Arxiv:2001.01120V1 [Physics.Chem-Ph] 4 Jan 2020 Ufiepouto Spooe Oocrtruhatwo-Electr Lit a for Through Mechanism Occur the to Cess

arXiv:2001.01120v1 [physics.chem-ph] 4 Jan 2020 ufiepouto spooe oocrtruhatwo-electr lit a for through mechanism occur The to cess. proposed reversed. is production is sulfide chemistry produc the sulfide charge, lithium in ing resulting ultimately ch lengths, containing di ing salts During sulfur into battery. reduced lithium-sulfur is a sulfur elemental of chemistry The 1. FIG. t alternative chemistries. practical lithium-ion a of-the-art as use ba potential its lithium-sulfur limits of and lithium ies realization the the provi to for intermediates challenges diffuse various unique can of solubility and The electrolyte anode. whic metal the of se- in some a salts, soluble through polysulfide are battery, lithium the intermediate in of discharge ries on proceed to known (Li metal lithium between trochemistry rdc,a lutae nFigure in illustrated as product, [ is density energy t theoretical (Li-ion up Wh/kg of capacity theoreti-cal high ∼ a with technology battery u (S fur 1 65mhgadhg hoeia nrydniyof density energy theoretical high and mAh/g 1675 – nodrt ans oto vrtebteytcnlg,the technology, battery the over control harness to order In ihu-ufrbteisaeapoiignext-generation promising a are batteries Lithium-sulfur 3 .Tervril atr ucinn sbsdo h elec- the on based is functioning battery reversible The ]. 8 ofr ihu ufie(Li sulfide lithium form to ) unu hmsr iuain fDmnn rdcsi Lith in Products Dominant of Simulations Chemistry Quantum eto u nweg fdpl oetcluain en pe f being using calculations device moment Valencia dipole Q of an IBM knowledge the molecule, our on of salt best molecule sulfur LiH lithium the a of for moment simulator quantum a on n ihu ufie(Li sulfide lithium and ioemmnsaerpre o ihu yrd LH,hydr (LiH), hydride lithium for reported are moments dipole unu hmsr iuain ffu nutilyreleva industrially four of simulations chemistry Quantum 3 ame G(ecdsBn as,GopRsac,HCG012- HPC Research, Group Cars), (Mercedes-Benz AG Daimler INTRODUCTION 2 ecdsBn eerhadDvlpetNrhAeia Sunn America, North Development and Research Benz Mercedes 1 B lae eerhCne,60HryRa,SnJs,C 951 CA Jose, San Road, Harry 650 Center, Research Almaden IBM ui .Rice, E. Julia ai Motta, Mario 1 2 h euto fslu is sulfur of reduction The . ) een edmntaeteaiiyt aclt ioem dipole calculate to ability the demonstrate we Herein, S). 2 )a h thermodynamic the as S) 0 n lmna sul- elemental and ) 1 ∼ av .Gujarati, P. Tanvi 1 5-0 Wh/kg) 350-500 nra Hintennach, Andreas iso vary- of ains in Dur- tion. scharge, state- o ∼ npro- on 2600 hium tter- des h o 1 ye .Takeshita, Y. Tyler qainhv enpooe,i atclrteVariational the particular [ in (VQE) Eigensolver proposed, Quantum been Schr¨odin the have to equation solutions accurate highly but approximate tecigfrLH H bond deter- LiH, along for moments in stretching dipole algorithms and energies quantum state of ground mining performance w the this in assess properties, we electrostatic and energies computing including areas, [ of chemistry number quantum a in computing advanta conventional provide over to potential enormous an has that problems, unl introduced. are computers, schemes classical approximation for expensive kno is exponentially that be problem Schr¨odin a to the interest, of solving molecules the requires for goal equation this Achieving impor- LiS of conceptual chemistry teries. deep the to with applicability problem significant and a tance, thus is molecules moment electr dipole of a external and energetics an of of computation to moment Accurate response field. dipole its The determines re- also transfer. molecule or re- electron attract generate and through easily gen- compounds actions other can In from polarity electrons batteries. valence high pel in with occurring molecules phenomena eral, of underst variety for critical a are ing moment, dipole their particularly eosrtdteueadacrc fVEi investigations [ in molecules of VQE variety of a accuracy of and use the demonstrated rdbtaeulkl rcse steei iteexperimen [ little reasons is entropic there for as and evidence processes consid 1-8 unlikely been of also are consists have but anion processes ered sulfur electron Four the which atoms. radi in sulfur generate salts, which Li-S processes, and electron cals two or eithe though produced. electron produced one products be to of assumed are mixture products Generally, the also confounds are which pathways active, electrochemical) (vs chemical batter the competing of operation (if during difficult characterize are to Intermediates impossible) field. co the highly in is debated environment and battery plex the in reduction sulfur mechanis for reaction the However, validated. int be reactive must of mediates identification and production for mechanism gnsld (H sulfide ogen tmlclsaerpre.Dsoito uvsand curves Dissociation reported. are molecules nt 3 oiae yteescess n yteiprac of importance the by and successes, these by Motivated mathematical of attack of mode a is computing Quantum and molecules, of distribution density electron-cloud The fre nqatmhardware. quantum on rformed eosrt h blt ocluaetedipole the calculate to ability the demonstrate d n enet .Garcia M. Jeannette and u uis hsi h rteapet the to example first the is This qubits. our B 13,Belne,Germany Boeblingen, 71034, BB, 2 ) ihu yrgnsld (LiSH) sulfide hydrogen lithium S), 2 o Latone, Joe vl,C 94085 CA yvale, 5 2 mnsuigu o2 qubits 21 to up using oments ,LS n Li and LiSH S, – 7 .Anme fhuitc oprovide to heuristics of number A ]. u-ufrBatteries ium-Sulfur 0 USA 20, 11 ]. 1 8 1 – 10 4 ]. .IMrsaceshave researchers IBM ]. 2 S. and y and- bat- ork not ger ger ess wn m- er- tal ge m ic s - r - 2

The calculations presented here are important for the devel- quantum processor, Valencia, where we employed readout- opment of quantum computing: to the best of our knowledge, error mitigation using Qiskit Ignis to correct measurement er- we present the first hard-ware simulation of dipole moments. rors. We also used a simple noise extrapolation scheme using Our results immediately generalize to expectation values of k- additional CNOT gates at the minimum energyVQE iterations body operators in materials, with significant implications for to account for errors introduced during the expensive 2-qubit the ability to study reactivity and electrostatics in batteries by entangling operations as was shown in Refs [24, 25]. quantum algorithms.

RESULTS AND DISCUSSION METHODOLOGY Quantum computing simulations on lithium-sulfur molecules The overall strategy for the calculations performed in this work involved initial preprocessing by classical quantum chemistry codes on conventional computers, to generate op- RHF CCSD q-UCCSD FCI timized mean-field orbitals and matrix elements of the Hamil- -7.3 0.3 A B tonian prior to performing computations with quantum simu- -7.4 0.2

lators or devices. The restricted Hartree-Fock (RHF) singlet -7.5 0.1 Ha] m state has been chosen as the initial state for all of the calcula- [ [Ha] -7.6 0.0 FCI E tions described here, since experience has indicated this state E -7.7 -0.1 − E as a good choice for a variety of chemical problems [12]. Re- -7.8 -0.2 stricted coupled-cluster with singles and doubles (CCSD) cal- -7.9 -0.3 culations were performed using Psi4 [13], at STO-3G level C D 5.0 0.2

of theory, using the frozen-core approximation for correlated ]

4.0 0.1 D [ calculations. The choice of the minimal basis is motivated by ] FCI D [ 3.0 0.0 the fact that only a few molecular orbitals can be simulated on D D 2.0 -0.1 − contemporary quantum hardware, since the number of avail- D able qubits is still relatively small, devices are noisy, and full 1.0 -0.2 quantum error correction techniques are not yet available. Ad- 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 -0.3 ˚ ˚ ditional details for the studied molecules are listed in Table I RLiH A RLiH A in the Supplemental Information. Having selected a set of single-electron orbitals for each of FIG. 2. Fig. 2 Calculations for lithium hydride, LiH. (A) Disso- ciation curve (ground state energy in Hartrees, as a function of in- the studied species, VQE computations were performed with teratomic distance, in angstroms) from RHF, CCSD, q-UCCSD and quantum simulators and devices. We use IBM’s open-source FCI. (B) Deviation from FCI energy, for CCSD and q-UCCSD, in Python library for quantum computing, Qiskit [14]. Qiskit millihartrees. (C) Dipole moment (in Debye) as a function of inter- provides tools for various tasks such as creating quantum cir- atomic distance, in angstroms). (D) Deviation from FCI dipole mo- cuits, performing simulations, and computations on real hard- ment, for CCSD and q-UCCSD, in Debye. Ha and D denote Hartree ware. Qiskit Aqua contains an implementation of the VQE and Debye respectively. algorithm, a hybrid quantum-classical algorithm that uses both quantum and classical resources to solve the Schr¨odinger The lithium hydride (LiH) bond can be represented by equation. It also provides a classical exact eigensolver algo- molecular (spatial) orbitals (MOs) with 2s, 2pz character for rithm (FCI) to compare results. lithium and an MO with 1s character for hydrogen, for a total In the VQE algorithm, we take our wavefunction in the of 6 spin-orbitals. The ground state energy and dipole mo- form of a quantum circuit, which is the unitary coupled cluster ments were calculated on the simulator through the use of with singles and doubles (UCCSD) [15–17], its quantum vari- six qubits and the hardware with four qubits after tapering ant (q-UCCSD) as defined in Ref [18] and the Ry variational [26, 27] techniques were applied. form provided in Qiskit [14]. As seen in Figure 2, results from CCSD, q-UCCSD and We then minimize the expectation value of the Hamiltonian FCI coincide for both energy and dipole moment along bond with respect to the parameters of our circuit. The minimiza- stretching, due to the fact that CCSD and q-UCCSD are equiv- tion is carried out through the classical optimization method, alent to FCI for two-electron systems. In the case of q- L BFGS B[19–21] on the simulator, and Simultaneous Per- UCCSD some small deviations from FCI are observed. Of turbation Stochastic Approximation (SPSA) [22, 23] on the course, with molecules that contain more than two electrons, device. Once the VQE is complete, we get the optimized vari- different results are observed for energies and other proper- ational form and the upper bound for the ground state energy. ties. Despite its simplicity, LiH has an interesting evolution in We ran our experiments on both the statevector and qasm the dipole moment along the dissociation curve from a regime simulators in Qiskit. We performed hardware experiments on with ionic character (polar, short R) to one without polarity IBMs Quantum Computation Centers new T-shaped 5-qubit (large R). 3

RHF CCSD q-UCCSD FCI estimates for energy and dipole are similar in accuracy along -394.0 0.3 A B bond stretching. 0.2 -394.1

0.1 Ha] RHF CCSD q-UCCSD FCI m [ -407.65 12.0

[Ha] -394.2 0.0 FCI A B E E 8.0 -0.1 − -407.75 -394.3 E -0.2 4.0 Ha] m [

[Ha] -407.85 0.0 -394.4 -0.3 FCI E C D E 0.2 -4.0 − 1.0 -407.95 E

] -8.0

0.1 D [

] 0.9 -408.05 -12.0 FCI D [ 0.0

D C D

D 0.8 − 0.2 -0.1 4.0 D ]

0.1 D 0.7 -0.2 [ ] 3.0 FCI D [ 0.0 0.6 -0.3 D 1.0 1.5 2.0 2.5 3.0 3.5 1.0 1.5 2.0 2.5 3.0 3.5 D 2.0 ˚ ˚ -0.1 − RSH A RSH A D 1.0 -0.2

FIG. 3. Fig. 3 Calculations for hydrogen sulfide, H2S. (A) Dis- 0.0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -0.3 sociation curve (ground state energy in Hartrees as a function of in- RLiS A˚ RLiS A˚ teratomic distance, in angstroms) from RHF, CCSD, q-UCCSD and FCI. (B) Deviation from FCI energy, for CCSD and q-UCCSD, in FIG. 4. Calculations for lithium sulfide, Li2S. (A) Dissocia- millihartrees. (C) Dipole moment (in Debye) as a function of in- tion curve (ground state energy in Hartrees as a function of inter- teratomic distance, in angstroms). (D) Deviation from FCI dipole, atomic distance, in angstroms) from RHF, CCSD, q-UCCSD and for CCSD and q-UCCSD, in Debye. Ha and D denote Hartree and FCI. (B) Deviation from FCI energy, for CCSD and q-UCCSD, in Debye respectively. millihartrees. (C) Dipole moment (in Debye) as a function of in- teratomic distance, in angstroms). (D) Deviation from FCI dipole, for CCSD and q-UCCSD, in Debye. Ha and D denote Hartree and Debye respectively. Hydrogen sulfide (H2S) is represented by MOs with 3s, 3px, 3py, 3pz, character for sulfur and an MO with 1s char- acter for hydrogen, totaling 12 spin orbitals and 8 electrons. Lithium hydrogen sulfide (LiSH) is represented by MOs After employing tapering and qubit reductions to exploit sym- with 3s, 3px, 3py, 3pz, character for sulfur, 1s character for metries in the molecule, the ground state energy and dipole hydrogen and 1s, 2s, 2px, 2py, 2pz, character for lithium, for moments were calculated on the simulator with 9 qubits. In a total of 18 spin orbitals and 8 electrons. After employing ta- Figure 3 we study the ground-state energy and dipole moment pering and qubit reductionspossible due to symmetry[26, 27], along bond stretching of H2S, using CCSD, q-UCCSD and the ground state energy and dipole moments were calculated FCI. q-UCCSD and CCSD are in agreement with FCI across on the simulator with 15 qubits. bond stretching (their maximum deviations from FCI being Results are shown in Figures 5 and 6. In Figure 5, the dis- 0.13 and 0.8 millihartree respectively), with q-UCCSD pro- sociation of the LiS bond is studied by varying RLiS with all viding a better estimate of the energy for large RHS. The other internal coordinates fixed at their experimental [28] val- dipole moment is seen to monotonically decrease with RHS ues. For all but the largest values of RLiS, CCSD is in better towards ∼0.6 D, and q-UCCSD values are in better agreement agreement with FCI for both the energy and the dipole mo- with FCI than CCSD ones (their maximum deviations from ment. As in the H2S case, the most challenging regime is the FCI being 1.7 millidebye and 11.5 millidebye respectively). intermediate dissociation region, where the energy is overes- As seen in Figure 3, differences between q-UCCSD and timated by 5 millihartrees and the dipole moment by ∼20%. FCI energies are always greater than zero. This can be under- A very different situation is seen for the dissociation of stood, since the q-UCCSD energies are variational in nature the covalent SH bond, illustrated in Figure 6. Description of [15–17] (only include singles and doubles) and are therefore the dissociation regime is more challenging for both CC fla- above the FCI values. It is also understandable that the most vors than that seen for the breaking of the Li-S bond, with challenging regime is the intermediate dissociation regime, q-UCCSD again performing better at large bond length. The where the ground-state wavefunction switches between differ- FCI dipole moment decreases monotonically with RSH, a ten- ent dominant determinants in its configuration interaction ex- dency that both CC flavors correctly reproduce at a qualita- pansion, exhibiting multireference character. Lithium sulfide tively, but not quantitatively for large bond length. In both (Li2S) is represented by MOs with 3s, 3px, 3py, 3pz, charac- cases, deviations from FCI signal the limited accuracy of the ter for sulfur and 2s, 2px, 2py, 2pz, character for lithium, for underlying ansatz, and take place in the intermediate dissoci- a total of 24 spin orbitals and 8 electrons. After employing ta- ation regime. pering and qubit reductions to account for molecular symme- Comparison between hydrogen-sulfur and lithium-sulfur tries, the ground state energy and dipole moments were cal- bond stretching in LiSH reveals interestingly different be- culated on the simulator with 21 qubits. UCCSD and CCSD haviors. In particular, when the dipole moment is plotted 4

RHF CCSD q-UCCSD FCI Quantum chemistry simulations on quantum devices -401.02 6.0 A B 4.0 -401.06

2.0 Ha]

m 3,raw 3,ext [ RHF FCI Ry Ry

[Ha] -401.10 0.0 FCI E

E -7.4

-2.0 −

-401.14 E -4.0 -7.5 -401.18 -6.0 C D -7.6 5.0 0.6 [Ha] ] E

4.0 0.3 D

[ -7.7 ] FCI D [ 3.0 -0.0 D

D -7.8 2.0 -0.3 − D 1.0 -0.6 -7.9 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 RLiH A˚ 0.01.0 2.0 3.0 4.0 5.0 6.0 1.0 2.0 3.0 4.0 5.0 6.0 -0.9 RLiS A˚ RLiS A˚ FIG. 7. Hardware-simulated ground-state energy calculations for FIG. 5. Simulations of lithium hydrogen sulfide, LiSH, when break- LiH on the IBM Q Valencia device (energy in Hartrees as a func- ing the Li-S bond. (A) Dissociation curve (ground state energy in tion of interatomic distance, in angstroms). Ha denotes Hartree. Hartrees as a function of interatomic distance, in angstroms) from RHF, CCSD, q-UCCSD and FCI. (B) Deviation from FCI energy, for CCSD and q-UCCSD, in millihartrees. (C) Dipole moment (in 3,raw 3,ext Debye) as a function of interatomic distance, in angstroms). (D) De- RHF FCI Ry Ry viation from FCI dipole, CCSD and q-UCCSD, in Debye. Ha and D 6.0 denote Hartree and Debye respectively. 5.0

4.0 RHF CCSD q-UCCSD FCI ]

-400.95 18.0 D A B [ 3.0 D -401.00 12.0 2.0

6.0 Ha] m

-401.05 [ 1.0

[Ha] 0.0 FCI E E -401.10

-6.0 − 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 -401.15 E -12.0 RLiH A˚

-401.20 -18.0 C D 0.6 FIG. 8. Hardware-simulated dipole moment calculations for LiH on 2.6 ] the IBM Q Valencia device (dipole moment in Debye, as a function

0.3 D [

] of interatomic distance, in angstroms). D denotes Debye. FCI D [ 2.2 -0.0 D D -0.3 − 1.8 D -0.6 Next, the dipole moment of lithium hydride was evaluated using quantum hardware as described in the Methods section. 1.4 1.0 2.0 3.0 4.0 5.0 1.0 2.0 3.0 4.0 5.0 -0.9 RSH A˚ RSH A˚ We employ the hardware-efficient Ry ansatz to estimate the ground-state energy and dipole moment at three representa- FIG. 6. Simulations of lithium hydrogen sulfide, LiSH, when break- tive values of R. We employed a depth=3 circuit with CX ing the S-H bond. (A) Dissociation curve (ground state energy in entangling gates with linear connectivity. To the best of our Hartrees as a function of interatomic distance, in angstroms) from knowledge, these are the first evaluations of dipole moments RHF, CCSD, q-UCCSD and FCI. (B) Deviation from FCI energy, by quantum hardware simulations. for CCSD and q-UCCSD, in millihartrees. (C) Dipole moment (in Debye) as a function of interatomic distance, in angstroms). (D) De- We chose to study four bond lengths representative of the viation from FCI dipole, CCSD and q-UCCSD, in Debye. Ha and D polar and dissociation regimes. Results from hardware simu- denote Hartree and Debye respectively. lations in presence of noise are shown. A Richardson extrap- olation [29, 30] is also conducted, with the aim of mitigating the impact of noise. As seen in Figures 7 and 8, the qualitative against the interatomic distance, very different behaviors are behavior of both energies and dipoles is correctly captured by observed. For instance, no maximum value is observed be- the hardware simulations upon extrapolation. In particular, the tween ∼2.5-3.0 Angstroms for S-H bond stretching; instead, dipole moment is accurate within ∼10% in the polar regime. a change in slope closer to ∼2.0 Angstroms is visible. We Future research will explore approaches to improve these re- reason that the differences observed result from a difference sults quantitatively, for example by the use of circuits with in the nature of the bond as Li-S is ionic in character whereas higher depth, by the exploration of multiple ansatze or by the S-H is covalent. use of different error mitigation techniques. 5

CONCLUSIONS

Herein, we have reported on the simulations of four molecules (LiH, H2S, LiSH and Li2S) that increase in com- plexity to Li2S which is relevant to the study of lithium-sulfur batteries. We showed that we could obtain ground state ener- gies for each of these molecules within chemical accuracy on a quantum simulator using 6, 9, 15, and 21 qubits respectively. Preliminary results on lithium sulfide (Li2S) are included.

Differences could be observed in the LiSH molecule de- pending on whether the hydrogen or lithium atom was disso- ciated from the sulfur atom. We hypothesize that these differ- ences can be explaineddue to the differencein ionic compared to covalent character of the bond. Furthermore, the effects of electronic distribution changes could be observed when eval- uating the dipole moment of lithium hydride. Additionally, we showed that the lithium hydride dipole moment could be qualitatively simulated on quantum hardware with four qubits.

This is a notable demonstration of the capabilities of the hardware since the dipole moment of lithium hydride changes from strongly polarized in its ionic state to effectively neu- tral at long bond distances over ∼2.5 A,˚ representing a highly entangled state at this distance.

In the future, these methods can be applied to larger calcu- lations on polyanions formed upon discharge in the Li-S bat- tery to help evaluate between radical and ionic mechanisms of the electrochemical reduction of sulfur to lithium sulfide with lithium metal.

ACKNOWLEDGMENTS

The authors gratefully acknowledge helpful discussions with Eunseok Lee (MBRDNA), as well as Donny Greenberg, Marco Pistoia, and Stephen Wood (IBM). 6

J. J. Wilke, M. L. Abrams, N. J. Russ, M. L. Leininger, C. L. Janssen, E. T. Seidl, W. D. Allen, H. F. Schaefer, R. A. King, E. F. Valeev, C. D. Sherrill, and T. D. Crawford, WIREs Com- [1] R. Van Noorden, Nature 498, 416 (2013). putational Molecular Science 2, 556 (2012). [2] D. Zheng, X. Zhang, J. Wang, D. Qu, X. Yang, and D. Qu, [14] G. Aleksandrowicz, T. Alexander, P. Barkoutsos, L. Bello, Journal of Power Sources 301, 312 (2016). Y. Ben-Haim, D. Bucher, F. Cabrera-Hern´andez, J. Carballo- [3] P. Gibot, M. Casas-Cabanas, L. Laffont, S. Levasseur, P. Car- Franquis, A. Chen, C. Chen, et al., Zenodo 16 (2019). lach, S. Hamelet, J.-M. Tarascon, and C. Masquelier, Nature [15] W. Kutzelnigg, The Journal of Chemical Physics 77, 3081 Materials 7, 741 (2008). (1982). [4] K. H. Wujcik, D. R. Wang, A. A. Teran, E. Nasybulin, T. A. [16] W. Kutzelnigg and S. Koch, The Journal of Chemical Physics Pascal, D. Prendergast, and N. P. Balsara, “Determination of re- 79, 4315 (1983). dox reaction mechanisms in lithiumsulfur batteries,” in Electro- [17] W. Kutzelnigg, The Journal of Chemical Physics 82, 4166 chemical Engineering (John Wiley & Sons, Ltd, 2018) Chap. 3, (1985). pp. 41–74. [18] P. K. Barkoutsos, J. F. Gonthier, I. Sokolov, N. Moll, G. Salis, [5] I. M. Georgescu, S. Ashhab, and F. Nori, Rev. Mod. Phys. 86, A. Fuhrer, M. Ganzhorn, D. J. Egger, M. Troyer, A. Mezzacapo, 153 (2014). S. Filipp, and I. Tavernelli, Phys. Rev. A 98, 022322 (2018). [6] I. Kassal, J. D. Whitfield, A. Perdomo-Ortiz, M.-H. Yung, and [19] R. H. Byrd, P. Lu, J. Nocedal, and C. Zhu, SIAM Journal on A. Aspuru-Guzik, Annual Review of Physical Chemistry 62, Scientific Computing 16, 1190 (1995). 185 (2011), pMID: 21166541. [20] C. Zhu, R. H. Byrd, P. Lu, and J. Nocedal, ACM Trans. Math. [7] Y. Cao, J. Romero, J. P. Olson, M. Degroote, P. D. Johnson, Softw. 23, 550560 (1997). M. Kieferov´a, I. D. Kivlichan, T. Menke, B. Peropadre, N. P. D. [21] J. L. Morales and J. Nocedal, ACM Trans. Math. Softw. 38, 7 Sawaya, S. Sim, L. Veis, and A. Aspuru-Guzik, Chemical Re- (2011). views 119, 10856 (2019). [22] J. C. Spall, Johns Hopkins apl technical digest 19, 482 (1998). [8] A. Peruzzo, J. McClean, P. Shadbolt, M.-H. Yung, X.-Q. Zhou, [23] J. C. Spall, in Proceedings of the 37th IEEE Conference on De- P. J. Love, A. Aspuru-Guzik, and J. L. O’Brien, Nature Com- cision and Control, Vol. 4 (1998) pp. 3872–3879 vol.4. munications 5, 4213 (2014). [24] E. F. Dumitrescu, A. J. McCaskey, G. Hagen, G. R. Jansen, [9] A. Kandala, A. Mezzacapo, K. Temme, M. Takita, M. Brink, T. D. Morris, T. Papenbrock, R. C. Pooser, D. J. Dean, and J. M. Chow, and J. M. Gambetta, Nature 549, 242 (2017). P. Lougovski, Phys. Rev. Lett. 120, 210501 (2018). [10] P. J. Ollitrault, A. Kandala, C.-F. Chen, P. K. Barkoutsos, [25] N. Stamatopoulos, D. J. Egger, Y. Sun, C. Zoufal, R. Iten, A. Mezzacapo, M. Pistoia, S. Sheldon, S. Woerner, J. Gam- N. Shen, and S. Woerner, arXiv:1905.02666 (2019). betta, and I. Tavernelli, arXiv:1910.12890 (2019). [26] S. Bravyi, J. M. Gambetta, A. Mezzacapo, and K. Temme, [11] Q. Gao, H. Nakamura, T. P. Gujarati, G. O. Jones, J. E. Rice, arXiv preprint arXiv:1701.08213 (2017). S. P. Wood, M. Pistoia, J. M. Garcia, and N. Yamamoto, [27] K. Setia, R. Chen, J. E. Rice, A. Mezzacapo, M. Pistoia, and arXiv:1906.10675 (2019). J. Whitfield, arXiv preprint arXiv:1910.14644 (2019). [12] J. Romero, R. Babbush, J. R. McClean, C. Hempel, P. J. Love, [28] R. D. Johnson III, NIST 101. Computational chemistry compar- and A. Aspuru-Guzik, Quantum Science and Technology 4, ison and benchmark database, Tech. Rep. (1999). 014008 (2018). [29] K. Temme, S. Bravyi, and J. M. Gambetta, Phys. Rev. Lett. [13] J. M. Turney, A. C. Simmonett, R. M. Parrish, E. G. Hohen- 119, 180509 (2017). stein, F. A. Evangelista, J. T. Fermann, B. J. Mintz, L. A. Burns, [30] A. Kandala, K. Temme, A. D. C´orcoles, A. Mezzacapo, J. M. Chow, and J. M. Gambetta, Nature 567, 491 (2019). 7

APPENDIX

species geometry equilibrium values symmetry spin-orbitals qubits frozen orbitals

LiH Li(0, 0, 0) RLiH = 1.595 A˚ C∞v 6 4 Li(1s,2px,2py) H(0, 0, RLiH)

H2S S(0, 0, 0) RSH = 1.356 A˚ C2v 12 9 S(1s,2s,2p) ◦ H(RSH cos(θHSH), RSH sin(θHSH), 0) θHSH = 92.11 H(−RSH cos(θHSH), RSH sin(θHSH), 0)

Li2S S(0, 0, 0) RLiS = 2.140 A˚ C2v 24 21 S(1s,2s,2p) ◦ Li(RLiS cos(θLiSLi), RLiS sin(θLiSLi), 0) θLiSLi = 134.97 Li(−RLiS cos(θLiSLi), RLiS sin(θLiSLi), 0)

LiSH S(0, 0, 0) RLiS = 2.140 A˚ Cs 18 15 Li(1s) Li(RLiS, 0, 0) RSH = 1.353 A˚ S(1s,2s,2p) ◦ H(RSH cos(θLiSH), RSH sin(θLiSH), 0) θLiSH = 92.11

LiH H2S

SH RLiH R θHSH

Li2S LiSH

RLiS R LiS θLiSLi R SH θHSLi

TABLE I. Table: List of the chemical species investigated in the present work. For each species, we list the investigated geometries, specified in terms of internal coordinates. Whenever internal coordinates are fixed at chosen values, the latter correspond with the quantities listed in column 3. For each species we show the frozen orbitals, molecular symmetry group and number of spin-orbitals and qubits. Figure: pictorial representation of the chemical species and geometries studied in the present work; white, purple and yellow spheres represent H, Li and S atoms respectively.