TURBOMOLE is a collaborative, multi-national software development project aiming to provide highly efficient and stable computational tools for quantum chemical simulations of molecules, clusters, periodic systems, and solutions. The TURBOMOLE software suite is optimized for widely available, inexpensive, and resource-efficient hardware such as multi-core workstations and small computer clusters. TURBOMOLE specializes in electronic structure methods with outstanding accuracy–cost ratio, such as density functional theory including local hybrids and the random phase approximation (RPA), GW-Bethe–Salpeter methods, second-order Møller–Plesset theory, and explicitly correlated coupled-cluster methods. TURBOMOLE is based on Gaussian basis sets and has been pivotal for the development of many fast and low-scaling algorithms in the past three decades, such as integral-direct methods, fast multipole methods, the resolution-of-the-identity approximation, imaginary frequency integration, Laplace transform, and pair natural orbital methods. This review focuses on recent additions to TURBOMOLE’s functionality, including excited-state methods, RPA and Green’s function methods, relativistic approaches, high-order molecular properties, solvation effects, and periodic systems. A variety of illustrative applications along with accuracy and timing data are discussed. Moreover, available interfaces to users as well as other software are summarized. TURBOMOLE’s current licensing, distribution, and support model are discussed, and an overview of TURBOMOLE’s development workflow is provided. Challenges such as communication and outreach, software infrastructure, and funding are highlighted.

We present a highly efficient implementation for density functional calculations of chemical shielding constants. It employs the multipole-accelerated resolution of the identity for the calculation of the Coulomb part, which complements the usage of low order scaling routines for the evaluation of the exchange-correlation part and the Hartree-Fock exchange part. Introduced errors for shifts of chemical shielding constants of H, C, F, and P are evaluated for respective test sets of molecules and are related to the accuracy of shifts obtained with hybrid and nonhybrid functionals of the generalized gradient approximation type as well as for meta-GGA functionals themselves. Efficiency is demonstrated for α-d-glucose chains with more than 2500 atoms on a single CPU as well as with an OpenMP parallelized version.

A quasi-relativistic implementation of NMR indirect spin− spin coupling constants is presented. The exact two-component (X2C) Hamiltonian and its diagonal local approximation to the unitary decoupling transformation (DLU) are utilized together with the (modified) screened nuclear spin−orbit approach. In a restricted kinetic balance, the finite nucleus model is available for both the scalar and vector potentials. The implementation supports density functionals up to the fourth rung of Jacob's ladder, i.e., (range-separated) hybrid and local hybrid functionals based on a seminumerical ansatz. We assess the quality of our quasi-relativistic X2C approach by comparison with "fully" relativistic four-component results for small main-group molecules and alkynyl compounds. The mean absolute error introduced by the DLU scheme is less than 0.05 × 10 19 T J −2 of the reduced coupling constant for the small main-group molecules and 0.5 Hz for the alkynyl compounds. Thus, the error is significantly smaller than finite nucleus size effects for heavy elements. The basis set convergence and the impact of different density functional approximations are further studied. We propose a simple scheme to develop segmented-contracted relativistic all-electron basis sets for NMR spin−spin couplings. Our implementation allows us to perform calculations of extended molecules with reasonable computational effort, which is illustrated for the 1 J( 119 Sn, 31 P) coupling constant of a low-valent tin phosphinidenide complex. The corresponding results are in good agreement with the experimental findings.

We present an efficient implementation for the computation of nuclear spin–spin coupling tensors within density functional theory into the TURBOMOLE software suite. Emphasis is put on methods to efficiently evaluate the Hartree–Fock exchange needed for hybrid functionals: resolution of the identity and seminumerical evaluation on a grid. Our algorithm allows for the selection of specific nuclei for the reduction of calculation times. Further, the accuracy of locally dense basis sets in the density functional theory framework is investigated. These features allow for the routine computation of coupling constants in systems comprising about 100 carbon atoms within less than one day on a single CPU and within a few hours when using the OpenMP variant. Based on seminumerical integration, the first implementation of local hybrid functionals for spin–spin couplings is reported. This has allowed a preliminary evaluation of position-dependent exact-exchange admixture in three local hybrid functionals for a set of 80 isotropic spin–spin couplings in 23 small main-group molecules against CC3 and MCSCF reference data. Two of the local hybrids (LH14t-calPBE and LH07t-SVWN) are the top performers in the overall statistical evaluation compared to several standard functionals (TPSS, TPSSh, B3LYP, PBE0, and BHLYP), in particular, as they do not exhibit notable outliers for specific coupling types.

We present the first steps to extend the Green’s function GW method and the Bethe–Salpeter equation (BSE) to molecular response properties such as nuclear magnetic resonance (NMR) indirect spin–spin coupling constants. We discuss both a nonrelativistic one-component and a quasi-relativistic two-component formalism. The latter describes scalar-relativistic and spin–orbit effects and allows us to study heavy-element systems with reasonable accuracy. Efficiency is maintained by the application of the resolution of the identity approximation throughout. The performance is demonstrated using conventional central processing units (CPUs) and modern graphics processing units (GPUs) for molecules involving several thousand basis functions. Our results show that a large amount of Hartree–Fock exchange is vital to provide a sufficient Kohn–Sham starting point to compute the GW quasi-particle energies. As the GW-BSE approach is generally less accurate for triplet excitations or related properties such as the Fermi-contact interaction, the admixture of the Kohn–Sham correlation kernel through the contracted BSE (cBSE) method improves the results for NMR coupling constants. This leads to remarkable results when combined with the eigenvalue-only self-consistent variant (evGW) and Becke’s half and half functional (BH&HLYP) or the CAM-QTP family. The developed methodology is used to calculate the Karplus curve of tin molecules, illustrating its applicability to extended chemically relevant molecules. Here, the GW-cBSE method improves upon the chosen BH&HLYP Kohn–Sham starting points.