Organic-Component Dependent Crystal Orientation and Electrical Transport Properties in ALD/MLD Grown ZnO-Organic Superlattices.

Two series of ZnO-organic superlattice thin films are fabricated with systematically controlled frequencies of monomolecular hydroquinone (HQ) or terephthalic acid (TPA) based organic layers within the ZnO matrix using the atomic/molecular layer deposition (ALD/MLD) technique. The two different organic components turn the film orientation to different directions and affect the electrical transport properties differently. While the TPA layers enhance the c-axis orientation of the ZnO layers and act as electrical barriers depressing the electrical conductivity even in low concentrations, adding the HQ layers enhances the a-axis orientation and initially increases the carrier concentration, effective mass, and electrical conductivity. The work thus demonstrates the intriguing but little exploited role of the organic component in controlling the properties of the inorganic matrix in advanced layer-engineered inorganic-organic superlattices.

Introduction

Inorganic–organic superlattices (SLs) are intriguing candidates for unique multifunctional materials since they comprise interfaces between two material components as different as inorganics and organics. As typical roles of the two components, we may imagine that the inorganic layers would be responsible for the specific electrical, optical or magnetic properties, and the organic layers could bring, e.g., mechanical flexibility to the hybrid material.[1−5] Also, the interface itself may have important functionality, e.g., as a barrier for heat conduction, ion transport, or gas permeability.[6−13] Introduction of organic interfaces within an inorganic matrix also provides us with a tool to nanostructurize the material, e.g., for quantum confinement effects.[14−16] An important dimension little investigated is the mutual effect between the two components; in the present work, we investigate such effects, in particular the role of the organic component in controlling the structural and electrical transport properties of the inorganic component in superlattices, where we introduce ultrathin monomolecular organic layers within a semiconducting inorganic matrix.

To synthesize such designer’s materials, the combined atomic/molecular layer deposition (ALD/MLD) thin-film technique[17−20] provides us with a number of unique benefits, as it yields—starting from mutually reactive gaseous inorganic and organic precursors—high-quality inorganic–organic thin films based on strong chemical bonding interactions, and facilitates the desired level of control on the introduction frequency of the organic layers within the superlattice structure.[21−24] The technique is moreover believed to be industry-feasible, based on the grounds set by the parent ALD (atomic layer deposition) technology for state-of-the-art ultrathin inorganic films.[25,26]

We chose ZnO as the inorganic base owing to its attractive functional properties and the facile ALD process.[27] Wurtzite-structured hexagonal ZnO is a wide-band gap transparent n-type semiconductor with high electron mobility, and has accordingly various potential applications in electronics and energy harvesters. The most commonly employed ALD process based on diethyl zinc (DEZ) and water yields polycrystalline ZnO thin films in a wide deposition temperature range.[27−29] The possibility to realize ZnO thin films with specific orientations or enhance their electrical characteristics would be of utmost importance for many of their applications. Just as one example, the photolysis property of ZnO has shown a clear dependence on the film orientation.[30] In this work, we will demonstrate that different monomolecular organic layers introduced within ALD-grown ZnO films may tune both the film orientation and the electrical transport properties in different ways. The two organic components investigated are terephthalic acid (TPA) and hydroquinone (HQ). Both of these precursors have been commonly used in previous ALD/MLD works.[17,19,21,31−34]

Methods

The ZnO–organic superlattice thin-film depositions were carried out on 3 × 3 cm2 silicon and glass substrates in a flow-type hot-wall ALD reactor (ASM Microchemistry; F-120) using the following precursors (sublimation temperatures in parentheses): diethyl zinc (DEZ; RT), deionized water (H2O; RT), hydroquinone (HQ; benzene-1,4-diol; 120 °C), and terephthalic acid (TPA; 1,4-benzene dicarboxylic acid; 180 °C). The number of deposition cycles for each film was set to 600 ± 2 to keep the overall film thickness in the range of 97–116 nm. All the films were deposited at 220 °C. The carrier and purge gas was N2; and the pulse/purge times (in sec.) for DEZ, H2O, TPA, and HQ were 1/1.5, 1.5/2, 10/30, and 8/12, respectively. Each deposition was started and completed with the (DEZ + H2O) ALD cycles for the ZnO block, and each organic layer between these blocks consisted of a single (DEZ + TPA/HQ) MLD cycle according to the [(DEZ + H2O) + (DEZ + TPA/HQ)] + (DEZ + H2O) scheme; the resultant SL thin film was named as TPA/HQ-n(m), where the number n tells us the total number of organic layers in the film, and m is the number of ALD cycles applied to grow the individual ZnO blocks. For example, TPA-6(85) denotes a film deposited with a total number of six TPA layers each deposited with a single (DEZ+TPA) cycle, between seven ZnO blocks in total, each deposited with 85 (DEZ + H2O) cycles; the total number of cycles is thus (6 × 85) + (6 × 1) + 85 = 601.

The film thicknesses and densities were determined by X-ray reflection (XRR; Panalytical XPert diffractometer, Cu Kα source). The experimental XRR patterns were fitted (X’Pert Reflectivity program v1.3 from PANalytical) not only for the overall film thickness but also for each layer in the SL structure separately; the fitting protocol is described in detail in the Supporting Information (SI). The same tool was used to collect the grazing-incidence X-ray diffraction (GIXRD) patterns for the samples with an incidence angle of 0.5°. To simplify the comparison of relative intensities of the peaks and removing various instrumental effects, the total intensities of GIXRD patterns were normalized based on their maximum intensity, except for the pure ZnO reference film.

Fourier transform infrared (FTIR; Bruker Alpha 2) spectra were collected for the samples to verify the organic species and examine the bonding modes between the inorganic and organic components. In these FTIR experiments, the chamber was continuously purged with N2 gas during the measurement, and a spectrum of blank Si was subtracted from the measured sample spectrum to compensate the interference caused by the substrate.

Electrical resistivity (ρ) and Hall voltage were measured for the films deposited on glass using the standard four point-probe technique (Physical Property Measurement System; PPMS; Quantum Design; 9T magnet). The Seebeck coefficient was measured in isothermal and open-circuit conditions. At a fixed temperature, a temperature gradient was created along the length of a rectangular bar-shaped sample by applying a small amount of heat to one end. In the steady-state condition, the temperature difference and the Seebeck voltage were measured simultaneously. The Seebeck coefficient was calculated by dividing the Seebeck voltage with the temperature difference. The Hall voltage was measured as a function of external magnetic field by sweeping it from −8 T to +8 T. The carrier concentration (n) and effective mass (m*) were calculated from the Hall coefficient (RH) and Seebeck coefficient (S) following a procedure described in detail in SI.[35,36]

Results and Discussion

Confirmation of Designed SL Structures

All our ZnO–organic thin films deposited with the pulsing sequence [(DEZ + H2O) + (DEZ + TPA/HQ)] + (DEZ + H2O) were highly uniform from visual inspection. The high quality and the intended structure of the films can be clearly seen from the GIXRD, XRR, and FTIR characterization data shown in Figure ; which presents the most indicative parts of the GIXRD patterns and FTIR spectra, while the full spectra/patterns are shown in the SI (Figure S1). Table summarizes the investigated SL thin films and their overall thicknesses (determined from XRR data); the films are denoted as TPA/HQ-n(m), where the number n tells us the total number of organic layers in the film, and m is the number of ALD cycles applied to grow the individual ZnO blocks. As expected, the overall film thickness was found to slightly increase with the increasing portion of the thicker organic layers.

Figure 1

(A) FTIR spectra, (B) XRR patterns, and (C) GIXRD patterns of the HQ- and TPA-based SL films; in (C) the (100), (002), and (101) diffractions are found at ca. 31.91°, 34.61°, and 36.29°, respectively.

Table 1

List of Films with Details

sample	total cycles m(n + 1)	film thickness (nm)
ZnO	600	97
TPA-6(85)	601	108
TPA-7(74)	599	105
TPA-8(66)	602	107
TPA-11(49)	599	110
TPA-14(39)	599	111
HQ-6(85)	601	113
HQ-9(59)	599	114
HQ-11(49)	599	116
HQ-14(39)	599	115

Figure B depicts the XRR patterns for the films; these patterns clearly confirm (i) the regular arrangement of the intervening organic layers within the ZnO matrix, and (ii) the targeted number n of these layers in each sample. The formation of the regular SL structures is seen from the presence of the regularly repeating more intense peaks in the patterns; such SL peaks are absent in the pattern of the pure ZnO reference thin film. Then, it is also clear that the number of the smaller peaks in between the SL peaks systematically increases with n, as designed.

Our ZnO–organic thin films thus consist of regular stackings of polycrystalline ZnO layers and monomolecular organic layers. From the GIXRD patterns (Figure C), the well-known hexagonal wurtzite structure can be readily confirmed for the ZnO layers. Then, the FTIR spectra reveal to us the mode of bonding of the organic molecules (benzene or terephthalic acid) at the ZnO–organic interfaces. First, no indications of the broad stretching absorption around 3500 cm–1 or the sharp bending vibration around 940 cm–1 due to hydroxyl groups are seen for any of the films, confirming that these groups have perfectly reacted with DEZ molecules during the (DEZ + H2O) and (DEZ + TPA/HQ) cycles, as expected for ideal ALD and MLD processes. This ascertains that the selected pulse and purge times are adequate for the targeted SL thin-film processes. However, the FTIR spectra clearly show the vibrations of the benzene ring and carboxyl group around 1400–1550 cm–1 for the TPA-based films, and those of the benzene ring at ca. 1490 cm–1 for the HQ-based films, the intensities of these features properly increasing with the number of organic layers. This vibrational area also tells us more details of the way the organic molecules are bonded to Zn atoms; see the Discussion in section 3.2.

Organic-Component Dependent Structural Changes

Having a closer look at the FTIR spectra for the TPA-based SL films reveals that the distance (Δ) between the two carboxylate peaks in the 1400–1700 cm–1 area, due to the symmetric and asymmetric stretchings of the carboxylate groups, is ca. 148 cm–1 (= 1550–1402 cm–1) and remains essentially constant for all n values. In literature, a splitting of these peaks in the range 130 < Δ < 200 cm–1 has been taken as an indication of bridging-type bonding between the carboxylate group and the metal atoms.[16,17,33,37] For the HQ molecule only a unidentate type bonding is possible. On the basis of these differences in bonding mode and also in the sizes of the two organic molecules, each TPA moiety requires twice as much space as one HQ moiety in the SL structure (Figure ). Interestingly, we could see this difference clearly in the density values determined for the organic layers separately from the XRR data fittings, i.e., ca. 2.6 g/cm3 for the HQ layer and ca. 1.3 g/cm3 for the TPA layer. Our XRR fittings also provided values for the thicknesses of the individual organic layers; the obtained values, i.e., 9.6 and 7.4 Å for the TPA and HQ layers, respectively, are in line with the anticipated (oxygen-to-oxygen distance) lengths of the TPA (8.1 Å) and HQ (6.1 Å) molecules.

Figure 2

Schematic representations of ZnO–organic supercells for (A) TPA-based, and (B) HQ-based SL films; color code for different atoms: Zn, gray; O, red; C, black; and H, white; distances in Å.

Most excitingly, the differences in the bonding modes of the two organic constituents were found to result in structural changes in the surrounding ZnO layers. From the GIXRD patterns shown in Figures C and S2, it can be seen that the relative intensities of the (100), (002), and (101) peaks are affected differently by the TPA and HQ layers. For the HQ-based SLs, the strongest diffraction is (100) and then (101), while for the TPA-based SLs, the situation is just the opposite, i.e., the strongest reflection is (002). In the case of the TPA layers, it is straightforward to understand that the difficulty of the TPA molecules to form bridging bonds (which they prefer) on the typical (100) oriented ZnO surface drives the underlining ZnO layers toward the c-axis orientation (Figure ), i.e., the (002) peak is enhanced. Alternatively, introduction of HQ layers into the ZnO matrix seems to enhance the (100) peak, compared to the polycrystalline quite randomly oriented ZnO reference film. Besides that, the relative intensity of the (101) reflection systematically increases with the increasing number of HQ layers.

The aforementioned changes in the crystallographic orientation of the ZnO layers are—quite impressively—also seen both in the overall growth rates of the SL films, and in the densities of the ZnO layers. From the fittings of the XRR patterns, the growth-per-cycle or GPC (total film thickness divided by the number of deposition cycles applied) value was 1.7 and 1.8 Å/cycle for the TPA- and HQ-based SL films, respectively, while the densities for the ZnO blocks in these SL films were 5.41 and 5.15 g/cm3, respectively.

Organic-Component Effect on Transport Properties

For the electronic transport property evaluation, we measure the resistivity (ρ) and Seebeck coefficient (S) values for all the samples in the 20 to 400 K temperature range. The Hall voltage was measured in the 50 to 350 K range in the steps of 50 K. For comparison, we also carried out parallel measurements for an ALD-grown ZnO film, which is a well-known n-type semiconductor. The measured data are presented in Figure (A and B), and the calculated charge carrier density (n) and effective mass (m*) are presented in Figure (C and D). A clear dependence of the electronic transport properties on the choice of the organic component is evident from Figure .

Figure 3

(A) Seebeck coefficient, (B) resistivity, (C) carrier concentration, and (D) effective mass against temperature for the TPA- and HQ-based SL films.

In general, the first assumption would be that adding organic barrier layers in the semiconducting inorganic matrix should increase ρ and make the (negative) S larger in magnitude. This expected trend is indeed exactly seen for the TPA-based SLs, for which the magnitudes of ρ and S systematically increase with the number of TPA layers. Also, the charge carrier density (n) extracted from the measured Hall coefficient (RH) decreases with the increasing number of TPA layers.

The HQ-based films behave differently. First of all, initially ρ decreases considerably compared to pure ZnO, but then slowly increases when more HQ layers are introduced. We believe that the initial decrease in ρ is due to electron doping, which is later compensated by the barrier effect caused by the more frequent organic layers. Indeed, the value of n extracted from RH behaves accordingly, increasing initially and then decreasing for the higher concentration of HQ layers. The value of S shows relatively small changes at low temperatures below 100 K with increasing organic content, however a significant increase is observed at high temperature, both for the TPA- and HQ-based SL films.

Previous computational calculations have indicated that the addition of HQ layers within the ZnO matrix decreases the band gap, generating two flat bands in the band gap of bulk ZnO and producing a significantly reduced indirect band gap.[11,14] The flatter bands near the Fermi level give rise to the larger effective mass for the charge carriers.[35] Another previous molecular calculation for isolated HQ molecules indicated that the HOMO of HQ is situated within the energy gap of bulk ZnO.[1,22] Hence, our present experimental results for the ZnO–HQ superlattices with a small amount of HQ match well with these theoretical findings.

Comparison between the results for the TPA- and HQ-based SLs reveals that the latter films are approximately 10 times less resistive. The electron doping effect of the HQ layers may originate from the generally known chemistry fact that the oxygen electrons in HQ are less stable in comparison to those in TPA, where they are relatively stabilized by the carbonyl groups. Therefore, at first adding HQ to the system facilitates the electron transfer and increases n and, consequently, reduces ρ. However, by adding more HQ layers, the barrier effect increases and overcomes the doping effect. For TPA, only the barrier effect is present, and ρ increases thorough the entire SL series with increasing number of organic layers.

Finally, we discuss the role of effective mass (m*) calculated by using the values of n and S for our ZnO–organic superlattices. From Figure , the increase in the organic content decreases not only n but also m* for all our SL films except those with the low frequency of HQ layers. Since the magnitude of (negative) Seebeck coefficient is directly related to m* but inversely to n (SI), and since |S| increases for TPA and decreases for HQ upon increasing the organic content, we tentatively conclude that the dominant factor for the transport properties of our ZnO–TPA SLs is m*, while for the ZnO–HQ SLs it is n. For the HQ-based films due to the less restricted electrons and the higher n, S decreases as more organic layers are added; the initial increase is explained by the flatter bands,[11,14] which means larger m*.

Conclusions

We have demonstrated that when combining inorganic and organic components into a single superlattice structure the organic component can have a considerable role in controlling both the structural and electronic transport properties of the inorganic matrix. Most importantly, such a level of control as demonstrated here can be achieved without additional dopants or external stimulants.

The system investigated here was a ZnO–organic superlattice with two different organic components, terephthalic acid and hydroquinone. For making such precisely tailored superlattice structures, the strongly emerging ALD/MLD thin-film technique is uniquely advantageous. Here we varied the number of monomolecular organic layers within the ZnO matrix. The superlattice films were then systematically characterized for the Seebeck coefficient, electrical resistivity, carrier concentration, effective mass, and crystal orientation alterations caused by the organic layers. We could see that, while the TPA layers enhanced the c-axis orientation of the ZnO layers and acted as electrical barriers depressing the electrical conductivity even in low concentrations, adding the HQ layers enhanced the a-axis orientation and initially increased the carrier concentration, effective mass, and electrical conductivity.

We are convinced that the deeper understanding of the possible active roles of organic components in multilayered inorganic–organic materials will stimulate the quest for next-generation advanced materials with tailorable properties; the present results should be readily transferable to many other related hybrid inorganic–organic materials.