The discovery of magnetic-induced synthesis of a ternary semiconductor is far beyond what has been observed to date using traditional solution chemistry, such as the solvothermal method, where temperatures as high as 200-300o C over time periods of several to dozens of hours are frequently required. The magnetic-induced crystal growth of Zn-doped CIS in this investigation should cause a self-heating effect originating from the developing Zn-CIS compound, since the CIS nanocrystals showed a stronger paramagnetic behavior after doping with Zn, Figure 6.9. The hysteresis loop for the resulting Zn-CIS nanocrystals was corrected for the diamagnetic contribution (inset of Figure 6.9) and shows typical paramagnetic behavior, with the magnetization essentially saturated above 10 kOe. The hysteresis loop data for the Zn-CIS nanocrystals at 300k and 5k (Figure 6.10) also evidenced the superparamagnetism of such Zn doped CIS nanocrystals. In theory, magnetic colloids in a magnetic field experience an internal stress as a result of the non-uniform distortion arising from magnetic forces, generating heat. Under magnetic induction,
-10000 -5000 0 5000 10000
the temperature of the Zn-CIS colloidal solution gradually rose from the very beginning and reached a steady-state temperature of several tens of Celsius, at which point the heat generation from the magnetic Zn-CIS nanocrystals was equilibrated with the heat dissipated to the environment.
Figure 6.9 Magnetization curves measured at room temperature for Zn-CIS QDs (hollow) and Zn-CIS nanocubes (solid). The inset shows the raw data and the data after subtracting the high-field diamagnetic component.
Figure 6.10 Magnetization curves measured at different temperature for Zn-CIS QDs.
Assuming magnetic heating involves only isolated, independent nanoparticles, RF field friction in and around a magnetic particle can be stemming from two sources:
[158] first, the particle may tumble causing frictional heating at the particle-solvent interface. The relaxation time for this mode can be estimated as the time required for Brownian motion over a characteristic distance of the order of one particle diameter.
Brownian relaxation may not be responsible for the frictional heating of the Zn-CIS, however, because the heat generated from this mechanism should be equally shared between the nanoparticle and solvent so it is unlikely for the Zn-CIS alone to reach a very high temperature. Friction may also arise from spin rotation without crystal lattice rotation within the crystal. The relaxation time for this mode (Neel relaxation) is the reciprocal of the spin flipping rate, too fast to contribute any significant friction in a RF field. As a first approximation, we can roughly approximate the heating rate RZCIS of the Zn-CIS (or ZCIS) nanoparticles in terms of RS using VZCISRZCISCZCIS= (1-VZCIS)RSCS. Here VZCIS is the volume fraction of the Zn-CIS nanocrystals relative to the solution, and CS and CZCIS are the volumetric specific heat of the solvent and ZCIS, respectively. Here we used VZCIS~0.01, meanwhile, referring to the specific heat of ZCIS and the solvent, we estimate CS/CZCIS ~1. Therefore, the calculated RZCIS is from 10o C/s to 100o C/s, i.e., it takes a few to few tens of seconds for the temperature to rise to several hundred degrees of Celsius in the Zn-CIS nanocrystals before a steady state is reached. However, interfacial heat transfer at a nanometer scale surrounding the nanocrystals is not well determined.
Referring back to the forming of such Zn-CIS nanocrystals under magnetic field, how do these particles form in the first place? In the solution without adding Zn precursor (which is virtually a Zn-S-containing precursor), the other two precursors (i.e., Cu and In) are unable to react in the presence of magnetic field (we have also found no any reaction with Zn-S-containing precursor alone). However, after the
Zn-S-containing precursor was prepared separately, i.e., completely dissolved in a solvent, we found no any precipitate in the solution and believed it may form a Zn-S-containing compound. While the first aliquot amount of the compound was added to the mixture solution, energetically effective collisions (under heating and agitating) among those precursors should cause the incorporation of Cu ion with the compound (due to the similar ionic size between Cu and Zn ions), upon which a chemically equivalent mixture of the those molecular precursors may develop into the first nanocrystal or “first compound” in the presence of magnetic field, such as Cu-containing ZnS compound, which allows to induce heating under exposure of the field. Since the electron configuration of Cu+ is d10, ZnS doped with Cu+ would be diamagnetic. If copper existed as Cu0 or Cu2+ in the ZnS, the system would be strongly paramagnetic. [159] In our study, the Cu precursor is CuCl, which means Cu+ could be pseudo-oxidized by the Zn-S-containing compound (i.e., S- in the Zn precursor, Figure 6.11).
Figure 6.11 Mechanism of “first paramagnetic nanocrystal” formation.
While the Cu-containing ZnS compound exhibits paramagnetism, the magnetic
Zn++
field can thus induce heat. To prove our hypothesis for the formation of a chemical compound, we mixed only Cu precursor into the solution of Zn-S-containing precursor, in the absence of In precursor. A secondary force may bring interaction between Cu and S ions, rather than forming any chemical bond, leading to a featureless of Cu-S bond using Raman spectroscopic analysis, in other words, there appears no chemical interaction between Cu and S in the mixture. In spite of such weak attraction, the reaction is somehow an ongoing procedure in the presence of magnetic field which was confirmed by a visual observation of the solution, turning from turbid appearance to transparent. This phenomenon indicated the precursors in the solution were actually affected by the magnetic field and consequently induced heating for further reaction. Although we have no direct evidence available for such an intermediate compound presently, the SQUID analysis of collected precipitation from the mixed Zn-S and Cu precursors demonstrated our hypothesis (see Figure 6.12). It is highly likely to believe that the “first” magnetic nanocrystal may preserve for a short period of time and form rapidly into a more stable nanocrystal with Zn-Cu-In-S near-stoichiometric chemistry under the presence of magnetic field due to the involvement of other ions, such as In. Once the first paramagnetic Zn-CIS nanocrystal was developed, a continuous growth was followed immediately and rapidly. Development of such a "first paramagnetic nanocrystal" in the solution should be time dependent, following a crystal growth kinetic, both scenarios result in a mixture of the nanocrystals of varying geometry and size over a certain time span of reaction, as evidenced in Figure 6.4, where the nanocrystals with different size and shape can be observed simultaneously.
-10000 -5000 0 5000 10000 -0.003
-0.002 -0.001 0.000 0.001 0.002 0.003
M (emu/g)
H (Oe)
Figure 6.12 Magnetization curve of precipitation from Zn and Cu precursors.
Accordingly, the formation of Zn-CIS nanocrystals arising from magnetic induction is associated with fast nucleation and growth kinetics. In other words, once the nuclei begin evolving, a rapid temperature rise in the Zn-CIS nuclei should promote fast atomic deposition from the surrounding precursor bath. The exchange of heat and mass between the growing hot nanocrystals and the surrounding cold solvent due to convection should accelerate the atomic deposition of the precursors. One important feature of this synthesis is that the hot nanocrystals are likely to impart energy to the impacting precursor molecules which carry an organic part, where a further thermal dissociation of the organic compartment at the points of deposition or collision should take place upon numerous collisions between the precursor molecules and growing nanocrystals. This results in a chemically-pure deposition, leaving behind the organic counterparts in the solvent medium, as is further evidenced in the HR-TEM photos, (Figure 6.4), where highly orderly-arranged texture of the Zn-CIS lattice for the nanocrystals from smallest to largest dimensions can be resolved at various time periods of synthesis. This result is not possible for solvothermal or similar techniques, except at relatively high temperatures, e.g., >200o C, for a long
2 nm 2 nm (b)
(1 1 2) (a)
synthesis duration, e.g., >10 hours, where the resulting ternary nanocrystals showed very poor crystallinity (Figure 6.13) and poor optical behaviors. Incorporation of organic molecules into the developing semiconductor nanocrystals may be one critical factor deteriorating the desired properties.
Figure 6.13 HR-TEM images of ZCIS nanocrystals synthesized through high-temperature organic solvent method (a) and HFMF (b). This comparison indicated that crystallinity of ZCIS nanoparticles was enhanced by HFMF.
It is also surprisingly to learn that a highly uniformly compositional development of the nanocrystals of various geometries can be achieved, which suggests that the co-deposition of all those molecular precursors over the time span of the synthesis is kinetically similar. In other words, once the nanocrystals were initially evolved, the potential of those precursors upon impingement to the growing nanocrystals should be energetically similar. Such a collision should be a matter of an interaction between the precursor and nanocrystals, as a result of interparticle attraction. This is further evidenced with a subsequent zeta potential measurement (supporting information Table 6.2) of the nanocrystals. The zeta potential value of Zn-CIS nanocrystals was slightly negatively charged in an average value of -10.8 + 3.2 mV for various
geometrical structures. These slightly negatively charged nanocrystals should exert weak attraction to the precursors, following a kinetically similar impingement to form final Zn-CIS crystals. However, the Cu-rich phase evolution of the nanocrystals suggests when incorporated into the lattice, the Cu precursor tends to form a lowest-energy solid solution compared to those with Zn precursor. Furthermore, the maximum amount of Cu substitution by Zn in the development of the energetically stable Zn-CIS nanocrystals in this finding is approximately 32.43%.
Table 6.2 Zeta potential values of different shape Zn-CIS nanocrystals. Particles are the sample obtained under HFMF duration for 30 and 45 seconds which show emission peaks at 590 and 630 nm.
Shape Particle (30sec)
Particle (45sec)
Cube/Pyramid Bar
Zeta potential (mV) 7.25 14.3 9.36 9.74
In order to distinguish the difference between magnetically-induced synthesis and the high-temperature organic solvent method (HTOSM), we compared our nanocrystal with the one reported by Nakamura et al. [75] The optical properties, including PL and UV-vis characters, are presented in Figure 6.14. There are two emission peaks located at 420 and 560 nm of the HTOSM sample. The 420 nm peak contributed from the ZnS indicates an insufficient energy transition in HTOSM to permit alloying of the Zn atom into the CIS lattice and form a secondary phase, ZnS.
This result confirmed our hypothesis of energy transform through magnetic induction.
Moreover, the UV-vis absorption spectra of the Zn-CIS nanocrystals prepared with magnetic induction shows a stronger and sharper peak compared to the Zn-CIS nanocrystals prepared through HTOSM. The relatively sharp exciton peak in the absorbance spectra indicates that the Zn-CIS nanocrystals are relatively size- and
(b) (a)
300 400 500 600 700 800 900
without HFMF under HFMF
Absorption (a.u.)
Wavelength/nm
400 500 600 700 800 900
without HFMF under HFMF
Relative Intensity (a.u.)
Wavelength/nm
Figure 6.14 Fluorescence (a) and absorption (b) spectra of the Zn-CIS nanocrystals obtained with (red) and without magnetic induction (black). The excitation wavelength for the fluorescence measurements was 366 nm.
shape- monodisperse. Furthermore, the optical properties are much better with band
edge PL and narrow peak widths compared with the one prepared from HTOSM.
While there are some studies have reported decent results of CIS [160-164] and CISe [164-165] (CuInSe2) nanocrystals recently, few of them exhibited quantum-confinement behaviors. Nevertheless, all of these reports need rigorous conditions such as high temperature (180~300oC), long crystal-growth duration (1~24 hours), and non-oxygen environment. Compared to those experiment conditions,
magnetically-induced Zn-CIS synthesis is much more easy and fast for such nanocrystals development. This distinct feature further confirms the promising development of magnetically-induced Zn-CIS synthesis over conventional high-temperature pathways.