Exploratory Workshop (Dec 2004): Report

DRAFT REPORT OF ACCENT ACCESS TO LABORATORY DATA EXPLORATORY WORKSHOP

Monday 13th of December 2004: Kaetsu Teaching Room, New Hall College, Cambridge

INTRODUCTION

An exploratory workshop of the ACCENT Access to Laboratory Data Work Package (WP5) was held on the 13th and 14th of December 2004, at New Hall College, Cambridge. The meeting, which was introduced by Tony Cox, had been proposed in order to explore priorities for:

(1) updating and improving access to databases, (2) data evaluation and (3) testing laboratory data in models.

The meeting was also to serve as a method for initiating collaboration between ACCENT members and with other research networks. Dr Cox gave an overview of the ACCENT project, briefly describing the joint research programme and the integration tasks. In particular, the 'Access to Data' work packages (Access to Emissions, Laboratory and Field data) were described, together with comments on the Quality Assurance and Modelling tasks.

Dr Cox then went on to describe the Access to Lab Data work package in more detail, including the programme aims and the progress so far. It was stated that the topic had been widened from its original focus of kinetic and photochemical data, to cover all physico-chemical laboratory measurements that are relevant to atmospheric chemistry.

Questions following the introduction focused on the ACCENT web portal and the use of ACCENT funds. It was agreed that the links given on the ACCENT Access to Lab Data web pages could be extended to include other data sets, as required. However, it was not clear as to how much detail should be given on the data resulting from chamber studies: should this be treated as field or laboratory data? Nevertheless, it would not be a large task to put links, to existing on-line databases from both relevant sections of the ACCENT web site. It was also noted that the web site gives links to International databases (i.e. outside of Europe).

The use of ACCENT funds for data evaluation was discussed. Whilst the evaluation itself cannot be funded, meeting costs etc. can be paid. Evaluation tasks need to be focussed, so that they are achievable.

SESSION 1: ACCESS TO GAS PHASE DATA

The Master Chemical Mechanism: An Overview Mike Jenkin

Dr Jenkin introduced the The Master Chemical Mechanism. It is employed by both users and suppliers of physico-chemical data. Data users can access the mechanism (and contextual information) and data suppliers can use the mechanism (or subsets thereof) to test the effects of their measured values. Where possible, the MCM uses measured data for rate coefficients etc. However, for the 1000 different RO2 radicals in the mechanism only 20 reactions have measured kinetic data. Hence, in all the other cases kinetic parameters and product branching ratios have been estimated. NB. For the reactions of RO2 + R'O2, a 1000 different RO2 radicals would result in 500000 reactions. Likewise, the reactions of most RO radicals have to be estimated, which requires making judgements on the branching ratio between reaction with O2, isomerisation and decomposition.

The MCM Website Andrew Rickard

Dr Rickard described the MCM dissemination through the web site (http://mcm.leeds.ac.uk/MCM/). It is possible to download box and trajectory models and browse the mechanism etc. The MCM web site, which is based on a MySQL database, has been improved with better navigation and search / extract tools. Contextual information (literature source, measurement method, parameter uncertainties etc.) is being added. The mechanism is also now available in XML and FORTRAN formats. Hence, it is more readily useable by the community (i.e. it is no longer necessary to purchase FACSIMILE). A brief demonstration on how to use the web site was given (for more details see the web site itself). A mirror site in the University of Western Australia is available as a back-up to the Leeds site.

New Internet-based UV-Vis Spectral database Geert Moortgat

Dr Moortgat described the database of UV-vis spectra that he has collated together with his colleague Dr Keller-Rudek, noting the marked changes in the capacity for data storage since the start of the project in the 1980's. Dr Keller-Rudek then gave a demonstration of the Web interface, which is planned to be publicly available in spring 2005.

In response to questions Dr Moortgat stated that the database is a compilation rather than a recommendation but that, where available, the IUPAC and JPL recommendations are given. He also acknowledged that the database does not currently include quantum yields: these will be added at a later date. Prof. Burrows stated that it would also be useful to be able to manipulate spectra (choosing resolution or making a composite spectra from different measurements made over different wavelength ranges).

Kinetic data base on reactions of oxygenated VOC Georges Le Bras

Dr Le Bras described the database of oxygenated VOC's that has been developed, by his colleague Dr Mellouki, in the EC IALSI project (http://www.gva.es/ceam/ialsi/), for use in EUPHORE experiments. It is Internet based and it is complementary to existing databases (the NIST database is no longer updated, IUPAC and the MCM are less complete for OVOC's and JPL have concentrated, at least initially, on the stratosphere). It contains rate coefficients for reactions of OVOC's with OH, O3 and NO3 at room temperature. Individual measurements and recommendations are given (where available). Future work includes expanding the number of OVOC's considered and extending the database to include (1) k as a function of T and (2) reactions of Cl with OVOC's.

This database was welcomed by the community but it was noted that it is important to carry out data evaluations (where there is sufficient data) and add the temperature dependency (as planned).

Some discussion then ensued on the topic of submission of new lab results to online databases. There appears to be a lack of availability of a Europe-wide data centre (or co-ordinated nationally funded data centres). BADC (http://badc.nerc.ac.uk/home/) act as a repository for NERC-funded data but not necessarily for EC funded projects. A Europe-wide data protocol and mechanism for data review and storage would be most useful.

IUPAC Web Site Tony Cox

Dr Cox described the work of the IUPAC subcommittee for gas kinetic data evaluation, which was set-up in 1977. Prior to 1998, the database was only available in hard copy format (published in J. Phys. Chem. Ref. Data), since then it has also been available on the Internet (http://www.iupac-kinetic.ch.cam.ac.uk/). Recently, part 1 of the database has also been published in ACP, three further parts will follow and will serve to act as an archive of the database, as it stands in 2004 / 2005. Dr Cox then described the limitations of the database. In particular, the heterogeneous processes (currently uptake coefficients) are given as a compilation only (and not an evaluation). Further work is also planned to:

" Complete further datasheets e.g. organohalogens, " Explore 'search' options, " Develop and implement protocols for the near-continuous updating of the evaluation (taking into account the needs of modellers, who may only be able to update models at larger time intervals) " Complete the hard-copy archive.

The discussion then focussed on evaluation in two areas: heterogeneous chemistry (some limited data evaluations are available from the JPL database) and SARs. It was agreed that both areas need work but resources are limited.

THALOZ Laboratory Studies John Burrows

Prof. Burrows gave a summary of some laboratory results from the EC-funded project THALOZ (http://www.atm.ch.cam.ac.uk/~thaloz/). The work involved spectroscopy and kinetics of iodine and iodine oxides, including excited state species and aerosols. The importance of the availability of high quality lab data (esp. spectra) together with a community approach to problem solving was emphasised. A data evaluation of iodine oxide chemistry has been initiated by THALOZ and it is proposed that this work should be completed, possibly within ACCENT.

After this talk, it was proposed that the ACCENT web portal should list relevant current, future and recently completed project web sites.

Laboratory work in Leuven Luc Vereecken

Dr Vereecken described the relevant activities carried out in his institute, including direct measurements of rate coefficients, reaction products and mechanisms and chemical activities. Dr Vereecken is currently working on SARs, and will publish his findings soon. He noted that the Leuven group could do the following (all of which would assist data evaluation):

" supplement experimental data; " merge different sets of data; " combine experimental and theoretical data, " extrapolate data to different conditions (T, P, reaction conditions).

Current mechanisms under consideration include terpene oxidation and reactions of oxygenates. Such work could contribute to mechanisms / databases. Contributions to data evaluation through SARs can also be made (i.e. systematic trends, exceptions and data gaps can be identified). Current SARs include: alkoxy radical decomposition, NO3 radical addition, site specific OH addition and H abstraction by OH.

The work at Leuven would interface with databases in two ways, firstly data would be imported (data mining) and secondly SAR results would be exported. Dr Vereecken noted a number of software tools would be required for this activity and that the databases would have to be suitable for interrogation using computer code (e.g. an SQL database together with an XML interface). Indexing, time stamps and versioning tools (with automatic marking of changes) would be required to ensure that the appropriate data is being accessed. Interfaces could be geared towards data evaluation (where sufficient data is available) with graphs, tables, statistics and outlier detection.

There was a general consensus that the community should move towards having low level interfaces for databases (i.e. away from the PDF data sheet).

One problem was noted with kinetic data in general: not all rate data fits the Arrhenius expression. It is difficult to know how to deal with these 'exceptions' (especially with automated systems).

Kinetic Model Framework for Aerosol Surface Chemistry and Gas-Particle Interactions: Part 1 - General Equations and Terminology Ulrich Pöschl

Dr Pöschl began his talk by reminding us that we shouldn't loose sight of the reasons for carrying out our research (i.e. health and environmental effects). He then went on to describe a simplified (yet detailed!) model of uptake and release, transformation and formation. It is based on a kinetic model framework and is designed to give consistent and universally applicable terminology. The model is required to be self-consistent (i.e. mass balance), flexible (i.e. processes can be added or condensed for computational purposes) and compatible (with existing kinetic descriptions).

Some discussion followed with users of the resistance model, which works well in many situations but shouldn't be used in all (as it is highly simplified). There was a general consensus that a more widely applicable model would be welcome. It was noted that there is a lack of experimental data in this area but Dr Pöschl considers that the required fluxes will be measured in the near future.

SESSION 2: ACCESS TO CONDENSED PHASE DATA

CAPRAM 3.0 an Aqueous Phase Model with Extended Organic Chemistry Hartmut Herrmann

Dr Herrmann introduced CAPRAM, which was developed to quickly test / include data measured in the lab (at Leipzig) in an aqueous phase model. A new version (CAPRAM 3.0) has recently been developed, which has more than quadrupled the number of organic processes, from the earlier version (2.4), which had 110 processes. The model will be made available on the Internet, once it has been published in the literature.

An example of the use of CAPRAM was shown. Field measurements at the Schmucke Mountain site indicate that the model simulates inorganic species well. In general, the model also simulates the concentration of organics in interstitial air very precisely. However, the measured cloud ontains more organics than the model would predict and the agreement is worst for species with low Henry's law coefficients. Hence, there is more work to be done!

Kinetic Model Framework for Aerosol Surface Chemistry and Gas-Particle Interactions: Part 2 - Exemplary Practical Markus Amman

Dr Amman carried on from Dr Pöschl's talk (session 1) by looking at the uptake of NO2 and O3 on soot over a long time period. The initial uptake is quicker but it rapidly falls off once the surface layer is 'full' (probably a monomolecular layer but other explanations are also possible). The uptake is presumably limited by uptake into the bulk (but here again other possibilities arise). In general it is easier to describe uptake onto the surface than it is into the bulk. Further lab measurements are required.

Data from AIDA aerosol and cloud chamber studies Ottmar Moehler

A number of different research topics are investigated at Karlsruhe. Including SOA formation, effect of coating layers on absorption of soot, optical properties of mineral dust, cloud microphysics, homogeneous and heterogeneous chemistry at low temperature, and much more besides. Data from chamber studies could be made available to interested parties within ACCENT (the chamber is also part of the access to infrastructures sub-project of ACCENT). At present no on-line database exists. However, access to results and data sets for e.g. modelling applications could be provided via the Internet. The AIDA team would be interested in contributing to discussions to harmonize data protocols and formats and to develop appropriate web portals within the ACCENT Access to Laboratory Data group.

Modelling to test laboratory data is also carried out at Karlsruhe. Dr Moehler is particularly interested in defining chemical ageing systems for aerosol in the atmosphere (this is also relevant to ACCENT aerosol and T&TP work packages).

Aerosol Inorganics Model Simon Clegg

Dr Clegg described the different versions of AIM that are currently accessible through the AIM web site (http://www.hpc1.uea.ac.uk/~e770/aim.html), where more details (including a tutorial) are available. Briefly, version 1 is purely acid, it is applicable from 180-330K and is used to investigate PSC formation; version 2 includes ammonium and is applicable over the same temperature range and version 3 includes NaCl and is applicable at 298.15K. Dr Clegg noted that it was a considerable effort to make the model accessible over the Internet but that it had been worthwhile, as it is used widely.

As part of ACCENT, Dr Clegg has tabulated the source data that he has used within AIM. The evaluations are those of Dr Clegg or others as referenced. A further objective is to add organics to the model.

It was noted, in the following discussion, that AIM is used by lab researchers to compare measurements with the model and hence it has some validation in this respect. There is a weakness at low temperatures, where fewer good lab measurements are made.

Thermodynamic modelling of mixed organic / inorganic aerosol particles Claudia Marcolli

Dr Marcolli presented her work on the thermodynamic modelling of mixed organic / inorganic aerosol particles, comparing UNIFAC simulations with lab measurements. In particular, polyols / water mixtures have been considered. Glycerol shows good agreement with the model but 2,3-butanediol shows poor agreement. It is necessary to refine the UNIFAC model.

In the following discussion it was asked if the model employed by Dr Marcolli was web-based: it is not but it could be considered as a future project. The different versions of the UNIFAC model were also discussed. It is commercially available (Dortmund consortium) but it is not particularly tailored for the atmospheric community. However, versions developed in the atmospheric science community may be behind the commercial version (as data is available later, through the relevant literature).

Data limitations in Aerosol and Cloud Equilibrium and Process Models.Gordon McFiggans

Dr McFiggans described his work to model aerosol from freshly polluted to aged polluted air, illustrating his talk with data from the NERC-funded TORCH campaign(s). Results from the UNIFAC model were compared with a model that had been improved for curved surfaces. The new model was found to be significantly better than the UNIFAC model. In particular, it is important to consider the radius dependency of organic compounds, as they have a notable effect on surface tension (i.e. organics surfactants).

Following the talk, Dr McFiggans was asked whether more measurements are required or whether systems can be reliably modelled. In his response he noted that, with more complex systems, there is insufficient data. He stated that it is now necessary to compile (and evaluate) existing data and propose the basic lab work required to give a more complete database.

Henry's Law constants Rolf Sander

Dr Sander introduced his talk by describing the different fields in which Henry's Law coefficients are used (i.e. cloud, aerosol, ocean and waste water chemistry). Unsurprisingly, different units are used within the different fields. It was suggested that representatives of different areas should meet to agree a standardisation of units and terminology. However, for existing data, Dr Sander has provided a unit converter on his web site (www.henrys-law.org), where a postscript file (> 100 pages) may be downloaded to view tabulated data for >900 species, with > 250 references.

Dr Sander invited individuals to send new papers / lab measurements for inclusion in the database.

In the following discussion Dr Pöschl asked about the conditions under which the Henry's law coefficients were obtained (i.e. how related to infinite dilution are they?). Dr Sander replied that he had collated the available measurements and if the measurements were made at high concentration, then this is noted.

Dr Cox wanted to know if 'effective' Henry's law coefficients are covered, i.e. where species react on dissolution (e.g. dissociation of acids or hydration of species such as CH2O). They are not covered under the Sander database. Hence, it was noted that other databases would be required (pKa values, data on hydration). Dr Cox pointed out that many modellers would require a 'one-stop shop'. It would be possible to link different databases from the ACCENT web site provided that sufficient information is given to assist individuals in the database use. It was also noted that there is not much data on reaction of organics in strongly acidified solutions (e.g. uptake and reaction of CH2O).

IUPAC task force on the thermochemistry of free radicals: Results to date and other data sources. Michel Rossi

Dr Rossi described an emerging database on the thermochemistry of free radicals. The database contains experimentally determined heats of formation and values determined from high-level ab initio calculations. A lot of data is available for some species (e.g. 30 pages are dedicated to CH3 alone). Others have much less data. The database will be published as a hard-copy in J Phys. Chem. Ref. Data and it will be in three parts. The first, which totals 150 pages is in press. The others will follow. All original references will be given in the publications. There are no current plans to have an on-line database but this might be a possibility for the future.

Dr Rossi invited participants to suggest sets of free radicals for detailed ab initio investigation.

Box model calculations of heterogeneous reaction systems as a gauge of their importance in atmospheric CRT models: from the stratosphere to marine aerosols.Michel Rossi

Dr Rossi gave a brief presentation on the use of box modelling to determine the impact of new lab results on the atmosphere (or inversely to show to which parameters the model is sensitive). Dr Rossi described his use of the chemkin interpreter in stratospheric and MBL scenarios. He noted that it can be used to determine whether further work (e.g. global modelling) is warranted.