Synthesis of sub-micron silver and silver sulfide particles via solvothermal silver azide decomposition

Article abstract of DOI:10.1016/S0025-5408(02)01028-0

Many transition-metal azides are thermodynamically unstable with respect to the elements and thus, may serve as energetic precursor sources in nanoscale metal particle synthesis. This report describes the use of silver azide (AgN₃) in nonaqueou

Full text of DOI:10.1016/S0025-5408(02)01028-0

Materials Research Bulletin 38 (2003) 213±220
Synthesis of sub-micron silver and silver sul®de particles
via solvothermal silver azide decomposition
Luke Grocholl, Jianjun Wang, Edward G. Gillan*
Department of Chemistry, Optical Science and Technology Center, University of Iowa, Iowa City, IA 52242-1294, USA
Received 17 July 2002; received in revised form 17 October 2002; accepted 7 November 2002
Abstract
Many transition-metal azides are thermodynamically unstable with respect to the elements and thus, may serve
as energetic precursor sources in nanoscale metal particle synthesis. This report describes the use of silver azide
(AgN3) in nonaqueous, solvothermal decomposition reactions to produce crystalline sub-micron silver particles
and interconnected structures. The thermal decomposition of AgN3 directly produces silver and N2 and no
secondary chemical reducing agent is required. This solvothermal conversion was examined in toluene,
tetrahydrofuran (THF), and trioctylamine below 250 8C. The coordinating solvents produced the smallest
particles (150±500 nm), while the toluene reaction products were near 1 mm in size. The addition of soluble
elemental sulfur to the THF reaction results in the growth of silver sul®de particles near 1 mm in size. The silver
and Ag2S products are crystalline by X-ray diffraction and show some faceting by scanning electron microscopy.
# 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Metals; Chalcogenides; Nanostructures; Chemical synthesis; Electron microscopy
1. Introduction
Nanoscale metal particles have been shown to exhibit unusual ductility [1±3] and undergo optically
and chemically induced shape changes [4,5]. Metal particle syntheses generally involve chemical
reduction of a metal salt with an external reductant, such as hydrazines, alkali borohydrides, and citrate
salts. Recent chemical methods have been reported for metal particles, such as Ni [6], Cu [7], Ag [8],
Au [9], and metal alloys [10,11]. The chemical reduction process is in¯uenced, to some extent, by
solution diffusion and may result in reactions that are hard to regulate, that produce undesirable
molecular byproducts, and that incorporate other species, such as counter-ions, into the metal structure.
*
Corresponding author. Tel.: ‡1-319-335-1308; fax: ‡1-319-335-1270.
E-mail address: edward-gillan@uiowa.edu (E.G. Gillan).
0025-5408/02/$ ± see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0025-5408(02)01028-0

214
L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
Metal azides are a well-studied class of energetic compounds [12,13], and several have utility in
organic syntheses (e.g. NaN3) [14] and as primary explosives (e.g. Pb(N3)2) [15]. Metal azides are often
thermodynamically unstable with respect to disproportionation and can violently decompose on
friction, pressure, or heating [16]. Recently molecular azides, such as Ga(N3)3 [17] and H2GaN3 [18]
have been utilized as precursors in low-temperature GaN powder and ®lm growth, respectively. We
recently showed that solvothermal environments moderate gallium azide decomposition during the
synthesis of GaN nanoparticles [19]. In the report below, we describe the straightforward production of
sub-micron silver and silver sul®de particles via the autoreductive decomposition of silver azide
(AgN3) in superheated organic solvents.
2. Experimental
2.1. Silver and silver sul®de synthesis
The AgN3 precursor was produced by the stoichiometric reaction of silver nitrate (AgNO3, Aldrich,
99‡%) with sodium azide (Aldrich, 99%). A white, microcrystalline AgN3 precipitate (Ksp ꢀ 2:8 Â
10À9 M2) forms immediately upon mixing 0.3 M aqueous solutions of these two reagents. The product
was washed with water, then diethyl ether, and dried in air. Crystalline silver azide was veri®ed by
powder X-ray diffraction (JCPDS #03-0906) and IR spectroscopy, and its thermal properties were
consistent with previously reports [12].
All solvents used in the solvothermal experiments were dried, distilled, and degassed prior to use. The
toluene (Fisher Scienti®c, ACS grade) and tetrahydrofuran (THF, C4H8O, Fisher Scienti®c, certi®ed)
reactions were performed in a high-temperature, high-pressure stainless steel reactor (125 ml Parr
Instruments model 4752) equipped with a silica- or Te¯on-coated stir bar. The AgN3 powder (1.4 g,
9.3 mmol) was loaded into the reactor that was ®lled to approximately two-thirds capacity (85 ml) with
the solvent, pressurized to 1 atm with N2, and heated by a temperature-controlled heating jacket (custom
Glas-Col mantle, Omega CN4800 temperature controller) under constant stirring. The reactor temp-
erature was raised to 130 8C (ꢀ10±20 8C/h) and held overnight, then the raised to 200 8C and held for 1
day. The onset of azide decomposition near 175 8C was noted by a steady increase in gas pressure inside
of the reactor, which was monitored with an attached pressure gauge. The maximum reaction temperature
was determined as the point where noticeable pressure evolution ceased. The ®nely dispersed powder
products were ®ltered from the solution, washed with fresh solvent, and dried under vacuum.
The ambient pressure decomposition of AgN3 (3.9 g, 26 mmol) was examined using 100 ml of
trioctylamine (Oct3N, [CH3(CH2)7]3N, Aldrich, 98%) in a stirred Schlenk ¯ask under a nitrogen purge.
This reaction was heated to 170 8C for 1 day, cooled for product/intermediate analysis, then reheated at
220 8C for 1 day, and ®nally raised to 250 8C and held overnight. The solid product was separated from
the solution, washed with triethylamine (Et3N) to remove any high boiling Oct3N residue, and dried
under vacuum.
Silver sul®de synthesis was run in the same manner as the THF reaction described above. In this case,
AgN3 (2.93 g, 19.6 mmol) was mixed with sulfur powder (EM Science, sublimed, 0.643 g, 20.1 mmol)
and loaded into the Parr reactor, followed by the addition of THF (85 ml). This reaction was heated to
150 8C (ꢀ20 8C/h) and held overnight, followed by a slow temperature increase to 190 8C, which was
maintained overnight to ensure reaction completion, as judged by the end of gas evolution. Note that

L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
215
initial gas evolution is observed below 100 8C in the silver sul®de reaction. The black powder product
was separated from the solvent, washed with clean solvent, and dried under vacuum. Safety note: Silver
azide and other metal azides are often thermally unstable and shock sensitive [12,20]. These
compounds should be treated as potentially explosive materials. Care should be taken whenever
working with reactions that may produce azides as products or intermediates. Reactions should be
performed on a small scale in high-pressure vessels with safety release valves. Proper protection should
be used when working with undecomposed azide intermediates.
2.2. Precursor and product characterization
The phase and crystallinity of the products were analyzed by powder X-ray diffraction using a
Siemens D5000 diffractometer. A crystalline silicon standard and 0.058 increments were used for
crystallite size measurements, which were calculated from peak line broadening (Ka2 removed) using
the Scherrer and Warren equations. Infrared spectra were taken on a Nicolet Nexus 670 spectrometer in
transmission mode using KBr pellets or a nitrogen-purged 10 cm gas cell. UV-Vis spectra were
obtained from powders suspended in ethanol solutions or coated on a silica glass slide using a Varian
Cary 100 spectrophotometer equipped with an integrating sphere accessory. Thermogravimetric-
differential thermal analysis (TG-DTA) was run on a Seiko Exstar 6300 system under ¯owing argon
with a 5 8C/min heating rate. Scanning electron microscopy was performed with a Hitachi S-4000 ®eld
emission system on powders adhered to aluminum holders. This system includes a Kevex or IXRF
elemental X-ray detector for energy dispersive spectroscopy.
3. Results and discussion
3.1. Silver particle synthesis
Polycrystalline silver azide (AgN3) is thermodynamically unstable (DHf ˆ ‡311 kJ/mol), begins to
degrade by 150 8C, and reportedly melts near 250 8C with nitrogen gas evolution; however, rapid
heating above 270 8C causes a violent and exothermic decomposition [12,16]. Other studies have
shown that AgN3 single crystals show evidence of silver particle formation after heating to 115 8C [21],
and AgN3 powder completely decomposes after 5 h at 226 8C [22]. The proposed overall
decomposition pathway is straightforward, as shown in Eq. (1). Though there is the possibility for
formation of an Ag3N intermediate, it is very unstable and will easily decompose to the elements.
AgN3ꢁs† ! Agꢁs† ‡ 32 N2ꢁg†
(1)
In the present study, a heated solvothermal environment moderates silver azide decomposition and
prevents violent de¯agration events from occurring. The AgN3 starting material consists of roughly 2±
4 mm particles as determined by optical microscopy and its decomposition was examined in
superheated toluene and THF up to 200 8C. Observable gas evolution was apparent by 190 8C in both
reactions. The total amount of evolved gas was roughly that expected for N2 based on Eq. (1), assuming
the full reactor volume and ideal gas behavior. The IR spectra of evolved gases in these two reactions
showed mainly solvent peaks. The toluene reaction also shows clear evidence of HN3 byproduct peaks
at 3330, 2130, and 1150 cmÀ1, but this gaseous azide is nearly absent from the THF IR spectrum. The

216
L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
Fig. 1. Powder X-ray diffraction of isolated products from the solvothermal decomposition of AgN3: (a) in toluene at
200 8CÐsilver peaks are marked with (h k l) indices, (b) in THF at 200 8C, (c) in Oct3N at 170 8CÐsilver peaks are noted by
an asterisk and all other peaks are AgN3 (JCPDS #03-0906), (d) in Oct3N at 250 8C, and (e) in THF at 190 8C with elemental
sulfurÐall peaks correspond to Ag2S (acanthite, JCPDS #14-0072).
silver products are homogeneous gray to light brown free-¯owing powders (>85% isolated yield based
on Ag). From powder X-ray diffraction (XRD) analysis, the products consist of crystalline cubic silver
Ê
(fcc, a ˆ 4:086 A) (Fig. 1a and b). Solid-state IR spectra of the silver powders show little or no
evidence for solvent surface coatings, as has been observed in other solvothermal azide reactions [19].
The toluene product shows some azide residue (2036 cmÀ1) that could be the result of HN3 reaction
with the silver surface.
The ambient pressure AgN3 decomposition was examined in trioctylamine, Oct3N (bp ˆ 366 8C).
The rapidly stirred silver azide suspension turns light gray by 130 8C and evolves gas bubbles
beginning at 170 8C. By XRD, the product at this point is a mixture of AgN3 and silver metal (Fig. 1c).
Additional gas evolution is observed upon heating to 250 8C and elemental silver is the only crystalline
solid remaining (Fig. 1d). The product's visible morphology consists of small porous, nearly spherical
1±5 mm re¯ective silver ``wif¯e-ball'' structures interspersed with a yellow±brown material. This
product shows clear, but broad IR signatures consistent with Oct3N, which suggests that the brown solid
contains some polymerized solvent component. TG-DTA (argon ¯ow) on the Oct3N silver product
shows a 5% weight loss between 150 and 200 8C that likely corresponds to loss of organic residues.
Scanning electron microscopy (SEM) shows that silver particles from the toluene reaction consists of
large irregular, faceted structures near 1 mm in size, similar to that of the starting AgN3 powder (Fig. 2A).
The silver particles formed in coordinating solvents (THF and Oct3N) have smaller crystallite and
particle sizes and smoother morphologies than those formed in the poorly coordinating toluene solvent

L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
217
Fig. 2. Scanning electron microscopy images of silver particles from AgN3 decomposition in (A) toluene at 200 8C and (B)
THF at 200 8C.
(Table 1). SEM shows that silver particles from THF are near 200 nm in diameter with smooth edges and
occasionally form into elongated rods around 750 nm in length (Fig. 2B). The silver product from Oct3N
consists of well-formed particles near 400 nm in diameter (Fig. 3A) with a signi®cant degree of
aggregation apparent at lower magni®cation (Fig. 3B), likely leading to the macroscopic millimeter-
sized structures noted above. After TG-DTA heating to 500 8C, the Oct3N product becomes a fused
silver mass that SEM shows is an interconnected network of 1 mm diameter elongated structures
(Fig. 3C). Note that this particle coalescence occurs well below the bulk silver melting point (962 8C).
The variation in XRD crystallite sizes generally track with SEM particle sizes, but they are
signi®cantly smaller (Table 1). Due to the relatively low synthesis temperatures (ꢂ250 8C), the
observed silver particles are likely comprised of smaller crystalline domains. In the solid state, the sub-
micron silver particles produced by all above methods are large enough to show both bulk silver UV-
Vis plasmon absorption near 310 nm and surface plasmon absorption near 400 nm (Table 1) [23]. The
latter peak is dominant in nanometer-sized silver colloids and slightly red shifts when Lewis bases
coordinate to the particle surface [24], consistent with the shifts seen by silver particles produced from
THF and Oct3N. Particle shape (spherical versus prismatic) has also been shown to cause shifts in the
surface plasmon peak [4].
While AgN3 has very limited solubility in most solvents, strongly polar systems have higher
solubilities than nonpolar ones. For example, its solubility is 2:5 Â 10À6 M in methanol, 5 Â 10À5 M in
water, and 1:4 Â 10À4 M in formamide (HCONH2) [25]. The THF and Oct3N solvents may improve
dispersion of the suspended AgN3 particles by breaking up aggregates in the precipitated starting
Table 1
Solvothermal silver azide decomposition reactions and products
Reagent(s)
Solvent
Maximum
temperature (8C)
XRD crystallite
size (nm)
SEM particle
size (nm)
UV-Vis absorption
(nm)
AgN3
AgN3
Toluene
THF
200
200
250
190
101
60
300±1100
150±250
300±600
450±1100
310, 385
310, 400
310, 395
300
AgN3
AgN3 ‡ S
Oct3N
THF
61
130

218
L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
Fig. 3. Scanning electron microscopy images of silver particles from AgN3 decomposition in (A) Oct3N at 250 8C, (B) higher
magni®cation image of (A), and (C) the fused particle product after TGA heating to 500 8C under argon ¯ow.
material, which could lead to smaller silver particles formed during solvothermal decomposition. We are
currently examining this silver particle growth process in strongly coordinating nonaqueous solvents.
3.2. Silver sul®de synthesis
Silver sul®de, Ag2S, has previously been synthesized by solution means, often using toxic H2S as the
sulfur source. Nonaqueous solution reactions between elemental sulfur and metal salts [26] and
elemental metals [27] have produced crystalline metal sul®des. Ag2S has also been synthesized from
AgNO3 and CS2 in alcohol solutions [28] and from Na2S in anhydrous ammonia [29].
In the present synthesis, soluble, elemental sulfur was added to an AgN3 suspension in THF and
produced a marked change in silver azide decomposition. Signi®cant gas evolution was observed by
100 8C and continued up to nearly 190 8C, consistent with a sulfur-assisted redox process (Eq. (2)).
2AgN3ꢁs† ‡ Sꢁdissolved† ! Ag2Sꢁs† ‡ 3N2ꢁg†
(2)
The gray±black product (95% isolated yield) is a ®nely dispersed powder with XRD re¯ections for
crystalline monoclinic Ag2S (Fig. 1e) and no evidence for any silver metal residue. The excess
dissolved sulfur was removed during solution ®ltration and washing. SEM images show that the Ag2S
particles are near 800 nm in size, much larger than the silver particles formed under the same

L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
219
Fig. 4. Scanning electron microscopy images of Ag2S particles from the reaction of AgN3 with S in THF at 190 8C.
conditions (Fig. 4A and Table 1). There is also evidence for larger organized structures nearly 10 mm in
size (Fig. 4B). Semiquantitative EDS data on the small and large structures are consistent with an Ag2S
composition, within the accuracy of this measurement.1
Since gas evolution occurs at much lower temperature (100 8C) in Ag2S versus Ag metal synthesis,
the dissolved sulfur likely attacks the suspended AgN3 particles before they decompose to silver
particles. To test this possibility, a bench top reaction of AgN3 (0.3 mmol) and sulfur (0.3 mmol) in
THF (25 ml) was brie¯y heated to re¯ux (ꢃ70 8C) with a heat gun. During this test, the sulfur powder
quickly dissolved and the white silver azide powder turned black and appeared to evolve gas. This
suggests that the sulfur containing solution reacts directly with the AgN3 suspended powders. This
conversion may initially involve reaction with low levels of dissolved AgN3 near the particle surface.
The isolated black powder was primarily poorly crystalline Ag2S by XRD. We are currently examining
this reaction in more detail.
4. Conclusions
Silver azide undergoes ef®cient autoreductive solvothermal decomposition to produce crystalline sub-
micron silver particles at low temperatures (ꢂ250 8C). Solvent-mediated thermolysis prevents rapid,
explosive AgN3 decomposition and produces dispersed particles. This silver synthesis requires no
secondary reduction source and is not subject to reagent diffusion limitations. In addition, elemental sulfur
reacts with an AgN3 suspension in THF to produce crystalline Ag2S particles. This azide elimination
method has potential utility in the solvothermal synthesis of other metal particles and metal compounds.
Acknowledgements
The authors acknowledge the University of Iowa and the Of®ce of Naval Research for partial support
for this work and Prof. MacGillivray for the use of his integrating sphere UV-Vis spectrophotometer.
1 EDS semiquantitative analysis found in wt.% (calculated): Ag 90.1 (87.1), S 9.9 (12.9).

220
L. Grocholl et al. / Materials Research Bulletin 38 (2003) 213±220
References
[1] K.H. Hahn, K. Vedula, Scripta Metall. 23 (1989) 7.
[2] T. Haubold, R. Bohn, R. Birringer, H. Gleiter, Mater. Sci. Eng. A 153 (1992) 679.
[3] M.J. Lang, M. Duarte-Dominguez, R. Birringer, R. Hempelmann, H. Natter, W. Arnold, Nanostuct. Mater. 12 (1999) 811.
[4] R. Jin, Y. Cao, C.A. Mirkin, K.L. Kelly, G.C. Schatz, J.G. Zheng, Science 294 (2001) 1901.
[5] P.L. Hansen, J.B. Wagner, S. Helveg, J.R. Rostrup-Nielsen, B.S. Clausen, H. Topsoe, Science 295 (2002) 2053.
[6] D.-H. Chen, S.H. Wu, Chem. Mater. 12 (2000) 1354.
[7] H. Ohde, F. Hunt, C.M. Wai, Chem. Mater. 13 (2001) 4130.
[8] E. Hutter, J.H. Fendler, Chem. Commun. 4 (2002) 378.
[9] R.L. Whetten, M.N. Shafigullin, J.T. Khoury, T.G. Schaaff, I. Vezmar, M.M. Alvarez, A. Wilkinson, Acc. Chem. Res. 32
(1999) 397, and references therein.
[10] J.A. Haber, N.V. Gunda, J.J. Balbach, M.S. Conradi, W.E. Buhro, Chem. Mater. 12 (2000) 973.
[11] C.B. Murray, S. Sun, H. Doyle, T. Betley, Mater. Res. Soc. Bull. 26 (12) (2001) 985.
[12] T.A. Richter, Synthesis and chemical properties, in: H.D. Fair, R.F. Walker (Eds.), Energetic Materials: Physics and
Chemistry of the Inorganic Azides, Plenum Press, New York, 1977 (Chapter 1).
[13] B.L. Evans, A.D. Yoffe, P. Gray, Chem. Rev. 59 (1959) 515.
[14] E.F.V. Scriven (Ed.), Azides and Nitrenes Reactivity and Utility, Academic Press, New York, 1984.
[15] R. Meyer, Explosives, 3rd ed. VCH, New York, 1987.
[16] K.G. Mason, The metallic azides, in: Mellor's Comprehensive Treatise on Inorganic and Theoretical Chemistry, vol. 8
(Suppl. 2, nitrogen part 2), Longmans, Green, and Co. Ltd., London, 1967 (Section 19).
[17] A. Frank, F. Stowasser, H. Sussek, H. Pritzkow, C.R. Miskys, O. Ambacher, M. Giersig, R. Fischer, J. Am. Chem. Soc.
120 (1998) 3512.
[18] J. McMurran, D. Dai, K. Balasubramanian, C. Steffek, J. Kouvetakis, J.L. Hubbard, Inorg. Chem. 37 (1998) 6638.
[19] L. Grocholl, J. Wang, E.G. Gillan, Chem. Mater. 13 (2001) 4290.
[20] J.L. Ennis, E.S. Shanley, J. Chem. Ed. 68 (1991) A6.
[21] A.C. McLaren, Proc. Phys. Soc. (London) B70 (1957) 147.
[22] B.E. Bartlett, F.C. Tompkins, D.A. Young, Proc. Royal Soc. London A 246 (1958) 206.
[23] W. Kaspar, U. Kreibig, Surf. Sci. 69 (1977) 619.
[24] V.V. Vukovic, J.M. Nedeljkovic, Langmuir 9 (1993) 980.
[25] O. Popovych, R.P.T. Tomkins, Nonaqueous Solution Chemistry, Wiley, New York, 1981, p. 125.
[26] Q. Lu, J. Hu, K. Tang, Y. Qian, X. Liu, G. Zhou, J. Solid State Chem. 146 (1999) 484.
[27] V. Dusastre, B. Omar, I.P. Parkin, G.A. Shaw, J. Chem. Soc. Dalton Trans. (1997) 3505.
[28] C. Wang, X. Zhang, X. Qian, W. Wang, Y. Qian, Mater. Res. Bull. 33 (1998) 1083.
[29] G.A. Shaw, I.P. Parkin, Inorg. Chem. 40 (2001) 6940.