Combustion of dodecane (a C12 straight-chain alkane) makes 12 CO2 molecules. A 747 jet consumes about 4 L of jet fuel per second when in flight. Based on the density and molecular weight of dodecane and the stoichiometry of its combustion reaction, you'd arrive at something like 0.01 ton of CO2 emitted per second of flight time.

4000 tons is a good estimate.

Back of the envelope...

Firstly, based off two full fuel tanks for a 777. Apparently a 777-200 has a 117350 L fuel tank, and the density of jet fuel is approximately 800 g / L.

The emitted carbon dioxide combines oxygen from the air with carbon from the jet fuel. Apparently the average chain length is 12, so it would be C_12H_10 for a straight hydrocarbon.

Factoring in the molecular masses, the amount of CO2 from two full tanks would be: (12 * 12.011 + 212.01115.999) / (1.0080 * 10 + 12 * 12.011) * 117350 * 800 * 2/ 1E6 = 643 tonnes.

Now, they might not use the full tank. Another estimate is that long haul flights use about 0.1 g / passenger / km. Apparently it is 4308 km in a straight line from SF to London, and a 777 seats about 350 people depending on the configuration, so multiplying by 2 for a return trip comes out at 302 tonnes of CO2 (that's probably an underestimate since it is a straight line).

So I think 4000 tonnes is an overestimate (and they said tons, which is even more than a metric tonne, at 4480), but the emissions of such a flight is still significant (and the global warming impact is higher because combustion also disperses water, which also has a warming effect at that altitude).

It’s about 850 US tons per full leg * 4 legs to make up the two round trips

(388 pax in a -300ER)