I do not have the recent or the fifth edition of Advanced Organic Chemistry: Part B but have the forth edition (Ref.1) to answer this question. First, this book is not for for studying stereochemistry, but for studying reactions used in synthesis. In that category, this book can't be a bad book because al most all universities in US use this book for advanced post-graduate teaching.
Said that, let's look at the question in hand. I suppose the authors expect the interesting students to look at the relevant reference (Ref.2) for further learning. The given transition states and relevant diastereo-selectivity $(de)$ are given as below:
The OP's question as follows:
when the 1-substituent contains benzyloxy group stereoselectivity is less than that of siloxy group, the problem in the following drawing is that reaction with the chelated TS shows that 2,2'-syn-2,3-syn is preferred than 2,2`-anti-2,3-syn but how is that possible with steric interaction between axial-like hydrogen in the aldehyde and (R-) group?
I highlighted the main group in the question for convenience, which is the $\ce{R'}$ group in all three transition states. The so-called chelated transition states are TS-II and TS-III (Ref.3). If you look at carefully, you'd see TS-II would have substantial steric hindrance due to $\ce{R}$ group and $\ce{i-Pr}$ group. The free rotation of $\ce{CHR-OR'}$ group is restricted due to the $\ce{OR'---Ti}$ chelation. As a consequence, $\ce{R}$ group is situated just above the bulky $\ce{i-Pr}$ group. On the other hand, under the same situation with chelating of $\ce{OR'---Ti}$ in TS-III, the steric hindrance for $\ce{R}$ group is minimal because there is no group below it. The same situation dictates the non-chelated-transition state, TS-I.
When $\ce{R'}$ group is $\ce{TBDMS}$ group, the 5-membered chelate is likely not so abundant in the equilibrium (Ref.4 and references therein). For instance, it has been stated in Ref.4 that:
Alkyl substitution on oxygen is known to permit chelation with common Lewis acids, e.g., $\ce{MgBr2, SnC14}$ or $\ce{TiC14}$, whereas substitution with trialkylsilyl apparently attenuates or even eliminates the ability of an oxygen to chelate.
Thus, it is safe to assume that When $\ce{R'}$ is $\ce{TBDMS}$ group, non-chelated transition states are likely dominate the stereochemical outcome with higher syn-syn isomer.
References:
- Francis A. Carey and Richard J Sundberg, In Advanced Organic Chemistry, Part B: Reactions and Synthesis; Fourth Edition, Kluwer Academic Publishers: New York, NY, 2002 (ISBN: 0-306-46244-3).
- Sergi Figueras, Ricardo Martín, Pedro Romea, Fèlix Urpí, Jaume Vilarrasa, "Highly stereoselective aldol reactions of titanium enolates from ethyl α-silyloxyalkyl ketones," Tetrahedron Lett. 1997, 38(9), 1637-1640 (DOI: https://doi.org/10.1016/S0040-4039(97)00108-1).
- Note that according to Ref.2, the meaning of the non-chelated- and chelated-transition states mentioned here is the five-membered chelation within enolate $\ce{O}$, $\ce{OR'}$, and the metal ion. In all three transition states, the six-membered transition chelation exists.
- S. D. Kahn, G. E. Keck, W. J. Hehre, "The Effect of Protecting Groups on Chelation Control," Tetrahedron Lett. 1987, 28(3), 279-280 (DOI: https://doi.org/10.1016/S0040-4039(00)95706-X).