M E N T S
6 1 A p p a r a t u s a n d reagents . . . . . . . . . 249
6 1.1 Teflon bomb construction . . . . . . . 250
6 1.2 Teflon bomb heating e q m p m e n t , . 250
0 1 3 Laboratory e q u i p m e n t . . 250
0.1.4 L a b w a r e , 250
0 1 5 Reagents . . . . 250
6 1 6 Sample size . , . 251
6 2 Total de~ o m p o s t t t o n p r o c e d u r e ..... 251
6 2.1 Storage of sample solutions 251
7 0 ATOMIC ABSORPTION ANALYSES OF SEDIMENTS
7,1 A p p a r a t u s a n d e q u i p m e n t .... 252
7 2 C a h b r a t l o n standards" . . . . . . 252
7 2 1 Ak Ca, Fe, K, Mg, and Na multi-element standards . 252
7.2 2 Cd, Co, Cu, LI, Mn, Nl, Pb, and Zn multi-element standards 253
7 2.3 SI single e l e m e n t standard . . . . . 253
7 2 4 TI single e l e m e n t standard . . . . . 253
7.2 5 Ba single e l e m e n t standard . . . . . 253
7 2.6 Be single e l e m e n t standard* .... 253
7 2 7 Cr single e l e m e n t s t a n d a r d . . 253
7.2 8 Rb single e l e m e n t standard 253
7 2 9 Sr single e l e m e n t standard . . . . . . . 254
7 2 10 V single e l e m e n t standard 254
7 3 A l t e r n a t t c e ~ a h b r a n o n s t a n d a r d s (preferred) 254
7 4 R e f e r e n c e materials' ........ 254
7 5 S a m p l e dth~ttom . . . . . . . . . . . . . . . . 255
7 5 1 Solutions for major and some minor e l e m e n t d e t e r m i n a t i o n s 256
7 5.2 Solutions for trace e l e m e n t d e t e r m i n a t i o n s 256
7 6 S t a n d a r d o p e r a t i o n a l c o n d i t i o n s / s e t t i n g s f o r f l a m e a t o m i c absorption analyses . . . 256
7 7 htterfi'rences . . . . . . . . . . . 256
7.8 G r a p h i t e f u r n a c e d e t e r m m a t t o n s . . . . . . . 256
7 8 1 Calibration standards . . . . . . . . . . . 257
7 8 2 Use of L'vov platforms . . . . . . 257
7 8 3 Automatic sample introduction . . . . . 257
7 8 4 Sample cups . . . . . . . . . . . . . 257
7 8 5 Optimization of the furnace program e.g Cd 257
7 9 htter/erences . . . . . . . 258
7 10 C o m b i n e d analytical s c h e m e . . . . . . . . . 258
7 11 Relatu'e a c c u r a Q a n d precision . . . . . . . . 258
7 12 C o m p a r i s o n o f m u ' r o w a c e a n d c o n c e n t l o n a l (bothng water) teflon b o m b digestion 259
8 0 C H E M I C A L PARTITION OF SEDIMENTS • ACETIC ACID (HOAc) E X T R A C T I O N O F SEDI-
MENTS
8 1 A p p a r a t u s attd reagents . . . . . . . . . . . . . . . . 261
8 1 I Laboratory e q u i p m e n t . . . . . 261
8 1 2 Labware . . . . . . 261
8.1.3 R e a g e n t s . . . . . . . 261
8.2 E r t r a c t t o n p r o c e d u r e . . . . . . . . . . . . 261
8,3 A t o m i c absorption analyses . . . . . . . . 262
8 3.1 A p p a r a t u s and e q u i p m e n t . . . . . . . . . . . . . 262
8 3 2 Calibration standards . . . . . . . . . 262
8 3 3 R e f e r e n c e materials . . . . . 262
MANUAL FOP, G E O C H E M I ( S A L ANAL't SES 237

8 3 4 Flame AAS determinations .... 262

8 3 5 Graphite furnace determinations . . . . 262
8 3.6 L ' v o v p l a t f o r m c o n d i t i o n i n g for C d . . . . . . . . . . . . . . . 262
8.3.7 F u r n a c e c o n t r o l p r o g r a m s t e p s . . . . . . . . . . . . . . . . . . . . .... 262
8.4 Calculattons o f n o n - d e t n t a l (actd soluble) and detrttal (acM msoluble) contrtbuttons . 263
9 0 FLAMELESS ATOMIC ABSORPTION ANALYSIS FOR MERCURY (Hg)
9.1 Apparatus a n d reagents •• 263
9 1.1 A p p a r a t u s 263
O 1 2 Mercury system setup ... 263
9 1 3 Reagents . . . . . . . . . . . 263
9 2 Sample preparation . . . . . . . . . . . . 263
9 3 Preparatton o f standards . . . . . . . 264
9 3 1 Hg 1000/xg/ml stock solution 264
9 4 Working standard,~ . . . . . . . 264
9 5 Deterrmnatton o f mercury ...... 264
9 6 Cahbratlon curt'e ........ 264
9 7 Relauce accuracy and prectston . . . . . . . . . . . . . . . . . 264
10.0 D E T E R M I N A T I O N OF READILY OXIDIZABLE ORGANIC MATTER
10.1 Apparatu3 and reagents . . . 265
101.1 A p p a r a t u s . 265
lII 1.2 R c a g e n t s ........ 265
10.2 D e t e r m m a t t o n o f organtc carbon . . . . . . . . . . . 265
10.3 C a k u l a t t o n o f results . . . . . . . . . . . . . . . 265
10.4 Standardization o f orgamc carbon d e t e r m m a t t o n s 266
10.5 Precision of d e t e r m m a t t o n s 266
1l 0 DETERMINATION OF CALCIUM CARBONATE
11.1 Apparatu3 and reagents 266
I1 1 I A p p a r a t u s . . . . . . . . . . . . 266
I I 12 Reagents . . . . . . . . . . . 266
1 l 2 Determination o f carbonate . . . . . . . . . . . . . . . . . . . 266
11 2 1 C a l c u l a t i o n o f c a l c m m c a r b o n a t e c o n t e n t . . . . 267
11 3 Precl,slon o f carbonate d e t e r m m a t l o n s . . 267
12 0 S U S P E N D E D P A R T I C U L A T E MATTER (SPM)
12 1 Fdter preparatton .... 268
12 1 1 F d t e r h o l d c r ~ ....... 268
12.2 Collectton of S P M samples . . . 268
12 2 1 O f f - h n e f i l t r a t i o n 268
12 2 2 O n - h n e f i l t r a t i o n . . . . 269
12 3 Fdter washing .... ! 269
12 4 Total H F dec omposttton o f suspended partwulate m a t t e r . . 269
12 5 Apparatus a n d reagents . . . . . . . . . . . . . 269
12 5 1 L a b o r a t o r y e q m p m e n t . . 269
12 5 2 L a b w a r e . . 269
12 5 3 R c a g c n t s . . 269
12 5 4 S a m p l e size 27O
12 6 Total d e c o m p o 3 m o n procedure . . . . . 27O
12.6 1 A p r c f e r r e d a l t e r n a t i v e m e t h o d for low c o n c e n t r a t i o n s o f m e t a l s . 27O
12 7 A t o m t c ab3orptton analyses o f S P M . . 270
12 7 1 A p p a r a t u s . . . . . . . 270
12 7 2 C a h b r a t l o n s t a n d a r d s . . . 270
12 7.3 R e f c r c n c e m a t e r i a l s .... 270
12 7.4 F l a m e A A S d e t e r m i n a t i o n s . . . 271
12 7 5 G r a p h i t e f u r n a c e d e t e r m i n a t i o n s . . 271
12 7 6 F u r n a c e c o n t r o l p r o g r a m s t e p s . . . . . . . 271
12 ~ Acettc acid extraction o f S P M 271
238 D H LORIN(; AND R T I RANTALA

12.9 A p p a r a t u s a n d reagents . . . . . 271

12 9.1 Extraction apparatus . • 271
12 9.2 L a b w a r e .... 271
12 9 3 R e a g e n t s . . . . . . . . . . . . . . . . . . . . . . . 271
12 9 4 S a m p l e size . . . . . . . . . . . . . . . . . 271
12.10 E x t r a c t i o n p r o c e d u r e . . . . . . . . . 272
12.11 A t o m t c absorption analyses . . . 272
12 11 1 Apparatus and equipment . . . . . . . 272
12 11 2 Cahbratlon standards . . . . . . 273
12 11.3 R e f e r e n c e m a t e n a l s . . 273
12 12 F l a m e A A S d e t e r m i n a t i o n s . . . 273
12 13 G r a p h i t e f u r n a c e d e t e r m m a t t o n s . . . . . . . 273
12 14 L "coc p l a t f o r m c o n d t t t o m n g f o r C d .... 274
12 15 F u r n a c e c o n t r o l p r o g r a m steps . . . . . . . . 274
12 16 D a t a reporting . . . . . . . . 274
12 17 Results o f m t e r c a h b r a t l o n exerctse (s) f o r s u s p e n d e d p a r t t c u l a t e m a t t e r 274
13 () G U I D E L I N E S F O R T H E N O R M A L I Z A T I O N OF GEOCHEMICAL DATA
13.1 I n t r o d u c t t o n . . . . . . . 274
13.2 R e q u t r e m e n t s f o r g e o c h e m w a l data 275
13 2 1 F i e l d samphng 275
13 2 2 C h e m i c a l a n a l y s e s . . . 276
13 2.3 Selection of metals . . . . 277
13.3 B a c k g r o u n d lel'el~ . 277
13 4 N o r m a h z a t t o n p r o c e d u r e s 277
13 4 1 S i m p l e n o r m a l i z a t i o n (Level I) . . . . 277
13 4 2 Mathematical normalization ( L e v e l II) . 278
13 4 3 M e t a l g r a i n size n o r m a h z a t l o n . 278
13 4.4 M e t a l reference metal normahzation 279
13.5 M u l t t - e l e m e n t n o r m a h z a t t o n .... • . o 281
13 6 S u m m a r y (~f n o r m a h z a t l o n p r o c e d u r e s 282
14 0 R E F E R E N C E S . . . . 282

1 0 INTRODUCTION
iments. As a result, the heavy metal data
In 1980, the International Council for the produced in different laboratories in the
Exploration of the Sea (ICES) formed a same country and in different countries was,
working group (WGMS) to examine the use and still is, mostly incomparable and unveri-
of sediments and suspended particulate mat- fiable (Loring and Rantala, 1988). On these
ter (SPM) for monitoring contamination in data, sedimentary environmental decisions
the marine environment. Since then, this have been, and in many countries are still
group has systematically reviewed and con- being, made.
stdered the physical, chemical and biological In 1961, the Canadian Department of
characteristics of sediments as well as the Fisheries and Oceans initiated a series of
processes that contribute to their role as a environmental geochemical studies to deter-
monitoring tool. mine the levels, behavior, and dynamics of
During these deliberations, it became clear heavy metals in eastern Canadian estuarine
that laboratories in each country represented and coastal sediments and suspended partic-
on the WGMS used different methods for ulate matter• For such studies it was neces-
the sampling and chemical analyses of sedi- sary to acquire and develop suitable sample
ments such as the use of different grain size collection, preparation techniques, and pre-
fractions and acid digestion methods for the cise and accurate analytical procedures for
determination of heavy metals in marine sed- the determinations of A1, Be, Ca, Cd, Co, Cr,
M A N U A L FOR G E O ( H E M I ( A L ANALYSES 239

Cu, Fe, Hg, K, Li, Mg, Mn, Na, Ni, Pb, Rb, natural distribution of metals in sediments
Si, Sr, Ti, V, and Zn. In addition, procedures and suspended particulate matter.
were developed to determine carbonates and This manual is based on the sampling and
organic carbon matter which influence the analytical experience we have acquired and

career he has published more than seventy five

scientific papers dealing with various aspects of
the geochemistry of marine sediments and SPM
from the Bay of Fundy, Gulf of St. Lawrence,
Baffin Bay, Greenland, the Tamar estuary (U.K.)
and the Wadden Sea (Germany). In addition, in
conjunction with Dr. D.J.G. Nota (Wageningen,
J The Netherlands) carried out the initial survey of
the morphology and sediments (granulometry,
Douglas H. Loring received his B.Sc. (1954) and mineralogy, chemistry) of the Gulf of St. Lawrence
M.Sc. (1956) in geology from Acadia University, and published the results in book form. Over the
Wolfville, N.S. After a year of post-graduate last ten years, he has also been involved in the
study at M.I.T. (Boston, Mass.), he went to International Council for the Exploration of the
University of Manchester, England as a Research Sea (ICES) and the Intergovernmental Oceano-
Assistant, and in 1960 received a Ph.D. in sedi- graphic Commission (IOC) working groups and
mentary geochemistry. Since 1960 he has been workshops on marine sediments in relation to
employed as a Research Scientist with the De- pollution.
partment of Fisheries and Oceans. During his

of the Department of Fisheries and Oceans at the

Bedford Institute of Oceanography, Dartmouth,
Nova Scotia. He specializes in atomic absorption
spectroscopy and over 23 years has been keenly
involved in the development of precise and accu-
rate methods for the determination of metals in
rocks, sediments and marine suspended particu-
late matter. This has resulted in the publication
of over 20 research papers and the development
Reijo T.T. Rantala was born in Finland, Decem- of the first commercially available all PTFE
ber 30, 1940. He immigrated to Canada in 1965. digestion bomb. He has also made significant
He was employed from 1965 to 1967 with the contribution to the National Research Council of
Geology Department of Dalhousle University, Canada's certified reference sediment program as
Halifax, Nova Scotia where he undertook rapid well as contributing to the United States Geologi-
silicate analyses and studied analytical chemistry. cal Survey manganese nodule and rock standards
Since 1967, he has been employed as a Geochem- program.
ical Analyst with the Marine Chemistry Division
240 D H I_ORING AND R l J RANI'ALA

developed in our geochemical studies over Chemical partition of sediments is used to

the past 30 years. It has been written to deduce the source and pathways by which
provide detailed instructions on the analyti- the natural and anthropogenic heavy metals
cal procedures required for environmental have entered the marine environment. Sec-
workers to acquire precise and accurate geo- tion 8 of this manual details the analytical
chemical data for making sound environmen- procedure used in our laboratory to estimate
tal decisions. the partition of the total metal concentra-
The general principles of sample collec- tions in sediments. It is an operationally de-
tion, field observations, sample preparation fined method based on the use of acetic acid
and grain size separation are covered at the to release the proportion of loosely bound
beginning of the manual in Sections 2, 3 and metals from such sites as ion-exchange posi-
4. Sampling strategy is covered briefly in tions,amorphous compounds of Fe and Mn,
Section 13. These procedures are not com- carbonates and those weakly held in organic
prehensive and those whose knowledge of matter. Data from such determinations have
them is limited should consult standard texts allowed us to make some deductions as to
for further detailed instructions. the carriers, transport mode and potential
The following three Sections (5, 6, and 7) bioavailability of metals (Loring et al., 1983,
of the manual deal with detailed instructions 1985).
for sample digestion and atomic absorption Accurate and precise analytical data for
determinations of total concentrations of ma- mercury (Hg) is essential because Hg is often
jor, minor and trace metals in sediments. a serious contaminant of sediments in many
Hydrofluoric acid (HF) in combination with harbors and estuaries. The analytical proce-
aqua regia is used in a sealed teflon decom- dures using flameless atomic absorption for
position vessel immersed in boiling water or the determination of mercury in sediments
in a microwave oven to decompose the sam- are detailed in Section 9.
ples for all metal determinations except Hg. Organic carbon matter is often deter-
The use of HF is essential because it is the mined to assess the role played by the or-
only acid that completely dissolves the sili- ganic fraction of sediments in the transport,
cate lattices and releases all the metals in- deposition and retention of metals in sedi-
cluding associated metals such as A1, Fe, and ments. Section 10 details the Walkey-Black
Li used for normalization. In addition, the titration procedure used in our laboratory
HF decomposition procedures allow for veri- for the determination of readily oxidizable
fication of the results by the use of certified organic matter using cold sulfuric acid and
reference materials. In the past, many labo- potassium dichromate.
ratories have used aqua regia or nitric acid Carbonate is sometimes an important
decompositions for heavy metal determina- component of sediments. Section 11 details a
tions. These types of digestions are not used simple method for the estimation of carbon-
or recommended here because they have ate contents of sediments.
been found to yield incomplete, inaccurate Suspended particulate matter (SPM) is of-
and unverifiable determinations of metals. ten the conduit by which heavy metals are
For those readers who are concerned carried, removed and deposited in the ma-
about the use of HF in general and in the rine environment. Section 12 details the pro-
method in detail, a 20 minute VHS video cedures used for the sampling and prepara-
(PAL, NTSC and SECAM systems) is avail- tion of SPM for metal analyses. Detailed
able on request from ICES (Palaegade 2-4, instructions are found in this section for the
DK-1261 Copenhagen K, DK) for the price decomposition of SPM samples with HF and
of copying. aqua regia in sealed teflon vessels, the ex-
MANUAL FOR GEOCHEMICAl ANALYSES 241

traction of loosely bound metals with acetic the sea floor and from the water column
acid, and the atomic absorption techniques using specific types of sampling apparatus.
used for the determination of the metals. This is done to ensure that the sample is
Since metals from natural and anthro- fully representative of the substrate or water
pogenic sources accumulate together, it is column at a specific geographical location
difficult to determine what proportion of the and that it is uncontaminated by any sub-
sedimentary metal load is natural and what stance that is being analyzed for or will inter-
proportion is anthropogenic. This is because fere with the subsequent treatment of the
of anthropogenic and natural metal inputs sample.
that vary by several orders of magnitude, Two types of sampling techniques are used
depending on the nature, grain size and for collecting sediments from the sea floor:
source of the metal r i c h / m e t a l poor miner- grab sampling which collects surface and near
als or compounds in the sediments. The final surface sediments and coring which collects a
part of the manual (Section 13) presents column of the subsurface sediment. In all
sampling, analytical and interpretative guide- grab and core operations, a slow approach to
lines for the normalization of geochemical the sea floor should be ensured to avoid the
data to account for the natural sedimentary creation of "'bow wave" that would disturb
variability of metals that should be consid- the sediment-water interface.
ered by environmental scientists in evaluat-
ing the extent, if any, of metal contamination 2.2.1 Grab sampling of surface sediments
of marine sediments. Undisturbed surface sediment samples can
provide an immediate assessment of the pre-
2.0 SAMPLING ESTUARINE AND COASTAL sent levels of contamination in the area in
SEDIMENTS
relation to the textural and geochemical
2.1 Estuarine sediment classification characteristics of the sediment. The sampler
used must consistently collect relatively
Estuarine and coastal sediments may be undisturbed samples to a required depth be-
regarded as a mixture of inorganic and or- low the sediment surface and in sufficient
ganic material that has arrived at the site of quantities for subsequent analyses.
deposition as solid particulate matter (detri- Tightly closing grab samplers of which
tus) or has been incorporated into the sedi- there are many designs (Bouma, 1969) are
ments from solution (non-detrital) in a vari- usually adequate for studies of the most re-
ety of ways. They may be classified according cently deposited layer. Special steps can be
to their grain size and relative proportion of taken to minimize contamination of the sam-
the sediment particles, their color, the rela- ple such as the use of a stainless steel grab
tive proportion of their organic and inorganic sampler with teflon coatings on all surfaces
components and their predominant chemical in contact with sediments, and polyethylene
(siliceous, calcareous) and mineralogical coated lowering cables. If the sampler is well
composition. In most northern latitude sedi- designed, no loss of the entrapped sediments
ments, detrital material dominates the non- and water should occur from the grab after
detrital material and determines the broad removal from the sea floor. It is especially
chemical composition of the sediments. The important to avoid leakage of fine-grained
organic component of such sediments is nor- sediments between the sea-surface and the
mally less than 10% by weight. deck, because erroneous grain size and com-
2.2 Sample collection and field observations positional analyses will result. The modified
Van Veen grab sampler used in our studies
Samples of marine sediments and sus- takes a sample of about 0.1 m 2 and pene-
pended particulate matter are collected from trates up to a depth of 15 cm depending on
242 D H LORING AND R T F RANTALA

the texture of the sediments. It has two small of this material should include color and
trap doors in the top that can be used to textural characteristics. To ensure a repre-
inspect the condition of the sample before sentative sample, about 100 g or more should
the grab is opened. be taken and placed in an numbered plastic
Onboard, the sediments contained in the bag. Since the sample contains iron and or-
grab sampler require special attention to en- ganic compounds that are easily oxidized, the
sure that essential components, observations, sample should be quickly frozen for return to
and measurements are not lost because of the laboratory. Samples of about 1 kg are
improper handling. The most critical sam- required for admixtures of gravel, sand and
pling and storage techniques relate to the mud.
avoidance of chemical contamination and
change in the physico-chemical characteris- 2.2.2 Barrel core sub-surface sediment sam-
tics of the sediments. pling
Initially, a visual inspection should be Sediment subsurface samples are often
made of the sample by means of the small taken using barrel, or box corers to deter-
trap doors in top of the grab to ensure that mine the change in lithology and chemical
the sample has been collected in an undis- composition with depth in order to assess
turbed state and to determine if water is environmental changes in metal fluxes with
present on top of the sample. If water is time. They are usually taken in areas having
present, it can be siphoned off with a glass fine-grained sediments but specialized cores
tube or slowly drained off so as not to wash are available for coarse-grained sediments.
the sample unduly. The main types of corers having cylindrical
Once the top of the sediment is exposed, barrels are: (a) the gral,ity corer which free-
visual estimates of grain size (coarse, falls from the ship and penetrates the sea
medium, fine grained), color (according to floor by gravity, and (b) the piston corer
the Munsell color chart code) and the rela- which is released a set distance above the sea
tive proportions of the components should floor, penetrates the sediment by free fall,
be made and recorded. In situ measurements and sucks the sediment into the core barrel
such as Eh or pH can be made by inserting by an upward moving piston as the core is
the appropriate electrodes into the sample. retrieved.
For example, most fine grained sediments Plastic core liners are placed inside the
usually have a thin, dark yellowish brown core barrels to contain the sediment core
surface layer resulting from the oxidization sample, and to avoid the problems of extru-
of iron compounds at the sediment-water sion and contamination that occur in unlined
interface. Since this layer represents the ma- barrels. In general, the greater the diameter
terial being deposited at the present time, it of the liner, the less will be the amount of
should be sampled carefully with a non-con- distortion of the subsurface sediment by the
taminating utensil such as a plastic spatula. corer penetrating the sediments. Core liners
About 1-3 grams for trace metal determina- with diameters > 50 mm are usually satisfac-
tions should be placed in a numbered tory for obtaining samples for geochemical
polyethylene vial, sealed and frozen for purposes.
transport to the laboratory. Samples for hy- After the corer is retrieved, the liners are
drocarbon analysis can be safely stored in capped at the bottom; the liner is removed
glass, teflon, or metal (stainless steel, or alu- from the barrel: the top is capped, and the
minum) containers. After the surface layer core stored in a vertical position until all the
has been sampled, the grab can be opened water inside the liner has risen to the top.
and an additional sample representative of The liner is cut off at the interface with the
the subsurface can be obtained. Observations water, capped and placed in a deep freezer
MANUAl. FOR GEOCHEMI(. AL ANALYSES 243

or a cold room (4°C) for transport to the pylene or borosilicate glass) should be used
laboratory. Visual observations and measure- for temporary storage. Samples for hydrocar-
ments of sediment core samples should in- bon analysis can be safely stored in glass,
clude information on the site number and teflon, or metal (stainless steel, or aluminum)
location, depth, time, core length, lithology, containers. Prior to use, containers and glass
stratigraphy and any distortions in sediment or plastic parts associated with the sampling
layers. equipment should be cleaned with detergent
In the laboratory, core sampling is best and then rinsed with metal-free water. The
carried out by extruding the core upwards samples should be stored frozen, or at a
and slicing off layers ( ~ 1 cm) using a non- sufficiently low temperature ( ~ 4°C) to limit
contaminatmg cutter (e.g. stainless steel, biological and chemical activity. It is recom-
plexiglass) or splitting the plastic core liners, mended that a minimum sample size of 50 g
avoiding the smeared zone around the inside be collected.
of the core liners and sampling the interior
section of the core. The samples are placed 2.2.3 Box core sediment sampling
in containers for transport to the laboratory; Rectangular sampling devices which ob-
plastic bags or wide-mouth jars (polypro- tain cores about 15-25 cm square and 15-60

OBTAINSAMPLE I
FROMGRABOR CORE

I REMOVEMATERIAL
>2mm U
I HOMOGENIZE SAMPLE ~

1 I I I I I

DISSOLUTIONWITH EXTRACTIONWITH DETERMINATION DETERMINATIONOF EXTRACTION ] GRAINSIZE

ORGAN C CARBON W TH HOAc I SEPARATONJ

DETERMINATION I DETERMINATION
OF MAJORAND OF Hg
TRACE METALS
lexcept Hg) [ o
Fig. l Sample preparanon and analyncal sequence for determinations of physical and chemical parameters in
marine sediments.
244 D H LORING AND R T T RANTALA

cm deep are known as box corers and can be trifuge of smaller capacity is also needed for
recommended for detailed sampling at or separating the acetic acid soluble fraction.
below the sediment-water interface. The ad- Polypropylene or glass centrifuge bottles
vantages of the various types of box or square
corers (Bouma, 1969) is that they can recover 3.1.2 Imtial preparation
the surface sediment and fauna virtually in- The selection of sub samples for sedimen-
tact. They can be subsampled by inserting tological and chemical analyses involves pro-
several 5 cm diameter tubes into them. When cedures for subsampling, drying, water con-
subsampling, however, the core material tent determination and removal of the sea
should be taken from the mid-part of the salts.
core to avoid any "edge effects". Such sam-
ples are treated in the same way as the core Extreme care must be taken during sample
samples described above. preparation, as always, to at,oid metalhc con-
tamination by metallic objects, undecontami-
3 0 SAMPLE PREPARATION PROCEDURES nated plastic and glassware and uncouered
hands.
Figure 1 illustrates schematically the labo-
ratory sequence for the determination of var- Initially, samples designated for chemical
ious physical and chemical parameters for analyses of the total sediment are unfrozen
marine sediments. and air-dried for 24 hours, and the coarse
material > 2 mm such as pebbles, coarse
organic fibers and shells are removed. The
3.1 Sample preparation remaining material is crushed and homoge-
nized before being split into subsamples for
Sample preparation for sedimentological the various chemical and physical analyses
and chemical analyses involves procedures and archiving by splitting and quartering.
for subsampling, drying, measurement of wa- Representative samples are then obtained
ter content, wet and dry sieving, crushing using plastic utensils.
and storage.
3.1.3 Drying sediments
3.1.1 Laboratory equtpment Sediment samples must be dried before
Agate mortar and pestle use for grain size and chemical analyses be-
Mechanical shaker cause analytical measurements are based on
Drying oven (drying ovens having teflon a constant dry weight.
coated surfaces and a filtered nitrogen atmo- At least two separate samples are required
sphere are sometimes used to ensure against for metal determinations:
sample contamination) (a) One sample (1-10 g) for trace metal
Nylon or stainless steel sieves (2000, 500, 63, determination, except for mercury (Hg), and
37 p,m diameter) other analyses should be oven dried or freeze
Atterburg cylinders dried to a constant weight.
Evaporating dishes (b) One sample for Hg determinations. Hg is
Analytical balance sensitive to 0.1 to 0.5% of sensitive to drying and requires a separate
the sample weight subsample, oven dried at 60°C or freeze dried
Centrifuge: A large capacity centrifuge is to avoid the loss of any volatile Hg.
highly desirable, although not absolutely nec- Determination of dry weight by the tradi-
essary, for washing samples free of sea salts tional oven-drying method has in recent years
and separating clay size fractions. The same been supplemented and in some cases super-
centrifuge with a smaller head or a cen- seded by freeze-drying.
M A N U A l FOR G E O C H E M I C A L ANALYSES 245

In oven drying, the samples are placed in quirement for constant wmght dry samples,
plastic or teflon containers and put into a the water content should be measured be-
drying oven for 24 h at a temperature of cause it allows a salt content correction fac-
105°C or 60°C (Hg) to remove their intersti- tor to be calculated in sediments of high
tial water. Oven drying at 100-110°C is suit- water content, and allows one to go from
able for preparing the sample for most types measurements of sedimentation rates ( m m /
of analyses but may cause the loss of some yr) to mass accumulation rates when radio-
volatile components and may partially de- chemical analyses are performed.
stroy the structure of some clay minerals.
Drying below 80°C is considered suitable for 3.1.5 Determination of water content in wet
retaining the clay mineral structure and the samples (Oven drymg)
most volatile components. For accurate dry (a) Weigh out accurately 1 to 10 grams of the
weights, samples dried in this way must be wet sample (x grams) into a plastic or teflon
corrected for their residual moisture con- weighing container that has been dried at
tents. 105°C and preweighed;
Freeze drying is one of the most useful (b) Place in drying oven (pre-heated to 105°C)
methods for drying sediments prior to most for 24 h;
chemmal analyses because of the lower risk (c) After drying, place the samples in a desic-
of losing the volatile elements and providing cator until cool;
a powdery material instead of hard clay ag- (d) When the samples reach room tempera-
gregates. Freeze drying, like oven drying, re- ture, reweigh them (y grams);
tains soluble sea salts that interfere with (e) Calculate the water content (expressed in
some major element analyses and prevent percentages) from the equation: water con-
some measurements of physical structure. tent% = weight loss ( x - y ) / x X100.
In addition, recent studies (Bartlett and
James, 1980; Thompson et. al., 1980; Kersten 3.1.6 Determination of water content in air
and Forstner, 1987) indicate that both air, dried samples (Ot,en drying)
oven and freeze drying introduce physical All the results should be expressed on a
and chemical changes in the composition of oven dry weight basis so that the results are
the original sedimentary material after it has independent of variations in air humidity.
been removed from its in situ environment. (a) Follow the procedure outlined for wet
These effects can be minimized by careful samples (Section 3.1.5).
sampling, storage and use of the material. (b) When using an air dried portion of the
Most of the changes relate to oxidation/re- sample, subtract percent air dry moisture
duction changes, and affect the proportion of found from the air dry sample weight to
metals weakly held in the sediments. For obtain the oven dry weight.
example, Kersten and Forstner (1987) found (c) Correction to obtain oven dry weight:
that oven drying and freeze drying, carried Weight of air dry sample = X grams
out separately under vacuum, change the Air dry moisture content = A%
original associations of the metals amongst Oven dry weight of samples = X - A / 1 0 0 X
the various sedimentary components. Since
3.1.7 Salt content correcnon
trace metal concentrations are relatively high
Corrections for salt content are required
in most sediments, such changes are most
for samples containing large amounts ( >
likely insignificant in relation to other errors
70%) of interstitial salt water because high
in total trace metal determinations.
salt contents introduce significant errors in
3.1.4 Water content metal determinations. In many cases, an av-
Fresh sediment samples contain 30 to 95% erage salinity (35%o) and a density of 1.025
by weight of water. In addition to the re- g / c m 3 may be assumed for pore water but
246 D H LORING ~ND R 1 F RANTALA

for sediments with a high water content ( > 3.1.9 Storage of samples
90%), the pore water should be squeezed out After drying, the various sub-samples are
and its salinity determined. placed in a desiccator until cool and stored
The corrections for salt content are made in air tight plastic vials until required for
from the data obtained from the determina- future use.
tion of the water content of wet samples
(Section 3.1.5). 4 0 S E D I M E N T O L O G I C A L ANALYSES
(a) Calculate the salt correction from the
following equation:
Sedimentological analyses involve proce-
Weight (wt.) of sample wet = x grams
dures for determining the granulometric and
Wt. of sample dry = y grams
mineralogical composition of the sediments.
(x - y ) = wt. of water = (wt. of sea-water -
Since this report is concerned primarily with
wt. of salt)
chemical analyses, only the outlines of the
Wt. of water + salt (g) = ( X - Y ) / 1 × 1.025
procedures for grain size determinations are
(density of pore water)
given below and the reader is referred to the
X-1.025(X-Y)=Wt. of dry mud (salt
standard textbooks on the subject for their
corrected).
details.

4.1 Grain size separation

3.1.8 Remot,al o f soluble sea salts
In addition to correcting for the salt con-
The chemical composition of marine sedi-
tent of sediments containing high water con-
ments changes with the grain size and miner-
tents, it is sometimes necessary to remove
alogical composition. Usually the trace metal
the soluble sea salts from the sediments be-
fore analyses. Removal of soluble sea salts is concentrations increase with decreasing grain
size of the material. This is because the host
required for samples which are used for the
minerals of the elements and the surface
determination of the major cations Na, K,
which are also present in sea-water. This is area of the particles making up the sediment
change with grain size. In addition, regional
done to ensure that there is no contamina-
comparisons of elemental concentrations can
tion of sediments by the sea water and that
only be made by using texturally equivalent
there are no chloride-induced interferences
in the chemical analyses. sediments a n d / o r size fractions. It is, there-
(a) Transfer 10 to 20 g of dry sample to a 1 I fore, often necessary to normalize for the
polypropylene centrifuge bottle; "grain size effect" and sometimes to deter-
(b) Disperse it by shaking for 30 rain with 1 1 mine the elemental concentrations in differ-
ent size fractions of the sediments.
of water (distilled and deionized water);
(c) Centrifuge the suspension at 2000 rpm
until clear; 4.1.1 Determinatton o f sand and mud stze
(d) Siphon the water off, repeating until the material
supernatant is chloride free (tested by adding Determination of the amount of sand and
AgNO 3 to a small water sample which pre- mud size material by weight is essential for
cipitates the chloride as AgCI2); the basic classification and identification of
(e) Transfer the sediment to an evaporating the sediment texture. The general scheme
dish and dry in the oven at 105°C; for the separation of the different grain size
(f) Crush the dried sample in an agate mor- fractions is shown in Fig. 2.
tar, mix well to avoid selective crushing, and (a) Use a preweighed dried sample to de-
store in air-tight vials until required. termine the amount of sand ( > 63 /xm) and
M~.NUAL FOR GEO(~ HEMI( AL ANALYSES 247

cal compositions of separated subfractions

< 500 p.m 500 #m SIEVE
are required, the dispersants must be non-
I contaminating such as a weak solution of
500 - 63 pm 63 gm SIEVE
metal-flee ammonium hydroxide (NH4OH).
Alternately, for samples designated for the
63 - 37 p.m 37 gm SIEVE analyses of trace metals in separate size frac-
I tions, the bulk sample can be unfrozen and
< 2 pm <2 !~m CENTRIFUGE the sample wet siet,ed though a 6 3 / z m nylon
I or stainless steel sieve. A separate determi-
16 2#rn < 16 p.m ATTERBERG CYLINDER nation of water content is required to calcu-
I late the amounts of the sand and mud size
37 16 }~m >16 um ATTERBERG CYLINDER material in the sample.
Fig 2 G e n e r a l scheme for gram size ~eparatlon
4.1.2 Detaded gram size separattons for trace
metal determmations
mud ( < 63/xm) size material in the sediment (a) Remove water soluble sea salts from
sample; 10-20 g of sample as described in Section
(b) place the sample in distilled or deion- 3.1.8, but do not dry (this step is not required
lzed water until it can be dispersed; for trace metal determinations, but is only
(c) wet sieve the sample through a 63 /xm necessary if the major cations Na, K, Mg,
sieve, brushing carefully to prevent particles and Ca are to be determined);
from aggregating or sticking to it; (b) wet sieve salt free sample through a set
(d) carefully remove the fraction retained of 500/xm, 63 /xm and 37/xm stainless steel
on the sieve and transfer it to a pre-weighed or nylon sieves, brushing carefully to prevent
teflon evaporating dish; particles from aggregating or sticking to
(e) dry it in the oven at 105°C, cool in a them;
desiccator, and reweigh; (c) collect the fractions from each sieve;
(f) calculate the sand content (expressed in (d) carefully remove the fraction retained
percentages) from the equation: sand con- on each sieve and transfer it to a pre-weighed
tent% = Wt. of sand fraction/wt, of sample teflon evaporating dish;
X 100; (e) dry each fraction at 105°C, cool in a
(g) retain the sand size material for de- desiccator and weigh (except the one from
tailed size analysis of this fraction; the > 500/xm sieve);
(h) collect the mud size material washed (f) resieve the dry > 5 0 0 /xm fraction
through the sieve, if required for further through a 2000 /xm sieve to remove rock
analysis, in a metal-free container; fragments > 2 mm and weigh;
(i) concentrate the mud size material ei- (g) calculate the dry weight percentages
ther by centrifuging, or by allowing time for for each size fraction from the equation: Size
settling; fraction = Wt. of fraction/wt, of sample X
(j) decant excess water and dry * in an 100:
oven or freeze dry. (h) retain the liquid passing through the 37
• Oven dried fine material tends to be /xm sieve:
very cohesive so freeze drying may be prefer- (i) pour into a centrifuge bottle, and make
able. If finer fractions such as the amount of up to 1 1 with 0.16 N N H 4 O H as the disper-
clay size fraction are to be determined (see sive agent;
below), the pipette method or settling cylin- (j) shake mechanically for 30 mln (an ul-
ders are used with chemical dispersants to trasonic probe could be used instead of
prevent aggregation. If, the absolute chemi- N H 4 O H to disperse the sample);
248 D H LORING AND R 1 F RANTALA

(k) separate the < 2 /xm fraction by cen- (s) unclamp the siphon tube and collect
trifuging (the centrifuging time is calculated the 2-16 /xm fraction in the centrifuge bot-
from Stoke's law, Tanner and Jackson, 1947). tle;
If a centrifuge is not available, the < 2 /xm (t) repeat until clear;
fraction can be separated in a beaker accord- (u) centrifuge off the excess water from
ing to the sampling times shown in Table 1; the 2 - 1 6 / x m fraction;
(l) Siphon the < 2 p.m material off and (v) transfer the fraction to a pre-weighed
store in centrifuge bottles; evaporation dish, dry at 105°C, cool in a
(m) add 0.16 N NH4OH again, shake the desiccator and weigh;
centrifuge bottle mechanically, and cen- (w) transfer the 16-37 /xm fraction re-
trifuge; maining in the Atterberg cylinder to a pre-
(n) repeat the whole process until the su- weighed evaporation dish, dry at 105°C, store
pernatant siphoned off is clear. Usually four in a desiccator until cool and weigh.
times is sufficient, although this depends on
the amount of clay present; 4.1.3 Storage of samples
(o) centrifuge the stored < 2 /xm fraction Dried fractions are crushed, mixed and
again to remove excess water; stored in vials until required for analyses.
(p) transfer it to a pre-weighed teflon dish,
dry at 105°C (60°C for Hg), cool in a desicca- 5 0 GEOCHEMICAL ANALYSES OF SEDIMENTS
tor and weigh;
(q) transfer the 37-2 p.m fraction from the Geochemical analyses of sediments in-
bottle into an Atterberg cylinder, marked 30 volves procedures for their decomposition
cm from the bottom of the siphon tube; and chemical analyses such as are outlined in
(r) make up to the mark with 0.16 N Fig. 1. The number of steps that are followed
NH4OH, shake, and let stand for the speci- will depend on the nature and extent of the
fied time intervals shown in Table 1; investigation. It is recommended that, at

TABLE 1
Sampling times for size fractions in b e a k e r or A t t e r b e r g cyhnder

Size < 2 #m < 4/xm < 8 p.m < 8 >m < 16 p.m
T e m p °C 5 cm * 5 cm 10 cm 30 cm 30 cm
l0 5 h 12' 1 h 18' 39' 0" 1' h 57' 29' 15"
ll 5 3 116 37 55 154 28 27
12 4 55 1 14 36 53 1 51 27 41
13 4 47 1 12 35 54 1 48 26 56
14 4 40 1 10 34 58 1 45 26 13
15 4325 1 8 34 4 142 25 32
16 425.5 1 65 33 12 139 24 53
17 4 19 1 5 32 21 137 24 15
18 4 12 1 3 31 32 134 23 39
19 4 6 1 15 3045 132 23 4
20 4 0 1 0 30 0 130 _.30
9")

21 3 54 58.5 29 16 1 28 21 57
22 3 40 57 28 31 1 25 21 24
23 3 44 55 5 27 46 20 51
24 3 39 54 27 01 20 18

h = hours
' = minutes
" = seconds
* Column from which particles larger than indmated have settled down m a specified time and temperature
M A N U A L FOR G E O C H E M I C A L ANALYSES 249

least, the amount of material < 63 /xm and (b) The proportion of metals dissolved is
the total trace concentrations be determined variable and depends on the sample type,
for each sample. matrix, and element.
(c) Accuracy of the results can not be
5.1 Sediment decomposttton methods determined since no reference materials are
certified for strong acid digestions.
In order to determine the major and trace
(d) Metal data obtained from strong acid
metal concentrations of marine sediments by
digestions are not intercomparable with total
wet chemical methods, it is necessary to dis-
metal data and are subJect to operationally
solve all or part of the sample. Sample diges-
defined bias.
tion methods commonly used are: (a) Total
decomposition; (b) strong acid digestion, or 5.1.3 Moderate or weak acid extracttons
(c) moderate or weak acid extractions. Moderate or weak acid extractions such as
hydrochloric acid (HC1) and acetic acid
5.1.1 Total decompositton methods
(HOAc) are often used for chemical parti-
Total decomposition methods use hydro-
tion studies (Loring, 1978, 1981). They are
fluoric acid (HF) in combinations with con-
strictly operationally defined. The acetic acid
centrated oxidizing acids such as aqua regia.
(HOAc 25% v/v) extraction procedure is
Alternatively, alkaline fusion followed by acid
described in Section 8 (marine sediments)
dissolution of the flux can be used.
and Section 12 (suspended particulate mat-
Hydrofluoric acid decomposition has the
ter).
following advantages:
(a) HF is the only acid that completely 60 TOTAL (HF) DECOMPOSITION OF SEDI-
dissolves the silicate lattices and releases all MENTS
the associated metals such as A1, Fe and Li
used for the gram size normalization of the In this method (Rantala and Loring, 1989;
data. Loring and Rantala, 1990), hydrofluoric acid
(b) Accuracy can be assessed by analyzing and aqua regia are used to release the total
reference materials certified for the total metal content from marine sediments into
metal content. solution in a sealed teflon decomposition
(c) Intercomparable data, free from opera- vessel referred to as a teflon bomb. The
tionally defined bias, can be obtained. teflon bomb decomposition procedure is
Some laboratories have been reluctant to shown m detail in a video available from
use HF due to its corrosive nature. In our International Counsel for the Exploration of
experience, after using it for 30 years, HF the Sea (ICES), Copenhagen, DK. It is also
poses no greater danger than other strong the method recommended for the ICES in-
acids when normal laboratory safety rules for tercalibration exercise on trace metals in sed-
handling acids are observed. iments (Loring, 1987).
The main advantages of the teflon bomb
5.1.2 Strong acid (non-HF) digestion decomposition are:
Strong acid digestions using nitric acid (a) Rapid decomposition.
(HNO 3) or aqua regia (HNO 3 + HCI) are (b) Reduced risk of contamination.
commonly used to decompose marine sedi- (c) Small volume of acid required.
ments. They are NOT recommended for the (d) No loss of volatile elements.
following reasons:
(a) Strong acid digestions without HF re- 6.1 Apparatus and reagents
sult in incomplete digestions because sili-
cates and other refractory oxides are not Teflon decomposition vessels (bombs), ap-
dissolved. paratus to heat the bombs, a mechanical
 250 D H LORING AND R T q RA N T A L A

shaker, and assorted plastic ware are the min. The temperature rise from room tem-
main items of equipment required to carry perature was found to be 20.9°C for the
out the decomposition procedure. microwave oven (Panasonic Model NE-7970C
with 700W full capacity) used in our labora-
O.1.1 Teflon bomb construction tory.
The construction of the bomb must be In microwave heating, some of the energy
such that no metallic contamination will oc- may be reflected back to the magnetron. As
cur; such risk is always present when steel the magnetron heats the output decreases.
clad bombs are used. LORRAN all teflon To maintain reproducible conditions, the
TFE decomposition vessels (Lorran Interna- magnetron should be allowed to cool be-
tional, Porters Lake, N.S., Canada, BOJ 2S0) tween the sample heating periods.
of 20 ml capacity are used by the authors. The heating time depends on the power
These bombs can be submerged in boiling output of the microwave oven and the load.
water and also heated in a microwave oven. The optimum heating time can be estimated
The bombs incorporate a pressure relief (Kingston and Jassie, 1986) and verified ex-
mechanism for safety. perimentally by analyzing certified reference
materials and comparing the results to those
6.1.2 Teflon bomb heating equtpment obtained by a 1 h decomposition in boiling
The teflon bomb m which the sediment water.
sample ~s decomposed is heated either in Teflon bombs without pressure relief
boiling water, or, in a more recent develop- should not be used for microwave heating
ment, in a microwave oven to provide the due to the potential danger of over-pressuri-
necessary heat and pressure for the total zation.
decomposition of the sample.
(a) For heating the bomb by submersion in 6.1.3 Laboratory equipment
boiling water: Mechanical shaker
Hotplate. Analytical balance
Covered ceramic dishes (25 cm × 25 cm × 10 Agate mortar and pestle
cm deep). Each dish can accommodate 8
6.1.4 Labware
bombs at a time.
Polypropylene volumetric flasks, 100 ml
Sufficient water to cover the teflon bombs
Polypropylene narrow mouth bottles
is placed into the ceramic dishes and heated
Polystyrene disposable weighing boats
to boiling on the hotplate.
Polypropylene graduated cylinders
(b) For microwave heating:
Polypropylene funnel
M~crowave pressure cooker to contain any
No glassware should be used in the pres-
possible leakage of acid fumes. Each cooker
ence of HF as contamination might be re-
wdl accommodate 4 bombs at a time.
leased from the glass.
Flat based microwave safe dish to compen-
All labware should be thoroughly cleaned
sate for the concave bottom of the cooker.
by soaking in dilute nitric acid and rinsing
Beaker, 100 ml.
with de-ionized water.
Microwave oven with turntable.
Although the corrosive nature of HF is
The microwave oven should be calibrated
diminished when complexed with boric acid
to detect any decrease in the power output in
(H3BO3), the above precautions should be
the future and to facilitate the calculation of
observed.
the optimum heating time for samples. The
calibration is carried out by measuring the 6.1.5 Reagents
temperature rise of 1 1 of distilled water in a Hydrofluoric acid [HF] (49%)
beaker with the oven set at full power for 2 Nitric acid [HNO~] (70%)
M A N U A L FOR G E O C H E M I C A L ANAL'I'SES 251

Hydrochloric acid [HCI] (37%) sure cooker; place the cooker along with a
Aqua Regia [HNOj-HCI] (1:3 v/v) beaker containing 50 ml water in the mi-
Boric acid crystals [H3BO 3] crowave oven; heat for 70 s at full power (700
Oe-~onized reverse osmosis water w);
All reagents must meet ACS analytical (f) remove the bomb from the heat source
reagent grade requirements. and cool it to room temperature in cold
water or an ~ce bath:
6.1.6 Sample size (g) weigh 5.6 g of H j B O 3 and transfer into
The size of sample required is dependent a 100 ml polypropylene volumetric flask;
mainly on the expected metal concentrations. (h) add 20 ml of H 2 0 and shake briefly;
In general a 0.1 g sample is used for the (i) remove the bomb from the cooling wa-
major elements and a 1 g sample is used for ter and dry it;
trace metal determinations if only flame AAS (j) open the bomb, (be sure to wipe off any
is available (Section 7.5). A 200 mg sample water found on the outside of the sealing
containing normal amounts of major and area) and transfer the contents into the 100
trace metals, however, has been found to be ml polypropylene flask;
sufficient for most metal concentrations if (k) rinse the bomb several times with
the sample is homogeneous and graphite fur- deionized water and add the rinsings to the
nace AAS ~s available. It would be prefer- flask;
able to make only one decomposition for all (I) shake the flask to complete the &ssolu-
elements using 0.1-0.3 g sample size and tion (black carbon residue may remain but
adjusting the dilutions (Section 7.5) accord- does not contain significant amounts of met-
ingly. Certain trace metals occurring in very als and does not interfere with subsequent
low concentrations may require larger sam- metal determinations);
ple sizes. (m) make the solution up to 100 ml with
Maximum sample size used in the mi- H20;
crowave decomposition is 200 mg. Larger (n) transfer the solution into a polypropy-
samples may be feasible although small sam- lene bottle for storage;
ples are best suited for microwave heating. (o) allow solutions obtained from 100-500
To obtain more concentrated solutions, the mg sample size to settle overmght; those
final volume could be reduced from 100 ml from 500-1000 mg sample size should settle
to 5(/ml. In that case, only 3 ml of HF and for several days m case borosllicate forms.
2.8 g of H3BO ~ should be used in the de- This is because the metals cannot be deter-
composition. mined m concentrated solutions ( > 500 mg
sample/100 ml) until the gelatinous precipi-
6.2 Total decomposition procedure tate of borosilicates has settled, leaving a
clear surface layer that can be analyzed. This
(a) Accurately weigh 100-1000 mg (100- process may take 7-14 days. When a smaller
200 mg for microwave heating) of finely sample is used ( < 50(1 mg sample/100 ml),
ground sample; such precipitation will not occur and the
(b) transfer to a teflon bomb; sample may be analyzed after the black car-
(c) add 1 ml of aqua regia (HNO 3:HC1, bon residue has settled overnight;
1:3 v/v); (p) analyze the solutions for trace metals
(d) add 6 ml of HF very slowly to avoid by flame or graphite furnace AAS.
excessive frothing;
(e) close the bomb tightly and submerge in 6.2.1 Storage of sample solutions
boiling water for a minimum of 1 h: OR Store sample solutions in precleaned
place the bombs (4) in the microwave pres- polypropylene bottles. Stored samples are
252 1) H [ . O R I N ( , A N D R I F R A N T A L A

extremely stable and ~t has been possible to deuterium background corrector, HGA-500
determine Cd, for example, accurately in a graphite furnace, AS-1 autosampler, R100A
clear surface layer after several years of stor- recorder, pyrolytically coated and uncoated
age. graphite tubes.
(c) Perkin-Elmer hollow cathode lamps,
7 (I ATOMIC ABSORPTION ANALYSES OF SEDI- except Cd, Pb and Ti which are Perkin-Elmer
MENTS E D L lamps.

Precise and accurate atomic absorption 7.2 Calibratton standards

analyses for the determination of metals in-
volves the use of flame (FAAS) and graphite Single-element standards can be used, but
furnace atomic absorption spectroscopy for convenience, a combined stock solution
(GFAAS). can be prepared for some metals. Single -ele-
FAAS is the method of choice of most ment standards are prepared for Si, Ti, Ba,
metal determinations when the metals are Be, Cr, Sr and Rb. Multi- element standards
present in sufficient concentrations to allow are prepared for A1, Ca, Fe, K, Mg, Na, Cd,
precise and accurate determinations. Co, Cu, Li, Ni, Pb and Zn. Single or multiple
GFAAS ~s generally used for the determi- element stock solutions are prepared from
nations beyond the sensitivity of FAAS. Due 1000 /xg/ml standards. The 1000 g g / m l
to improved accuracy and precision, GFAAS standards are either prepared in the labora-
should be used for all Cd, Cr and N1 deter- tory or obtained commercially.
minations. Pb should be determined by FAAS To compensate for matrix effects between
when concentrations are adequately high; samples and standards, a decomposition blank
otherwise GFAAS should be used. is required for preparing the working stan-
dards and should contain 1 ml of aqua regia,
7 1 Apparatus and equtpment 6 ml concentrated HF and 5.6 g H~BO3/100
ml.
Atomic absorption spectrophotometer. In addition, KC1 is added m solution or as
Hollow cathode lamps and E D L lamps. a solid to some of the working standards and
Pyrolytically coated and uncoated graphite sample solutnons to suppress lOnlzat,on dur-
tubes. ing flame absorption.
Digital diluter. Details for preparing stock solutnons are
Polystyrene disposable culture tubes. usually found ,n the instructions provided by
Eppendorf digital pipettor 100-1000/xl. the manufacturer of the atomic absorption
Various pipettes equipment.
Volumetric flasks, polypropylene and glass.
7.2.1. Al, Ca, Fe, K, Mg and Na multt-ele-
A variety of atomic absorption spec- ment standards
trophotometers and equipment is available. Prepare a combined stock standard solu-
In our laboratory the following spectropho- tion of 3 0 0 / x g / m l A1 and 100 ~ g / m l Ca, Fe,
tometers are in use: K, Mg and Na from the 1000 /xg/ml stock
(a) Perkin Elmer Model 4000 atomic ab- solutions by:
sorption spectrophotometer equipped with (a) Pipettmg 30 ml of the A1 stock solu-
d e u t e r i u m / t u n g s t e n background corrector, tion, and 10 ml each of the Ca, Fe, K, Mg
corrosion resistant nebulizer, 3 slot and N,O and Na stock solutions into a 100 ml volu-
burner heads, PRS-10 printer sequencer. metrnc flask;
(b) Perkin-Elmer Model 306 atomic ab- (b) diluting the combined solution to 100
sorption spectrophotometer equipped with a ml with de-ionized water;
M A N U A L I~OR G E O C H E M I C A L ANALYSES 253

further dilute the combined stock solution Make up the working standards in the
25, 50, 100, 200 and 400 times to prepare the decomposition blank. Add solid KCl to each
working standards. All working standards standard so the that the final concentration
should contain 1500 p~g/ml K which is added is 1500 p~g/ml K.
as a solid KC1, except the standards for K
which should contain 1500 p~g/ml Na as 7.2.5 Ba smgle element standard
NaCl. Dilute the 1000 # g / m l stock solution to
Working standards obtained for Ca, Fe, K, obtain working standards of 0.1-2.0 p~g/ml
Mg, and Na have the following concentra- Ba;
tions of metals ( ~ g / m l ) : 0.25, 0.50, 1.00, 2.00 Make up the working standards in the
and 4.00 /.Lg/ml. For A1 the concentrations decomposition blank. Add solid KCI to each
are 0.75, 1.50, 3.00, 6.00 and 12.00 t~g/ml. standard so the that the final concentration
The decomposition blank contains 1 ml of is 1500 p~g/ml K.
aqua regla, 6 ml HF and 5.6 g H 3 B O J 1 0 0
m]. 7.2.6 Be single element standard *
Dilute the 1000 p~g/ml stock solution to
obtain working standards of 5-30/.Lg/I Be in
7.2.2 Cd, Co, Cu, Li, Mn, Nt, Pb and Zn
the decomposition blank.
multi-element standards
Prepare a combined stock standard solu- • In preparing Be standards, it is advis-
able to use commercially available stock so-
tion containing 50 # g / m l Cd, Co, Cu, Li,
Mn, Ni, Pb and Zn from the individual 1000 lutions because beryllium dust is potentially
p~g/ml stock solutions by pipetting 5 ml of toxic.
each and diluting to 100 ml with de-ionized
7.2. 7 Cr single element standard
water.
Prepare working standards of 0.1-2.0 Dilute the 1000 / z g / m l stock solution to
p~g/ml for these elements, except Cd, from obtain working standards of 0.1-1.5 ~ g / m l
the combined stock solution (50 /~g/ml) by Cr in the decomposition blank.
dilution using the decomposition blank. Further dilute the working standards with
For Cd, prepare working standards of 0.5- H~O for GFAAS determinations.
10 ~ g / l Cd in the decomposition blank.
7.2.8 Rb single element standard
Dilute the 1000 p~g/ml stock solution to
7.2.3 Sl smgle element standard obtain working standards of 0.2-3.0 /~g/mi
A 1000 /~g/ml Si standard is prepared Rb in the decomposition blank containing
from a pure silica rod by the same decompo- sufficient K added to match the sample K
sition procedure as described for the samples concentrations.
(Section 6.2). The reason for careful matching is that the
Dilute the 1000 # g / m l stock solution to absorption of Rb ~s suppressed by the de-
obtain working standards of 2 0 0 - 4 0 0 / ~ g / m l composition blank solution and enhanced by
Si in the decomposition blank solution. K. The enhancement by K is progressive to
about 600 ~ g / m l K, above which suppres-
7.2.4 Tl smgle element standard sion by K occurs. Thus the addition of exces-
A 1000 p~g/ml Ti standard is prepared sive amount of K results in a vastly sup-
from TiO 2 by the same decomposition proce- pressed absorption signal and is not recom-
dure as described for the samples (Section mended. Enhancement by Na is, however,
6.2). small. For example, when 200 ~ g / m l of Na
Dilute the 1000 # g / m l Ti standard to and K are added to a 1 /.~g/ml Rb standard
obtain working standards of 2-10 ~ g / m l Ti. in the decomposition blank, the enhance-
254 DH LORINGAND RT1 RANTALA

ment is only about 2% more than in 200 into the standard and sample set to assess
p~g/ml K. Since the determination of Rb is the accuracy of the results.
highly dependent on K concentration, it is
first necessary to determine the K content of 7.4 Reference materials
the samples and match it to the standards for
the most accurate determinations. Since the Accuracy of an analytical determination
absorption of Rb is linear to about 3 ~ g / m l can be established in one of two ways (Veil-
Rb for a particular K concentration, several Ion, 1986). The first would be to analyze the
standard curves with varying K concentra- same sample by two or more independent
tions can be drawn by using just one stan- methods. This is not always feasible. The
dard in each case for fast but less accurate second would be to analyze reference mate-
Rb determination. The Rb concentration can rials. These are materials whose analyzed
then be read from the standard curve having content has been established by two or more
the K concentration closest to that of the independent methods and whose matrix is as
sample. When large numbers of samples are nearly identical to that of the samples as
analyzed, therefore, it may not be practical possible.
to match the concentration of K in each The reference materials are certified for
standard, but the Rb can be interpolated total metal contents only. Therefore, they
from other standard curves. can only be used as an accuracy check of the
metal concentrations determined in a solu-
7.2.9 Sr smgle element standard tion obtained from H F dissolution of the
Dilute the 1000 t~g/ml stock solution to sample.
obtain working standards of 0.1-1.0 /.~g/ml The samples should not be analyzed until
Sr in the decomposition blank as for Ba. the value obtained from the CRM meets the
accuracy requirements for the determination.
7.2.10 V smgle element standard
Generally, sediment CRM's are used al-
Dilute the 1000 /~g/ml stock solution to
though rock CRM's are also suitable.
obtain working standards of 0.3-5.0 ~ g / m l
For normal metal levels, the N R C C estu-
V in the decomposition blank to which 100
arine sediments BCSS-1, MESS-l, BEST-1
p~g/ml AI and 500/.Lg/ml K have been added
(Hg only) and USGS marine mud MAG-1
to enhance the V absorption. No further
are suitable.
enhancement of V has been observed above
For high metal levels, the N R C C harbor
100 ~ g / m l A1 and 500 ~ g / m l K.
sediment PACS-1 and NIST river sediment
7.3 Alternattce cahbration standards (Pre- SRM 2704 are available.
ferred) USGS rock standards are also useful as
they g~ve a wide range of metal concentra-
An alternative means of calibration is to tions.
use certified reference materials (CRM's) as For a compilation of working values for
calibration standards because the concentra- 272 geostandards, the reader is referred to a
tion of each metal has been established within recent publication by Govlndaraju (1989).
a certain degree of confidence. A minimum NRCC CRM's are available from: Marine
of 3 sediment or rock CRM's should be Analytical Chemistry Standards Division,
chosen for calibration to ensure that the National Research Council, Montreal Road,
calibration curve generated from the analy- Ottawa, Canada K I A 0R6. NIST CRM's are
ses of these materials is correct. The CRM's available from: Standard Reference Material
are prepared and decomposed and the met- Program, Building 202, Room 204, National
als determined in the same way as the un- Institute of Standards and Technology,
known samples. A fourth CRM is introduced Gaithersburg, MD 20899 USA. USGS refer-
M A N U A L FOR G E O C H E M I C A L ANALYSES
255

TABLE 2
Sample dilution scheme for flame atomic absorpnon analysis

Sample wt. DiluUon Element

grams
01 1000 × Mn, $1
1000 × m 1500/~g/ml K Ba, Sr, Tl
20000 × m 1500/zg/ml K AI, Ca, Fe, Mg, Na
20000 × in 1500/xg/ml Na K
10 100 × Be, Cd, Co, Cr, Cu
Li, N1, Pb, Rb, V, Zn

e n c e m a t e r i a l s are available f r o m J.S. K a n e , e l e m e n t s in s a m p l e solutions p r o v i d i n g the

U.S. G e o l o g i c a l Survey, R e s t o n , V A 22092 c o n c e n t r a t i o n s are a d j u s t e d so that t h e y fall
USA. within the linear p o r t i o n o f the c a l i b r a t i o n
curve a n d are c o n f i r m e d by the use o f refer-
7.5 Sample dtlutions e n c e solutions d e r i v e d f r o m the dissolution
of C R M ' s . F o r m o s t m a j o r e l e m e n t s this
F l a m e A A S m a y be u s e d to d e t e r m i n e r e q u i r e s a f u r t h e r dilution o f the s a m p l e
precisely a n d a c c u r a t e l y the m a j o r a n d t r a c e solution ( T a b l e 2). A KCI solution is o f t e n

TABLE 3
Standard conditions for flame atomic absorption analysis

Perkm-Elmer 4000
Element Wavelength Slit Background Oxidant Reducing
(nm) settmg correction flame
(nm)
AI 309 3 0.7 OFF N20 USE
Ca 422.7 0.7 OFF N20
Fe 248 3 0.2 OFF AIR
K 766.5 2.0 OFF AIR
Mg 285.2 0.7 OFF N20
Mn 279 5 0.2 OFF AIR
Na 589 0 0.7 OFF AIR
Si 251.6 0.2 OFF N20 USE
T1 365.3 0.2 OFF N~O USE
Ba 553 6 0.2 OFF N20 USE
Be 234 9 07 OFF N20 USE
Co 240.7 0.2 ON AIR
Cr 357.9 0.7 OFF N20 USE
Cu 324.8 0.7 OFF AIR
Ll 670.8 0.7 OFF AIR
NI 232.0 0.2 ON AIR
Pb 283.3 0.7 ON AIR
Rb 780.0 2.0 OFF AIR
Sr 460.7 0.7 OFF N20 USE
V 318.3 07 OFF N20 USE
Zn 213.9 0.7 ON AIR
 25(3 D H LORING AND R T T RA N T A L A

used as a diluent not only to adjust the 7. 7 Interferences

concentrations to the optimum range but also
to suppress ionization of certain metals (Ta-
Matrix interferences for major and trace
ble 2). Such dilutions are made rapidly by
using an adjustable diluter and disposable elements are reduced by diluting the sample
polystyrene culture tubes. solutions in such a way as to adjust the
concentrations to the linear portion of the
Table 2 shows that the 0.1 gram sample
absorbance curve.
solution from the teflon bomb decomposition
is initially diluted to 100 ml (1000 × dilution). Ionization is controlled by additions of
1500 /xg/ml K or Na to standards and sam-
ples (Section 7.5; Table 2). For example,
7.5.1 Solutions for major and some minor ionization in Ba and Sr determinations is
element determinations controlled by adding 1500 /xg/ml K to sam-
(a) Mn and Si are determined directly ples and standards.
without any other dilutions. Chemical suppression of elements such as
(b) Solid KC1 is added to separate aliquots Si and A1 on Ca and Mg absorption is avoided
to give a K concentration of 1500/xg/ml and by the use of nitrous oxide-acetylene flame.
these are analyzed for Ba, Sr and Ti. This eliminates the need for lathanum which
(c) For the determinations of A1, Ca, Fe, is essential in an air-acetylene flame. No
Mg and Na; a further 20 x dilution is made interferences were observed for Mn and Si in
to give the desired dilution factor of 2 × 10 4 a fluoboric acid matrix.
in the KCI ( 1 5 0 0 / x g / m l K) matrix. A similar Standards for Rb and V require special
dilution is also made for the determination attention as discussed under standard prepa-
of K by substituting NaC1 as the diluent. ration (Sections 7.2.8 and 7.2.10).
Molecular absorption is corrected by the
Deuterium Background Corrector in Co, Ni
7.5.2 Solutzons for trace element determina-
and Pb determinations.
tions
GFAAS should be used for Ni if available.
For the trace element determinations, the
In our experience, large positive or negative
separate 1 g sample decomposed in the
errors can result from the use or non-use of
Teflon bomb is also initially diluted to 100 ml
the deuterium background corrector in the
(100 x dilution).
FAAS determinations of Ni.
(a) From this solution, Be, Cd, Cr, Cu, Li,
No interferences have been observed for
Ni, Pb, V and Zn are determined directly.
Be, Cu, Li, Mn, Si and Zn in a fluoboric-
As indicated in Section 6.2, 1 g/100 ml
boric acid mixture.
decomposition requires several days for the
separation of a clear layer for the analysis.
7.8 Graphite furnace determinattons
7.6 Standard operational conditions/settings
for flame atomzc absorption analyses
GFAAS determinations require the use of
L'vov platforms, preferably automatic sample
The standard operational conditions and introduction, optimization of instrumental
instrumental settings for the different ele- parameters, background corrections for Cd,
ments using the Perkin-Elmer model 4000 Ni and Pb, and correct application of appro-
for atomic absorption analysis is summarized priate reference materials.
in Table 3. For other instruments, readers In our laboratory, Cd, Cr, Ni and low
should consult their own manuals for the levels of A1 and Pb are determined by
correct instrumental settings. GFAAS.
M A N U , M . F O R G E O C H E M I ( AL A N A L Y S E S 257

7.8.1 Calibration standards 7.8.4 Sample cups

Calibration standards are diluted from Watch for any formation of air bubbles at
stock standard solutions to the appropriate the bottom of the sample cups and eliminate
range in the same reagent matrix as the them either by emptying the cup and refilling
sample. it, or by tapping the cup to drive out the
bubbles.
7.8.2 Use of L 'L'ot' platforms
7.8.50ptimtzatton of the furnace program
A L'vov platform placed in a graphite tube
e.g. Cd
permits the atomization of the sample under
Optimizing the furnace program requires
nearly isothermal conditions. This results in
the use of a blank, a standard and a refer-
the reduction of interferences and permits
ence material solution.
direct comparison with aqueous standards
The following is an example of the proce-
(Sturgeon et al., 1982). L'vov platforms for
dure for optimizing the furnace program for
Cd and Pb should be made of uncoated
Cd determinations.
graphite. In earlier work, without the plat-
The Cd concentration of the standard so-
form, Rantala and Loring (1980) found that
lution should match that of the reference
pyrolytically coated tubes were unsatisfactory
material. Following the thermal conditioning
for Cd as they produced constantly changing
of the tube and the platform as indicated by
sensitivity and performed well only after >
a low blank reading, the standard and the
100 firings. Findings with platforms made in
reference material solutions are run in se-
the laboratory from the grooved end of the
quence until the furnace program is opti-
tube (Hinderberger et al., 1981) confirm that
mized (Rantala and Loring, 1987).
uncoated platforms perform the best. Un-
The furnace program has been optimized
coated platforms are generally not available
when maximum peak heights are obtained
commercially, but are prepared in the labo-
for the standard and the reference material
ratory. Platforms can be made by cutting the
solutions, and the peak height of the refer-
two ends of a graphite tube into eight (4
ence material solution is within _+ 10% of the
from each end) 7 mm x 5 mm grooved,
standard at the 0 . 0 0 2 / x g / m l Cd level (at the
curved sections as described by Hinderberger
linear range).
et al (1981). For A1, Cr and Ni, the platforms
To assist in optimizing, the following
are made from coated tubes. They can be
guidelines should be followed:
inserted into either coated or uncoated tubes.
(a) Drying: The platform requires a rela-
The platform must be centered under the
tively high drying temperature of approxi-
sample introduction hole and visually in-
mately 270°C. Ramptime must be sufficiently
spected to ensure that no sample will run off
slow to avoid spattering of the sample.
the platform.
(b) Charring: Charring temperature, ramp
and hold times must be optimized to avoid
7.8.3 Automatic sample introduction loss of metals during this cycle.
Automatic sample introduction of 20/xl is (c) Atomization: The fastest possible
recommended as it will provide better preci- ramptime to atomization temperature should
sion than manual pipetting. The tip of the be used. The atomization temperature is de-
sampler must be adjusted in such a manner termined experimentally. The gas interrupt is
that it will enter far enough into the sample usually required.
introduction hole of the graphite tube so that Instruments having a maximum power
no sample solution can creep around the mode feature such as the Perkin Elmer
hole. At the same time the analyst must take HGA-500 require temperature control cali-
care that the tip does not touch the platform. bration. During calibration, be sure to press
258 D H LORING AND R T T RANTALA

TABLE 4
GFAAS parameters

Metal Platform Background Dry Char Atomize Burn out Coohng

type corrector (°C) (°C) (°C) (°C) (°C)
Al pyrolytic OFF 270 1500 2400 2700 20
Cd non-pyrolytic ON 270 400 1600 27/)0 20
Cr pyrolytic OFF 270 900 2700 27(10 20
N~ pyrolytic ON 270 1000 2700 27(10 20
Pb non-pyrolytic ON 270 550 1800 2700 20
Cu OFF 120 900 2100 27(10 -
Ramp(s) 10 20 max power 1 1
Hold(s) 20 20 3 3 15

Tube type: Pyrolytic

Sample volume: 20 ~ L
Argon flow during atomization: 50 m l / m m
Gas interrupt mode is used for low concentrations of Cd and Pb.

the manual temperature button long enough, to suppression of Cr absorption by matrix

while adjusting the calibration control. With elements unless the solutions are diluted
the platform, the time required to adjust the 10 × with water.
calibration control at 1600°C could be as Interferences by matrix elements are re-
long as 30-40 s. duced significantly by using the L'vov plat-
The GFAAS parameters for the HGA-500 form.
graphite furnace are shown in Table 4. The method of standard additions is un-
necessary as direct calibration with standards
7. 9 Interferences can be used. We do not recommend the
method of additions because it is time con-
Sample solutions for A1 and Cr determina- suming and can produce imprecise and inac-
tions are diluted with water prior to GFAAS curate results (Tyson, 1984; Slavin, 1988)
determinations. This is important for Cr as,
in our experience, Cr recoveries using 7.10 Combmed analytical scheme
GFAAS can be reduced as much as 60% due
A combined analytical scheme utilizing
TABLE 5 both FAAS and GFAAS with appropriate
Combined analytzcal scheme for FAAS and GFAAS dilutions is shown in Table 5 for sediments
determinations containing natural levels of heavy metals.
Element Further dilution AAS Mode Initial solution: 0.2 g/100 ml.
m H20 *
AI 500 x GFAAS 7.11 RelatiL,e accuracy and prec&ion
Cd NIL GFAAS
Cr 10 × GFAAS Tables 6 and 7 show the relative accuracy
NJ NIL GFAAS of the major and trace element data for the
Pb NIL GFAAS
Cu NIL FAAS
marine sediments: MAG-1, BCSS-1, MESS-l,
Fe 25 × FAAS PACS-1 and rock samples: BIR-1, DNC-1
L1 NIL FAAS and W-2 by comparison of our values with
Mn NIL FAAS certified or consensus values reported. Preci-
Zn NIL FAAS sion is expressed as a standard deviation of
* Imtml dilution = 0.2 g/100 ml the mean value.
M A N U A l F O R GEOCHIc-M1CAL A N A L Y S E S 259

7.12 Comparison of microwave and conven- Heavy metals, along with other elements,
tional (boiling water) teflon bomb digestion are introduced into estuarine environments
in solution and as part of or in association
with, solid particles from natural and anthro-
For this experiment, three sediment sam-
pogenic sources, mainly via rivers, but some-
ples containing natural levels of metals were
times through outfalls and dumping. Very
digested (HF + aqua regia) in triplicate using
little is known, however, about the carriers
0.2 g samples by both 70 second microwave
and transport modes of these particulate
and 1 hour conventional heating methods.
metals or their response to estuarine condi-
Table 8 compares the results obtained us-
ing the combined analytical scheme (Table 5) tions.
Measurement of total sedtment or particu-
by the two methods.
late metal concentrations is a poor method
The data show that the microwave mean
of determining such characteristics because
concentrations and precision compare well
part of the metal load is loosely bound to the
with those obtained for the conventional di-
particles and part is locked up physically
gestion.
a n d / o r chemically in detrital particles and
minerals. Selective chemical methods using
80 CHEMICAL PARTITION OF SEDIMENTS: acetic acid (Loring, 1978), weak HCI solu-
ACETIC ACID (HOAc) EXTRACTION OF SEDI- tions (Luoma and Jenne, 1976) and even
MENTS sequential extractions (Tessier et al., 1979)
have been developed and used to partition
Chemical partition of sediments is used to the total metal concentrations into their
deduce the source and pathways by which loosely bound (non-detrital) and residual
the major and trace metals have entered the (detrital) phases. It should be noted that all
marine environment. these extraction methods are strictly opera-

TABLE 6
Major and m i n o r constituents of s t a n d a r d reference materials

Oxide PACS- 1 W-2 BIR- 1 DNC-1

sd % sd % sd °/c sd
AlcOa 12 19 ° 0.12 14 83 ° 0 28 15 19 0 19 17 94 0 24
12.23 c 0.22 15.35 c~ 0.23 15.35 0 51 18 30 0.49
CaO 2 84 0.01 10.98 0.06 13 35 0.19 11 40 0 20
2 92 0.13 10.87 0.29 13 24 0 29 11 27 0 22
Fe20 ~ 6 98 0.04 10 60 0.09 11 14 0 12 9 84 0.08
6 96 0.12 10.74 0.23 11.26 0.23 9.93 0 14
K~O 1.51 0.02 0.60 0 01 0.012 0.004 0.20 0.00
1.50 0.09 0.63 0.02 0 027 0.010 0 23 0 02
MgO 2.48 0.02 6 24 0.05 9 44 0.13 9 96 0 19
2 41 0.09 6 37 0 18 9.68 0 28 10 05 0.33
Na20 4 45 0.03 2 18 0.01 1.77 0 04 1 87 0.03
440 0 11 2.14 0 12 1.75 0 11 1 87 009
S10, 55 27 0 19 52.47 0 50 47.94 0 45 47 16 0 35
55 70 0.50 52.44 0.64 47 77 0 51 47 04 0 64
TiO, - 1 06 0.01 0 94 0 03 0 48 0 01
1.06 0 03 0.96 0.03 0.48 0.02

° = O u r value (W-2, BIR-1 a n d DNC-1 from Flanagan, 1984)

= CerUfied values for PACS-1 ( N R C C )
c, = C o n s e n s u s m e a n values for W-2, BIR-1 a n d DNC-1 ( G l a d n e y and Roelandts, 1988a).
260 DH L O R I N G AND R F T RA N T A L A

tionally defined. Sequential extractions are weakly held in organic matter. It leaves the
time consuming and of limited value because silicate lattices intact and does not attack the
they fail to provide actual particulate specia- resistant iron and manganese minerals or
tion (Nirel and Morel, 1990). organic compounds
The acetic acid method was chosen be- The proportion of the total metal concen-
cause it is one of the weakest chemical treat- tration removed by the extraction is opera-
ments that can be used to remove effectively tionally defined as the non-detrital (acid sol-
the weakly bound part of the total metal uble) metal fraction of the sediment.
concentrations in sediments (Loring, 1978; The proportion of a metal remaining in
Loring and Rantala, 1988) and particulate the residual fraction is operationally defined
matter (Loring et al., 1983, 1985; Loring and as the detrital (acid insoluble) fraction of the
Asmund, 1989). Acetic acid (25% v/v) re- material.
moves metals held in ion exchange positions, Such fractionation allows some deductions
easily soluble amorphous compounds of iron as to the carriers, transport mode and poten-
and manganese, carbonates and those metals tial bio-availability of metals entering and
TABLE 7

Trace element concentrations in standard reference matermls

Element MAG-I MESS-1 BCSS-1 PACS-1

mg/kg sd mg/kg sd mg/kg sd mg/kg sd

Ba 480 ° 28 - - -
4 7 9 c~ 41 - - -
Be 36 0 1 - - -
3.2 O4 - - -
Cd 0.19 0.01 0.59 ° 0.05 0.25 0.01 2.25 0 10
0.20 0.03 0 59 c 0 10 0 25 0 04 2 38 0.20
Co 19 3 1.9 - - 17.7 0 5
204 1 6 - - 175 1.1
Cr 112 3 72 1 - 106 1
97 8 71 11 - 113 8
Cu 31 4 27.0 0.6 19.5 0 8 441 6
30 3 25.1 3.8 18.5 2.7 452 16
Hg - 0.179 0.006 0.153 0,009 -
- 0.171 0 014 0.129 0.012 -
L~ 78 6 0.5 44 9 0.06 46.5 0.5 33 6 0 5
79 4 - - -
Nl - - - 44 5 1 4
- - - 44.1 2.0
Pb 28 5 31.1 1.1 23.7 2 6 362 6
24 3 34.0 6 1 22.7 3.4 404 20
Rb 152 2 - - -
149 6 - - -
Sr 133 3 - - -1
146 15 - - -
V 145 4 - - 130 1
140 6 - - 127 5
Zn 133 3 180 2 112 1 839 21
130 6 191 17 119 12 824 22

° = Our value ( M A G - 1 from Rantala and Loring, 1 9 7 8 , 1 9 8 0 )

cs = Consensus mean values for M A G - 1 (Gladney and Roelandts, 1 9 8 8 b )
c = Certified values for M E S S - l , B C S S - 1 , a n d P A C S - 1 ( N R C C ) . Mercury values were decertified for MESS-1 and
BCSS-1 in 1 9 9 0 b y N R C C .
MANU~.L FOR G E O C H E M I C A L ANALYSES 261

within these different systems (Loring, 1981; 8.2 Extraction procedure

Loring et al., 1985)
(a) Place a portion of dry sample in an
agate mortar. Do not grind it. Simply crush
8.1 Apparatus and reagents
the lumps;
(b) weigh 2 grams of sample and transfer it
8.1.1 Laboratory equipment into a propylene centrifuge tube;
Centrifuge (c) weigh the tube + the sample and record
Mechanical shaker the weight;
Analytical balance (d) add 25 ml of 25% v / v HOAc;
Agate mortar and pestle (e) cap the tube and shake slowly in a
mechanical shaker for 6 h;
8.1.2 Labware (f) balance the centrifuge tubes and cen-
Polypropylene centrifuge tubes, 30 ml with trifuge for 10 min at 2500 RPM or until the
caps supernatant is clear;
Polypropylene narrow mouth bottles (g) pour the supernatant HOAc into a 50
Glass volumetric flasks, 50 ml ml volumetric flask;
Funnel (h) wash the sediment with 10 ml of water
and shake the tube briefly on the shaker;
8.1.3 Reagents (i) separate the wash water by centrifuging
Glacial acetic acid [HOAc], 25% v / v (ACS and add it to the flask;
analytical grade) (j) rinse the funnel and make up the HOAc
De-ionized reverse osmosis water solution to a volume of 50 ml;

TABLE 8
C o m p a r i s o n of metal d e t e r m i n a t i o n s from conventional (boding water) and microwave digestions

Element Sample A Sample B Sample C

sd sd sd
AI ( % ) 7.10 c 0 14 7.16 0.19 7.53 0.09
7.10 m 0 11 7.19 0.08 7.58 0.20
Fe ( % ) 3.69 0.01 3.36 0.01 3.84 0.01
3.68 0.01 3.29 0.01 3.79 0.07
Cd ( m g / k g ) 0.11 0.01 0.10 0.02 0.27 0.00
0 11 0.01 0 09 0.00 0 27 0.01
Cr 84 0.6 72 1.7 89 2.3
82 2 1 68 10 85 1.5
Cu 27 0.6 24 0.6 36 0.6
27 1.0 23 1.0 37 1.2
Li bl 00 49 06 60 0.6
61 06 49 06 60 0.6
Mn 446 38 478 06 415 10
442 12 463 1.5 414 6.7
Pb 26 0.6 32 2.1 40 0.6
25 0.0 28 0.6 39 0.0
Zn 92 15 85 2.0 127 10
92 12 83 12 126 2.0

c = C o n v e n t i o n a l m e t h o d : 1 h b o m b digestion ( H F + aqua regla) m b o l h n g w a t e r (n = 3)

m = Microwave m e t h o d 70 s b o m b digestion ( H F + aqua regla) m microwave oven (n = 3)
262 D H LORING AND R T T RANTALA

(k) dry the tube containing the residue in solution with a final concentration of 1500
the oven at 105°C and place in the dessicator ~ g / m l K.
to cool; (c) Solution for K determinations is di-
(1) when the tube reaches room tempera- luted 10 x with the NaCI solution with a
ture, weigh it and calculate the percent by final concentration of 1500/zg/ml Na.
weight contribution of the residual fraction;
(m) transfer the dry residue to the mortar
and use a portion of the ground up sample 8.3. 5 Graphtte furnace determinations
for the HF teflon bomb decomposition to (a) Solutions for A1 determinations are
determine the trace metals held in the acid diluted with HzO 10 × or more to bring the
insoluble or detrital fraction of the sedi- concentrations to the linear portions of the
ments; A1 calibration curve.
(n) by flame or graphite furnace A_AS; (b) Cd, Cu, Ni and Pb are generally deter-
analyze the acetic acid soluble and residual mined directly from solution.
solutions for trace metals. (c) Uncoated L'vov platforms are used for
Cd and Pb determinations.
8.3 Atomic absorptton analyses (d) Coated L'vov platforms are used for A1
and Ni determinations.
(e) No platform is used for Cu which is
FAAS is used to determine: Ca, Fe, Mg,
determined using a pyrolytically coated tube.
Mn, Na, K, Cu, Li and Zn.
GFAAS is used to determine: A1, Cd, Ni,
Pb and Cu when it is too low to determine by
8.3.6 L 'roy platform condittoning for Cd
FASS.
It was found that the platform had to be
conditioned for the Cd determination in the
8.3.1 Apparatus and equtpment acetic acid matrix to ensure reproducible re-
Same as for sediments. sults. Reproducible results can be obtained
by running a conditioning solution made up
8.3.2 Calibration standards of a HOAc sediment leachate (2 g / 5 0 ml) at
(a) Calibration standards are diluted from the beginning of the analyses and between
commercial 1000/zg/ml stock solutions. each sample and standard.
(b) Final working standards are prepared
in the same matrix as the sample to be ana-
lyzed, i.e. HOAc, KC1, NaCI or H20. 8.3. 7 Furnace control program steps
The furnace controller programs used are
shown in Table 4. In addition, two other
8.3.3 Reference materials steps are used.
No reference materials are certified for
(a) The residual matrix is removed by a 3 s
weak or strong acid solutions.
firing at 2700°C and the platform is cooled
for 15 s at 20°C before the next sample is
8.3.4 Flame AAS determinations introduced into the furnace.
(a) Mn, Cu, Li and Zn are determined (b) It is also necessary to lower the atom-
directly from solution, but further dilutions ization temperature for Cd from 1600°C in
are made if required to bring the metal con- the H3BO 3 matrix to ll00°C in the acetic
centrations to the linear portions of the cali- acid matrix.
bration curves. (c) Internal argon flow is reduced to 50
(b) Solutions for Fe, Ca, Mg and Na deter- m l / m i n during atomization. Gas interrupt
minations are diluted 10 x with the KCI mode is used for low Cd concentrations.
MANUAL FOR GEOCHEMICAL ANALYSES 263

8.4 Calculattons of non-detrital (acid soluble) Absorption Cell. Constructed from borosili-
and detntal (acid msoluble) contributions cate glass tubing, 25 mm o . d . x 15 cm with
quartz windows (25 mm x 2 mm thickness).
The total metal concentration is the sum Gas inlet and outlet ports attached 2 cm
of the acetic acid soluble and residual contri- from each end.
butions. Variac, 0 to 120 V for controlling pump
For example: Zn. speed
Let: Pump with peristaltic pumping head
Sample wt. for H O A c leach = 2.000 g Water bath
Wt. of H O A c residue = 1.800 g Aerator: Perkin-Elmer
Wt. of residue for decomposition = 0.300 g BOD bottles
Vol. of flask for H O A c solution = 50 ml Tygon tubing
Vol. of flask for decomposition solution = 100 Digital pipetter 100-1000/xl
ml Sulfuric acid desiccator: Made by snipping a
Concentration of Zn in H O A c solution = 0.20 diffuser from a Perkin-Elmer aerator so that
b~g/ml it extends about halfway into a standard BOD
Concentration of Zn in decomposition solu- bottle. The bottle is filled within 1 cm of the
tion = 0.25 /xg/ml aerator tube with concentrated sulfuric acid
Non-detrital contrib, of Zn = 50/2.000 x (Kothandaraman and Dallmeyer, 1976).
0.20 = 5 / x g / g
Detrital contrib, of Zn = 1.800/2.000 x 9.1.2 Mercury system setup
100/0.3 × 0.25 = 7 5 / x g / g The mercury system ~s assembled by at-
Therefore: taching tygon tubing as follows: a e r a t o r - -
Total Zn concentration = 5 /~g/g + 7 5 / x g / g sulfuric acid desiccator--absorption c e l l - -
= 80/xg/g p u m p h e a d - - a e r a t o r . W h e n the aerator is at-
tached to the reaction vessel (BOD bottle) a
9 0 FLAMELESS ATOMIC ABSORPTION ANALY- closed system is obtained. After removing
SIS FOR MERCURY (Hg) the aerator from the reaction vessel, mercury
remaining in the system is vented out by
The method used is a modified form of means of an exhaust fan.
that used by Hatch and Ott (1968). It consists
of oxidizing the sediment sample (10-500 mg 9.1.3 Reagents
dry weight) by digesting it with a mixture of Nitric acid [HNO 3] Concentrated (low in Hg)
concentrated nitric acid ( H N O 3) and sulfuric Sulphuric acid [H2SO4] Concentrated (low in
acid (H2SO 4) at 60°C for 1 h and reducing Hg)
Hg to its elemental state with stannous chlo- Stannous chloride [SnCI 2] 10% w / v in 6 N
ride (SnCl2). The mercury vapor is then HC1
passed through a quartz absorption cell of an
AAS where its concentration is measured 9.2 Sample preparatton
and recorded by means of a strip chart
recorder. (a) Oven dry the sediment at 60°C;
(b) Crush lumps and grind lightly to homoge-
9.1 Apparatus and reagents nize the sample;
(c) Weigh accurately 100 mg of sample (less
9.1.1 Apparatus if the sample might contain > 1 mg Hg/kg).
Atomic Absorption Spectrometer Samples larger than 100 mg can be used, but
Mercury Hollow Cathode Lamp may result m excessive frothing during aerat-
R e c o r d e r 10 mv full deflection ing.
264 D H LORING AND R TT RANTALA

(d) Transfer weighed sample to BOD bottle; Prior to analyses, find the aeration rate
(e) Add 10 ml of concentrated HNO 3 fol- that gives the best absorption by turning the
lowed by 20 ml of concentrated. H 2 S O 4. Do variac while aerating 1 ml of the 0.1 /xg/ml
not stopper. Hg standard. This setting doe~ not need to
(f) Place the BOD bottle in the water bath be adjusted again although a new portion of
and digest for 1 h at 60°C; tubing needs to be placed in the pump head
(g) Remove from the water bath and add when lower absorption readings are observed
water up to the top of the etched label on for the standards.
the BOD bottle, insert the stopper, and place (a) Pipette 10 ml of stannous chloride so-
in freezer until cold. lution with a fast delivering pipette into the
Note cold sample solution in the BOD bottle.
Before adding acid to any sample, the blank (b) Attach the aerator Immediately and
solution must be analyzed for Hg to assure observe the sample peak height on the
that the reagent blank is not excessively high. recorder.
Another blank should be run with the sam- (c) When it reaches a plateau, detach the
ples. aerator and rinse it with water.
(d) After the peak falls back to baseline,
9.3 Preparation of standards continue with the next sample.
(e) Standards should be run at even inter-
9.3.1 Hg: 1000 txg / m l stock solution vals between the samples.
(a) Dissolve 0.1354 g of mercury (II) chlo-
ride into 100 ml of 1 N sulfuric acid. 9.6 Cahbration curt,e
(b) From this solution prepare a 10 ixg/ml
solution in 5% HNO 3 and 0.01% K2Cr20 v The calibration curve is constructed by
(Feldman, 1974). The 10 g g / m l solution plotting peak height, versus /,Lg Hg (stan-
prepared in this way is stable for several dards). The amount of Hg in the unknown
months. samples is read from,that curve. The ~g Hg
found from the curve is converted i n t o / x g / g
Hg in the sediment by dividing by the sample
9.4 Working standards
weight in grams.

Daily, dilute the 1 0 / z g / m l Hg solution to 9. 7 Relatzt,e accuracy and precision

obtain a 0.1 p~g/ml solution in 5% HNO 3
and 0.01% K2Cr20 7. Replicate analyses (12) of a sample with
Prepare the working standards for each an average of 0.180 ~zgHg/g gave a standard
batch of samples by dispensing 0.10-1.00 ml deviation of +0.023 txgHg/g (coefficient of
of the 0.1 /xg/ml Hg solution from a digital variation of 12.8%). At a level of 0.410
pipetter into BOD bottles. Add the same /xgHg/g, the standard deviation was found
volume of reagents as to the samples. Diges- to be _+0.030 txgHg/g (n = 6), coefficient of
tion of the standard is not normally neces- variation of 7.3%. At a level of 9.5 ~ g H g / g ,
sary although it may be safer to put it through the standard deviation of replicate analyses
the same procedure as the samples. (n : 12) was +0.95 /xgHg/g (coefficient of
variation of 10%).
9.5 Determination o f mercury Accuracy of the analyses was established
by the use of certified reference materials
Adjust the atomic absorption settings: (NRCC Marine sediments BCSS-1 and
Wavelength 253.7 nm; slit 0.7 nm; back- MESS-l). Certified values for BCSS-1 and
ground corrector On. MESS-1 are 0.129+0.012 txg/g Hg and
MANUAl_ FOR G E O C H E M I C A L ANALYSES 265

0.171 _+ 0.014 /~g/g Hg respectively. Values Standard 1 N K 2 C r 2 0 7 solution (dissolve

obtained by this method (n = 6) are 0.153 _+ 49.04 g of K2Cr20 v in water; dilute to 1 1)
0.009/~g/g Hg for BCSS-1 and 0.179 +_ 0.006 0.5 N Ferrous solution (dissolve 196.1 g of
for MESS-1. However, NRCC has recently Fe(NH4)2SO 4 6H20 in 800 ml of water con-
decertified BCSS-1 and MESS-1 for mercury taining 20 ml of concentrated H2SO4; dilute
because of suspected contamination. In their to 1 1)
place, NRCC has certified a new estuarine Diphenylamine Indicator (dissolve approx-
sediment BEST-1 for mercury with a value of imately 0.5 g of reagent grade diphenylamine
0.092 +_ 0.009 ~ g / g H g . For BEST-l, we ob- m 20 ml of water and 100 ml of concentrated
tained a value of 0.095 _+ 0.006 p.g/gHg. H2SO 4)

10.0 DETERMINATION OF READILY OXIDIZ- 10.2 Determination of organtc carbon

ABLE ORGANIC MATTER
(a) Place 0.5 g of dried and sieved (200
Organic carbon matter is determined to ~zm' sieve) sediment sample in a 500 ml Er-
assess the role played by the orgamc fraction lenmeyer flask;
of sediments In the transport, deposition and (b) add exactly 10 ml of 1 N K2Cr20 7
retention of trace metals. solution by burette and 20 ml of concen-
The readily oxidizable organic carbon con- trated H 2 S O 4 with Ag2SO 4 and mix by gen-
tent is determined by the Walkey-Black tly rotating the flask for about 1 min;
method (1947), adopted and modified from Note This should be done carefully, to
Jackson (1958, pp. 219-221). This method insure complete mixing of the reagents with
differentiates humus matter from extraneous the sediment, while avoiding splashing the
sources of organic carbon such as graphite sediments onto the sides of the flask out of
and coal. Gaudette et al. (1974) found that contact with the reagents.
this method provided excellent agreement (c) allow the mixture to stand for 30 rain;
with the LECO combustion method of or- (d) a standardization blank without sedi-
ganic carbon analysis. ment should be run with each new batch of
The Walkey-Black m e t h o d utilizes samples;
exothermic heating and oxidation with potas- (e) after 30 rain, add 200 ml distilled wa-
sium dichromate and concentrated H2SO 4 of ter; 10 ml of 85% H3PO 4 and 0.2 g NaF;
the sample, followed by the titration of ex- (f) add 15 drops (0.5 ml) of the dipheny-
cess dichromate with 0.5 N ferrous ammo- lamine indicator to the sample flask;
nium sulfate solution to a sharp, 1 drop, end (g) back titrate the solution with the 0.5 N
point. Oxidation of C1- can be prevented by ferrous ammonium sulfate solution to a one-
the use of A g z S O 4 in the digestion mixture. drop end point (brilliant green).
Note: The color of solution will progress
10.1 Apparatus and reagents from an opaque green-brown, to green upon
the addition of approximately 10 ml of the
lO. 1.1 Apparatus ferrous solution. The color will continue to
2-50 ml burette with 0.1 ml graduations shift upon titration to a bluish-black-grey; at
Magnetic stirrer
this point the addition of 10-20 drops of
51)0 ml Erlenmeyer flasks ferrous solution will shift the color to a bril-
liant green giving one-drop end point
lO. 1.2 Reagents
85% H3PO 4 10.3 Calculation of results
Solid NaF
Concentrated ,H2SO 4 with A g 2 S O 4 (dis- % organic matter (readily oxidizable)=
solve 2.5 g Ag2SO 4 in 1 1 of H2SO 4) 10(1 T / S ) x F
-
266 DH LORIN(-JAND R F1 RANTALA

S = standardization blank titration, ml of 11 0 DETERMINATION OF CALCIUM CARBON-

ferrous solution ATE
T = sample titration, ml of ferrous solu-
tion Carbonate is often an important compo-
F = factor derived as follows: nent of marine sediments and has been found
F = (1.0 N ) × 1 2 / 4 0 0 0 × 1.72 × to be an important indicator of provenance
1 0 0 / s a m p l e weight and dispersal of terrigenous material in the
= 1.03 when sample weight is exactly 0.5 g Gulf of St Lawrence (Loring and Nota, 1973)
Where 1 2 / 4 0 0 0 = meq wt. carbon and For a simple determination of calcium car-
1.72 = factor for organic matter from carbon. bonate (CaCO3), the sediment is placed in a
Note: In the back titration, the ferrous pre-weighed stoppered flask and treated with
solution reduces the dichromate that has not HCI. By adding excess HC1 to C a C O 3, a
been used up in the oxidation processes. certain volume of CO 2 is evolved, while an
Therefore, if it takes less than 4 ml of the equal volume of air is expelled. The loss of
ferrous solution to reach the end point, then weight due to the escape of air expelled by
more than 8 ml of the available 10 ml of the evolved CO 2 is determined. As both CO 2
dichromate has been consumed in the oxida- and air follow the B o y l e - G a y Lussac Law,
tion. If this is the case, then it is necessary to and weights are determined instead of vol-
repeat the determination using less sedi- umes, the temperature and pressure have no
ment. influence on the analysis, assuming that these
quantities do not change between weighings.
10.4 Standardlzatton of organtc carbon deter-
mmations 11.1 Apparatus and reagents
Dextrose (C~H~20~6) is used as the stan-
11.1.1 Apparatus
dard. It should contain 39.99% carbon.
Conical flask, 250 ml
(a) Weigh out exactly 0.01 g of dextrose
R u b b e r stopper fitted with a glass tube
and treat in the same manner as the sedi-
filled with CaC12 to remove water vapour
ment sample.
and plugged at both ends with cotton. An-
(b) The carbon in dextrose is calculated as
other similar tube is attached to the first
follows:
tube by means of plastic tubing. This tube is
% C = 10(1 - T/S) x F
used to ensure that outside moisture does
F = ( 1.0 N ) × 12/4000 X 1 0 0 / s a m p l e weight
not enter the main tube.
= 30 for 0.01 g of dextrose.
Glass vials: 7 ml
The theoretical value is 39.99% C in one
Desiccator
gram of dextrose.
Analytical balance
Following standardization with dextrose it
is r e c o m m e n d e d that a well homogenized
11.1.2 Reagents
sample be analyzed and be used for stan-
HC1 4N: dilute 330 ml of concentrated
dardizanon instead of dextrose.
HCI to 1 I with water.
10.5 Prectston of determinattons CaCO3: reagent grade, oven dried and
stored in the desiccator.
Replicate analyses (10) of a sample with
an average of 0.36% organic matter gave a 11.2 Determmation of carbonate
standard deviation of 0.04% (coefficient of
variation of 11%). At a level of 3.17% or- (a) Transfer 1 g of oven dry ground sedi-
ganic matter the standard deviation was ment into a 250 ml conical flask (see A in
0.05%, (coefficient of variation of 1.6%). Fig. 3);
MANUAl FOR GEO(_HEMICAL ANALYSES
267

11.3 Precision of carbonate determinations

Replicate analyses (6) of a sample with an

average of 6.84% C a C O 3 gave a standard
deviation of _+0.18% (coefficient of variation
of 2.6%).

12.0 SUSPENDED PARTICULATE MATTER (SPM)

Samples of suspended particulate matter

are collected to determine their concentra-
tion and distribution in the water column as
well as their inorganic and organic composi-
tion.
Suspended particulate matter is commonly
defined in relation to a minimum diameter,
d e p e n d m g on the pore size of the filter being
used. Since the build up of particles modifies
Fig 3 Apparatus used for the determination of CaCO~
A = Flask, B = Rubber stopper fitted with glass tube the effective pore size of the filter, the mate-
filled with CaCI:; C = Second glass tube filled with rlal retained on the filter includes additional
CaCI~ attached to tube B by means of plastic tubing; particles smaller than the original pore size.
D = Vial filled with 4N HCI In practice, suspended particulate matter is
defined operationally as that which is re-
(b) Place vial D filled with 5 ml of 4N HC1 tained by the type of filter being used.
into the flask; Suspended particulate matter in this re-
(c) Close the flask with a rubber stopper port is defined as that retained by Nuclepore
containing tube B filled with CaC12; 0.4 tzm polycarbonate filter membranes.
(d) D e t e r m i n e the total weight; Nuclepore filters are widely used for gravi-
(e) Mount tube C also filled with CaCI~ metric and chemical determinations. They
on top of tube B with the aid of a piece of have well defined pore sizes that give them a
plastic tubing; relatively precise cutoff in the size of parti-
(f) Tilt the flask so that the HCI in vial D cles that they retain. The polycarbonate com-
is poured out onto the sediment sample; position of the filters is relatively metal free
(g) During the next 2 h, swirl the contents and hydrophobic which makes them easy to
of the flask occasionally so that the acid is in tare and reweigh after sample collection. The
contact with the sediment; major disadvantage of the Nuclepore filters
(h) After 2 h, remove tube C and weigh; is that they clog at relatively low loadings
(1) Calculate loss of weight in grams (P); which limits the amount of material that can
(j) R e p e a t the procedure using 100 mg of be collected.
CaCO3 instead of the sediment and calculate The collection of suspended particulate
the loss of weight in grams (Q). matter involves procedures for the prepara-
tion of the filters, collection of SPM samples,
11.21 Calculation of calcium carbonate washing of the filter to remove salt, and
con ten t reweighlng. The following subsections con-
~, C a C O 3 = P / Q × 0 . 1 0 0 / weight of dry tain a summary of these procedures.
sediment × 100 For detailed instructions, the reader
Note: Tubes filled with CaCI~ should be should consult the report by Yeats and Brug-
stored in a desiccator and the CaC12 re- mann (1990) on the collection of SPM for
placed if it is observed to be hard. gravlmetric and trace metal analysis.
268 D H LORING AND R T T RANTALA

12.1 Filter preparatton are thoroughly rinsed with deionized water

between each use before the preweighed fil-
All preparation should be done in as dust ter is reloaded into the filter holder.
free an environment as possible because the
filters tend to build up a static charge and 12.2 Collection of SPM samples
thus attract particles from the air. A clean
room is the best environment for handling The basic collection technique consists of
the filters, but a laminar flow clean bench is pressure filtration of the sea water through
sufficient. It is essential to use an anti-static the Nuclepore filters in clean plastic filter
source to remove any static charge immedi- holders.
ately before weighing because of the static Depending on the SPM levels, sample col-
charge build-up. lection procedures will either be direct on
For trace metal analyses, the filters should line filtration from the Go-Flo, Niskin bottle
be soaked for 24 h in 2 N H N O 3 (Merck, or other sampler, or an off-line filtration of
pro analsi) and then rinsed several times in an initial sample of unfiltered sea-water sam-
Milli-Q water. The filters are stored in poly- ple in a separate filtration apparatus.
carbonate petri dishes that have been rinsed,
then soaked in 6 N HCI and rinsed again 12. 2.1 Off-Line filtration
and then dried in a microwave oven. When For highly turbid waters, rapid settling of
the filters have returned to room tempera- particles in the water sampler results in an
ture they can be transferred with plastic underestimation of the true SPM concentra-
tweezers to a microbalance capable of weigh- tions if sampled directly from the water sam-
ing an ~ 20 mg filter to +0.01 rag. After pler. To compensate for this problem, about
weighing, the filters should be stored flat in 2 1 or less of unfiltered water is withdrawn
individual n u m b e r e d precleaned petri dishes from the sampler and filtered separately. To
because they will have to be loaded onto flat ensure a representative subsample, the water
filter holders. sampler is agitated by inverting it and return-
ing it to its upright position to ensure a
12.1.1 Filter holders uniform distribution of SPM before subsam-
Polycarbonate filter holders for 47 mm piing. After the subsample is withdrawn, it is
diameter Nuclepore filters with plastic fit- then gently agitated again to ensure that the
tings and air vents such as the Millipore SPM remains in suspension before pouring it
Swinnex, Nuclepore or Sartorius model into the filtration apparatus. The time
SM16541B are all suitable. Since the filter elapsed between sampling and filtration is
holders are reused, they must be thoroughly kept to a minimum ( < 24 h).
cleaned in the laboratory between each The off-line filtration apparatus used at
oceanographic cruise. An ultrasonic bath BIO consists of a 500 ml polypropylene sepa-
containing laboratory detergent may be used ratory funnel that has been adapted for pres-
to remove particles that have become at- sure filtration by welding a fitting for a pres-
tached to the filter holders. The filter hold- sure line into the lid. A preloaded Sartorious
ers are then soaked for 24 h in 6 N HCI (model SM16541B) filter holder is attached
after removal of the silicon O-rings. The to the bottom of the funnel with a short
silicon O-rings are soaked for 1 h in 1 N piece of silicone tubing. An aliquot of the
HC1. After rinsing several times in deionized unfiltered water is poured into the funnel,
water, the filter holders are then air dried in the top is screwed on and the pressure ap-
a clean air bench and loaded with the plied. An air filter is placed in the nitrogen
preweighed filters using plastic tweezers. Af- gas line to prevent the introduction of partic-
ter their initial use at sea, the filter holders ulate contaminanon through the line. The
M A N U A l FOR G E O C H E M I C A L ANALYSES 269

sample is filtered until the filter clogs or all 12.4 Total HF decomposition of suspended
of the water sample has been passed through particulate matter
the filter.
For water samples containing much less The Nuclepore filters are decomposed in
that 0.5 mg/1 of SPM, an on-line filtration teflon bombs in a manner similar to sedi-
procedure is used because insufficient sam- ments, but with smaller amounts of reagents.
ple is collected by the off-line filtration pro- The filters themselves are not decomposed
cedure unless large quantities of water are (Rantala and Loring, 1977, 1985). The num-
processed. ber of steps that are followed will depend on
the nature and extent of the investigation. It
is recommended that, at least, the total trace
12.2.2 On-Line ftltration metal concentrations except chromium (Cr)
For the lower SPM concentration, an on- be determined for each sample. Chromium
line direct filtration from the Go-Flow or cannot be determined accurately in the SPM
Niskin bottle is used. In this procedure, the because of the high content of this element
filter in its holder is attached to the modified in Nuclepore filters (Yeats and Dalziel, 1987).
teflon drain spigot of the Go-Flo sampler
with a piece of clean silicon tubing and nitro- 12.5 Apparatus and reagents
gen pressure (maximum ~ 0.8 atm) is ap-
plied by attaching the pressure line to an
For the suspended particulate matter de-
adapter designed to hold the nitrogen line.
composition, the same teflon bombs and
The sample is filtered until the filter clogs or
heating equipment are used as described in
all of the water sample has been passed
Sections 6.1.1 and 6.1.2.
through the filter. To inhibit particle settling,
the water bottles are occasionally inverted
during sampling. 12.5.1 Labora tory Equipment
Mechanical shaker
Microbalance (readability 0.01 mg)
12.3 Filter washing
12.5.2 Labware
While still in their holders, each filter is Polypropylene volumetric flasks, 25 ml.
then washed by attaching the holder to the Polypropylene narrow mouth bottles
off-line separatory funnel as described in Polypropylene funnel
section 12.2.1 and rinsed with two 50 ml Plastic tweezers
aliquots of filtered (0.2 /~m) Milli-Q water
while applying nitrogen pressure. The washed 12.5.3 Reagents
filters are dried for 12 h at 60 °C in an oven Hydrofluoric acid [HF] (46%)
or m a microwave oven, cooled in a desicca- Nitric Acid [HNO3] (71%)
tor, and then reweighed to obtain the weight Hydrochloric acid [HC1] (37%)
of the SPM retained on the filter. After Aqua Regia: [HNO3-HCI ] (1:3 v/v)
weighing, the filters are stored flat in individ- Boric acid crystals
ual numbered precleaned petri dishes until High purity water (Milli-Q or De-ionized
required for chemical analyses. The SPM reverse osmosis water)
concentration in m g / l in the water column is Ultra-high purity grade acids equivalent to
calculated from the weight of the SPM re- the J.T. Baker Ultrex grade must be used for
tained on the filter and the volume of water the SPM decomposition because the reagent
filtered. grade acids contain large amounts of Fe and
270 D H I.ORING A N D R 1 1" R A N T A L A

Sl. Boric acid crystals, however, can be of (b) add 0.5 g of boric acid and 1 ml of
ACS analytical reagent grade. water into the bottle;
(c) transfer the contents of the bomb (ex-
12.5.4 Sample stze cept the filter) into the bottle; Rinse well and
The amount of SPM sample required for add to the bottle;
precise and accurate determinations of the (d) add H 2 0 until the weight of the solu-
elements will depend on their concentration tion equals 10.6 g (10 ml);
in the SPM. A minimum of 1 mg (dry weight) (e) shake the bottle to complete the disso-
~s usually required to determine most of the lution;
metals at their natural background levels. (f) analyze the solutions for trace metals
by flame or graphite furnace AAS.
12.6 Total decomposttlon procedure
12. 7 Atomtc absorptton analyses of SPM
(a) Using plastic tweezers, transfer a filter
into a Teflon bomb and squash it into the FAAS is used to determine: Si, Ca, Fe,
bottom of the vessel; Mg, Mn, Na, K, Cu, Li and Zn.
(b) add 1 ml of aqua regia, 1 ml of H F and G F A A S is used to determine: AI, Cd, Ni,
close the bomb tightly; Pb and also Cu when it is too low in concen-
(c) submerge the bomb in boiling water for tration to determine by FAAS.
1 h, or heat for 40 s in a microwave oven
(For further details see Section 6.2); 12. 7.1 Apparatus
(d) after cooling, decant the contents
through a polypropylene funnel into a 25 ml The same as for sediments.
polypropylene volumetric flask containing
0.93 g of boric acid and approximately 5 ml 12. 7.2 Calibration standards"
of water; (a) Calibration standards are diluted from
(e) wash the filter remaining in the Teflon commercml 1000 p~g/ml stock solutions.
bomb several times with small volumes of (b) Final working standards are prepared
water, each time collecting the washings in in the same matrix as the sample to be ana-
the 25 ml flask (some broken pieces of the lyzed, i.e. decomposition blank, ultra high
filter may enter the flask, but will settle and purity KC1, NaC1 or H , O .
do not interfere with subsequent metal de- (c) Alternatively, certified reference mate-
terminations); rials (CRM's) may be used for calibration. A
(f) finally, shake the volumetric flask to minimum of three sediment or rock CRM's
complete the dissolution and make up to 25 should be chosen to cover the expected con-
ml with de-~onized water; centration range. For such a calibration, 10
(g) prepare the blank solution in the same mg of each CRM is decomposed in the same
manner, omitting the sample; m a n n e r as the SPM samples and a fourth
(h) store the solutions in polypropylene CRM Is used as an accuracy check.
bottles:
(i) analyze the solutions for trace metals 12. 7.3 Reference materials
by flame or graphite furnace AAS. Certified reference materials (CRM's) are
used to confirm the accuracy when calibra-
12.6.1 A preferred alternatu,e method for tion standards are prepared from 1000
low concentrations of metals /xg/ml stock solutions. To prepare the CRM
After step (c) above standards: 10 mg of each CRM is decom-
(a) weigh a small precleaned polypropy- posed in the same m a n n e r as the SPM sam-
lene bottle; ples. Metal concentrations in the C R M solu-
MANUA.L F O R G E O C H E M I C A L A N A L Y S E S 271

tions should be in the concentration range of Pb.

the SPM solutions.
12.8 Acetic acid extracnon of SPM
12. 7.4 Flame A A S determinattons
(a) Si, Mn, Cu, Li, and Zn are determined The chemical technique to partition the
particulate metals into their non-detrital and
directly from solution.
(b) Solutions for Fe, Ca, Mg, and Na de- detrital phases involves the leaching of the
terminations are diluted 10 × with the KC1 particulate matter on Nuclepore filters with
solution with a final concentration of 1500 25% v / v acetic acid in a modified Millipore
/zg/ml K. Sterifil vacuum filtration apparatus (Rantala
(c) Solution for K determinations is di- and Loring, 1985). The residue remaining on
luted 10 × with the NaC1 solution with a the filters is decomposed with a combination
final concentration of 1500 ~ g / m l Na. of ultra-high purity hydrofluoric acid and
(d) Further dilutions may be required to aqua regia m the teflon bombs.
bring the metal concentrations into line with 12. 9 Apparatus and reagents
the linear portions of the calibration curves.
(e) Si, Ca, and Mg are determined in the 12. 9.1 Extraction apparatus
N20-acetylene flame. Polycarbonate Millipore Sterifil vacuum
(f) The other metals are determined in the filtration funnel and receiver flask attached
air-acetylene flame. to a polypropylene Millipore Swinnex-47 mm
filter holder base * (Fig. 4).
12. 7.5 Graphite furnace determmations
Silicone tube containing a short glass rod
(a) Solutions for A1 determinations are
which is attached to the filter outlet.
diluted with H 2 0 10 × or more to bring the
Vacuum pump
concentrations into line with the linear por-
• It is necessary to replace the red silicone
tion of the calibration curve.
O-ring with a teflon O-ring because the sili-
(b) Cd, Cu, Ni and Pb are generally deter-
cone O-ring has been found to contaminate
mined directly from solution.
the filtrate with Zn. In addition, some filter
(c) Uncoated L'vov platforms are used for
supports have been found to contain metallic
Cd and Pb determinations because they pro-
particles that release Zn during leaching. To
vide better sensitivity than the coated plat-
prevent contamination from this source, the
forms.
supports should be examined for such parti-
(d) Coated L'vov platforms are used for AI
cles with a microscope and be replaced if
and Ni determinations.
particles are found to be present.
(e) No platform is used for Cu because
greater sensitivity is provided by using a py- 12. 9.2 Labware
rolytically coated tube. Glass volumetric flasks, 10 ml
Polypropylene narrow mouth bottles
12. 7.6 Furnace control program steps
Polypropylene funnel
The furnace controller programs used are
Plastic tweezers
shown in Table 4 (Section 7.8.5). In addition,
two other steps are used. 12. 9.3 Reagents
(a) The residual matrix is removed by a 3 s Glacial acetic acid [HOAc] 25% v/v
firing at 2700°C and the platform should be ultra-h~gh purity
cooled for 15 s at 20°C before the next sam- De-ionized reverse osmosis water
ple is introduced into the furnace.
(b) Internal argon flow is reduced to 50 12. 9. 4 Sample size
m l / m i n during atomization. The gas inter- The amount of SPM sample required for
rupt mode is used for low levels of Cd and precise and accurate determinations of the
272 D H LORING AND R TT RANTALA

(b) screw the filtration funnel tightly on to

the base and insert the glass rod into the
silicone tubing;
(c) attach the filter holder base to the
' ! ~-~FILTRATION FUNNEL
receiver flask;
(d) place 5 ml of 25% v / v acetic acid in
the funnel;
(e) cover the funnel and let it stand for 24
h;
~ ~ TEFLONO-RING (f) after 24 h, remove the glass rod and
quickly insert the silicone tubing into a 10 ml
( ~ ~ = FILTER SUPPORT
volumetric flask placed in the receiver flask

~ ~ FILTER HOLDERBASE
(no loss of acetic acid solution from the
tubing before its insertion into the flask will
occur unless the filter has been pierced or
badly seated on the support);
~ SILICONE TUBING (g) apply vacuum to the receiver flask;
(h) after all the acetic acid has been fil-
~]~ GLASSROD tered, wash the filter twice with 2 ml of
water, each time applying vacuum;
(i) after the final wash, remove the funnel
and transfer the filter to a Petri dish for
drying at 60°C.
(j) under vacuum, suck the rest of the
FLASK
filtrate trapped in the filter holder into the
flask;
.UMETRIC FLASK
(k) make the flask up to a volume of 10 ml
with water, transfer the solution into a
polypropylene bottle, and store for AAS
Fng. 4. Apparatus for acetic acid extraction of SPM analyses.
(From Loring and Rantala, 1990).
(1) treat the Nuclepore filter blanks in the
same manner.
(m) decompose the residual matter on the
elements will depend on their concentration dried filters with HF + aqua regia (Section
in the SPM. A minimum of 3 mg (dry weight) 12.6).
is usually required under this procedure to
partition most of the particulate metals into
their non-detrital and detrital phases at their 12.11 Atomic absorption analyses
natural background levels. Samples smaller
than 3 mg can be used, but it may not be FAAS is used to determine: Ca, Fe, Mg,
possible to determine the detrital (acid insol- Mn, K, Na, Cu, Li and Zn.
uble) metal concentration for some trace GFAAS is used to determine: AI, Cd, Ni,
metals. Pb and Cu when it is too low to determine by
FAAS.
12.10 Extraction procedure

(a) Center the sample filter and lay it flat 12.11.1 Apparatus and equipment
on the moistened filter support; The same as for sediments.
M A N U A L FOR G E O C H E M I C A L ANALYSES 273

12.11.2 Calibration standards (b) Solutions for Fe, Ca, Mg and Na deter-
(a) Calibration standards are diluted from minations are diluted 10 × with the KC1
commercial 1000/xg/ml stock solutions. solution with a final concentration of 1500
(b) Final working standards are prepared /xg/ml K.
in the same matrix as the sample to be ana- (c) Solution for K determinations is di-
lyzed i.e. HOAc, KCI, NaCI or H20. luted 10 × with the NaC1 solution with a
final concentration of 1500 p~g/ml Na.
12.11.3 Reference materials (d) Ca and Mg are determined in the
No reference materials are certified for N20-acetylene flame. The other metals are
weak or strong acid solutions. determined in the air-acetylene flame.

12.12 Flame AAS determinations

12.13 Graphite furnace determinations
(a) Mn, Cu, Li and Zn are determined
directly from solution, but further dilutions (a) Solutions for AI determinations are
are made if required to bring the metal con- diluted with H 2 0 10 × or more to bring the
centrations into line with the linear portions concentrations into line with the linear por-
of the calibration curves. tions of the AI calibration curve.

TABLE 9
Results of m t e r c a l i b r a t l o n exercise(s) for particulate metals

1 ICES i n t e r c a h b r a t i o n exercise for particulate metals in sea w a t e r

Sample Metal (%) sd Metal (mg/kg) sd
Fergusons Cove AI 3.26 o 0 44 Cd 1.8 1.2
2.91 c 0.41 3.5 16
Bedford Basra 2 56 0.25 18 0.3
2.41 0.41 3.5 1.2
Fergusons Cove Fe 2.06 0.21 Cu 55 17
2.55 0.27 64 7
Bedford Basin 2.47 0.21 135 32
2.49 0.23 100 36
F e r g u s o n s Cove Mn 0.162 0.012 Pb 152 39
0.168 0.060 121 37
Bedford Basra 0.808 0 116 288 29
0.780 0.107 220 59
F e r g u s o n s Cove Zn 250 46
197 60
B e d f o r d Basin 525 41
417 104

o = O u r value (n = 4: sample weight = 0.80-1.27 mg S P M on N u c l e p o r e Filters)

c = C a h b r a t l o n exercise m e a n value (Yeats a n d Dalzlel, 1987).

2. G E M S I m t e r c a h b r a t l o n exercise for particulate trace metals m rwer water" c o o r d i n a t o r H. W m d o m

A1 sd Fe sd Mn sd Cu sd N1 sd Pb sd Zn sd
mg/l mg/l /zg/l /zg/l /zg/l /zg/l /zg/l
4.4 o 0.5 1.7 0.2 20 3 2 1 0.1 1.9 0.3 1.3 0.2 5.1 08
3.5 c 1.0 1.9 03 21 6 1.8 0.6 1.1 0.7 1.0 0.6 56 19
o = O u r value (n = 6)
c = C a l i b r a t i o n exercise m e a n value.
274 D H LORING AND R T T RANTALA

(b) Cd, Cu, Ni, and Pb are generally deter- valuable data that otherwise would not be
mined directly from solution. obtained. SPM concentration in the water
(c) Uncoated L'vov platforms are used for column can now be calculated and metal
Cd and Pb determinations. concentrations expressed both on a dry
(d) Coated L'vov platforms are used for A1 weight and on a unit per volume basis.
and Ni determinations. This makes it possible to determine if the
(e) No platform is used for Cu which is metal levels are due to the amount of SPM
determined using a pyrolytically coated tube. in the water column or due to the concentra-
nons in the particles themselves. It also al-
12.14 L 't,ot' platform conditioning for Cd lows comparison with sediment trap and bot-
tom sediment data.
It was found that the platform had to be
condinoned for the Cd determination in the
acetic acid matrix to ensure reproducible re- 12.17 Results of intercalibration exercise (s)
sults. Reproducible results can be obtained for particulate metals
by running a conditioning solution made up
of a HOAc sediment leachate (2 g / 5 0 ml) at The data shown in Table 9 compares the
the beginning of the analyses and between values obtained from the determination of
each sample and standard. various particulate metals in samples used
for intercalibration exercises using the proce-
12.15 Furnace control program steps dures shown above.
The results show that the values obtained
The furnace controller programs used are in this laboratory compare well with those
shown in Table 4. In addition, two other obtained by other workers.
steps are used.
(a) The residual matrix is removed by a 3 s
13.0 GUIDELINES FOR THE NORMALIZATION
firing at 2700°C and the platform should be OF GEOCHEMICAL DATA
cooled for 15 s at 20°C before the next sam-
ple is introduced into the furnace.
13.1 Introduction
(b) It is also necessary to lower the atom-
ization temperature for Cd from 1600°C in
the H3BO 3 matrix to 1100°C in the acetic In many estuaries and coastal regions ad-
acid matrix. jacent to industrial and urban high latitude
(c) Internal argon flow is reduced to 50 areas of eastern North America ( > 42 °) and
m l / m m during atomization. The gas inter- western Europe ( > 52°), sediments are the
rupt mode is used for low levels of Cd and largest repository and potential source of
Pb. metallic contaminants in the marine environ-
ment.
12.16 Data reporting Since metals from natural and anthro-
pogenic sources accumulate together, it is
SPM metal data may be reported on a dry difficult to determine what proportion of the
weight or unit per volume basis or both. sedimentary metal load is natural and what
Unit per volume requires least work since proportion is anthropogenic. This is because
the weighing of the filter is not necessary and of variable anthropogenic inputs and natural
only the volume of the water passing through sedimentary metal loads that can vary by
the filter is measured. This method, however, several orders of magnitude, depending on
also gives the least information. the nature, grain size distribution and prove-
The additional time to weigh the filters nance of metal-rich/metal-poor minerals/
before and after SPM collection results in compounds in the sediments.
M A N U A L FOR G E O C H E M I C A L ANALYSES 275

In most areas, hydraulic and mineralogical statistical analysis of the variance within and
(chemical) particulate fractionation usually between samples.
results in increasing heavy metal concentra- Three levels of sampling can be used to
tions with decreasing sedimentary grain sizes. provide increasing amounts of environmental
It is essential, therefore, to understand, and information.
normalize for, the effects of grain size distri- (a) The first is the simplest and least ex-
butions and provenance on natural metal pensive and involves the use of grab samples
variability, before the effects of anthro- (Section 2.2.1) of the top surface layer to
pogenic metallic inputs can be determined. provide an immediate assessment of the sedi-
This section reviews various granulometric ment characteristics and the presence or ab-
and geochemical approaches that can be used sence of contamination in an area. Relatively
to determine background levels of heavy undisturbed samples of the sediment surface
metals and to normalize for the grain size layer can be obtained if precautions are taken
effects on natural metal variability in differ- in soft muds to avoid contamination of the
ent samples so that anthropogenic metal samples and to ensure that fines have not
contributions may be quantified. Its purpose been lost on sampler recovery. Since deposi-
is to demonstrate how to collect sufficient tion rates cannot be determined from such
data to normalize for the grain size effect samples, this approach does not allow an
and to allow detection, at various levels, of estimate of the time trend, if any, of contam-
anomalous metal concentrations within estu- ination. Rough baseline levels, as will be
arine and coastal sediments. discussed later, however, can be established
from a grab sample.
13.2 Requtrements for geochemical data (b) Level 2 sampling involves box or core
sampling (Sections 2.2.2 and 2.2.3), in which
For any study of sediments, a basic amount representative samples from successive
of information on their physical and chemi- depths are obtained. This allows more reli-
cal characteristics is required before an as- able surface and sub-surface samples to be
sessment can be made on the presence or collected from which time trend information
absence of anomalous metal concentrations. can be obtained as well as information on
The levels at which contamination can be the deposition sequence and the extent of
detected depends on the sampling level and biological mixing of the sediments.
the number of physical and chemical vari- (c) Level 3 sampling is the same as that of
ables that are determined for individual sam- level 2 but with the significant addition of
pies. radiochemical or other measurements to de-
termine the rate of deposition and biological
13.2.1 Field sampling mixing rates. Such a level of sampling is
Briefly, repeated sampling of sediments is required to understand the rate and mecha-
required to establish spatial and temporal nisms of contamination build-up, if any, in
trends in trace metal concentrations of a the sedimentary environment. In most cases,
given area. A gridded sampling pattern is samples of the most recently deposited sedi-
often used when sampling for spatial trends, ments, usually the top 1-2 cm, obtained from
but sometimes a pattern based on submarine grab samplers are sufficient to determine the
morphology is necessary to ensure that rep- present trace metal concentrations. Since
resentative samples of all sediment types are such concentrations may vary with time, short
obtained. The high variability of the sedi- cores ( ~ 10 cm) sampled at 1 cm intervals
ment characteristics requires the considera- will provide sufficient samples to establish
tion of a sufficient number of samples. This trace metal time trends, if any, in fine grained
number can be evaluated by an appropriate sediments.
276 D H LORING A N D R T T RA N T A L A

13.2.2 Chemical analyses For example, the Canadian Ocean Dump-

It is recommended that the whole sample ing Control Act (ODCA) limits dumping ma-
be digested and analyzed. This is because terial containing > 0.6 m g / k g Cd and > 0.75
total metal concentrations determine the true m g / k g Hg in the total sample.
extent of metal levels in the sediments. It is A size fraction of the total sediment may
the criteria by which the extent, if any, of be used for subsequent analysis if required to
contamination is evaluated and on which na- determine the absolute metal concentrations
tional and international dredging and dump- in that fraction, providing that its contribu-
ing regulations are based. tion to the total is kept in perspective when

TABLE 10
Summary of normalization factors

Normal. Factor Size Indicator Role

Textural: (/zm)
Grain Size 2000- < 2 Granular variations Determines physical
of metal bearing minerals/ sorting and depositional
compounds pattern of metals
Sand 2000-63 Coarse grained metal-poor Usually diluent of trace
minerals/compounds metal concentrations
Mud < 63 Silt and clay size metal Usually overall concentrator
bearing minerals/compounds of trace metals *
Clay <2 Metal-rich clay mmerals Usually fine grained
accumulator of trace metals *
Chemical
Si Amount and distribution of Coarse grained diluter of
metal poor quartz trace metal concentrations.
AI A1 sihcates, but used to Chemical tracer of Al-
account for granular slhcates, particularly
variations of metal r~ch the clay minerals *
fine silt + clay s~ze
Al-silicates
Fe Metal-rich silt + clay Chemical tracer for Fe-rlch
size Fe bearing clay clay minerals.
minerals, Fe rich heavy
minerals and hydrous
Fe oxides
Sc Sc structurally combined Tracer of clay minerals which
m clay minerals are concentrators of trace
metals
Cs Cs structurally combined in Tracer of clay minerals which
clay minerals and feldspars are concentrators of trace
metals
LI Ll structurally combined Tracer of clay mmerals,
in clay minerals and micas particularly in sediments
containing Al-sdicates in
all size fractions.
Orgamc Carbon Free grained organic Sometimes accumulator of
matter trace metals like Hg and Cd
* except in sediments derived from glacial erosion of igneous rocks
MANUAL FOR GEOCHEMICAL ANALYSES 277

interpreting the data. Such size fraction data 13.3 Background let,els
might be useful in tracing the regional dis-
persal of metals associated with a specific It is necessary to establish natural back-
size fraction, providing the provenance of the ground metal levels before the extent, if any,
material remains the same. of heavy metal contamination can be esti-
A number of subsamples are required for mated. Such background levels are sub-
analyses (Fig. 1). A 0.1-1 gram subsample is tracted from the total values to yield an
required for the determination of the total estimate of the anthropogenic contribution.
metal concentrations and should be well ho- Background levels can be estimated by:
mogenized. (a) Average metal concentrations of textu-
In order to determine the total metal con- rally-equivalent sediments reported in the lit-
centration, by atomic absorption spectrome- erature.
try (AAS) (Section 7), or inductively coupled (b) Direct measurements of metal concen-
plasma atomic emission spectrometry (ICP), trations in recent texturally and mineralogi-
it is necessary to decompose the sediment cally-equivalent sediments from a known
sample in such a way as to release the metals pristine region.
from minerals and compounds with which (c) Direct measurements of metal concen-
they reside. The use of a reliable and repeat- trations in texturally-equivalent sub-surface
able decomposition technique (Section 5) in- core samples obtained from a depth below
volving HF + aqua regia has a decisive effect any possible contamination or biological mix-
on the amount and quality of analytical data ing.
and is the basic condition for obtaining pre-
cise and accurate data (Loring and Rantala,
1988). 13. 4 Normalization procedures

Simple to more complex approaches can

be used to normalize geochemical data for
13.2.3 Selectton of metals
the grain size effect and identify anomalous
The trace metals to be determined depend
metal concentrations. The approach used de-
on the purpose of the investigation. Usually
pends on the requirements and resources of
Hg and Cd are considered to be the most
the investigation. Table 10 summarizes the
environmentally critical metals, followed by
role of various factors used for normalization
Pb. In addition, the concentrations of Zn and
of trace metal data.
Cu are usually determined although they sel-
dom reach levels in the sediments that are
toxic to organisms. 13.4.1 Simple normalization (Level I)
For normalization purposes, determina- The simple approach to the normalization
tions of at least Li a n d / o r Al are also rec- of geochemical data for grain size effects is
o m m e n d e d followed by Fe and Mg in order to compare the total metal concentrations of
to account for the metal variations in respect surface grab samples with background levels
to the variations of the aluminosilicate min- with those estimated by the means suggested
eral fraction. Determination of the carbonate in Section 13.3. Direct determination of metal
and organic matter contents are also recom- concentrations in texturally and mineralogi-
mended because carbonate may be an impor- cally equivalent sediments provides the best
tant diluent (carrier under certain condi- background levels. The degree of contamina-
tions) and organic matter is sometimes an tion and time trends of contamination at
important concentrator of trace metals, par- each sampling location can be improved upon
ticularly Hg and Cd in the sediments. by making a comparison with metal levels in
278 O H LORING AND R FT RANFALA

texturally equivalent sub-surface core sam- which can be attributed to mineralogical

ples. The texture of the sediments can be variations in the sediments.
roughly classified as to their sandy or muddy It should be emphasized that the validity
nature visually and by feel, or preferably by of the background values and enrichment
separation into their sand and mud size com- factors will depend on the accuracy of the
ponents by wet sieving (Section 4.1). The background term i.e. the direct measure-
< 53 ~ m fraction was used by Loring and ments of background values either in textu-
Nora (1973) and Loring (1978) to mark the rally equivalent sediments or size fractions
separation between sand and mud before the are more accurate than values obtained from
< 63 ~ m fraction was established as the pre- the literature.
ferred fraction. The < 63 p~m size fraction is
now used and recommended. In the Gulf of 13.4.2 Mathematical normalization (Let,el
St. Lawrence (Loring, 1978), for example, li)
muddy silicate sediments containing > 70% It is necessary to reduce or eliminate grain
by weight material < 53 ~ m have average size effects on chemical data from different
background levels of 0.2 mgCd/kg, 0.2 samples in order to deduce the real trends
mgHg/kg, 21 m g P b / k g and 85 mg Z n / k g . rather than those superimposed by grain size
Sandy sediments ( > 70% > 53 txm by weight) effects. In this approach, it is assumed that
have total background levels that average 0.1 the relationship between the metal such as
mg Cd/kg, 0.1 mg Hg/kg, 19 mg Pb/kg, and Zn and the particle sizes or a proxy for
33 mg Z n / k g . If a factor of 3 is introduced particle size in the sediment is linear. The
to account for mineralogical variations in idea being that, should the concentration of
background values, it is therefore likely that the metal be related to changing sediment
comparable muddy sediments in this region particle size, the concentration will change
are significantly contaminated by one or more with a constant relation to grain size or its
of the following metals if Cd is > 0.60 mg/kg, proxy. For linear mathematical normaliza-
Hg > 0.60 mg/kg, Pb > 63 mg/kg, or Zn is tion to be of value, the following require-
> 255 mg/kg. Conversely, total Cd levels of ments should be met:
> 0.3 mg/kg, Hg levels of > 0.3 mg/kg, Pb (a) Significant granular variations occur
levels of > 57 mg/kg, or Zn levels of > 99 between sediment samples.
m g / k g in sandy sediments from the region (b) A strong significant relationship, at
would signify contamination. Such informa- least at the 95% confidence level (the 99%
tion would allow an immediate rough assess- level would be preferred), should exist be-
ment of the surface contamination and its tween the metal content and the relevant
spatial distribution. parameter.
Some estimate of the degree, if any, of (c) The regression line which expresses the
contamination can also be calculated by us- significant relationship should ideally follow
ing the so-called index of geoaccumulation the Equation y = a x instead of the form
(Igeo) proposed by Muller (1979). This index y = a x +b.
is defined as Igeo = log 2 C n / l . 5 Bn where (d) It should be possible to provide accu-
Cn is the measured concentration of the rate and precise analysis of the metal and the
metal "n'" in sediment or size fraction and relevant parameter to limit errors of mea-
Bn represents the background concentration surements of x and y that may mask some
of the metal "n"-either found in the litera- of the natural variability of the parameters.
ture or measured directly in texturally equiv-
alent uncontaminated sediments or size frac- 13.4.3 Metak Grain size normahzation
tions. The factor of 1.5 is introduced to ac- One common procedure used to account
count for variations in background values for trace metal variability is the mathemati-
MANLIAI FOR (,EO( HEMI(AI ANALYSES 279

cal normalization to grain size. Application

to trace metal data usually shows that de-
creasing grain size with increasing metal con- #
centrations occurs consistently but the
~t
strength of the relationship depends on the + +

~f, ~+
metal and the sedimentary environment be-
cause of variations in mineralogical composi-
tion related to origin rather than particle
size. l
To establish the relationship between the
metal and grain s~ze, the concentrations of z,
1
the metal are plotted, for example, against ¢ V> ~,c) ~O [0o
~lud
the percentage of the mud size fraction.
F,g 5 Zn mud scatterplol tor the St Lawrence estuary
In most cases, some sort of linear relation-
(n = 49 r = 0.82) Sohd hne represents the regression
ship emerges of the general form y = ax + b, hne. Parallel dashed hnes define the 95% confidence
the strength of which is measured by the band Data from Lormg, 1978
correlation coefficient (r). A linear relation-
ship of the form y = a x is uncommon be- Data points falling inside the 95% confi-
cause the sand size fractions usually contain dence band would be considered to be part
some trace metals, the amount of which can of the natural population whereas those out-
be estimated from the intercept of the Y side would be considered to belong to an
axis. anomalous population. For example, an
If the relationship is significant ( p < 0.05 anomalous Z n / m u d ratio occurring outside
or 0.01), a regression line should be calcu- the graphed predlcnon bands is easily identi-
lated and graphed along with a 95% or 99% fied in Fig. 5. Such samples with anomalous
prediction band so that the natural geochem- m e t a l : g r a i n size ratios do not necessarily
ical population of that metal in relation to indicate anthropogenic inputs as they might
gram size changes can be defined. This means be the result of plotting errors, analytical
that there is a 9 5 ~ or 99% probability that errors, or anomalous concentrations of detri-
the points which fall outside the prediction tal heavy minerals containing the metal, such
band are from a different or anomalous pop- as Cr bearing magnetite or chromite. Such
ulanon. normalization for grain size, of course, is of
For example, Fig. 5 shows a linear plot little value if the grain size of the sediments
that takes the form of y = a x + b for Zn containing the trace metals is essentially the
concentrations in relation to the mud same, or a selected grain size fraction is
(material < 5 3 ~ m in diameter) content used. In both cases, the variability of the
(percentage by weight) in sediments from the trace metal concentrations reflect mineralog-
St. Lawrence estuary (Loring, 1978). It shows ical and source variability, but in the latter,
Zn varies significantly (r = 0.82, p < 0.001) the contribution to the total sediment is also
with the mud content of the sediments. The an important consideration.
proportional changes in Zn concentrations
with changes m mud content within a narrow 13.4.4 Metal: reference metal normahzatton
confidence band indicate that the data meet Metal: reference metal normalization can
the criteria for mathematical metal-grain size be used in addition to, or in lieu of grain size
normalization and allow for the compensa- normalization. The assumpnon ~s that the
tion of the grain size effect on the natural Zn reference metal used such as A1 or L1 repre-
populanon. sents a certain mineral fraction of the sedi-
28(} D H LORING AND R FT RANTALA

ments such as the use of Al as a proxy for 120

the granular variations of the aluminosilicate b~ 100

fraction, particularly the clay fraction, or a 80

specific mineral or mineral group such as the 6O

use of LI as a proxy for micas, a n d / o r the 4O

clay minerals (Table 10). If a metal proxy for

20-

grain size such as AI or Li is used in conjunc- - ~ & ,2, ~

0 ~ . i . i • J • i . i
tion with grain size data, then it should be
established that AI or other reference metal 250 -

varies significantly with grain size before it is

200 •
used. The normalizing element must,there-
fore, be an important constituent of one or 150 •

more of the major fine-grained trace metal I00 ,

carrier(s) and reflect their granular variabil-

ity in the sediments. They are "'conservative" 50

in that they have a uniform flux from crustal 0 L i i ' i • i • i

rock sources and so compensate for changes

in the input rates of various diluents or varia- 10-

9
tions m sedimentation rates. 8

In many cases, separate determinations of 7

the grain size distributions are not made or 5

not d e e m e d necessary and a metal proxy for

granular variations such as AI a n d / or Li is
used instead. This approach is valid and AI
has been successfully used by Windom (1987),
Windom et al. (1989) and many others for Fig. 6 Changes of L1, Zn and Al concentrations with
normalization of sediments that have not gram size in 18 samples from different parts of the
been derived from the glacial erosion of ig- Gulf of St. Lawrence. Data from Lormg, 1978.
neous rocks. It has been shown, however,
that LI (Loring, 1990) is superior to A1 for
the normalization of trace metal data from the Li and Zn bearing minerals with decreas-
glacial sediments and for the identification of ing grain size. The spread of Li concentra-
anomalous metal concentrations. Lithium tions and Zn within each fraction reflects the
determinations are also both precise and ac- effects of mineralogical differences (prove-
curate by F A A S (Rantala and Loring, 1990). nance) on the granular Li and Zn distribu-
The close statistical covarience of the non.
metal distributions with Li and grain size Figure 7 illustrates the use of LI for nor-
rather than A1 for such glacial sediments can malization of Zn data from the St Lawrence
be demonstrated from direct determinations estuary. This scatterplot of Zn versus Li was
of the trace metals, Li, and AI in individual constructed and a regression line with the
sediment size fractions from different parts 95% confidence band graphed to define the
of the Gulf of St Lawrence (Loring, 1990). natural geochemical population of Zn in re-
Figure 6 shows, for example, the close co- lation to the normalizer element Li in the
varience of Zn with Li but not with AI con- sediments. The confidence band was in-
centrations as the metals increase with the cluded to indicate a 95% probability that the
decreasing size fractions. The close covari- points that fall within the prediction interval
ence of Li and Zn with changes in the size belong to the normal population and those
fractions in this case, reflects the increase in outside belong to an anomalous population.
MANUAL FOR GEOCHEMICAl. ANALYSES 281

fact of the analysis. If real, its presence would

suggest that one of the sediments is influ-
enced by anthropogenic inputs although the
+ + +/ - possibility of a natural enrichment of a Zn
,_< I + ++++~¢;~ ++ mineral carrier could not be ruled out with-
i
+>7+ out further investigation.
E~q + i ~+ +
~.//++ + ÷ One of the major disadvantages is that the
+
m e t a l / L i or any other reference element
normalization yields a ratio value instead of
a real concentration. This can be overcome
I in varying degrees by standardizing the con-
ol tents to a reference material and defining an
1~ ~o ~0 - To ~o
Ll mg/kg enrichment factor.
Fig. 7 Z n ' L 1 scatterplot for the St Lawrence estuary The relative enrichment of Zn or any other
(n = 49' r = 0 87). Solid line represents the regression metal can be estimated by standardizing
line. Parallel dashed lines define the 95% confidence
anomalous metal values to those in some
band Data from Lorlng, 1978.
background reference sediment. For exam-
ple, an enrichment factor of Zn relative to Li
Figure 7 shows that Zn varies significantly can be defined from (Zn/Li,ample) /
(r = 0.87, p < 0.001, n = 49) with Li (Loring, (Zn/L1 reference sediment )"
1990) within a narrow confidence band. The The validity of such an enrichment factor
choice of Li as a proxy or normalizer for size varies with values used for the reference
variations in grain size can be confirmed by material. Most authors have used values for
its significant correlation with mud content crustal abundances which are of little value
(r = 0.88, p < 0.001, n = 49) as with Zn (Fig. because they do not represent the regional
5) in these sediments. background levels and because of the analyt-
In order to use such a scattergraph as in ical uncertainty in their measurements. To
Fig. 7 to identify anomalous values of Zn, for be of any value, the reference ratio must
example, in other samples from this area it represent the material for which a compari-
would be necessary to: son ~s required.
(a) define and remove any outlier values
from the data set representing the natural 13.5 Multi-element normalizatton
metal population;
(b) recalculate and graph the trimmed data A m u l t i - e l e m e n t / c o m p o n e n t study in
set; which the major and trace metals along with
(c) determine the Li and Zn concentra- grain size and organic carbon contents have
tions at other sites; been measured allows the inter-relationships
(d) plot the Zn versus Li data on the between the variables to be estabhshed in
scatterplot. the form of a correlation matrix. From such a
Data points falling inside the 95% confi- matrix, the most significant trace metal: rele-
dence band would be considered to be part vant parameter(s), be it grain size, A1, Li, or
of the natural population where as those organic carbon, can be determined and used
outside would be considered to belong to an for normalization, identification of metal car-
anomalous population. For example, an riers, and detection of anomalous trace metal
anomalous Z n / L i ratio occurring outside the values. Factor analyses can sort all the vari-
graphed prediction bands is easily identified ables into groups (factors) that are associa-
in Fig. 7. Firstly, it should be determined if tions of highly correlated variables so that
the anomalous ratio is real and not an arti- specific a n d / o r nonspecific textural, miner-
282 DH IORIN(-, AND RT I RANIALA

alogical a n d c h e m i c a l f a c t o r s c o n t r o l l i n g the Gaudette, H.E., Flight, W.R., Toner, L and Folger,

trace m e t a l variability m a y be i n f e r r e d f r o m D.W, 1974 An |nexpenslve titration method for the
determination of organic carbon m recent sedi-
the d a t a set ( S p e n c e r et al., 1968; L o r i n g
ments. J. Sediment Petrol. 44:249-253
19781. Gladney, E and Roelandts, I, 1988a 1987 compilation
of elemental concentration data for USGS BIR-1,
13. 6 Summary of normalization procedures DNC-1, and W-2 Geostandards Newsl., 12 63-118
Gladney, E. and Roelandts, I, 1988b 1987 compdatlon
T h e use o f the g r a n u l o m e t r i c m e a s u r e - of elemental concentration data for USGS BHVO-1,
MAG-I, QLO-1, RGM-1, SCO-I, SDC-I, SRG-1,
ments, m e t a l / A 1 , m e t a l / L i or o t h e r m e t a l /
and STM-I Geostandards Newsl, 12 253-362
r e f e r e n c e e l e m e n t ratios are all useful ap- Govmdaraju, K, 1989 1989 compllatton of working
p r o a c h e s t o w a r d s c o m p l e t e n o r m a l i z a t i o n of values and sample descrlpuon lor 272 geostandards
g r a n u l a r a n d m i n e r a l o g i c a l variations, a n d Geostandard Newsl., 13, Spcc Iss, 113 pp
identification of a n o m a l o u s m e t a l c o n c e n t r a - Hatch, W R and On, W L, 1968. Determination of
sub-microgram quantltms of mercury by atomm ab-
tions in s e d i m e n t s . T h e i r use r e q u i r e s that a
sorption spectrophotometry Anal Chem, 40
large a m o u n t of g o o d analytical data be col- 2(185-2087
lected a n d specific g e o c h e m i c a l c o n d i t i o n s Hlnderberger, E J., Kaiser, M.L and Kolrtyohann S.R.,
be met b e f o r e all the n a t u r a l m e t a l v a r i a b d - 1981 Furnace atomic absorption analysis of biologi-
ity is a c c o u n t e d for, and the a n o m a l o u s m e t a l cal samples using the L'vov platform and matrix
levels can be d e t e c t e d . A n o m a l o u s metal lev- modification Atomic Spectrosc, 2 1--7
Jackson, M L, 1958 Soil Chemical Analysis Prentice
els, however, m a y not always be a t t r i b u t e d to
Hall, New York, N.Y, 485 pp.
c o n t a m i n a t i o n but r a t h e r c o u l d easily be a Kersten, M and Forstner, U, 1987 Effect of sample
reflection of d i f f e r e n c e s in s e d i m e n t prove- pretreatment on the rehabdlty of ,sohd speclatlon
nance. data of heavy metalsqmphcatlons lot the study of
G e o c h e m i c a l studtes that involve the de- early dlagenctmprocesses Mar ('hem, 22 299-312.
Kingston, H.M and Jassle, J B, 1986 Microwave en-
t e r m i n a t i o n o f m a j o r a n d heavy metals, grain
ergy for acid decomposmon at elevated tempera-
size p a r a m e t e r s , o r g a n i c m a t t e r , c a r b o n a t e , tures and pressures using biological and botanical
and m i n e r a l o g i c a l c o m p o s i t i o n in the sedi- samples Anal Chem, 58 2534-2541
m e n t s are, however, m o r e suitable for d e t e r - Kothandaraman, P and Dallmeyer, J F., 1976 Im-
m i n m g the factors that c o n t r o l the trace proved desiccator for mercury cold vapour tech-
m e t a l distribution t h a n m e a s u r e m e n t of ab- tuque At. Absorpt. Newsl, 15 120-121
Lorlng, D H, 1975 Mercury In the sediments o1 the
solute trace m e t a l c o n c e n t r a t i o n s an a spe-
Gulf of St Lawrence Can J Earth Scl, 12 1219-
cific size f r a c t i o n s or the use of m e t a l / r e f e r - 1237
ence metal ratios alone. This is b e c a u s e such Lormg, D.H, 1978 Geochemistry ol zinc, copper, and
s t u d m s can identify the factors that c o n t r o l lead in the sediments of the estuary and Gulf of St
the trace m e t a l v a r t a b d i t y in the sediments. Lawrence. Can J Earth Scl, 15 757-772
Lorlng, D.H, 1981 Potentml b|oavaflabll|t!¢ of metals
m eastern Canadian estuarme and coastal sedi-
14 (I REFERENCES
ments. Rapp P V Reun Cons Int Explor Mer,
181 93-101
Bartlett, R and James, B, 1980 Studying dried, stored Lormg, D H, 1987 A final report on the ICES mter-
soil samples-some pitfalls Sod Scl Am, 44 721- cahbratlon for trace metals m marine sediments
724 (1/TM/MS) Int. Explor Mer. (Copenhagen, DK),
Bouma, A H, 1969 Methods for the Study of Sedi- 143 134 pp
mentary Structures, Robert E Krmger Pubhshmg Lormg, D H, 19911 Llthmm- a new approach for the
Co., Huntington, New York, 157 pp. granulomemc normalization of trace metal data.
Fcldman, C., 1974 Preservatmn of dilute mercury solu- Mar Chem., 29 155-168
tions Anal. Chem., 46 99-102. Lorlng, D H. and Asmund, G, 1989. Heavy metal
Flanagan, F J, 1984 Three USGS maflc rock refer- contamination of a Greenland fjord system by mine
ence samples, W-2, DNC-1, and BIR-1 U S Geol wastes. Environ. Geol Water Scl, 14 61-71.
Surv Bull 1623, 54 pp Loring, D H. and Nota, D J G, 1973 Morphology and
M A N I . I A I F O R ( J E ( ) ( H E M I ( A L ANALX~SES 283

sediments of the Gulf of St Lawrence. Fish Res Slavln, W., 1988 Graphite furnace AAS for biological
Board Can Bull, 182 147 pp. materials Scl Total Environ, 71 17-35
Lorlng, D H and Rantala, R T.T., 1988. An lntercah- Spencer, D W , Degens, E T. and Kulblckl, G , 1968.
braUon exercise for trace metals in marine sedi- Factors affecting element distribution in sediments
ments Mar C h e m , 24 13-28. In L.H Athens (Editor), Symposium on the Origin
Lorlng, D H and Rantala, R.T.T, 1990 Sediments and Distribution of the Elements Pergamon Press,
and suspended particulate matter" Total and partial New York, pp. 981-999
methods of digestion ICES Tech. Mar Environ Sturgeon, R E , Desaulnlers, J A , Berman, S S and
Scl 9 Int. Explor Met (Copenhagen, DK) 14 pp Russell, D S, 1982 Determination of trace metals
I,orlng, D H., Rantala, R T T., Morris, A W., Blake, in estuarlne sediments by graphite furnace atomic
A.J , and Howland, R J M , 1983 Chemical compo- absorption spectrometry Anal Chlm Acta, 134
sition of suspended particles in an estuarlne turbid- 283- 291
lly maximum zone Can J Fish. Aquatic Scl, 40, Tanner, C B and Jackson, M L, 1947 Nomographs of
Supp No l 201-206 sedimentation times for ,,oil particles under gravity
Lorlng, D H , Rantala, R T T Morns, A W and As- o! centrifugal acceleration Soil Scl Soc Proc, 6(1-
round G , 1985 Non-detrltal and detrltal particulate 65
metal transport In estuarlne systems In. D E Cald- Tessler, A , Campbell, G C and Bisson, M , 1970
well, J A Brlerlcy and C L Brlerley (Editors), Plan- Sequential extraction procedure for the speclatlon
etary Ecology Van Nostrand Reinhold Company, of particulate trace metals Anal Chore, 51 844-
N x( , pp 187-197 851
Luoma, S N and Jcnne, E.A., 1 9 7 6 Eshmatlng Thompson, K.C, 1978. Shape oI the atomic-absorption
bloavallablllty of sediment-bound trace metals with calibration graphs for chromium using an alr-
chemical extractants In D.D Hemphlll (Editor), acctylenc flame Analyst, 103 1258-1262
Trace Substances m Environmental Health. Unl,~ Thompson, E A , Luoma, S N., Cain, D J and Johans-
Missouri Press, Columbia, MO, pp 343-351 son, C , 1980 The effect of sample storage on the
Muller, G , 1979 Schwermetalle in den sedlmenten des extraction ot Cu, Zn, Fe, Mn and organic material
Rhelms-Veranderungen selt Umschau, 79' 778-783 trom oxidized estuarme sediments Water, Air, Soil
Nlrel, P V M and Morel, F.M M., 1990 Pitfalls of Pollut, 14 215-233
sequential extractions Water Res, 24' 1055-1056 Tyson, J F , 1984 A critical look at calibration proce-
Rantala, R T T and Lorlng, D.H., 1977. A rapid deter- dures for flame atomic absorption spectrometry
ruination ol 10 elements in marine suspended par- Analyst, l(19 313-317
hcnlatc matter by atomic absorption spcctropho- Velllon, C , 1986 Trace element analysis ol biological
tometry At Abs Ncwsl, 16 51-52 samples. Anal Chem., 58 851A-866A
Rantala, R T T and Lorlng, D H , 1978 Atomic ab- Walkey, A , 1947 A critical examination of a rapid
sorption analysis of USGS reference sample Ma- method for determining organic carbon m soil Soil
rine Mud MAG-I Geostandards Newsl., 2 125-127 S c l , 6 3 251-263
Rantala, R T T and Lorlng, D . H , 1980 Direct deter- Windom, H L , 1987 Guide to the Interpretation of
ruination ol cadmium m slhcates from a fluoborlc- reported metal concentrations m cstuarlnc sedi-
boric acid matrix by graphite furnace atomic absorp- ments Florida Dcpt Environ Regulation (Talla-
turn spectrometry. At Spectrosc, 1' 163-165 hassee, Florida), 43 pp
Rantala, R T T. and Lormg, D H., 1985. Partition and Wlndom, H.L, Schropp, S J Calder, F D , Ryan, J D ,
,

determination of cadmium, copper, lead and zinc m Smith, Jr.R G , Burney, L C , Lewis, F G and Rawl-
marine suspended particulate matter Intern J En- mson, C . H , 1989 Natmal Trace metal concentra-
viron Anal C h em , 19 165-173 tions m estuarlne and coastal marine sediments ot
Rantala, R T T and Lormg, D . H , 1987 Cadmium in the southeastern United States Environ Scl Tech-
marine sediment,s. Determination by graphite fur- nol, 23:314-320
nacc atomic absorption spectroscopy. ICES Tech Yeats, P A and Brugmann, L, 199(I Suspended partic-
Mar Environ Sol 3 lnt Explor Met (Copenhagen, ulate matter Collection methods for gravlmetrlc
DK), 9 pp and trace metal analysis ICES Tech Mar Environ
Rantala, R T T and Lormg, D H , 1989 Teflon bomb Scl 7 Int Explor Mer (Copenhagen, DK) 9 pp.
decomposition of slhcate materials in a microwave Yeats, P A and DalzJel, J A 1087 ICES mtercahbra-
,

oven Anal Chlm Acta, 220 263-267 tlon for metals m suspended particulate matter. J.
Rantala, R T T and Lormg, D H , 1990 Determina- Cons Int Explor Mer., 43 272-278
tion of Lithium m NRCC sediments: MESS-l,
BCSS-1, PACS-I and the NIST river sediment SRM
27(14 Geostandards Newsl., 14 475-476