Comment on "Formation of Molecular Chlorine from the Photolysis of Ozone and Aqueous Sea-Salt Particles" by K. Oum et al. (submitted to Science, 1998, rejected, click here for details)

Rolf Sander, Max-Planck Institute for Chemistry, Airchemistry Division, Postfach 3060, 55020 Mainz, Germany

In a recent publication K. Oum et al. (1) presented the experimental observation of the production of molecular chlorine via ozone photolysis in the presence of sea-salt particles. They proposed a mechanism to explain this observation and extrapolated it to atmospheric conditions. From their extrapolation they predict a chlorine production of 310 pmol/mol in 10 hours. In their mechanism Cl2 is formed via the self-reactions of the aqueous-phase chlorine radicals Cl and Cl2.

All reactions leading to chlorine radicals in their mechanism are included in our MOCCA model (2) and the results of our model runs show negligible amounts of Cl and Cl2. To analyze this discrepancy, I did some calculations based on the Oum et al. mechanism and typical conditions in the marine boundary layer (mbl).

I make the following assumptions: For the sea salt aerosol I assumed a particle radius of r = 10-6 m and a liquid water content of L = 3*10-11 m3(aq)/m3(gas). At this liquid water content, an aqueous-phase concentration of 1 M (= 1 mole/liter) is equivalent to a gas-phase mixing ratio of 720 pmol/mol. Photolysis rates were assumed to be J(H2O2) = 5*10-6 s-1 and J(O3-O(1D))=7.2*10-6 s-1. In the gas phase I assumed 35 nmol/mol O3, 1 nmol/mol H2O2 and 106 molecules/cm3 OH. The Oum et al. mechanism depends on scavenging of either O3, H2O2, or OH from the gas phase into the sea-salt particles. Each of these possibilities are analyzed separately below. In addition, I looked at the consequences of the proposed mechanism on aerosol acidity.

1) O3 scavenging: Given a Henry's law coefficient of kH(O3) = 1.3*10-2 M/atm, a gas-phase mixing ratio of 35 nmol/mol yields an aqueous-phase concentration of 4.6*10-10 M. Thus we have 4.6*10-10*720 = 3.3*10-7 pmol/mol O3 in the aqueous phase. With J(O3-O(1D)) = 7.2*10-6 s-1 the rate of aqueous-phase OH production is 7.2*10-6*3.3*10-7 = 2.4*10-12 pmol/(mol*s). This is equal to 8.5*10-8 pmol/mol over 10 hours and cannot explain Cl2 production even if aqueous-phase OH is completely converted to Cl2.

2) H2O2 scavenging: Given a Henry's law coefficient of kH(H2O2) = 105 M/atm, a gas-phase mixing ratio of 1 nmol/mol yields an aqueous-phase concentration of 10-4 M. Thus we have 10-4*720 = 7.2*10-2 pmol/mol H2O2 in the aqueous phase. With J(H2O2) = 5*10-6s-1 the rate of aqueous-phase H2O2 photolysis is 3.6*10-7 pmol/(mol*s). Assuming instantaneous conversion of H2O2 to Cl2, the production rate of Cl2 is still only 0.013 pmol/mol over 10 hours.

3) OH scavenging: Assuming infinite solubility of OH (i.e. irreversible dissolution), the scavenging rate is determined by the following equation:

dcg/dt = -L*kt*cg (1)

where cg = gas-phase concentration, t = time, and kt = transfer coefficient. The maximum transfer coefficient can be obtained assuming that gas-phase diffusion does not limit the transfer and that the accommodation coefficient is alpha = 1. Using the equation given by Schwartz (3), this leads to:

kt = 3v/(4r) = 4.5*108s-1 (2)

where v = mean molecular speed = 600 m/s for OH. Inserting cg = 106 molecules/cm3 and L = 3*10-11 yields: dcg/dt = -1.4*104 molecules/(cm3s). This is equal to -20 pmol/mol over 10 hours. An OH loss of 20 pmol/mol over 10 hours can only produce 10 pmol/mol of Cl2 over 10 hours and not 310 pmol/mol over 10 hours as postulated. Introducing estimates for the Henry's law coefficient, the accommodation coefficient, and the gas-phase diffusion constant of OH, scavenging of OH will be even slower.

4) Aerosol pH: Aqueous-phase OH is produced after scavenging of either O3, H2O2, or OH. The postulated overall reaction in the aqueous phase is: 2 OH + 2 H+ + 2 Cl- --> 2 H2O + Cl2 Thus, when 310 pmol/mol of Cl2 are produced in 10 hours, then 620 pmol/mol of H+ are consumed. At a liquid water content of L = 3*10-11, this is equivalent to almost 1 M H+. Thus, this reaction needs an aerosol pH of zero to start with. Alternatively, starting with a neutral particle, the final aerosol pH would be pH=14 if the reaction proceeds via 2 OH + 2 Cl- --> 2 OH- + Cl2.

From these calculations it looks like the laboratory results and the proposed mechanism of K. Oum et al. cannot be extrapolated to atmospheric conditions in the mbl. Although there is strong experimental evidence from field studies for active gas-phase chlorine species (see references 16-21 in (1)) the mechanism by which these are produced remains uncertain.


For helpful discussions about the mechanism I would like to thank J. Crowley, P. J. Crutzen, B. J. Finlayson-Pitts, R. von Glasow, K. Oum, and R. Vogt.


Why did Science reject this paper?

First, I would like to stress that I don't claim that there is anything wrong with the experiments conducted by Oum et al. I'm only saying that the proposed mechanism cannot be extrapolated to atmospheric conditions.

The above manuscript was submitted to Science and then rejected by the editor without any scientific explanation pretending that they didn't have space for it. I asked them why they refuse to publish a comment that points to an important error in a publication. After only 8 months (!) I received an answer saying that the rejection "does not mean that we think it is incorrect". Instead, the rejection was explained by the fact that "we judge it to be of less interest to our readers than others we received". I have thus decided to make the full text available on this web page instead. For your information my correspondence with Science is given below. I leave it up to you to make your own opinion about the prestigious journal Science.

25 June 1998: Manuscript submitted to Science.

8 Sep 1998: Manuscript rejected

Dear Dr. Sander:

Thank you for your letter of 25 June.  I regret to say we are 
not able to give space to a discussion of your ideas, as we 
receive many more items than we can accommodate. Perhaps a 
specialty journal would have more available space.

We appreciate your interest in SCIENCE.


Christine Gilbert
Letters Editor

15 September 1998: Request for reconsideration of the decision

Dear Dr. Gilbert,

Thank you for your email of 8 September. I am surprised that you reject my
technical comment on K. Oum et al. (SCIENCE, 279, 74-77, 1998) merely on the
basis of available space without any scientific evaluation. I'd appreciate it
if you could reconsider your decision taking into account the following

The Oum et al. paper has already been cited several times in the recent
literature without a critical analysis of the mechanism. See for example:
Spicer et al. (nature, 394, 353-356, 1998), Keene & Savoie (Geophys. Res.
Lett., 25, 2181-2184, 1998), Volpe et al. (Geophys. Res. Lett., in press 1998),
Clegg & Toumi (J. Geophys. Res., in press 1998), and Jacob (Atmos. Environ.,
submitted 1998). The large number of citations shows that the topic is of major
interest to the research community. If, as my calculations suggest, the Oum et
al. mechanism indeed turns out to be of negligible importance in the marine
boundary layer, SCIENCE should be the first journal to mention this.

If space is a problem, I could shorten the text and provide additional
information to interested readers upon request.

Publication in a different journal as you suggested would be unsuitable since
my manuscript is a direct comment on a recent article that was published in

        Yours sincerely
              Rolf Sander

September 1998 - May 1999: no reply from Science in spite of several inquiries

17 May 1999: An answer

Dear Dr. Sander,

I am sorry that you received no answers to your inquires about 
resubmission of your comment on a paper by K. Oum et al. We are often 
too inundated with work to answer requests for reconsideration, having 
spent considerable time making the original decision. 

Your comment was carefully evaluated the first time we received it, and 
our editors decided against publication. They agree that more work is 
required to understand Oum et al.'s observations, but they would prefer 
to see further experimentation backing up any explanations.

Please understand that we publish only a small fraction of the comments 
sent to us, and a recommendation to send a comment such as yours to a 
specialty journal does not mean that we think it is incorrect, only 
that we judge it to be of less interest to our readers than others we 

Best wishes,

Katrina Kelner
Deputy Managing Editor, Compass

Rolf Sander (01 Jun 2011)