Simultaneous analysis of *15N, *13C and *34S and NCS

Report
Performance update for soil and sediment
samples and their simultaneous analysis of
δ15N, δ13C, δ34S and NCS concentrations
using an Elementar Vario Isotope EA and
Isoprime 100 IRMS.
By Paul D. Brooks, Stefania Mambelli, Kari Finstad, Joey
Pakes and Todd E. Dawson.
Univ. of California, Berkeley
Disclaimer
• The product names used in this presentation are for
information only and do not constitute a promotion
or endorsement by the University of California,
university affiliates or employees.
Acknowledgements
•
•
•
•
•
•
•
The authors would like to thank:
Dr. Brian Fry, formerly Univ. of Hawaii.
Dr. Andreas Rossmann, Isolab Germany.
Steve Silva, USGS Menlo Park, Ca.
Scott Hughes, Elementar Americas Inc.
Robin Sutka, formally of Elementar Americas Inc.
Everyone who has replied to questions on
Isogeochem and has attended ASITA or earlier
CFIRMS conferences.
Instruments used
• All information in this presentation were
generated using:
• An Elementar vario ISOTOPE cube
interfaced to:
• A Isoprime 100 mass spectrometer.
Why analyze NCS isotopes in one
sample?
• Analysis of food web can greatly benefit from the addition of 34S by
adding an additional dimension to the analysis. Food web studies
usually require a large number of samples to reduce the noise level
of the data.
• Analysis for NCS concentration, then weighing out individual
aliquots of each sample for separate N, C, S isotope analysis
reduces the number of field samples that can be analyzed.
• Many samples are so small that it is impractical to subsample them
into two different aliquots for analysis by two different methods.
• Combining samples results in the loss of the field sample noise
which is usually critical to answering the experimental hypothesis in
Ecosystem sciences.
Example of NCS data for individuals from stream population. Note noise level in
populations and change in S ratio when N and C do not change. Samples from student
Hiromi Uno.
sample name
Ephemerella maculata
Ephemerella maculata
Ephemerella maculata
Ephemerella maculata
Ephemerella maculata
parasitic nematode
Ephemerella maculata
Ephemerella maculata
Ephemerella maculata
Ephemerella maculata
Ephemerella maculata
soapsoapsoapsoapsoapavg=
stdev=
soap+
soap+
soap+
soap+
soap+
avg=
stdev=
Timpanoga_nymph
Timpanoga_nymph
Timpanoga_nymph
avg=
stdev=
Timpanoga_subimago
Timpanoga_subimago
Timpanoga_subimago
avg=
stdev=
Timpanoga_imago
Timpanoga_imago
avg=
%N
ug N in
capsule
d 15N
%C
mg C in
capsule
d 13C
%S
ug S in
capsule
d 34S
13.85
12.33
13.48
13.84
15.70
13.84
1.08
12.75
12.10
13.28
12.60
12.81
13.11
12.78
0.41
10.36
9.87
10.60
10.28
0.30
11.82
11.87
10.92
11.54
0.44
11.89
12.33
12.11
205
188
199
183
212
198
11
157
208
182
178
197
173
187
13
137
140
148
142
4
160
183
181
175
10
183
191
187
5.75
7.58
4.42
1.42
3.56
4.54
2.07
3.99
1.20
5.12
1.61
1.73
8.17
3.57
2.70
0.63
0.57
0.16
0.45
0.21
1.34
0.96
1.51
1.27
0.23
1.23
1.26
1.24
58.40
62.36
56.56
56.95
60.12
58.88
2.14
57.11
54.21
55.52
59.61
59.38
58.11
57.37
2.15
42.87
41.91
43.34
42.71
0.60
55.44
55.23
60.21
56.96
2.30
58.16
64.19
61.18
0.87
0.95
0.84
0.75
0.81
0.84
0.07
0.71
0.93
0.76
0.84
0.91
0.77
0.84
0.07
0.57
0.60
0.60
0.59
0.02
0.75
0.85
1.00
0.87
0.10
0.90
0.99
0.95
-28.06
-29.48
-26.99
-26.27
-27.85
-27.73
1.08
-25.72
-22.25
-28.51
-25.70
-25.25
-27.59
-25.86
2.17
-19.94
-19.66
-19.07
-19.56
0.36
-16.17
-17.06
-16.92
-16.72
0.39
-17.50
-17.19
-17.34
0.80
0.78
0.74
0.81
0.82
0.79
0.03
0.92
0.70
0.73
0.74
0.76
0.79
0.74
0.03
0.66
0.63
0.66
0.65
0.01
0.82
0.81
0.70
0.78
0.05
0.82
0.80
0.81
12
12
11
11
11
11
0
11
12
10
10
12
10
11
1
9
9
9
9
0
11
12
12
12
1
13
12
13
-1.61
-2.79
-1.11
-1.22
-1.53
-1.65
0.60
-1.30
-0.78
-2.38
-0.61
-1.30
-1.55
-1.32
0.63
2.44
3.00
3.05
2.83
0.28
2.37
2.91
2.88
2.72
0.25
2.46
2.71
2.58
• Large number of samples
required.
• Concentration required.
34
•  S may significantly improve
source identification.
To be useful, the NCS isotope analysis
must meet these requirements
• Be capable of a high throughput of over 60
unknown samples per day in order to analyze
many field samples and reduce field noise level.
• Costs, sample preparation and ease of analysis
should not be excessively higher than 15N 13C
analysis.
• The analysis system must be able to analyze a
wide sample range with different concentrations
range of N, C and S.
• Precision and accuracy must be similar to
conventional NC and S methods.
Problems solved for high throughput
NCS isotope analysis
• S analysis usually uses one combined combustion
reduction column with short lifetime. Solution: Use
separate combustion and reduction columns
connected with a heated quartz bridge. Only fill 110
mm center of reduction tube with Cu and heat to 880
°C.
• Variable 18 O in samples interferes with SO2 mass 66 as
66 can be due to 34S or 18O. Solution: Use a
magnesium perchlorate water trap immediately after
the 1st reduction tube followed by a 900 °C quartz
buffering tube with CuO at center to buffer O. (Fry et
al. 2002).
O buffering of SO2 with quartz buffering tube.
raw d34S
Magnesium perchlorate drying tube before quartz tube.
17
16
15
14
13
12
11
Silver sulfide with varying
amounts of EDTA added to
change C/S ratio with no
quartz buffering tube.
NOTE Y SCALES ARE DIFFERENT
0
100
C/S
200
300
Silver sulfide with varying
amounts of EDTA added to
change C/S ratio with quartz
buffering tube.
On every analysis we measure a
AgS2 standard with and without
added sucrose with no
difference in  34S.
Data from Robin Sutka.
Mitigation S memory
• Memory effects for S. Mitigation: Use a drying tube
immediately after the 1st Cu tube to trap water.
• Hypothesis, this may be due to SO2 and H2O being in
equilibrium with H2SO3. Keeping the water trap hot
may prevent H2O from condensing.
• This may be why a combined comb/red column or
heated connection between separate combustion and
copper tube is necessary for SO2.
• Could SO2 be dissolving in a H2O film?
SO2 + H2O
H2SO3
Test of various standards for memory effect.
Currently 0.11-0.15 µgS on UCB system.
S memory test
30.00
25.00
20.00
delta 34S
original delta
15.00
carryover
corrected delta
10.00
5.00
0.00
-5.00
0
20
40
Analysis Number
60
Water trap split to allow daily changes
of magnesium perchlorate.
Split water trap in place over Cu column
Prevent SO2 trailing, fully reduce NOx
• Problem: SO2 begins to trail as ash build up in
combustion tube.
• Solution: Trap SO2 and release after all SO2 is
collected.
• Problem: NOx is not fully reduced in 880 °C Cu
reduction column needed to pass SO2.
• Solution: Use a second 650 °C Cu reduction
column after the SO2 trap. (Brian Fry, personal
communication.)
Analyze N 29/28 and S 66/64 on same
triple collector.
• Problem: As the N 29/28 ratio is much smaller
than S 66/64, careful sample size selection
based on prior knowledge of the N and S
concentration is necessary to avoid saturating
S on mass 66 or insufficient N on mass 28 with
10 volt AD converters.
• Solution: Use an IRMS with 100 volt AD
converters for wide dynamic range.
Get accurate concentrations for NCS
• Concentration of N, C and S is not as precise
using the IRMS as from the EA.
• Solution: Interface the MS and EA software so
sample names and weights are input
automatically into the EA software and the
TCD concentration and IRMS isotope results
are combined in one final Excel file.
Calibration requires a large number of standards.
• Preferred range of sample size is 30-1000 µgN, 0.2-5
mgC (adjustable with different dilution) and 10-140
µgS in a capsule.
• Calibration of all three isotopes requires a large
number of standards.
• Solution: Use 120-place auto-sampler, a 10 minute
per analysis method, and analyze 133 capsules with
46 standards per analysis and 81 unknowns, 3
standards and 3 blanks at beginning to stabilize
system.
• 133 total capsules takes ≈22.2 hours.
• There is potential for reducing the number of
standards required.
Preferred sample range 30-1000 µgN
Isoprime 100 test of N range with NIST 1547 peach leaves and NIST 1577b bovine liver,
400 µAmp tuning
9
7
5
delta 15N
Bovine liver
3
Peach leaves
1
-1
Use dual mixing model to correct for small samples < 30 µgN
-3
-5
0
200
400
600
approx ug N in tin capsule
800
1000
Final schematic for
NCS isotope analysis
2nd Cu reduction tube 650°C
Large size
CO2 trap.
Heated quartz
Quartz buffering tube 900°C
WO3
Tungsten oxide comb tube 1150°C
Cu reduction tube 880 °C
Cu
Magnesium perchlorate trap
P2O5
trap
quartz
SO2
trap
CuO
P2O5
trap
CO2
trap
Bypass
valves
for SO2
P2O5
trap
quartz
TCD
To MS
Combustion tube, 1st reduction tube, and
magnesium perchlorate water trap.
Use tungsten oxide in long ash finger to
mitigate long term memory and
Increase combustion tube life.
TCD chromatogram
MS chromatogram
Is an added oxidant needed?
V2O5 is very toxic and we do not allow its use by
our undergraduates who weigh most of our
samples and standards.
Nb2O5 or WO3 are used as substitutes but do not
seem to work as well (Steve Silva personal
communication). Could this be because of
melting temperature?
V2O5 melting temp 690 °C.
Nb2O5 melting temp 1512 °C.
WO3 melting temp 1473 °C.
Is an extra oxidant needed?
• Oxidants seem to be added to help mitigate trailing
problems with SO2. This may not be necessary if the
SO2 is trapped, but depends on material (see later
slides).
%N
avg stdev
d15
N
avg stdev
%C
avg stdev
d 13C
stde
avg
v
%S
avg stdev
d 34S
avg stdev
peach w/Nb2O5
3.02 0.02 1.92 0.03 49.76 0.72 -25.87 0.03 0.21 0.02 8.56 0.08
peach no Nb2O5
2.96 0.01 2.06 0.03 48.41 0.61 -25.86 3.00 0.20 0.01 8.10 0.06
Yolo soil w/Nb2O5
0.10 0.00 3.74 0.21 0.87 0.01 -25.43 0.11 0.02 0.00 -2.67 0.44
Yolo soil no Nb2O5
0.10 0.00 4.31 0.14 0.85 0.01 -25.57 0.02 0.02 0.00 -2.31 0.13
Joey Pakes data.
15N
and 13C results are the same in NC
and NCS mode
• Since the second Cu reduction tube was
added 15N results have been the same as in NC
mode.
• 13C results have always been the same.
S is more challenging (difficult).
• Mass 66 saturates at about 140 ug S with current
system, potential exists to gain shift and increase the
range.
• If the samples are all similarly small size then sample
less than 4 µgS are feasible.
• The memory effect of the current system models at
about 0.11-0.15 µg S as estimated by fitting a dual
mixing model to the data. This may limit the
precision and accuracy of small samples with big
differences in isotope ratio.
• There is a phantom blank effect equivalent to about
0.8 µg S.
34S vs µ S for standard
9.00
8.80
8.60
delta 34S
8.40
8.20
8.00
rep dS
7.80
y=ax2+xb+c
7.60
corr
reported avg = 7.98 stdev= 0.36
7.40
actual
corrected avg = 8.10 stdev = 0.12
7.20
7.00
2
4
6
8
10
µS
12
14
16
18
≈ 0.8 µg S
Standardization procedure
• Use a calibration standard of 3.8-4.2 mg (32
µgS) bovine liver every 12 samples to correct
for drift, large size minimizes carryover.
• Put a variable weight bovine liver after the
calibration standard to use for QC.
• Put in 10 variable weight standards each of
fishmeal and spirulina to check carryover,
adjust linearity and normalize isotope values.
Post analysis calculation
• Drift correct between calibration standards using
peak to peak correction.
• Check S carryover using variable weight fishmeal and
spirulina standards.
• Summarize different standards data and move to dual
mixing model spreadsheet for linearity and
normalization correction.
• Check blank correction for S using fishmeal and
spirulina.
Soils and sediment analysis.
• Soil analyze well for N and C, but S may be
difficult for some soils and sediments.
• For example, SRM 1646a appears to have a
slow release of S resulting in a big memory
effect.
• This effect may in turn affect S analysis of later
samples.
Soils analysis for NCS shows no bias with size and without
V2O5 show good agreement with other analysis.
Sample
EM high organic B2151
EM high organic B2151
EM high organic B2151
EM high organic B2151
EM high organic B2151
Target Actual
mg
mg
5
5
10
10
10
5.08
4.90
10.12
9.89
10.15
EM low organic B2153
EM low organic B2153
EM low organic B2153
EM low organic B2153
EM low organic B2153
EM low organic B2153
avg
std
CERTIFIED VALUE
70
70
70
140
140
140
70.94
70.79
70.27
140.64
139.93
139.81
δ 13C
%S
ug S
Fry δ 34S
4.42
4.44
4.56
4.65
4.61
9.12
9.26
9.33
9.36
9.39
-26.36
-26.31
-26.35
-26.42
-26.41
0.75
0.76
0.73
0.73
0.73
43
42
81
79
81
4.49
4.50
4.41
4.25
4.12
4.54
0.10
4.42 +_ .29
9.29
0.11
-26.37
0.05
-26.27 +- 0.15
0.74
0.02
6.95
6.87
6.91
6.79
6.91
6.91
1.55
1.54
1.56
1.54
1.53
1.55
-27.57
-27.50
-27.42
-27.39
-27.31
-27.25
0.02
0.02
0.02
0.02
0.02
0.02
6.89
0.06
6.7 +- .15
1.55
0.01
-27.41
0.12
-27.46 +_ .11
0.02
0.00
ug N
δ 15N
0.69
0.72
0.67
0.67
0.67
35
35
68
67
69
0.68
0.02
avg
std
CERTIFIED VALUE
% C mg C
%N
0.14
0.14
0.14
0.14
0.14
0.14
0.14
0.00
98
97
97
190
189
189
0.46
0.45
0.94
0.93
0.95
1.10
1.09
1.09
2.17
2.15
2.16
4.35
0.16
4.20
19
18
18
36
36
36
4.80
4.66
4.67
4.64
4.54
4.49
4.63
0.11
4.94 +_ 1.4
Soils analysis for NCS shows no bias with size and without
V2O5 show good agreement with other analysis.
Sample
Target Actual
mg
mg
%N
ug N
δ 15N
% C mg C
δ 13C
%S
ug S
Fry δ 34S
Icacos soil
Icacos soil
Icacos soil
Icacos soil
Icacos soil
Icacos soil
Icacos soil
Icacos soil
Icacos soil
avg
std
40
40
40
60
60
60
80
80
80
40.41
41.70
39.73
59.17
60.31
60.81
80.96
80.88
79.92
0.29
0.29
0.28
0.29
0.29
0.29
0.27
0.27
0.28
0.28
0.01
117
120
112
170
173
174
221
219
222
3.50
3.41
3.45
3.49
3.43
3.50
3.43
3.44
3.51
3.46
0.04
5.10
5.05
4.96
5.13
5.10
5.10
5.00
4.98
5.04
5.05
0.06
2.06
2.11
1.97
3.03
3.07
3.10
4.05
4.03
4.03
-28.25
-28.27
-28.24
-28.29
-28.24
-28.25
-28.09
-28.01
-28.07
-28.19
0.10
0.05
0.06
0.06
0.06
0.06
0.06
0.05
0.05
0.05
0.05
0.00
24
25
25
37
38
39
50
51
51
15.38
15.36
15.57
15.38
15.20
15.42
15.96
15.93
15.95
15.57
0.29
Malachite lake Mud
Malachite lake Mud
Malachite lake Mud
Malachite lake Mud
Malachite lake Mud
Malachite lake Mud
5
5
5
10
10
10
5.19
5.01
4.98
10.01
10.09
10.00
2.12
2.13
2.05
2.00
2.02
1.96
110
107
102
200
204
197
-0.35
-0.31
-0.26
-0.16
-0.15
-0.22
40.72
40.71
40.16
40.47
40.41
40.12
2.11
2.04
2.00
4.05
4.08
4.01
-27.08
-27.05
-26.99
-26.23
-26.29
-26.25
0.21
0.23
0.23
0.25
0.26
0.28
12
13
12
28
29
31
8.78
8.52
8.63
8.07
7.96
7.62
-0.24
0.08
-0.37
40.43
0.26
39.48
-26.65
0.43
-26.44
0.24
0.03
0.21
avg
std
Andrew Schauer vlaues
2.05
0.07
1.93
8.26
0.45
8.26
Note analysis works well up to 140 mg
of soil, and possibly higher.
Sample
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
Yolo soil
average
std
Target Actual
mg
mg
50
50
50
70
70
70
100
100
100
120
120
120
140
140
140
50.16
50.10
50.02
70.04
70.07
70.04
100.03
100.13
100.13
120.04
120.16
120.08
140.09
140.07
140.03
%N
ug N
δ 15N
0.11
0.11
0.11
0.11
0.11
0.11
0.11
0.11
0.11
0.11
0.11
0.11
0.10
0.10
0.10
0.11
0.00
55
56
56
80
76
79
108
109
110
129
134
135
145
144
144
4.48
4.36
4.50
4.46
4.41
4.46
4.54
4.50
4.51
4.56
4.53
4.40
4.54
4.45
4.51
4.48
0.06
% C mg C
δ 13C
%S
ug S
Fry δ 34S
0.88
0.88
0.88
0.87
0.87
0.88
0.98
0.98
0.98
0.88
0.88
0.84
0.84
0.84
0.83
0.89
0.05
-25.46
-25.45
-25.41
-25.42
-25.39
-25.38
-25.41
-25.46
-25.40
-25.45
-25.46
-25.38
-25.48
-25.54
-25.52
-25.44
0.05
0.02
0.02
0.02
0.01
0.02
0.02
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.00
7
8
8
10
11
11
15
16
16
19
20
19
22
23
23
-3.41
-3.60
-3.30
-3.37
-3.60
-3.76
-3.93
-3.52
-3.56
-4.19
-3.63
-3.26
-3.67
-3.55
-3.38
-3.58
0.24
0.44
0.44
0.44
0.61
0.61
0.61
0.98
0.98
0.98
1.05
1.05
1.01
1.17
1.17
1.17
SRM 1646a sediment appears to introduce
a memory effect.
Memory Icacos soil alternate with SRM 1646a
20.00
Icacos soil
28-32 µgS
15.00
Delta 34S
10.00
5.00
0.00
0
10
20
30
40
50
60
-5.00
-10.00
SRM 1646a
18-22 µgS
-15.00
-20.00
run #
70
A 2 stage memory dual mixing model corrected the memory
effect. But how would the analyst know what correction to
apply?
memory correction 2 stage 0.8 µgS
20.00
Icacos soil
28-32 µgS
15.00
10.00
Ddelta 34 S
5.00
0.00
0
10
20
30
40
50
60
70
-5.00
-10.00
SRM 1646a
18-22 µgS
-15.00
-20.00
run #
Anoxic sediment had a sever carryover and
even appears to adsorb S from the next
sample. N and C results looked good.
run # name
104
105
106
107
108
109
110
111
112
113
114
anoxic sediment #1
anoxic sediment #2
anoxic sediment #3
Blank
bovliver
var_bovliver
blank
blank
blank
blank
blank
S pk ht
S pk ht should be
16.01
11.04
10.77
3.11
1.73
1.37
0.62
0.53
0.29
0.15
0.11
(0.08)
(4.07)
(2.78)
(0.08)
(0.08)
(0.08)
(0.08)
(0.08)
δ 34S
δ 34S should be
14.48
17.01
16.84
16.68
10.63
8.33
7.07
7.55
9.04
8.45
6.01
( 7.6)
( 7.6)
How to further improve NCS isotope
analysis (and current NC analysis?)
• Provide at least 3 standards with all NCS
values either heavy, light, and one to use as a
QC in between.
• Treat soils for S analysis carefully, especially
anoxic sediments.
Conclusions 1
• The system can analyze 133 total capsules (samples
including standards) in 22.2 hours.
• NCS mode requires additional standards, so 81
unknowns can be analyzed in 22.2 hours.
• Precision in a size range of 30-1000 µgN, 0.2-5 mgC and
10-140 µgS in a capsule compares well with separate
NC and S analysis.
• The only significant additional maintenance compared
to NCS is the changing of the 1st Cu reduction tube and
small water trap daily.
• S analysis is improved with capability to analyze 10
variable weight samples of 3 different 34S isotope
standards for a total of 30 normalization and QC
standards.
Conclusions 2
• S analysis should be over 10 µgS and better over 20 µgS which
minimizes problems with blank correction and carryover.
• This is not difficult to achieve as the Vario Isotope Cube is
easily capable of burning samples weighing of at least 10 mg.
The upper limit on sample size has not been explored.
• Soil samples up to at least 140 mg can be analyzed.
• Some soil or sediment samples do not analyze well for S even
though results for N and C are good.
• We have not tried to measure these problem sediment
samples adding V2O5 or other accelerants such as ammonium
nitrate.
• We hypothesize that anaerobic sediments are problematic for
S analysis as they have a large carryover.
References
• Fry, Brian. 2007. Coupled N, C and S stable isotope measurements using a
dual-column gas chromatograph system. Rapid Communications in Mass
Spectrometry. 21:750-756.
• Fry, B., et al. 2002. Oxygen isotope corrections for online 34S analysis.
Rapid Communications in Mass Spectrometry. 16:854-858.
• Sieper, Hans-Peter et al. 2006. A measuring system for the fast
simultaneous isotope ratio and elemental analysis of carbon, hydrogen,
nitrogen and sulfur in food commodities and other biological material.
Rapid Communications in Mass Spectrometry. 20:2521-2527.
• Hansen, T. et al. 2009. Simultaneous 15N, 13C and 34S measurements of
low biomass samples using a technically advanced high sensitivity
elemental analyzer connected to an isotope ratio mass spectrometer.
Rapid Communications in Mass Spectrometry. 23:2521-2527.

similar documents