Multi-step slow annealing perovskitefilms for high performance
planar perovskite solar cells
Like Huang,Ziyang Hu n,Jie Xu,Ke Zhang,Jing Zhang,Yuejin Zhu n
Department of Microelectronic Science and Engineering,Ningbo Collabrative Innovation Center of Nonlinear Harzard System of Ocean and Atmosphere,
Ningbo University,Ningbo315211,China
a r t i c l e i n f o
Article history:
Received24February2015
Received in revid form
31May2015
Accepted9June2015
Available online25June2015
Keywords:
Perovskite solar cells
One-step direct annealing
Multi-step slow annealing
九年级上册英语单词表
Tightly-distributed performance parameters
a b s t r a c t
The morphology,structure,optical and electrical properties of perovskitefilms treated by two different
annealing methods with different annealing temperature ramp and their corresponding device perfor-
mance have been studied and compared.Annealing temperature ramp significantly influences the sur-
face morphology and optical properties of perovskitefilms which determines the performance of solar
cells which determines the performance of solar cells.The perovskitefilms treated by one-step direct
annealing method tend to exhibit irregular and weak ultraviolet–visible
absorption spectrum,which can
easily result in great variation in thefinal performance of solar cells.While multi-step slow annealing is
beneficial for preparing highly uniform and well-crystallized perovskitefilms,and thus the devices
prent tightly-distributed performance parameters.The best device treated by multi-step slow
annealing method showed a short circuit current density of21.49mA/cm2,an open circuit voltage of
0.988V,afill factor of64.86%,and a power conversion efficiency(PCE)of13.58%,which is a57%
enhancement of the overall PCE relative to8.65%of the device treated by one-step annealing metho
d.
Thefindings suggest that optimized slow temperature ramp is necessary to prepare high-efficient and
well-reproducible perovskite solar cells.
&2015Elvier B.V.All rights rerved.
1.Introduction
Hybrid organic/inorganic ,CH3NH3PbI3and the
mixed-halide perovskite CH3NH3PbI3Àx Cl x)bad solar cells(PVK
SCs)are considered to be economically low cost and have potential
competitiveness with any other type of solar cells becau of the
remarkably high power conversion efficiencies(PCEs)combined
with simple low-temperature and solution-procesd capacity[1–
6].Perovskite materials have almost all the excellent mi-
conducting characteristics required for solar cell,such as small
band-gaps,high extinction coefficients,high carrier mobility,and
large charge diffusion length etc.[2,3].Wherein,CH3NH3PbI3Àx Cl x
is the preferred materials due to its large carrier lifetime and dif-
fusion length that makes it very suitable for planar heterojunction
solar cells(PH PVK SCs)where the perovskitefilm is sandwiched
between the hole and electron lective contacts.Currently,in
order to simplify the preparation process of PVK SCs the simple
planar heterojunction structure is widely adopted[4–6].To realize
high efficient PH PVK SCs,the morphology of the perovskitefilm is
considered to be one of the most critical issues[7].There are many
efforts reported to control the morphology of the perovskite to
obtain high quality perovskitefilm,including controlling the
annealing condition[8–10],adjusting the organic/inorganic pre-
cursor ratio[11],adding additives into the precursor solution[12],
and solvent engineering[13,14].As important process parameters,
annealing temperature and time have a great effect on perovskite
film morphology thus the structural,electrical and optical prop-
erties in perovskite material[7,8,15].For typical one-step solution-
processing of perovskitefilm,annealing temperature and time of
100°C/45min is adopted since the work reported by Lee et al.[16].
For instance,Yang et al.demonstrated a highest efficiency of
$19.3%in PH PVK SCs by annealing the perovskite precursor
solution directly at90°C for1h and100°C for25min[6].
Meanwhile,Snaith et al.investigated influence of the thermal
processing protocol upon the crystallization and photovoltaic
performance of perovskites[17].Shen et al.obrved a60%
increa in PCE for optimized device procesd with two-step
thermal annealing method relative to that of the device prepared
汽车商标using a one-step process(90°C for30min)[18].Kim et al.pro-
pod a stepwi ramp annealing method to control the solvent
evaporation rate to obtain high surface coverage perovskitefilms
[19].According to the previously reported results,we can conclude
that the quality of the perovskitefilms cloly interrelates with the
annealing temperature and duration,and hence the efficiency of
江南逢李龟年翻译Contents lists available at ScienceDirect
journal homepage:/locate/solmat
Solar Energy Materials&Solar Cells
dx.doi/10.1016/j.solmat.2015.06.018
0927-0248/&2015Elvier B.V.All rights rerved.
n Corresponding authors.Tel.:þ8657487600770;fax:þ8657487600744.
E-mail address:huziyang@nbu.edu(Z.Hu),zhuyuejin@nbu.edu(Y.Zhu).
Solar Energy Materials&Solar Cells141(2015)377–382
the device has a great varied distribution.In practice,we have
found that the as spin-coated perovskite precursors on the sub-
strate can easily generate perovskitefilms with abnormal color
variation if being annealed directly at temperature of90–100°C.
Moreover,such perovskitefilms with abnormal light absorption
can result in large deviation in thefinal device performance.This
can result in a low chance of preparing high quality perovskite films,which will hamper the fabrication of large-area perovskite solar cell modules[20].Therefore,exploring a universal annealing
process to realize high-efficient and well-reproducible perovskite
solar cells is the main motivation in this paper.
In this regard,two different annealing process including one-
step(OS)direct annealing method and multi-step(MS)slow
annealing method were introduced to treat the perovskitefilms.写毛笔字的好处
By comparing the corresponding morphology,structure,optical
and electrical properties of the treated perovskitefilms,we found
that the MS method is an universal annealing process to realize
high-efficient and well-reproducible PH PVK SCs.We demon-
strated a maximum PCE of13.58%in MS annealing PVK SCs,
accompanied by a57%enhancement of the overall PCE relative to
8.65%of the OS annealing device.
2.Experimental ction
2.1.Materials
Lead chloride(PbCl2,99.999%),Diethanolamine(98%),4-tert-
Butylpyridine and TiCl4were purchad from Sigma-Aldrich,
CH3NH3I from Shanghai Materwin New Materials Co.Ltd.,Titanium
(IV)isopropoxide(98þ%)and Li-bis(trifluoromethanesulfonyl)
imide(Li-TFSI)from Acros,spiro-OMeTAD from Lumtec,dimethyl-
formamide(DMF),acetonitrile,isopropanol,ethanol and chlor-
obenzene from Shanghai Chemical Agent Ltd.,China(Analysis purity
grade).All materials were ud as received.
2.2.Synthesis of CH3NH3I and TiO2nanocrystals
Methyl ammonium iodide(CH3NH3I)was synthesized by the
available process as reported in literature[6].First,18.7mL
(0.15mol)methylamine(Sigma-Aldrich,33wt%in absolute etha-
nol)and19.8mL(0.15mol)hydroiodic acid(Sigma-Aldrich,
99.99%,57wt%in water)at a1:1equimolar ratio were stirred in an
ice bath for2h.Then the precipitate was collected by evaporating
at50°C for2h.Finally,a white powder was received by washing
with diethyl ether and ethanol three times and then drying at
100°C in a vacuum oven for24h.
The TiO2nanocrystals were synthesized from a sol–gel method
in the ambient air[21].Simply,0.675mL of titanium(IV)iso-
propoxide was added to18mL of isopropanol and0.25g of die-
thanolamine;18μL of deionized water were added before stirring for5min at room temperature,then the sol was left to age for half
an hour before using.
2.3.Solar cell device fabrication
FTO coated glass substrates(Nippon,14Ω/□)were cleaned with deionized water,ethanol and acetone and then subjected to an ozone-ultraviolet treatment for15min.A40nm thick TiO2compact (c-TiO2)layer was spin-coated on the substrates using the sol–gel solution synthesized above with a spin speed and time of4000rpm/ 25s.The c-TiO2layer was then annealed at450°C for30min.Before using,the substrates were treated in a aqueous solution of TiCl4 (0.04M)at70°C for30min,then rind with deionized water and dried at120°C for15min.To deposit the perovskite layer,a1:3ratio of PbCl2/CH3NH3I(0.8M and2.4M)was mixed in DMF.The solution was spin-coated on the FTO/c-TiO2substrates at2000rpm for40–
50s and then treated with two different annealing methods descri-电脑卡顿怎么解决
bed below.Once perovskite thinfilms grow well,a hole transport
layer(HTL)solution was spin-coated at2800rpm for30s,in which
1mL spiro-OMeTAD/chlorobenzene(72.3mg/mL)solution was
employed with addition of18m L Li-TFSI/acetonitrile(520mg/mL),
and29m L4-tert-butylpyridine.Lastly,a120nm thick silver layer was
thermally evaporated on top of the device under a pressure of
5Â10À6Torr to form the back contact.Apart from the c-TiO2layer,all
functional layers were prepared in a nitrogenfilled glove box.The
devices fabricated have a layered p–i–n configuration glass/FTO/c-
TiO2/CH3NH3PbI3Àx Cl x(PVK)/spiro-MeOTAD/Ag,which consists of $450nm thick FTO electrode,$40nm of TiO2,$400nm of per-ovskite,$400nm of spiro-MeOTAD,and$120nm of Ag,as shown
in Fig.1(a and b).The schematic diagram for fabricating such a device
is depicted in Fig.1(c).
2.4.Characterizations and measurements
The UV–vis absorption spectra of the perovskitefilms were
recorded with a VARIAN Cary5000UV–vis–NIR spectrophotometer.
The X-ray diffraction(XRD)pattern(2θscans)were obtained from perovskitefilms deposited on the FTO/c-TiO2substrates using a
Bruker AXS D8Advance X-ray diffractometer using Cu-Kαradiation (λ¼1.54050Å).The top view images and thickness of the depos-ited perovskitefilms were confirmed by a Hitachi SU-70scanning electron microscope(SEM).Atomic force microscope(AFM)topo-graphic images of the perovskitefilms deposited on the FTO/c-TiO2 substrates were taken using a Bruker Dimension5000Scanning Probe Microscope(SPM)in tapping mode.Steady-state photo-luminescence spectroscopy(PL)measurements were acquired using an Edinburgh Instruments FLS920fluorescence spectrometer with an excitation wavelength of460nm.
The current density–voltage(J–V)measurements were con-
ducted under simulated AM1.5G sunlight of100mW/cm2using
an AM1.5G typefilter(Newport,81904,USA).The light intensity
was adjusted by using a standard Si cell.J–V curves were obtained
by applying an external bias to the cell,and measurements were
recorded with a Keithley model2400digital source meter at room
temperature in the ambient air.The incident photo-current con-
version efficiency(IPCE)spectrum was measured by a IPCE system
(Newport2936-c power meter)in the300–800-nm wavelength清时有味是无能
range under the irradiation of a300W xenon light source with an
Oriel Cornerstone2601/4monochromator in DC mode at room
temperature.The effective area of the cell was defined to be
0.07cm2by a non-reflective metal mask.
3.Results and discussion
Two different annealing process,OS and MS,were adopted
for perovskitefilm treatment in fabrication of PH PVK SC devices
as shown in Fig.1.The ints in Fig.1(d)correspond to0min(1),
1min(2),10min(3),20min(4),30min(5),45min(6)while
tho of Fig.1(e)correspond to0min(1),15min(2),40min(3),
60min(4),80min(5),105min(6).Generally,a transparent yel-
lowishfilm forms originally once deposition of the precursor
solution on the FTO/c-TiO2substrate is completed.Leaving thefilm
at room temperature for veral minutes,the color changes to red,
then deep yellow andfinally deep black[22].This change in color
corresponds to the solvent evaporation,reaction of the precursor
and sublimation of excess organic CH3NH3Cl as reaction byproduct
[7].Obviously,as shown in Fig.1(d,e),the color change of per-
ovskitefilms treated with the OS method is faster than that of
perovskitefilms treated with the MS method.It is reasonable that
precursor reaction and byproducts sublimation occurred faster
L.Huang et al./Solar Energy Materials&Solar Cells141(2015)377–382 378
due to elevated temperature since the beginning of annealing.Becau the melting point of precursor CH 3NH 3I (about 70°C)is signi ficantly lower than the boiling point of DMF (about 150°C),rapid annealing such as the OS method may cau too fast and uncontrollable evaporation of solvent accompanied with CH 3NH 3I,which may result in reaction stoichiometry ratio mismatch of the
0102030405060708090100
102030405060708090100
Time (min)
Time (min)T e m p e r a t u r e C )
6
5
4
3
凤箫吟2
1Fig.1.Device structure (a),scanning electron microscope cross-ctional image of the device (b),and process steps for the fabrication of the perovskite solar cells (c),and two different annealing methods adopted for the treatment of perovskite films (d),OS and MS methods (e).The ints are the photographs of the perovskite films.
Fig.2.(a)UV –vis absorption spectra of two reprentative perovskite films annealed with different methods,(b)XRD patterns for perovskite films annealed with different methods,(c)the photovoltaic performance of their corresponding devices,and (d)the corresponding incident photo-current conversion ef ficiency spectra.
L.Huang et al./Solar Energy Materials &Solar Cells 141(2015)377–382379
precursor.Actually,we didfind that the u of OS method is dif-ficult to obtain high quality perovskitefilms compared with the MS method.The color of a typical high quality perovskitefilm is reddish-brown as shown in the int of Fig.1(e),which is also reported by the previous reports[6,7,22].However,in our experiments thefilms treated with the OS method exhibit grayish-purple or grayish-green as shown in the int of Fig.1(d).Even under a prolonged annealing time,the color was not changed.
We characterized thefilms prepared using the two different methods by UV–vis spectroscopy as shown in Fig.2(a).Compared to the absorption spectrum curves of perovskitefilms treated with the MS method,the absorption curve shape and intensity of the per-ovskitefilms treated by the OS method is irregular and weak.The perovskitefilms exhibit the760nm absorption peak corresponding to the direct band gap transition from thefirst valence band(VB1)to conduction band(CB)[16],regardless of the different annealing methods adopted.While the absorption spectrum of the grayish-purple and grayish-green perovskitefilms treated by the OS method scarcely contains the480nm absorption peak corresponding to the transition from the cond valence band(VB2)to CB[3].Note that the absorption spectra curves of the grayish-purple and grayish-green perovskitefi
lms are very similar with tho of the initial annealingfilms(annealed at95°C/65min)with the MS method.The stable non-red-brown color regardless of annealing time implies that thefilms suffered irreversible changes at a direct annealing tem-perature of95°C,possibly resulting from too fast CH3NH3I eva-poration accompanied with the solvent DMF.Since light absorption is thefirst step for a solar cell as a photoelectric conversion device,one can anticipate that such a great variation in UV–vis absorption can lead to significant deviation in thefinal device performance.
However,the XRD measurements(Fig.2(b))do not show any difference in diffraction peaks position between the perovskite films treated by the two annealing methods.The peaks at14.2°, 28.5°,43.3°and59.0°are assigned to the(110),(220),(330)and (440)planes[6,16],respectively,similar to tho reported for the tetragonal perovskite structured CH3NH3PbI3crystals.While the intensity of each diffraction peak of the perovskitefilms treated by the OS method is weaker than that of perovskitefilms treated by the MS method and no new peaks or peak shifts were obrved. This reveals that the crystal structure of the two differently annealed perovskitefilms is the same and the only difference between them is the degree of crystallinity.The slow solvent evaporation in MS method facilitates atomic rearrangement which leads to increa in crystalline fraction and hence the enhanced peak intensities.
According to the different light absorption spectra and crys-tallinity of thefilms treated by the two methods,we can expect different photovoltaic performance.Fig.2(c)shows the current density–voltage(J–V)measurements of their corresponding PVK SC devices.For device OS,it contains perovskitefilm treated by the OS method,thefilm is abnormally grayish-purple,the typical device shows PCE¼3.60%,J SC¼8.42mA/cm2,V OC¼0.847V,and FF¼49.63%.For the device MS,it contains perovskitefilm treated by the MS method,thefilm is normally reddish-brown,the typical device shows PCE¼10.14%,J SC¼17.56mA/cm2,V OC¼0.931V,and FF¼60.97%.The dark J–V curve of a solar cell device generally reflects its rectifying characteristics as a diode.Compared to the device OS,the device MS behaved as well defined diodes with better rectifying characteristics as its dark J–V curve exhibits larger turn-on voltage.Additional,for the device OS,although the per-ovskitefilm UV–vis absorption spectrum contains only one absorption peak of760nm,it still has a large V OC of0.847V,as if the V OC is decided only by thefirst valence band(VB1)and is independent of the cond valence band(VB2),regardless of other factors that affect V OC[23].Fig.2(d)prents the incident photo-current conversion efficiency(IPCE)spectra.Taking into account the fact that the smaller IPCE value of the device OS relative to that of the device MS in the entire visible wavelength,we can conclude that the grayish-purple and grayish-green perovskitefilms are not the desirablefilms for high performance.
We further examined the influence of the different annealing processing on the morphology and structure of the perovskite layer formed on the c-TiO2layer by SEM as shown in Fig.3.It is believed that for efficient PVK SCs,aflat,highly uniform,pin-hole-free and high surface coverage CH3NH3PbI3Àx Cl xfilm is of extreme significance,as required in other thinfilm solar cells.We can e that thefilms produced by the OS method(Fig.3(a))contain dif-ferent size grains with an incomplete coverage on the substrate, while thefilms produced by the MS method(Fig.3(b))are very uniform.A pinhole with a feature size of$2m m has been high-lighted in Fig.3(a).It should be noted that the prence of pinholes thus the lower coverage of the perovskitefilm could have a sig-nificant impact on thefinal UV–vis absorption spectra measured, as light may go directly through the pinholes without being absorbed by perovskite.This may be one reason that the above measured perovskite thinfilms prepared by the OS method showed low absorption intensity.Otherwi,we can expect that the coverage of perovskite would have a significant effect on J SC of thefinal devices.It is the reason that the device MS prents the double J SC of the device OS(Fig.2(c)).In addition,we notice that many small particulates are gathered at the step edges of the layered perovskite crystals(Fig.3S).We suppo that the par-ticulates were precipitated from the perovskitefilms becau of the rapid annealing process.Actually,we can reduce the pin-holes and particles by surface solvent treatment[14].
The surface morphology of the perovskitefilms was also char-acterized by the AFM measurement as shown in Fig.4(a,b).The AFM images(8Â8μm2)show that thefilms produced by the
MS Fig.3.SEM images for perovskitefilms annealed with OS method(a),MS method(b).The ints are high-magnification SEM images.
掌故是什么意思L.Huang et al./Solar Energy Materials&Solar Cells141(2015)377–382
380
method has low roughness with a root-mean-square (RMS)of 35nm while the film prepared by the OS method has a high roughness with a RMS of 72nm.Consistent with the SEM obr-vations,the crystal arrangement in the film produced by the MS method is more compact as no pinholes can be found,which ensures suf ficient light absorption and avoids possible short-circuit occurrence.It is believed that high temperature and rapid annealing treatment can result in the rapid growth of perovskite crystals and lead to the formation of large crystalline islands and the associated large gaps.Lower temperature and slow annealing treatment allow the perovskite crystals to grow slowly and uniformly from a large number of nucleation sites and result in uniform crystal structure with a small number of internal voids or pinholes.Thus,slow annealing may be necessary to fabricate highly uniform perovskite films without pinholes.
The film quality may also have a great impact on the electron transport properties of perovskite film.As shown by SEM images,the two films contain varying degrees of surface defects which will have effects on the device performance.In order to compare the crystal defects inside the films,steady-state photoluminescence spectroscopy (PL)was conducted as shown in Fig.4(c).The enhanced PL spectroscopy of the perovskite films treated by the MS method con firms the improved crystallinity of the perovskite films.This can be attributed to the number of crystal defects is well redu
ced,therefore the non-radiative recombination paths are greatly elimi-nated.Thus,we can anticipate that the carrier can be collected by the corresponding lective contacts ef ficiently without too much recombination induced by crystal defects .The perovskite films treated by the OS method exhibit weak fluorescence intensity,which suggests that the films contains many physical defects or bulk traps that act as carrier recombination centers.Fig.5(a)displays the histograms of PCE of the 48devices from 4different batches.It is shown that the PCEs were distributed within 8–14%for the MS method procesd PVK SCs and 0–9%for the OS method procesd PVK SCs,respectively.Obviously the PCE values of the former exhibits a narrow distribution with a small deviation about 12%from the average values.While the PCE values of the later show a relatively wide distribution with a large deviation of about 48%from the average values.The origin of this variation in PCE values of the device OS here is primarily attributed to the abnormal characteristics of the perovskite films that are treated by the OS methods.Furthermore,the distribution of PCE values of the solar cells was fabricated using the MS method is fairly narrow compared to that of the OS method procesd PVK SCs.
Finally,we compare the stability of the two kinds of devices treated with different annealing methods.The tested OS and MS devices without encapsulation were stored at a glove box full with d
ry nitrogen and tested outside at every 24h for 7days.The results are shown in Fig.5(b).The devices OS retain 20–25%of the initial performance after 72h,and 10–15%after 7days,respec-tively.While the devices MS retains more than 60%of the initial performance after 72h,and 30–35%after 7days,respectively.Here,the enhanced stability of the devices MS was attributed to the slower decomposition of the MS annealing perovskite films induced by fewer molecular defects as the PL measurement reveals.The growth of a more stable perovskite material is of virtual signi ficance for photovoltaic applications.Even in the abnce of humidity,a decomposition of the perovskite structure can take place through the statistical formation of molecular defects with a non-ionic character,who volatility at surfaces should break the thermodynamic defect equilibriums [24].The strategies that can substantially prolong the lifetime of the material were also reported [14,25,26].Here,by comparing
the
Fig.4.AFM images for perovskite films annealed with OS method (a),MS method (b),and PL spectroscopy of the perovskite films
(c).
Fig.5.(a)Histograms of the photoelectric conversion performance of 48devices with perovskite films annealed with two different methods from 4different batches,and (b)stability investigation on the unencapsulated devices OS and MS as a function of time.
L.Huang et al./Solar Energy Materials &Solar Cells 141(2015)377–382381
