『Abstract
　In light of recent work on the reactivity of specific sites on large (hydr)oxo-molecules and the evolution of surface topography during dissolution, we examined the ability to extract molecular-scale reaction pathways from macroscopic dissolution and surface charge measurements of powdered minerals using an approach that involved regression of multiple datasets and statistical graphical analysis of model fits. The test case (far-from-equilibrium quartz dissolution from 25 to 300℃, pH 1-12, in solutions with [Na⁺]≦0.5 M) avoids the objections to this goal raised in these recent studies. The strategy was used to assess several mechanistic rate laws, and was more powerful in distinguishing between models than the statistical approaches employed previously. The best-fit model included three mechanisms - two involving hydrolysis of Si centers by H2O next to neutral (＞Si-OH⁰) and deprotonated (＞Si-O^-) silanol groups, and one involving hydrolysis of Si centers by OH^-. The model rate law is
　dSi/dt (mol/m²s) = e^-8.9±0.8 T e^{(-67.5±2.7 kJ/mol／RT)} (θ>SiOH) + e^3.6±0.7 T e^{(-82.8±2.1 kJ/mol／RT)} (θ>SiO^-) + e^6.7±1.8 T e^{(-77.5±6.0 kJ/mol／RT)} aOH^- 　　　　　　　　　　　　　　(±0.7 log units),
where θ>SiOH andθ>SiO^- are the fraction of surface silanol groups in the neutral and deprotonated forms, and aOH^- is the bulk activity of OH^-. The fitted ΔH^‡ value (67.5 kJ/mol) for the dominant low pH mechanism indicates that the model lacks a fourth mechanism involving protonation of bridging oxygens on siloxane (＞Si-O-Si＜) groups, which cannot be included because the acidity of bridging oxygens is unknown. Further progress on this and other, more complex systems requires development of more predictive and realistic models of surface speciation.』

1. Introduction
2. Background
3. Methods
　3.1. Surface speciation
　3.2. Aqueous species and their near-surface activities
　3.3. Fitting and statistical analysis
　3.4. Model evaluation
4. Results
5. Discussion
6. Conclusions
Acknowledgments
Appendix A. Supplementary data
References