Volume 139, Issue 13, 07 October 2013

In this work, the extension of the previously developed domain based local pairnatural orbital (DLPNO) based singles and doubles coupled cluster (DLPNOCCSD) method to perturbatively include connected triple excitations is reported. The development is based on the concept of triplesnatural orbitals that span the joint space of the three pair natural orbital (PNO) spaces of the three electron pairs that are involved in the calculation of a given tripleexcitation contribution. The truncation error is very smooth and can be significantly reduced through extrapolation to the zero threshold. However, the extrapolation procedure does not improve relative energies. The overall computational effort of the method is asymptotically linear with the system size O(N). Actual linear scaling has been confirmed in test calculations on alkane chains. The accuracy of the DLPNOCCSD(T) approximation relative to semicanonical CCSD(T0) is comparable to the previously developed DLPNOCCSD method relative to canonical CCSD. Relative energies are predicted with an average error of approximately 0.5 kcal/mol for a challenging test set of medium sized organic molecules. The triples correction typically adds 30%–50% to the overall computation time. Thus, very large systems can be treated on the basis of the current implementation. In addition to the linear C150H302 (452 atoms, >8800 basis functions) we demonstrate the first CCSD(T) level calculation on an entire protein, Crambin with 644 atoms, and more than 6400 basis functions.
 ARTICLES

 Theoretical Methods and Algorithms

Natural triple excitations in local coupled cluster calculations with pair natural orbitals
View Description Hide DescriptionIn this work, the extension of the previously developed domain based local pairnatural orbital (DLPNO) based singles and doubles coupled cluster (DLPNOCCSD) method to perturbatively include connected triple excitations is reported. The development is based on the concept of triplesnatural orbitals that span the joint space of the three pair natural orbital (PNO) spaces of the three electron pairs that are involved in the calculation of a given tripleexcitation contribution. The truncation error is very smooth and can be significantly reduced through extrapolation to the zero threshold. However, the extrapolation procedure does not improve relative energies. The overall computational effort of the method is asymptotically linear with the system size O(N). Actual linear scaling has been confirmed in test calculations on alkane chains. The accuracy of the DLPNOCCSD(T) approximation relative to semicanonical CCSD(T0) is comparable to the previously developed DLPNOCCSD method relative to canonical CCSD. Relative energies are predicted with an average error of approximately 0.5 kcal/mol for a challenging test set of medium sized organic molecules. The triples correction typically adds 30%–50% to the overall computation time. Thus, very large systems can be treated on the basis of the current implementation. In addition to the linear C150H302 (452 atoms, >8800 basis functions) we demonstrate the first CCSD(T) level calculation on an entire protein, Crambin with 644 atoms, and more than 6400 basis functions.

On the large interelectronic distance behavior of the correlation factor for explicitly correlated wave functions
View Description Hide DescriptionIn currently most popular explicitly correlated electronic structure theories, the dependence of the wave function on the interelectronic distance r ij is built via the correlation factor f (r ij ). While the shortdistance behavior of this factor is well understood, little is known about the form of f (r ij ) at large r ij . In this work, we investigate the optimal form of f (r 12) on the example of the helium atom and heliumlike ions and several wellmotivated models of the wave function. Using the RayleighRitz variational principle, we derive a differential equation for f (r 12) and solve it using numerical propagation or analytic asymptotic expansion techniques. We found that for every model under consideration, f (r 12) behaves at large r ij as and obtained simple analytic expressions for the system dependent values of ρ and B. For the ground state of the heliumlike ions, the value of B is positive, so that f (r 12) diverges as r 12 tends to infinity. The numerical propagation confirms this result. When the HartreeFock orbitals, multiplied by the correlation factor, are expanded in terms of Slater functions r ^{ n } e ^{−βr }, n = 0,…,N, the numerical propagation reveals a minimum in f (r 12) with depth increasing with N. For the lowest triplet state, B is negative. Employing our analytical findings, we propose a new “rangeseparated” form of the correlation factor with the short and longrange r 12 regimes approximated by appropriate asymptotic formulas connected by a switching function. Exemplary calculations show that this new form of f (r 12) performs somewhat better than the correlation factors used thus far in the standard R12 or F12 theories.

The multilayer multiconfiguration timedependent Hartree method for bosons: Theory, implementation, and applications
View Description Hide DescriptionWe develop the multilayer multiconfiguration timedependent Hartree method for bosons (MLMCTDHB), a variational numerically exact ab initio method for studying the quantum dynamics and stationary properties of general bosonic systems. MLMCTDHB takes advantage of the permutation symmetry of identical bosons, which allows for investigations of the quantum dynamics from few to manybody systems. Moreover, the multilayer feature enables MLMCTDHB to describe mixed bosonic systems consisting of arbitrary many species. Multidimensional as well as mixeddimensional systems can be accurately and efficiently simulated via the multilayer expansion scheme. We provide a detailed account of the underlying theory and the corresponding implementation. We also demonstrate the superior performance by applying the method to the tunneling dynamics of bosonic ensembles in a onedimensional double well potential, where a singlespecies bosonic ensemble of various correlation strengths and a weakly interacting twospecies bosonic ensemble are considered.

Timeresolved photoelectron imaging spectra from nonadiabatic molecular dynamics simulations
View Description Hide DescriptionWe present an efficient method for the simulation of timeresolved photoelectron imaging (TRPEI) spectra in polyatomic molecules. Our approach combines trajectorybased molecular dynamics that account for nonadiabatic effects using surface hopping, with an approximate treatment of the photoionization process using Dyson orbitals as initial and Coulomb waves as final electron states. The method has been implemented in the frame of linear response timedependent density functional theory. As an illustration, we simulate time and energyresolved anisotropy maps for the furan molecule and compare them with recent experimental data [T. Fuji, Y.I. Suzuki, T. Horio, T. Suzuki, R. Mitrić, U. Werner, and V. BonačićKoutecký, J. Chem. Phys.133, 234303 (2010)]. Our method can be generally used for the interpretation of TRPEI experiments allowing to shed light into the fundamental photochemical processes in complex molecules.

General implementation of the resolutionoftheidentity and Cholesky representations of electron repulsion integrals within coupledcluster and equationofmotion methods: Theory and benchmarks
View Description Hide DescriptionWe present a general implementation of the resolutionoftheidentity (RI) and Cholesky decomposition (CD) representations of electron repulsion integrals within the coupledcluster with single and double substitutions (CCSD) and equationofmotion (EOM) family of methods. The CCSD and EOMCCSD equations are rewritten to eliminate the storage of the largest fourindex intermediates leading to a significant reduction in disk storage requirements, reduced I/O penalties, and, as a result, improved parallel performance. In CCSD, the number of ratedetermining contractions is also reduced; however, in EOM the number of operations is increased because the transformed integrals, which are computed once in the canonical implementation, need to be reassembled at each Davidson iteration. Nevertheless, for large jobs the effect of the increased number of ratedetermining contractions is surpassed by the significantly reduced memory and disk usage leading to a considerable speedup. Overall, for mediumsize examples, RI/CD CCSD calculations are approximately 40% faster compared with the canonical implementation, whereas timings of EOM calculations are reduced by a factor of two. More significant speedups are obtained in larger bases, i.e., more than a twofold speedup for CCSD and almost fivefold speedup for EOMEECCSD in ccpVTZ. Even more considerable speedups (67fold) are achieved by combining RI/CD with the frozen natural orbitals approach. The numeric accuracy of RI/CD approaches is benchmarked with an emphasis on energy differences. Errors in EOM excitation, ionization, and electronattachment energies are less than 0.001 eV with typical RI bases and with a 10^{−4} threshold in CD. Errors with 10^{−2} and 10^{−3} thresholds, which afford more significant computational savings, are less than 0.04 and 0.008 eV, respectively.

A hybrid stochastic hierarchy equations of motion approach to treat the low temperature dynamics of nonMarkovian open quantum systems
View Description Hide DescriptionThe hierarchical equations of motion technique has found widespread success as a tool to generate the numerically exact dynamics of nonMarkovian open quantum systems. However, its application to low temperature environments remains a serious challenge due to the need for a deep hierarchy that arises from the Matsubara expansion of the bath correlation function. Here we present a hybrid stochastic hierarchical equation of motion (sHEOM) approach that alleviates this bottleneck and leads to a numerical cost that is nearly independent of temperature. Additionally, the sHEOM method generally converges with fewer hierarchy tiers allowing for the treatment of larger systems. Benchmark calculations are presented on the dynamics of two level systems at both high and low temperatures to demonstrate the efficacy of the approach. Then the hybrid method is used to generate the exact dynamics of systems that are nearly impossible to treat by the standard hierarchy. First, exact energy transfer rates are calculated across a broad range of temperatures revealing the deviations from the Förster rates. This is followed by computations of the entanglement dynamics in a system of two qubits at low temperature spanning the weak to strong systembath coupling regimes.

Heterogeneous nucleation in multicomponent vapor on a partially wettable charged conducting particle. I. Formulation of general equations: Electrical surface and line excess quantities
View Description Hide DescriptionThermodynamics is applied to formulate general equations for internal energies and grand potential for a system consisting of a dielectric liquid nucleus of a new phase on a charged insoluble conducting sphere within a uniform macroscopic one or multicomponent mother phase. The currently available model for ioninduced nucleation assumes complete spherical symmetry of the system, implying that the seed ion is immediately surrounded by the condensing liquid from all sides. We take a step further and treat more realistic geometries, where a capshaped liquid cluster forms on the surface of the seed particle. To take into account spontaneous polarization of surface layer molecules we introduce the electrical surface and line excess quantities.

Heterogeneous nucleation in multicomponent vapor on a partially wettable charged conducting particle. II. The generalized Laplace, GibbsKelvin, and Young equations and application to nucleation
View Description Hide DescriptionBased on the results of a previous paper [M. Noppel, H. Vehkamäki, P. M. Winkler, M. Kulmala, and P. E. Wagner, J. Chem. Phys.139, 134107 (2013)], we derive a thermodynamically consistent expression for reversible or minimal work needed to form a dielectric liquid nucleus of a new phase on a charged insoluble conducting sphere within a uniform macroscopic one or multicomponent mother phase. The currently available model for ioninduced nucleation assumes complete spherical symmetry of the system, implying that the seed ion is immediately surrounded by the condensing liquid from all sides. We take a step further and treat more realistic geometries, where a capshaped liquid cluster forms on the surface of the seed particle. We derive the equilibrium conditions for such a cluster. The equalities of chemical potentials of each species between the nucleus and the vapor represent the conditions of chemical equilibrium. The generalized Young equation that relates contact angle with surface tensions, surface excess polarizations, and line tension, also containing the electrical contribution from triple line excess polarization, expresses the condition of thermodynamic equilibrium at threephase contact line. The generalized Laplace equation gives the condition of mechanical equilibrium at vaporliquid dividing surface: it relates generalized pressures in neighboring bulk phases at an interface with surface tension, excess surface polarization, and dielectric displacements in neighboring phases with two principal radii of surface curvature and curvatures of equipotential surfaces in neighboring phases at that point. We also reexpress the generalized Laplace equation as a partial differential equation, which, along with electrostatic Laplace equations for bulk phases, determines the shape of a nucleus. We derive expressions that are suitable for calculations of the size and composition of a critical nucleus (generalized version of the classical KelvinThomson equation).

A unifying modecoupling theory for transport properties of electrolyte solutions. I. General scheme and limiting laws
View Description Hide DescriptionWe develop a general method for calculating conductiondiffusion transport properties of strong electrolyte mixtures, including specific conductivities, steadystate electrophoretic mobilities, and selfdiffusion coefficients. The ions are described as charged Brownian spheres, and the solventmediated hydrodynamic interactions (HIs) are also accounted for in the noninstantaneous ion atmosphere relaxation effect. A linear response expression relating longtime partial mobilities to associated dynamic structure factors is employed in our derivation of a general mode coupling theory (MCT) method for the conductiondiffusion properties. A simplified solution scheme for the MCT method is discussed. Analytic results are obtained for transport coefficients of pointlike ions which, for very low ion concentrations, reduce to the DebyFalkenhagenOnsagerFuoss limiting law expressions. As an application, an unusual nonmonotonic concentration dependence of the polyion electrophoretic mobility in a mixture of two binary electrolytes is discussed. In addition, leadingorder extensions of the limiting law results are derived with HIs included. The present method complements a related MCT method by the authors for the electrolyte viscosity and shear relaxation function [C. ContrerasAburto and G. Nägele, J. Phys.: Condens. Matter24, 464108 (2012)], so that a unifying scheme for conductiondiffusion and viscoelastic properties is obtained. We present here the general framework of the method, illustrating its versatility for conditions where fully analytic results are obtainable. Numerical results for conductiondiffusion properties and the viscosity of concentrated electrolytes are presented in Paper II [C. Contreras Aburto and G. Nägele, J. Chem. Phys.139, 134110 (2013)].

A unifying modecoupling theory for transport properties of electrolyte solutions. II. Results for equalsized ions electrolytes
View Description Hide DescriptionOn the basis of a versatile modecoupling theory (MCT) method developed in Paper I [C. Contreras Aburto and G. Nägele, J. Chem. Phys.139, 134109 (2013)], we investigate the concentration dependence of conductiondiffusion linear transport properties for a symmetric binary electrolyte solution. The ions are treated in this method as charged Brownian spheres, and the solventmediated ionion hydrodynamic interactions are accounted for also in the ion atmosphere relaxation effect. By means of a simplified solution scheme, convenient semianalytic MCT expressions are derived for the electrophoretic mobilities, and the molar conductivity, of an electrolyte mixture with equalsized ions. These expressions reduce to the classical DebyeFalkenhagenOnsagerFuoss results in the limit of very low ion concentration. The MCT expressions are numerically evaluated for a binary electrolyte, and compared to experimental data and results by another theoretical method. Our analysis encloses, in addition, the electrolyte viscosity. To analyze the dynamic influence of the hydration shell, the significance of mixed slipstick hydrodynamic surface boundary conditions, and the effect of solvent permeability are explored. For the stick boundary condition employed in the hydrodynamic diffusivity tensors, our theoretical results for the molar conductivity and viscosity of an aqueous 1:1 electrolyte are in good overall agreement with reported experimental data for aqueous NaCl solutions, for concentrations extending even up to two molar.

Efficiency at maximum power of a chemical engine
View Description Hide DescriptionA cyclically operating chemical engine is considered that converts chemical energy into mechanical work. The working fluid is a gas of finitesized spherical particles interacting through elastic hard collisions. For a generic transport law for particle uptake and release, the efficiency at maximum power ηmp takes the form , with 1/2 a universal constant and Δμ the chemical potential difference between the particle reservoirs. The linear coefficient c is zero for engines featuring a socalled left/right symmetry or particle fluxes that are antisymmetric in the applied chemical potential difference. Remarkably, the leading constant in ηmp is nonuniversal with respect to an exceptional modification of the transport law. For a nonlinear transport model, we obtain ηmp = 1/(θ + 1), with θ > 0 the power of Δμ in the transport equation.

Solution of the Dirac Coulomb equation for heliumlike ions in the PoetTemkin model
View Description Hide DescriptionThe DiracCoulomb equation for the helium atom is studied under the restrictions of the PoetTemkin model which replaces the 1/r 12 interaction by the simplified 1/r > form. The effective reduction in the dimensionality made it possible to obtain binding energies for the singlet and triplet states in this model problem with a relative precision from 10^{−8} to 10^{−10}. The energies for the singlet state were consistent with a previous configuration interaction calculation [H. Tatewaki and Y. Watanabe, Chem. Phys.389, 58 (2011)]. Manifestations of BrownRavenhall disease were noted at higher values of nuclear charge and ultimately limited the accuracy of the PoetTemkin model energy. The energies from a nopair configuration interaction (CI) calculation (the negativeenergy states for the appropriate hydrogenlike ion were excluded from the CI expansion) were found to be different from the unrestricted Bspline calculation.

Alternative separation of exchange and correlation energies in multiconfiguration rangeseparated densityfunctional theory
View Description Hide DescriptionThe alternative separation of exchange and correlation energies proposed by Toulouse et al. [Theor. Chem. Acc.114, 305 (2005)] is explored in the context of multiconfiguration rangeseparated densityfunctional theory. The new decomposition of the shortrange exchange–correlation energy relies on the auxiliary longrange interacting wavefunction rather than the Kohn–Sham (KS) determinant. The advantage, relative to the traditional KS decomposition, is that the wavefunction part of the energy is now computed with the regular (fully interacting) Hamiltonian. One potential drawback is that, because of double counting, the wavefunction used to compute the energy cannot be obtained by minimizing the energy expression with respect to the wavefunction parameters. The problem is overcome by using shortrange optimized effective potentials (OEPs). The resulting combination of OEP techniques with wavefunction theory has been investigated in this work, at the HartreeFock (HF) and multiconfiguration selfconsistentfield (MCSCF) levels. In the HF case, an analytical expression for the energy gradient has been derived and implemented. Calculations have been performed within the shortrange local density approximation on H2, N2, Li2, and H2O. Significant improvements in binding energies are obtained with the new decomposition of the shortrange energy. The importance of optimizing the shortrange OEP at the MCSCF level when static correlation becomes significant has also been demonstrated for H2, using a finitedifference gradient. The implementation of the analytical gradient for MCSCF wavefunctions is currently in progress.

Solving the Schroedinger equation using Smolyak interpolants
View Description Hide DescriptionIn this paper, we present a new collocation method for solving the Schroedinger equation. Collocation has the advantage that it obviates integrals. All previous collocation methods have, however, the crucial disadvantage that they require solving a generalized eigenvalue problem. By combining Lagrangelike functions with a Smolyak interpolant, we device a collocation method that does not require solving a generalized eigenvalue problem. We exploit the structure of the grid to develop an efficient algorithm for evaluating the matrixvector products required to compute energy levels and wavefunctions. Energies systematically converge as the number of points and basis functions are increased.

Twocomponent Kramers restricted complete active space selfconsistent field method with relativistic effective core potential revisited: Theory, implementation, and applications to spinorbit splitting of lower pblock atoms
View Description Hide DescriptionThe relativistic twocomponent complete active space selfconsistent field theory in Kramers restricted formalism (KRCASSCF) through the framework of the spinorbit relativistic effective core potential is implemented into the KPACK package. This paper continues the development previously reported [Y. S. Kim and Y. S. Lee, J. Chem. Phys.119, 12169 (2003)] and extends the theory by means of adding timereversal symmetry into the relevant expressions so as to complete the course of theoretical development. We retained the usage of elementary spinor excitation operator for defining the spinor rotation operator and derived the gradient and Hessian in simpler forms than previously found. To eliminate redundant computation resulting from repeating sums in the derivatives, a suitable decomposition method is proposed, which also facilitates the implementation. The twostep near secondorder approach is employed for convergence. The present implementation is applicable for both closed and openshell systems and is used to calculate the atoms of lower pblock. The results for 5p and 6p are in good agreement with the experiments, and those for 7p are comparable to multireference configuration interaction results, showing that KRCASSCF is a versatile tool for the relativistic electronic structure calculation of molecules containing moderateweight through superheavy elements.
 Advanced Experimental Techniques

Highprecision drop shape analysis on inclining flat surfaces: Introduction and comparison of this special method with commercial contact angle analysis
View Description Hide DescriptionDrop shape analysis is one of the most important and frequently used methods to characterise surfaces in the scientific and industrial communities. An especially large number of studies, which use contact angle measurements to analyse surfaces, are characterised by incorrect or misdirected conclusions such as the determination of surface energies from poorly performed contact angle determinations. In particular, the characterisation of surfaces, which leads to correlations between the contact angle and other effects, must be critically validated for some publications. A large number of works exist concerning the theoretical and thermodynamic aspects of two and triphase boundaries. The linkage between theory and experiment is generally performed by an axisymmetric drop shape analysis, that is, simulations of the theoretical drop profiles by numerical integration onto a number of points of the drop meniscus (approximately 20). These methods work very well for axisymmetric profiles such as those obtained by pendant drop measurements, but in the case of a sessile drop onto real surfaces, additional unknown and misunderstood effects on the dependence of the surface must be considered. We present a special experimental and practical investigation as another way to transition from experiment to theory. This procedure was developed to be especially sensitive to small variations in the dependence of the dynamic contact angle on the surface; as a result, this procedure will allow the properties of the surface to be monitored with a higher precession and sensitivity. In this context, water drops onto a 111 silicon wafer are dynamically measured by video recording and by inclining the surface, which results in a sequence of nonaxisymmetric drops. The drop profiles are analysed by commercial software and by the developed and presented highprecision drop shape analysis. In addition to the enhanced sensitivity for contact angle determination, this analysis technique, in combination with innovative fit algorithms and data presentations, can result in enhanced reproducibility and comparability of the contact angle measurements in terms of the material characterisation in a comprehensible way.
 Atoms, Molecules, and Clusters

Exploring the dynamics of C/H and C/Cl exchanges in the C(^{3}P) + C_{2}H_{3}Cl reaction
View Description Hide DescriptionThe dynamics of the C(^{3}P) + C2H3Cl reaction at collision energy 3.8 kcal mol^{−1} was investigated in a crossed molecularbeam apparatus using synchrotron vacuumultraviolet ionization. Timeofflight spectra of products C3H2Cl, C3H3, and Cl were recorded at various laboratory scattering angles, from which translationalenergy distributions and angular distributions of product channels C3H2Cl + H and C3H3 + Cl were derived. Cl correlates satisfactorily with C3H3 in linear momentum and angular distributions, which confirms the production of C3H3 + Cl. The Hloss (Clloss) channel has average translationalenergy release 14.3 (8.8) kcal mol^{−1} corresponding to a fraction 0.30 (0.14) of available energy into the translational degrees of freedom of product HCCCHCl + H (H2CCCH + Cl). The branching ratio of channel H to channel Cl was determined approximately as 12:88. The measurements of translationalenergy releases and photoionization thresholds cannot distinguish HCCCHCl from H2CCCCl because both isomers have similar enthalpy of formation and ionization energy; nevertheless, the RiceRamspergerKasselMarcus calculation prefers HCCCHCl. The measurement of photoionization spectra identifies product C3H3 as H2CCCH (propargyl). Both products C3H2Cl + H and C3H3 + Cl might correlate to the same triplet intermediate H2CCCHCl but have distinct angular distributions; the former is nearly isotropic whereas the latter is forward biased. A comparison with the C(^{3}P) + C2H3F reaction is stated.

Dissociation of chloromethanes upon resonant σ^{*} excitation studied by xray scattering
View Description Hide DescriptionThe dissociation process following the Cl Kshell excitation to σ^{*} resonances is studied by high resolution spectroscopy of resonant elastic and inelastic xray scattering on CH3Cl, CH2Cl2, CHCl3, and CCl4 molecules. Calculations employing the transition potential and DeltaKohnSham DFT approach are in good agreement with the measured total fluorescence yield and show the presence of a second quasidegenerate group of states with σ^{*} character above the lowest σ^{*} unoccupied molecular orbital for molecules with more than one Cl atom. A bandwidth narrowing and a nonlinear dispersion behavior is extracted from the Kα spectral maps for both σ^{*} resonances. The fitted data indicate that the widths of the FranckCondon distributions for the first and second σ^{*} resonances are comparable for all the molecules under study. In addition, an asymmetric broadening of the emission peaks is observed for resonant elastic xray scattering with zero detuning on both σ^{*} resonances. This is attributed to the fast dissociation, transferring about 0.15 of the scattering probability into higher vibrational modes.

The molecular structure of and interconversion tunneling in the argoncis1,2difluoroethylene complex
View Description Hide DescriptionGuided by ab initio predictions, the structure of the gasphase complex formed between cis1,2difluoroethylene and an argon atom in a pulsed molecular jet is determined using microwave spectroscopy in the 5.7–21.5 GHz region of the spectrum. This is a nonplanar, symmetric species, with the argon atom located in the FCCF cavity of the difluoroethylene. The transitions in the microwave spectrum are observed to be split by an interconversion tunneling motion between the two equivalent configurations for the complex with the argon atom located either above or below the difluoroethylene molecular plane. Both one and twodimensional discrete variable representation calculations of the tunneling splitting using the ab initio interaction potential for the complex suggest that the barrier to interconversion is overestimated by theory.

Highresolution infrared spectroscopy of atomic bromine in solid parahydrogen and orthodeuterium
View Description Hide DescriptionThis work extends our earlier investigation of the nearinfrared absorption spectroscopy of atomic bromine (Br) trapped in solid parahydrogen (pH2) and orthodeuterium (oD2) [S. C. Kettwich, L. O. Paulson, P. L. Raston, and D. T. Anderson, J. Phys. Chem. A112, 11153 (2008)]. We report new spectroscopic observations on a series of double transitions involving excitation of the weak Bratom spinorbit (SO) transition (^{2}P1/2 ← ^{2}P3/2) in concert with phonon, rotational, vibrational, and rovibrational excitation of the solid molecular hydrogen host. Further, we utilize the rapid vapor deposition technique to produce pH2 crystals with a nonequilibrium mixture of face centered cubic (fcc) and hexagonal closed packed (hcp) crystal domains in the freshly deposited solid. Gentle annealing (T = 4.3 K) of the pH2 sample irreversibly converts the higher energy fcc crystal domains to the slightly more stable hcp structure. We follow the extent of this conversion process using the intensity of the U 1(0) transition of solid pH2 and correlate crystal structure changes with changes in the integrated intensity of Bratom absorption features. Annealing the pH2 solid causes the integrated intensity of the zerophonon Br SO transition to increase approximately 45% to a value that is 8 times larger than the gas phase value. We show that the magnitude of the increase is strongly correlated to the fraction of hcp crystal domains within the solid. Theoretical calculations presented in Paper II show that these intensity differences are caused by the different symmetries of single substitution sites for these two crystal structures. For fully annealed Bratom doped pH2 solids, where the crystal structure is nearly pure hcp, the Bratom SO transition sharpens considerably and shows evidence for resolved hyperfine structure.