8 September 2009

How accurately do we know13C/12C, 18C/16O and 17O/16O ratios in CO2and their corresponding δ values?

Jan Kaiser

School of Environmental SciencesUniversity of East AngliaNorwichUnited Kingdom

Recommended inter-laboratory comparability (1σ)

14th WMO/IAEA Meeting of Experts on Carbon Dioxide Concentration and Related Tracers Measurement Techniques, 10-13 September 2007 in Helsinki, Finland

CO2 ±0.1 ppm (±0.05 ppm in southern hemisphere)13δ(CO2) ±0.01 ‰18δ(CO2) ±0.05 ‰

How do we measure carbon and oxygen isotope ratios with mass spectrometers?

1//

r44

r45

444545 −=

IIII

Iδ

Measure ion current ratios of sample relative to reference (NBS 19-CO2 as stand-in for VPDB-CO2)m/z 44 12C16O2

+

m/z 45 13C16O2+ 12C17O16O+

m/z 46 12C18O16O+ 13C17O16O+ 12C17O2+

We assume that all minor isotopologues have the same relative ionisation efficiency. Then, the ion current delta is the same as the isotopologue delta.

1//

r44

r45

44454545 −==

RRRR

I δδ

"Conventional" isobaric correction using isotopologue and isotope ratios R

r17

r13

r45 2 RRR +=

2r

17r

17r

13r

18r

46 22 RRRRR ++=

RRR 171345 2+=21717131846 22 RRRRR ++=

528.01817 RAR = 528.0r

18r

17 RAR =

Solve for 3 unknowns 13R, 17R, 18R

1//

r44

r45

444545 −=

RRRRδ

"New" isobaric correction using δ values (Kaiser & Röckmann 2008)

δδδ 17r

1713r

1345r

45 2 RRR +=

)2()(22 217172r

1717131713r

17r

1318r

1846r

46 δδδδδδδδ +++++= RRRRR

111 18

1813

17

1717

13

1313 −=−=−=

rrr RR

RR

RR δδδ

Solve for 3 unknowns 13δ, 18δ, 17δ(including possible 17O isotope excess aka oxygen isotope anomalies)

1)1)(1( 528.0181717 −++= δΔδ

"New" isobaric correction using δ values (Kaiser & Röckmann 2008)

1)1)(1( 528.0181717 −++= δΔδ

)(2 17454513 δδδδ −+= C

[ ]217174517464618 3)1()42()22()2( δδδδδδδ CCCCD +++−−−++=

r18

r17

r13

r13

r17

2 RRRD

RRC ==

How to measure C (and D)?

⇒ no isotope ratios required for δ calculation

See also Poster P10

C from δ measurements only

MOW/VCONBS1917

/wCOOw45

/wCOO45

/wCONBS1945

/VSMOWOw17

O/wO17

CO/VPDBCONBS1913

COVPDB13

VSMOW17

2

222216

222216

22

2 1)1)(1(2/)1(

S

RR

−−−−

−−−

−−

−−−

−

++

+=

δδδ

δδδδ

MOW/VCONBS1917

/wCOVSMOW45

/wCOOH45

/wCONBS1945normalised

O/VSMOWH17

CO/VPDBCONBS1913

COVPDB13

VSMOW17

2

2216

2

2162

22

2 1)1(

2/)1(

S

RR

−−−

−

−−

−−−

−

+

+=

δδδ

δδδ

derived from Li et al. 1988

derived from Assonov & Brenninkmeijer 2003

Values for the ratio 17R/13R for VPDB-CO2

103 C Reference

33.8±0.1 Craig (1957)

35.7±0.7 Santrock et al. (1985)34.6±0.2 Li et al. (1988)

35.16±0.05 Assonov & Brenninkmeijer (2003)34.17±0.04 Valkiers et al. (2007)

orange: derived from δ measurements only

Influence of C on the computed 13δ value

-0.10

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

-100 -80 -60 -40 -20 0 20 4045δ /‰

Δ13δ /‰

CraigValkiersLiSantrockAssonov

Influence of 17Δ on the computed 13δ value

...528.01)1)(1( 1817528.0181717 ++≈−++= δΔδΔδ

Δδδ

Δδδ

δδδ

171845

171845

174513

0703.00371.00703.122528.0)21(

2)21(

−−≈

−×−+≈

−+=

CCCCC

17Δ(atm. O2) = –0.25 ‰ (Barkan & Luz 2000)17Δ(trop. CO2) = +0.65 ‰ (Hoag et al. 2005)

This could lead to systematic offsets in 13δbetween –0.05 and 0.02 ‰, depending on CO2 source.

Influence of the relative ionisation efficiency of 12C17O16O versus 13C16O2

Assume 13δ = –10 ‰, 18δ = 40 ‰ and 17Δ = 0 ‰.

Relative difference inionisation efficiency

45δ / ‰

0 –7.9680.001 –7.9660.01 –7.9490.1 –7.772

How good is the assumption that relative ionisation efficiencies are the same?

14th WMO/IAEA CO2 Experts Meeting (2007)accuracycloseness of agreement between a test result and the accepted reference value (VIM 2nd edition,1993).

measurement accuracy (2.13)closeness of agreement between a measured quantity value and a true quantity value of a measurand

measurement trueness (2.14)closeness of agreement between the average of an infinite number of replicate measured quantity values and a reference quantity value (5.18)

VIM 3rd edition (2008)

VIM 3rd edition (2008)

reference quantity valuereference value

quantity value used as a basis for comparison with values of quantities of the same kind

NOTE 1 A reference quantity value can be a true quantity value of a measurand, in which case it is unknown, or a conventional quantity value, in which case it is known.

Can secondary reference materials improve this situation?

Scale normalisation withL-SVEC (Li2CO3) 13δ = –46.6 ‰

(Coplen et al. 2006)

I think this is a conventional reference value.Not a true one.Nevertheless, it does improve our trueness.But not necessarily the accuracy.

13R(VPDB-CO2)

106 13R

Original (±1σ) References

11237.2 Craig (1957); Nier (1950)

IAEA Recommendation

13R(VPDB-CO2)

106 13R

Original (±1σ)

106 13R

Re-evaluated (Kaiser, 2008)

References

11237.2±30 11232.3±39 Craig (1957); Nier (1950)

13R(VPDB-CO2)

106 13R

Original (±1σ)

106 13R

Re-evaluated (Kaiser, 2008)

References

11237.2±30 11232.3±39 Craig (1957); Nier (1950)

11090.1±6 Nørgaard et al. (1999)

11101.4±9 11086.4±3 Russe et al. (2004); Coplen et al. (2006)

11179.7±28 11182.4±16 Chang & Li (1990)

11137.6±1.6 11137.6±1.6 Valkiers et al. (2007)

Average of last 4 values: (11124±45) × 10–6

18R(VPDB-CO2)

106 18R

Original (±1σ) References

2088.35 Baertschi et al. (1976); O’Neil et al. (1975)

IAEA Recommendation

18R(VPDB-CO2)

106 18R

Original (±1σ)

106 18R

Re-evaluated (Kaiser, 2008)

References

2088.35±0.80 2089.32±1.5 Baertschi et al. (1976); O’Neil et al. (1975)

18δVPDB-CO2/VSMOW = (42.0±0.7) ‰ (with 18δSLAP/VSMOW = –56.18 ‰)

Re-evaluated (Kaiser, 2008)

18δVPDB-CO2/VSMOW = 41.47 ‰ (with 18δSLAP/VSMOW = –55.5 ‰)Recommendation (IAEA, 1995)

18R(VPDB-CO2)

106 18R

Original (±1σ)

106 18R

Re-evaluated (Kaiser, 2008)

References

2088.35±0.80 2089.32±1.5 Baertschi et al. (1976); O’Neil et al. (1975)

2079.00±2.5 2078.71±2.5 Craig (1957); Nier (1950)

2107.69±2.4 Valkiers & De Bièvre (1993)

2087.54±1.6 Nørgaard et al. (1999)

2088.24±0.48 2088.24±0.48 Valkiers et al. (2007)

Average of rows 3, 5 & 6: (2088.37±0.90) × 10–6

17R(VPDB-CO2)

106 17R

Original (±1σ) References

395.11±0.94 Assonov & Brenninkmeijer (1993); Craig (1957)

IAEA Recommendation

17R(VPDB-CO2)

106 18R

Original (±1σ)

106 18R

Re-evaluated (Kaiser, 2008)

References

(395.11±0.94) (395.11±1.66) Assonov & Brenninkmeijer (1993); Craig (1957)

17R(VPDB-CO2)

106 18R

Original (±1σ)

106 18R

Re-evaluated (Kaiser, 2008)

References

(395.11±0.94) (395.11±1.66) Assonov & Brenninkmeijer (1993); Craig (1957)

379.95±0.5 380.18±0.2 Craig (1957); Nier (1950)

394.67±2 Valkiers & De Bièvre (1993)

(389.88±0.2) Nørgaard et al. (1999);Assonov & Brenninkmeijer (1993)

380.58±0.48 380.58±0.48 Valkiers et al. (2007)

Why should we care about R?

Nice to know …To make isotope measurements SI traceable[but: precision of R always worse than δ ] To cross-check with data-reduction coefficients from other molecules (O2, CO, N2O, SO2)[B, C, D, E, F, (G, H, I) – see poster P10]To callow calculation of isotopologue mixing ratios

Calculation of isotopologue mixing ratios

)1)(1()CO()OC( 181713

22

1612

RRRxx

+++=

x(CO2) = 390 μmol/mol18R = 2089.3 × 10–6

C = 0.0351613R = 11237.2 × 10–6 x(12C16O2) = 383.76 μmol/mol13R = 11086.4 × 10–6 x(12C16O2) = 383.82 μmol/mol

Δx = 0.06 μmol/mol

Is this important? You decide …

Conclusions

Agreed isobaric correction and normalisation schemes improve reproducibility (and maybe trueness) of δ values.

Several minor issues remain for the accuracy of δ valuesconfirmation of C (and D)17Δ value of tropospheric CO2

relative ionisation efficiency of 12C17O16OMajor issues remain for the accuracy of R values

13R(VPDB-CO2): converging?18R(VPDB-CO2): converged17R(VPDB-CO2): diverging [but can be linked to 13R]