On capitulations of even $K$-groups and pseudo-null submodules in $\mathbb{Z}_p^d$-extensions

Technical Summary: Capitulations of Even K-Groups and Pseudo-Null Submodules in $\mathbb{Z}_p^d$-Extensions

Problem Statement
The paper investigates the relationship between the capitulation of ideals in $\mathbb{Z}_p^d$-extensions and the existence of non-trivial pseudo-null submodules within classical Iwasawa modules. Previous work by Ozaki (for $\mathbb{Z}_p$-extensions) and Fujii (for $\mathbb{Z}_p^d$-extensions with $d \ge 2$) established that the non-triviality of the maximal finite submodule of the inverse limit of ideal class groups is equivalent to the non-triviality of the kernel of the capitulation map for some finite subextension. However, Fujii's results for $d \ge 2$ required strict constraints: the base field $F$ must have only one prime above $p$, and this prime must be totally ramified in the extension.

This paper seeks to establish analogous results for even algebraic K-groups ($K_{2i}$) rather than ideal class groups. Specifically, it aims to determine if the non-triviality of the maximal pseudo-null submodule of the inverse limit of $K_{2i}$ groups ($X_{2i, F_\infty}$) is equivalent to the non-triviality of the capitulation kernel for some finite subextension, and whether these K-theoretic results can yield new sufficient conditions for the existence of pseudo-null submodules in classical Iwasawa modules without the restrictive ramification assumptions required by Fujii.

Methodology
The author employs a combination of algebraic K-theory and Iwasawa theory, leveraging the following key tools:

1. Identification with Galois Cohomology: Utilizing the $p$-adic Chern class maps and the deep results of Rost and Voevodsky, the paper identifies the $p$-primary part of even K-groups, $K_{2i}(\mathcal{O}_L)[p^\infty]$, with the second Galois cohomology group $H^2(G_S(L), \mathbb{Z}_p(i+1))$. This allows the translation of K-theoretic problems into the language of Iwasawa cohomology.
2. Iwasawa Cohomology and Descent: The inverse limit module $X_{2i, F_\infty}$ is identified with the second Iwasawa cohomology group $H^2_{Iw}(F_\infty/F, \mathbb{Z}_p(i+1))$. The proof strategy relies heavily on Tate's descent spectral sequence for Iwasawa cohomology (Proposition 2.3), which relates the cohomology of the full extension to that of subextensions.
3. Module-Theoretic Analysis: The paper analyzes the structure of finitely generated torsion modules over the Iwasawa algebra $\mathbb{Z}_p[[G]]$ (where $G \cong \mathbb{Z}_p^d$). Key lemmas (3.3–3.6) establish properties regarding pseudo-null submodules, their behavior under restriction to subgroups, and the projective dimension of these modules.
4. Galois Descent Properties: A crucial methodological distinction noted by the author is that even K-groups enjoy stronger Galois descent properties compared to ideal class groups. This property allows the author to bypass the strict constraints on the number of $p$-adic primes and their ramification that were necessary in Fujii's work on class groups.

Key Contributions and Results

• Analogue of Ozaki's Theorem ($d=1$): Theorem 1.1 (Theorem 3.1) establishes that for a $\mathbb{Z}_p$-extension $F_\infty$, the maximal pseudo-null submodule $X^0_{2i, F_\infty}$ is non-zero if and only if there exists a finite subextension $L$ such that the capitulation kernel $\ker(K_{2i}(\mathcal{O}_L)[p^\infty] \to A_{2i, F_\infty})$ is non-zero.
• Analogue of Fujii's Theorems ($d \ge 2$):
  • Theorem 1.2 (Theorem 3.2): If $X^0_{2i, F_\infty} \neq 0$ for a $\mathbb{Z}_p^d$-extension ($d \ge 2$), then the capitulation kernel is non-trivial for some finite subextension. This result holds without the assumption that $F$ has only one prime above $p$ or that it is totally ramified.
  • Theorem 1.3 (Theorem 3.7): Conversely, if $K_{2i}(\mathcal{O}_F)[p^\infty]$ is a non-trivial cyclic group and the capitulation kernel is non-trivial for some finite subextension, then $X^0_{2i, F_\infty} \neq 0$.
• Projective Dimension Result: As a byproduct (Theorem 1.4/Theorem 3.8), the paper proves that if the capitulation kernel is non-trivial for a $\mathbb{Z}_p^d$-extension ($d \ge 2$), the projective dimension of $X_{2i, F_\infty}$ as a $\mathbb{Z}_p[[G]]$-module is at least 2. This is presented as a structural observation for the second Iwasawa cohomology group that goes beyond previously known results.
• Application to Classical Iwasawa Modules: The paper applies Theorem 3.7 to derive a sufficient condition for the existence of non-trivial pseudo-null submodules in the classical Iwasawa module $X_{E_\infty}$ (where $E_\infty = F_\infty(\mu_p)$). Under the assumption that primes of $F(\zeta + \zeta^{-1})$ above $p$ do not split in $F(\mu_p)$ (Condition $S_p$), and given specific conditions on the capitulation of $K_{2i}$ groups and the cyclicity of $K_{2i}(\mathcal{O}_F)[p^\infty]$, the paper concludes that $X^0_{E_\infty} \neq 0$.

Significance and Claims
The paper claims that its primary significance lies in extending the known relationship between capitulation and pseudo-null submodules from ideal class groups to even K-groups. By utilizing the stronger Galois descent properties of K-groups, the author successfully removes the restrictive ramification conditions (single prime, total ramification) required in Fujii's theorems for $d \ge 2$.

Furthermore, the paper posits that the structural result regarding the projective dimension of $X_{2i, F_\infty}$ (Theorem 3.8) represents a novel insight into the structure of second Iwasawa cohomology groups, which are central to the Main Conjecture of Iwasawa theory but remain poorly understood in general.

The author explicitly notes modesty regarding the application to classical class groups: while the K-theoretic results provide a sufficient condition for the existence of pseudo-null submodules in $X_{E_\infty}$, the paper does not establish a direct equivalence between the capitulation of K-groups and the capitulation of ideal class groups. The relationship remains partial due to the identification of K-groups only with specific eigenspaces of class groups modulo $p$-powers, leaving the full structural relationship as an open problem for future work.