RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 14, 335343 (2000)

Determination of trace elements by inductively

coupled plasma mass spectrometry of biomass
and fuel oil reference materials using milligram
sample sizes
Herve Lachas, Raphaelle Richaud, Alan A. Herod*, Denis R. Dugwell and Rafael Kandiyoti
Department of Chemical Engineering and Chemical Technology, Imperial College (University of London), Prince Consort
Road, London SW7 2BY, UK

Most of the analytical techniques used to quantify elements associated with solid samples suffer from high
detection limits and cannot be used for trace elements in biomass samples, particularly when only 20 mg are
available for analysis. Inductively coupled plasma mass spectrometry (ICP-MS) can achieve detection limits
of parts-per-trillion with liquid sample introduction by solution nebulisation. This technique was therefore
tested with two standard biomass reference materials: oriental tobacco leaves and cabbage leaves. Two
preparations successfully used on coal standards were used to digest the solid samples: a total digestion
method (wet ashing digestion) and a partial leaching (microwave extraction). The concentrations of up to
seventeen elements (As, Ba, Be, Cd, Co, Cr, Cu, Ga, Mn, Mo, Ni, Pb, Sb, Se, Sn, V and Zn) were measured
after the two preparations. The accuracy and sensitivity of the measurements improved when the dilution
factor decreased from 5000 to 1000 and to 500. Since the proportion of mineral matter in biomass samples is
small (5%), the microwave digestion extracted elements that are generally not completely extracted from
coal samples (e.g. Sb). However, some trace element concentrations were below the limit of quantification
after microwave extraction, even with a reduced dilution factor (As, Se and Mo) and could not be quantified.
A fuel oil was also digested. The trace element concentrations were very low (between 28 and 0.1 mg g1) but
acceptable results were obtained by applying a dilution factor of 100. Only six elements in the fuel oil (As,
Ba, Co, Ni, Se and V) had certified or indicated values. Factors affecting the accuracy and sensitivity of the
analyses are discussed. The reproducibility of analysis of the tobacco leaf standard was checked over a
period of nine months by both digestion methods. The wet ashing method gave acceptable reproducibility
for Ba, Cd, Co, Cu, Ga, Mn, Mo, Ni, Pb, V and Zn but poor precision for Cr, Se and Sn and showed evidence
of residual chloride interference for As. The microwave extraction gave good reproducibility for As, Ba, Cd,
Co, Cr, Cu, Mo, Ni and Zn but poor precision for Se and low recoveries for Ga, Mn, Sn and V. In spite of the
small quantities of material analysed, it proved possible to determine the trace elements at levels down to
0.1 mg g1 in the reference materials. Copyright # 2000 John Wiley & Sons, Ltd.
Received 30 September 1999; Revised 30 November 1999; Accepted 3 October 1999

Studies of the roles of metals in biomass have had a variety
of aims. The influence of growth conditions on biomass
production has been investigated. These studies include the
soil-pH,1 the concentration of trace elements in the soil,26
the concentration of trace elements applied to growing
foliage7 and the growth of plants on fly-ash basins.8
Mechanisms by which plants may store toxic, non-essential
metal ions such as Cd inside the cell have been
investigated9 and sunflowers have been used successfully
to clean uranium-contaminated waters.10 Enzymes, proteins
or porphyrins bearing elements such as Co, Cu, Mn, Mo,
Ni, V and Zn have been characterised.11,12 Organometallic
compounds containing Se are known13 and the element has
an important role in the growth of some plants2,11 and is an
*Correspondence to: A. A. Herod, Department of Chemical Engineer-
ing and Chemical Technology, Imperial College (University of
London), Prince Consort Road, London SW7 2BY, UK.
E-mail: a.herod@ic.ac.uk
Contract/grant sponsor: The European Union; Contract/grant number:
JOF3-CT95-0018 and JOF3-CT95-0024.
Contract/grant sponsor: NERC.
essential element in animals and humans.13 Trace element
concentrations in aerosol-particulates from forest fires14,15
have been investigated, as well as the pyrolysis,16 combus-
tion and gasification17 of biomass samples. Information
about the release of inorganic elements such as Na, K, Ca, P
or Cl can be found in the literature, but little information on
trace metal release from biomass in these systems is
apparently available.
Reference materials such as tobacco or cabbage leaves
are available, with certified values for trace elements
(supplied by the Institute of Nuclear Chemistry and
Technology, Warszawa, Poland and by the Institute of
Physics and Nuclear Techniques, Cracow, Poland, respec-
tively). Studies have been published on the quantification of
trace elements in biomass samples, such as cigarette
tobacco,18,19 spice samples,20 lichens,21 lucerne, melon,
maize, barley, carnation and oil palm,22 foods such as bread,
fruit and vegetables23 and on standard reference materials
alone.2427 Inductively coupled plasma mass spectrometry
(ICP-MS) has been the technique of choice in these studies
with detection limits down to parts-per-trillion, with

Copyright # 2000 John Wiley & Sons, Ltd.

336 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS

introduction of the sample by solution nebulisation after Table 1. Instrumental parameters and operating conditions used in
dissolution of the solid sample.28 In our previous work,29 ICP-MS
two digestion methods, wet ashing digestion and microwave
Forward power 1.3 kW
extraction, have been tested successfully to extract up to 17
Reflected power <5 W
elements (As, Ba, Be, Cd, Co, Cr, Cu, Ga, Mn, Mo, Ni, Pb, Plasma gases All argon
Sb, Se, Sn, V and Zn) from coal and ash standard reference Nebuliser gas flow rate 0.92 L min1
materials. The accuracy of the determination for Sn could Coolant gas flow rate 14.0 L min1
not be checked in coal standards29 because of the poor Auxiliary gas flow rate 1.5 L min1
precision for this element. Wet ashing with sulfuric acid was Sample uptake flow rate 0.5 mL min1 (pumped)
used as the first digestion step for coal to destroy Nebuliser De Galan V-groove
carbonaceous materials, followed by solution in several Spray chamber Water cooled (5 C)
mineral acids to destroy and dissolve the minerals remain- Detector Pulse counting
ing. However, the loss of volatile elements may occur when Dwell time 320 ms
Ion lenses Optimised on 115In
the samples are subject to relatively high temperatures
Sample time 66 s
during the open vessel acid attack.30,31 Microwave extrac- Washout time 60 s
tion was used31 as an alternative sample preparation. The Number of replicates 3
digestion of biomass and organic samples by microwave Scan region 5.6210.4 u
methods has been investigated21,22 and the use of HF acid in Skipped regions 11.449.6, 55.456.6, 79.480.6, 139.4
the digestion has been found to be necessary if the plant 200.6 u
material contains substantial quantities of silicates. In our Calibration mode External calibration
work32 organic samples such as solvent extracts and Curve fit Linear through zero
distillate material, which do not contain mineral matter,
do not require the wet ashing treatment and a microwave
extraction using nitric acid alone would seem suitable. while the microwave digestions were carried out in a
In the various programmes of work undertaken in our Milestone MLS microwave oven system (Sorisole, Italy)
laboratory to investigate the partitioning of particular trace provided with a rotating 10-position carousel containing
elements during co-combustion and co-gasification of coal 120-mL Teflon vessels from the CEM Company with
and biomass, a muffle furnace, low temperature ashing pressure release valves resistant up to 110 bars.
equipment and small bench-scale reactors, (wire mesh
reactor,33,34 hot rod35 and fluidised bed34) have been used.
Products formed under various conditions in these reactors
Sample solution preparation for analysis by ICP-MS
(ashes, tars and chars) are prepared in very small quantities,
sometimes as little as 1 mg. The small sample size available Wet ashing digestion. A mixture of raw fuel oil or biomass
required some novelty of approach as sample preparation sample and sulfuric acid was heated in a platinum crucible
normally involves relatively large quantities, usually on a hot plate at 150 C for 3 h, and then ashed in a muffle
between 200 and 500 mg as recommended by the reference furnace at 520 C overnight. The ashed product was treated
material distributors. This paper outlines the development of with perchloric and hydrofluoric acid and heated at 250 C
both extraction techniques toward the quantification of up to to incipient dryness. When fume emissions ceased, the
17 trace elements (As, Ba, Be, Cd, Co, Cr, Cu, Ga, Mn, Mo, residue was dissolved in dilute nitric acid and washed into a
Ni, Pb, Sb, Se, Sn, V and Zn) from standard reference volumetric flask. The nitric acid concentration allowed the
materials using sample quantities as small as 10 mg. introduction of a 2% nitric acid solution into the mass
Results obtained by two sample preparation methods, wet spectrometer without further dilution. Volumes (millilitres)
ashing digestion and HNO3 leaching in a microwave, prior initially used with gram sample sizes were reduced to
to analysis by ICP-MS, are presented for two biomass and microlitres for milligram sample sizes. A further reduction
one fuel oil standard. The aims of this work were to study of the volumetric flask volume, which would have allowed
the effect of factors such as the influence of sample size, the the attainment of a smaller dilution factor, proved
preparation methods, the limits of quantification (LOQ) and impossible since 8 mL was the minimum volume necessary
the dilution factor (DF) on the accuracy, precision and to rinse the crucibles. Dilution factors of 1000 and 500 were
reproducibility of the results for the different biomass reached by the introduction of 10 and 20 mg of biomass
matrices. The LOQ was defined as 10 times the standard standard, respectively. A single drop of the fuel oil standard
deviation of the procedural blank, multiplied by the dilution weighed 20 mg so a dilution factor of 500 was reached with
factor29 and expressed as a concentration in the solid sample that standard. The introduction of 100 mg, however, gave a
(mg g1). dilution factor of 100 which was necessary to measure the
concentrations of trace elements which were close to the
limits of quantification.
EXPERIMENTAL
Microwave extraction. For sample preparation using 10 mg
Instrumentation
of sample (the standard), nitric acid and hydrogen peroxide
All analyses by ICP-MS were performed on a commercially were introduced in a clean 7-mL PTFE vessel with a screw
available PQ 2 STE from Fisons Instruments (Wythen- top. 10-mL of concentrated HNO3 were added to the 120-
shawe, UK) using the standard operating conditions shown mL vessel to maintain the pressure inside the vessel. The
in Table 1. It was equipped with a de Galan V-groove small vessel was loosely closed and introduced into the
nebuliser and a Gilson Minipuls 3 peristaltic pump. A larger one, which was then placed on the carousel. A set
standard PC 486 computer provided overall system control. microwave programme was run: 1 min at 250 W, 2 min rest,
The wet ashing digestions were performed in Pt crucibles 5 min at 250 W, 5 min at 500 W and 5 min at 600 W.

Rapid Commun. Mass Spectrom. 14, 335343 (2000) Copyright # 2000 John Wiley & Sons, Ltd.
 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS
Maximum power available was 1000 W. After 5 min rest
under ventilation, the carousel was taken out and placed at
18 C for 1 hr to minimise the loss of volatiles and gases
on opening. The sample was then poured from the small
vessel into a 10-mL volumetric flask. For the biomass
standards, the yellowish solution was finally filtered to
remove the remaining solids (mineral matter such as
silicates that did not dissolve in the hot nitric acid). Dilution
factors of 1000 or 500 were reached by introducing 10 and
20 mg of standard, respectively. The introduction of 100 mg
of fuel oil was not possible using this procedure because the
volume of acid necessary to wet the sample would have
exceeded the volume of the small vessel.
Dilution factor (DF)
The total dissolved solid (TDS) in the digestion solutions
required control for analyses by ICP-MS; TDS levels should
generally be lower than 0.010.02% to prevent plasma
suppression effects. Dilution of the final solution was
therefore necessary prior to analyses by ICP-MS. At the
beginning of the study, the trace elements were deter-
mined32,29 using a solid to final solution dilution factor of
5000.
Biomass samples generally contain less than 5% ash,
while organic samples should be free of ash apart from the
residual metal oxides from porphyrins or enzymes; most of
the sample is lost as carbon dioxide and water during the
ashing in the presence of sulfuric acid. The ash level defines
the TDS level which is clearly lower than if all of the solid
materials were retained in solution, and therefore less
diluted solutions can be analysed. Since the minimum
volume of solution used to wash the Pt crucibles or the
PTFE vessels proved to be 8 mL, solutions prepared from a
10 mg sample were held in 10-mL volumetric flasks,
equivalent to a dilution factor of 1000. This gave an
increase in the sensitivity of the determination for all the
trace elements in biomass and fuel oil samples and
particularly for those at the lowest concentration. Trace
element concentrations are often low in biomass samples, so
experiments were also carried out using 20 mg of biomass
standards while conserving the final volume of 10 mL so the
dilution factor was around 500. For the fuel oil, a dilution
factor of 100 was used as the concentrations to be
determined were even lower. These increases in sample
sizes were possible because the raw unknown biomass
samples to be analysed were available in relatively large
quantities and so were the organic samples analysed during
laboratory work.
ICP-MS operating conditions
Typical instrumental parameters and operating conditions
are listed in Table 1; they have been used in previous
work.29 The instrumental background for ICP-MS was
checked at 5 u and 220 u. The instrument was optimised for
multi-element determination from 6Li to 210Pb. The ion
lenses were optimised using an isotope in the middle of the
mass range (115In), so that the count rate was relatively
uniform across the mass range considered. As shown in
Table 1, mass regions corresponding to elements of no
interest for this study were skipped; only selected ion ranges
were monitored. Owing to the wide dynamic range of the
technique (calibration curves in ICP-MS are linear over five
or six orders of magnitude), quantification was obtained for
Copyright # 2000 John Wiley & Sons, Ltd.
337
all the elements by external standard calibration curves
drawn through the origin and three points: 10, 20 and 30 mg
L1, using multi-element 2 N HNO3 aqueous standard
solutions. Isotopes used for the analyses were: As 75; Ba
137; Be 9; Cd 111; Co 59; Cr 52; Cu 65; Ga 71; Mn 55, Mo
95; Ni 60; Pb 208; Sb 121; Se 82; Sn 118; V 51 and Zn 66.
Each sample determination was the mean of three separate
sequential determinations, and a reagent blank correction
was made. Signal drift was monitored by running the 10 mg
L1 standard solution after every three samples and a linear
drift correction was applied to the raw count integrals. The
analytical procedure consisted of a calibration blank and a
procedural blank, calibration solutions, a batch of 3 or 4
solutions (standard and unknown samples with similar
matrices), a drift solution check, a batch of 3 or 4 unknown
samples, followed by drift solutions and 3 samples, repeated
as necessary, and finally a drift solution check. All results
are given with the relevant confidence intervals. The
precision has been expressed as 95% confidence limits
over three replicates.
Standard reference materials and reagents
Three standard reference materials were used in this study:
oriental tobacco leaves CTA-OTL-1 (certified by the
Institute of Nuclear Chemistry and Technology, Warszawa,
Poland), cabbage leaves CL-1 (certified by the Institute of
Physics and Nuclear Techniques, Cracow, Poland), and a
fuel oil SRM 1634c (certified by the National Institute of
Standards and Technology (NIST), Washington DC, USA).
These samples had certified values for many elements,
indicated or informational values for others, and no values
given for others. The indicated values corresponded to
estimates of concentrations which had not been so
rigorously compared by different analysts or analytical
methods as the certified values.
Sulfuric acid (98% w/w), hydrofluoric acid (40% w/w),
nitric acid (69% w/w), perchloric acid (60% w/w) and
hydrogen peroxide (30% w/v) were all of BDH (Poole, UK)
Aristar grade. Distilled water (Milli-Q Micropore grade)
was used throughout the work for rinsing vessels and
preparing samples. All the labware was cleaned using 10%
Decon 90 solutions and 5% HNO3 solutions. Standard
solutions (SPEX solutions supplied by Glen Spectra,
Stanmore, UK) at 10 mg mL1 were used to prepare multi-
element aqueous calibration standards.
RESULTS AND DISCUSSION
Analysis of the oriental tobacco leaf sample
Solutions prepared by wet ashing digestion. The results are
shown in Table 2. Eight of the trace elements (Ba, Co, Cu,
Ga, Mn, Pb, V and Zn) were determined accurately
following the wet ashing digestion of the oriental tobaccco
leaves (CTA-OTL-1), despite the use of the high dilution
factor of 5000. However, As, Be, Cd, Cr, Mo, Ni, Sb, Se and
Sn could not be quantified because they were present at
concentrations below the LOQ, as found for these elements
in coal and ash standards at this dilution factor for similar
levels of occurrence.29
The sensitivity improved when the dilution factor was
reduced to 1000 allowing the quantification of Cd, Cr, Ni
and Sb, but it was obviously not sufficient to quantify the
trace elements Mo and Se with certified concentrations less
Rapid Commun. Mass Spectrom. 14, 335343 (2000)
338 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS

Table 2. ICP mass spectrometric analysis of biomass CTA-OTL-1 after wet ashing digestion (in mg g1)
Element Certified values DF = 5000 DF = 1000 DF = 500 DF = 500 DF = 500
As 0.539  0.060 <0.4 1.15  0.24 1.18  0.15 1.18  0.19 1.90  0.27
Ba 84.2  11.5 93.8  4.9 95.4  11.2 91.8  5.1 94.7  10.1 85.9  10.4
Be <1 <1 0.11  0.03 0.11  0.04 <0.2
Cd 1.12  0.12 <1.4 1.20  0.11 0.69  0.22 0.70  0.08 1.16  0.27
Co 0.879  0.039 0.98  0.20 0.84  0.05 0.89  0.08 0.95  0.09 1.00  0.13
Cr 2.59  0.32 <4.4 3.68  0.74 2.23  0.11 1.79  0.40 3.11  0.65
Cu 14.1  0.5 15.2  1.8 13.5  0.5 12.4  0.7 13.3  2.5 11.1  1.5
Ga (a) 0.549  0.036 0.64  0.21 0.54  0.08 0.55  0.06 0.57  0.12 0.58  0.09
Mn 412  14 420  21 397  20 542  175 371  37 354  110
Mo [0.26] <2 <1 0.37  0.14 0.27  0.05 0.33  0.10
Ni 6.32  0.65 <2 5.26  1.04 3.50  0.27 5.68  0.82 5.71  0.70
Pb 4.91  0.80 4.28  0.37 4.48  0.80 4.97  0.15 3.93  0.39 4.14  0.64
Sb [0.075] <0.4 0.11  0.03 0.11  0.02 0.07  0.02 0.12  0.03
Se 0.153  0.018 <19 <3 <7 <1 <5
Sn <2 <3 0.19  0.05 0.16  0.07 <0.2
V 3.08  0.42 4.49  1.52 3.06  0.25 3.63  0.40 3.38  0.62 4.02  0.53
Zn 49.9  2.4 43.1  0.5 71.7  8.5 46.7  2.2 <54 44.9  4.8

[] = Informational value only.

(a) = Certified value from Ref. 26.

than the LOQ, or Be and Sn with no certified concentrations.
The value for Zn at 1000 dilution factor in Table 2 (71.7 mg/
g) was greater than the certified value or that at dilution
factor 5000 and it is presumed this preparation was
contaminated.
The high concentrations observed for As in Table 2,
approximately twice the certified value, indicate the
presence of residual chlorides as observed for ash and coal
samples;29 the problem arises from the use of perchloric
acid with incomplete removal of chloride ion as HCl vapour
and is particularly liable to affect low concentrations of the
element.
In an attempt to reduce the dilution factor, two further
approaches were considered: (i) reducing the final solution
volume or (ii) increasing the amount of sample used in the
digestion procedure. It was impossible to reduce the final
solution volume below 10 mL because of the requirement to
wash the crucibles sufficiently at the end of the digestion
procedure, but an increase in the sample size was possible
for the reference materials since sample was available in
gram quantities. Clearly, this last would not be possible for
samples produced in small quantities in the laboratory but
permitted the assessment of the experimental method to
produce accurate results; concentrations less than the LOQ
for real samples could then be accepted as the best estimate
available.
Accordingly, the amount of sample taken for analysis was
increased to 20 mg, in order to reduce the dilution factor to
500 and increase the sensitivity of the analysis. Improve-
ments were noticed for Be, Mo and Sn in Table 2 although,
even then, the concentrations of Be and Sn were less than
the LOQs in the right hand column of the table and the Se
concentration always remained below the LOQ for that
element.

Table 3. ICP mass spectrometric analysis of biomass CTA-OTL-1 after microwave extraction (in mg g1)
Element Certified values DF = 1000 DF = 1000 DF = 500 DF = 500 DF = 500
As 0.539  0.060 <0.7 <0.6 <1 0.83  0.11 0.41  0.13
Ba 84.2  11.5 78.2  9.6 82.6  19.6 71.2  3.2 86.7  5.0 78.9  7.8
Be ; <0.1 <1 <0.1 0.09  0.01 <0.04
Cd 1.12  0.12 1.13  0.54 1.01  0.24 1.37  0.72 1.27  0.07 1.88  0.24
Co 0.879  0.039 0.92  0.21 0.84  0.06 0.74  0.20 0.90  0.08 0.81  0.14
Cr 2.59  0.32 <1 1.33  0.32 2.75  0.09 1.68  0.32 2.10  0.28
Cu 14.1  0.5 13.8  2.9 11.1  0.8 14.0  0.4 14.6  1.3 12.3  1.3
Ga (a) 0.549  0.036 0.36  0.10 0.34  0.03 0.21  0.03 0.28  0.03 0.19  0.04
Mn 412  14 448  215 438  96 369  68 374  29 304  43
Mo [0.26] 0.53  0.18 0.45  0.06 0.36  0.11 0.34  0.08 0.30  0.07
Ni 6.32  0.65 8.51  3.42 6.53  0.61 5.44  0.97 6.30  0.56 6.64  1.86
Pb 4.91  0.80 4.38  1.16 3.45  0.97 3.60  0.69 4.40  0.18 3.51  0.10
Sb [0.075] <0.04 0.06  0.01 0.07  0.02 0.05  0.01 0.06  0.02
Se 0.153  0.018 <3 <1 <5 <1.4 <2
Sn <0.3 <0.5 <0.2 <0.2 <0.1
V 3.08  0.42 1.73  0.36 1.51  0.16 1.18  0.07 1.34  0.09 0.83  0.13
Zn 49.9  2.4 139  40 74.0  7.6 41.9  7.6 44.6  4.2 58.2  11.5

[] = Informational value only.

(a) = Certified value from Ref. 26.

Rapid Commun. Mass Spectrom. 14, 335343 (2000) Copyright # 2000 John Wiley & Sons, Ltd.
 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS 339

Table 4. ICP mass spectrometric analysis of biomass CL-1 (in mg g1)

Element Wet ashing DF = 5000 Wet ashing DF = 1000 Wet ashing DF = 500 Certified values Microwave DF = 1000 Microwave DF = 500
As <0.6 <1 <0.6 [0.13] <0.6 <1
Ba <5 3.07  0.14 3.56  0.10 [3.24] 4.39  0.64 2.30  0.78
Be <0.2 <1 <0.08 0.03  0.01 <0.09
Cd 0.22  0.08 <0.6 <0.4 0.23  0.04 <0.5 0.31  0.07
Co <0.4 <1 0.22  0.01 [0.73] 0.15  0.03 0.15  0.02
Cr 9.00  0.50 0.81  0.19 1.21  0.13 [0.98] <0.5 0.65  0.17
Cu 9.71  0.70 3.39  0.73 2.25  0.23 3.3  0.25 <0.7 2.72  0.24
Ga 3.02  0.91 <0.1 <0.1 0.03  0.01 <0.08
Mn 67.7  15.0 50.1  6.0 58.2  3.6 57.6  2.1 57.4  4.8 41.5  6.7
Mo <0.6 <0.2 0.37  0.15 [0.22] 0.37  0.04 0.23  0.10
Ni 6.01  0.91 <2 <2 3.0  0.30 2.48  0.49 <4
Pb <0.5 0.44  0.13 0.29  0.11 0.26  0.05 <0.1 0.22  0.01
Sb <0.4 <0.07 <0.04 <0.01 <0.05
Se <11 <3 <3 0.20  0.03 <1 <5
Sn <0.7 <0.3 <0.2 <0.5 <0.2
V 1.80  0.30 0.81  0.24 0.86  0.08 [0.52] <0.07 <0.1
Zn 43.0  3.7 33.5  4.8 37.5  1.5 38.5  1.9 63.1  1.5 23.3  5.1

[] = Informational value only.

Solutions prepared by microwave extraction. The results are
shown in Table 3. Ba, Cd, Co, Cu, Mn, Ni and Pb
concentrations were measured accurately following the
microwave extraction involving a 1000 dilution factor, but
Sb was less accurately measured. The recovery of V was
poor while Zn was overestimated, apparently due to an
unidentified interference ormore probablysample con-
tamination. The low concentrations of As and Se in the
sample were less than the LOQs at the dilution factor of
1000. The informational value for Mo of 0.26 mg/g was less
than the measured values, 0.53 and 0.45, although of similar
magnitude. No certified values are available for Be and Sn
but the wet ashing results of Table 2 (0.1 and 0.2 mg/g,
respectively) indicate their concentrations to be similar to or
less than the LOQs for these elements. The Cr concentration
was underestimated by both experiments at a dilution factor
of 1000.
A further reduction in the dilution factor to 500 allowed
the accurate estimation of As, Cr, Mo and Zn, in addition to
those already estimated accurately in the more dilute
solution, while V was poorly quantified; Se and Sn were
still below the LOQ. No certified value was quoted in the
standard reference sample for Be although the two
procedures used here indicate the same concentration of
about 0.1 mg/g. Results for Ga appeared to decline by nearly
a factor two with decreasing dilution factor, giving a result
less than half the certified value and this result cannot be
explained.
The results for the standard tobacco leaf show that the
digestion procedures lead to correct estimations of con-
centration for all the elements with the exceptions of As by
wet ashing and Ga and V by microwave. Neither method
was sensitive enough to allow estimation of Se and,
although both methods gave similar results for Be and Sn,
the results cannot be confirmed as accurate in the absence of
a certified concentration. The methods for digestion have
been scaled down in sample weights and acid volumes from
larger scale preparations29 and at the lowest dilution factors
are limited by the volumes of acid needed to react with the
sample size and wash out the crucibles used. LOQ values for
the two digestion methods have been shown29 for dilution
factors 5000 and 1000; values for DF500 were generally
about half those for DF1000. Therefore, it may be expected
that the procedures would allow accurate estimation of the
elements at concentrations (mg/g) above the values 0.3 (As),
0.4 (Ba), 0.1 (Be), 0.3 (Cd), 0.05 (Co), 0.5 (Cr), 0.7 (Cu),
0.05 (Ga), 0.2 (Mn), 0.1 (Mo), 1.2 (Ni), 0.1 (Pb), 0.05 (Sb),
1.5 (Se), 0.15 (Sn), 0.1 (V) and 2 (Zn). Comparison of the
results of Tables 2 and 3 with these values show that other
factors such as adventitious contamination of samples or
solutions sometimes prevent attainment of these ideal
values.
Analyses of the standard reference cabbage leaf sample
CL-1
Results are shown in Table 4, and indicate that some limited
improvements in accuracy were obtained in the determina-
tions of Ba, Cr, Cu, Mn and Pb, when the dilution factor was
reduced from 5000 to 1000 and later to 500, following
digestion by wet ashing. However, results at low dilution
factors were close to the certified values where the LOQ was
sufficiently low (Cu, Mn, Pb and Zn); results for elements
with values quoted as informational value only (Ba, Co,
Cr, Mo, V) were similar to the informational values and,
apart from Ba, were at values of less than 1 mg/g.
The microwave extraction results did not improve by
using a 500 dilution factor except for Cd, Cr, Cu, Mo and
Pb. Both digestion methods indicated that concentrations of
Be, Ga, Sb and Sn were very low, <0.1 mg/g for Be, Ga, Sb
and <0.2 mg/g for Sn. The concentrations of the elements
As, Cd, Co, Cr, Cu, Ga, Pb and V in CL-1 were too low to be
determined routinely. It was decided that the oriental
tobacco leaf sample (CTA-OTL-1) would be preferred as
the standard reference material in later studiesto test the
sample preparation methods and instrument behaviour,
when analysing biomass feeds.
Analyses of the fuel oil reference material
Our recent trace element analyses of coal tars and extracts of
coal and biomass in organic solvents have suggested that
some of the trace elements may actually occur in organic
association.32,36 A fuel oil derived standard reference
material was the closest matrix commercially available for
tars or organic solvent extracts. Of the 17 elements of

Copyright # 2000 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 14, 335343 (2000)
340 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS

Table 5. ICP mass spectrometric analysis of fuel oil SRM 1634c (in mg g1)
Element Wet ashing DF = 100 Wet ashing DF = 100 Certified values Microwave DF = 5000 Microwave DF = 500 Microwave DF = 500
As 0.15  0.03 0.13  0.03 0.14  0.01 <2 <0.3 <0.3
Ba 1.79  0.10 1.69  0.19 [1.8] 1.71  0.50 1.41  0.37 1.39  0.05
Be <0.02 <0.01 <0.4 <0.04 <0.04
Cd <0.01 <0.04 <2 <0.13 <0.13
Co 0.14  0.01 0.13  0.02 0.15  0.01 <0.9 0.20  0.03 0.19  0.03
Cr 0.22  0.03 0.21  0.06 <4 1.68  0.21 1.52  0.05
Cu 0.10  0.02 0.54  0.08 <2 1.03  0.27 1.49  0.30
Ga 0.02  0.01 0.02  0.01 <0.9 0.04  0.01 <0.05
Mn 0.26  0.01 0.54  0.07 <3 0.45  0.12 0.41  0.08
Mo 0.07  0.03 0.07  0.01 <3 0.14  0.06 0.15  0.02
Ni 17.9  0.1 18.0  2.1 17.5  0.2 12.3  5.9 19.8  2.7 20.8  2.3
Pb 0.55  0.02 0.50  0.04 <0.6 0.77  0.04 0.73  0.11
Sb 0.02  0.01 <0.01 <0.5 <0.05 0.02  0.01
Se <0.8 <0.6 0.102  0.004 <23 <6 <6
Sn <0.02 0.03  0.01 <1 0.18  0.05 0.18  0.03
V 22.9  0.2 23.5  2.9 28.2  0.4 22.0  2.9 16.6  1.6 17.5  1.1
Zn 3.42  0.23 2.84  0.39 <7 <2 <2

[] = Informational value only.

interest, only five elements have certified values (As, Co, conclusions. The low estimates for V after either prepara-
Ni, Se and V) and one has an indicated value (Ba). The tion were not fully understood.
sample was prepared by wet ashing and digested by
microwave extraction.

Analysis of solutions prepared by microwave extraction.

The first analysis carried out by microwave extraction with a
dilution factor of 5000 clearly showed that more concen-
trated solutions were required (Table 5). Only Ba, Ni and V
could be adequately quantified; the concentration found
for V was slightly lower than the certified value whilst the
concentrations found for Ni and Ba agreed with the certified
values. The dilution factor was then reduced to 500 directly
because the weight of the smallest drop that was easy to
produce, of this dense and viscous fuel oil standard, was
20 mg. A significant improvement was observed: many
more elements could be quantified but only the accuracy of
the Co concentration could be compared with an available
certified value. A further reduction in the dilution factor was
not possible for the microwave extraction without a
complete revision of the experimental procedure. While a
larger sample weight could be introduced into the sample
holder, the volume of acid required to react with the sample
and extract the elements was too large for the vessel.
Solutions prepared by wet ashing digestion. Here, the
dilution factor could be reduced to 100. This successfully
reduced the limits of quantification and the concentrations
of Co and As were determined accurately. Other elements
(Cr, Cu, Ga, Mn, Mo, Pb and Zn) were also quantified but
comparison with certified or indicated values was not
possible.
A comparison of the wet ashing and microwave results
for the elements without certified or indicated concentration
values indicated that for the five elements, Mn, Ga, Mo
(factor two difference between methods), Sb and Pb, the
concentrations found by the two sample preparation
methods (wet ashing and microwave digestion) agreed,
indicating a level of confidence in the results. For several
other elements (Cr, Cu, Sn and Zn), the differences were too
great and further analyses were needed before drawing firm
Limits of quantification (LOQ)
Taking into account both instrumental sensitivity and
uncertainties in sample preparation, the limits of quantifica-
tion (LOQ) were taken as the concentration equal to 10
times the standard deviation () of the procedural blank,
10, multiplied by the appropriate dilution factor. Values of
LOQ are thus quoted as concentration in the original solid or
liquid sample (mg g1) for each preparation procedure.
Within this framework the LOQ values represent the lowest
level of occurrence that could be reliably determined during
these analyses. Clearly diminishing the dilution factor to
1000 or to 500 reduced the LOQ values compared with a DF
value of 5000.
It was not surprising therefore to find that LOQ values for
the microwave extractions were usually lower than for wet
ashing digestion; more reagents were added to the sample
during the wet ashing procedure: three (or four) different
acids were added in an open vessel compared with using
only nitric acid and hydrogen peroxide in a closed
microwave digestion vessel. Thus the possibility of
adventitious contamination during the wet ashing digestion
procedure was higher. Contamination by reagents has been
reported31 for elements such as As, Cr, Cu, Mn, Ni, Pb, Sb,
Sn and V. The formation of polyatomic interferences for
51 52
V , Cr, 75As and 77Se during the actual analysis by
ICP-MS has been discussed elsewhere.29
The poor precision of the Se determination was
undoubtedly linked to the insufficient counts recorded
during the calibration of the instrument: 1500 counts s1
were usually recorded for the 10 mg L1 multi-element
standard solution in comparison with - say - 15 000 for Ba
(lowest number of counts after Se) and 200 000 counts s1
for V (largest number of counts). Contamination by Zn was
the probable explanation for the high limits of quantification
for that element.

Rapid Commun. Mass Spectrom. 14, 335343 (2000) Copyright # 2000 John Wiley & Sons, Ltd.
 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS 341

Table 6. Long-term reproducibility (% RSD) for the determination of 17 elements in biomass CTA-OTL-l by ICP-MS after wet ashing
digestion or microwave extraction (mass uptake = 20 mg; dilution factor = 500)
Wet ashing digestion Microwave extraction
Element Mean (mg g1) Long-term reproducibility (% RSD) Certified values Mean (mg g1) Long-term reproducibility (% RSD)
As Cl interference 0.539  0.060 0.49 43
Ba 91.4 4 84.2  11.5 78.6 8
Be 0.10 15 0.06 31
Cd 0.82 21 1.12  0.12 1.38 26
Co 0.95 6 0.879  0.039 0.82 9
Cr Poor precision 2.59  0.32 2.01 25
Cu 12.6 12 14.1  0.5 13.4 8
Ga (a) 0.57 8 0.549  0.036 0.23 18
Mn 415 21 412  14 364 12
Mo 0.38 22 [0.26] 0.32 14
Ni 6.03 30 6.32  0.65 6.00 12
Pb 4.34 14 4.91  0.80 3.83 11
Sb 0.10 16 0.06 23
Se Poor precision 0.153  0.018 Poor precision
Sn Poor precision Low recovery
V 3.41 15 3.08  0.42 1.07 20
Zn 44.2 9 49.9  2.4 46.5 17

Poor precision: RSD >50%.

Cl interference for As.
Low recovery: mean below the LOQ.
(a) = Certified value from Ref. 26.

Accuracy reference materials showed that both the digestion methods

Our previous work carried out on coal and ash standards29
has shown that reducing the amount of sample used for
analysis from 250 mg to 1020 mg does not necessarily lead
to reduction in the accuracy of the analysis of such samples.
The overall conclusion was that no obvious change in the
analytical accuracy could be detected and neither were
changes in procedure called for, when either the sample size
was reduced or the sample type changed, providing that the
concentrations to be determined were above the limits of
quantification. Major differences were observed between
results from the two different sample preparation proce-
dures adopted; the success of a particular method of sample
solution preparation also appeared to be dependent on the
nature of the sample and directly related to the mode of
occurrence of the trace elements. Results reported for ash
reference materials29 indicate that the wet ashing digestion
gives far superior results compared with the microwave
procedure, which does not contain an element of HF
digestion. Wet ashing was shown to allow the determination
of nearly all elements investigated with good accuracy,
often better than 10% of the value, except Se which was
present at concentrations (5mg/g) lower than the LOQ of
16mg/g.
Arsenic and selenium were the only elements to be
completely extracted in various ash samples by the HNO3
leaching during our microwave extraction procedure
which did not include treatment with HF. Results for all the
other elements by this method proved lower than the
expected values by a factor of two, apparently because
silicates were not destroyed. A reduction in the dilution
factor from 5000 to 1000 improved the sensitivity to
elements (Cd, Cu, Mo, Se, Sn or Zn) in low concentrations
by reducing the LOQs, but did not allow the quantification
of elements held in silicate minerals (Ba, Be, Cd, Co, Cr,
Cu, Mn, Mo, Ni, Pb, Sb, Sn and V).
In this work, results obtained for biomass and fuel oil
were suitable for those matrices, providing that the
concentrations to be measured were above the limits of
quantification. Although biomass may contain silicates34 it
appears from the present work that silicates in the tobacco
leaf or cabbage leaf standards did not trap all the trace
elements (As, Ba, Cd, Co, Cr, Cu, Mo, Ni, Pb, Sb, and Zn)
and only the values for Ga, Mn and V were low (Table 3), by
factors 2.4, 1.2 and 2.7, respectively. It is expected that use
of HF in the microwave digestion would improve the
estimation of Ga, Mn and V in these standard materials as
observed elsewhere34 for Cd, Cu, Pb and Zn. For the fuel oil
(SRM 1634 c), where silicate minerals are expected to be
absent, four (As, Co, Ni and V) of the five certified (As, Co,
Ni, Se and V) elements and the indicated concentration
element (Ba) were accurately estimated by the wet ashing
procedure, with Ba, Co and Ni accurately estimated by the
microwave method. The As concentration was below the
LOQ by microwave and the V result was low, but for some
reason (volatility possibly) other than trapping in silicates.
For both digestion methods, the certified Se concentration
(0.102 mg/g) was below the LOQ values of 0.6 (wet ashing)
and 6 (microwave).
Reproducibility
The long-term reproducibility was assessed by analysing
four different digestions of the tobacco leaf biomass
reference material CTA-OTL-1 over a nine month period.
Relative standard deviations (% RSD), calculated as the
standard deviation of the analysis over 12 results (4
preparations with 3 measurements each), divided by the
mean of these twelve determinations and expressed as a
percentage are presented in Table 6.
For the wet ashing results, the relative standard deviations
were below 10% for Co, Zn, Ga and Ba. The concentrations
of Zn and Ba were high compared with the LOQs, while
those of Co and Ga were around 20 times the LOQs (0.05

Copyright # 2000 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 14, 335343 (2000)
342 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS
and 0.03 mg g1 for the two elements, respectively). For
Be, Cu, Pb, Sb and V, the long-term precision could be
considered as relatively good (RSDs below 16%) although
for two elements (Be and Sb) the concentrations determined
(no certified values were available) were only about twice
the LOQ values. The relatively poorer reproducibility for
Mn (21%) may be due to signal drift during one of the
analyses. In this case, the three determinations were 583,
583 and 461 ppm so the short-term reproducibility was poor
and the mean significantly different from the other three
(542 compared with 365, 346 and 406 ppm). The poor
reproducibility observed for Mo and Cd may be due to
memory effects, or contamination carried over from
previous analyses (for Mo31), but the concentrations of
these elements in the sample were close to the LOQ values.
The 30% RSD calculated for Ni is likely to have resulted
from the presence of Ni in some of the component parts of
the instrumentation (e.g. major component of the interface
cones, between the plasma and the mass spectrometer). The
long-term reproducibilities of Cr, Se and Sn were above
50%, indicating that the values were widely spread (the
LOQs for Cr and Se were similar to or greater than the
certified concentrations and the Sn concentration was close
to the LOQ), while the high mean for As indicated the
presence of residual chlorides for these preparations leading
to interference in the plasma in the instrument from ArCl
in spite of the precautions taken to remove chloride as HCl.
The long-term reproducibility following microwave
extraction was similar to that after wet ashing digestion:
Co, Cu and Ba showed RSD values below 9%. RSD values
for V, Mn, Ni, Zn, Ga, Cd and Pb below 20%; whilst the
RSDs of As, Be, Cd, Cr and Sb were poor and that of Se was
above 50%. The LOQs of Be, As, Se, Cd and Sb were close
to the concentrations to be determined in the standard. The
reason for the high Cr RSD is not understood, particularly
since the mean of the twelve values was close to the certified
value and approximately ten times the LOQ.
For both sample solution preparation methods, the long-
term reproducibility was reasonable (considering the low
concentrations analysed) for all the elements except Se and
Sn where the LOQs were similar to or greater than the
concentrations. The LOQ depends on the purity of reagents
and the cleanliness of the laboratory environment, as well as
interferences coming from the plasma itself.
CONCLUSIONS
The accuracy and reproducibility of trace element content
determinations by ICP mass spectrometric analysis have
been investigated. Trace element contents found by two
different sample solution preparation procedures have been
compared using a small set of standard reference samples:
two solid biomass samples and a heavy liquid sample.
The mixture of acids (H2SO4, HF, HClO4 and HNO3)
used in the wet ashing digestion procedure proved to be the
most effective sample solution preparation method for most
of the elements in biomass samples. The microwave HNO3
leaching procedure appears more convenient for volatile
elements such as arsenic and for minimising polyatomic
interferences and contamination during sample preparation.
However, in the absence of HF, the microwave digestion
procedure did not completely extract trace elements from
silicates in the biomass.
Organic samples (including the fuel oil reference material
studied here) do not contain a mineral phase and can be
Rapid Commun. Mass Spectrom. 14, 335343 (2000)
digested by microwave extraction. However the trace
elements were present at relatively low concentrations,
similar to or less than the LOQs in many cases; both
digestion methods gave similar results except for V, which
appeared to be lost from the microwave digestion. The level
of agreement between the methods was much better than
observed previously for coals and ash samples. The dilution
factor needs to be reduced as much as possible to enhance
the sensitivity to trace elements at low concentrations, but
the reduction is limited by the opposing requirements of
increased sample size, reduced acid volumes and volume of
the digestion vessel. The wet ashing digestion achieved a
dilution factor of 100 compared with 500 for microwave
digestion, mainly because the wet ashing procedure
involves drying and dissolution steps with the final step
defining the limiting volume of solution.
By combining results obtained for the two procedures
evaluated in this work, up to 15 elements have been
determined accurately in biomass samples. For two other
elements (Be and Sn) there were no certified reference
values. The results indicate that the reduction of the sample
size to 20 mg or less did not affect the accuracy or the
precision of the analysis, provided that solutions of
sufficiently high concentration in trace elements were
prepared.
The analytical methods allowed the accurate determina-
tion of concentrations below 0.1 mg/g for the fuel oil at the
lowest dilution factor of 100 and, by selection of the
appropriate dilution factor, it appears possible to obtain
accurate and reproducible analyses of biomass and organic
samples. Also, it is likely that for samples which are not
standard reference materials, the methods should give
analyses which reflect the trace element composition of
the sample, including those concentrations below the LOQ
on the day of the analysis.
Acknowledgements
The authors thank the European Union for support under grant
numbers JOF3-CT95-0018 and JOF3-CT95-0024 and the NERC for
support for use of the ICP-MS facility at Silwood Park.
REFERENCES
1. Oborn I, Johnsson G, Johnsson L. Water, Air, Soil Pollut. 1995; 85:
835.
2. Lindstrom K. J. Phycology 1991; 27: 207.
3. Madan M, Kamra N. Acta Biotechnologica 1991; 11: 141.
4. Rida A. Soil Biol. Biochem. 1996; 28: 1037.
5. Gong X, Chen F. J. Appl. Phycology 1997; 9: 437.
6. Redente EF, Richards JL. Arid Soil Research and Rehabilitation
1997; 11: 353.
7. Khan SH, Bingham RL, Felker P. Agroforestry Systems 1997; 37:
155.
8. Dosskey MG, Adriano DC. Soil Biol. Biochem. 1993; 25: 1547.
9. Clemens S, Kim EJ, Neumann D, Schroeder JI. EMBO 1999; 18:
3325.
10. Dushenkov S, Vasudev D, Kapulnik Y, Gleba D, Fleisher D, Ting
KC, Ensley B. Environ. Sci. Technol. 1997; 31: 3468.
11. Steward FC. Plant Physiology, A Treatise, vol. III, Inorganic
Nutrition of plants, Academic Press: New York and London, 1963.
12. Kendrick MJ, May MT, Plishka MJ, Robinson KD. Metals in
Biological Systems, Ellis Horwood Ltd: Chichester, 1992.
13. Greenwood NN, Earnshaw A. Chemistry of the Elements,
Pergamon: Oxford 1989.
14. Artaxo P, Rabello MLC, Watt F, Grime G, Swietlicki E. Nucl.
Instrum. Methods 1993; B75: 521.
15. Gaudichet A, Echalar F, Chatenet B, Quisefit JP, Malingre G,
Cachier H, Buatmenard P, Artaxo P, Maenhaut W. J. Atmospheric
Chem. 1995; 22: 19.
Copyright # 2000 John Wiley & Sons, Ltd.
 ICP-MS OF TRACE ELEMENTS IN BIOMASS AND FUEL OIL REFERENCE MATERIALS
16. Agblevor FA, Besler S, Wiselogel AE. Energy Fuel 1995; 9: 635.
17. Dayton DC, Frederick Jr., WJ. Energy Fuels 1996; 10: 284.
18. Garg AN, Singh V, Chutke NL, Ambulkar MN. J. Radioanal.
Nucl. Chem. Articles 1995; 195: 161.
19. Mierzwa J, Adeloju SB, Dhindsa HS. Analyst 1997; 122: 539.
20. Galas W, Kita A. Chemia Analityczna 1997; 42: 403.
21. Bettinelli M, Spezia S, Baroni U, Bizzari G. Multielement
characterization of lichens and vegetable materials by ICP-MS
In Plasma Source Mass Spectrometry Holland G, Tanner SD.
(eds). Royal Society of Chemistry: Cambridge, UK, 1997; 211.
22. Curdova E, Szakova J, Koplik R, Suchanek M. Plant material
decomposition methods and determination of selected elements by
ICP-MS Elan 6000 In Plasma Source Mass Spectrometry, Holland
G, Tanner SD. (eds). Royal Society of Chemistry: Cambridge, UK,
1997; 223.
23. Baxter MJ, Crews HM, Robb P, Strutt PR. Quality control in the
multi-element analysis of foods using ICP-MS In Plasma Source
Mass Spectrometry, Holland G, Tanner SD. (eds). Royal Society of
Chemistry: Cambridge, UK, 1997; 95.
24. Dybczynski R, Polkowskamotrenko H, Samczynski Z, Szopa Z.
Fresenius J. Anal. Chem. 1993; 345: 99.
25. Dybczynski R, Kulisa K, Polkowskamotrenko H, Samczynski Z,
Szopa Z, Wasek M. Chemia Analityczna 1997; 42: 815.
Copyright # 2000 John Wiley & Sons, Ltd.
343
26. Smodis B, Dermelj M, Jacimovic R. J. Radioanal. Nucl. Chem.
Articles 1995; 190: 3.
27. Kucera J. Chemia Analityczna 1995; 40: 405.
28. Jarvis KE, Gray AL, Houk RS. Handbook of Inductively Coupled
Plasma Mass Spectrometry, (2nd edn). Chapman & Hall: London
1995.
29. Lachas H, Richaud R, Jarvis KE, Herod AA, Dugwell DR,
Kandiyoti R. The Analyst, London, 1999; 124: 177.
30. Lichte FE. Analysis of Coal by Laser Ablation ICP-MS, Proc. Int.
Conf. Elemental Analysis of Coal and its By-Products, September
911, 1991, Barren River Resort, USA, Vouvopoulos V (ed) 1991.
31. Fadda S, Rivoldini A, Cau I. Geostandards Newsletter, 1995; 19:
41.
32. Richaud R, Lachas H, Lazaro M-J, Jarvis KE, Herod AA, Gibb TC,
Kandiyoti R. Fuel 1999; 79: 57.
33. Li C-Z, Wu F, Cai H-Y, Kandiyoti R. Energy Fuels 1994; 8: 1039.
34. Li C-Z, Wu F, Xu B, Kandiyoti R. Fuel 1995; 74: 37.
35. Pindoria RV, Lim J-Y, Hawkes JE, Lazaro M-J, Herod AA,
Kandiyoti R. Fuel 1997; 76: 1013.
36. Richaud R, Lazaro M-J, Lachas H, Miller BB, Herod AA, Dugwell
DR, Kandiyoti R. Rapid Commun. Mass Spectrom. 2000; 14: 317.
Rapid Commun. Mass Spectrom. 14, 335343 (2000)