Direct Mapping of Charge Distribution during Lithiation of GeNanowires Using Off-Axis Electron HolographyZhaofeng Gan,† Meng Gu,‡ Jianshi Tang,§ Chiu-Yen Wang,§,∥ Yang He,⊥ Kang L. Wang,§

Chongmin Wang,*,‡ David J. Smith,*,† and Martha R. McCartney*,†

†Department of Physics, Arizona State University, Tempe, Arizona 85287, United States‡Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, Richland, Washington 99352, United States§Device Research Laboratory, Department of Electrical Engineering, University of California, Los Angeles, California 90095, UnitedStates∥Department of Materials Science and Engineering, National Taiwan University of Science and Technology, Taipei City, Taiwan10607, Republic of China⊥Department of Mechanical Engineering and Materials Science, University of Pittsburgh, Pittsburgh, Pennsylvania 15261, UnitedStates

*S Supporting Information

ABSTRACT: The successful operation of rechargeable batteries relies on reliable insertion/extraction of ions into/from the electrodes. The battery performance and the response of theelectrodes to such ion insertion and extraction are directly related to the spatial distributionof the charge and its dynamic evolution. However, it remains unclear how charge isdistributed in the electrodes during normal battery operation. In this work, we have used off-axis electron holography to measure charge distribution during lithium ion insertion into aGe nanowire (NW) under dynamic operating conditions. We discovered that the surfaceregion of the Ge core is negatively charged during the core−shell lithiation of the Ge NW,which is counterbalanced by positive charge on the inner surface of the lithiated LixGe shell.The remainder of the lithiated LixGe shell is free from net charge, consistent with its metalliccharacteristics. The present work provides a vivid picture of charge distribution and dynamicevolution during Ge NW lithiation and should form the basis for tackling the response ofthese and related materials under real electrochemical conditions.

KEYWORDS: Lithium ion battery, charge distribution, in situ TEM, electron holography

Lithium ion batteries (LIBs) are indispensable nowadays forpowering portable electronics and electric vehicles, and

they are emerging, in combination with other multivalence ionbatteries, as a means for storing renewable energy such as windand solar.1,2 The LIB recharging process is rather simple inphysical and chemical terms, relying on insertion/extraction ofions into/from a host lattice (the electrode) with concurrentflow of electrons to compensate the charge balance. Thematerials used as cathode and anode for the rechargeablebattery range from metal to semiconductor and insulator. It hasbeen well-documented that the electrochemically driveninsertion of lithium ions into a semiconductor or insulatorwill lead to amorphization (typically Li into Si, Ge, ZnO2, andSnO2),

3−6 a process that is not attainable with other knownsolid-state or wet-chemistry synthesis methods. Some materials,instead of being subject to solid-state amorphization, alloydirectly with the inserted ions to form new crystalline phases(typically Li with Al or Sn).7,8 The distinctively differentresponses of these materials to ion insertion and retractionappear to be related to the local charge distribution. Based oncomputational modeling,9 it has been proposed that the localelectron enrichment contributes to the solid-state amophization

of the materials upon lithium ion insertion into silicon andgermanium (semiconductor), while for metals (conductor) theion insertion leads to alloying. Furthermore, the charge anddischarge rates of a battery critically depend on the rate of ionmobility in the electrode, which has been found to be closelyrelated to the electronic conducting characteristics of theelectrode.10,11 Apparently, during the functioning of arechargeable battery, the local charges show dynamic evolution,which critically governs the response of the electrode to ioninsertion/extraction and ultimately determines the batteryperformance. However, associated with the dynamic, evolvingnature of the electrode, direct observation of charge distributionunder battery operating conditions has never been possible,although some attempts have been reported.12

In this work, we use a combination of in situ transmissionelectron microscopy (TEM) and off-axis electron holography todirectly map the charge distribution across a Ge/LixGe core/shell nanowire (NW) formed during the lithiation of the Ge

Received: March 14, 2016Revised: April 28, 2016Published: May 18, 2016

Letter

pubs.acs.org/NanoLett

© 2016 American Chemical Society 3748 DOI: 10.1021/acs.nanolett.6b01099Nano Lett. 2016, 16, 3748−3753

NW under dynamic battery operating conditions. Wediscovered that, during the core−shell lithiation of the GeNW, the outer part of the Ge core is negatively charged, whichis counterbalanced by positive charge on the inner surface ofthe lithiated LixGe shell. The remaining lithiated LixGe shell isfree from net charge, indicating its metallic characteristics. Thepresent work provides a vivid picture of charge distribution anddynamic evolution during Ge NW lithiation and should formthe basis for tackling the response of these materials under realelectrochemical conditions.Off-axis electron holography is an electron-microscopy-based

interferometric technique.13,14 By using interference betweenthe reference wave which has passed only through vacuum, andthe object wave, which has passed through the sample, anelectron hologram is formed. Phase and amplitude images ofthe sample can then be reconstructed. In the absence of anymagnetic field, the projected phase shift of the electron wavecaused by electrostatic fields within the sample can be simplyexpressed as

∫φΔ = · +x y C V x y z V x y z z( , ) ( ( , , ) ( , , )) dE 0 bi (1)

where CE is an electron-energy-dependent interaction constantwith the value of 0.00653 rad V−1 nm−1 for 300 keV electrons,V0 is the mean inner potential (MIP) of the sample, and Vbi isthe built-in potential due to any charge distribution. By usingthis equation, the phase shift measured by holography can bedirectly related to the electrostatic potential distribution of thesample.The single-crystal Ge NWs studied here were grown along

[111] directions on Si(001) substrates, as described in detailelsewhere.15 After growth, one Ge NW was detached from thesubstrate and attached to a Pt tip using silver glue while Limetal was attached to a second Pt tip. Both electrodes werethen installed in a Nanofactory STM holder for the in situlithiation experiments. Before insertion into the TEM lenscolumn, the STM holder was exposed to air for a few seconds,and a Li2O layer was formed on the Li metal surface during thisshort period, which then serves as the solid electrolyte duringlithiation. The Ge NW was moved to make physical contactwith the Li source, forming an open cell battery structure, asillustrated in Figure 1. A bias of ∼ −2 V was then applied to the

Ge NW, while the Li metal was kept grounded during thelithiation process. The off-axis electron holography experimentswere performed using an FEI Titan 80-300 electron micro-scope, which was operated at 300 keV and equipped with fieldemission gun, probe corrector, electrostatic biprism, Lorentzmini-lens, and Gatan Imaging Filter (Quantum 965). For theholography experiments, the objective lens was switched off,and the Lorentz mini-lens was used to obtain a larger field ofview. The typical biprism voltage was 120 V, the fringe spacingwas ∼3.4 nm, and the hologram exposure time was 2s.

Based on in situ TEM studies, it has been well-establishedthat lithiation of the Ge NW proceeds in two phases, leadinginitially to the formation of an LixGe shell that surrounds theremaining core of the Ge.6 This process is also demonstratedfor the present case, as shown by the in situ TEM movieincluded with the Supporting Information (Movie S1).10,16,17

Figure 2 shows TEM images of the Ge NW taken during

lithiation. Figure 2a shows the Ge NW in contact with the Limetal. Application of −2 V bias causes Li2O to diffuse onto theGe NW surface as a solid electrolyte. Figure 2b indicates that Liions have diffused into the surface of the Ge NW, forming acrystalline Ge/amorphous LixGe core/shell structure. Aslithiation continued, the shell became thicker, while the coreregion shrank. The total volume of the NW increased, as clearlyvisible in Figure 2c, while the core size decreased further. At thecompletion of lithiation, the Ge core had completelydisappeared, and the NW became polycrystalline, as shown inFigure 2d. Figure 3 shows STEM HAADF and EELS mappingof the Ge/LixGe core/shell structure, which were recordedimmediately after Figure 2b. In the HAADF image, theintensity of the NW shell is lower, compared to the core, whichmeans that the shell has lower average atomic number. The Liionization edge arrowed in Figure 3b was used to map thedistribution of Li across the NW, as shown by Figure 3c. Theintensity distribution in the mapping indicates that the Li wasdistributed homogeneously in the NW shell.Electron holography observations were made during the

lithiation process. Figure 4a, d, and g shows holograms taken atalmost the same times as the corresponding images shown inFigures 2b, c, and d, respectively. Figure 4b, e, and h are thecorresponding reconstructed phase images, using pseudocolorto emphasize the local changes of phase. The color scale barsare shown at top right. The center part of the NW has higher

Figure 1. Experimental setup: Schematic diagram showing exper-imental setup used for in situ observations of the Ge NW lithiationprocess.

Figure 2. TEM images showing Ge NW during lithiation: (a) Beforestart of lithiation; (b) formation of Ge/LixGe core/shell structure; (c)further growth of LixGe shell as lithiation continues; (d) highermagnification image taken after completion of lithiation, showingformation of polycrystalline LixGe NW.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.6b01099Nano Lett. 2016, 16, 3748−3753

3749

phase compared to the outer parts, which is due to thecombination of greater thickness and higher mean innerpotential of crystalline Ge, compared to LixGe. Phase profileswere extracted along the white arrows in the reconstructedphase images and are shown at the right in Figure 4c, f, and i,respectively. The Ge core in Figure 4c is about half thediameter of the NW, and both the core and shell parts havenearly cylindrical shapes. In Figure 4f, the Ge core has shrunkto a smaller size and appears to be faceted, as suggested by thetriangular profile at its center. Meanwhile, the LixGe shell stillretains a round cross section. In Figure 4h, the entire Ge NWhas been lithiated, the Ge core has disappeared, and thelithiated LixGe (x ∼ 3.75) NW again resembles a cylindricalNW shape. During this process, the NW diameter has increasedfrom 176 nm in Figure 4c, to 258 nm in Figure 4i.The phase profiles across the NW result from superposition

of the mean inner potentials of the materials and any built-inpotential. As a first step toward interpretation of the phaseprofiles, it is initially assumed that there are no trapped chargesin the NW and that the phase changes are due only to meaninner potential Vshell and the change of thickness. Models thatinclude trapped charges are discussed below. Assuming that theNW has a cylindrical shape, then the phase change due to theshell part can be calculated and compared with theexperimental results, as shown by the red curves in Figure 4c,f, and i. From the fitting for the best fitted Vshell, it seems that acylindrical NW shape fits reasonably well with the experimentalresults. The best fitted values of Vshell for these three stages oflithiation are shown in Table 1. As the lithiation continued, themean inner potential of Vshell decreased from 7.6 to 5.1 V. Thisdrop indicates that the Li concentration in the LixGe shell hasincreased during lithiation, which is to be expected since Li is alight element and has a smaller mean inner potential relative toGe. The bias applied to the NW should not affect thisconclusion since the bias was kept fixed at −2 V throughout theentire process. Similar fitting was applied for the core part inFigure 4c, as shown by the green curve. This model also fits

closely with the experimental data. However, the best fittedVcore is only 10.6 ± 0.1 V, which is in contrast to the value of14.3 ± 0.2 V for crystalline Ge published in the literature.18

This substantial difference of 3.7 ± 0.3 V suggests thelikelihood that charges are trapped in the NW during thelithiation process.In order to determine the amount of trapped charge in the

NW, a variety of models were considered, including: (a) sheetelectrons at the Ge core surface and sheet positive charge at theinner shell surface; (b) bulk electrons in the Ge core and sheetpositive charge at the inner shell surface; (c) sheet electrons atthe Ge core surface and sheet positive charge at the outer shellsurface; (d) bulk electrons in the Ge core and sheet positivecharges at the outer shell surface; (e) sheet electrons at the Gecore surface and bulk positive charge in the shell (see additionalfigures in the Supporting Information). The model thatprovides the closest fit with the experimental results is shownin Figure 5a. This model assumes that Li+ ions haveaccumulated at the inner surface of the LixGe shell, whileelectrons are trapped near the outer surface of the Ge core.In this model, the LixGe shell is grounded, while the Ge core

is negatively biased through the TEM biasing holder. Thus,electrons are distributed near the Ge surface, causing the GeNW core to become an equipotential, while the Li+ ions aredistributed on the inner surface of the LixGe shell and balancethe charge of electrons in the Ge core. Thus, an electric field isformed at the core−shell interface, which is attributed to apossible band alignment. The LixGe shell is also anequipotential so that there are no free charges in its bulk,which would be consistent with the idea that the LixGe shell hasmetallic characteristics. Based on this model, the built-inpotential for the Ge core is −3.7 V, while the LixGe shellremains grounded, so that the potential offset between the coreand shell would be ∼3.7 V. Thus, the lithiation process seemsto be associated with local electron enrichment, as predicted bytheoretical modeling,9 and suggests that the rate of NWlithiation might be limited by the LixGe/Ge interface reaction,

Figure 3. EELS mapping of Ge/LixGe core/shell structure: (a) STEM HAADF image, where bright contrast corresponds to remaining Ge core; (b)EELS spectrum showing Li ionization edge (arrowed) used to map out the Li distribution. Energy window used for background subtraction alsoshown; (c) Li spectrum mapping of core/shell structure.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.6b01099Nano Lett. 2016, 16, 3748−3753

3750

which is consistent with the kinetics studies of Ge and Sireported in the literature.19−21

It is generally appreciated that it is very difficult to quantifythe Li concentration, as indicated by the value x of Li in theLixGe, especially by in situ methods as the reaction proceeds.The present measurements of the change in mean innerpotential with progression of the lithiation provides a viablemeans of quantifying x. This can be done from the shell/corevolume ratio and the measured mean inner potential using thefollowing expression:

· + · =⎛⎝⎜⎜

⎞⎠⎟⎟x V V

VOLVOL

V( )Li GeGe

Li GeLi Ge

xx

(2)

where VGe and VLixGe are the mean inner potentials for crystalGe and LixGe, respectively, VOLGe and VOLLixGe are thevolumes for Ge and LixGe, respectively, and VLi is the meaninner potential changed when one Li atom is added to the Geunit cell while keeping the total volume unchanged. Thisequation also assumes that the MIP for Ge does not changewhile the lithiation process takes place.The measured radius values of the NW core and the

complete NW are shown in Table 2, where case A is beforelithiation, and cases B, C and D are for almost the same timesas when Figure 4a, d, and g, respectively, were taken. The caseD of Figure 4g is used for calculation of the MIP of Li, where xis taken to be 3.75 for the fully lithiated phase, and the volume

ratio VOLVOL x

Ge

Li Gewas calculated to be 0.13. The value for VLi is then

Figure 4. Electron holography observations of Ge/LixGe core/shell NW during lithiation: (a, d, and g) Holograms of NW; (b, e, and h)Corresponding reconstructed phase images, shown in pseudocolor (scale bar shown at top right in units of radian); (c, f, and i) Phase profiles alongthe white arrows in b, e, and h, respectively.

Table 1. Measured Potential of the LixGe Shell

number Vshell (V)

1 7.6 ± 0.12 6.4 ± 0.13 5.1 ± 0.1

Nano Letters Letter

DOI: 10.1021/acs.nanolett.6b01099Nano Lett. 2016, 16, 3748−3753

3751

calculated to be 6.34 V, using eq 2 and the MIP value forcrystalline Ge of 14.3 V.18 Since the core has fully disappeared,there should be no trapped charge in the NW, which mightotherwise influence the result.After calibration, the concentration of Li in LixGe, as given by

the value of x, can be calculated, again using eq 2. For case B atFigure 4a, the x value in the NW shell is estimated to be 2.4 forthe measured MIP of 7.6 V, and the effect of trapped charge hasbeen ignored. These results indicate that the intermediatelithiated state for the NW shell is LixGe, where x is significantlylower than 3.75. As the lithiation process continued, more Lidiffused into the LixGe shell, and thus x increased until itreached 3.75 and NW became fully lithiated. The case A cannotbe evaluated to confirm this result because it appears that thecore has a faceted shape, meaning that its cross-section areacannot be calculated accurately.Using a combination of in situ TEM and electron

holography, we have probed the dynamic evolution of chargedistribution during the Ge NW lithiation. With the increase oflithium concentration in the lithiated LixGe shell, the meaninner potential for the LixGe shell decreases. More importantly,during the progress of the Ge core−LixGe shell lithiation, it wasalso discovered that the Ge core had lower mean innerpotential than its theoretical value. This difference wasattributed to accumulation of trapped charge near the surfaceof the Ge core. Moreover, this charge within the Ge core wascounterbalanced by positive charge at the very inner surface ofthe lithiated LixGe shell, which was otherwise an equipotential.The present results provide a direct route to observations ofcharge distribution and its dynamic evolution during batterycharging and open up a promising new avenue for studyingbattery electrode dynamics involving ion insertion and electrontransport under battery operating condition.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.nano-lett.6b01099.

Different possible models for fitting the experimentalresults (PDF)In situ TEM movie showing the core−shell lithiation ofGe nanowire (AVI)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: david.smith@asu.edu.*E-mail: chongmin.wang@pnnl.gov.*E-mail: molly.mccartney@asu.edu.Author ContributionsZ.G. and M.G. contributed equally to this work.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe electron holography studies have been supported by DoEGrant DE-FG02-04ER46168. The Ge NW synthesis wassupported by the Ministry of Science and Technology throughGrant MOST 103-2218-E-011-007-MY3. The authors acknowl-edge use of facilities in the John M. Cowley Center for HighResolution Electron Microscopy at Arizona State University.This work at PNNL is supported by the Assistant Secretary forEnergy Efficiency and Renewable Energy, Office of VehicleTechnologies of the U.S. Department of Energy underContract No. DE-AC02-05CH11231, Subcontract No. 18769and DE-AC-36-08GO28308 under the Advanced BatteriesMaterials Research. A portion of the electron holographystudies was supported by the Laboratory Directed Research andDevelopment Program as part of the Chemical ImagingInitiative at Pacific Northwest National Laboratory (PNNL).The work was conducted in the William R. Wiley Environ-mental Molecular Sciences Laboratory (EMSL), a nationalscientific user facility sponsored by DOE’s Office of Biologicaland Environmental Research and located at PNNL. PNNL is

Figure 5. Model for trapped charges in Ge/LixGe core/shell structure: (a) Schematic diagram of the model; (b) experimental data (black), and bestfitting results (red).

Table 2. Measured Radius for NW Core and Whole NW

case core (nm) whole NW (nm)

A 66B 45 88C 110D 129

Nano Letters Letter

DOI: 10.1021/acs.nanolett.6b01099Nano Lett. 2016, 16, 3748−3753

3752

operated by Battelle for the DOE under Contract DE-AC05-76RLO1830.

■ REFERENCES(1) Armand, M.; Tarascon, J. M. Nature 2008, 451, 652−657.(2) Arico, A. S.; Bruce, P.; Scrosati, B.; Tarascon, J.-M.; vanSchalkwijk, W. Nat. Mater. 2005, 4, 366−377.(3) Huang, J. Y.; Zhong, L.; Wang, C. M.; Sullivan, J. P.; Xu, W.;Zhang, L. Q.; Mao, S. X.; Hudak, N. S.; Liu, X. H.; Subramanian, A.;Fan, H. Y.; Qi, L. A.; Kushima, A.; Li, J. Science 2010, 330, 1515−1520.(4) Limthongkul, P.; Jang, Y. I.; Dudney, N. J.; Chiang, Y. M. ActaMater. 2003, 51, 1103−1113.(5) Kushima, A.; Liu, X. H.; Zhu, G.; Wang, Z. L.; Huang, J. Y.; Li, J.Nano Lett. 2011, 11, 4535−4541.(6) Liu, X. H.; Huang, S.; Picraux, S. T.; Li, J.; Zhu, T.; Huang, J. Y.Nano Lett. 2011, 11, 3991−3997.(7) Liu, Y.; Hudak, N. S.; Huber, D. L.; Limmer, S. J.; Sullivan, J. P.;Huang, J. Y. Nano Lett. 2011, 11, 4188−4194.(8) Winter, M.; Besenhard, J. O. Electrochim. Acta 1999, 45, 31−50.(9) Wang, Z.; Gu, M.; Zhou, Y.; Zu, X.; Connell, J. G.; Xiao, J.; Perea,D.; Lauhon, L. J.; Bang, J.; Zhang, S.; Wang, C.; Gao, F. Nano Lett.2013, 13, 4511−4516.(10) Liu, X. H.; Zhang, L. Q.; Zhong, L.; Liu, Y.; Zheng, H.; Wang, J.W.; Cho, J. H.; Dayeh, S. A.; Picraux, S. T.; Sullivan, J. P.; Mao, S. X.;Ye, Z. Z.; Huang, J. Y. Nano Lett. 2011, 11, 2251−2258.(11) Zhang, L. Q.; Liu, X. H.; Liu, Y.; Huang, S.; Zhu, T.; Gui, L.;Mao, S. X.; Ye, Z. Z.; Wang, C.-M.; Sullivan, J. P.; Huang, J. Y. ACSNano 2011, 5, 4800−4809.(12) Yamamoto, K.; Iriyama, Y.; Asaka, T.; Hirayama, T.; Fujita, H.;Fisher, C. A. J.; Nonaka, K.; Sugita, Y.; Ogumi, Z. Angew. Chem., Int.Ed. 2010, 49, 4414−4417.(13) McCartney, M. R.; Smith, D. J. Annu. Rev. Mater. Res. 2007, 37,729−767.(14) McCartney, M. R.; Agarwal, N.; Chung, S.; Cullen, D. A.; Han,M.-G.; He, K.; Li, L.; Wang, H.; Zhou, L.; Smith, D. J. Ultramicroscopy2010, 110, 375−382.(15) Tang, J. S.; Wang, C.-Y.; Xiu, F. X.; Lang, M. R.; Chu, L.-W.;Tsai, C.-J.; Chueh, Y.-L.; Chen, L.-J.; Wang, K. L. ACS Nano 2011, 5,6008−6015.(16) Liu, X. H.; Huang, J. Y. Energy Environ. Sci. 2011, 4, 3844−3860.(17) Liu, X. H.; Liu, Y.; Kushima, A.; Zhang, S.; Zhu, T.; Li, J.;Huang, J. Y. Adv. Energy Maters. 2012, 2, 722−741.(18) Li, J.; McCartney, M. R.; Dunin-Borkowski, R. E.; Smith, D. J.Acta Crystallogr., Sect. A: Found. Crystallogr. 1999, 55, 652−658.(19) Gu, M.; Yang, H.; Perea, D. E.; Zhang, J.-G.; Zhang, S.; Wang,C.-M. Nano Lett. 2014, 14, 4622−4627.(20) McDowell, M. T.; Ryu, I.; Lee, S. W.; Wang, C. M.; Nix, W. D.;Cui, Y. Adv. Mater. 2012, 24, 6034−6041.(21) Yang, H.; Huang, X.; Liang, W.; van Duin, A. C. T.; Raju, M.;Zhang, S. Chem. Phys. Lett. 2013, 563, 58−62.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.6b01099Nano Lett. 2016, 16, 3748−3753

3753