Solution synthesis and characterization of indium-zinc formate precursors for transparent conducting oxides

Article abstract of DOI:10.1021/ic902430w

A series of In-Zn formate mixtures were investigated as potential precursors to amorphous In-Zn-oxide (IZO) for transparent conducting oxide (TCO) applications. These mixtures were prepared by neutralization from formic acid and characterized by elemental analysis, IR spectroscopy, powder X-ray diffraction, and thermogravimetry-differential scanning calorimetry (TG-DSC) measurements. Thermal analysis revealed that a mixture of In and Zn formates reduced the overall decomposition temperature compared to the individual constituents and that OH-substitution enhanced the effect. In terms of precursor feasibility, it was demonstrated that the decomposition products of In-Zn formate could be directed toward oxidation or reduction by controlling the decomposition atmosphere or with solution acid additives. For TCO applications, amorphous IZO films were prepared by ultrasonic spray deposition from In-Zn formate solutions with annealing at 300-400 °C.

Full text of DOI:10.1021/ic902430w

5424 Inorg. Chem. 2010, 49, 5424–5431
DOI: 10.1021/ic902430w
Solution Synthesis and Characterization of Indium-Zinc Formate Precursors for
Transparent Conducting Oxides
Robert M. Pasquarelli,*,† Calvin J. Curtis,‡ Alexander Miedaner,‡ Maikel F.A.M. van Hest,‡ Ryan P. O’Hayre,† and
David S. Ginley‡
†Colorado School of Mines, Metallurgical and Materials Engineering, 1500 Illinois Street, Golden, Colorado
80401, and ‡National Renewable Energy Laboratory, 1617 Cole Boulevard, Golden, Colorado 80401
Received December 7, 2009
A series of In-Zn formate mixtures were investigated as potential precursors to amorphous In-Zn-oxide (IZO) for
transparent conducting oxide (TCO) applications. These mixtures were prepared by neutralization from formic acid
and characterized by elemental analysis, IR spectroscopy, powder X-ray diffraction, and thermogravimetry-differential
scanning calorimetry (TG-DSC) measurements. Thermal analysis revealed that a mixture of In and Zn formates
reduced the overall decomposition temperature compared to the individual constituents and that OH-substitution
enhanced the effect. In terms of precursor feasibility, it was demonstrated that the decomposition products of
In-Zn formate could be directed toward oxidation or reduction by controlling the decomposition atmosphere or with
solution acid additives. For TCO applications, amorphous IZO films were prepared by ultrasonic spray deposition from
In-Zn formate solutions with annealing at 300-400 °C.
Introduction
typically in the form of acetates, chlorides, nitrates, or
acetylacetonates.9-19 Precursors for In and Zn are of parti-
cular interest as many now consider amorphous In-Zn-oxide
(a-IZO) to be the archetype amorphous oxide semiconductor
material.6-8,20
While examples of the use of acetate precursors are
numerous,9-16 the use of formate precursors for TCOs is
rare.21 Here we investigate the synthesis and properties of the
formates of In and Zn as potential IZO precursors. Formate
is the simplest carboxylate form, thus having the smallest
steric effect. It can bind to metal cations in various ways as
mono- or polydentate ligands. The formates of divalent 3d
transition metals (M = Zn, Mg, Cu, Mn, Ni, Cd)22-29 are
Transparent conducting oxide (TCO) thin films play a
critical role in many current and emerging opto-electronic
devices. Their unique combination of high transparency in
the visible region of the spectrum and tunable electronic
conductivity is ideal for applications in displays, transparent
thin-film transistors (TTFTs), and solar cells.1-3 TCO com-
position-space is composed of mixed binary and ternary
oxides of Sn, In, Zn, Ga, and Al.4-8 Solution deposition
routes to these metal oxides employ chemical precursors,
*To whom correspondence should be addressed. E-mail: robert.pasquarelli@
nrel.gov.
(1) Lewis, B. G.; Paine, D. C. MRS Bull. 2000, 25, 22.
(2) Ginley, D. S.; Bright, C. MRS Bull. 2000, 25, 15.
(3) Fortunato, E.; Ginley, D. S.; Hosono, H.; Paine, D. C. MRS Bull.
2007, 32, 242.
(15) Wienke, J.; van der Zanden, B.; Tijssen, M.; Zeman, M. Sol. Energy
Mater. Sol. Cells 2008, 92, 884.
(16) Kim, G. H.; Shin, H. S.; Ahn, B. D.; Kim, K. H.; Park, W. J.; Kim,
H. J. J. Electrochem. Soc. 2009, 156, H7.
(17) Agashe, C.; Hupkes, J.; Schope, G.; Berginski, M. Sol. Energy Mater.
Sol. Cells 2009, 93, 1256.
(18) Maldonado, A.; de la Luz Olvera, M.; Tirado Guerrab, S.;
Asomozaa, R. Sol. Energy Mater. Sol. Cells 2004, 82, 75.
(19) Moses Ezhil Raja, A.; Lalithambikab, K. C.; Vidhyac, V. S.; Rajagopalc,
G.; Thayumanavand, A.; Jayachandranc, M.; Sanjeevirajae, C. Physica B 2008,
403, 544.
(4) Minami, T. MRS Bull. 2000, 25, 38.
(5) Coutts, T. J.; Young, D. L.; Li, X. MRS Bull. 2000, 25, 58.
(6) Taylor, M. P.; Readey, D. W.; Teplin, C. W.; van Hest, M. F. A. M.;
Alleman, J. L.; Dabney, M. S.; Gedvilas, L. M.; Keyes, B. M.; To, B.;
Perkins, J. D.; Ginley, D. S. Meas. Sci. Technol. 2005, 16, 90.
(7) Taylor, M. P.; Readey, D. W.; van Hest, M. F. A. M.; Teplin, C. W.;
Alleman, J. L.; Dabney, M. S.; Gedvilas, L. M.; Keyes, B. M.; To, B.;
Perkins, J. D.; Ginley, D. S. Adv. Funct. Mater. 2008, 18(20), 3169.
(8) Perkins, J. D.; van Hest, M. F. A. M.; Taylor, M. P.; Ginley, D. S. Int.
J. Nanotechnol. 2009, 6(9), 850.
(9) Jin-Hong, L.; Byung-Ok, P. Thin Solid Films 2003, 426, 94.
(10) Chang-Hyun, K.; Jin-Hong, L.; Byung-Ok, P. Mater. Sci. Forum
2004, 449-452, 469.
(11) Seung-Yup, L.; Byung-Ok, P. Thin Solid Films 2005, 484, 184.
(12) de la Luz Olvera, M.; Gomez, H.; Maldonado, A. Sol. Energy Mater.
Sol. Cells 2007, 91, 1449.
(20) Hosono, H. J. Non-Cryst. Solids 2006, 352, 851.
(21) Seki, S.; Sawada, Y.; Ogawa, M.; Yamamoto, M.; Kagota, Y.; Shida,
A.; Ide, M. Surf. Coat. Technol. 2003, 167-170, 525.
ꢀ
(22) Osaki, K.; Nakai, Y.; Watanabe, T. J. Phys. Soc. Jpn. 1963, 18, 919.
(23) Burger, V. N.; Fuess, H. Z. Kristallogr. 1977, 145, 346.
(24) de With, G.; Harkema, S.; van Hummel, G. J. Acta Crystallogr. 1976,
B32, 1980.
(13) Choi, C. G.; Seo, S.-J.; Bae, B.-S. Electrochem. Solid-State Lett. 2008,
11, H7.
(14) Wienke, J.; Booij, A. S. Thin Solid Films 2008, 516, 4508.
(25) Bukowska-Strzyzewska, M. Acta Crystallogr. 1965, 19, 357.
(26) Radhakrishna, P.; Gillon, B.; Chevrier, G. J. Phys.: Condens. Matter
1993, 5(35), 6447.
r
pubs.acs.org/IC
Published on Web 05/24/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5425
well-known and form isostructural-monoclinic crystals with
and 30 s scan time. Elemental analysis of carbon and hydrogen
was performed by Huffman Laboratories Inc. (Golden, CO)
with a combustion microanalyzer with coulometric detection. A
Bruker AXS D8 Discover X-ray diffraction (XRD) system with
a Cu KR radiation source and large area GADDS detector was
used for structural analysis. The calculated molecular formulas
of the products are based on a combination of metals, C-H, IR,
and thermogravimetric analyses.
the general formula M(HCOO)2 2H2O. In their dihydrate
3
form, the formate ions bind μ2. Conversely, in their anhy-
drous form M(HCOO)2 (M = Zn, Mg, Cu, Mn),30-33 the
formate ions coordinate μ3. Zn(HCOO)2 2H2O has been
3
synthesized by neutralizing the oxide34 and hydroxide
carbonate35 with formic acid. Trivalent metal formates have
been studied less extensively than divalent metal formates.
The indium formates consist of InO6 octahedra and exhibit
various formate and hydroxyl bridging modes. In(HCOO)3
(hexagonal), In2(HCOO)5(OH) (monoclinic), and In(HCOO)2-
(OH) (tetragonal) have been produced by hydrothermal
synthesis.36 In(HCOO)3 hasalso beenmadeby neutralization
of the oxide.21
Precursor solutions consisted of 0.1 M In-Zn formate (3a) in
a methanol-water-acid (6:1:0.2 vol) mixture. HCOOH and
conc. HNO3 were used as acid additives. Drop-casting was
performed onto microscope glass at 300 °C in air. Films were
deposited using a custom ultrasonic spray system consisting of a
Sono-Tek Model 8700 spray head driven at 120 kHz, a N2 fed
directional nozzle, a Fluid Metering Inc. VMP TRI pump, and a
heated platen mounted on X-Y slide stages. Films were annealed
in a Lindberg Model 59344 tube furnace with a 3 h ramp and
20 min hold at 300-400 °C under Ar-4%H2.
In this article, we report on the characterization of a series
of co-synthesized In-Zn formate mixtures exhibiting various
degrees of hydration and OH-substitution prepared by neu-
Synthesis of Zn(HCOO)2 2H2O (1). HCOOH (0.66 mol,
3
tralization from formic acid: Zn(HCOO)2 2H2O þ hexago-
3
25 mL) was added with stirring to zinc hydroxide carbonate
(7.36 mmol, 4.04 g) in a 50 mL round-bottom flask fitted with a
condenser vented to atmosphere. The mixture bubbled upon
addition, and a white suspension resulted. The mixture was
heated at 80 °C in an oil bath for 5 h. After cooling to room
temperature, the precipitate was collected via suction filtration,
washed with tetrahydrofuran (THF), and then dried under
aspiration to remove residual formic acid. Small white crystals
were obtained (3.26 g, 59% yield). IR spectrum (cm-1): 3346m,
3268m, 3170m, 2984vw, 2900w, 1667w, 1560vs, 1396m, 1374s,
1352vs, 878w, 836w, 761m, 722m, and 565m. Anal. Calcd for
nal In(HCOO)2(OH), anhydrous Zn(HCOO)2 þ hexagonal
In(HCOO)3, and Zn(HCOO)2 2H2O þ tetragonal In-
3
(HCOO)2(OH) with compositions ranging 70-86 cation%
In. Thermal analysis revealed that a mixture of In and Zn
formates reduced the overall decomposition temperature
compared to the individual constituents and that OH-sub-
stitution enhanced the effect. As a precursor, it was demon-
strated that the decomposition products of In-Zn formate
could be directed toward oxidation or reduction by control-
ling the decomposition atmosphere or with solution additives
such as acids. Additionally, as demonstration for TCO
applications, amorphous IZO films were solution deposited
using In-Zn formates.
Zn(HCOO)2 2H2O (FW 191.47): C 12.55, H 3.16. Found: C
3
12.60, H 3.26.
Synthesis of In(HCOO)3 (2). HCOOH (0.66 mol, 25 mL) was
added with stirring to In2O3 (10.9 mmol, 3.04 g) in a 50 mL
round-bottom flask fitted with a condenser. The mixture was
heated at 80 °C in an oil bath for 2 weeks, over which the
suspension gradually whitened in appearance. The product was
collected and washed in the same way as 1. A white powder with
yellow tint was obtained (5.45 g, 99% yield). IR spectrum
(cm-1): 2986vw, 2913w, 1633m, 1572s, 1543vs, 1402m, 1387s,
1363vs, and 792vs. Anal. Calcd for In(HCOO)3 (FW 249.87): C
14.42, H 1.21. Found: C 14.08, H 1.21.
Experimental Section
Materials and Instrumentation. Zinc hydroxide carbonate
[ZnCO3]2 [Zn(OH)2]3 and In2O3 were obtained from Riedel-de
3
€
Haen and Materials Research Corporation, respectively. Formic
acid (HCOOH) and solvents were from Sigma-Aldrich. Starting
materials were used as supplied without further purification.
Infrared spectroscopy was performed on a Nicolet Nexus 870
FTIR spectrometer equipped with a Pike Technologies MIRacle
diamond ATR at 2 cm-1 resolution. To determine the decom-
position behavior, simultaneous thermogravimetry (TG) and
differential scanning calorimetry (DSC) were performed under
dry air and N2 atmospheres in alumina pans with a heating rate
of 10 °C/min using a TA Instruments SDT Q600 analyzer.
Precursor metal ratios were determined by inductively coupled
plasma atomic emission spectroscopy (ICP-AES) and con-
firmed by X-ray fluorescence (XRF). ICP-AES solutions in
5% HNO3-deionized water were analyzed on a Varian Liberty
150 spectrometer. XRF measurements were performed on a
Matrix Metrologies MaXXi 5 with W anode, 800-μm collimator,
Synthesis of InxZn(1-x)(HCOO)(2þx-y)(OH)y zH2O (3). Pre-
3
pared with a starting In:Zn of 65:35 at % (3a). HCOOH (0.80
mol, 30 mL) was added with rapid stirring to fine mixture of
In2O3 (13.0 mmol, 3.61 g, 26 mmol In) and zinc hydroxide
carbonate (2.81 mmol, 1.54 g, 14 mmol Zn) in a 50 mL round-
bottom flask fitted with a condenser. The reaction was heated at
80 °C in an oil bath for 2 weeks. Over the initial several hours, the
mixture turned off-white and thickened. The final product was
collected and washed in the same way as 1. A white powder was
obtained (8.24 g, 95% yield). IR spectrum (cm-1): 3415w,
2892w, 3342vw, 3277vs, 1550vs, 1388m, 1370s, 1353vs, 1040m,
969w, 880w, 795m, 760m, 753m, 635w, and 567s. Anal. Calcd
for In0.7Zn0.3(HCOO)2(OH)0.7 0.6H2O (FW 212.74): C 11.29,
H 1.85. Found: C 11.68, H 1.82.
3
(27) Kageyama, H.; Khomskii, D. I.; Levitin, R. Z.; Vasil’ev, A. N. Phys.
Rev. B 2003, 67(22), 224422.
(28) Post, M. L.; Trotter, J. Acta Crystallogr. 1974, B30, 1880.
(29) Stoilova, D. J. Mol. Struct. 2006, 798, 141.
(30) Viertelhaus, M.; Anson, C. E.; Powell, A. K. Z. Anorg. Allg. Chem.
2005, 631(12), 2365.
(31) Dollimore, D.; Tonge, K. H. J. Inorg. Nucl. Chem. 1969, 29, 621.
(32) Sapina, F.; Burgos, M.; Escriva, E.; Folgado, J. V.; Marcos, D.;
Beltran, A.; Beltran, D. Inorg. Chem. 2002, 32(20), 4337.
(33) Viertelhaus, M.; Henke, H.; Anson, C. E.; Powell, A. K. Eur. J.
Inorg. Chem. 2003, 2003(12), 2283.
(34) Stoilova, D.; Baggio, R.; Garland, M. T.; Marinova, D. Vib.
Spectrosc. 2007, 43(2), 387.
(35) Vassileva, V. Z. Croat. Chem. Acta 2005, 78(2), 295.
(36) Su, J.; Wang, Y.; Yang, S.; Li, G.; Liao, F.; Lin, J. Inorg. Chem. 2007,
46, 8403.
3b was prepared in the same way as 3a but with a starting In:
Zn of 70:30 at % using 14.0 mmol In2O3 and 2.40 mmol zinc
hydroxide carbonate. IR spectrum (cm-1): 2986vw, 2915w,
2892vw, 1633m, 1572s, 1550vs, 1402m, 1387s, 1363vs, 1329s,
1152w, and 792vs. Anal. Calcd for In0.74Zn0.26(HCOO)2.74 (FW
225.32): C 14.61, H 1.23. Found: C 14.81, H 1.34.
Synthesis of InxZn(1-x)(HCOO)(2þx-y)(OH)y zH2O with
3
H2O (4). Prepared in the same way as 3a but using a HCOOH-
deionized water (1:1 vol) addition and a reaction period of
1 week. 2-3 mm long crystals were observed growing out of
the precipitate. The white precipitate (4a) (6.50 g, 75% yield) and
crystals (4b) (71.7 mg) were separated and washed. IR spectrum
of 4a (cm-1): 3415m, 3355w, 3272w, 2978vw, 2908w, 1604m,

5426 Inorganic Chemistry, Vol. 49, No. 12, 2010
Pasquarelli et al.
Table 1. Comparison of Cation Ratio for Reagent Mixture and Final Product
In and that Zn(HCOO)2 2H2O crystals grew out of the
solution upon cooling (4b).
3
cation ratio (at % In)
Second, not only was the composition difficult to
reproduce between batches, but it also varied within a
batch. XRF provided a non-destructive method to ran-
domly sample each powder within a ∼800 μm spot size to
examine compositional uniformity. Compositional vari-
ation was 2.6-5.4% (relative percent) over different spots
within a batch. This is significantly larger than the
instrumental precision of 0.9%. For example, the con-
centration of 3a (1) at nominally 71 cation% In actually
varied by (2.1 absolute cation% (a 2.9% relative
variation) depending on what portion of the powder
sample was analyzed. The non-uniformity can be attrib-
uted to the reaction involving a suspended solid phase as
opposed to a homogeneous solution. Furthermore, the
collected material was a mixture of two separate com-
pounds. The effect was most pronounced in the case of 3a
(2), where difficultly was encountered stirring the reaction
mixture, resulting in a variation of 8.6%.
Overall, the difficulties in tailoring the final cation
composition and maintaining compositional uniformity
bring into question the practicality of co-synthesized
In-Zn formates as potential single-source precursors
for TCOs. Investigations into more reproducible syn-
thetic routes are ongoing. While such control is desirable
for large-scale production consistency, a single composi-
tion is not required for a-IZO phase formation from a
materials perspective. Taylor et al. has shown that the
material is amorphous over a large composition region
from 55 to 84 cation% In when sputtered. Furthermore,
maximum conductivity can be achieved in this region by
not only tuning the cation ratio but also by adjusting
oxygen content in the sputter gas.7,8 This behavior in the
sputtered material suggests that a variation of cation
composition in a solution processed material because of
the precursor could be compensated for by controlling
oxygen content in the film during deposition or with a
post-anneal treatment to achieve the desired electronic
and optical properties.
starting material product by ICP-AES product by XRF
3a (1)
3a (2)
3b
4a
4b
65
65
70
65
65
69.8 ( 0.6
66.6 ( 1.2
73.9 ( 0.7
86.1 ( 0.5
1.05 ( 0.05
71.4 ( 2.1
73.1 ( 6.3
74.1 ( 1.9
82.7 ( 4.5
1.55 ( 0.25
1543vs, 1396w, 1383m, 1352vs, 1036s, 879w, 838w, 796s,
767s, 634s, and 566s. Anal. Calcd for In0.83Zn0.17(HCOO)2-
(OH)0.83 0.34H2O: C 11.09, H 1.63. Found: C 11.06, H 1.68. IR
3
spectrum of 4b (cm-1): 3346m, 3268m, 3170m, 2984vw, 2900w,
1667w, 1560vs, 1396m, 1374s, 1352vs, 878w, 836w, 761m, 722m,
and 565m. Anal. Calcd for Zn0.99In0.01(HCOO)2 2H2O: C 12.55,
H 3.15. Found: C 12.62, H 3.06.
3
Results and Discussion
Synthesis. The series of compounds were synthesized
by neutralization. The bubbling observed during the
synthesis of 1 using [ZnCO3]2 [Zn(OH)2]3 was expected,
3
as CO2 and H2O are byproducts of the neutralization with
heating. Bubbling stopped shortly after the stoichio-
metric amount of HCOOH was added. Similar initial
bubbling was observed in the case of the mixed cation
reactions (3 and 4), which also used [ZnCO3]2 [Zn-
3
(OH)2]3, suggesting the zinc formate establishes itself in
solution much earlier than the indium formate. Its pre-
sence appeared to accelerate the indium formate forma-
tion, as evident when comparing the consistency of the
indium formate synthesis to the mixed cation synthesis.
While reaction mixture 2 showed no visible difference
(a yellow suspension of In2O3) after 5 h, reaction mixtures
for 3 and 4 had significantly thickened. Reaction 4 was
carried out with an equal volume of HCOOH and water,
which further altered the process because of solubility
differences, as noted by the presence of crystals (4b)
growing out of the precipitate (4a) upon cooling. All
products were washed with THF instead of water to
remove residual HCOOH without washing away product.
1 was especially soluble in water.
IR. Figure 1 shows the FTIR absorbance spectra of the
compounds. Peak listings are provided in the Experimen-
tal Section. There were two principle diagnostic regions
in the IR spectra: the O-H stretch region (3600-3000
cm-1) and the formate bands. The former provided
information about the hydrate structure and OH-substi-
tution. Samples were dried in-vacuo for 36 h to minimize
interference from adsorbed water. The latter was distin-
guished by characteristic formate vibrational modes:
C-H stretching (∼2900 cm-1), asymmetric O-C-O
stretching (∼1550 cm-1), symmetric O-C-O stretching
(∼1370 cm-1), and symmetric O-C-O bending (∼790
cm-1).37,38
Cation Composition. As the conductivity and phases of
the In-Zn-oxide system are dependent on the cation
ratio, knowing the ratio in the precursors was critical.6,7
The compounds were synthesized with a target of 65-80
cation% In. Table 1 compares the cation ratio of the
reagents to the measured ratio of the products. Reported
uncertainties are the standard deviations of five replicates
for ICP-AES and six spots per sample for XRF, with 3a
(1) and 3a (2) denoting two separate synthesis attempts.
The XRF results are in agreement with ICP-AES given
their uncertainties.
The cation composition results reveal several impor-
tant factors. First, the indium fraction of the products was
higher than the mixed fraction in the reagents. The
preferential solubility of zinc formate resulted in the zinc
being retained in solution and not incorporated into the
product precipitate, hence the product being zinc defi-
cient. Therefore, it was hard to guarantee a final compo-
sition by simply mixing a desired starting stoichiometry.
The effect was enhanced in the case of reaction 4. Given
the high solubility of 1 in water, it was not surprising
that the precipitate (4a) was more In-rich with 86 cation%
The spectrum of 1 was typical of metal formate
dihydrates.29,39 Conversely, 2 showed no O-H bands,
which is consistent with the compound being anhydrous.
The mixed cation formate compounds exhibited various
degrees of hydration and OH-substitution. 3a showed
(37) Heyns, A. M. J. Mol. Struct. 1985, 127, 9.
(38) Moura, M. R.; Paschoal, C. W. A.; Ayala, A. P.; Guedes, I.; Leyva,
A. G.; Polla, G.; Vega, D.; Perazzo, P. K. J. Raman Spectrosc. 2002, 33, 273.
(39) Heyns, A. M. J. Mol. Struct. 1985, 127, 217.

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5427
Figure 1. FTIR absorbance spectra of 1, 2, 3a, 3b, 4a, and 4b. Vertical
lines provided as visual aid.
weak stretching peaks for both hydration and OH-sub-
stitution for formate anions in the structure, the latter
being indicated by a distinct peak at higher frequency
(∼3415 cm-1) and bending at ∼1040 cm-1. Su et al. noted
similar peaks for In2(HCOO)5(OH) and In(HCOO)2-
(OH).36 Interestingly, 3b was anhydrous and did not show
any O-H peaks for water nor OH-substitution even
though it was synthesized similarly to 3a. The cause for
the differences in hydration between the products is now
under investigation but was possibly due to a lack of
experimental control on water content in the reaction
mixture, either as a neutralization product or from the
solvent. While a small fraction of 1 was detected in 4a, the
remaining vibrational pattern most closely matches that
of In(HCOO)2(OH).36 The spectrum of 4b matched 1.
Structure. XRD powder patterns of the compounds
are provided in Figure 2 and further aided in the identi-
fication of the mixtures. Powder patterns of In(HCOO)3
and In(HCOO)2(OH) are included as reference. They
were simulated using CrystalDiffract 5.1 from the crystal-
lographic information files (CIF) provided in the Sup-
porting Information of Su et al.36 The XRD pattern of
2 was in agreement with hexagonal In(HCOO)3. 4a
strongly matched with the peak positions of tetragonal
In(HCOO)2(OH). Interestingly, the pattern for 3a was
consistent with 2 and not 4a. This was unexpected given
that IR results showed 3a to be OH-substituted, and
elemental analysis was consistent with a ∼1:1 In:OH
ratio. Thus, the powder diffraction results suggest that
the OH-substituted In formate in 3a adopted the hexa-
gonal In(HCOO)3 rather than tetragonal In(HCOO)2-
(OH) crystal structure. The exact nature of OH incor-
poration could not be determined from the given data,
Figure 2. XRD powder patterns of 1, 2, 3a, 3b, 4a, and ground crystals
of 4b. Simulated patterns of In(HCOO)3 and In(HCOO)2(OH) are
included as reference.
and a suitable sample for single crystal diffraction was not
obtained.
It should be noted that it was difficult to detect peaks
from Zn(HCOO)2 2H2O in the mixed formates because
of the small fraction of zinc material present. The only
distinguishable peak for Zn(HCOO)2 2H2O occurred at
3
3
34.6° 2θ in 3a and 33.8° 2θ in 4a. XRD results of the In-
Zn formate compounds 3a and 4a suggest they were not
solid solutions but instead mixtures of Zn(HCOO)2
3
2H2O and OH-substituted In formate. The XRD pattern
of the ground crystals of 4b was consistent with Zn-
(HCOO)2 2H2O (1), with a slight variation in relative
peak intensity because of texturing from the fine crystal-
line nature of the powder.
3
The XRD pattern for 3b was very similar to 2, which is
consistent with the IR results for a mixture containing
mostly In(HCOO)3. The XRD data was inconclusive
about the nature of Zn in 3b, although the IR results
clearly indicate that it was not in its hydrated form. The
thermal decomposition data (presented in the following
section) suggested a mixture of anhydrous Zn(HCOO)2
and In(HCOO)3. A list of compounds identified in the
products is summarized in Table 2.
Thermal Analysis. The thermal analysis curves of the
compounds under N2 and air atmospheres are shown in
Figures 3 and 4, respectively. The curves progress in order
of increasing indium content from zinc formate dihydrate
(1) to indium-zinc formate (3a, 3b, 4a) to indium formate

5428 Inorganic Chemistry, Vol. 49, No. 12, 2010
Table 2. List of Compounds Identified in Products
compounds identified
Pasquarelli et al.
1
2
3a
3b
4a
4b
Zn(HCOO)2 2H2O
In(HCOO)3 (hex)a
3
Zn(HCOO)2 2H2O þ In(HCOO)2(OH) (hex)a
3
Zn(HCOO)2 þ In(HCOO)3 (hex)a
Zn(HCOO)2 2H2O þ In(HCOO)2(OH) (tet)b
3
3
Zn(HCOO)2 2H2O
a Hexagonal crystal structure (hex). b Tetragonal crystal structure (tet).
Figure 4. TG (solid) and DSC (dashed) curves of 1 (a), mixed powders
of 1 and 2 with 50 cation% In (b) and 70 cation% In (c), 3a (d), 3b (e), 4a
(f), and 2 (g) performed under air. Curves are presented in order of
increasing In content. TG-derivative (dotted) curves and a line at 270 °C
are provided as a guide.
broad O-H stretching associated with a hydrated struc-
ture in all but 2 and 3b. Similar thermal behavior was
observed for dehydration with a weight loss occurring at
105-110 °C corresponding to an endothermic process.
The measured weight losses were in agreement with the
proposed mixture compositions.
The second decomposition stage of the formate com-
pounds was more complex and showed distinct differences.
First, there was a significant difference of ∼20 °C between
the decomposition of the individual zinc and indium for-
mates. 1decomposed between 270 and 310 °C (peak 290 °C),
but 2 decomposed at a lower temperature in air (peak
270 °C). Second, the thermal decomposition seemed to take
place in two steps for the mixed formates, as indicated by the
splitting of DSC peaks and a slope change in the TG curves.
It is important to note that the decomposition behavior was
not a superposition of the individual compounds. Instead,
a shift to lower temperature by ∼5 °C was observed. This
was reproducible and not an artifact of the measurement.
Figure 3. TG (solid) and DSC (dashed) curves of 1 (a), mixed powders
of 1 and 2 with 50 cation% In (b) and 70 cation% In (c), 3a (d), 3b (e), 4a
(f) and 2 (g) performed under N2. Curves are presented in order of
increasing In content. TG-derivative (dotted) curves and a line at 270 °C
are provided as a guide.
(2). Ground mixtures of 1 and 2 at 50 and 70 cation% In
were included in the set to compare with the co-synthesized
In-Zn formate. Table 3 summarizes the measured and
calculated weight losses and decomposition temperatures.
Decomposition occurs by two principle transitions:
dehydration, followed by decomposition of the anhy-
drous formate. Dehydration was observed for 1, 3a, 4a,
and the mixed individual formates, but not for 2 or 3b.
Thisis consistent with the IRresults (Figure1) thatshowed

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5429
Table 3. TG-DSC Critical Temperatures and Weight Losses
1
Mix 50 cation% In
8.18 {8.16}
Mix 70 cation% In
4.57 {4.65}
3a
3b
4a
2
dehydration
loss (wt %)a
decomposition
temperature
(°C)b
18.80 {18.82}
5.07 {5.08}
4.58 {2.83}
270-310
[290]
240-300
[258] [285]
240-300
[261] [285]
230-290
[258] [274]
240-300
[265] [285]
230-290
[254] [268]
260-300
[288] (N2)
[270] (air)
51.05 (N2)
42.68 (air)
<44.44>
decomposition
loss (wt %)c
55.48 (N2)
56.68 (air)
<57.48>
53.60 (N2)
52.20 (air)
<50.96>
52.88 (N2)
50.41 (air)
<48.36>
49.10 (N2)
42.43 (air)
<42.84>
53.06 (N2)
50.11 (air)
<45.02>
44.63 (N2)
42.08 (air)
<40.44>
a Calculated value for dehydration weight loss { }. b Decomposition peak temperatures from TG derivative curve under N2 [ ]. c Calculated value
assuming ZnO and/or In2O3 products < >.
The progression of the two decomposition steps with tem-
perature and composition is most evident upon examination
of the TG derivative curves (Figure 3). The mixtures of 1and
2 both showed weight losses between 240 and 300 °C with
peak splitting at ∼258 and 285 °C. The decomposition step
temperatures were independent of the cation% In.
process exhibited an endothermic to exothermic transi-
tion under air.
The energetics of 3a and 4a are the most revealing. The
decomposition was endothermic under N2 but appeared
solely exothermic under air. Increasing the heating rate
from 10 °C/min to 20 °C/min under air resulted in a DSC
curve change, displaying an endotherm transitioning to
an exotherm similar to that of the mixed powders. The
change could be due to the reaction kinetics. With a faster
heating rate, the compound may reach the required
temperature and decompose (endothermic reduction)
faster than oxygen can diffuse in for reaction (exother-
mic oxidation). The previously noted lower decomposi-
tion temperature and above-mentioned different DSC
behavior of 3a and 4a compared to the mixed powders
could be the result of co-synthesis achieving finer mixing
than was achieved by physically mixing the individual
formates with grinding. In principle, finer mixing would
allow thelocalenergeticsof indium formatedecomposition
to more effectively influence the zinc formate. However, it
is more probable that the effect was due to the presence of
OH-substituents, which was unique to these two samples.
This is supported by the fact that 4a, with a larger fraction
of In(HCOO)2(OH), exhibited an even greater decomposi-
tion temperature shift than3a. Furthermore, 3b, which was
anhydrous and not OH-substituted, showed a less pro-
nounced shift and was very similar to the mixed individual
formate powders in terms of TG and DSC behavior.
Finally, thermal analysis of the compounds revealed
that reduction and oxide products form. The measured
losses under N2 were 53.60 wt % (Calcd 50.96 wt %) and
52.88 wt % (Calcd 48.36 wt %) for the 50 and 70 cation%
In mixtures respectively. The measured losses were signi-
ficantly larger than was expected for the decomposition
into stoichiometric ZnO and In2O3 products. This effect
was more significant under the reducing N2 environment,
especially in the case of 3a with losses of 49.43 wt % (N2)
versus 42.43 wt % (air). The XRD powder patterns
indicate that the decomposition products contain In
metal, In2O3, and ZnO (Figure 5). A full list of XRD
patterns for the TG decomposition products after heating
to 600 °C under N2 and air are provided in Figures S1 and
S2 of the Supporting Information. Likewise, images of
decompositions products are provided in Figure S3 of the
Supporting Information. As shown in Figure 5, 3a yielded
a significant amount of In metal following decomposition
after 300 °C and that it was converted to mostly oxide
with longer heating up to 600 °C. Samples that favored
oxidation (3a and 4a) showed a more pronounced yellow
color, indicative of In2O3, while samples that favored
While 3a also exhibited peak splitting, it decomposed at
an even lower temperature (230-290 °C) with peak split-
ting at 258 and 274 °C. Interestingly, the second peak for
the co-synthesized formate occurred at a significantly
lower temperature (∼10 °C) than the mixed individual
formateswiththe same cation composition. Conversely, 3b
showed splitting similar to the mixed individual formates.
This was not surprising given that the mixed individual
formates were anhydrous after 110 °C. Both 3b and the
mixture were effectively Zn(HCOO)2 and In(HCOO)3
before the last decomposition stage. 4a decomposed at
slightly lower temperature than 3a, peaking at 254 and
268°C. Thermal analysisresultsfor4ashowed inconsistent
weight losses to the degree that an accurate formula
determination could not be carried out, but elemental
analysis, IR, and XRD all support it being a mixture of
Zn(HCOO)2 2H2O and In(HCOO)2(OH). In general, it
3
was observed that a mixture of In and Zn formates reduced
the decomposition temperature and that OH-substitution
reduced the temperature of the second step.
We concluded that the shift to lower decomposition
temperature was due to a synergistic effect between the In
and Zn species. The results suggest the initial decomposi-
tion of the indium formate was assisting the decomposi-
tion of its zinc counterpart. The initial decomposition
step may provide sufficient local energy to heat the zinc
formate higher than the applied temperature during the
analysis, thus lowering the required temperature of the
second step. Such energetics are apparent when contrast-
ing the DSC results measured under N2 (Figure 3) and air
(Figure 4). In the case of 2, the decomposition process
was endothermic under N2 but exothermic under the
air environment. This difference was attributed to atmo-
sphere dependent reactions, with the more oxidiz-
ing atmosphere favoring formation of In2O3 over In
metal (discussed further later). In fact, the gradual and
then sudden steepness in the TG slope (almost vertical)
observed under air is indicative of a highly exothermic
process in which the sample heats faster than the
applied heating rate. As the zinc compound exhibited
only oxide formation, no change in DSC behavior was
observed under different atmospheres. In the case of
the mixed individual formates, the two decomposition
steps were both endothermic under N2. However, the

5430 Inorganic Chemistry, Vol. 49, No. 12, 2010
Pasquarelli et al.
with indium-containing compounds even in an oxidizing
environment was confirmed and exploited by drop-cast-
ing from solutions with different acid additives. Aqueous
solutions of 3a with HCOOH and HNO3 were drop-cast
onto microscope glass at 300 °C in air. Figure 5 shows
typical XRD profiles of the resulting material. The addi-
tion of excess formic acid drove the reaction toward the
formation of indium metal, as indicated by the relatively
large In (101) peak at 32.9° 2θ. Conversely, the addition
of HNO3 (an oxidizing acid) yielded the oxide. ZnO was
observed in both cases. The results suggest that In-Zn
formate can be used as either an indium metal or an oxide
precursor simply by taking advantage of the additive’s
chemistry.
Likewise, indium reduction can occur by a red-ox
reaction with zinc. The reduction of In(III) to In metal,
with its more positive reduction potential, is electroche-
mically favored in the presence of Zn metal:
In3þ þ 3e- f In0
Zn2þ þ 2e- f Zn0
E° - 0:34 V vs SHE
E° - 0:76 V vs SHE
Zhang et al. employed the process to synthesize metallic
indium hollow spheres and nanotubes by reacting InCl3
with Zn metal powder.40 In the case of the In-Zn
formates, if Zn metal were to form during decomposition,
it would react with any In(III) present and produce In
metal and ZnO, which may account for 3b yielding a
larger fraction of In metal than 2. This is supported by the
presence of In metal, In2O3, ZnO, but not Zn metal in the
post-TG decomposition products and drop-cast material.
Sprayed IZO Films. To investigate the potential of
In-Zn formates as precursors for a-IZO, films were
deposited by ultrasonic spray onto microscope glass
slides at 210 °C from solutions of 3a in methanol and
water with HNO3. The method employed an ultrasonic
nozzle and carrier gas to generate and direct an aerosol of
precursor solution onto the substrate. It differed from
other chemical vapor deposition and spray pyrolysis
methods since neither vaporization of the precursor nor
full decomposition of the precursor upon contact with the
substrate occurred. Substrate surface temperatures were
below the decomposition temperature, and annealing was
required for oxide formation.
Thin films (100-200 nm) were produced with good optical
transmittance (>80%) and conductivities of ∼50 S/cm.
The films were amorphous after thermal processing at 300-
400 °C, as shown by the XRD profile in Figure 5e of
a typical film after processing. The broad “hump” in the
spectrum (28-39° 2θ) was indicative of amorphous mate-
rial and strongly resembled the glass substrate. The ability
to produce amorphous IZO by solution processing, as
demonstrated here, is especially important for material
applications. When sputtered, IZO is the most conductive
in the amorphous state.6,7 Additionally, its amorphous
structure may make it more suitable for flexible substrates
compared to other crystalline TCOs. To the best of our
knowledge, this is the first reported instance of solution
deposited a-IZO films. A full analysis of the structural,
optical, and electronic properties of these films is the
subject of a future paper.
Figure 5. XRD powder patterns of resulting material from drop-cast
3a-HCOOH solution (a), 3a powder heated to 300 °C (b), 3a powder
heated to 600 °C (c), drop-cast 3a-HNO3 solution (d), and sprayed 3a-
HNO3 film after 300 °C anneal (e). Indexed patterns of In metal, In2O3,
Zn metal, and ZnO are included as reference.
reduction to In metal were grayer (2 and 3b). 3b strongly
favored reduction, producing almost entirely gray mate-
rial even when the analysis was performed under air.
The measured losses of 53.06 wt % (N2) and 50.11 wt
% (air) were significantly larger than was expected for
the decomposition into stoichiometric oxide products
(45.02 wt %).
As decomposition of formate may produce the redu-
cing gas CO, it is not surprising that In(III) is reduced in
situ to In metal. Similarly, Su et al. noted reduction
products in the decomposition of In(HCOO)3, In2-
(HCOO)5(OH), and In(HCOO)2(OH) and that the ratio
of indium metal to oxide decreased as the degree of OH-
substitution increased. Furthermore, they observed CO,
CO2, and H2O during decomposition using mass spectro-
metry.36 CO can exhaust local oxygen content during
decomposition, thus allowing for reduction products
even in air, as was observed in the case of 3b.
Precursor Chemistry. The reduction behavior of the
formates is unique compared to the more commonly used
acetate precursors. During TCO processing, films are
often annealed in a reducing environment (such as form-
ing gas) to create a non-stoichiometric oxide and thus
increase electronic carrier concentration. Similarly, the
presence of an inherently reducing ligand may have the
advantage of chemically aiding in the introduction of
carriers in lieu of or complementary to a reducing anneal.
In terms of solution processing, additives not only
influence solution properties (such as wetting) but also
influence the chemistry. The reducing properties of formate
(40) Zhang, Y.; Li, G.; Zhang, L. Inorg. Chem. Commun. 2004, 7(3), 344.

Article
To investigate the effect of the nitric acid additive,
Inorganic Chemistry, Vol. 49, No. 12, 2010 5431
It was found that the presence of both In and Zn species
had a synergistic effect that resulted in an overall reduction in
the decomposition temperature. OH-substitution further
reducedthetemperature. As a precursor, itwas demonstrated
that the decomposition products of In-Zn formate could
be directed toward oxidation or reduction by controlling the
decomposition atmosphere or with solution additives such as
acids. While the focus of the paper was characterization of
the precursors, it was also demonstrated that amorphous
IZO films could be prepared by solution processing from
In-Zn formates for TCO applications.
comparative experiments were performed using solutions
from only nitrate salts and from nitrate salts with formic
acid as an additive. Films sprayed by these routes and
processed under similar conditions exhibited extensive
pinholing and were half as conductive. While this work
demonstrates the potential for the formate system, addi-
tives other than HNO3 that do not contribute a significant
fraction of counterions are being explored so that a
formate-only system can be studied.
Conclusions
In summary, a series of In-Zn formate mixtures exhibiting
Acknowledgment. This work was supported by the
National Center for Photovoltaics(NCPV) atthe National
Renewable Energy Laboratory (NREL) through U.S.
Department of Energy under Contract No. DE-AC36-
08GO28308. Thanks to Lynn Gedvilas and Phil Parilla at
NREL for assistance with the IR and SDT measurements.
various degrees of hydration and OH-substitution were
synthesizedbyneutralization. IndividualZn(HCOO)2 2H2O
3
(1) and hexagonal In(HCOO)3 (2) were prepared. Mix-
tures were identified as Zn(HCOO)2 2H2O and hexagonal
3
In(HCOO)2(OH) with 70 cation% In (3a), anhydrous
Zn(HCOO)2 and hexagonal In(HCOO)3 with 74 cation%
In (3b), and Zn(HCOO)2 2H2O and tetragonal In(HCOO)2-
3
Supporting Information Available: Additional IR analysis.
Characterization of 4b. XRD powder patterns of decomposition
products under air and N2, and images of the decomposition
products. Preparation and characterization of OH-substituted
In(HCOO)3-x(OH)x from 2. This material is available free of
charge via the Internet at http://pubs.acs.org.
(OH) with 83 cation% In (4a). XRD powder diffraction
results suggest that the OH-substituted 3a adopted the
hexagonal In(HCOO)3 rather than tetragonal In(HCOO)2-
(OH) crystal structure. No evidence was found for solid
solution formation in the In-Zn formate system.