The different precipitates in ASW and the NaCl medium in the absence of PO4 indicate that PO4 is not crucial for ikaite formation in ASW. It has been reported

(Bischoff et al., 1993 and Fernández-Díaz et al., 2010) that Mg2 + and SO42 − ions in seawater could also inhibit the formation of more stable phases of calcium carbonate, and thus could favor ikaite formation. This might explain why ikaite was also found in sea ice even at very low PO4 concentrations (Dieckmann et al., 2010). According to the evolution curves of log (IAP) under all the experimental conditions, we can conclude that τ is mainly controlled by the rates of log (IAP) evolution and also greatly affected by the kinetic effect, such as inhibitor ions. In the following sub-sections, the effect of experimental conditions on ikaite precipitation will focus on the factors controlling the rates of log (IAP) evolution as well as the kinetic effect. In ASW at a constant salinity of DAPT mouse 70 Bafilomycin A1 mw and temperature of 0 °C, the activity coefficients of both Ca2 + and CO32 − do not change. Therefore, we only need to focus on the change in CO32 − concentration with variations of pH. According to the calculation results from

CO2SYS, under the same conditions, the results obtained by using constants_a and constants_b show a similar trend (Fig. 6a). The increase in pH can greatly increase the CO32 − fraction in this studied pH range, resulting in a much faster approach to ikaite solubility (Fig. 5a).

However, the decrease in τ with pH is not linear, which is much faster at low pH than at high pH. This is because the CO32 − fraction cannot increase infinitely; the increase in the CO32 − fraction will slow down at high pH and the CO32 − fraction will approach 1. We can speculate that above a certain pH (depending on the salinity and temperature conditions, since the CO32 − fraction is also affected by them, as is discussed in 4.3.2 and 4.3.3), the increase in pH will not have an impact on the CO32 − fraction, and therefore has no effect on ikaite precipitation. We notice that Ω in this studied pH range increases from 3.02 to 5.37 with increasing pH (Table 2). This indicates that if the evolution of log (IAP) is slow, ikaite could be precipitated at a much Cell press lower supersaturation level. This is also confirmed by a second study, which shows that at different pumping rates of Ca2 + and DIC, Ω is low at slow pumping rates (Hu et al., submitted). The different trends in τ in ASW and the NaCl medium indicate that the effect of salinity on ikaite precipitation is not straightforward. First, according to the calculation results from CO2SYS, although there is large uncertainty in predicting the exact CO32 − fraction change with salinity at high salinities, both the results obtained from two sets of constants show a similar trend (Fig. 6b): the CO32 − fraction increases with salinity (referred to as a positive effect).