Sonochemical synthesis can lead to a dramatic increase in the kinetics of formation of polymer spheres (templates for carbon spheres) compared to the modified Stöber silica method applied to produce analogous polymer spheres. Reactive molecular dynamics simulations of the sonochemical process indicate a significantly enhanced rate of polymer sphere formation starting from resorcinol and formaldehyde precursors. The associated chemical reaction kinetics enhancement due to sonication is postulated to arise from the localized lowering of atomic densities, localized heating, and generation of radicals due to cavitation collapse in aqueous systems. This dramatic increase in reaction rates translates into enhanced nucleation and growth of the polymer spheres. The results are of broad significance to understanding mechanisms of sonication induced synthesis as well as technologies utilizing polymers spheres.

Recent advances in colloidal synthesis techniques have enabled unprecedented control over size, pore structure, surface morphology, and chemical composition of carbonaceous nano spheres (CSs) with far-reaching applications in energy.1–3 Synthesis of polymer spheres (PSs)4,5 (that act as templates for synthesizing carbon spheres) via a modified Stöber method,6 as pioneered by Liu and co-workers,7 has attracted a lot of interest owing to the high porosity, excellent thermal properties, and significant electrical conductivity of the as-prepared CS.8–11 In the traditional Stöber method, tetraethyl orthosilicate (TEOS) is hydrolyzed and condensed using the ammonia catalyst in an alcohol-water solvent to produce silica spheres.6 Liu et al. successfully modified this technique to prepare polymer spheres (PSs) by replacing TEOS with resorcinol (R) and formaldehyde (F).7 Using this modified Stöber method, they could synthesize PSs of diameter 500–900 nm via catalytic hydrolysis and condensation of RF at a high temperature over ∼24 h.7 These PSs then act as templates for producing CSs using a heat-treatment (curing) procedure in a N2 atmosphere at ∼900 °C.7 The reaction conditions, including catalysts, solution concentration, drying method, and temperature of curing, strongly influence the size, shape, and morphology of the synthesized PS.8,9 These methods hold a distinct disadvantage due to their prolonged synthesis time (24 h), as well as stiff challenges involved in controlling the shape and morphology of spheres.8,9

Here, we found that mediating the modified Stöber synthesis method with ultrasonic radiation enhances the kinetics of RF condensation dramatically and yields PSs with narrow size distribution (∼900 nm diameter) within 5 min. Such accelerated reaction rates can be attributed to the extreme localized pressure and temperature (over nm length scales) conditions caused by continuous bubble growth and collapse during sonication.12,13 Although there have been significant advances in continuum scale models to describe bubble dynamics,14,15 these mathematical models cannot accurately capture complex non-linear effects surrounding cavitation. In particular, the coupled effect of sonication-induced localized density variations (due to bubble formation) and elevated temperatures over nanometer length scales on the nucleation and growth of the PS is not well understood. Using reactive molecular dynamics (RMD), we find that in addition to elevated temperatures and extreme cooling rates achieved by sonication induced hot spots, localized reduction in density via cavitation bubbles plays a key role in the enhanced kinetics of PS formation.

We synthesize PSs using the modified Stöber technique7 with solid resorcinol (1 g) and formaldehyde solution (1.4 ml; 37 wt. % in aqueous solution stabilized by 10% methanol) as chemical precursors. The solvent is prepared by homogeneously mixing 100 ml of deionized water (H2O), 40 ml of absolute ethanol, and 0.5 ml of aqueous ammonia catalyst (NH4OH, 25 wt. %); note that the ammonia catalyst is added all at once. The prepared solution of R and F in the water-alcohol solvent is transparent as shown in Fig. 1(a). Using this solution, we perform two sets of synthesis experiments: (a) mediated by ultrasonic radiation (Ultrasonic Generator, model GSCVP 150) and (b) with continuous stirring using a magnetic stir bar without any sonication. For ultrasonication, we employ a piezoelectric ultrasonic 20 kHz probe (Sonics & Materials, Inc.) with 40% intensity. The high energy acoustic waves were applied to 150 ml glass beaker, while the ammonia catalyst (0.5 ml) is added during the sonication. To prevent ammonia vaporization, room temperature (∼27 °C) was mentioned using an ice bath around the sonicating beaker. Upon ultrasonic irradiation, the transparent solution [Fig. 1(a)] turns pale brown within ∼4 min [Figs. 1(b) and 1(c)]. Eventually, after a total 5–6 min of sonication, the RF hydrolysis and condensation reactions proceed to completion and result in a milky white precipitate [Fig. 1(d)]. This solid product is then recovered, washed with water/ethanol, and centrifuged to remove excess ammonia. Finally, the as-obtained carbonaceous solid is vacuum-dried at 50 °C for 4 h. Scanning electron microscopy (SEM; S-4800, Hitachi Co. Ltd.;10 kV) images of the obtained solid indicates formation of smooth, spherical particles with a diameter of ∼900 nm [Figs. 1(e) and 1(f)]; the surface smoothness of the sonochemically derived PS is further confirmed by transmission electron microscopy (TEM; JEOL Model JEM-2100F; 200 kV) images as indicated in Fig. 1(g). The polydispersity of the particle sizes of the sonochemically produced PS is narrow (0.16), which we obtained via dynamic light scattering as well as SEM images. On the other hand, in the absence of sonication, the RF hydrolysis and condensation require much longer reaction times ∼24 h, even upon increasing the reaction temperature to ∼100 °C. In addition, the final carbonaceous particles obtained in the no-sonication case are much larger (∼2 μm) with polydispersity index >0.4 and are mainly aggregates lacking spherical shape [Figs. 1(h) and 1(i)]. This along with the drastic drop in reaction kinetics strongly suggests that the formation of nuclei (that eventually lead to PSs) is severely impaired upon removing ultrasonic mediation.

FIG. 1.

Synthesis of PSs from resorcinol and formaldehyde using a modified Stöber method. (a)–(d) Temporal evolution of RF condensation reaction under ultrasonic irradiation. Electron microscopy images of the obtained carbonaceous particles (e)–(g) with ultrasonic mediation, and (h), (i) in the absence of sonication. Panels (e), (h), (i) are imaged using the scanning microscopy mode, while panels (f), (g) are obtained in the transmission mode.

FIG. 1.

Synthesis of PSs from resorcinol and formaldehyde using a modified Stöber method. (a)–(d) Temporal evolution of RF condensation reaction under ultrasonic irradiation. Electron microscopy images of the obtained carbonaceous particles (e)–(g) with ultrasonic mediation, and (h), (i) in the absence of sonication. Panels (e), (h), (i) are imaged using the scanning microscopy mode, while panels (f), (g) are obtained in the transmission mode.

Close modal

To uncover the atomic scale processes responsible for the accelerated kinetics of PS formation under ultrasonic radiation, we employ reactive molecular dynamics (RMD) simulations using the LAMMPS package.16 The interactions between the C, H, and O atoms are described using a general bond-order based reactive force field (ReaxFF),17 which accurately captures the dissociation, and formation of chemical bonds in a wide variety of organic compounds.18 During ultrasonic radiation, the collapse of bubbles causes intense local heating producing hot spots (of nm sizes) with an effective temperature of 5000–6000 K;5,12,13 such localized heating is short-lived resulting in extreme cooling rates of 1010–1013 K s−1.12,13,19 To mimic such sonication induced nanoscale hot spots, we increase the temperature of our system from 300 K to 6000 K over 0.5 ns and hold it at the elevated temperature for an additional 0.5 ns using canonical (constant number of atoms, volume, and temperature; i.e., NVT) RMD simulations. Thereafter, the system is cooled down to room temperature over 2 ns, i.e., with a cooling rate of 2.35 × 1011 K s−1.

RMD simulations indicate that the sudden surge in temperatures (to as high as ∼6000 K) over nanometer length scales, induced by ultrasonic radiation, facilitates rapid dissociation of solvent water molecules into H and OH radicals (within ∼0.1 ns). This finding is consistent with previous studies, which report that hot spots generated during ultrasonic radiation possess sufficient energy to cause bond cleavage events at such an exaggerated pace.13 The enormous kinetic energy available to the hot spots causes bond breaking and produces excited states of C, H, and O atoms, which in turn enables polymerization reactions, and rapid gelation.19 Additionally, we also observe that ∼ 90% of resorcinol and formaldehyde molecules decompose into atomic fragments as well as shorter hydrocarbons, similar to our previous works.20 Upon cooling, however, the polymerization of carbon atoms is severely impeded. The carbon atoms either remain strongly bound to nearby hydrogen/oxygen atoms or form short isolated linear hydrocarbon chains containing at most 3–5 carbon atoms. These linear hydrocarbon chains, once formed, persist throughout the cooling time period (2 ns). More interestingly, no cyclization, i.e., formation of graphite-like rings, is observed; such graphitic rings have been previously reported as crucial precursors in the growth of carbonaceous spheres.20 This clearly shows that the accelerated kinetics of PS formation in the sonication-mediated modified Stöber synthesis technique cannot be explained as a mere manifestation of elevated temperatures and exaggerated cooling rates associated with collapse of cavitation bubbles.

In addition to the formation of hot spots, the continuous nucleation, growth, and collapse of cavitation bubbles induced by ultrasonic radiation also result in drastic fluctuations in density over nanometer length scales.12,13 To systematically investigate the impact of cavitation on the nucleation/growth of PSs from RF polymerization, we introduce variations in the overall atomic density ρ of our system. This is achieved by randomly removing H and OH radicals from the equilibrated system at 6000 K to obtain a certain desired target density. At every density investigated in this work, the H/O ratio is maintained at a constant value (∼1.95) to avoid extraneous effects owing to variations in the overall OH content; in addition, the initial configurations at various ρ values are equilibrated at 6000 K using NVT-RMD simulations for 0.1 ns. The equilibrated configurations at various values of ρ are then subjected to cooling from 6000 K to 300 K using NVT-RMD simulation for 2 ns.

Figure 2 shows the temporal evolution of the system during the cooling from 6000 K to 300 K (over 2 ns) at various overall atomic density ρ with values ranging from 80 atoms/nm3 (solvent density: 1 g/cc) to 7.2 atoms/nm3 (only C atoms; complete loss of H and OH radicals). Initially (t = 0), at 6000 K, the atoms are randomly distributed at all values of ρ such that configurations resemble liquid-like amorphous state. Upon cooling, the polymerization of C atoms begins at ∼4500 K (at t ∼ 0.5 ns) irrespective of the value of ρ employed [Figs. 2(a), 2(d), 2(g), and 2(j)]. Beyond this early stage of nucleation, however, the formation of carbonaceous particles exhibits a density-dependent behavior. At high values of ρ, i.e., ρ30.6 nm3, only linear hydrocarbon chains containing a maximum of 3–5 C atoms form; no cyclization/ring formation is observed even when the cooling phase is completed, i.e., t = 2 ns [Figs. 2(b) and 2(c)]. A slight reduction in ρ to 14.2 atoms/nm3 results in formation of few graphitic rings although the polymerization of C atoms mainly yields isolated linear hydrocarbons. A direct visualization of the obtained structure at 300 K shows that only <1% of the C atoms are involved in the formation of graphite-like rings at ρ=14.2 atoms/nm3 [Figs. 2(e) and 2(f)]. Upon further reduction of atomic density to ∼9.5 atoms/nm3, a pronounced affinity to form graphitic structures is observed, as indicated by the surge in the fraction of total C atoms (∼70%) involved in graphite-like structures [Fig. 2(h)]. This results in polymerization of C atoms into several graphitic flakes (few nm in size) along with a few linear hydrocarbon chains as shown in Fig. 2(i). In the extreme case, wherein all the H and OH radicals leave the system, the preference for formation of C rings is greatly enhanced resulting in extremely rapid polymerization of C atoms [Figs. 2(k) and 2(l)]. In fact, in this case, at 300 K, the C atoms organize themselves into a highly ordered, spherical, and onion-like structure, which is composed mainly of graphitic rings [Fig. 2(l)].

FIG. 2.

Structural organization of the carbon atoms upon cooling an aqueous solution of resorcinol and formaldehyde from elevated temperatures attained due to sonication (6000 K) down to 300 K, as obtained from our RMD simulations. To investigate the role of sonication-induced cavitation on the RF condensation, we simulate various overall density ρ values (a)–(c) 30.6, (d)–(f) 14.2, (g)–(i) 9.5, and (j)–(l) 7.2 atoms/nm3. At each density, representative RMD snapshots at three different times, namely, 0.5 ns (4575 K) [Panels (a), (d), (g), (j)], 0.75 ns (2130 K) [Panels (b), (e), (h), (k)], and 2 ns (300 K) [Panels (c), (f), (i), (l)] are presented. For clarity, only the carbon atoms are shown.

FIG. 2.

Structural organization of the carbon atoms upon cooling an aqueous solution of resorcinol and formaldehyde from elevated temperatures attained due to sonication (6000 K) down to 300 K, as obtained from our RMD simulations. To investigate the role of sonication-induced cavitation on the RF condensation, we simulate various overall density ρ values (a)–(c) 30.6, (d)–(f) 14.2, (g)–(i) 9.5, and (j)–(l) 7.2 atoms/nm3. At each density, representative RMD snapshots at three different times, namely, 0.5 ns (4575 K) [Panels (a), (d), (g), (j)], 0.75 ns (2130 K) [Panels (b), (e), (h), (k)], and 2 ns (300 K) [Panels (c), (f), (i), (l)] are presented. For clarity, only the carbon atoms are shown.

Close modal

We further evaluate radial distribution functions (RDFs) g(r) in the final nanostructures obtained after cooling (i.e., 300 K) at different values of ρ [Fig. 3(a)]. At ρ30.6 nm3, the RDF for C-C pairs exhibit peaks arising from nearest-neighbors at separation distances r=1.31.5 Å, which is in good agreement with typical C-C bond lengths in sp3/sp2 hybridization environments.20 The higher order peaks are non-existent indicating the formation of isolated hydrocarbons and complete lack of long-range spatial order. Consistent with this finding and our direct visualization [Fig. 2(c)], a quantitative ring analysis further evidences that most of the C atoms are either bound to nearby H/O atoms or form short chain hydrocarbons without any cyclization [Fig. 3(b)]. Upon reducing the value of ρ to 14.6 atoms/nm3, an additional small peak appears at r ∼ 2.45 Å in the RDF of C-C indicating the formation of cyclic C rings,20 which is confirmed by our ring analysis indicating the formation of ∼10 C rings. A significantly high fraction (∼ 40%) of the C rings formed contains 6 carbon members, i.e., they possess graphitic order. As the value of ρ is further lowered to 9.5 atoms/nm3, the nearest neighbor peak becomes centered around r ∼ 1.47 Å, and the intensity of peak at r ∼ 2.45 Å in the C-C RDF progressively increases, indicating a profound increase in the formation of graphite-like C hexagons; furthermore, the higher order peaks become more prominent, indicating a long-range spatial order [Fig. 3(a)]. Our ring count analysis confirms this finding, indicating the formation of ∼2000 C-rings, out of which a large fraction (∼58%) consist of 6 C members [Fig. 3(b)]. Another noteworthy observation is the appearance of a shoulder around r 2.85 Å in the C-C RDF, which corresponds to the intra-hexagon spacing [Fig. 2(a)] and suggests formation of compact, less defective structure with graphitic order. When the system is devoid of any H/OH radical, i.e., ρ=7.2 atoms/nm3, all the C atoms present in the system are involved in cyclization forming predominantly hexagonal rings (∼65% of the total ∼3200 rings are 6-membered) [Fig. 3(b)]. The resultant onion-like structure possesses a high crystalline graphitic order, as indicated by well-defined peaks in the C-C RDF [Fig. 3(a)].

FIG. 3.

Structural order of the carbonaceous particles obtained from our RMD simulations (a) radial distribution function of C-C pairs in the final configuration (300 K) and (b) relative fraction of the carbon rings of various sizes at various overall density values. The inset in panel (b) shows the number of rings (with size < 12 C atoms) as a function of density. In panels (c)–(e), we show magnified atomic snapshots of the final configuration at 300 K obtained from our RMD simulations illustrating the extent of saturation of graphitic regions by hydrogen. The C, H, and O atoms are represented by black, cyan, and red spheres, respectively.

FIG. 3.

Structural order of the carbonaceous particles obtained from our RMD simulations (a) radial distribution function of C-C pairs in the final configuration (300 K) and (b) relative fraction of the carbon rings of various sizes at various overall density values. The inset in panel (b) shows the number of rings (with size < 12 C atoms) as a function of density. In panels (c)–(e), we show magnified atomic snapshots of the final configuration at 300 K obtained from our RMD simulations illustrating the extent of saturation of graphitic regions by hydrogen. The C, H, and O atoms are represented by black, cyan, and red spheres, respectively.

Close modal

Finally, we focus on the physical factors that govern the dependence of the morphology of the obtained carbonaceous particles on the variations in the atomic density of the system induced by ultrasonic radiation. At very high atomic densities, i.e., the cases in which most of the radicals formed via dissociation of solvent molecules are present in the system, each C atom is surrounded by an abundance of H and OH radicals. Consequently, these C atoms bind to the H and O atoms available in close proximity to them resulting in the formation of isolated short hydrocarbons, as shown in Fig. 3(c). As the overall density of the system reduces, the number of available H and OH radicals decreases, which in turn increases the propensity of the C atoms to cyclize into graphite-like rings. At intermediate density values, e.g., ρ=14.2 atoms/nm3, graphitic rings nucleate. However, before these small graphitic domains gain a critical size, their edges become saturated via binding with the nearby H/OH radicals obfuscating their further growth to form PS particles [Fig. 3(d)]. At low density values, such as ρ=9.5 atoms/nm3, there is a significant reduction in the number of H/OH radicals, which enables the growth of graphitic nuclei to grow into polymer/carbonaceous particles, several nanometers in size, within a few ns [Fig. 3(e)].

In summary, we have elucidated the atomistic mechanism that drives the sonochemical synthesis of polymer spheres from resorcinol and formaldehyde precursors. Our synthesis experiments employing ultrasonic probe reveal accelerated growth kinetics of PSs along with formation of cavitation bubbles and radicals. Reactive MD demonstrates that the hot spots generated during ultrasonic radiation possess sufficient energy to cause bond cleavage events accelerating the dissociation of precursor molecules. The high kinetic energy in conjunction with localized density lowering due to cavitation facilitates rapid nucleation and growth of PSs. Analysis of structural features indicates high degree of graphitic order with increased cavitation.

Use of the Center for Nanoscale Materials was supported by the U.S. Department of Energy, Office of Science, and Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357. This work was supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Office of Vehicle Technologies of the U.S. Department of Energy under Contract No. DE-EE0006832 under the Advanced Battery Materials Research (BMR) Program.

1.
Y. Z.
Jin
,
Y. J.
Kim
,
C.
Gao
,
Y. Q.
Zhu
,
A.
Huczko
,
M.
Endo
, and
H. W.
Kroto
,
Carbon
44
,
724
(
2006
).
2.
G. S.
Chai
,
S. B.
Yoon
,
J. H.
Kim
, and
J. S.
Yu
,
Chem. Commun.
2004
,
2766
.
3.
G.
Jianping
,
H.
Tuan
,
H.
Yongxing
, and
Y.
Yadong
,
Nano Lett.
8
,
931
(
2008
).
4.
K. S.
Suslick
and
G. J.
Price
,
Annu. Rev. Mater. Sci.
29
,
295
(
1999
).
6.
W.
Stober
,
A.
Fink
, and
E.
Bohn
,
J. Colloid Interface Sci.
26
,
62
(
1968
).
7.
J.
Liu
,
S. Z.
Qiao
,
H.
Liu
,
J.
Chen
,
A.
Orpe
,
D.
Zhao
, and
G. Q.
Lu
,
Angew. Chem., Int. Ed.
50
,
5947
(
2011
).
8.
A. M.
ElKhatat
and
S. A.
Al-Muhtaseb
,
Adv. Mater.
23
,
2887
(
2011
).
9.
V. G.
Pol
,
L. K.
Shrestha
, and
K.
Ariga
,
ACS Appl. Mater. Interfaces
6
,
10649
(
2014
).
10.
A. D.
Roberts
,
X.
Li
, and
H.
Zhang
,
Chem. Soc. Rev.
43
,
4341
(
2014
).
11.
J.
Liu
,
N. P.
Wickramaratne
,
S. Z.
Qiao
, and
M.
Jaroniec
,
Nat. Mater.
14
,
763
(
2015
).
12.
K. S.
Suslick
,
Y.
Didenko
,
M. M.
Fang
,
T.
Hyeon
,
K. J.
Kolbeck
,
W. B.
McNamara
,
M. M.
Mdleleni
, and
M.
Wong
,
Philos. Trans. R. Soc., A
357
,
335
(
1999
).
13.
K. S.
Suslick
,
Ultrasound: Its Chemical, Physical, and Biological Effects
(
VCH Publishers
,
New York
,
1988
).
14.
W.
Lauterborn
,
T.
Kurz
,
R.
Mettin
, and
C. D.
Ohl
, “
Experimental and theoretical bubble dynamics
,” in
Advances in Chemical Physics
, edited by
I.
Prigogine
and
S. A.
Rice
(
John Wiley and Sons, Inc.
,
1999
), Vol.
110
, Chap. 5, p.
295
.
15.
C. D.
Ohl
,
T.
Kurz
,
R.
Geisler
,
O.
Lindau
, and
W.
Lauterborn
,
Philos. Trans. R. Soc., A
357
,
269
(
1999
).
16.
S.
Plimpton
,
J. Comput. Phys.
117
,
1
(
1995
).
17.
K.
Chenoweth
,
A. C. T.
van Duin
,
I.
Goddard
, and
A.
William
,
J. Phys. Chem. A
112
,
1040
(
2008
).
18.
See supplementary material at http://dx.doi.org/10.1063/1.4959885 for details of the reactive molecular dynamics simulations.
19.
Y. T.
Didenko
,
W. B.
McNamara
, and
K. S.
Suslick
,
J. Am. Chem. Soc.
121
,
5817
(
1999
).
20.
S. A.
Deshrnukh
,
B.
Narayanan
,
G.
Kamath
,
V. G.
Pol
,
J.
Wen
,
D. J.
Miller
, and
S. K. R. S.
Sankaranarayanan
,
Carbon
96
,
647
(
2016
).

Supplementary Material