Journal of Materials Science: Materials in Electronics

, Volume 17, Issue 12, pp 1047–1053

Hybrid organic on inorganic semiconductor heterojunction

Authors

  • C. H. Chen
    • Department of Electrical and Computer EngineeringMcGill University
    • Department of Electrical and Computer EngineeringMcGill University
Article

DOI: 10.1007/s10854-006-9038-y

Cite this article as:
Chen, C.H. & Shih, I. J Mater Sci: Mater Electron (2006) 17: 1047. doi:10.1007/s10854-006-9038-y

Abstract

Hybrid organic on inorganic semiconductor heterojunctions with a sandwich structure have been fabricated and studied using conjugated polymers. The inorganic semiconductor was n-type silicon substrate. The conjugated polymers used include poly(2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene) containing polyhedral oligomeric silsesquioxanes (MEH-PPV POSS), regioregular poly(3-hexylthiophene) (RR-P3HT) and poly(3,4-ethylenedioxythiophene) (PEDOT). Current density–voltage and capacitance–voltage measurements were performed. All of the devices displayed a rectifying characteristic. Among these devices, the first ever reported PEDOT doped with BF3 on n-Si heterojunction devices showed the best performance with a rectification ratio around 5.7 × 105 at  ± 2 V and an ideality factor of 2.3. The results showed better device performance with decreased potential barrier height at the organic–inorganic interface. Results also suggested that smaller energy level offset between the HOMO of the conjugated polymer and the work function of anode metal will improve device performance.

1 Introduction

During the past decades, intensive research on semiconductor heterostructures has led to the discoveries of many new physical phenomena and device concepts such as semiconductor lasers and photodetectors. A new class of heterostructures reported involves contacts between organic and inorganic semiconductors. Several heterojunctions with a thin layer of a molecular semiconductor deposited onto the surface of an inorganic semiconductor substrate such as Si, GaAs, or InP have been found to form rectifiers with characteristics similar to ideal p–n junctions [1].

The main advantages of organic materials include simple and low-temperature thin film processing through inexpensive techniques such as spin coating. In addition, the flexibility of organic chemistry enables the formation of organic molecules with useful luminescent and conducting properties. For example, the carrier mobilities of organic channel layers in organic field-effect transistors (OFETs) have increased dramatically from 10−4 to 1 cm2/V s (comparable to those of amorphous silicon) over the last few years [2]. Many applications of the organic heterojunctions devices have been reported including photovoltaic [3] and electroluminescent [4, 5] devices. However, the electrical properties of organic semiconductor devices often degrade when exposed to unfavourable environments. Moreover, organic semiconductors may be affected by many of the chemicals used in conventional device processing. These two factors have limited the practical use of organic semiconductors despite their demonstrated applications to optoelectronic devices.

Nevertheless, considerable progress has been made in realizing practical active electronic and optoelectronic devices where an organic material forms an integral part of the device structure. One promising approach employs a thin organic film that is layered onto the surface of a conventional inorganic semiconductor substrate to form a hybrid organic on inorganic semiconductor heterojunction device. The main advantage of such hybrid devices is the possibility of altering the composition of the organic film to effect large changes in its optical and electronic properties. Furthermore, different combinations of organic and inorganic semiconductor can be utilized to obtain different desirable properties or applications.

The purpose of this paper is to investigate the characteristics of the hybrid organic on inorganic semiconductor heterojunction devices by current-voltage and capacitance–voltage measurements. Organic polymers chosen for the fabrication include poly(2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene) containing segments of polyhedral oligomeric silsesquioxanes (MEH-PPV POSS), regioregular-poly(3-hexylthiophene) (RR-P3HT) and poly(3,4-ethylenedioxythiophene) (PEDOT), while the inorganic counterpart is n-type silicon substrate.

2 Experimental

2.1 Device preparation

N-type (111) silicon wafers are first cut into smaller substrates of rectangular shape. The Si substrates are then cleaned using acetone, trichloroethylene and de-ionized water in an ultrasonic bath. A wet oxidation is performed to grow a thick silicon dioxide layer (∼0.5 μm) on top of the substrate. The purpose of this oxide layer is to reduce the probability of probe needle punching through the top metal contact and the organic polymer during measurement.

Photolithography process is carried out to open windows for active regions on the substrate using a first mask. The SiO2 is then etched away with buffered HF. Immediately after the substrate has been rinsed and dried; the organic polymer (MEH-PPV POSS, P3HT or PEDOT) is spin-coated at 2,000 rpm for 20 s onto the substrate to create a uniform layer. MEH-PPV POSS used in this work was prepared by dissolving 20 mg of the organic polymer in 8 g of chloroform while RR-P3HT solution was prepared by dissolving 56 mg of the P3HT in 8 g of chloroform to obtain a concentration of 0.7 wt%. The PEDOT was purchased from Sigma Aldrich.

The samples are then mounted onto the substrate holder of the vacuum system. A second mask is aligned carefully with each substrate so as to create the top metal electrode with thermal evaporation of gold and chromium. After evaporation, the samples are removed from the vacuum chamber. Wood’s alloy is applied to the backside of the silicon substrate to create the bottom electrode of the device. The fabricated device is now ready for measurements. Figure 1 illustrates the cross-sectional view of the devices after each fabrication step.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig1_HTML.gif
Fig. 1

Cross-sectional views of the organic on inorganic semiconductor heterojunction device after different fabrication steps. (a) Si substrate after oxidation; (b) active region window opening by the first mask; (c) spin-coating of polymer; and (d) thermal evaporation of top electrode by the second mask

2.2 Measurement setup

Typical active area of the hybrid organic on inorganic semiconductor heterojunction devices was 0.2 cm2. Current density–voltage (J–V) characteristics were obtained with a Hewlett-Packard model 4145A semiconductor parameter analyzer. The devices to be tested were mounted onto a holder stage with two probing needles connected to the top and bottom electrodes of the devices respectively. The voltage was scanned from  − 2.0 to 2.0 V in steps of 0.05 V.

The measurement setup for determining the capacitance–voltage (C–V) characteristic is very similar to that of the J–V measurements. The C–V measurements were carried out using a Hewlett-Packard multi-frequency HP model 4274A LCR meter. The dc bias voltage is monitored using an HP model 3478A multimeter. In the C–V measurements, the bias voltage was varied from  − 2.5 to 0 V at room temperature with a frequency of 100 kHz. All the devices were characterized in atmospheric ambient at room temperature under dark condition unless specified otherwise.

3 Results and discussion

3.1 Current density–voltage (J–V) characteristics

In this section, the electrical characteristics of the hybrid organic on inorganic semiconductor heterojunction devices with different organic polymers and top metal electrodes are discussed. A comparison of the parameters obtained from the J–V characteristics for different sets of heterojunctions were summarized in Table 1.
Table 1

Summary of the J–V characteristics for the organic on inorganic semiconductor heterojunctions

Organic polymer

MEHPPV POSS

RR P3HT

PEDOT

PEDOT:BF3

Anode contact

Al

Al

Au

Au

Rectification ratio (|1 V|)

6.4 × 10

1.1 × 103

1.2 × 104

5.1 × 104

Rectification ratio (|2 V|)

4.7 × 102

1.4 × 103

2.0 × 104

5.7 × 104

Saturation current density J0 (nA/cm2)

23

22

15

9.7

Leakage current density (μA/cm2)

31

2.8

0.44

0.29

Series resistance (Ω)

7.0 × 106

3.8 × 106

9.6 × 105

9.9 × 105

Ideality factor (n)

1.7

2.0

2.9

2.3

Almost all the organic on inorganic semiconductor heterojunctions regardless of the organic polymer or anode contact displayed a rectifying characteristic. From the various energy-band diagrams (Figs. 2, 3, 4) that have been constructed for the three sets of heterojunction devices in this work, it clearly shows that a potential barrier existed between the organic and inorganic semiconductor interface. During forward bias operation, electrons injected from the cathode into the n-Si will encounter a potential barrier between the inorganic semiconductor and the organic polymer. The electrons in forward bias will have to overcome this barrier in order to flow to the opposite terminal. On the other hand, holes injected from the anode will encounter a smaller energy barrier as shown in figures and therefore become the dominant transport carriers for these heterojunction devices. Thus, the height of this potential barrier will affect the performance of the devices.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig2_HTML.gif
Fig. 2

(Top) Energy-band diagram for MEH-PPV POSS on n-Si heterojunction devices before Fermi level alignment. (Bottom) Energy-band diagram for MEH-PPV POSS on n-Si heterojunction devices after Fermi level alignment

https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig3_HTML.gif
Fig. 3

(Top) Energy-band diagram for RR-P3HT on n-Si heterojunction devices before Fermi level alignment. (Bottom) Energy-band diagram for RR-P3HT on n-Si heterojunction devices after Fermi level alignment

https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig4_HTML.gif
Fig. 4

(Top) Energy-band diagram for PEDOT on n-Si heterojunction devices before Fermi level alignment. (Bottom) Energy-band diagram for PEDOT on n-Si heterojunction devices after Fermi level alignment

Investigation of the effect of the contact metal on the organic on inorganic semiconductor heterojunction properties was carried out. The aluminium (Al) anode contact was replaced by lower work function magnesium–silver (Mg–Ag) alloy. The J–V characteristic of the MEH-PPV POSS on n-Si heterojunctions with magnesium–silver (Mg–Ag) anode was measured. Basically, the heterojunctions with Mg–Ag or Al contact display a similar rectifying characteristic. The average rectification ratio RR evaluated at applied voltage of 1 V and 2 V are 45 and 150 respectively. Both values are smaller than when Al contact was used. The average magnitude for the saturation current density J0 and reverse leakage current density were determined to be 55 nA/cm2 and 1.5 ×  102 μ A/cm2. The average series resistance RS was calculated to be around 25 MΩ. The heterojunctions have an average n factor of 1.9. Comparing the results, it is evident that MEH-PPV POSS on n-Si heterojunction devices perform slightly better as a rectifying junction when Al anode was used. As the Al anode contact has a work function of around 4.28 eV and on the other hand, Mg–Ag has a work function of around 3.7 eV [6]. In Fig. 2, the use of the low-work function metal contributes a high energy offset between the Fermi energy at the anode and the HOMO level in the MEH-PPV POSS (which is 5.3 eV [7]) to result in an increased series resistance. This effectively reduces the number of injected holes at the anode and thus the performance of the devices was affected. In addition, as this energy barrier existed between the metal and organic polymer interface, the charge particles have a probability to tunnel through it. A more suitable metal for the anode contact could be Gold (Au) since it has a work function of 5.1 eV, which will match closely with the HOMO of the organic polymer MEH-PPV POSS.

Another interesting property of the heterojunction devices that has been investigated was the sensitivity of the devices to optical illumination. The illuminated J–V characteristics were taken at room temperature under Oriel tungsten lamp with an optical power density of 100 mW/cm2. Optical sensitivity measurements were performed on some of the MEH-PPV POSS on n-Si heterojunctions and the I–V characteristics for the devices carried out in dark or under illuminated source showed little or no difference. The J–V characteristics comparison was shown in Fig. 5. It is not conclusive whether the organic on inorganic semiconductor heterojunctions are insensitive to light or there exists a high series resistance at the interface between the organic polymer and the anode, which prevents the resulted small photocurrent to be measured accurately. Therefore, it cannot be concluded whether the organic on inorganic semiconductor heterojunctions in this work are sensitive to illumination or not.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig5_HTML.gif
Fig. 5

Forward ln(J)–V characteristics for MEH-PPV POSS on n-Si heterojunctions with Al contact in dark and under illumination

Similar to MEH-PPV polymer, RR-P3HT has a reported HOMO of 5.2 eV while work function of the Al anode is 4.28 eV. Therefore, there exists an energy-level offset at the contact and polymer interface which will affect the hole-injection efficiency and thus performance of the devices. Furthermore, known disadvantage of the P3HT polymers such as sensitivity to oxygen and moisture [8] could lead to the degradation observed from some of the devices as the measurement were carried out in air at room temperature over a period of time.

When the organic polymer PEDOT was doped (30%) with a p-type dopant boron trifluoride (BF3), the heterojunctions displayed a better performance than those with undoped PEDOT. In Fig. 6, an increase in the magnitude of the output current density for PEDOT:BF3 on n-Si heterojunctions by 2∼3 times was observed.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig6_HTML.gif
Fig. 6

Comparison of the J–V characteristics for PEDOT on n-Si heterojunctions and PEDOT:BF3 on n-Si heterojunctions

Among the various heterojunctions, the PEDOT:BF3 on n-Si heterojunctions with Au contact seem to have the best performance with high rectification ratio, small saturation current density and leakage current density. This could be accounted for by the lowest potential barrier among the three heterojunctions and almost little or no energy level offset between the organic polymer and gold anode. However the ideality factor calculated was slightly larger compared to the MEH-PPV POSS or RR-P3HT on n-Si heterojunction devices. All the devices have fairly large series resistances in the range of MΩ.

3.2 Capacitance–voltage (C–V) characteristics

For organic polymer on inorganic semiconductor heterojunctions, measurements of the differential capacitance–voltage can provide knowledge about the fixed charge concentration and built-in voltage. Any variation of the charge within a p–n diode with an applied voltage variation yields a capacitance which must be added to the circuit model of a p–n diode. The junction capacitance dominates for the reversed-biased diodes, while the diffusion capacitance dominates in strongly forward-biased diodes. Figure 7 illustrates the variation of the differential junction capacitance with the bias voltage at a frequency of 100 kHz for two sets of 3 PEDOT on n-Si heterojunction devices and 3 PEDOT:BF3 on n-Si heterojunction devices respectively.

From Fig. 7, the built-in voltage, Vbi can be obtained by extrapolating the curve to the point where 1/C2 = 0. Assume a one-sided junction model, the 1/C2 vs. V relationship can be simplified to so that the slope of the curve is inversely proportional to the doping concentration of the lightly doped region in the junction. The large reverse bias region or the so called deep region is considered for the calculation of these parameters as the curves are more linear due to the less pronounced effect of the interface states in the deep region.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig7_HTML.gif
Fig. 7

1/C2 vs. voltage of PEDOT and PEDOT:BF3 on n-Si heterojunction devices

For PEDOT on n-Si heterojunction devices, the average built-in voltage, Vbi was calculated to be approximately 1.4 V while the doping concentration, ND, of the lightly doped n-type silicon was roughly 6.1 × 1015 cm−3. On the other hand, the average Vbi for PEDOT:BF3 on n-Si heterojunction devices was slightly higher, 1.7 V while the doping concentration was 4.3 × 1015 cm−3. If the lightly doped n-region becomes depleted at a high-applied reverse-bias, it is possible to obtain the width of the n-region, W (cm), from its capacitance, C (F/cm2), using the following equation:
$$ W=\frac{\varepsilon _S \cdot A}{C} $$
where A is area of the active region (cm−2) and ɛS is the substrate permittivity. With the doping concentration and the depletion layer width, it is possible to obtain the doping profile of the lightly doped region in the n-type silicon. Figure 8 shows the approximate doping profile for a PEDOT on n-Si heterojunction device.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-006-9038-y/MediaObjects/10854_2006_9038_Fig8_HTML.gif
Fig. 8

Doping profile for a PEDOT on n-Si heterojunction

4 Conclusions and discussion

Organic on inorganic semiconductor heterojunction devices were fabricated with a sandwich structure on n-type silicon wafers. Electrical characterizations (J–V and C–V) of three sets of organic on inorganic semiconductor heterojunctions devices were carried out. The heterojunction devices include MEH-PPV POSS on n-Si, RR-P3HT on n-Si and PEDOT on n-Si. All the three organic on inorganic semiconductor heterojunctions regardless of the conjugated polymer or anode contact displayed a rectifying characteristic. From the energy-band diagrams constructed for the three heterojunctions, there is a potential barrier existed at the organic and inorganic semiconductor interface. The potential barrier for the electrons in forward bias is determined by the energy difference between the LUMO of the conjugated polymer and the electron affinity of the n-Si. The performance of the devices appeared to be inversely proportional to the height of the barrier. The PEDOT:BF3 on n-Si heterojunctions exhibit the highest rectification ratio. This is believed to be the first time that PEDOT doped with BF3 on n-Si heterojunction devices have been reported. Doping of PEDOT polymer with BF3 p-type dopant has improved the performance of the corresponding heterojunctions and an increase in the current magnitude by 2∼3 times was observed.

From the present work, the organic on inorganic semiconductor heterojunction device with the best performance was obtained by PEDOT:BF3 on n-Si with gold anode. These hybrid organic on inorganic heterojunctions still have not been studied extensively in this work. For example, annealing the devices at different temperatures may lead to a better organic–inorganic interface, resulting in an improvement to the device electrical characteristics. In addition, it is evident that the series resistance greatly affects the device performance. Therefore, it is important to study the contact resistance between different metal anodes and the conjugated polymers and hopefully reduce this contact resistance so that it does not limit the performance of the devices. Finally, more detailed information on the material characterisation of the organic polymers such as thickness, interface composition, and mobility will greatly improve the performance of the organic on inorganic semiconductor heterojunction devices.

Copyright information

© Springer Science+Business Media, LLC 2006