Synthesis and characterization of volatile, fluorine-free β-ketoiminate lanthanide MOCVD precursors and their implementation in low-temperature growth of epitaxial CeO2 buffer layers for superconducting electronics

Article abstract of DOI:10.1021/ic020299h

A new class of volatile, low-melting, fluorine-free lanthanide metal-organic chemical vapor deposition (MOCVD) precursors has been developed. The neutral, monomeric Ce, Nd, Gd, and Er complexes are coordinatively saturated by a versatile, multidentate ether-functionalized β-ketoiminato ligand series, the melting point and volatility characteristics of which can be tuned by altering the alkyl substituents on the keto, imino, and ether sites of the ligand. Direct comparison with conventional lanthanide β-diketonate complexes reveals that the present precursor class is a superior choice for lanthanide oxide MOCVD. Epitaxial CeO₂ buffer layer films can be grown on (001) YSZ substrates by MOCVD at significantly lower temperatures (450-650 °C) than previously possible by using one of the newly developed cerium β-ketoiminate precursors. Films deposited at 540 °C have good out-of-plane (Δω = 0.85°) and in-plane (Δφ = 1.65°) alignment and smooth surfaces (rms roughness ～ 4.3 A). The film growth rate decreases and the films tend to be smoother as the deposition temperature is increased. High-quality yttrium barium copper oxide (YBCO) films grown on these CeO₂ buffer layers by pulsed organometallic molecular beam epitaxy exhibit very good electrical transport properties (T_c = 86.5 K, J_c = 1.08 x 10⁶ A/cm² at 77.4 K).

Full text of DOI:10.1021/ic020299h

Inorg. Chem. 2002, 41, 5005−5023
Synthesis and Characterization of Volatile, Fluorine-Free â-Ketoiminate
Lanthanide MOCVD Precursors and Their Implementation in
Low-Temperature Growth of Epitaxial CeO Buffer Layers for
2
Superconducting Electronics
†
†
†
†
†
Nikki L. Edleman, Anchuan Wang, John A. Belot, Andrew W. Metz, Jason R. Babcock,
†
†
†
†
†
Amber M. Kawaoka, Jun Ni, Matthew V. Metz, Christine J. Flaschenriem, Charlotte L. Stern,
Louise M. Liable-Sands, Arnold L. Rheingold, Paul R. Markworth, Robert P. H. Chang,
Michael P. Chudzik, Carl R. Kannewurf, and Tobin J. Marks*
‡
‡
§
§
|
|
,†
Department of Chemistry and the Materials Research Center, Northwestern UniVersity,
EVanston, Illinois 60208, Department of Chemistry and Biochemistry, UniVersity of Delaware,
Newark, Delaware 19716, Department of Materials Science and the Materials Research Center,
Northwestern UniVersity, EVanston, Illinois 60208, and Department of Electrical and
Computer Engineering and the Materials Research Center, Northwestern UniVersity,
EVanston, Illinois 60208
Received April 29, 2002
A new class of volatile, low-melting, fluorine-free lanthanide metal−organic chemical vapor deposition (MOCVD)
precursors has been developed. The neutral, monomeric Ce, Nd, Gd, and Er complexes are coordinatively saturated
by a versatile, multidentate ether-functionalized â-ketoiminato ligand series, the melting point and volatility
characteristics of which can be tuned by altering the alkyl substituents on the keto, imino, and ether sites of the
ligand. Direct comparison with conventional lanthanide â-diketonate complexes reveals that the present precursor
class is a superior choice for lanthanide oxide MOCVD. Epitaxial CeO buffer layer films can be grown on (001)
2
YSZ substrates by MOCVD at significantly lower temperatures (450−650 °C) than previously possible by using
one of the newly developed cerium â-ketoiminate precursors. Films deposited at 540 °C have good out-of-plane
(
∆ω ) 0.85°) and in-plane (∆φ ) 1.65°) alignment and smooth surfaces (rms roughness ∼ 4.3 Å). The film
growth rate decreases and the films tend to be smoother as the deposition temperature is increased. High-quality
yttrium barium copper oxide (YBCO) films grown on these CeO buffer layers by pulsed organometallic molecular
beam epitaxy exhibit very good electrical transport properties (T
2
6
2
c
) 86.5 K, J
c
) 1.08 × 10 A/cm at 77.4 K).
-12
Introduction
multilayer device buffer layers (e.g., CeO
tric constant (high-k) materials (e.g., Gd
2
),5 high dielec-
1
3,14
O
2 3
),
high-
The growth and properties of lanthanide-containing oxide
films is of great current interest to the chemistry, materials
science, and electronics communities.1-4 Applications include
(
4) Lewkebandara, T. S.; Winter, C. H. Chemtracts: Inorg. Chem. 1994,
271-282.
5) Jimenez, C.; Weiss, F.; Senateur, J. P.; Abrutis, A.; Krellmann, M.;
Selbmann, D.; Eickemeyer, J.; Stadel, O.; Wahl, G. IEEE Trans. Appl.
Supercond. 2001, 11, 2905-2908.
(6) Selvamanickam, V.; Carota, G.; Funk, M.; Vo, N.; Haldar, P.;
Balachandran, Y.; Chudzik, M.; Arendt, P.; Groves, J. R.; DePaula,
R.; Newnam, B. IEEE Trans. Appl. Supercond. 2001, 11, 3379-3381.
(7) Knierim, A.; Auer, R.; Geerk, J.; Linker, G.; Meyer, O.; Reiner, H.;
Schneider, R. Appl. Phys. Lett. 1997, 70, 661-663.
(8) Norton, D. P.; Goyal, A.; Budai, J. D.; Christen, D. K.; Kroeger, D.
M.; Specht, E. D.; He, Q.; Saffian, B.; Paranthaman, M.; Klabunde,
C. E.; Lee, D. F.; Sales, B. C.; List, F. A. Science 1996, 274, 755-
757.
(
*
To whom correspondence should be addressed. E-mail: t-marks@
northwestern.edu.
†
Department of Chemistry and the Materials Research Center, North-
western University.
‡
Department of Chemistry and Biochemistry, University of Delaware.
Department of Materials Science and the Materials Research Center.
Department of Electrical and Computer Engineering and the Materials
§
|
Research Center.
(
1) Bunzli, J.-C. G.; Andre, N.; Elhabiri, M.; Muller, G.; Piquet, C. J.
Alloys Compd. 2000, 303-304, 66-74.
(2) Tiitta, M.; Niinisto, L. Chem. Vap. Deposition 1997, 3, 167-182.
3) Gun’ko, Y. K.; Edelmann, F. T. Comments Inorg. Chem. 1997, 19,
(
(9) Philips, J. M. J. Appl. Phys. 1996, 79, 1829-1848.
(10) Suzuki, M.; Ami, T. Mater. Sci. Eng. 1996, B41, 166-173.
153-184.
1
0.1021/ic020299h CCC: $22.00 © 2002 American Chemical Society
Inorganic Chemistry, Vol. 41, No. 20, 2002 5005
Published on Web 09/10/2002

Edleman et al.
7-δ),1
5-17
phos-
pentadienyls ) or fluorinated ligand substituents
35
22,36
often
temperature superconductors (e.g., LnBa
phor dopants,
materials.23
2
Cu
3
O
18-21
22
solid-state oxide fuel cells, and magnetic
leads to impurity incorporation which compromises desired
film properties.
37,38
MOCVD (metal-organic chemical vapor deposition)
offers many attractions for oxide film growth, including low
equipment costs, straightforward scale-up, conformal deposi-
tion on a variety of complex substrates, low growth tem-
peratures, and rapid growth rates. However, the success of
an MOCVD process depends critically on the availability
of volatile, thermally stable precursors which exhibit constant
vapor pressure and the capacity to selectively form the
desired phase at the substrate surface. Low-melting precur-
sors are preferred because solid compounds can be handled
with ease at room temperature, while the liquid form at
reservoir operating temperatures affords constant surface area
for stable vapor delivery to the reactor.24-32
No known lanthanide precursor satisfies all of the above
criteria, although the extensive body of known lanthanide
3
9-42
complexes
offers many potential candidates to the
synthetic chemist. A common structural motif employs
nonfluorinated â-diketonate ligands (I). Metal-organic
The design and realization of new molecular MOCVD
precursors offers a considerable synthetic challenge. For
many applications, the metal ion of interest is of large ionic
radius and therefore has many coordination sites which
require saturation to prevent volatility-depressing oligomer-
ization. Large ion size is often further complicated by a low
ionic charge, demanding the use of multidentate anionic and/
or donating ligands in order to ensure a monomeric, neutrally
complexes such as I suffer falling volatility over film growth
times due to oligomerization, thermal decomposition, and
22,36,43-45
sintering.
In addition, these sources frequently
36,46,47
require undesirably high growth temperatures.
The
volatilization temperature of type I precursors must often
be continuously ramped in order to provide a constant
33,34
charged metal complex.
In addition, the use of metal-
48
precursor vapor pressure in the film growth reactor.
carbon bonded ligand frameworks (e.g., substituted cyclo-
â-Diketonate precursor volatility and thermal stability have
been improved by the use of fluorinated ligand substituents
(
11) Van Wijck, M. A. A. M.; Verhoeven, M. A. J.; Reuvekamp, E. M. C.
(
II); however, formation of fluoride phases in the resulting
M.; Gerritsma, G. J.; Blank, D. H. A.; Rogalla, H. Appl. Phys. Lett.
2
2
-
1996, 68, 553-555.
films is problematic, and F species can corrode certain
(
(
12) Schulz, D. L.; Marks, T. J. AdV. Mater. 1994, 6, 719-730.
13) Kwo, J.; Hong, M.; Kortan, A. R.; Queeney, K. L.; Chabal, Y. J.;
Opila, R. L., Jr.; Muller, D. A.; Chu, S. N. G.; Sapjeta, B. J.; Lay, T.
S.; Mannaerts, J. P.; Boone, T.; Krautter, H. W.; Krajewski, J. J.;
Sergent, A. M.; Rosamilia, J. M. J. Appl. Phys. 2001, 89, 3920-3927.
14) Hong, M.; Kwo, J.; Kortan, A. R.; Mannaerts, J. P.; Sergent, A. M.
Science 1999, 283, 1897-1900.
i.e., metal) substrates and reactor components.49
(
A major research effort has focused on the synthesis of
alternative precursors for lanthanide oxide MOCVD with the
goal of overcoming the difficulties associated with common
sources as described above. Lanthanide tris[bis(trimethylsi-
lyl)amides] (III) have been used (primarily for ALE),5
(
(15) Rosner, C. H. IEEE Trans. Appl. Supercond. 2001, 11, 39-48.
(16) Norton, D. P. Annu. ReV. Mater. Sci. 1998, 28, 299-343.
(17) Macmanus-Driscoll, J. L. AdV. Mater. 1997, 9, 457-473.
(18) Kushida, T. Springer Series Solid-State Sci. 2000, 128, 1-35.
(19) Teren, A. R.; Wessels, B. W. Mater. Res. Soc. Symp. Proc. 2000,
0-52
(34) Babcock, J. R.; Benson, D. D.; Wang, A.; Edleman, N. L.; Belot, J.
A.; Metz, M. V.; Marks, T. J. Chem. Vap. Deposition 2000, 6, 180-
183.
597, 15-20.
(
(
(
(
(
20) Rack, P. D.; Naman, A.; Holloway, P. H.; Sun, S. S.; Tuenge, R. T.
(35) Nukeaw, J.; Tanagisawa, J.; Matsubara, N.; Fujiwara, Y.; Takeda, Y.
Appl. Phys. Lett. 1997, 70, 84-86.
MRS Bull. 1996, 21, 49-58.
21) Ronda, C. R.; Justel, T.; Nikol, H. J. Alloys Compd. 1998, 277, 669-
(36) Becht, M.; Gerfin, T.; Dahmen, K. H. Chem. Mater. 1993, 5, 137-
676.
144.
22) McAleese, J.; Plakatouras, J. C.; Steele, B. C. H. Thin Solid Films
(37) Greenwald, A. C.; Rees, W. S., Jr.; Lay, U. W. Mater. Res. Soc. Symp.
Proc. 1993, 301, 21-26.
(38) Schmaderer, F.; Huber, R.; Oetzmann, H.; Wahl, G. Appl. Surf. Sci.
1990, 46, 53-60.
(39) Anwander, R.; Hermann, W. A. Top. Curr. Chem. 1996, 179, 2-32.
(40) Anwander, R. Top. Curr. Chem. 1996, 179, 33-112.
(41) Anwander, R. Top. Curr. Chem. 1996, 179, 149-244.
(42) Joshi, K. C.; Pathak, V. N. Coord. Chem. ReV. 1977, 22, 37-122.
(43) Graboy, I. E.; Markov, N. V.; Maleev, V. V.; Kaul, A. R.; Polyakov,
S. N.; V. L. Svetchnikov, V. L.; Zandbergen, H. W.; Dahmen, K. H.
J. Alloys Compd. 1997, 251, 318-321.
1996, 280, 152-159.
23) Collocott, S. J.; Dunlop, J. B.; Lovatt, H. C.; Ramsden, V. S. Mater.
Sci. Forum 1999, 315-317, 77-83.
24) Lindler, J.; Schumacher, M.; Dauelsberg, M.; Schienle, F.; Miedl, S.;
Burgess, D.; Merz, E.; Strauch, G.; Juergensen, H. AdV. Mater. Opt.
Electron. 2000, 10, 163-167.
(
25) Gleizes, A. N. Chem. Vap. Deposition 2000, 6, 155-173.
26) Hinds, B. J.; McNeely, R. J.; Studebaker, D. L.; Marks, T. J.; Hogan,
T. P.; Schindler, J. L.; Kannewurf, C. R.; Zhang, X. F.; Miller, D. J.
Mater. Res. 1997, 12, 1214-1236.
(
(27) Marks, T. J.; Belot, J. A.; Reedy, C. J.; McNeely, R. J.; Studebaker,
(44) Ikegawa, S.; Motoi, Y. Thin Solid Films 1996, 281-282, 60-63.
(45) Hiskes, R.; DiCarolis, S. A.; Jacowitz, R. D.; Lu, Z.; Fiegelson, R.
S.; Ronte, R. K.; Young, J. L. J. Cryst. Growth 1993, 128, 781-787.
(46) Becht, M.; Wang, F.; Wen, J. G.; Morishita, T. J. Cryst. Growth 1997,
170, 799-802.
D. L.; Neumayer, D. A.; Stern, C. L. J. Alloys Compd. 1997, 251,
243-252.
(28) Rees, W. S., Jr., Ed. CVD of Nonmetals; VCH Publishers: Weinheim,
Germany, 1996.
(29) Kodas, T.; Hampden-Smith, M. The Chemistry Of Metal CVD; VCH
Publishers: Weinheim, Germany, 1994.
(47) Lu, Z.; Hiskes, R.; DiCarolis, S. A.; Nel, A.; Ronte, R. K.; Fiegelson,
R. S. J. Cryst. Growth 1995, 156, 227-234.
(
30) Wessels, B. W. Annu. ReV. Mater. Sci. 1995, 25, 525-546.
(48) Liang, S.; Chern, C. S.; Shi, Z. Q.; Lu, P.; Lu, Y.; Kear, B. H. J.
Cryst. Growth 1995, 151, 359-364.
(31) Dahmen, K. H.; Gerfin, T. Prog. Cryst. Growth Charact. Mater. 1993
27, 117-161.
(49) Chadwick, D.; McAleese, J.; Senkiw, K.; Steele, B. C. H. Appl. Surf.
Sci. 1996, 99, 417-420.
(
32) Marks, T. J. Pure Appl. Chem. 1995, 67, 313-318.
(33) Belot, J. A.; Neumayer, D. A.; Reedy, C. J.; Studebaker, D. B.; Hinds,
B. J.; Stern, C. L.; Marks, T. J. Chem. Mater. 1997, 9, 1638-1648.
(50) Gordon, R. G.; Becker, J.; Hausmann, D.; Suh, S. Chem. Mater. 2001,
13, 2463-2464.
5006 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
(
VII),6
3-72
although additional processing steps are required
63,68,69
to remove and/or prevent fluoride phase formation,
and
although the working lifetime of these materials is limited
by the appreciable air-sensitivity. Winter and co-workers
have developed a series of complexes with pyrazolato ligand
-
the threat of metal corrosion by F species is again a
7
3
concern.
5
3
The goal of this research effort was to design a multiden-
tate, fluorine-free ligand system that would address the
general precursor considerations discussed above and could
successfully be applied to the entire lanthanide series, with
trivalent 8-coordinate ionic radii ranging from 1.16 to 0.977
systems (IV); however, these compounds have been used
only to dope GaAs semiconductor films and have not been
demonstrated in pure oxide deposition.5
4-56
Rees et al.
recently communicated the synthesis of a bidentate â-ke-
toiminato Yb complex (V).57 However, thermogravimetric
3
+
3+ 74
Å (La -Lu ). In addition, a precursor that is a liquid at
typical reactor operating temperatures would avoid the
sintering problem commonly encountered with the above-
mentioned solid I and II sources. Therefore, the synthesis
and characterization of low-melting, nonfluorinated lan-
analysis (TGA) data show that this complex volatilizes at
rather high temperatures (∼600 °C), and no film growth
using V has been reported.
Other groups have made significant synthetic advances in
developing classes of lanthanide precursors based on a
combined â-diketonate + glyme ligation approach. This
synthetic strategy was originally demonstrated for alkaline
3
+
3+
3+
3+
thanide (Ln ) Ce , Nd , Gd , and Er ) complexes
bearing homoleptic multidentate â-ketoiminato ligands (VIII)
3
3,58
earth compounds
and is successful because the addition
of a donating polyether chain leads to monomeric complexes,
in contrast to the parent â-diketonates. Several groups have
3
detailed the synthesis of nonfluorinated Ln(â-diketonate) -
glyme) compounds (VI),5
9-62
although structural data show
(
that, in some cases, the glyme ligand bridges two metal
centers, and decomposition of these complexes upon heating
has been documented,5
9,61
as well as the need for high
precursor bubbler temperatures and irreproducible precursor
delivery rates.61 Fluorination of the â-diketonato skeleton
improves the thermal stability of this precursor class
has been a central goal.7
5-78
This ligand system was
(
(
(
51) Just, O.; Rees, W. S., Jr. AdV. Mater. Opt. Electron. 2000, 10, 213-
21.
52) Rees, W. S., Jr.; Just, O.; Van Derveer, D. S. J. Mater. Chem. 1999,
2
originally explored by our laboratory for coordinative satura-
9, 249-252.
53) Pfeiffer, D.; Ximba, B. J.; Liable-Sands, L. M.; Rheingold, A. L.;
Heeg, M. J.; Coleman, D. M.; Schlegel, H. B.; Kuech, T. F.; Winter,
C. H. Inorg. Chem. 1999, 38, 4539-4548.
(63) McAleese, J.; Darr, J. A.; Steele, B. C. H. Chem. Vap. Deposition
1996, 2, 244-247.
(64) Baxter, I.; Darr, J. A.; Hursthouse, M. B.; Malik, K. M. A.; McAleese,
J.; Mingos, D. M. P. Polyhedron 1998, 17, 1329-1341.
(65) Malandrino, G.; Bettinelli, M.; Speghini, A.; Fragal a` , I. L. Eur. J.
Inorg. Chem. 2001, 1039-1044.
(66) (a) Malandrino, G.; Nigro, R. L.; Benelli, C.; Castelli, F.; Fragal a` , I.
L. Chem. Vap. Deposition 2000, 6, 233-238. (b) Lo Nigro, R.;
Malandrino, G.; Fragal a` , I. L. Chem. Mater. 2001, 13, 4402.
(67) Condorelli, G. G.; Gennaro, S.; Fragal a` , I. L. Chem. Vap. Deposition
2000, 6, 185-192.
(68) Malandrino, G.; Fragal a` , I. L. Chem. Mater. 1998, 10, 3765-3768.
(69) Malandrino, G.; Benelli, C.; Castelli, F.; Fragal a` , I. L. Chem. Mater.
1998, 10, 3434-3444.
(70) Malandrino, G.; Fragal a` , I. L.; Aime, S.; Dastru, W.; Gobetto, R.;
Benelli, C. J. Chem. Soc., Dalton Trans. 1998, 1509-1512.
(71) Malandrino, G.; Incontro, O.; Castelli, F.; Fragal a` , I. L. Chem. Mater.
1996, 8, 1292-1297.
(72) Malandrino, G.; Licata, R.; Castelli, F.; Fragal a` , I. L. Inorg. Chem.
1995, 34, 6233-6234.
(73) McAleese, J.; Steele, B. C. H. Corros. Sci. 1998, 40, 113-123.
(74) Shannon, R. D. Acta Crystallogr. 1976, A32, 751-767.
(
(
(
(
(
(
(
(
(
54) Cederberg, J. G.; Culp, T. D.; Bieg, B.; Pfeiffer, D.; Winter, C. H.;
Bray, K. L.; Kuech, T. F. J. Appl. Phys. 1999, 85, 1825-1831.
55) Cederberg, J. G.; Culp, T. D.; Bieg, B.; Pfeiffer, D.; Winter, C. H.;
Bray, K. L.; Kuech, T. F. J. Cryst. Growth 1998, 195, 105-111.
56) Culp, T. D.; Cederberg, J. G.; Bieg, B.; Kuech, T. F.; Bray, K. L.;
Pfeiffer, D.; Winter, C. H. J. Appl. Phys. 1998, 83, 4918-4927.
57) Rees, W. S., Jr.; Just, O.; Castro, S. L.; Matthews, J. S. Inorg. Chem.
2000, 39, 3736-3737.
58) Teren, A. R.; Belot, J. A.; Edleman, N. L.; Marks, T. J.; Wessels, B.
W. Chem. Vap. Deposition 2000, 6, 175-177.
59) Drake, S. R.; Lyons, A.; Otway, D. J.; Slawin, A. M. Z.; Williams,
D. J. J. Chem. Soc., Dalton Trans. 1993, 2379-2386.
60) Darr, J. A.; Mingos, D. M. P.; Hibbs, D. E.; Hursthouse, M. B.; Malik,
K. M. A. Polyhedron 1996, 15, 3225-3231.
61) McAleese, J.; Plakatouras, J. C.; Steele, B. C. H. Thin Solid Films
1996, 286, 64-71.
62) Baxter, I.; Drake, S. R.; Hursthouse, M. B.; Malik, K. M. A.;
McAleese, J.; Otway, D. J.; Plakatouras, J. C. Inorg. Chem. 1995, 34,
1384-1394.
Inorganic Chemistry, Vol. 41, No. 20, 2002 5007

Edleman et al.
tion of alkaline earth ions,79-82 the MOCVD applications of
which offer many of the same challenges as lanthanide ions.
A further aim of the present study was to elaborate and
low-temperature film deposition is requisite for multilayer
device fabrication, all CeO
2
MOCVD processes reported
prior to this contribution have required very high growth
3
6,46,47,93
investigate the effect of various alkyl substituents at the R
1
,
temperatures (680-800 °C).
to improve these CeO growth parameters by using special
reactors and/or other methods;
temperatures are still required.
2
Efficient, reproducible MOCVD CeO growth processes
Efforts have been made
R
2
, and R sites of ligand VIII on precursor melting point
3
2
48,94
and volatility characteristics. In this contribution, lanthanide
â-ketoiminate complexes of type VIII are named according
however, high deposition
to the substitution at the R
and the noncoordinated protonated ligands are simply
H(R ).”
CeO , with a cubic fluorite crystal structure (a ) 5.411
1
, R
2
, and R
3
1 2 3 3
sites, “Ln(R R R ) ,”
depend crucially on the availability of high-purity metal-
organic precursors having high and stable vapor pressures.
As discussed above, it has been found that the current
generation fluorine-bearing and fluorine-free â-diketonate
cerium complexes have inherently poor thermal stability.
“
1 2 3
R R
2
83
Å), has been widely investigated as a protective buffer layer
for high-temperature superconductor (HTS) films in current-
carrying applications5-9 and as an insulating layer in mul-
Some of them [e.g., Ce(dpm) , dpm ) 2,2,6,6-tetramethyl-
4
tilayer electronic devices.1
0-12
Compared to YSZ, CeO
Cu
YBCO) because of the smaller lattice mismatch and similar
thermal expansion coefficient. Thus far, many growth
2
is a
O
3 7-δ
3,5-heptanedione] decompose within 1 h at temperatures
necessary to maintain a useful vapor transport rate. Such high
precursor evaporation temperatures (above 200 °C) also
present a serious materials complication in MOCVD reactor
design(e.g., Teflon components cannot be used). Further-
more, as noted, all of the aforementioned Ce precursors
require growth temperatures in excess of 680 °C for high-
more attractive buffer layer for superconducting YBa
2
(
8
4,85
86,87
techniques including MBE,
pulsed laser deposition,
sputtering,8
8-90
and electron beam evaporation91,92 have been
utilized for CeO
molecule-based metal-organic chemical vapor deposition
MOCVD) techniques offer many attractive features such
2
film growth. However, as noted above,
2
quality CeO thin film growth, which is undesirable in
(
multilayer device fabrication. In this contribution, we report
the first in situ MOCVD deposition of ceria thin films at
temperatures as low as 450 °C using one of the newly
developed liquid cerium precursors of type VIII, and we
demonstrate its inherent advantages as a buffer layer for
YBCO growth by MOCVD.
as conformal coverage, ability to coat complex shapes,
simplified apparatus, and lower growth temperatures for thin
film growth and device fabrication.2
4-32
Nevertheless, MOCVD
2
deposition. Although
has not been fully optimized for CeO
(75) Edleman, N. L.; Belot, J. A.; Babcock, J. R.; Metz, A. W.; Metz, M.
V.; Stern, C. L.; Marks, T. J. Mater. Res. Soc. Symp. Proc. 2000,
6
23, 371-376.
Experimental Section
(
76) Wang, A.; Belot, J. A.; Marks, T. J.; Markworth, P. R.; Chang, R. P.
H.; Chudzik, M. P.; Kannewurf, C. R. Physica C 1999, 320, 154-
General Synthetic Procedures. All glassware was carefully
washed in distilled water (to minimize metal contaminants), rinsed
with acetone, and dried for at least 2.0 h in an oven prior to use.
Standard Schlenk techniques and either a Vacuum Atmospheres
or M. Braun nitrogen-filled glovebox were used in the isolation
and handling of all lanthanide reagents and complexes. Diethyl ether
and THF solvents were dried over NaK-benzophenone ketyl and
distilled immediately prior to use. Pentane and toluene solvents
were either dried and distilled from NaK-benzophenone ketyl or
passed through a dual-column (alumina and Q-5 copper catalyst)
Solvtek solvent purification system. Xylenes were dried over and
distilled from molten Na.
160.
(
77) Belot, J. A.; Wang, A.; McNeely, R. J.; Liable-Sands, L.; Rheingold,
A. L.; Marks, T. J. Chem. Vap. Deposition 1999, 5, 65-69.
78) Belot, J. A.; Wang, A.; Edleman, N. L.; Babcock, J. R.; Metz, M. V.;
Marks, T. J.; Markworth, P. R.; Chang, R. P. H. Mater. Res. Soc.
Symp. Proc. 1999, 574, 37-43.
79) Studebaker, D. B.; Neumayer, D. A.; Hinds, B. J.; Stern, C. L.; Marks,
T. J. Inorg. Chem. 2000, 39, 3148-3157.
80) Neumayer, D. A.; Belot, J. A.; Feezel, R. L.; Reedy, C. J.; Stern, C.
L.; Marks, T. J.; Liable-Sands, L. M.; Rheingold, A. L. Inorg. Chem.
(
(
(
1
998, 37, 5625-5633.
(
(
(
81) Schulz, D. L.; Hinds, B. J.; Neumayer, D. A.; Stern, C. L.; Marks, T.
J. Chem. Mater. 1993, 5, 1605-1617.
82) Schulz, D. L.; Hinds, B. J.; Stern, C. L.; Marks, T. J. Inorg. Chem.
1993, 32, 249-250.
Reagents. The reagents 2-methoxyethylamine, pinnacolone, ethyl
acetate, ethyl propionate, chlorotrimethylsilane, and NH Cl (99.998%)
4
were purchased from Aldrich, and 2,2,6,6-tetramethyl-3,5-hep-
tanedione was purchased from Lancaster Synthesis or Oakwood
83) Galasso, F. S. Structure and Properties of Inorganic Solids; Interna-
tional Series of Monographs in Solid State Physics; Pergamon: New
York, 1970; Vol. 7, p 93.
(
84) Inoue, T.; Yamamoto, Y.; Koyama, S.; Suzuki, S.; Ueda, Y. Appl.
Phys. Lett. 1990, 29, 1332-1333.
Chemicals. High-purity (99.999%) Ln
was obtained from Cerac. Gelest supplied KN(SiMe
solution in toluene). Ce(dpm) and Er(dpm) were purchased from
. All reagents were
2
O
3
(Ln ) Ce, Nd, Gd, Er)
(
85) Yoshimoto, M.; Shimozono, K.; Maeda, T.; Ohnishi, T.; Kumagai,
M.; Chikyow, T.; Ishiyama, O.; Shinohara, M.; Koinuma, H. Jpn. J.
Appl. Phys. 1995, 34, L688-L690.
3 2
)
(11%
4
3
(
86) Denhoff, M. W.; McCaffrey, J. P. J. Appl. Phys. 1991, 70, 3986-
Strem and were handled and stored under N
2
3988.
used as received.
Physical Measurements. H NMR spectra were recorded in
(
87) Wu, X. D.; Dye, R. C.; Muenchansen, R. E.; Foltyn, S. R.; Maley,
1
M.; Rollett, A. D.; Garcia, A. R.; Nogar, N. S. Appl. Phys. Lett. 1991,
58, 2165-2167.
3 6 6
CDCl or dry, deoxygenated C D on either a Varian Gemini 300
(
(
(
(
(
88) Yaegashi, S.; Kurihara, T.; Hoshi, H.; Segawa, H. Jpn. J. Appl. Phys.
MHz, Varian Mercury 400 MHz, or Varian Unity 500 MHz
spectrometer. Chemical shifts were referenced to solvent signals.
Elemental analyses were performed by Midwest Microlabs, Inc.
1994, 33, 270-274.
89) Guo, S.; Arwin, H.; Jacobson, S. N.; Jarrendahl, K.; Helmersson, U.
J. Appl. Phys. 1995, 77, 5369-5376.
90) Wang, F.; Kunkel, G.; Coppetti, C.; Kohlstedt, H.; Wordenweber, R.
Appl. Supercond. 1993, 2, 155-1158.
(Indianapolis, IN). Melting points were determined in sealed
91) Inoue, T.; Ohsuna, T.; Luo, L.; Wu, X. D.; Maggiore, C. J.; Yamamoto,
Y.; Koyama, S.; Chang, H. Appl. Phys. Lett. 1991, 59, 3604-3606.
92) Kotelyanskii, I. M.; Luzanov, V. A.; Yu, M.; Dikaev, M.; Kravchenko,
V. B.; Melekh, B. J. Thin Solid Films 1996, 280, 163-166.
(93) Becht, M.; Morishita, T. Chem. Vap. Deposition 1996, 2, 191-197.
(94) Merchant, P.; Jacowitz, R. D.; Tibbs, K.; Taber, R. C.; Laderman, S.
S. Appl. Phys. Lett. 1992, 60, 763-765.
5008 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
capillary tubes with an uncalibrated MelTemp melting point
apparatus. Thermogravimetric analysis data were collected under
CH
2
OCH
3
], 5.33 [s, 1 H, C(O)CHC(NH)], 11.62 [br s, 1 H, NH].
N: C, 69.67; H, 11.27; N, 5.80. Found:
Anal. Calcd for C14
27 2
H O
5
.0 (( 0.1) Torr of N
2
with a TA Instruments SDT 2960
C, 69.54; H, 11.35; N, 5.95.
3
+
simultaneous DTA-TGA instrument. Weight loss data and sub-
sequent activation data were collected at a ramp rate of 1.5 °C/
min, while isothermal data were recorded at 155 °C.
Synthesis of Anhydrous Lanthanide Chlorides, LnCl , Ln
3
) Ce3 , Nd , Gd , Er . Anhydrous LnCl3 reagents were
+
3+
3+
3+
obtained in near-quantitative yield by reaction of the corresponding
Synthesis of â-Ketoimine Ligands [H(R
R
1 2
R
3
)]. The â-dike-
Ln
2 3 4
O with excess NH Cl (>10 equiv for Ln ) La-Gd; >12 equiv
tones 2,2-dimethyl-3,5-hexanedione and 2,2-dimethyl-3,5-hep-
tanedione were synthesized according to a literature procedure.95
â-Ketoimines based on 2,2,6,6-tetramethyl-3,5-heptanedione (Hdpm)
were prepared by neat reaction of the trimethylsilyl enol ether
derivative of Hdpm with the amino ether following general literature
procedure.96 All other â-ketoimine ligands were synthesized by
condensation of the appropriate â-diketone with the corresponding
amino ethers in benzene.97 The pure â-ketoimine ligands were dried
for Ln ) Tb-Lu, Y, Sc) via modification of a published
procedure. Commercial oxides and NH
beaker and dissolved in 4:3 H O:concentrated HCl. The resulting
98
4
Cl were placed in a large
2
solution (0.3 M in Ln) was slowly evaporated to dryness with
careful stirring. The solid obtained was then ground in a mortar
and pestle and placed in a Schlenk sublimation tube, which was
then placed in a tube furnace and heated according to the following
schedule in vacuo (∼1 Torr): 1 day, 140 °C (to drive off excess
over MgSO
4
or molecular sieves overnight and then distilled in
water); 1 day, 250 °C [to form (NH
(NH LnCl
the complex chlorides and drive off any excess NH
was then allowed to cool and taken directly into the glovebox, where
4 2 5
) LnCl for Ln ) La-Gd or
vacuo directly onto molecular sieves in a nitrogen-filled storage
tube to ensure anhydrous quality.
4
)
3
6
for La ) Tb-Lu, Y, Sc]; 1 day, 360 °C (to decompose
Cl). The tube
4
Synthesis of 2,2-Dimethyl-5-N-(2-methoxyethyl-imino)-3-hex-
t
the solid LnCl was collected and stored.
Synthesis of Lanthanide Tris[bis(trimethylsilyl)amides], Ln-
anone, [H( BuMeMe)] (1). In a 250 mL round bottom flask fitted
3
with reflux condenser and Dean-Stark trap, 27.4 g (194 mmol) of
,2-dimethyl-3,5-hexanedione was dissolved in 100 mL of benzene.
3 2 3
) ]
, Ln3 ) Ce , Nd , Gd , Er . The lanthanide
+
3+
3+
3+
3+
[
N(SiMe
amides were synthesized following a modified literature procedure.
KN(SiMe was reacted in a mixture of 1:1 THF:toluene with
3 2 3
excess anhydrous LnCl to afford the corresponding Ln[N(SiMe ) ]
2
99
While the reaction mixture was stirred rapidly, 16.0 g (214 mmol)
of 2-methoxyethylamine was added and the reaction mixture
refluxed overnight. Water was collected in the Dean-Stark trap as
the reaction progressed. Residual water was removed azeotropically,
followed by distillation of benzene. The resulting yellow liquid was
then distilled in vacuo at 115 °C/300 mTorr to afford 28.7 g of
3 2
)
3
complex. The crude solid was then purified by repeated fractional
vacuum sublimation. Final recrystallization from pentane (-78 °C)
after sublimation led to purer Ln[N(SiMe ) ] reagents. Impurities
3 2 3
[unreacted alkali metal amide and/or HN(SiMe ) ] were typically
1
compound 1 as a pale yellow liquid (74.3% yield). H NMR (300
MHz, CDCl , δ): 1.15 [s, 9 H, C(CH ], 1.99 [s, 3 H, C(NH)-
CH ], 3.36 [s, 3 H, OCH ], 3.42 [m, 2 H, NHCH CH OCH ], 3.53
m, 2 H, NHCH CH OCH ], 5.12 [s, 1 H, C(O)CHC(NH)], 10.99
br s, 1 H, NH]. Anal. Calcd for C11 N: C, 66.29; H, 10.62;
3 2
present at 3-5 mol % in the final product as judged by NMR.
3
3 3
)
1
SiMe
3
H NMR shifts found for these amides are as follows (500
, δ): Ce[N(SiMe , -3.39; Nd[N(SiMe , -6.25;
Gd[N(SiMe , -11.07; Er[N(SiMe , 62.89.
3
3
2
2
3
MHz, C
D
6 6
3
)
]
2 3
3 2 3
) ]
[
[
2
2
3
3
)
2
]
3
3 2 3
) ]
21 2
H O
N, 7.03. Found: C, 66.46; H, 10.73; N, 7.13.
1 2 3 3
Synthesis of Lanthanide Tris-â-ketoiminates, Ln(R R R ) ,
Ln3 ) Ce , Nd , Gd , Er . Tris[2,2-dimethyl-5-N-(2-meth-
+
3+
3+
3+
3+
Synthesis of 2,2-Dimethyl-5-N-(2-methoxyethyl-imino)-3-hep-
t
t
oxyethyl-imino)-3-hexanato]cerium(III), [Ce( BuMeMe) ] (4). In
tanone, [H( BuEtMe)] (2). This compound was prepared analo-
3
the glovebox, a two-neck, 100 mL round bottom flask fitted with
a rubber septum, magnetic stir bar, and reflux condenser with
vacuum adapter was charged with 1.49 g of Ce[N(SiMe ) ] (2.40
3 2 3
gously to compound 1, starting with 30.0 g (192 mmol) of 2,2-
dimethyl-3,5-heptanedione and 15.8 g (211 mmol) of 2-methoxy-
ethylamine. The resulting yellow liquid was distilled at 82 °C/200
mmol). The reaction vessel was then removed from the glovebox
and immediately interfaced to a Schlenk line, and 24 mL of xylenes
was added via syringe. The reaction flask was then immersed in a
preheated oil bath (∼150 °C) and the reaction mixture stirred until
homogeneous. Immediately, 1.57 g of ligand 1 (7.90 mmol) was
injected into the reaction flask, and the solution was refluxed
overnight (14 h), gradually changing color from golden yellow to
a deeper amber. The volatiles were then removed in vacuo (1 Torr),
and the resulting waxy amber solid was recrystallized from pentane
mTorr to afford 28.7 g of compound 2 as a pale yellow liquid
1
(
78.7% yield). H NMR (300 MHz, CDCl
3
, δ): 1.15 [s, 9 H,
], 2.28 [q, 2 H, CH CH ], 3.39 [s,
], 3.43 [m, 2 H, NHCH CH OCH ], 3.54 [m, 2 H,
OCH ], 5.18 [s, 1 H, C(O)CHC(NH)], 11.02 [br s, 1
H, NH]. Anal. Calcd for C12 N: C, 67.57; H, 10.87; N, 6.57.
C(CH
H, OCH
NHCH CH
3
3
) ], 1.16 [t, 3 H, CH
2
CH
3
2
3
3
3
2
2
3
2
2
3
23 2
H O
Found: C, 68.39; H, 10.94; N, 5.97.
Synthesis of 2,2,6,6-Tetramethyl-5-N-(2-methoxyethyl-imino)-
t
t
3
-heptanone, [H( Bu BuMe)] (3). In a 250 mL round bottom flask
fitted with reflux condenser was placed at 26 g (102 mmol) of
-ene-2,2,6,6-tetramethylsiloxy-3-heptanedione. While the reaction
at -30 °C to afford 1.50 g of amber crystalline 4 (85% yield); mp
1
9
1
5-98 °C. H NMR (300 MHz, C
6 6
D
, δ): -10.8, -8.4, -2.4, 1.2,
Ce: C, 53.93;
4
60 6 3
.6, 2.6, 2.8, 5.1, 13.5. Anal. Calcd for C33H O N
mixture was stirred rapidly, 8.4 g (112 mmol) of 2-methoxyethyl-
amine was added and the reaction mixture heated to 90-95 °C
overnight. The resulting yellow liquid was distilled twice in vacuo
H, 8.23; N, 5.72. Found: C, 53.60; H, 8.26; N, 5.44.
Tris[2,2-dimethyl-5-N-(2-methoxyethyl-imino)-3-hexanato]-
t
neodymium(III), [Nd( BuMeMe)
3
] (5). This Nd complex was
at 87 °C/500 mTorr to afford 9.1 g of compound 3 as a pale yellow
1
prepared analogously to compound 4, starting with 1.50 g of Nd-
N(SiMe (2.40 mmol) and 1.57 g of ligand 1 (7.90 mmol). The
liquid (37.0% yield). H NMR (300 MHz, CDCl
3
, δ): 1.15 [s, 9
], 3.41 [s, 3 H,
], 3.65 [m, 2 H, NHCH
[
3 2 3
) ]
3 3 3 3
H, C(O)C(CH ) ], 1.26 [s, 9 H, C(NH)C(CH )
reaction mixture changed color from bright blue to a blue-green.
The resulting waxy blue-green solid was recrystallized from pentane
at -30 °C to afford 1.51 g of blue-green crystalline 5 (85% yield);
OCH
3
], 3.59 [m, 2 H, NHCH
2
CH
2
OCH
3
2
-
(
(
(
95) Swamer, F. W.; Hauser, C. R. J. Am. Chem. Soc. 1950, 72, 1352-
1356.
96) Shin, H.-K.; Hampden-Smith, M. J.; Kodas, T. T.; Rheingold, A. L.
J. Chem. Soc., Chem. Commun. 1992, 217-219.
(98) Meyer, G.; Garcia, E.; Corbett, J. D. Inorg. Synth. 1989, 25, 146-
150.
97) Moffett, R. B.; Hoehn, W. M. J. Am. Chem. Soc. 1947, 69, 1792-
(99) Bradley, D. C.; Ghotra, J. S.; Hart, F. A. J. Chem. Soc., Dalton Trans.
1973, 1021-1023.
1794.
Inorganic Chemistry, Vol. 41, No. 20, 2002 5009

Edleman et al.
1
mp 78-81 °C. H NMR (300 MHz, C
-
6
D
6
, δ): -18.8, -14.6, -3.2,
Nd: C,
78 6 3
-31.4, -19.6, -1.3, 8.7, 15.5, 21.8. Anal. Calcd for C42H O N -
Nd: C, 58.30; H, 9.09; N, 4.86. Found: C, 58.48; H, 9.14; N, 4.60.
0.3, 1.2, 1.6, 3.0, 7.9, 16.2. Anal. Calcd for C33
H
60
O
6
N
3
5
3.63; H, 8.18; N, 5.69. Found: C, 53.84; H, 8.10; N, 5.62.
Tris[2,2,6,6-tetramethyl-5-N-(2-methoxyethyl-imino)-3-hep-
t
t
Tris[2,2-dimethyl-5-N-(2-methoxyethyl-imino)-3-hexanato]-
tanato]erbium(III), [Er( Bu BuMe)
prepared analogously to compound 4, starting with 1.00 g of Er-
N(SiMe (1.50 mmol) and 1.23 g of ligand 3 (5.10 mmol). The
3
] (12). This Er complex was
t
gadolinium(III), [Gd( BuMeMe)
prepared analogously to compound 4, starting with 1.60 g of Gd-
N(SiMe (2.50 mmol) and 1.65 g of ligand 1 (8.30 mmol). The
3
] (6). This Gd complex was
[
3 2 3
) ]
reaction mixture changed color from pink to orange. The resulting
waxy orange solid was recrystallized from pentane at -30 °C to
[
3 2 3
) ]
reaction mixture changed color from colorless to a light orange.
The resulting waxy orange solid was recrystallized from pentane
at -30 °C to afford 0.95 g of light orange crystalline 6 (51% yield);
mp 76-78 °C. H NMR studies were not possible due to the highly
paramagnetic nature (4f ) of this complex. Anal. Calcd for
afford 0.93 g of bright orange crystalline 12 (70% yield); mp 135-
1
1
43 °C. H NMR (400 MHz, C
6 6
D , δ): -8.0, -5.5. Anal. Calcd
1
42 78 6 3
for C H O N Er: C, 56.79; H, 8.85; N, 4.73; Found: C, 56.25;
H, 8.71; N, 4.29.
7
C
33
H
60
O
6
N
3
Gd: C, 52.70; H, 8.04; N, 5.59. Found: C, 52.70; H,
Single-Crystal X-ray Diffraction Studies. X-ray data for single
crystals of lanthanide complexes 4-7, 9, 10, and 12 [grown from
pentane solution and mounted on a glass fiber with Paratone-N
(Exxon)] were collected on a CCD area detector with graphite-
monochromated Mo KR radiation. Reflections were collected with
a Bruker SMART detector and processed with SAINT-NT from
Bruker. Data were corrected for Lorentz and polarization effects.
The structures were solved by direct methods and expanded using
Fourier techniques. The non-hydrogen atoms were refined aniso-
tropically. Hydrogen atoms were included but not refined. All
calculations were performed using the Bruker SHELXTL crystal-
lographic software package. The space groups for 4, 5, 6, 7, 9, 10,
8
.05; N, 5.44.
Tris[2,2-dimethyl-5-N-(2-methoxyethyl-imino)-3-hexanato]er-
t
bium(III), [Er( BuMeMe)
analogously to compound 4, starting with 1.56 g of Er[N(SiMe
3
] (7). This Er complex was prepared
3 2 3
) ]
(
2.40 mmol) and 1.57 g of ligand 1 (7.90 mmol). The reaction
mixture changed color from pink to orange. The resulting waxy
orange solid was recrystallized from pentane at -30 °C to afford
1
.55 g of bright orange crystalline 7 (85% yield); mp 65-68 °C.
1
H NMR (300 MHz, C
6
D
6
, δ): -47.9, -15.8, -2.5, 0.8, 1.6, 1.8,
Er: C, 52.01; H, 7.94;
2.6, 5.2, 35.5. Anal. Calcd for C33H O N
60 6 3
N, 5.51. Found: C, 51.89; H, 8.21; N, 5.61.
and 12 were determined to be P 1h , P2
1
/c, P2
1
1
/n, P 1h , P 1h , P2 /c, and
Tris[2,2-dimethyl-5-N-(2-methoxyethyl-imino)-3-heptanato]-
P2 /c, respectively, based on a statistical analysis of intensity
1
t
cerium(III), Ce( BuEtMe)
analogously to compound 4, starting with 1.49 g of Ce[N(SiMe
3
] (8). This Ce complex was prepared
distributions, and the successful solution and refinement of the
structures. Important experimental parameters for these structure
determinations are tabulated in Tables 1-3. The asymmetric unit
of complex 4 contains four crystallographically independent, but
chemically equivalent, molecules. Also, the two tert-butyl groups
at C(21) and C(44) and the methoxy group at C(50) in compound
3 2 3
) ]
(
2.40 mmol) and 1.68 g of ligand 2 (7.90 mmol). The reaction
mixture changed color from golden yellow to a deeper amber. The
resulting waxy amber solid was recrystallized from pentane at -30
°
C to afford 0.89 g of amber crystalline 8 (49% yield); mp 99-
1
101 °C. H NMR (500 MHz, C
6
D
H
6
, δ): -10.3, -8.2, -2.0, 1.2,
Ce: C, 55.65; H, 8.56;
5
are disordered over two equally occupied positions. The effect
2.1, 5.1, 12.9. Anal. Calcd for C36
66 6 3
O N
on the metrical parameters of interest is negligible. Data for all
these complexes has been deposited at the Cambridge Crystal-
N, 5.41. Found: C, 55.57; H, 8.74; N, 5.52.
Tris[2,2-dimethyl-5-N-(2-methoxyethyl-imino)-3-heptanato]-
100
lographic Database Center.
t
neodymium(III), [Nd( BuEtMe)
3
] (9). This Nd complex was
2
Film Growth Experiments. MOCVD of CeO thin films was
prepared analogously to compound 4, starting with 0.81 g of Nd-
N(SiMe (1.30 mmol) and 0.92 g of ligand 2 (4.30 mmol). The
carried out in the low-pressure, horizontal, cold-wall MOCVD
[
)
3 2
]
3
26
system described elsewhere. The metal-organic precursor and
reaction mixture changed color from bright blue to a blue-green.
The resulting waxy blue-green solid was recrystallized from pentane
at -30 °C to afford 0.63 g of blue-green crystalline 9 (62% yield);
the reactant gas (100 sccm of O ; sccm ) standard cubic centimeters
2
per minute) vapor streams were mixed immediately upstream of
the reactor. The precursor was maintained at 150-160 °C and
introduced into the reactor using ultrahigh-purity Ar flowing at 70
sccm. Deposition was carried out at a system pressure of 3.5 Torr,
and the single-crystal (001) YSZ substrates (purchased from
Materials Technology International, Inc.) were externally heated
by an IR heater to 450-650 °C. To verify the viability of low-
1
mp 85-88 °C. H NMR (400 MHz, C
6
D
6
, δ): -19.3, -15.1, -2.9,
Nd: C, 55.35; H, 8.52;
0.7, 7.9, 16.0. Anal. Calcd for C36
H
66
O
6
N
3
N, 5.38; Found: C, 55.43; H, 8.45; N, 5.47.
Tris[2,2-dimethyl-5-N-(2-methoxyethyl-imino)-3-heptanato]-
t
erbium(III), [Er( BuEtMe)
analogously to compound 4, starting with 1.57 g of Er[N(SiMe
3
] (10). This Er complex was prepared
) ]
3 2 3
temperature CeO MOCVD deposition for HTS buffers, pulsed
2
(
2.40 mmol) and 1.70 g of ligand 2 (7.90 mmol). The reaction
organometallic beam epitaxy (POMBE) was employed to deposit
mixture changed color from pink to orange. The resulting waxy
YBCO films at 725 °C on the MOCVD-derived CeO /YSZ (001)
2
orange solid was recrystallized from pentane at -30 °C to afford
structures. Details of the YBCO deposition process have been
described elsewhere.1 Growth parameter dependence of deposited
film structure, crystallinity, and thickness were investigated by Ni-
filtered Cu KR θ-2θ and φ scan X-ray diffraction (XRD; Matrix
and Scintag XDS 2000 instruments) and cross-sectional transmission
01
0
.84 g of bright orange crystalline 10 (44% yield); mp 81-82 °C.
1
H NMR (500 MHz, C , δ): -81.0, -49.4, -16.7, -2.7, 1.8,
D
6 6
3
5.6. Anal. Calcd for C36H O N Er: C, 53.77; H, 8.27; N, 5.23;
66 6 3
Found: C, 53.52; H, 8.46; N, 5.27.
Tris[2,2,6,6-tetramethyl-5-N-(2-methoxyethyl-imino)-3-hep-
t
t
(100) Supplementary crystallographic data for complexes 4 (114935), 5
tanato]neodymium(III), [Nd( Bu BuMe)
was prepared analogously to compound 4, starting with 1.10 g of
Nd[N(SiMe (1.80 mmol) and 1.40 g of ligand 3 (5.80 mmol).
3
] (11). This Nd complex
(
174766), 6 (175017), 7 (174765), 9 (178184), 10 (178185), and 12
(178272) can be obtained free of charge from the Cambridge Crystal-
3 2 3
) ]
lographic Database Center (CCDC) via www.ccdc.cam.ac.uk/conts/
retrieving.html (or from the CCDC, 12 Union Road, Cambridge CB2
The reaction mixture changed color from bright blue to a blue-
green. The resulting waxy blue-green solid was recrystallized from
1
EZ, U.K.; fax, +44 1223 336033; e-mail, deposit@ccdc.cam.ac.uk).
(
101) Duray, S. J.; Buchholz, D. B.; Song, S. N.; Richeson, D. S.; Ketterson,
J. B.; Marks, T. J.; Chang, R. P. H. Appl. Phys. Lett. 1991, 59, 1503-
1505.
pentane at -30 °C to afford 0.74 g of blue-green crystalline 11
1
(
49% yield); mp 125-127 °C. H NMR (400 MHz, C
6 6
D , δ):
5010 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
Table 1. Crystal Data and Structure Refinement for Compounds 4,
Table 2. Crystal Data and Structure Refinement for Compounds 7,
9, and 10
5
, and 6
t
t
t
t
t
t
complex identity
Ce( BuMeMe)
3
,
Nd( BuMeMe)
3
,
Gd( BuMeMe)
3
,
complex identity
Er( BuMeMe)
3
,
Nd( BuEtMe)
3
,
Er( BuEtMe)
3
,
4
5
6
7
9
10
empirical formula
C H N
33 60 3
6
O Ce
C
33
H
60
N
3
6
O Nd
C
H
33 60
3 6
N O Gd
empirical formula
33 60 3
C H N O
6
Er
C
36
H
66
N
3
6
O Nd
36 66 3 6
C H N O Er
fw
temp (K)
734.96
198(2)
739.08
198(2)
752.09
153(2)
fw
temp (K)
762.10
293(2)
781.16
153(2)
804.18
153(2)
wavelength (Å)
cryst syst,
space group
0.71073
triclinic,
P1h
0.71073
monoclinic,
P2 /c
0.71073
monoclinic,
P2 /c
wavelength (Å)
cryst syst,
space group
0.71073
triclinic,
P1h
0.71073
triclinic,
P1h
0.71073
monoclinic,
P2 /c
1
1
1
unit cell dimens (Å)
unit cell dimens (Å)
a
b
c
17.5040(2)
21.1598(2)
23.0715(2)
13.3606(2)
17.1475(2)
16.4002(2)
10.6115(6)
25.2146(15)
14.1149(9)
a
b
c
10.0926(1)
11.0198(1)
18.5054(1)
11.5887(9)
12.5994(9)
14.6711(11)
21.0002(12)
9.8643(6)
19.3381(11)
unit cell angles (deg)
unit cell angles (deg)
R
â
γ
112.8853(6)
90.9699(11)
99.5012(7)
7733.63(12)
8, 1.262
90.00
98.9027(7)
90.00
3712.02(8)
4, 1.322
90.00
100.8080(10)
90.00
3709.7(4)
4, 1.347
R
â
γ
102.882(1)
97.346(2)
108.370(1)
1859.92(2)
2, 1.361
97.6820(10)
97.7390(10)
104.2170(10)
2026.4(3)
90
91.3530(10)
90
4004.8(4)
4, 1.334
3
3
vol (Å )
vol (Å )
Z, calcd
Z, calcd
density (g/cm )
abs coeff (mm
F(000)
2, 1.280
3
3
density (g/cm )
abs coeff (mm-
F(000)
1
)
1.218
3080
1.441
1548
1.830
1564
-1
)
1.218
790
1.323
822
2.138
1676
cryst size (mm)
0.20 × 0.20 ×
0.30 × 0.20 ×
0.31 × 0.20 ×
cryst size (mm)
0.40 × 0.30 ×
0.36 × 0.16 ×
0.228 × 0.180 ×
0
.10
0.15
0.16
0.15
0.18
0.070
θ range for data
collection (deg)
limiting indices
0.96-28.13
1.73-26.00
1.62-28.95
θ range for data
collection (deg)
limiting indices
2.03-24.00
1.42-28.30
0.97-28.34
-22 e h e 23
-17 e h e 15
0 e k e 19
0 e l e 21
8579/4865
-14 e h e 14
-32 e k e 33
-19 e l e 18
33706/9106
-11 e h e 10
-12 e k e 11
-21 e l e 20
10894/5714
[R(int) ) 0.0357] [R(int) ) 0.0217] [R(int) ) 0.0312]
97.8
-15 e h e 15
-16 e k e 16
-19 e l e 19
18669/9373
-27 e h e 26
-13 e k e 12
-24 e l e 25
35724/9618
-
28 e k e 27
29 e l e 23
-
reflns collected/
unique
49148/32224
reflns collected/
unique
[R(int) ) 0.0407] [R(int) ) 0.1134] [R(int) ) 0.0278]
85.1
completeness to
θ ) 28.13° (%)
abs correction
refinement meth
data/restraints/
params
98.7
92.6
completeness to
θ ) 28.13° (%)
abs correction
refinement meth
data/restraints/
params
92.9
96.3
none
DIFABS
full-matrix least-squares on F
none
none
face centered and SADABS
full-matrix least-squares on F
2
2
32224/0/1549
4735/0/402
9106/0/388
5714/0/389
9373/0/415
9618/0/415
GOF on F2
1.120
1.667
1.058
GOF on F
2
1.149
1.062
1.049
final R indices
R1 ) 0.0528
wR2 ) 0.1157
R1 ) 0.0884
wR2 ) 0.1326
0.987 and
R1 ) 0.0851
wR2 ) 0.2306
R1 ) 0.1159
wR2 ) 0.2713
0.822 and
R1 ) 0.0324
wR2 ) 0.0780
R1 ) 0.0403
wR2 ) 0.0816
2.462 and
final R indices
R1 ) 0.0318
wR2 ) 0.0957
R1 ) 0.0360
wR2 ) 0.1026
1.869 and
R1 ) 0.0299
wR2 ) 0.0736
R1 ) 0.0356
wR2 ) 0.0764
1.324 and
R1 ) 0.0269
wR2 ) 0.0649
R1 ) 0.0358
wR2 ) 0.0683
2.342 and
I > 2σ(I)]a
[I > 2σ(I)]
a
[
R indices
R indices
(all data)
all data)a
a
(
largest diff peak
largest diff peak
and hole
-2.307
-1.100
-0.710
and hole
-0.754
-0.701
-0.605
-
-3
-
-3
(
e /Å
)
(e /Å )
a
R1 ) ∑||Fo| - |Fc||/∑|Fo| and wR2 ) {∑[w(Fo - Fc ) ]/∑[w(Fo ) ]}1/2,
2
2 2
2 2
a
R1 ) ∑||Fo| - |Fc||/∑|Fo| and wR2 ) {∑[w(Fo - Fc ) ]/∑[w(Fo ) ]}1/2,
2
2 2
2 2
2
2
2
2
2
2
2
2
2
2
where w ) 1/[σ (Fo ) + (0.1551P) + 0.0000P] and P ) (Fo + 2Fc )/3.
where w ) 1/[σ (Fo ) + (0.1551P) + 0.0000P] and P ) (Fo + 2Fc )/3.
electron microscopy (TEM; Hitachi S4500 FE instrument). Con-
ventional four-probe methods were used for measuring the YBCO
ligand (see Figure 1 for the nomenclature scheme) in high-
boiling xylenes overnight (eq 1). In this procedure, the
film critical current density, J
c
, and the YBCO superconducting
transition temperature, T , was measured inductively. Electrical
c
characterization techniques have been outlined elseware.26 Film
surface roughness and morphology were analyzed with a Digital
Instruments Nanoscope IIIa atomic force microscope (AFM) used
in the contact mode.
Results
The synthesis of monomeric lanthanide complexes of type
VIII is outlined here. These compounds are low-melting,
thermally stable, and volatile, despite the fact that they
contain no fluorinated substituents. Studies of alkyl substitu-
tion at the ligand keto, imino, and ether sites and the resulting
lanthanide complex melting point and volatility variation are
also detailed. Implementation of a new type VIII precursor
2
in heteroepitaxial CeO thin film growth for superconductor
current-carrying applications is also described.
MOCVD Precursor Synthesis. The present lanthanide
â-ketoiminate complexes are synthesized in a single-step
amine elimination reaction by refluxing the corresponding
lanthanide tris[bis(trimethylsilyl)amide] with excess free
desired species are readily isolated, since the only reaction
3 2
byproduct is volatile HN(SiMe ) . The products are purified
Inorganic Chemistry, Vol. 41, No. 20, 2002 5011

Edleman et al.
Table 3. Crystal Data and Structure Refinement for Compound 12
t
t
complex identity
empirical formula
fw
Er( Bu BuMe)3, 12
C42H78N3O6Er
888.33
temp (K)
153(2)
wavelength (Å)
0.71073
cryst syst, space group
monoclinic, P21/c
unit cell dimens (Å)
a
b
c
10.3100(7)
27.3356(17)
17.4357(11)
unit cell angles (deg)
R
â
γ
90
105.4000(10)
90
4737.5(5)
4, 1.245
3
vol (Å )
3
Z, calcd density (g/cm )
abs coeff (mm
-
1
)
1.814
F(000)
1868
cryst size (mm)
θ range for data collection (deg)
limiting indices
0.36 × 0.39 × 0.30
1.42-28.30
-13 e h e 13
-
36 e k e 35
17 e l e 22
-
reflns collected/unique
completeness to θ ) 28.30° (%)
abs correction
34982/11218 [R(int) ) 0.0190]
95.3
face and SADABS
full-matrix least-squares on F2
11218/0/469
refinement meth
data/restraints/params
2
GOF on F
1.038
a
final R indices [I > 2σ(I)]
R1 ) 0.0250, wR2 ) 0.0597
R1 ) 0.0295, wR2 ) 0.0621
1.812 and -0.696
a
R indices (all data)
largest diff peak and hole (e /Å-3)
-
a
R1 ) ∑||Fo| -|Fc||/∑|Fo| and wR2 ) {∑[w(Fo - Fc ) ]/∑[w(Fo ) ]}1/2,
2
2 2
2 2
2
2
2
2
2
where w ) 1/[σ (Fo ) + (0.1551P) + 0.0000P] and P ) (Fo + 2Fc )/3.
Figure 2. Idealized representation of the lanthanide â-ketoiminato
complexes synthesized in this study. Note that not all ether oxygen atoms
are coordinated in all complexes.
Table 4. Synthetic Yield, Melting Point, and Coordination Number of
New â-Ketoiminate Lanthanide MOCVD Precursors
synthetic
yield (%)
coord
no.
a
complex name
mp (°C)
t
4
5
6
7
8
, Ce( BuMeMe)3
85
85
51
85
49
62
44
49
70
95-98
78-81
8
8
8
7
8
8
7
t
, Nd( BuMeMe)3
t
, Gd( BuMeMe)3
76-78
t
, Er( BuMeMe)3
65-68
t
b
, Ce( BuEtMe)3
99-101
85-88
t
Figure 1. Schematic representation of the â-ketoimine ligands synthesized
9, Nd( BuEtMe)3
t
in this study.
10, Er( BuEtMe)3
1, Nd( Bu BuMe)3
12, Er( Bu BuMe)3
81-82
t
t
b
1
125-128
135-143
8
7
t
t
by recrystallization from pentane to yield homoleptic,
3
+
monomeric, colorful lanthanide complexes (Ce , amber;
a
As determined from single-crystal X-ray diffraction. b Assumed.
3+
3+
3+
Nd , blue-green; Gd , light orange; Er , bright orange;
see Figure 2 for a schematic representation of the complexes
isolated). Synthetic yields are listed in Table 4. The com-
plexes are stable under an inert atmosphere and can be readily
and easily sublimed at 80-110 °C/10 Torr. Melting points
are in the 60-130 °C range (Table 4), affording liquid,
constant surface area, thermally stable precursors at operating
MOCVD precursor reservoir temperatures.
substitution effects at a particular molecular site. As sum-
marized in Table 4 and shown in Figure 3, two melting point
trends are noteworthy. First, the melting point decreases with
ionic radius across the lanthanide series for a given ligand
system (i.e., mp 4, 95-98 °C; 5, 78-81 °C; 6, 76-78 °C;
7, 65-68 °C; see Figure 3a). Second, introduction of larger
alkyl groups (e.g., substitution of ethyl or tert-butyl for
-4
Precursor Melting Point Trends. Alkyl ligand substit-
uents were deliberately selected such that lanthanide precur-
sor physical properties could be assessed with respect to
2
methyl) at the R site results in an increase in melting point
(i.e., mp 5, 78-81 °C; 9, 85-88 °C; 11, 125-128 °C; see
Figure 3b). Attempts to synthesize complexes with substit-
5012 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
Table 5. Selected Bond Lengths (Å) and Angles (deg) for Complexes
and 5
4
4
5
Ce(1A)-O(41A)
Ce(1A)-O(21A)
Ce(1A)-O(1A)
Ce(1A)-N(21A)
Ce(1A)-N(41A)
Ce(1A)-N(1A)
Ce(1A)-O(2A)
Ce(1A)-O(22A)
2.349(3) Nd(1)-O(1)
2.374(3) Nd(1)-O(41)
2.385(3) Nd(1)-O(21)
2.616(4) Nd(1)-N(1)
2.663(4) Nd(1)-N(41)
2.684(4) Nd(1)-N(21)
2.694(3) Nd(1)-O(22)
2.823(3) Nd(1)-O(2)
2.278(1)
2.289(1)
2.315(9)
2.518(9)
2.568(1)
2.582(1)
2.649(8)
2.813(9)
O(41A)-Ce(1A)-O(21A) 85.75(1) O(1)-Nd(1)-O(41) 85.3(4)
O(41A)-Ce(1A)-O(1A) 147.42(1) O(1)-Nd(1)-O(21) 121.8(4)
O(21A)-Ce(1A)-O(1A) 119.37(1) O(41)-Nd(1)-O(21) 147.1(3)
O(41A)-Ce(1A)-N(21A) 87.26(1) O(1)-Nd(1)-N(1)
70.9(3)
O(21A)-Ce(1A)-N(21A) 69.82(1) O(41)-Nd(1)-N(1) 89.3(3)
O(1A)-Ce(1A)-N(21A) 83.22(1) O(21)-Nd(1)-N(1) 83.7(3)
O(41A)-Ce(1A)-N(41A) 69.66(1) O(1)-Nd(1)-N(41) 143.5(4)
O(21A)-Ce(1A)-N(41A) 143.31(1) O(41)-Nd(1)-N(41) 70.9(3)
O(1A)-Ce(1A)-N(41A) 94.51(1) O(21)-Nd(1)-N(41) 91.1(4)
N(21A)-Ce(1A)-N(41A) 132.61(1) N(1)-Nd(1)-N(41) 133.6(3)
O(41A)-Ce(1A)-N(1A) 141.54(1) O(1)-Nd(1)-N(21) 79.2(3)
O(21A)-Ce(1A)-N(1A) 80.30(1) O(41)-Nd(1)-N(21) 139.8(3)
O(1A)-Ce(1A)-N(1A)
68.04(1) O(21)-Nd(1)-N(21) 69.2(3)
N(21A)-Ce(1A)-N(1A) 120.19(1) N(1)-Nd(1)-N(21) 119.3(4)
N(41A)-Ce(1A)-N(1A) 102.01(1) N(41)-Nd(1)-N(21) 101.2(4)
O(41A)-Ce(1A)-O(2A) 79.79(1) O(1)-Nd(1)-O(22) 73.5(3)
O(21A)-Ce(1A)-O(2A) 74.47(1) O(41)-Nd(1)-O(22) 78.0(3)
O(1A)-Ce(1A)-O(2A)
124.62(1) O(21)-Nd(1)-O(22) 124.6(3)
N(21A)-Ce(1A)-O(2A) 142.74(1) N(1)-Nd(1)-O(22) 143.0(3)
N(41A)-Ce(1A)-O(2A) 74.67(1) N(41)-Nd(1)-O(22) 74.7(3)
N(1A)-Ce(1A)-O(2A)
62.00(1) N(21)-Nd(1)-O(22) 62.1(3)
129.9(3)
O(41A)-Ce(1A)-O(22A) 77.28(1) O(1)-Nd(1)-O(2)
O(21A)-Ce(1A)-O(22A) 129.81(1) O(41)-Nd(1)-O(2) 78.4(3)
O(1A)-Ce(1A)-O(22A) 70.66(1) O(21)-Nd(1)-O(2) 69.9(3)
N(21A)-Ce(1A)-O(22A) 62.58(1) N(1)-Nd(1)-O(2)
62.0(3)
N(41A)-Ce(1A)-O(22A) 72.01(1) N(41)-Nd(1)-O(2) 72.9(3)
N(1A)-Ce(1A)-O(22A) 137.62(1) N(21)-Nd(1)-O(2) 138.5(3)
O(2A)-Ce(1A)-O(22A) 144.41(1) O(22)-Nd(1)-O(2) 144.7(3)
monodentate (Figures 16 and 17). Comparisons to known
structures are made in the Discussion section.
Precursor Volatility Trends. Several of the present
lanthanide precursors were further characterized by vacuum
thermogravimetric analysis (TGA) at pressures approximat-
ing those used in film growth. The new precursors were
Figure 3. Melting point trends of â-ketoiminato lanthanide complexes as
a function of (a) Ln identity and (b) alkyl substituent at ligand imino
R2 site.
3
+
uents bulkier than a methyl group at the ether R
3
site were
t
successful [i.e., Ln( BuEtEt)
3
and Ln(MeMeEt)
3
]; however,
directly compared to commercially supplied Ce(dpm)
Er(dpm) . Figure 11a shows the raw comparative weight loss
versus temperature data for cerium complexes 4, 8, and Ce-
(dpm) . An Arrhenius analysis (sublimation rate versus
4
and
only qualitative assessment of physical characteristics was
possible owing to the thick, viscous nature of the liquid
products.
3
4
Precursor Molecular Structures. Single-crystal X-ray
diffraction studies were carried out on lanthanide complexes
inverse temperature; Figure 11b) for these Ce complexes
allows direct comparison of the molar volatility/transport rate
2
6
4-7, 9, 10, and 12. Important bond distances and angles for
characteristics. Finally, the Ce complexes were character-
ized in isothermal experiments (Figure 11c). Note that
complex 4 exhibits the highest molar volatility under
MOCVD reactor conditions (155 °C; lowest apparent activa-
tion energy; see Figure 11b). Similar plots are shown for Er
these structures are tabulated in Tables 5-8, and relevant
lanthanide ionic radii are listed in Table 9. Generally, the
3+
3+
3+
crystal structures of the 8-coordinate Ce , Nd , and Gd
complexes [i.e., 4, 5, 6, and 9] reveal that two of the three
ligands are coordinated to the metal through all three possible
donor sites; however, the third ligand is bidentate, with a
noncoordinating ether moiety (Figures 4 and 5, 6 and 7, 8
and 9, and 10 and 11, respectively). Conversely, the solid-
state structures of Er3 complexes 7 and 10 are 7-coordinate,
with two dangling ether groups (Figures 12 and 13, and 14
and 15, respectively). The crystal structure of complex 12
reveals that this Er3 compound is also 7-coordinate, although
two ligands act as tridentate donors and the third is
complex 10 and Er(dpm) (Figures 12a-c) and reveal that
3
10 has constant vaporization characteristics with little ap-
parent thermal instability.
CeO
2 2
Film Growth Experiments. Epitaxial CeO films
+
were deposited in situ on (001)-oriented YSZ substrates using
t
3
newly developed Ce precursor 4, Ce( BuMeMe) . The
reservoir containing 4 was maintained at 150-160 °C, and
Ar flowing at 70 sccm carried the precursor into the reactor
+
hot zone, where it reacted with 100 sccm of O
2
to form the
Inorganic Chemistry, Vol. 41, No. 20, 2002 5013

Edleman et al.
Table 6. Selected Bond Lengths (Å) and Angles (deg) for Complexes
and 7
Table 7. Selected Bond Lengths (Å) and Angles (deg) for Complexes
9 and 10
6
6
7
9
10
Gd(1)-O(1)
Gd(1)-O(3)
Gd(1)-O(10)
Gd(1)-N(11)
Gd(1)-N(1)
Gd(1)-N(3)
Gd(1)-O(2)
Gd(1)-O(4)
2.2362(2) Er(1)-O(41)
2.2773(2) Er(1)-O(21)
2.193(3)
2.195(3)
2.212(3)
2.437(3)
2.439(3)
2.465(3)
2.535(3)
Nd(1)-O(1B)
Nd(1)-O(1A)
Nd(1)-O(1C)
Nd(1)-N(5B)
Nd(1)-N(5A)
Nd(1)-N(5C)
Nd(1)-O(9C)
Nd(1)-O(8B)
2.303(2)
2.304(2)
2.336(2)
2.543(2)
2.589(2)
2.647(2)
2.663(2)
2.758(2)
Er(1)-O(1A)
Er(1)-O(1C)
Er(1)-O(1B)
Er(1)-N(5B)
Er(1)-N(5C)
Er(1)-O(8C)
Er(1)-N(5A)
2.185(2)
2.196(16)
2.199(16)
2.438(2)
2.445(2)
2.476(2)
2.559(2)
2.280(2)
2.537(2)
2.540(2)
2.549(2)
2.554(2)
2.7520(2)
Er(1)-O(1)
Er(1)-N(41)
Er(1)-N(1)
Er(1)-O(2)
Er(1)-N(21)
O(1)-Gd(1)-O(3)
O(1)-Gd(1)-O(10)
O(3)-Gd(1)-O(10)
O(1)-Gd(1)-N(11)
O(3)-Gd(1)-N(11)
144.22(7) O(41)-Er(1)-O(21) 160.10(1)
O(1B)-Nd(1)-O(1A) 83.23(7)
O(1A)-Er(1)-O(1C) 83.78(6)
94.98(8)
O(41)-Er(1)-O(1)
114.88(1)
82.92(1)
O(1B)-Nd(1)-O(1C) 122.76(7) O(1A)-Er(1)-O(1B) 158.86(6)
O(1A)-Nd(1)-O(1C) 146.78(7) O(1C)-Er(1)-O(1B) 113.67(6)
111.39(7) O(21)-Er(1)-O(1)
83.58(8)
82.14(7)
O(10)-Gd(1)-N(11) 72.17(7)
O(41)-Er(1)-N(41) 76.33(1)
O(1B)-Nd(1)-N(5B) 70.93(7)
O(1A)-Nd(1)-N(5B) 84.56(7)
O(1C)-Nd(1)-N(5B) 85.43(7)
O(1A)-Er(1)-N(5B) 96.91(6)
O(1C)-Er(1)-N(5B) 78.32(6)
O(1B)-Er(1)-N(5B) 76.25(6)
O(21)-Er(1)-N(41) 100.69(1)
O(1)-Er(1)-N(41)
O(41)-Er(1)-N(1)
76.88(1)
77.32(1)
118.18(1)
73.58(1)
126.66(1)
95.18(1)
81.01(1)
123.09(1)
159.86(1)
67.31(1)
O(1)-Gd(1)-N(1)
O(3)-Gd(1)-N(1)
O(10)-Gd(1)-N(1)
N(11)-Gd(1)-N(1)
O(1)-Gd(1)-N(3)
O(3)-Gd(1)-N(3)
O(10)-Gd(1)-N(3)
N(11)-Gd(1)-N(3)
N(1)-Gd(1)-N(3)
O(1)-Gd(1)-O(2)
O(3)-Gd(1)-O(2)
O(10)-Gd(1)-O(2)
N(11)-Gd(1)-O(2)
N(1)-Gd(1)-O(2)
N(3)-Gd(1)-O(2)
O(1)-Gd(1)-O(4)
O(3)-Gd(1)-O(4)
O(10)-Gd(1)-O(4)
N(11)-Gd(1)-O(4)
N(1)-Gd(1)-O(4)
N(3)-Gd(1)-O(4)
O(2)-Gd(1)-O(4)
72.70(7)
134.58(7) O(21)-Er(1)-N(1)
78.25(7)
140.03(7) N(41)-Er(1)-N(1)
95.57(7)
74.43(7)
148.87(7) O(1)-Er(1)-O(2)
138.16(7) N(41)-Er(1)-O(2)
77.12(7)
O(1B)-Nd(1)-N(5A) 138.57(7) O(1A)-Er(1)-N(5C) 120.21(7)
O(1A)-Nd(1)-N(5A) 70.37(7)
O(1C)-Nd(1)-N(5A) 95.04(7)
O(1C)-Er(1)-N(5C) 73.51(6)
O(1B)-Er(1)-N(5C) 77.89(6)
O(1)-Er(1)-N(1)
N(5B)-Nd(1)-N(5A) 134.22(7) N(5B)-Er(1)-N(5C) 129.44(7)
O(41)-Er(1)-O(2)
O(21)-Er(1)-O(2)
O(1B)-Nd(1)-N(5C) 80.42(6)
O(1A)-Er(1)-O(8C) 81.72(6)
O(1A)-Nd(1)-N(5C) 141.40(7) O(1C)-Er(1)-O(8C) 122.40(6)
O(1C)-Nd(1)-N(5C) 69.08(6)
O(1B)-Er(1)-O(8C) 97.37(6)
N(5B)-Nd(1)-N(5C) 121.75(7) N(5B)-Er(1)-O(8C) 158.66(6)
N(5A)-Nd(1)-N(5C) 100.55(7) N(5C)-Er(1)-O(8C) 67.06(6)
N(1)-Er(1)-O(2)
139.37(7) O(41)-Er(1)-N(21) 87.04(1)
74.39(7)
70.77(7)
O(1B)-Nd(1)-O(9C) 74.11(6)
O(1A)-Nd(1)-O(9C) 80.40(6)
O(1A)-Er(1)-N(5A) 73.14(6)
O(1C)-Er(1)-N(5A) 147.69(6)
O(21)-Er(1)-N(21) 73.07(1)
O(1)-Er(1)-N(21) 145.90(1)
O(1C)-Nd(1)-O(9C) 123.85(6) O(1B)-Er(1)-N(5A) 86.03(6)
N(5B)-Nd(1)-O(9C) 143.24(7) N(5B)-Er(1)-N(5A) 82.29(6)
124.15(7) N(41)-Er(1)-N(21) 84.06(1)
67.26(7)
82.29(7)
70.35(6)
73.88(6)
137.55(6)
66.84(7)
129.47(7)
73.51(6)
144.15(6)
N(1)-Er(1)-N(21)
O(2)-Er(1)-N(21)
139.33(1)
77.21(1)
N(5A)-Nd(1)-O(9C) 70.51(6)
N(5C)-Nd(1)-O(9C) 61.57(6)
O(1B)-Nd(1)-O(8B) 130.29(6)
O(1A)-Nd(1)-O(8B) 77.20(7)
O(1C)-Nd(1)-O(8B) 70.11(6)
N(5B)-Nd(1)-O(8B) 62.14(7)
N(5A)-Nd(1)-O(8B) 75.01(6)
N(5C)-Nd(1)-O(8B) 138.35(6)
O(9C)-Nd(1)-O(8B) 143.54(6)
N(5C)-Er(1)-N(5A) 137.85(6)
O(8C)-Er(1)-N(5A) 76.93(6)
desired films. The overall working pressure was 3.5 Torr.
Epitaxial films were successfully obtained at susceptor
temperatures ranging between 450 and 650 °C, and the
effects of temperature variation on film microstructure were
closely examined (vide infra). Film thickness varied between
observed, repeating every 90° and in the same positions as
those from the YSZ substrates (Figure 13c), demonstrating
a high level of in-plane epitaxy. The fwhm from the φ scan
of {113} CeO
the orientation relationship between the film and the YSZ
substrate is therefore CeO (001)[ 1h 10]//YSZ (001)[ 1h 10].
AFM images of the as-grown CeO films show that the
2
is 1.65°. From the θ-2θ and φ XRD scans,
∼
100 and 200 nm, with a dependence on growth temperature
vide infra).
CeO Film Characterization. A typical θ-2θ X-ray
diffraction pattern of an as-grown CeO film (Figure 13a)
consists of only CeO and substrate (00l) reflections. The
intensity of diffraction from films grown for a fixed time
3.5 h) was found to decrease with increasing deposition
temperature, and diffraction from CeO films deposited at
50 °C is quite weak. The film out-of-plane orientation
characteristics were analyzed by an ω-rocking scan of the
002) reflection, and the fwhm, determined by least-squares
2
(
2
surfaces consist of a granular structure. At low-growth
temperatures (450 °C), the tetragonal-pyramidal shaped
hillock-type grains have sizes in the range 50-110 nm, while
the rms roughness is about 15.5 Å (Figure 14a). At higher
growth temperatures (above 540 °C), the surface becomes
much smoother (rms roughness ∼ 4.3 Å), and hillock surface
features are not detectable (Figure 14b).
2
2
2
(
2
6
Cross-sectional high-resolution electron microscopy
(HREM) was also used to investigate the epitaxial growth
(
fitting, was found to decrease from 1.0° to 0.2° with
increasing deposition temperature (450-650 °C). The ω-rock-
ing curves of the films grown at high temperature (>520
of the CeO films on the (001) YSZ substrates. The CeO2
film grown at 540 °C exhibits sizable regions of single
crystal, defect-free and epitaxial growth (Figure 15). HREM
images and selected area diffraction (Figure 16) also reveal
2
°
C) are in fact a superposition of a broad underlying peak
with a narrow one sitting on top. With increase of deposition
temperature, the narrow component tends to dominate. The
that the CeO films and the YSZ substrates have the same
orientational relationship as that defined by the XRD
2
ω scan fwhm from the CeO
.85° (Figure 13b).
XRD φ scans were performed to assess the quality of the
in-plane epitaxy. Four equivalent planes of reflection are
2
films deposited at 540 °C is
experiments (vide supra). The interface area between the
0
CeO and YSZ substrates contains atomic stacking disloca-
2
tions having a periodicity of ∼44 Å. As the deposition
temperature increases, the columnar size remains almost
5014 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
Table 8. Selected Bond Lengths (Å) and Angles (deg) for Complex 12
Er(1)-O(1C)
Er(1)-O(1A)
Er(1)-O(1B)
Er(1)-N(5A)
Er(1)-N(5B)
Er(1)-O(8B)
Er(1)-O(8A)
O(1A)-C(2A)
O(8A)-C(9A)
O(8A)-C(7A)
N(5A)-C(4A)
N(5A)-C(6A)
C(2A)-C(3A)
C(2A)-C(10A)
C(3A)-C(4A)
C(4A)-C(14A)
C(6A)-C(7A)
O(1B)-C(2B)
O(8B)-C(9B)
2.165(1)
2.173(1)
2.201(2)
2.415(2)
2.448(2)
2.478(2)
2.484(2)
1.300(2)
1.435(3)
1.437(3)
1.308(3)
1.467(3)
1.374(3)
1.535(3)
1.437(3)
1.565(3)
1.509(4)
1.275(3)
1.423(3)
O(8B)-C(7B)
N(5B)-C(4B)
N(5B)-C(6B)
C(2B)-C(3B)
C(2B)-C(10B)
C(3B)-C(4B)
C(4B)-C(14B)
C(6B)-C(7B)
O(1C)-C(2C)
O(8C)-C(9C)
O(8C)-C(7C)
N(5C)-C(4C)
N(5C)-C(6C)
C(2C)-C(3C)
C(2C)-C(10C)
C(3C)-C(4C)
C(4C)-C(14C)
C(6C)-C(7C)
1.429(3)
1.314(3)
1.469(3)
1.385(3)
1.540(3)
1.432(3)
1.563(3)
1.512(4)
1.319(2)
1.409(3)
1.419(3)
1.277(3)
1.450(3)
1.352(3)
1.544(3)
1.470(3)
1.542(3)
1.513(3)
O(1C)-Er(1)-O(1A)
98.84(6)
O(1A)-Er(1)-O(8B)
77.83(5)
O(1C)-Er(1)-O(1B) 176.56(6)
O(1B)-Er(1)-O(8B) 102.69(6)
N(5A)-Er(1)-O(8B) 150.64(6)
O(1A)-Er(1)-O(1B)
O(1C)-Er(1)-N(5A)
82.69(6)
90.58(6)
N(5B)-Er(1)-O(8B)
O(1C)-Er(1)-O(8A)
O(1A)-Er(1)-O(8A) 143.04(6)
O(1B)-Er(1)-O(8A)
N(5A)-Er(1)-O(8A) 68.14(6)
N(5B)-Er(1)-O(8A) 78.90(6)
O(8B)-Er(1)-O(8A) 139.13(5)
66.64(6)
90.01(5)
O(1A)-Er(1)-N(5A) 75.91(6)
O(1B)-Er(1)-N(5A)
86.80(6)
O(1C)-Er(1)-N(5B) 107.74(6)
O(1A)-Er(1)-N(5B) 130.48(6)
86.94(6)
O(1B)-Er(1)-N(5B)
73.24(6)
N(5A)-Er(1)-N(5B) 142.28(6)
O(1C)-Er(1)-O(8B) 80.68(6)
Table 9. Average Lanthanide-Ligand Atom Bond Distances and Bond
Distance Trends (Å) for Complexes 4-7, 9, 10, and 12
Figure 4. (a) Solid-state crystal structure of 8-coordinate cerium complex
coord Ln3 ionic
no.
+
Ln-
Ln-
Ln-
t
4, Ce( BuMeMe) . (b) Immediate coordination environment of 8-coordinate
3
complex
radiusa
O(keto) N(imino) O(ether)
cerium complex 4, Ce( BuMeMe) .
t
3
t
4
5
6
7
9
1
1
, Ce( BuMeMe)3
8
8
8
7
8
7
7
8
8
8
7
1.143
1.109
1.053
0.945
1.109
0.945
0.945
1.143
1.109
1.053
0.945
2.371(6) 2.655(2) 2.760(3)
2.294(4) 2.556(4) 2.731(9)
2.264(6) 2.543(9) 2.653(1)
2.200(3) 2.470(6) 2.465(3)
2.314(5) 2.593(2) 2.710(7)
2.194(2) 2.480(7) 2.476(2)
2.180(1) 2.432(2) 2.481(2)
2.371(6) 2.655(2) 2.760(3)
2.304(5) 2.574(8) 2.721(3)
2.264(6) 2.543(9) 2.653(1)
2.191(5) 2.461(2) 2.474(2)
posited by the pulsed MOCVD POMBE101 technique on the
present low-temperature (540 °C) MOCVD-derived epitaxial
t
, Nd( BuMeMe)3
t
, Gd( BuMeMe)3
t
, Er( BuMeMe)3
t
CeO
shown in Figure 17a, consists of bilayer film features with
c-axis oriented YBCO growth on the CeO film. The ω scan
2
films. XRD θ-2θ data for the YBCO film deposition,
, Nd( BuEtMe)3
t
0, Er( BuEtMe)3
2, Er( Bu BuMe)3
Ce av
Nd av
Gd av
Er av
t
t
2
3
+
3
+
fwhm of the YBCO films is 0.80° for the (005) reflection
(inset, Figure 17a). In φ scans on the present specimens, four
equivalent (102) YBCO planes of reflection are observed,
repeating every 90° (Figure 17b). The fwhm from the φ scan
of the {102} YBCO planes is as narrow as 1.30. Four-probe
transport measurements on these multilayer structures reveal
a transition temperature of 86.5 K (transition width ∼ 2.0
K; Figure 18a). The critical current density at 77.4 K is
3
+
3+
a
Ionic radii for corresponding coordination number.74
unchanged but the film surface becomes smoother. Film
thickness was also ascertained from cross-sectional TEM
samples. With an increase in deposition temperature, the
CeO
C to ∼6.5 Å/min at 540 °C.
Despite the columnar structure, the present CeO
2
growth rate decreases slightly from ∼10 Å/min at 450
6
2
1
.08 × 10 A/cm , as shown in Figure 18b.
°
Discussion
2
films are
very dense, and no fissures are evident either at the interface
or within the film (Figure 15). In fact, the column boundaries
are difficult to observe in the HREM. In contrast, the YSZ
substrates are found not to be atomically smooth. On the
The present synthetic coordination chemistry and film
growth study was undertaken to develop a new class of
lanthanide MOCVD precursors that would be thermally
stable, volatile, fluorine-free, and liquid at typical reservoir
temperatures for useful oxide growth conditions. No existing
commercial or previously synthesized precursor in the
published literature possesses all of these qualifications. The
discussion below shows that the broad new class of lan-
thanide â-ketoiminate complexes reported here satisfies all
of the desired characteristics and proves to be suitable for
oxide MOCVD, as demonstrated by growth of high-quality
2
step planes, the CeO films grow with an atomically abrupt
interface. However, at the YSZ step edges, slight bending
of crystal planes parallel to the substrate surfaces (arrow 1
in Figure 15) formed near the step. Simultaneously, slight
bending of the crystal planes parallel to the substrate surface,
indicated by arrow 2 in Figure 15, can also be observed near
the columnar boundaries in the films.
YBCO Film Growth on MOCVD-Derived CeO
2
Buffer
2
CeO thin films and their implementation as buffer layers
Layers. Well-aligned YBCO films were successfully de-
for the growth of high-quality YBCO films.
Inorganic Chemistry, Vol. 41, No. 20, 2002 5015

Edleman et al.
Figure 5. (a) Solid-state crystal structure of 8-coordinate neodymium
t
complex 5, Nd( BuMeMe)3. (b) Immediate coordination environment of
t
8
-coordinate neodymium complex 5, Nd( BuMeMe)3.
Precursor Synthesis, Structures, and Properties. The
one-pot syntheses of the present ether-functionalized lan-
thanide â-ketoiminate complexes are uncomplicated, and
Figure 6. (a) Solid-state crystal structure of 8-coordinate gadolinium
t
complex 6, Gd( BuMeMe)3. (b) Immediate coordination environment of
t
8
-coordinate gadolinium complex 6, Gd( BuMeMe)3.
3 2
isolation by in vacuo removal of the HN(SiMe ) byproduct
followed by recrystallization of the crude material from
pentane is straightforward. The waxy solids obtained (Figure
Single-crystal X-ray diffraction experiments reveal that all
lanthanide complexes formed with the present â-ketoiminate
ligands are monomeric, a characteristic preferred for optimum
volatility (Figures 4-17). In fact, the three potentially
tridentate ligands exceed the coordinative demands of the
2) are characteristically highly colored, display melting points
below those reported for commonly used â-diketonate
1
02
precursors, and can be sublimed at suitably low temper-
3
+
3+
3+
3+
3+
Ln ion in all cases studied (Ln ) Ce , Nd , Gd ,
-
4
atures (80-110 °C) under reduced pressure (10 Torr).
Additionally, substitution of various alkyl moieties on the
â-ketoiminate ligand backbone allows tailoring of the â-ke-
toiminate melting point, as summarized in Table 4. The
3
+
Er ), resulting in 8-coordinate complexes for Ce, Nd, and
Gd (4, 5, 6, and 9; Figures 4 and 5, 6 and 7, 8 and 9, and 10
and 11, respectively), compared to 7-coordinate Er com-
pounds (7, 10, and 12; Figures 12 and 13, 14 and 15, and
observed melting point decreases across the lanthanide series
1
6 and 17). This difference is doubtless a result of the
t
for a given ligand system [i.e., for the Ln( BuMeMe)
3
3
+
lanthanide contraction, with the smaller Er ion being
coordinatively saturated at a lower coordination number. All
the 8-coordinate structures display two ligands acting in
tridentate fashion, with the third being bidentate with a
noncoordinating ether group, consistent with the ether
functionality being the only neutrally charged donor site.
series: mp 4, 95-98 °C; 5, 78-81 °C; 6, 76-78 °C; 7,
5-68 °C; see Figure 3a], a manifestation of the lanthanide
contraction. This melting point trend is also observed for
6
102
the lanthanide â-diketonates, Ln(dpm)
alkyl group with greater steric bulk at the imino location R
e.g., ethyl or tert-butyl for methyl) leads to a higher melting
point (i.e., mp 5, 78-81 °C; 9, 85-88 °C; 11, 125-128
3
.
Substitution of an
2
(
Further of note is the anomalous solid-state structure of
Er complex 12 (Figure 16), in which two of the three ligands
act as tridentate donors, while the third is only monodentate,
having both noncoordinating imino and ether components.
This is in contrast to the other Er structures (7 and 10; Figures
°
R
R
1
C; see Figure 3b). It can be speculated that a greater R to
t
t
2
size disparity (i.e., R
1
) Bu, R
2
) Me versus R ) Bu,
1
t
2
) Et or Bu) lowers the symmetry of the complex,
affording a lower-melting precursor.
1
2 and 13, and 14 and 15, respectively), which display one
tridentate ligand and two bidentate ligands with dangling
ether moieties, a bonding scheme similar to that observed
(
102) Eisentraut, K. J.; Sievers, R. E. J. Am. Chem. Soc. 1965, 87, 5254-
5256.
5016 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
Figure 8. (a) Solid-state crystal structure of 7-coordinate erbium complex
t
7
, Er( BuMeMe)3. (b) Immediate coordination environment of 7-coordinate
t
erbium complex 7, Er( BuMeMe)3.
The average bond distances observed in Ce structure 4
Figure 7. (a) Solid-state crystal structure of 8-coordinate neodymium
(Table 9) can be compared to those reported for the
t
complex 9, Nd( BuEtMe)3. (b) Immediate coordination environment of
8
-coordinate cerium of the dimer Ce
2
(etbd)
6
(tetraglyme)
t
8
-coordinate neodymium complex 9, Nd( BuEtMe)3.
6
4
(IX). The average Ce-O(keto) bond length of 4 [2.371(6)
3+
3+
3+
for the 8-coordinate Ce , Nd , and Gd complexes (vide
supra). The increased steric bulk of a tert-butyl group at the
ligand imino site R apparently restricts the “bite” of the
2
charged â-ketoimine portion of this ligand. Close examina-
tion of the bond distances for complex 12 (Table 8, Figure
1
9) indicates that the intraligand metrical parameter trends
3+
are consistent with Er coordination, since the keto oxygen-
to-carbon bond length is longer, and the imino nitrogen-to-
carbon bond length is shorter for the monodentate ligand
than for the other two (tridentate) ligands.
Important bond distances and angles for all structures
analyzed in this study (complexes 4-7, 9, 10, and 12) are
compiled in Tables 5-8. Average bond distance values and
bond distance trends based on lanthanide identity are
compiled in Table 9. In all cases (Ln ) Ce, Nd, Gd, Er), the
Ln-O(keto) bond length is shorter than the Ln-O(ether)
distance, consistent with the fact that the keto oxygen atom
is a component of the anionic â-ketoiminato ligand fragment
while the ether oxygen atom is a neutrally charged pendant
donor. Additionally, the trend across the lanthanide series is
for all Ln-O and Ln-N bond distances to decrease as the
Å] is slightly shorter than that of IX [2.440(4) Å], while the
average Ce-O(ether) bond length is longer [2.760(3) vs
2.605(0) Å]. This result is not surprising, in that fluorinated
ligands are known to induce tighter binding of polyether
oxygens, owing to greater induced electrophilicity at the
3+
Ln ion becomes heavier, an obvious manifestation of the
lanthanide contraction. Variation in ligand substitution pattern
effects only minor deviations in bond distances for a
3
+
particular Ln ion.
Inorganic Chemistry, Vol. 41, No. 20, 2002 5017

Edleman et al.
Figure 10. (a) Solid-state crystal structure of 7-coordinate erbium complex
t
t
1
2, Er( Bu BuMe)3. (b) Immediate coordination environment of 7-coordinate
t
t
Figure 9. (a) Solid-state crystal structure of 7-coordinate erbium complex
erbium complex 12, Er( Bu BuMe)3.
t
1
0, Er( BuEtMe)3. (b) Immediate coordination environment of 7-coordinate
t
erbium complex 10, Er( BuEtMe)3.
ether moiety to the â-ketoiminate backbone renders this
ligand more compact, resulting in the shorter average bond
lengths observed for 6.
The overall average Er-O(keto) and Er-O(ether) bond
distances (averaged values for complexes 7, 10, and 12; Table
metal center. Complex 4 can also be compared to the reported
structure of 8-coordinate Ce(acac) (phenanthroline) (X).
3
Here the average Ce-O(keto) bond length is 2.464(2) Å,
longer than that of 4 [2.371(6) Å], and the average Ce-N
bond length of 2.741(5) Å is comparable to that of complex
1
03
9
) are 2.191(5) and 2.461(2) Å, respectively. The average
Er-N bond distance for the present three complexes (7, 10,
and 12) of 2.474(2) Å can be compared to those in structure
4
[2.760(3) Å].
The average Gd-O bond lengths in Gd complex 6 [Gd-
53
2
XII. Here the η -pyrazolato nitrogen atoms bind at a shorter
average Er-N distance of 2.296(4) Å.
O(keto), 2.264(6) Å; Gd-O(ether), 2.543(9) Å, Table 9] can
be compared to those in complex XI. This Gd(dpm)
3
-
From the foregoing discussions, it is not surprising that
introduction of alkyl substituents with greater steric bulk at
the ether R ligand site lowers the precursor melting point
3
dramatically, judging from the fact that at least one of three
possible ether components is not coordinated to the metal
center in all cases characterized by X-ray diffraction, and
from the symmetry-breaking arguments advanced above.
t
Complexes of the type Ln( BuEtEt)
3
and Ln(MeMeEt)
3
(substitution of ethyl for methyl at R
3
) were also synthesized
in this study; however, attempts to purify the resulting
viscous liquids were unsuccessful. Qualitatively, these crude
complexes are lower-melting than the crude products ob-
tained for complexes 4-12 as discussed above. Note that a
precursor which exists in the liquid state at room temperature
is undesirable because of difficulties in manipulation and
transfer. Complexes 4-12 display optimum melting points,
affording room-temperature solids that can easily be weighed
and transferred to MOCVD precursor reservoirs, while also
6
2
(
monoglyme) compound has a longer average Gd-O(keto)
bond length [2.333(4) Å] and a bit longer average Gd-
O(ether) bond length [2.575(2) Å]. Clearly, the â-ketoiminate
ligand of complex 6 is more tightly bound through the
oxygen atoms to Gd3 than in the â-diketonate/glyme ligand
combination of complex XI. Perhaps directly appending the
+
(
103) Christidis, P. C.; Tossidis, I. A.; Paschalidis, D. G.; Tzavellas, L. C.
Acta Crystallogr. 1998, C54, 1233-1236.
5018 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
Figure 12. (a) Weight loss TGA volatility comparison between the new
t
Figure 11. (a) Weight loss TGA volatility comparison between the new
precursors Ce( BuMeMe)3 (4) and Ce( BuEtMe)3 (8), and commercially
available Ce(dpm)4. Data were collected at a ramp rate of 1.5 °C/min under
precursor Er( BuEtMe) (10) and commercially available Er(dpm) . Data
3
3
t
t
were collected at a ramp rate of 1.5 °C/min under 5.0 (( 0.1) Torr of N2.
(b) Thermal activation TGA plot showing direct volatility comparison
t
5
.0 (( 0.1) Torr of N2. (b) Thermal activation TGA plot showing direct
between the new precursors Er( BuEtMe) (10) and commercially available
3
t
volatility comparison between the new precursors Ce( BuMeMe)3 (4) and
Ce( BuEtMe)3 (8), and commercially available Ce(dpm)4. Data were
Er(dpm) . Data were collected at a ramp rate of 1.5 °C/min under 5.0 ((
3
t
0.1) Torr of N . (c) Isothermal TGA data showing volatility comparison
2
t
collected at a ramp rate of 1.5 °C/min under 5.0 (( 0.1) Torr of N2. Arrow
points to experimental MOCVD reservoir temperature used for precursor
between the new precursor Er( BuEtMe) (10) and commercially available
3
Er(dpm) . Data were collected at 155 °C under 5.0 (( 0.1) Torr of N .
3
2
4
(
. (c) Isothermal TGA data comparing volatility of the new precursors Ce-
BuMeMe)3 (4) and Ce( BuEtMe)3 (8), and commercially available Ce-
t
t
Thermogravimetric analysis (TGA) of the new â-ketoim-
inate precursors gives insight into volatility characteristics,
since these analyses are carried out at a typical experimental
MOCVD reactor pressure [5.0 (( 0.1) Torr]. Ce complexes
(dpm)4. Data were collected at 155 °C under 5.0 (( 0.1) Torr of N2.
melting at low temperatures to allow liquid-state operation
and constant delivery into the reactor during film growth.
Inorganic Chemistry, Vol. 41, No. 20, 2002 5019

Edleman et al.
2
Figure 14. AFM images over a 2 × 2 µm area of CeO2 films grown at
(
a) 450 °C, with an rms roughness of ∼15.5 Å, and (b) 540 °C, with an
rms roughness of ∼4.3 Å.
Figure 13. XRD of a 540 °C MOCVD-derived CeO2 film grown on (001)
YSZ: (a) θ-2θ scan showing only (00l) reflections of both the film and
substrate; (b) ω-rocking curve of the (002) reflection showing the film to
have a high degree of out-of-plane epitaxy; (c) φ scan of the {113} peak
revealing that the film has a high degree of in-plane microstructural
alignment.
4
and 8 and Er complex 10 were directly compared with
commercially available Ce(dpm) and Er(dpm) . Figure 11a
4
3
presents the weight-loss data for the Ce complexes, and
although 4 and 8 appear to sublime with ∼8% residue, we
believe that this is a result of unavoidable exposure to
ambient atmosphere during TGA handling. The thermal
activation plot of sublimation rate is more directly related
to MOCVD reactor behavior, and these data show that 4
and 8 clearly have substantially higher vaporization rates
under typical film growth conditions and at lower temper-
atures than does Ce(dpm)
optimum film growth (Figure 11b). The isothermal data
Figure 11c) reveal that all three Ce complexes show some
4
, a characteristic preferred for
Figure 15. HREM image of the CeO2/YSZ interface. Solid arrows mark
the bent planes; open arrow points to the step on the substrate surface; “I”
is for interface.
(
degree of depressed volatility over time.
TGA weight-loss data for Er complex 10 and Er(dpm)
show that both complexes have favorable volatility charac-
teristics, with the new complex subliming with a small
amount of residue (as in the Ce case), most likely due to
ambient-air exposure (Figure 12a). The thermal activation
plot (Figure 12b) reveals that both Er complexes provide
stable transport characteristics. Most significantly, the iso-
thermal data show that 10 exhibits a stable evaporation rate,
3
3
while Er(dpm) sublimes more quickly initially, but then
5020 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
Figure 16. Composite electron diffraction pattern of CeO2/YSZ.
Figure 18. (a) Tc (with a transition width of ∼2.0 K) and (b) Jc mea-
surements on an YBCO/CeO2/YSZ(001) multilayer.
Figure 19. Schematic of the monodentate ligand of complex 12. Arrows
and the table express the comparative shortening and lengthening of the
monodentate ligand bonds as compared to the two tridentate ligands of
complex 12.
orientation and processing requirements. Thus, the present
CeO films grown in situ at 450-650 °C by MOCVD were
2
subjected to a battery of microstructural characterization
techniques to assess crystalline quality. The XRD θ-2θ
diffraction pattern (Figure 13a) shows that the films consist
of only (00l) oriented components on single-crystal (001)
YSZ substrates. Such orientation is critical for subsequent
deposition of a highly conductive (001)-oriented supercon-
ducting layer. A decrease in the XRD peak intensity is
observed as the growth temperature is increased. The full-
width-at-half-maximum (fwhm) of the ω scan assays the
alignment/perfection of the film growth plane with respect
to the substrate surface. In this case, the ω-rocking curve
fwhm of the (002) reflection decreases from 1.0° to 0.2° with
Figure 17. XRD of a YBCO/CeO2/YSZ(001) multilayer structure: (a)
θ-2θ scan pattern; inset displays the ω-rocking curve of the (005) YBCO
reflection, revealing good out-of-plane epitaxy; (b) φ scan of the {102}
YBCO reflection showing excellent in-plane alignment.
displays a sudden fall in transport rate (Figure 12c). The
superior MOCVD precursor is certainly one that provides a
constant transport rate, as this implies stable delivery of the
metal-containing species into the reactor hot-zone, hence
compositionally reproducible thin film growth.
CeO
2
Film Growth. Buffer layers suitable for supercon-
ducting current-carrying applications must meet stringent
Inorganic Chemistry, Vol. 41, No. 20, 2002 5021

Edleman et al.
increasing growth temperature. These values are comparable
to those of films grown at considerably higher tempera-
crystal YSZ substrates, in contrast, are not atomically smooth,
and although CeO grows with an abrupt interface on the
2
step planes, crystal plane bending parallel to the surface is
observed at the step edges (arrow 1 in Figure 15). Owing to
the slight lattice mismatch, stress may build up when the
4
6,47,93
tures.
The higher-temperature film growth ω scan
actually consists of a superposition of a narrow peak on top
of a broader one, and the narrow component dominates as
4
6,93
the temperature increases, which suggests thinner films.
2
CeO film atomic planes on the lower part of the step attempt
This observation, together with the deposition temperature
dependence of the XRD intensity, clearly indicates lower
growth rates at higher temperatures. Although the 0.85° ω
to match the neighboring substrate planes, resulting in the
observed bending. Crystal plane bending parallel to the
surface is also observed at the columnar boundaries of the
scan fwhm from the CeO
2
films deposited at 540 °C (Figure
film (arrow 2 in Figure 15), in a fashion similar to HREM
3b) is slightly larger than those previously reported,46,47,93
observations on CeO
films grown on SrTiO
46
Such types
1
2
3
.
this result is impressive considering the low deposition
temperature.
of dislocations could alter the XRD ω scan fwhms. When
the CeO films are thin, perfectly aligned CeO film portions
2
2
XRD φ scans demonstrate that the present CeO
a high degree of in-plane epitaxy with the underlying YSZ
Figure 13c). The fwhm from the φ scan of {113} CeO
which greatly affects YBCO in-plane alignment and thus the
2
films have
on the step planes contribute mainly to the narrow portion
of the XRD ω-rocking curve, while the broad underlying
portion may arise principally from the crystal plane bending
(
2
,
of CeO
thickening of the CeO
2
film portions above the substrate steps. With
films, the existences of atomic plane
104
J
c
,
is 1.65°. Thus, from both the θ-2θ and φ scans, the
orientation of the CeO film on the YSZ substrate has the
relationship CeO (001)[ 1h 10]//YSZ (001)[ 1h 10].
AFM studies of the present CeO films give more detail
2
bending near the columnar boundaries would broaden the
XRD ω scan curves, and they would appear as a conventional
mosaic, but broader than those of the underlying substrates.
YBCO Film Growth. YBCO films were successfully
2
2
2
on surface morphology and roughness. Upon increasing the
deposition temperature, the grain diameter remains ap-
proximately the same (50-110 nm); however, the root-mean-
square (rms) film roughness decreases. At low growth
temperatures (450 °C; Figure 14a), the rms roughness is
deposited on the 540 °C MOCVD-derived CeO
2
using a
pulsed MOCVD technique, POMBE. XRD θ-2θ (Figure
7a) and ω scan (inset, Figure 17a) data show that the YBCO
101
1
films are (001)-oriented with a high degree of out-of-plane
crystallinity. This compares rather well with data for
POMBE-derived YBCO (700 °C growth) and MOCVD-
∼
15.5 Å. Close observation of the grain particles indicates
that the grains are composed of tetragonal-pyramidal shaped
1
05
93
derived YSZ (850 °C) bilayer films. A high level of in-
plane epitaxy is also revealed by φ scan XRD data (Figure
hillocks, which are thought to consist of four (111) planes.
These hillock formations are not observed in the 540 °C
sample (Figure 14b), and the surface of the films is much
smoother (rms ∼4.3 Å), most likely owing to the slower
growth rates at higher temperatures.
1
7b). The multilayer therefore has the following heteroepi-
taxial relationship: (YBCO(001)[010]//CeO (001)[ 1h 10]//
YSZ(001)[ 1h 10]), with the YBCO rotated 45° in the basal
plane of the CeO underlayer. The narrow 1.30° fwhm of
the {102} φ scan should correlate with high critical current
2
2
2
Cross-sectional HREM (Figure 15) shows that a CeO film
grown at 540 °C exhibits a high content of single-crystal
regions with low defect densities. HREM images and
selected-area electron diffraction (Figure 16) confirm the
1
04
density (J
c
). Indeed, four-probe transport measurements
) of 86.5
of 1.08 × 10 A/cm (Figure 18, plots a and b,
respectively). Considering the low deposition temperature
of the CeO , which simplifies processing and may be critical
reveal a superconducting transition temperature (T
c
6
2
K and a J
c
high-quality epitaxial relationship between the CeO
YSZ, as also determined using XRD. The single-crystal-like
CeO electron diffraction pattern is characteristic of a high-
quality epitaxial film. Owing to the slight lattice mismatch
∼5%), the interface area between the film and substrate
contains atomic stacking dislocations having a periodicity
2
and
2
2
for multilayer (superconductor-insulator-superconductor)
device fabrication, this result represents a significant im-
provement.
(
3
6,47,94
of ∼44 Å [(020) layers].
Consistent with the AFM grain
Conclusions
size observations, the columnar size remains unchanged;
however, the smoothness increases as the deposition tem-
perature is increased. This increased smoothness is doubtless
a result of the slower growth rate at higher temperatures,
which was determined to decrease from ∼10 Å/min at 450
A new series of fluorine-free lanthanide MOCVD precur-
sors has been prepared which exhibit good volatility and
lower melting points versus commercial and other previously
reported lanthanide sources. The â-ketoiminate ligand is a
versatile, multidentate system, producing coordinatively
saturated, monomeric lanthanide complexes which are proven
here to be excellent MOCVD precursors. Substitution of
various alkyl moieties at the keto, imino, and ether sites of
the â-ketoiminate ligand offers significant tuning of the
melting point and volatility characteristics, facilitating the
design of customized lanthanide precursors.
°
C to ∼6.5 Å/min at 540 °C.
Further analysis of the HREM images (Figure 15) reveals
2
that the CeO films are densely packed with no cracks evident
either at the interface with the substrate or within the film,
consistent with the narrow fwhms of the XRD φ and ω scans,
and in spite of the columnar microstructure. The single-
(
105) Kaatz, F. H.; Markworth, P. R.; Dai, J. Y.; Buchholz, D. B.; Liu,
X.; Chudzik, M. P.; Belot, J. A.; Kannewurf, C. R.; Marks, T. J.;
Chang, R. P. H. Chem. Vap. Deposition 1998, 4, 99-102.
(
104) Dimos, D.; Chaaudhari, P.; Mannhart, J. Phys. ReV. B 1990, 41,
4038-4049.
5022 Inorganic Chemistry, Vol. 41, No. 20, 2002

â-Ketoiminate Lanthanide MOCVD Precursors
CeO
2
buffer layer films have been epitaxially deposited
2
attractive for low-temperature growth of CeO buffer layers
on (001) YSZ in the temperature range 450-650 °C by
MOCVD using the newly developed Ce( BuMeMe) precur-
3
sor. With an increase in deposition temperature, the film
growth rate declines while the films tend to be smoother.
for coated conductors.
t
Acknowledgment. We thank the NSF for support of this
research through Grant CHE-0076097 and the MRSEC
(Materials Research and Engineering Center) Program
through the Northwestern Materials Research Center (Grant
DMR-0076097).
Supporting Information Available: An X-ray crystallographic
file in CIF format containing information for complexes 4-7, 9,
10, and 12. This material is available free of charge via the Internet
at http://pubs.acs.org.
High-quality epitaxial CeO
atures as low as 540 °C with smooth surfaces (rms roughness
4.3 Å) and good in-plane (∆φ ) 1.65°) and out-of plane
∆ω ) 0.85°) alignment. YBCO films grown on these CeO
buffer layers by POMBE are epitaxial and exhibit very good
2
films can be grown at temper-
∼
(
2
6
2
transport properties (T
7
c
) 86.5 K, J
c
) 1.08 × 10 A/cm at
7.4 K). The new precursor has thus proven stable and is
IC020299H
Inorganic Chemistry, Vol. 41, No. 20, 2002 5023