1 Analysis of BBCRDS Spectra: Inferred Upper Limits for Water Dimer Absorption A.J.L. Shillings 1, S.M. Ball 2 and R.L. Jones 1 1 University of Cambridge,

1

Analysis of BBCRDS Spectra: Inferred Upper Limits for Water Dimer

Absorption

A.J.L. Shillings1, S.M. Ball2 and R.L. Jones1

1University of Cambridge, Department of Chemistry, Lensfield Road, Cambridge, CB1 2EW2 University of Leicester, Department of Chemistry, Leicester, LE1 7RH

2

Outline

1. Quick overview of BBCRDS

2. BBCRDS Measurements

• Measurement rationale

• near-IR (750 nm) results

• Orange (612 nm) results

3

Traditional CRDS

))/exp(( 0 ctttII ot

Xttc o

111

•CRDS very sensitive technique - very long path lengths are achievable

•BUT: Tricky to unambiguously identify a particular species form a single measurement (Other Absorbers?)

•SOLUTION: BBCRDS combined benefits of CRDS with multi species ability of DOAS (Differential Optical Absorption Spectroscopy)

4

BB (Broadband) CRDS

30km

60km

90km

Laser Pulse

5

Spectral Regions Investigated

Orange

612 nm 16340 cm-1

Near-IR

750 nm 13400 cm-1

6

Rationale of Approach

• Record Spectra of high concentration water vapour in air at elevated temperatures.

•Calculate (explicitly treat non Beer-Lambert behaviour) and then subtract water monomer contribution to reveal ‘contiuum’ features below.

7

near-IR Database Comparison

•Subtle (but potentially significant) differences between databases

•New linelist from UCL (J.Tennyson et al) seems to do best job but some monomer structure remains.

8

Deriving Residuals from Measurements

•Fit includes calculated water monomer spectrum and a high order polynomial.

•Fitted Polynomial then added onto residual from full fit = everything apart from water monomer

9

Deriving Residuals from Measurements

•Seems to suggest UCL_08 database includes a component of the ‘continuum’ absorption in this region

•Residual from Hitran_04 broadly consistent with MT_CKD continuum – but also clearly contains errors

10

Investigation of Pfeilsticker (Science ‘03) Dimer Signal

11

near –IR Theoretical Spectra

•Position, Shape and Width of dimer feature are uncertain.

•In the following analysis, we assume a Lorentzian shape

12

Deriving Detection Limits

•Fit a Lorentzian function (variable centre and width) to the measured residual.

•Leads to an upper limit for water dimer absorption.

13

Fitted Peak Absorption to Measurement Residual

Measured

High Keq

Low Keq

eqKOHthlinestrengTheory

22

Plot above shown for Lorentzian centred at 13404 cm-1

Residual from measurement at 312K [H2O] = 1.41x1018 molecules/cm3

14

High Keq

Low Keq

•Structure results from water monomer mis-fitting.

•Yellow (S&K 03), Red (L&K 99) and Green (S&K 07) dots show lower and upper predicted peak dimer absorbance from theory.

Peak absorbance fitted to

measurement residual

Fitted Peak Absorption to Measurement Residual (2)

15

Retrieved Upper Linestrength

eqKOH

absorptionpeakthlinestreng

22

Curtiss et al (79) Keq

Low Keq

High Keq

Plot shows data for

13404 cm-1

16

Retrieved Upper Linestrength (2)

Surface calculated using Curtiss et al (79) Keq

17

Effects of Water Monomer Database

Plot shows data for

13404 cm-1

18

near-IR Conclusions

• UCL_08 database appears to give the best quality spectral fits in this spectral region.

• However, some structure due to poorly characterised water monomer absorption remains.

• Measurements suggest UCL_08 database can account for a (significant) proportion of the MT_CKD continuum in this region.

• Using this database, BBCRDS measurements give an upper limit for water dimer absorption very much at the lower end of the theoretical predictions.

19

Orange Region Theoretical Spectra

Spectra cover region in between two water absorption bands – very weak monomer absorption.

610nm = 16393 cm-1

20

Orange Database Comparison

•UCL_08 data again seems to be doing a better job, but discrepancies remain.

•Widely used Hitran (06) data appears to be very poor in this weak region between the main water bands.

•Neither database gives perfect quantitative agreement with hygrometer measurement (6.1x1018 molecules/cm3)

•UCL_08 = 6.6x1018 molecules/cm3

•Hitran 06 = 8.2x1018 molecules/cm3

21

Orange Residual Spectra

22

Fitted Peak Absorption to Measurement Residual

Plot above shown for Lorentzian centred at 16236 cm-1

Residual from measurement at 361K [H2O] = 6.1x1018 molecules/cm3

High Keq

Low Keq

eqKOHthlinestrengTheory

22

Measured

23

Fitted Peak Absorption to Measurement Residual (2)

•Structure results from water monomer mis-fitting.

•Yellow (S&K 03), Red (L&K 99) dots show lower and upper predicted peak dimer absorbance from theory.

Peak absorbance fitted to

measurement residual

High Keq

Low Keq

24

Retrieved Upper Linestrength

Curtiss et al (79) Keq

Low Keq

High Keq

eqKOH

absorptionpeakthlinestreng

22

Plot shows data for

16236 cm-1

25

Retrieved Upper Linestrength (2)

26

Effects of Water Monomer Database

Plot shows data for

16236 cm-1

27

Orange Conclusions

•UCL_08 database seems to do a better job than Hitran but discrepancies still remain.

•Despite structure introduced by water monomer database errors, BBCRDS measurements exclude the upper theoretical predictions for water dimer absorption.

•Water monomer databases contain significant discrepancies.

•Despite monomer structure in residuals, upper limit for water dimer absorption excludes the higher theoretical predictions.

•Hopefully these measurements (with their assumptions) will provide a useful constraint for further theoretical work.

Summary