{"id":"5de22a2e-5f75-4144-9919-1400e4f34310","arxiv_id":"2508.15559","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"A user-oriented review of CP2K collecting input recipes for DFT, quantum chemistry, GW/BSE, spectroscopy, and embedding methods, plus a benchmark of the UZH basis and pseudopotential protocol against all-electron FP-LAPW.","lead":"CP2K, a widely used open-source program for simulating atoms and molecules, gets a practical user's guide showing how to compute energies, spectra, and dynamics with ready-to-use input templates. The review consolidates hands-on knowledge for the package's quantum and classical methods, lowers the barrier for new users, and documents how the package's basis sets and pseudopotentials compare with all-electron reference calculations.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"UZH protocol 'highly transferable' claim rests on unary-cubic-crystal benchmark only; compound/low-symmetry transferability untested.","rationale":"The reader and I identify the same weakest assumption. The paper's central claim is descriptive, and the disclosed limitations (CPU-only HFX, k-point gaps, g-tensor normalization, SOSEX caveat) are honest. However, the single quantitative pillar—the UZH protocol benchmark—validates only unary cubic crystals. The adjective 'highly transferable' is a claim about condensed-phase generality, and the provided evidence does not support it across the chemical space the protocol is recommended for. This is not a reason to reject the review; the recipes and limitations are useful as documented. But it is a reason to request either a scope-qualifying sentence in Section 2.1.2 or a small set of binary-compound benchmark points, so that the advertised reliability matches the evidence. Hence CONDITIONAL acceptance rather than unconditional ACCEPT.","tokens_in":56293,"tokens_out":5588,"duration_ms":67285,"concrete_test":"Reproduce the Fig. 1 comparison for 8–10 binary compounds spanning ionic, covalent, and mixed bonding (e.g. MgO, LiF, GaAs, SiC, AlN, TiO2 rutile, ZrO2, MoS2) using the same Quickstep UZH TZV2P-MOLOPT/GTH and SIRIUS FP-LAPW settings, computing Eq. 7 over the same volume range. If the mean binary epsilon is within the range of the unary-crystal values shown in Fig. 1 (and stated in the paper), the transferability concern is resolved; if it is systematically larger, the claim should be narrowed to unary/high-symmetry systems or supplemented with element-/environment-specific caveats.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 2.1.2 calls the GTH PPs of the UZH protocol 'highly transferable and norm-conserving,' and the paper's only quantitative support is Fig. 1 (Eq. 7): Quickstep/TZV2P-MOLOPT+GTH versus all-electron FP-LAPW/SIRIUS for four mono-elemental cubic crystals per element up to Rn. That validation samples only unary, high-symmetry, mostly s/p/metallic bonding environments. It does not test ionic/covalent binary compounds (e.g. oxides, nitrides), low-symmetry distortions, magnetic ordering, or f-electron open shells—precisely where GTH pseudopotentials and atom-fitted MOLOPT bases are known to need special care (the paper itself notes separate lanthanide/actinide and nonlinear-core-correction PPs exist, Section 2.1.2). The paper is candid about many software limitations, but it does not qualify the transferability claim to the tested subset. If the atomic-fitted parameters fail in any of these unrepresented environments, the advertised 'UZH protocol' reliability is overstated, and the benchmark agreement in Fig. 1 is not representative. This is an external-validity gap, not an internal inconsistency; it is the most load-bearing assumption because the protocol is recommended as a general default throughout the review.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript is a practical review of the CP2K program package, organized by target property rather than by method. It covers total-energy and force methods (GPW/GAPW DFT, HFX and its ADMM/RI variants, MP2/RPA, DFT+U), band-structure methods (SIRIUS PP-PW/FP-LAPW, GW with k-point and Gamma-only variants), embedding approaches (SCCS, QM/MM, DFET), magnetic and optical spectroscopies, real-time excited-state dynamics, X-ray spectroscopy, and related workflows. The text provides numerous ready-to-use input templates and is unusually candid about implementation limits, including CPU-only HFX, missing k-points for DFT+U, and the restriction of periodic GW-k to 2D cells. The main quantitative support is the \"UZH protocol\" benchmark in Section 2.1.2, which compares Quickstep/GTH-MOLOPT equations of state against all-electron FP-LAPW/SIRIUS results using the metric of Eq. (7).","tokens_in":56477,"tokens_out":6384,"duration_ms":72104,"significance":"If the descriptions are accurate, this review will be a valuable practical reference that lowers the barrier for new CP2K users and complements the earlier theory/code paper. Its strengths include a clear property-oriented organization, a large set of concrete input templates, explicit statements of unsupported combinations, and quantitative convergence/accuracy statements for GW and BSE (e.g. 10 meV and 5 meV claims). The UZH-protocol benchmark, if properly quantified, would also provide a useful default-basis validation for condensed-phase GPW calculations. The paper does not present new methodology, but for a software review that is appropriate.","major_comments":[{"comment":"The benchmark metric as written does not remove a constant energy offset. The numerator is Σ_i [E_a(V_i) − E_b(V_i)]^2, while the denominator normalizes by the energy spreads of each curve. Since Quickstep/GTH and SIRIUS/FP-LAPW total energies have arbitrary absolute references, any constant offset C between the two methods contributes N·C² to the numerator. Unless the energies are explicitly aligned (e.g. shifted to a common minimum, or the numerator is centered by subtracting the mean difference), the metric will be dominated by the offset rather than by shape agreement of the equations of state. The text does not state any alignment procedure. Because Fig. 1 is the only quantitative support for the UZH protocol accuracy claim, this needs to be clarified or the metric corrected.","section":"Section 2.1.2, Eq. (7)"},{"comment":"The text calls the UZH-protocol GTH pseudopotentials and MOLOPT basis sets 'highly transferable' and recommends the protocol as a general default, but the quantitative validation in Fig. 1 covers only four mono-elemental cubic crystals per element up to Rn. This samples unary, high-symmetry, mostly s/p-metallic environments. It does not test ionic/covalent compounds, low-symmetry distortions, magnetic ordering, or f-electron open shells, where atom-fitted pseudopotentials and molecularly optimized basis sets are known to require special care. The paper itself notes that separate lanthanide/actinide and nonlinear-core-correction pseudopotentials exist. The transferability claim is therefore broader than the evidence. Please qualify the claim to the tested subset or add representative compound/low-symmetry benchmarks. This is an external-validity gap rather than an internal inconsistency,","section":"Section 2.1.2, UZH protocol transferability"}],"minor_comments":[{"comment":"The input example closes a subsection with '&EBD SUBSYS'; this should be '&END SUBSYS'.","section":"Section 2.2.2, RI-HFX input snippet"},{"comment":"The main text says 'a large number of different unary crystal structures', while the caption specifies 'four mono-elemental cubic crystals per element up to Rn'. Please harmonize the wording so that 'large number' is understood as the number of elements times four, not as structural diversity.","section":"Section 2.1.2, Fig. 1 caption"},{"comment":"The sentence 'k-points are not available with DFT+U, yet' is slightly awkward; consider 'k-points are not yet available with DFT+U'.","section":"Section 2.4, DFT+U"},{"comment":"The statement that a 32×32 k-point mesh 'is expected to reach convergence of the GW band gap within 10 meV for a 2D material' would be easier to use if accompanied by an example material or a reference to the specific convergence test in Ref. 125.","section":"Section 3.2.4, GW k-point sampling"}],"recommendation":"major_revision","confidential_remarks":"The paper is a broad software review, and the descriptive portions appear consistent with the method literature. The main risk is the quantitative benchmark: Eq. (7) as written is sensitive to absolute energy offsets, and the transferability claim in Section 2.1.2 goes beyond the tested unary cubic systems. Both issues are fixable in revision without changing the scope of the paper. I would not reject over these points, but they should be addressed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The review part of the paper is a solid, property-organized user guide to CP2K: it collects scattered practical knowledge into concrete input recipes, and it is more candid about implementation limits than most software papers of this kind. It flags CPU-only HFX, no k-points for DFT+U, 2D-only periodic GW-k, and even the g-tensor additivity bug and the SOSEX self-correlation caveat. That honesty is the paper's main virtue and it earns the 'made simple' title.\n\nWhat is actually new is the organization and the consolidated UZH protocol as a recommended default, plus the benchmark against all-electron FP-LAPW via SIRIUS. That benchmark is meaningful but narrow. The paper calls the UZH GTH pseudopotentials 'highly transferable and norm-conserving,' and the only quantitative support is Fig. 1: four mono-elemental cubic crystals per element up to Rn. That samples only unary, high-symmetry, mostly s/p environments. It does not test the ionic/covalent compounds, low-symmetry distortions, magnetic ordering, or f-electron open shells where transferability is usually the worry. The paper itself notes separate lanthanide/actinide and NLCC potentials exist, which undercuts the unqualified 'highly transferable' phrasing. So the external-validity gap is real, but it is a qualification problem, not a load-bearing defect. The central descriptive claim — that the documented recipes work and the listed limitations are accurate — holds.\n\nI did not execute the code or verify the accuracy numbers, and sections 9-11 (EDA, MLPs, vibrational spectroscopy) were outside my text budget, so my confidence is moderate rather than high. But for a review, the relevant bar is descriptive accuracy and honesty about limits, and the paper clears that bar. It should go to referees; the referee can ask for a one-line qualification of 'highly transferable' and a pointer to future validation on compounds. I'd cite it as the practical CP2K reference and would probably use the input snippets as starting points.","headline":"A useful, unusually candid CP2K user guide; the one real weakness is that the 'highly transferable' UZH pseudopotential claim is only benchmarked on unary cubic crystals.","tokens_in":57323,"tokens_out":2217,"would_cite":true,"duration_ms":23302,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The review claims CP2K's standard GTH/MOLOPT protocol reproduces all-electron FP-LAPW equations of state across the periodic table, while its property-organized templates cover the package's main static and dynamical methods.","keywords":["CP2K","Quickstep","Gaussian and plane wave (GPW)","GAPW","GTH pseudopotentials","MOLOPT basis sets","density functional theory","GW approximation"],"falsifier":"Run the same epsilon(a,b) comparison for a set of binary compounds (e.g., oxides, nitrides, or magnetic f-electron materials) using TZV2P MOLOPT with the corresponding GTH pseudopotentials against all-electron FP-LAPW. If the metric for several binary or low-symmetry systems exceeds the range observed for unary cubic crystals, the protocol's implied transferability to general condensed-phase environments is not supported. Alternatively, take any documented property template in the review and compare output energies or excitation energies against the stated reference values; a failure at the st","tokens_in":56046,"feed_emoji":"⚛️","tokens_out":8332,"duration_ms":87809,"temperature":0.7,"pith_summary":"CP2K is presented as a single open-source package that spans gas-phase molecules, low-dimensional materials, crystals, liquids, and soft matter, combining quantum and classical force methods with geometry optimization, transition-state search, and sampling. The review's load-bearing quantitative claim is that the standard protocol--MOLOPT Gaussian basis sets paired with GTH pseudopotentials--reproduces all-electron FP-LAPW equations of state for a large set of elements, as measured by the epsilon(a,b) metric of Eq. 7. If true, a user can take the paper's property-organized input templates as reliable starting points for total energies and forces, band structures, GW/BSE excited states, NMR/EPR parameters, X-ray spectra, and dynamical simulations. The paper is organized as a practical guide: each section introduces just enough theory, then gives a concrete CP2K input snippet.","feed_headline":"Default CP2K protocol matches all-electron solids, element by element","feed_subtitle":"The same GTH/MOLOPT recipe that matches FP-LAPW equations of state also documents GW, NMR, and excited-state dynamics.","key_machinery":"The carrying mechanism is the mixed Gaussian/plane-wave representation (GPW) inside Quickstep: Kohn-Sham orbitals are expanded in atom-centered Gaussian functions while the density is represented on a plane-wave grid, giving O(M log M) electrostatics via FFTs while keeping a compact localized basis. GAPW extends the same machinery to all-electron calculations by separating hard and soft density contributions. The accuracy argument is carried by the protocol--MOLOPT generally contracted basis sets paired with GTH norm-conserving pseudopotentials fitted to atomic scalar-relativistic reference data--and quantified by the epsilon(a,b) metric of Eq. 7, which compares energy-volume curves while re","core_discovery":"The paper's claim is that CP2K's Quickstep engine, built on the Gaussian-and-plane-wave (GPW) representation and its all-electron GAPW extension, provides one coherent framework for essentially the full range of static and dynamical atomistic properties, and that the accompanying protocol--MOLOPT basis sets plus the corresponding norm-conserving GTH pseudopotentials--makes that framework accurate out of the box. The supporting evidence is a benchmark in which Quickstep with a TZV2P MOLOPT basis and GTH pseudopotentials is compared against all-electron FP-LAPW calculations from the integrated SIRIUS code for four mono-elemental cubic crystals per element up to Rn, using the dimensionless epsi","pith_inferences":["The unary-crystal benchmark leaves open a natural extension: running the same epsilon(a,b) comparison for binary compounds, low-symmetry structures, and magnetic or f-electron systems would test whether the atomic-fitted protocol transfers to chemically richer environments; the paper does not report such tests.","The review's property-first organization implies that the method hierarchy is secondary to the observable; a user could mix, say, DFT+U for correlated oxides and RI-RPA for dispersion, but the paper does not discuss cross-method consistency checks.","Because the templates include specific convergence keywords, the recipes carry an implicit reproducibility claim: a competent user with the same inputs should obtain the stated accuracy, and a systematic regression of each template against the reported references would make that claim testable.","The GW/BSE guidance (evGW0@PBE, specific basis extrapolation) suggests a transferable protocol for molecular excitations, but the paper only validates it on selected test sets; extending to disordered or heterogeneous environments remains open."],"forward_implications":["A single input framework, organized by the property to be computed, covers isolated molecules, periodic solids, surfaces, interfaces, liquids, and amorphous systems.","The protocol gives a default starting point for DFT calculations: TZV2P MOLOPT with corresponding GTH pseudopotentials reproduces all-electron FP-LAPW energy-volume curves for the tested elements, so users do not need to revalidate baselines for those cases.","For properties that depend on core electrons, GAPW with all-electron basis sets is the required route, enabling NMR/EPR parameters, hyperfine couplings, and X-ray absorption/emission spectra.","GW-BSE and RT-TDDFT implementations put quasiparticle band structures, optical absorption, exciton descriptors, and excited-state dynamics within reach of the same code, with documented convergence settings giving agreement at the few-to-tens of meV level on molecular test sets.","Low-scaling RI-RPA/SOS-MP2 and ADMM/RI-HFX variants extend post-HF and hybrid calculations to larger periodic systems, with the paper noting the trade-offs in scaling, memory, and accuracy."],"supporting_citations":[{"why":"Define the Quickstep module and the GPW method that all energy/force and property calculations build on.","marker":"14, 22, 23"},{"why":"Introduce GAPW, the all-electron counterpart required for core-sensitive spectroscopies and small-core pseudopotentials.","marker":"24, 25"},{"why":"Supplies the molecularly optimized MOLOPT Gaussian basis sets that the protocol's accuracy depends on.","marker":"30"},{"why":"Provide the GTH pseudopotentials used in the protocol, whose transferability is asserted.","marker":"31–33"},{"why":"Describes the atomic all-electron scalar-relativistic reference calculations against which GTH parameters are fitted.","marker":"34"},{"why":"Defines the epsilon(a,b) metric and the unary-crystal comparison dataset used for the FP-LAPW benchmark.","marker":"55"},{"why":"Introduces the orbital transformation SCF method that the convergence recipes are built around.","marker":"62"},{"why":"Provides the GW100-based evidence and guidance for choosing evGW0@PBE as the GW starting point.","marker":"124"},{"why":"Describes the BSE implementation and the Thiel-set comparison giving the few-meV agreement cited.","marker":"201"}],"fun_headline_variants":["CP2K's default settings: all-electron accuracy for every solid","Same CP2K recipe: accurate for solids from hydrogen to radon","One CP2K protocol for gas, liquid, and crystal—all accurate","CP2K's out-of-box settings: no tuning, just all-electron accuracy","CP2K's default recipe: all-electron quality from H to Rn"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The whole protocol rests on parameters fitted to isolated-atom all-electron reference data, and the review assumes those fits remain accurate in arbitrary periodic condensed-phase environments, though its benchmark covers only a set of mono-elemental cubic crystals.","fun_headline_variants_meta":{"raw":{"variants":["CP2K's default settings: all-electron accuracy for every solid","Same CP2K recipe: accurate for solids from hydrogen to radon","One CP2K protocol for gas, liquid, and crystal—all accurate","CP2K's out-of-box settings: no tuning, just all-electron accuracy","CP2K's default recipe: all-electron quality from H to Rn"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000802,"raw_usage":{"total_tokens":3312,"prompt_tokens":644,"completion_tokens":2668,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":388,"completion_tokens_details":{"reasoning_tokens":2570}},"tokens_in":388,"tokens_out":2668,"duration_ms":22061,"temperature":1.0,"reasoning_tokens":2570,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T17:50:27.828185+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same epsilon(a,b) comparison for a set of binary compounds (e.g., oxides, nitrides, or magnetic f-electron materials) using TZV2P MOLOPT with the corresponding GTH pseudopotentials against all-electron FP-LAPW. If the metric for several binary or low-symmetry systems exceeds the range observed for unary cubic crystals, the protocol's implied transferability to general condensed-phase environments is not supported. Alternatively, take any documented property template in the review and compare output energies or excitation energies against the stated reference values; a failure at the st","supporting_citations":[],"review_version":1}