Spontaneous Generation of Electromotive Force in Thin Film Al/Nanosilicon/Al Structures

batteries

Article Spontaneous Generation of Electromotive Force in Thin Film Al/Nanosilicon/Al Structures

Sergey G. Dorofeev 1,* ID , Nikolay N. Kononov 2, Sergei S. Bubenov 1, Pavel A. Kotin 1, Aleksandr N. Zolotykh 1 and Denis V. Grigoriev 1 1 Chemistry Department, Lomonosov Moscow State University, Leninskie Gory, Moscow 119234, Russia; [email protected] (S.S.B.); [email protected] (P.A.K.); [email protected] (A.N.Z.); [email protected] (D.V.G.) 2 Prokhorov General Physics Institute, Russian Academy of Sciences, Moscow 119333, Russia; [email protected] * Correspondence: [email protected]; Tel.: +7-916-641-0442

Received: 8 September 2017; Accepted: 7 October 2017; Published: 10 October 2017

Abstract: Contemporary pursuits in electronics include the miniaturization as well as flexibilization of devices. Although there are a large number of different thin and flexible electrochemical batteries, only a few can boast the possibility of working in high humidity conditions. This paper reports on the fabrication of structures consisting of films of silicon nanoparticles encased between two aluminium electrodes. The value of electromotive force (emf) measured depends on the temperature of the sample and on the pressure of water vapor in the storage atmosphere and reaches approximately 1 V. Volt-ampere characteristics were investigated at different conditions to yield a model of emf generation in these structures. It was found that the reaction of water with silicon nanoparticles is the prime reason behind emf generation. Such a source may be introduced into electronic paper, and employed in the next generation of smart cards. The structure may also be manufactured directly on the surface of silicon chips, such as on the back of crystals in microschemes.

Keywords: thin ﬁlm battery; nanosilicon; silicon-water element; spontaneous emf generation

1. Introduction The rapid advances in technology for the fabrication of miniature electronic devices call for the development of sufficiently small sources and accumulators of energy (on-chip power sources). Micro-fuel cells and supercapacitors are the most viable options for this purpose [1–3]. The use of an external oxidizing agent (air) allows for a significant decrease in the size/weight of electrochemical cells; therefore, metal-air batteries (e.g., Zn-air, Li-air, Al-air) [4–6] offer a prospective solution in terms of creating autonomous energy sources. In these batteries, charge separation occurs via oxidation of the metal in contact with an electrolyte and reduction of oxygen entering the battery through a porous electrode. Among other representatives of the class, it is worth mentioning silicon-air (Si-air) batteries based on crystalline silicon (c-Si) wafers [7–10]. Due to a four-electron oxidation reaction, their specific energy is 3822 mAh·g−1, which is on par with that of a Li-air battery (3860 mAh·g−1)[10], while their material (silicon) is abundant, non-toxic, and eco-friendly. An essential component of a Si-air or a metal-air battery is an alkaline electrolyte or an ionic liquid containing fluorides (e.g., 1-ethyl-3-methylimidazolium hydrofluoride), which enables charge transfer between electrodes. These electrochemical cells have a considerable thickness (over 1 cm) and, therefore, cannot be considered flexible energy sources. Great progress has been achieved recently in the field of electrochemical power sources. A lot of developments and investigations have occurred in incorporating different materials, including carbon

Batteries 2017, 3, 31; doi:10.3390/batteries3040031 www.mdpi.com/journal/batteries Batteries 2017, 3, 31 2 of 10 nanotubes as conductors and electrodes [11], silicon nanowires as anodes [12], and cellulose—which is the most abundant renewable polymer source available—as a substrate for electrode materials [13–15]. Thus, prominent trends in the research of power sources include miniaturization (introduction of low-scaleBatteries materials 2017, 3, 31 such as nanoparticles or graphene) and the achievement of high flexibility2 of 10 in combinationcarbon withnanotubes the abundance, as conductors non-toxicity, and electrodes and eco-friendliness[11], silicon nanowires mentioned as anodes earlier [12], [11 –and18]. Thiscellulose—which work deals withis the themost fabrication abundant renewable of an ultra-thin polymer source battery available—as composed a ofsubstrate a film for of silicon nanoparticleselectrode encased materials between [13–15]. two Thus aluminium, prominent planar trends electrodes. in the research The netof power thickness sources of such include a structure (disregardingminiaturization the thickness (introduction of the substrate)of low-scale is materials less than such 2 microns, as nanoparticles which makesor graphene) it a perfect and the candidate achievement of high flexibility in combination with the abundance, non-toxicity, and for a flexible energy source, assuming that a flexible substrate is used. eco-friendliness mentioned earlier [11–18]. PreviousThis investigations work deals with of siliconthe fabrication nanoparticles of an ultra-thin between battery interdigitated composed aluminium of a film of electrodessilicon [19] showednanoparticles a positive effectencased of waterbetween on two electrical aluminium conductivity, planar electrodes. and it wasThe net established thickness thatof such oxidation a of the surfacestructure of Si (disregarding nanocrystals the (nc-Si)thickness diminishes of the substrate) the is electrical less than 2 conductivity microns, which of makes the films it a perfect irreversibly. However,candidate the advantages for a flexible of energy the application source, assumi of suchng that structure a flexible in substrate a humid is environmentused. in itself deserves further study.Previous At the investigations same time, of the silicon authors nanoparticles of [19] didn’t between observe interdigitated emf generation, aluminium dueelectrodes to the [19] symmetry showed a positive effect of water on electrical conductivity, and it was established that oxidation of of electrodes. Asymmetry of the structure is crucial for emf generation. The battery produced in this the surface of Si nanocrystals (nc-Si) diminishes the electrical conductivity of the films irreversibly. work isHowever, truly an autonomousthe advantages current of the application source, and of issuch essentially structure ain nc-Si/water a humid environment element. in In itself the element reporteddeserves here, thefurther film study. of silicon At the nanoparticlessame time, the author workss of simultaneously [19] didn’t observe as emf fuel, generation, as a charge due transportto membrane,the symmetry and as an of electrodeelectrodes. separator.Asymmetry of the structure is crucial for emf generation. The battery produced in this work is truly an autonomous current source, and is essentially a nc-Si/water 2. Experimentalelement. In Sectionthe element reported here, the film of silicon nanoparticles works simultaneously as fuel, as a charge transport membrane, and as an electrode separator. Silicon nanoparticles produced via laser-induced silane pyrolysis [20] were used to obtain silicon films. The2. Experimental mean diameter Section of the nc-Si was determined to be 15 nm by means of transmission electron microscopySilicon (TEM). nanoparticles TEM was performedproduced via on laser-induced a JEM2100 microscopesilane pyrolysis (JEOL, [20] Tokyo,were used Japan) to obtain with 200 kV acceleratingsilicon voltage. films. The The mean mean diameter size of was the calculatednc-Si was determined by averaging to be 15 of nm diameters by means of of 100transmission particles. After storageelectron in air, the microscopy surface of(TEM). nc-Si TEM becomes was performed partially oxidized,on a JEM2100 and microscope its composition (JEOL, mayTokyo, be Japan) described as with 200 kV accelerating voltage. The mean size was calculated by averaging of diameters of 100 SiOx (0 < x < 2). particles. After storage in air, the surface of nc-Si becomes partially oxidized, and its composition Fabrication of sandwich structures started with vacuum deposition of a bottom aluminium may be described as SiOx (0 < x < 2). electrode (denotedFabrication later of sandwich as the bottom structures contact, started see with Figure vacuum1) on deposition a glass substrate. of a bottom Films aluminium of nc-Si were then depositedelectrode on(denoted the bottom later as contactthe bottom by contact, means see of Figu eitherre 1) electrophoresis on a glass substrat ore. spin-coating Films of nc-Si were from sols of nc-Si inthen polar deposited solvents on the (water bottom or contact acetone). by means The procedureof either electrophoresis was finalized or spin-coating with a vacuum from sols deposition of of top aluminiumnc-Si in polar electrodes.solvents (water A or few acetone). structures The procedure were annealed was finalized in vacuumwith a vacuum (10− 5depositionTorr) for of 1 h at a −5 temperaturetop aluminium of 700 K electrodes. before the A deposition few structures of top were aluminium annealed in electrodes. vacuum (10 Torr) for 1 h at a temperature of 700 K before the deposition of top aluminium electrodes.

Figure 1. Scheme of Al/nc-Si/Al structure. Figure 1. Scheme of Al/nc-Si/Al structure. Batteries 2017, 3, 31 3 of 10

As a result, the cell work area was 1.4 cm2 for each contact. The height of the cell was determined by the glass substrate (1.2 mm). The thickness of the Al contacts (bottom and top) and that of nc-Si film were 0.3 µm and 0.2–0.5 µm, respectively. To discern the electronic and protonic conductivity, Nafion-containing samples were produced. Batteries 2017, 3, 31 3 of 10 Nafion is a sulfonated tetrafluoroethylene-based fluoropolymer-copolymer. It is commonly used as a protonicAs conductor a result, forthe protoncell work exchange area was membrane 1.4 cm2 for in each fuel contact. cells. ThisThe structureheight of the was cell manufactured was throughdetermined vacuum by deposition the glass substrate of an aluminium (1.2 mm). The electrode thickness on of Nafionthe Al contacts film, followed (bottom and by top) deposition and of a film ofthat nc-Si of nc-Si on the film other were side,0.3 µm and and covering0.2–0.5 µm, it respectively. with the top aluminium electrode. This resulted in a flexible structureTo discern (without the electronic substrate), and protonic where conducti the Nafionvity, membraneNafion-containing was between samples were the bottomproduced. contact (3) Nafion is a sulfonated tetrafluoroethylene-based fluoropolymer-copolymer. It is commonly used as and thea filmprotonic of siliconconductor nanocrystals for proton exchange (2) (see membrane Figure1). in fuel cells. This structure was manufactured Thethrough quality vacuum of the deposition silicon nanoparticle of an aluminium films waselectr controlledode on Nafion with film, a scanning followed electron by deposition microscope of a (SEM) NVisionfilm 40 of (Carl nc-Si Zeiss, on theOberkochen, other side, and Germany) covering it withwith the an top acceleration aluminium voltage electrode. 5 kVThis and resulted 2 nm in resolution. a The thicknessflexible structure of the (without films was substrate), determined where the with Nafion use membrane of a TalyStep was between profilometer the bottom (Taylor-Hobbson,contact (3) Leicester,and England).the film of silicon nanocrystals (2) (see Figure 1). The quality of the silicon nanoparticle films was controlled with a scanning electron microscope Emf measurements were carried out in the temperature range of 300 to 400 K. The temperature (SEM) NVision 40 (Carl Zeiss, Oberkochen,◦ Germany) with an acceleration voltage 5 kV and 2 nm was registeredresolution. with The anthickness error ofof 0.1the filmsC. For was signal determined registration, with use high-ohmic of a TalyStep 24-bit profilometer 4-channel digital recorders(Taylor-Hobbson, Expert-001 (Ekoniks,Leicester, England). Moscow, Russia) were used. The studies of effects of atmosphere compositionEmf on measurements electronic characteristicswere carried out in of the the temperature samples range were of performed 300 to 400 K. in The a temperature gas flow chamber allowingwas pressure, registered gaswith composition an error of 0.1 and °C. humidityFor signal registration, control. Relative high-ohmic partial 24-bit pressures 4-channel of digital water vapor and oxygenrecorders were Expert-001 determined (Ekoniks, with Moscow, HIH 4000 Russia) (Honeywell, were used. Lincolnshire, The studies of IL, effects USA) of and atmosphere Oxic 12 sensors composition on electronic characteristics of the samples were performed in a gas flow chamber (Oxonium, Saint-Petersburg, Russia) to an accuracy of 3.5 and 1%, respectively. allowing pressure, gas composition and humidity control. Relative partial pressures of water vapor and oxygen were determined with HIH 4000 (Honeywell, Lincolnshire, IL, USA) and Oxic 12 3. Results and Discussion sensors (Oxonium, Saint-Petersburg, Russia) to an accuracy of 3.5 and 1%, respectively. Initially, the investigation of M/nc-Si/M structures was concerned with the use of nc-Si film as a 3. Results and Discussion solid separator and as an electrolyte. Thus, different bottom and top contact materials were considered. With the resultsInitially, obtained, the investigation it was found of M/nc-Si/M that the structures change was of the concerned material with of electrodesthe use of nc-Si didn’t film significantlyas a solid separator and as an electrolyte. Thus, different bottom and top contact materials were affect emf generation. Moreover, it was established that a more important role in this process is played considered. With the results obtained, it was found that the change of the material of electrodes by water;didn’t therefore, significantly we affect focused emf ourgeneration. attention Moreover, on equal-metal it was established structures, that a and more aluminium important role proved in to be the mostthis convenient process is played material. by water; therefore, we focused our attention on equal-metal structures, and aluminium proved to be the most convenient material. 3.1. SEM Investigation 3.1. SEM Investigation In Figure2, SEM images of the structures without the top electrode (Figure2a), and with the aluminiumIn top Figure contact 2, SEM (Figure images2b) of are the presented. structures with Theout top the aluminium top electrode contact (Figure mostly 2a), and consists with the of faceted aluminium top contact (Figure 2b) are presented. The top aluminium contact mostly consists of ± metal nanoparticlesfaceted metal nanoparticles with an average with ansize average of 75 size of10 75 nm. ± 10 nm.

(a) (b)

Figure 2. Scanning electron microscope (SEM) investigation of different structure layers. (a) nc-Si Figure 2. Scanning electron microscope (SEM) investigation of different structure layers. (a) nc-Si ﬁlm film without the top Al electrode, (b) the surface of the top Al electrode. Images represent the top withoutview the of top the Alstructures. electrode, (b) the surface of the top Al electrode. Images represent the top view of the structures.

Batteries 2017, 3, 31 4 of 10

Evidently, the fabricated material is very porous. Porosity of the top aluminium electrode enablesBatteries water 2017, permeation.3, 31 Moreover, it was established that this electrode configuration4 is offormed 10 spontaneously during the deposition process. Evidently, the fabricated material is very porous. Porosity of the top aluminium electrode On the contrary, porosity of nc-Si film may lead to poor charge transfer. In order to mitigate this, enables water permeation. Moreover, it was established that this electrode configuration is formed some samples were annealed in vacuum before the top aluminium electrode deposition (the conditions spontaneously during the deposition process. are mentionedOn the incontrary, the Experimental porosity of nc-Si section). film may As a lead result, to poor emf valuescharge transfer. increased In moreorder thanto mitigate two times (see below).this, some samples were annealed in vacuum before the top aluminium electrode deposition (the conditionsThe use of are annealing mentioned to in increasethe Experimental the conductivity section). As ofa result, nc-Si emf films values was increased already knownmore than [21 ,22]. Possibletwo mechanismstimes (see below). include sintering, particle transfer, and even sublimation of nanoparticles from a hot substrateThe use to aof cold annealing one [23 to]. increase This process the conducti makesvity it possible of nc-Si tofilms increase was already density known and conductivity. [21,22]. OurPossible investigations mechanisms showed include that sintering, the thickness particle oftransfer, nc-Si and films even decreased sublimation twofold of nanoparticles upon annealing from at 300 ◦aC hot for substrate 30 min. Afterto a cold that, one the [23]. thickness This process remained makes constant.it possible to No increase sublimation density wasand conductivity. observed during Our investigations showed that the thickness of nc-Si films decreased twofold upon annealing at 300 °C this process, and we propose that an increase of nc-Si film density occurs. for 30 min. After that, the thickness remained constant. No sublimation was observed during this 3.2. Theprocess, Influence and we of Temperaturepropose that andan increase Humidity of nc-Si on Emf film Generation density occurs. 3.2.Methods The Influence of nc-Si of Temperature film deposition and Humidity (electrophoresis on Emf Generation or spin-coating) do not yield significantly differentMethods results inof termsnc-Si film of thedeposition final emf (electrophoresis value. The otheror spin-coating) two parameters do not yield are temperaturesignificantly and humidity.different Both results temperature in terms andof the humidity final emf have valu ae. positive The other impact two onparameters the emf. are While temperature they affect and charge transferhumidity. processes, Both moisturetemperature can and also humidity serve as anhave oxidizer a positive component, impact on which the emf. ensures While the they maintenance affect of chemicalcharge transfer processes. processes, moisture can also serve as an oxidizer component, which ensures the maintenanceAn investigation of chemical of a series processes. of structures in dry Ar/O2 mixtures has shown that the emf increases with the oxygenAn investigation amount. of Thea series growth of structures is not significant—noin dry Ar/O2 mixtures more has than shown 10 percent that the emf after increases the increase with the oxygen amount. The growth is not significant—no more than 10 percent after the increase of oxygen from 0 to 100 percent. As the influence of oxygen was even lower when using wet Ar/O2 of oxygen from 0 to 100 percent. As the influence of oxygen was even lower when using wet Ar/O2 mixtures, the variation of this parameter wasn’t studied in greater detail. mixtures, the variation of this parameter wasn’t studied in greater detail. EveryEvery sample sample studied studied exhibited exhibited a certain a certain potential potential difference difference between between the the top top and and bottom bottom contacts, withcontacts, the exact with value the ranging exact value between ranging 30 mV between and 1 30 V, dependingmV and 1 V, on depending temperature on andtemperature humidity. and For all samples,humidity. the potential For all differencesamples, the between potential the differ bottomence and between top contacts the bottom was positive.and top Currentcontacts thatwas flows frompositive. the bottom Current electrode that flows to the from top electrodethe bottom in electrode external to circuit the top is positive.electrode in external circuit is positive. 3.2.1. Temperature Dependence 3.2.1. Temperature Dependence The plotting of temperature-emf curves allows one to formulate a model of emf generation in the structuresThe investigated. plotting of temperature-emf In Figure3, we present curves theallows semi-log one to plotformulate of the a emf model generated of emf generation by the sample in in air atthe constant structures humidity investigated. as a function In Figure of reciprocal3, we present temperature. the semi-log The plot dependence of the emf givengenerated is characteristic by the sample in air at constant humidity as a function of reciprocal temperature. The dependence given is of all structures studied, and isn’t fit well by a single exponent emf ~e−E/kT, where E is the activation characteristic of all structures studied, and isn’t fit well by a single exponent emf ~ e−E/kT, where E is energy.the Therefore,activation energy. the following Therefore, model the follow (Figureing 3model, insert) (Figure is proposed. 3, insert) is proposed.

Figure 3. Dependence of emf generated by the structure on reciprocal temperature, semi-log plot, the Figure 3. Dependence of emf generated by the structure on reciprocal temperature, semi-log plot, dashed line shows an approximation of the dependence; (insert) - equivalent scheme of the model. the dashed line shows an approximation of the dependence; (insert) - equivalent scheme of the model.

Batteries 2017, 3, 31 5 of 10 Batteries 2017, 3, 31 5 of 10

The modelmodel suggests suggests that that there there are twoare chargetwo char transferge transfer processes processes associated associated with emf generation:with emf generation:one series to one the series source to (possesses the source an (possessesrs resistance), an whichrs resistance), is associated which with is associated internal charge with transfer;internal chargeand one transfer; leak process and one (rleak leak, shunting), process which(rleak, shunting), probably appearswhich probably in contacts appears of different in contacts layers. of Activation different layers.energies Activation behind these energies processes behind (E stheseand E processesleak, respectively) (Es and differEleak, respectively) greatly, because differ the greatly, energy because of short thecircuit energy tends of toward short circuit zero. Therefore,tends toward the followingzero. Therefore, relation the is true:followingEs >> relationEleak. This is true: implies Es >> that Eleak parallel. This impliesresistance that (r leakparallel) is less resistance dependent (rleak on) is temperatureless dependent in theon studiedtemperature temperature in the studied interval. temperature However, interval.the series However, process is the of activationalseries process character, is of activa andtional one can character, get the emf and from one thecan followingget the emf formula: from the following formula: rleak × ε0 em f = (1) rs + rleak (1) E/kT takingtaking into accountinto account that rthat= r 0r ×= re0 × eE/kTand and neglecting neglectingEleak Eleak: : r0s em f = ε0 /11 + exp(Es/kT⁄) (2)(2) r 0leak

The values of EEss, r0s/r/r0leak0leak andand εε0 0areare fit fit parameters. parameters. The The proposed proposed model model leads leads to to an accurate approximation of emf dynamics data for the structure, as is shown in Figure 33.. TheThe fitfit valuesvalues forfor thisthis sample are: ε0 ≈≈ 1.11.1 V, V, EEs s≈≈ 0.280.28 eV. eV. According According to to [24], [24], there there are are surface surface states states for for Si Si that that are are situated insideinside the band gap 0.5 eV below the bottom of the conduction band. This value is close to Es,, and and we we suppose that injection ofof electronselectrons happenshappens through through these these states states with with consequent consequent thermal thermal activation activation in inthe the conduction conduction band. band. From thethe analysisanalysis of of the the temperature temperature dependences dependence of thes of emf, the it emf, follows it thatfollows dissipative that dissipative processes processesoccur in these occur structures, in these structures, in parallel in with parallel the process with the of process emf generation. of emf generation. Obviously, Obviously, such processes such processesshunt the shunt emf, andthe emf, lead and to a lead decrease to a decrease in its effective in its effective value. Tovalue. find Tor0 andfind rrleak0 andvalues, rleak values, a series a seriesof measurements of measurements of load of characteristicsload characteristics was performed,was performed, in which in which the structurethe structure (measured (measured in the in theprevious previous investigation) investigation) was connectedwas connected to an externalto an exte resistancernal resistance R, the valueR, the of value which of could which be could changed be changed(the scheme (the of scheme the circuit of the and circuit the results and the can results be found can be in Figurefound 4in). Figure 4).

Figure 4. InvestigationInvestigation of of load load characteristics of of the thin-filmthin-film structure: dependence dependence of of reciprocal voltage on the electrodes of the stru structurecture as a function of the reciproc reciprocalal of the external resistance in −1 −1 logarithmlogarithm scales. The dasheddashed lineline inin thisthis figurefigure shows shows the the approximating approximating function function V −V1(R(R−1)) which which is is constructed using the model presented on the insetinset of this figure;figure; (insert)—equivalent(insert)—equivalent circuit.circuit.

By measuring voltage V at the external resistance, we obtained the dependence V−1(R−1) (see By measuring voltage V at the external resistance, we obtained the dependence V−1(R−1) Figure 4). To find the unknown parameters (rs and rleak), a linear fit was performed for the following (see Figure4). To ﬁnd the unknown parameters ( r and r ), a linear ﬁt was performed for the coordinates: s leak following coordinates: 1 rs 11 rs +r = × + leak (3)(3) V V ε0 RR rleak ×ε0

Batteries 2017, 3, 31 6 of 10 Batteries 2017, 3, 31 6 of 10

− − From thethe analysisanalysis ofof thethe V V−1(R(R−1)1 )dependence dependence and and taking taking into into account account our our previous previous investigation 7 7 ((εε00 = 1.1 V) the values of rs andand rrleakleak werewere derived derived to to be be 6.9 6.9 ×× 10107 andand 3.9 3.9 × 10× 710 Ohm,Ohm, respectively. respectively.

3.2.2.3.2.2. InfluenceInfluence of Humidity As waswas describeddescribed above, one of thethe prominentprominent peculiaritiespeculiarities of the structures is their reaction to waterwater vapor.vapor. ItIt waswas establishedestablished thatthat therethere isis nono spontaneousspontaneous emfemf generationgeneration inin vacuum.vacuum. When water vaporvapor isis introduced introduced into into the measurementthe measurement chamber chamber and the and humidity the ofhumidity the atmosphere of the surroundingatmosphere thesurrounding structures the increases structures accordingly, increases emf accordingly, is spontaneously emf is produced, spontaneously leading produced, to a potential leading difference to a betweenpotential thedifference electrodes between encasing the theelectrodes nc-Si film. encasing the nc-Si film. To quantifyquantify thethe minuteminute amountamount ofof waterwater absorbedabsorbed byby nc-Sinc-Si film,film, aa quartz microbalance basedbased on oscillationoscillation frequency measurementmeasurement (12(12 MHz)MHz) waswas used.used. Each value waswas obtainedobtained byby comparingcomparing thethe frequencyfrequency ofof the the test test sample sample with with an an empty empty reference reference sample. sample. Humidity Humidity was increasedwas increased stepwise stepwise from anfrom argon an argon atmosphere atmosphere (see Figure (see 5Figurea). After 5a). reaching After reaching the maximum, the maximum, the humidity the humidity was lowered was lowered to zero, butto zero, a considerable but a considerable amount of amount water was of caughtwater bywas the caught structure. by the The structure. final stage The of our final experiment stage of wasour −5 vacuumexperiment degassing was vacuum (400 K, degassing 2 h, 10 Torr).(400 K, The 2 h, mass 10−5 value Torr). reached The mass inthis value process reached was in treated this process as the baselinewas treated mass as ofthe the baseline pure nc-Si mass film of the (Figure pure 5nc-Sia, green film square), (Figure 5a, and green water square), mass was and identified water mass as thewas excess.identified The as structure the excess. was foundThe structure to absorb was between found 1 to and absorb 5 mass between % of water, 1 and depending 5 mass on% humidityof water, (seedepending Figure5 ona). humidity (see Figure 5a).

(a) (b)

Figure 5. (a) Relative mass of water absorbed by nc-Si film during the humidity experiment. (b) Figure 5. (a) Relative mass of water absorbed by nc-Si film during the humidity experiment. Dependences of the emf of different samples on relative humidity of surrounding atmosphere at T = (b) Dependences of the emf of different samples on relative humidity of surrounding atmosphere at 300 K. Sample 1 was annealed in vacuum before; Sample 2 and Sample 3 are different samples on the T = 300 K. Sample 1 was annealed in vacuum before; Sample 2 and Sample 3 are different samples on thesame same substrate. substrate. Sample Sample 4 contains 4 contains Nafion Nafion film. film.

The influence of relative humidity on the emf of different samples is shown in Figure 5b. We annealedThe influence Sample of1 before relative the humidity investigation on the emf using of different the technique samples described is shown inearlier. Figure Samples5b. We annealed 2 and 3 Samplewere made 1 before on the the investigationsame substrat usinge, and the the technique measurements described indicate earlier. reproducibilitySamples 2 and 3of were our mademethod. on theSample same 4 substrate, contained and a theNafion measurements film. All results indicate are reproducibility subject to the of oursame method. law, close Sample to 4linear. contained The asignificantly Nafion film. higher All results emf value are subject of Sample to the 1 same is ev law,idence close of the to linear. positive The effect significantly of annealing. higher emf value of SampleAs was 1 is mentioned, evidence of there the positive is no spontaneous effect of annealing. emf generation in vacuum. Sample 1 is not an exceptionAs was to this mentioned, rule, and there the high is no value spontaneous of emf at humidity emf generation close to in 16% vacuum. is due to Sample the higher 1 is density not an exceptionof the sample to this after rule, annealing, and the highand its value greater of emf ability at humidity to hold water close molecules. to 16% is due After to thekeeping higher Sample density 1 ofin thea desiccator sample after for several annealing, weeks, and the its greateremf value ability dropped to hold to waterzero. molecules. After keeping Sample 1 in a desiccatorThe nature for of several the emf weeks, dependence the emf on value humidity dropped for to Nafion zero. structure (Sample 4) allows us to stateThe that naturethe process of the of emf emergence dependence of electrical on humidity potentials for at Nafion the electrodes structure of (Sample the studied 4) allows structure us tois stateconnected that the with process ion transport. of emergence of electrical potentials at the electrodes of the studied structure is connected with ion transport.

Batteries 2017, 3, 31 7 of 10 Batteries 2017, 3, 31 7 of 10

3.3. Comparative Analysis of Current-VoltageCurrent-Voltage (I-V) Characteristics Investigation of of volt-ampere volt-ampere characteristics characteristics (VACs) (VACs) allowed allowed us usto find to ﬁnd an influence an inﬂuence of humidity of humidity and andtemperature temperature on emf on and emf discharge and discharge current currentvalue. VACs value. of VACsall samples of all were samples asymmetrical were asymmetrical (Figure 6), (Figurewhich allows6), which one allows to ascertain one to the ascertain existence the of existence electrochemical of electrochemical processes in processes the structure. in the Oxidation structure. Oxidationprocesses on processes the top on and the reduction top and reductionprocesses processeson the bottom on the electrodes bottom electrodes lead to positive lead to positivecurrent currentvalues. In values. fact, this In fact, corresponds this corresponds to the fast to thedischarge fast discharge of a galvanic of a galvanic structure. structure. Negative Negative current currentvalues valuescorrespond correspond to reverse to reverse processes. processes.

Figure 6.6. InvestigationInvestigation ofof volt-amperevolt-ampere characteristicscharacteristics (VAC) (VAC) ofof Al/nc-Si/AlAl/nc-Si/Al structures structures measured measured at 300K: (black curve) before heating,heating, (blue(blue curve)curve) afterafter heatingheating andand recooling.recooling.

It can be clearly seen in Figure 6 (black curve) that initial VACs at ambient conditions possessed It can be clearly seen in Figure6 (black curve) that initial VACs at ambient conditions possessed significantly higher discharge current values (at the same voltage) and emf than after heating the significantly higher discharge current values (at the same voltage) and emf than after heating the sample to 340 K and recooling to 300 K (see Figure 6, blue curve). Recooling the sample didn’t allow sample to 340 K and recooling to 300 K (see Figure6, blue curve). Recooling the sample didn’t allow us to get initial emf and discharge current values. This feature shows that the influence of humidity us to get initial emf and discharge current values. This feature shows that the influence of humidity is is very large. very large. The possibility of porous silicon oxidation in pure water (pH ~ 7) coupled with molecular The possibility of porous silicon oxidation in pure water (pH ~7) coupled with molecular hydrogen hydrogen formation has been demonstrated before [25], the effectiveness of the process is associated formation has been demonstrated before [25], the effectiveness of the process is associated with the with the very high specific surface area of nanocrystalline silicon. We would also like to note Ref. [26], very high specific surface area of nanocrystalline silicon. We would also like to note Ref. [26], which is which is related to hydrogen generation with the use of particles obtained via laser-induced silane related to hydrogen generation with the use of particles obtained via laser-induced silane pyrolysis. pyrolysis. The particles studied had a mean size similar to that of those considered in this paper. It The particles studied had a mean size similar to that of those considered in this paper. It was shown that was shown that the etching of nc-Si in water is isotropic and occurs at a rate of 3 × 10−4 µm/s at pH = 7. the etching of nc-Si in water is isotropic and occurs at a rate of 3 × 10−4 µm/s at pH = 7. Accordingly, Accordingly, we consider the following reactions of nc-Si oxidation (the potential values [27] are we consider the following reactions of nc-Si oxidation (the potential values [27] are given for the same given for the same reactions in aqueous solutions, and are present as a reference): reactions in aqueous solutions, and are present as a reference): Si + 3 H2O → H2SiO3 + 4H+ + 4 e−; E = 1.19 V, pH = 7; (4) + − Si + 3 H2O → H2SiO3 + 4H + 4 e ; E = 1.19 V, pH = 7; (4) Si + 2 H2O → SiO2 + 4 H+ + 4 e−; E = 1.27 V, pH = 7. (5) + − Si + 2 H2O → SiO2 + 4 H + 4 e ; E = 1.27 V, pH = 7. (5) A supply of water molecules is needed in each process, and we suggest that sample drying leadsA to supply a decrease of water of current molecules values isneeded because in of each a decli process,ne in reaction and we suggestrate. It is that also sample important drying that leads the tosecond a decrease reaction of currentshows a values tendency because to saturation of a decline [25,28]. in reaction That is rate. the Itresult is also of importantformation that of a theSiO second2 layer reactionon the silicon shows surface a tendency that inhibits to saturation contact [25between,28]. That silicon is the lattice result and of water formation molecules. of a SiO This2 layer effect on is thenot siliconprominent surface for nc-Si that inhibits because contact of the high between specific silicon surface lattice area and of water the material. molecules. This effect is not prominentThe emf for generation nc-Si because in Al/nc-Si/Al of the high structures specific surface is initiated area ofby the the material. diffusion of water vapor through the porousThe emf contact, generation leading in Al/nc-Si/Al to oxidation structures of nc-Si. isThe initiated electrons bythe injected diffusion in the of waterconduction vapor throughband of thenc-Si porous as a result contact, of leadingoxidation, to oxidationreach the oftop nc-Si. contac Thet electronsby hopping, injected and increate the conductiona negative bandcharge of on nc-Si it, aswhereas a result protons of oxidation, diffuse reach to the the bottom top contact contact, by where hopping, they and are createreduced a negative to hydrogen charge (see on Figure it, whereas 7): protons diffuse to the bottom contact, where they are reduced to hydrogen (see Figure7): 2 H+ + 2 e− → H2; E = −0.41 V, pH = 7. (6)

Batteries 2017, 3, 31 8 of 10

+ − 2 H + 2 e → H2; E = −0.41 V, pH = 7. (6)

Batteries 2017, 3, 31 8 of 10

Figure 7. Scheme of the structure investigated, and imaging of potential difference formation. Figure 7. Scheme of the structure investigated, and imaging of potential difference formation. Should the electrodes be almost symmetric (e.g., interdigitated structures, where electrodes are ShoulddepositedFigure the at electrodesthe 7. Schemesame time), of be the almost electrodestructure symmetric investigated, potentials (e.g., wouldand imaging interdigitated be identical of potential and structures, difference emf observation formation. where electrodeswould be are depositedimpossible. at the We same suggest time), that electrode there is potentialsa natural al woulduminium be oxide identical layer andon the emf bottom observation Al electrode, would be impossible.whichShould was We formed the suggest electrodes before that nc-Sibe there almost film is deposition.asymmetric natural aluminium(e.g.,This layerinterdigitated would oxide create structures, layer a barrier on the where between bottom electrodes nc-Si Al electrode, andare the bottom electrode. At the same time, the boundary between the top electrode and nc-Si doesn’t whichdeposited was formed at the before same time), nc-Si electrode film deposition. potentials This would layer be identical would create and emf a barrier observation between would nc-Si be and impossible.contain an aluminium We suggest oxide that layer.there is a natural aluminium oxide layer on the bottom Al electrode, the bottom electrode. At the same time, the boundary between the top electrode and nc-Si doesn’t whichThe was aluminium-Al formed before2O nc-Si3-electrolyte film deposition. contact possesses This layer rectifying would create properties. a barrier Electrons between move nc-Si from and contain an aluminium oxide layer. thethe bottomAl electrode electrode. to electrolyte, At the same lead totime, electrochemical the boundary reactions between on the the top Al 2electrodeO3 surface, and and nc-Si can’t doesn’t return Thecontainto Al. aluminium-Al This an aluminiumeffect is widely2O oxide3-electrolyte used layer. in electrolytic contact ca possessespacitors. Due rectifying to this properties.feature, processes Electrons of hydrogen move from the Alreduction electrodeThe aluminium-Al can to take electrolyte, place2O on3-electrolyte leadthis electrode, to electrochemical contact and possessessignificant reactions rectifying asymmetry on properties. the of Al the2O investigated3 Electronssurface, and movestructure can’t from is return to Al.therealized. This Al electrode effect is widelyto electrolyte, used lead in electrolytic to electrochemical capacitors. reactions Due on to the this Al feature,2O3 surface, processes and can’t of return hydrogen reductionto Al.A canThis significant takeeffect place is portion widely on thisofused charge electrode, in electrolytic transport and cain significant pacitors.the film happensDue asymmetry to this through feature, of diffusion theprocesses investigated of ofH +hydrogen ions structure[19]. is realized.reductionThis is supported can take byplace studies on this of electrode,impedance and spectra significant in our asymmetry previous work of the [29] investigated and the possibility structure ofis Arealized.emf significant generation portion in Nafion of structures. charge transport Proton motion in the can film occur happens by transferring through diffusionalong hydrogen of H+ bondions [19]. grids, or molecules of water adsorbed between nc-Si. + This is supportedA significant by studies portion ofof impedancecharge transport spectra in the in ourfilm previoushappens through work [29 diffusion] and the of possibilityH ions [19]. of emf This isOur supported final investigation by studies of humidityimpedance influence spectra inwas our vacuum previous measurement work [29] and (~10 the−2 possibilityTorr). It was of generation in Nafion structures. Proton motion can occur by transferring along hydrogen bond grids, emfestablished generation that in the Nafion emf value structures. tends toProton zero (Figuremotion 8).can occur by transferring along hydrogen bond or molecules of water adsorbed between nc-Si. grids, or molecules of water adsorbed between nc-Si. −2 OurOur final final investigation investigation of of humidity humidity influence influence waswas vacuum measurement measurement (~10 (~10−2 Torr).Torr). It was It was establishedestablished that that the the emf emf value value tends tends to to zero zero(Figure (Figure8 8).).

Figure 8. VACs of Al/nc-Si/Al structures measured at 300 K: (blue curve) at ambient conditions, (black curve) in vacuum (~10−2 Тоrr).

4. ConclusionsFigure 8. VACs of Al/nc-Si/Al structures measured at 300 K: (blue curve) at ambient conditions, Figure 8. VACs of Al/nc-Si/Al structures measured at 300 K: (blue curve) at ambient conditions, (black (black curve) in vacuum (~10−2 Тоrr). curve)It inwas vacuum shown ((~10 for the−2 Torr).first time that a thin film Al/nc-Si/Al structure with a porous top contact may spontaneously generate emf upon interaction with water vapor in the surrounding atmosphere. 4. Conclusions

It was shown for the first time that a thin film Al/nc-Si/Al structure with a porous top contact may spontaneously generate emf upon interaction with water vapor in the surrounding atmosphere.

Batteries 2017, 3, 31 9 of 10

4. Conclusions It was shown for the first time that a thin film Al/nc-Si/Al structure with a porous top contact may spontaneously generate emf upon interaction with water vapor in the surrounding atmosphere. The most important fact is that such structures are spontaneously formed by a self-assembly process during layer coating. The high porosity of nc-Si film isn’t suitable for efficient charge transfer. Therefore, annealing is needed prior to top electrode deposition. The value of emf increases with the temperature of the structure. Upon analysis of the activational dependences of the emf, it was established that electrical charge leak processes significantly affect the value of emf. We suggest that water reacts with the nc-Si surface, resulting in the formation of an oxygen silicon compound. It is this process that is the origin of energy generation in the structures. Water is thus necessary for the chemical reactions and proton conductivity that lead to emf generation. The role of proton conductivity was confirmed by separation of the structure by a Nafion membrane. It is important to mention that the possibility of emf observation is related to the asymmetry of the two aluminium electrodes. There is an oxide layer between the bottom aluminium electrode and the nc-Si film; on the other hand, there is no such layer between the top electrode and the nc-Si. This feature leads to the impossibility of electron transfer to the bottom electrode, but does not prevent hydrogen reduction on it. Due to the simplicity of the thin structures obtained, such a source may be introduced into electronic paper and employed in the next generation of smart cards. Moreover, it is not necessary to use a liquid electrolyte, and structures may easily be manufactured directly on the surface of silicon chips.

Acknowledgments: This work was supported by Russian Foundation for Basic Research grant No. 15-02-09135. Author Contributions: Dorofeev S.G. and Kononov N.N. conceived and designed the experiments, Bubenov S.S., Kotin P.A., Grigoriev D.V. and Zolotykh A.N. performed the experiments and analyzed the data; Kononov N.N., Bubenov S.S., Kotin P.A. and Dorofeev S.G. wrote the paper. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Zhang, L.L.; Zhao, X.S. Carbon-based materials as supercapacitor electrodes. Chem. Soc. Rev. 2009, 38, 2520–2531. [CrossRef][PubMed] 2. Overmeere, Q.V.; Kerman, K.; Ramanathan, S. Energy Storage in Ultrathin Solid Oxide Fuel Cells. Nano Lett. 2012, 12, 3756–3760. [CrossRef][PubMed] 3. Sang-Min, L.; Kim, Y.-J.; Eoma, S.-W.; Choi, N.-S.; Kim, K.-W.; Cho, S.-B.; Ramanathan, S. Improvement in self-discharge of Zn anode by applying surface modiﬁcation for Zn–air batteries with high energy density. J. Power Sources 2013, 227, 177–184. [CrossRef] 4. Tominaka, S.; Ohta, S.; Obata, H.; Momma, T.; Osaka, T. On-Chip Fuel Cell: Micro Direct Methanol Fuel Cell of an Air-Breathing, Membraneless, and Monolithic Design. J. Am. Chem. Soc. 2008, 130, 10456–10457. [CrossRef][PubMed] 5. Kraytsberg, A.; Ein-Eli, Y. Review on Li–air batteries—Opportunities, limitations and perspective. J. Power Sources 2011, 196, 886–893. [CrossRef] 6. Zhang, Z.; Zuo, C.; Liu, Z.; Yu, Y.; Zuo, Y.; Song, Y. All-solid-state Al–air batteries with polymer alkaline gel electrolyte. J. Power Sources 2014, 251, 470–475. [CrossRef] 7. Cohn, G.; Starosvetsky, D.; Hagiwara, R.; Macdonald, D.D.; Ein-Eli, Y. Silicon–air batteries. Electrochem. Commun. 2009, 11, 1916–1918. [CrossRef] 8. Cohn, G.; Macdonald, D.D.; Ein-Eli, Y. Remarkable Impact of Water on the Discharge Performance of a Silicon–Air Battery. ChemSusChem 2011, 4, 1124–1129. [CrossRef][PubMed] 9. Zhong, X.; Zhang, H.; Liu, Y.; Bai, J.; Liao, L.; Huang, Y.; Duan, X. High-Capacity Silicon–Air Battery in Alkaline Solution. ChemSusChem 2011, 5, 177–180. [CrossRef][PubMed] Batteries 2017, 3, 31 10 of 10

10. Park, D.-W.; Kim, S.; Ocon, J.D.; Abrenica, G.H.A.; Lee, J.K.; Lee, J. Controlled Electrochemical Etching of Nanoporous Si Anodes and Its Discharge Behavior in Alkaline Si–Air Batteries. ACS Appl. Mater. Interfaces 2015, 7, 3126–3132. [CrossRef][PubMed] 11. Kaempgen, M.; Chan, C.K.; Ma, J.; Cui, Y.; Gruner, G. Printable thin film supercapacitors using single-walled carbon nanotubes. Nano Lett. 2009, 9, 1872–1876. [CrossRef][PubMed] 12. Chan, C.K.; Peng, H.; Liu, G.; McIlwrath, K.; Zhang, X.F.; Huggins, R.A.; Cui, Y. High-performance lithium battery anodes using silicon nanowires. Nature Nanotech. 2008, 3, 31–35. [CrossRef][PubMed] 13. Jabbour, L.; Bongiovanni, R.; Chaussy, D.; Gerbaldi, C.; Beneventi, D. Cellulose-based Li-ion batteries: a review. Cellulose 2013, 20, 1523–1545. [CrossRef] 14. Jabbour, L.; Gerbaldi, C.; Chaussy, D.; Zeno, E.; Bodoardod, S.; Beneventi, D. Microfibrillated cellulose–graphite nanocomposites for highly flexible paper-like Li-ion battery electrodes. J. Mater. Chem. 2010, 20, 7344–7347. [CrossRef] 15. Jabbour, L.; Destro, M.; Gerbaldi, C.; Chaussy, D.; Penazzi, N.; Beneventi, D. Aqueous processing of cellulose based paper-anodes for flexible Li-ion batteries. J. Mater. Chem. 2012, 22, 3227–3233. [CrossRef] 16. Hua, L.; Choia, J.W.; Yanga, Y.; Jeongb, S.; La Mantiaa, F.; Cuia, L.-F.; Cui, Y. Highly conductive paper for energy-storage devices. PNAS 2009, 106, 21490–21494. [CrossRef][PubMed] 17. Hu, L.; Wu, H.; Mantia, F.; Yang, Y.; Cui, Y. Thin, Flexible Secondary Li-Ion Paper Batteries. ACS Nano 2010, 4, 5843–5848. [CrossRef][PubMed] 18. Koo, M.; Park, K.; Lee, S.H.; Suh, M.; Jeon, D.Y.; Choi, J.W.; Kang, K.; Lee, K.J. Bendable Inorganic Thin-Film Battery for Fully Flexible Electronic Systems. Nano Lett. 2012, 12, 4810–4816. [CrossRef][PubMed] 19. Rastgar, N.; Rowe, D.J.; Anthony, R.J.; Merritt, B.A.; Kortshagen, U.R.; Aydil, E.S. Effects of Water Adsorption and Surface Oxidation on the Electrical Conductivity of Silicon Nanocrystal Films. J. Phys. Chem. 2013, 117, 4211–4218. [CrossRef] 20. Kuz’min, G.P.; Karasev, M.E.; Khokhlov, E.M.; Kononov, N.N.; Korovin, S.B.; Plotnichenko, V.G.; Polyakov, S.N.; Pustovoy, V.I.; Tikhonevitch, O.V. Nanosize Silicon Powders: The Structure and Optical Properties. Laser Phys. 2000, 10, 939–945. 21. Lechner, R.; Stegner, A.R.; Pereira, R.N.; Dietmueller, R.; Brandt, M.S.; Ebbers, A.; Trocha, M.; Wiggers, H.; Stutzmann, M. Electronic properties of doped silicon nanocrystal films. J. Appl. Phys. 2008, 104, 053701. [CrossRef] 22. Drahi, E.; Blayac, S.; Borbely, A.; Benaben, P. Impact of ink synthesis on processing of inkjet-printed siliconnanoparticle thin films: A comparison of Rapid Thermal Annealing and photonic sintering. Thin Solid Films 2015, 574, 169–176. [CrossRef] 23. Chao, Y.; Siller, L.; Krishnamurthy, S.; Coxon, P.R.; Bangert, U.; Gass, M.; Kjeldgaard, L.; Patole, S.N.; Lie, L.H.; O’Farrell, N.; Alsop, T.A.; Houlton, A.; Horrocks, B.R. Evaporation and deposition of alkyl-capped silicon nanocrystals in ultrahigh vacuum. Nature Nanotech. 2007, 2, 486–489. [CrossRef][PubMed] 24. Seidel, H.; Csepregi, L.; Heuberger, A.; Baumgärtel, H. Anisotropic Etching of Crystalline Silicon in Alkaline Solutions: I. Orientation Dependence and Behavior of Passivation Layers. J. Electrochem. Soc. 1990, 137, 3612–3626. [CrossRef] 25. Goller, B.; Kovalev, D.; Sreseli, O. Nanosilicon in water as a source of hydrogen: size and pH matter. Nanotechnology 2011, 22, 305402. [CrossRef][PubMed] 26. Erogbogbo, F.; Lin, T.; Tucciarone, P.M.; LaJoie, K.M.; Lai, L.; Patki, P.N.; Prasad, P.N.; Swihart, M.T. On-Demand Hydrogen Generation using Nanosilicon: Splitting Water without Light, Heat, or Electricity. Nano Lett. 2013, 13, 451–456. [CrossRef][PubMed] 27. Zhang, X.G. Silicon/Electrolyte Interface. In Electrochemistry of Silicon and Its Oxide, 1st ed.; Kluwer Academic Publisher: Dordrech, The Netherlands, 2004; ISBN 978-0-306-46541-3. [CrossRef] 28. Kale, P.; Gangal, A.C.; Edla, R.; Sharma, P. Investigation of hydrogen storage behavior of silicon nanoparticles. Int. J. Hydrogen Energy 2012, 37, 3741–3747. [CrossRef] 29. Kononov, N.N.; Dorofeev, S.G. Chapter 19: Dielectric and Transport Properties of Thin Films Deposited from Sols with Silicon Nanoparticles. In Smart Nanoparticles Technology, 1st ed.; Hashim, A.A., Ed.; InTech Croatia: Rijeka, Croatia, 2012; Volume 19, pp. 407–440. ISBN 978-953-51-0500-8.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).