Near-Infrared Organic Photodiodes
Organic photodiodes (OPDs) are attractive as solution-processed devices for sensing applications. Industrial and medical sensors often have the requirement to operate in the near-infrared (NIR) spectrum between 650 and 900 nm and are ideally visible-blind. Due to the tailored spectral sensitivity of the organic semiconductors, OPDs are attractive as filter-free solidstate alternative. In addition, the large active areas of the OPDs potentially allow fabricating lens-free light-barrier and reflective sensors. In this paper, we discuss different approaches toward NIR sensitive OPDs with a large active area up to 1 cm/sup2sup/ applying polymers and small molecules as light absorbers. We demonstrate that with layer stacks optimized to the solution-processed semiconductor properties photodiodes with bulk heterojunctions with a minimum external quantum efficiency peak >40% in the NIR and a rectification ratio of ∼10/sup5sup/ can be achieved, which match industrial sensing requirements.
I. INTRODUCTION
SOLUTION-processed OPDs with NIR sensitivity are an attractive replacement for solid-state photosensors due to the cost-effective processing. Si or InGaAs sensors are usually applied to cover the industrially-relevant spectral region between ∼650 nm and ∼900 nm wavelengths beyond the spectral sensitivity of the human eye. Unfortunately, these solid-state devices show a broad spectral response in both the visible and the NIR range which requires filtering. For many applications such as light barriers or reflective sensors only the NIR sensitivity is needed since photodetectors are usually used in ambient condition with visible light superimposed to the NIR signal. For such applications the dark current has less relevance but the spectral sensitivity is usually required to be >600 nm wavelength. Furthermore, the high rectification ratio of diodes rather than the photoconducting behaviour is favourable for low power dissipating devices and improves charge extraction due to the built-in voltage [1], [2]. Stateof-the-art solid-state solutions often require a visible light filter to reduce the spectral sensitivity in the visible. Using organic photodiodes instead of solid-state solutions is potentially beneficial due to the tunable absorption properties of the organic materials. In addition, OPDs are of interest because of the solution processing possibilities on large arbitrary-shaped active areas of several cm2, even on flexible substrates at low processing temperatures [2]. The general OPD stack includes an anode, an interlayer (IL) as electron blocker, a semiconductor layer and a cathode (Figure 1a). The organic semiconductor layer is usually a BHJ with polymer/fullerene composites as donor/acceptor system that provides interfaces for exciton dissociation sandwiched between electrodes with different work functions for efficient charge extraction. In general, the light absorbing layer is part of the BHJ and the photo-generated carriers are collected at the reverse biased electrodes. Electrons are collected at the low work function electrode and holes are transported to the high work function electrode. A very common material for organic photovoltaics and photodiodes is the poly(3-hexylthiophene):[6], [6]-phenyl C61 butyric acid methyl ester (P3HT:PCBM) blend [3]. Previous reports have demonstrated that OPDs based on P3HT:PCBM blend can reach low dark current densities of 10−4 mA cm−2, rectification ratios as high as ∼105 and EQEs higher than 70 % in the visible range [4] with the EQE as the ratio of the number of charge carriers collected at the electrodes to the number of incident photons of a given energy shining the device. For light barrier and reflective sensor applications similar NIR diode characteristics are of interest. Unfortunately, the sensitivity of P3HT:PCBM thin film diodes is limited to the visible spectrum.
There are different strategies to obtain NIR sensitivity with organic materials. First, a small band gap material absorber can be applied for the BHJ replacing the common visible absorber P3HT (band gap of 1.9–2.0 eV). The rich chemistry of semiconducting polymers offers powerful methods to tune the HOMO and LUMO levels and modify the band gap of the material. A proven strategy to reduce the band gap of an organic π-conjugated material consists of incorporating electron-rich and electron-deficient units in an alternating fashion in a polymer chain. This approach allows synthesizing absorbers with band gaps in the 1.2-1.6 eV range [5]–[7]. Recent progress in the field of organic photovoltaics report a variety of potential polymers suitable for NIR detections such as poly[2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta [2,1-b;3,4-b] dithiophene)-alt-4,7-(2,1,3-benzothiadiazole)] (PCPDTBT) [8]–[10], polythieno[3,4-b]thiophene (PTT) [11], poly{5,7-bis [3,4-di(2-ethylhexyloxy)2-thienyl]-2,3-diphenylthieno [3,4-b]pyrazine} (PBEHTT) [12], poly{5,7-di-2-thienyl-2,3-bis(3,5di (2-ethylhexyloxy)phenyl)-thieno[3,4-b] pyrazine} (PTBEHT) [12] and LBPP-1 [13]. Encouraging results for long wavelength absorption (λ = 800-1000 nm) are obtained with PTBEHT and by LBPP-1 as absorbers in BHJs with diodes showing EQEs of ∼18 % [12] and ∼10 %, respectively [13]. Furthermore, PCPDTBT and PTT exhibit competitive EQEs of ∼35 % [8]–[10] and ∼44 %, respectively [11]. Second, the absorbing propriety of the P3HT:PCBM blend can be tuned for NIR sensitivity. It has recently been demonstrated that a thick layer (> 500 nm) of P3HT:PCBM can absorb up to λ = ∼750 nm [14]. Last, the p-type polymeric material in the BHJ can be replaced with a low molecular weight absorber like squaraine derivatives [15]–[18]. In this letter we compare NIR photodiodes with different absorber materials and we discuss for each approach the respective advantages and disadvantages. P3HT, PCPDTBT and Diguajazulenesquaraine (DaSQ) [19] are applied as model absorbers to demonstrate how to extend and to modulate the organic diode bandwidth in the NIR range. Figure 1b shows the molecular structure of the absorbers presented in this study. P3HT is a well-studied hole transporting material showing a good solubility in many organic solvents at ambient conditions. The P3HT used in this study has molecular weight (MW) of ∼41000 g mol−1 and regioregularity of ∼93 %. Reported P3HT hole mobilities at room temperature are in the range of 10−5–10−2 cm2 V−1s−1 with neglectable electron mobility values [20], [21]. The polymer shows an absorption peak for visible wavelengths corresponding to the π − π* absorption band of the molecule. The P3HT can be used either as light absorber or as IL. Reported P3HT HOMO and LUMO levels are ∼5.1 eV and ∼3.2 eV, respectively [22]. However, the energy levels depend on several properties, such as the film crystallinity [23]. Broad spectral absorption from the ultraviolet to ∼900 nm wavelength is reported for low band gap molecule PCPDTBT. The absorption peak from a pristine PCPDTBT film is ∼775 nm and shows a bathochromic shift when the polymer is in solution or after an annealing step [8]. The PCPDTBT band gap is estimated to be ∼1.4 eV [6], [8], [24] with reported energy levels of ∼3.5 eV for the LUMO and ∼4.9 eV for the HOMO [10], [9], [25], [26]. For small molecule absorber, DaSQ is applied with a low molecular weight of 474 g mol−1. DaSQ HOMO and LUMO are estimated to be ∼4.9 eV and ∼3.3 eV, respectively, [17] and the absorber shows a high decomposition temperature of ∼245 °C [19]. Compared to polymeric absorbers, small molecule materials have a distinct advantage with respect to the tedious polymer purification process, which is critical for the price of the semiconductor. In general, we observe that with higher purity of the absorber material, OPD characteristics, such as dark current, improve dramatically. Figure 1c shows the flat band diagram of the studied stacks. For all the BHJ the n-type material PCBM is used. PCBM HOMO (6.1 eV) and LUMO (3.7 eV) are taken from literature [27], [28]. To visualize the spectral region we are interested in, Figure 2 shows the normalized spectral sensitivities of OPDs with P3HT and a low bandgap polymer PCPDTBT as light absorbers.
A thin layer of P3HT:PCBM (∼500 nm thick) absorbs mainly in the visible range between ∼350 nm and ∼650 nm. Replacing the light absorber with a low-band gap material the diode spectral bandwidth extends to the range of interest. The EQE of the device with PCPDTBT:PCBM shows two absorption peaks, one in the NIR and one in the ultraviolet (UV). The absorption of the PCPDTBT:PCBM in the NIR between ∼700 nm and ∼850 nm is a result of the interband π −π* transition of the PCPDTBT [10], [25] while the PCBM transition is in the UV [25]. The green area in Figure 2 represents the relevant spectral range for industrial NIR sensing between ∼650 nm to ∼900 nm.
II. NIR SENSITIVITY WITH LOW BAND GAP POLYMER PCPDTBT
Our first approach to NIR OPDs uses a low-band gap material as P3HT substitute. Reports have shown that a BHJ composite concept can be equally applied with a low-bandgap semiconducting polymer such as PCPDTBT, since it is generally assumed that a small offset between the LUMO levels of the electron and acceptor material is required for efficient photoinduced charge transfer [6], [8]. With PCPDTBT:PCBM composites, electron acceptors seem to be necessary to provide a balanced and efficient charge transport in the films [26]. In particular it has been demonstrated that the photocurrent increases by about two orders of magnitude when PCBM is added to PCPDTBT [9]. According to [8], the optimum lifetime of photogenerated carriers is observed at the weight ratio of ∼1:3. This result is confirmed by solar power conversion efficiency measurements [10]. In Figure 3a normalized absorption measurements on PCPDTBT:PCBM films with ratio from 1:1 to 1:3 are presented.
Pristine PCPDTBT and PCBM are added as references. PCPDTBT shows two NIR absorption peaks at ∼710 nm and ∼750 nm and a visible peak at ∼415 nm wavelengths in agreement with literature [8] while PCBM has a low absorption in the NIR range. In the PCPDTBT:PCBM blend one dominating peak in the NIR is observed. A small red shift of the peak is visible with increasing PCBM content [8]. To confirm the literature data we fabricated different OPDs by spray-coating with PCPDTBT:PCBM as active material with different weight ratios. The OPD layer stack is ITO/PEDOT:PSS/PCPDTBT:PCBM/Ca/Ag. Currentvoltage characteristics of the diodes show that with a 1:1 ratio of PCPDTBT:PCBM the OPD exhibits moderate dark current densities of ∼10−3 mA cm−2 at −5 V reverse bias (not shown). With increasing PCBM content a significantly higher photoresponse is observed. At the same time increasing PCBM content significantly reduces the serial resistivity of the diode. In addition, the NIR illumination further decreases the resistivity due to the photoconductance of the film. High PCBM contents blend in the range from 1:2.5 to 1:3 show comparable EQE characteristics in the NIR of ∼45 % at −5 V reverse bias (Figure 3b). EQE measurements on OPDs with PCPDTBT:PCBM as active material show spectral sensitivities up to ∼1000 nm. However, the spectral sensitivity covers a broad portion of the visible spectrum, which is not ideal for our NIR sensor application. Visible wavelength filtering can be achieved with an external low-pass filter in series with the photodiode. However, more intriguing is to integrate the visible wavelength filter as an intrinsic element of the OPD layer stack. In Figure 4 we show the influence of the interlayer material on the OPD spectral and the current – voltage (IV) characteristic.
As a reference Figure 4a shows the IV of an OPD with spin-coated PEDOT:PSS IL (black line) and spraycoated PCPDTBT:PCBM BHJ (weight ratio 1:2.5) in the dark and under polychromatic illumination (>870 nm). The diode shows favorable properties such as dark current densities of ∼10−4 mA cm−2 at −5 V reverse bias with a rectification ratio at +/− 2 V of ∼4 × 105 and a series resistance of ∼1 k measured at +2 V. Compare to the reference diode with PEDOT:PSS IL the OPDs with spray-coated P3HT IL match the dark and photocurrent properties though series resistance rises. A sufficiently thick P3HT IL reduces the dark current as being reported for the P3HT:PCBM blend [29], [30] which allows to achieve dark current densities of ∼10−5 mA cm−2 at −5 V reverse bias. In Figure 4b the spectral characteristic of the diodes are compared. The P3HT IL filters the visible part of the spectrum. For the reference diode an EQE maximum at −5 V reverse bias (Figure 4a, black line) of ∼44 % is observed at λ = 700 nm while the EQE minimum is for the green wavelength λ = 520 nm of ∼32 %. The EQE at 520 nm reduces from 32 % (PEDOT:PSS IL) to 28 % (240 nm P3HT IL) and to 5.3 % (520 nm P3HT IL). The EQE minimum shows a red shift from λ = 520 nm (PEDOT IL) to λ = 540 nm (240 nm P3HT IL) to λ = 550 nm wavelength (52 nm P3HT IL). Reduction of the EQE amplitude in the NIR is expected to be related to the total absorber thickness of both BHJ + IL which leads to an increased carrie’s recombination. NIR EQE at λ = 750 nm drops from ∼43 % (PEDOT:PSS IL) to ∼32 % (24 nm P3HT IL) to ∼25.3 % (520 nm P3HT IL). With P3HT IL we achieve good NIR sensitivity but some residual sensitivity for wavelengths < 400 nm remains that cannot be filtered by the glass substrate or encapsulation. We observe also a blue-shifting peak at the range from 600 to 700 nm which probably correlates with the P3HT IL thickness. To understand the reason for the EQE peak of the NIR diodes between 600 to 700 nm we remove the PCBM from the BHJ fabricating a diode with P3HT only. The diode stack is the following: ITO/PEDOT:PSS/P3HT/Al. P3HT is spray-coated from a xylene solution with a weight ratio of 1 % resulting in a mean film thickness of ∼1 μm. Fig. 5a shows the measured IV characteristic. In the dark the device behaves as an open circuit with a dark current density of ∼10−5 mA cm−2 for both direct and reverse polarizations which is expected for a photoconductor. Under illumination at λ = 530 nm the OPD is a weakly rectifying diode because of the photodoping at the Schottky contacts. More interestingly Figure 5b shows the EQE of the diode with only P3HT measured for different voltages ranging from 0 V to −5 V reverse bias.
The device shows a sharp EQE peak centred at ∼640 nm (photon energy of ∼1.93 eV). Such peak can clearly be attributed to the P3HT layer, which confirms our interpretation of Figure 4 above. The EQE of the OPD with P3HT IL is indeed a superposition of the absorption of the independent layers. This result indicates that only a low intermixing occurs in between the IL and the BHJ as a result of the spray-coating process. The EQE in Figure 4b can be interpreted as a superposition of the EQE of both the BHJ and the P3HT IL. The red shift of the EQE peak is discussed in Figure 8.
III. NIR SENSITIVITY WITH P3HT AS VISIBLE ABSORBER
Furthermore, we show that NIR sensitivity of the diodes can also be obtained with the P3HT:PCBM blend only (Figure 6 to 10). With one single spray coating step a maximum of ∼500 nm thick P3HT:PCBM BHJ layer is fabricated due to the solubility limitations of the BHJ materials. For thicker OPDs multiple spray steps are required. Figure 6a shows microscopy images of a spray-coated P3HT:PCBM (1:0.75 wt.) blend with different mean thicknesses ranging from ∼500 nm to ∼70 μm, the last requiring ∼150 spraycoating steps.
The BHJ roughness increases significantly with increasing layer thickness. Due to the absorption, the BHJ color gradually changes from pink to dark-brown, as shown in Figure 6b for thicknesses up to ∼2 μm. We observe a moderate roughness until a film thickness of ∼5 μm (Figure 7a).
Even thicker films show a very pronounced increase of the film roughness due to agglomeration although we controlled that after each spray-coating step the deposited material was dried. In Figure 7b the BHJ transmittance (T %) vs. BHJ layer thickness is presented. The spectra exhibit the characteristic π–π* band due to the electronic transitions in P3HT (maximum at λ = ∼500 nm) [32]. The shoulder near 2.1 eV is due to interchain excitons, resulting from π-stacking of the polymer chains in lamllar aggregates [31], [33]. The peak at 710 nm and the absorption of the blend at higher photon energy (λ < 400 nm) is due to the PCBM, as evident comparing the spectra in Figure 7b (P3HT:PCBM blend) with the spectra in Figure 7d (pristine P3HT). Supplementary Figure 1 shows additional transmission measurements on pristine PCBM films with varying thickness from ∼70 nm to ∼7 μm to confirm the nature of the stable peak at ∼710 nm. As reported by [14] the PCBM absorption shows a red-shift with increasing thickness. With ∼500 nm BHJ thickness the P3HT:PCBM blend absorbs >95 % of the incident visible light. Important for our discussion of NIR diodes is the pronounced absorption in the NIR with layer thicknesses above ∼1 μm. Photons with a wavelength of 550 nm are absorbed in the first ∼500 nm of the film while with increasing wavelength the penetration depth increases to several tens of μm for excitation at 750 nm. Photons with an energy 10 μm (Figure 7c). This observation fits to the Lambert-Beer law. Here, we apply the P3HT:PCBM absorption properties of the thick films for NIR sensing. For thick OPDs fabrication we use PEDOT:PSS as IL and P3HT:PCBM (1:0.75) as active material. The BHJ is prepared from a xylene solution and is spray-coated onto the IL with a mean thickness ranging from ∼500 nm to ∼36 μm. Al is thermally evaporated as cathode (see experimental section for details). Independently from the film roughness as result of the spray-coating process, the diode characteristics show comparable low dark current densities of ∼10−5 mA cm−2 at −5 V reverse bias (Figure 8a).
Photocurrents are measured with illumination through the bottom electrode at 532 nm with an irradiation intensity of ∼780 μW cm−2. The series resistance measured in the dark at +2 V gradually increases from 1.1 k for the 500 nm thin diode to ∼8.4 M for the 36 μm thick OPD. At the same time with increasing BHJ thickness, the rectification ratio decreases due to the increasing series resistance (Figure 8b). With increasing active layer thickness the device characteristic is dominated by photoconductance. Note that the photodoping effect at forward bias is relatively low for all diode thicknesses. In supplementary Figure 2 one can observe that due to the mobility the diode cut-off frequency decreases exponentially with the film thickness. Figure 8c shows photocurrents measurements on the same diodes with a polychromatic NIR light >870 nm. The photocurrent with 4 μm show NIR sensitivity only. Results are in good agreement with reported spectral sensitivities [14], [34]. In 2010 Chen et al. [14] reported a similar spectral red shift on dropcasted diodes indicating that the spectral shift is a material property not related to the OPD fabrication process. Yang et al. [34] observed that such NIR absorption exists only in the blend and discussed this as a charge transfer exciton phenomena. Similar results were reported by [35] and [36]. EQE peak at λ = ∼640 nm suggests that in thin OPDs charge transport is related to both PCBM and P3HT [3] while with thick diodes the P3HT transport dominates (cf. Figure 9, 4b, 5b and 8e).
The transition indicates that recombination losses lead to an in-built filtering of the visible spectrum. Supplementary Table 1 correlates the OPD BHJ mean thickness with the wavelength corresponding to the maximum EQE and the transmittance at the same wavelength. Increasing the BHJ thickness the maximum EQE amplitude decreases while the corresponding wavelength shows a right shift from ∼530 nm to ∼730 nm. At the same time the BHJ transmittance for the same wavelengths decreases with increasing BHJ thickness. Assuming negligible reflection losses at the anode interface, we can derive the losses by recombination mechanism subtracting the transmittance and the EQE from the total amount of light applied to the device, so 100 (%) – T (%) – EQE (%). The calculated losses are shown in Table 1. A trend can be identified, with the losses increasing with increasing the BHJ thickness. It is important to note that charge carrier recombination increases with the BHJ thickness or distance that the charges need to travel to their respective electrodes. When the BHJ thickness is higher than the photon penetration deep, which is true for >2 eV photons, excitons are generated in the first layers of the film while the remaining BHJ behaves as a pure resistor were recombination reduces the diode current. Low energy photons (2 eV photons, with lower recombination probability resulting in an increased photocurrent. Last, Figure 8f shows the correlation between peak wavelength and peak EQE vs. BHJ thickness. The EQE peak rolls off from 73 % for a ∼500 nm thin diode to 17.5 % for a 36 μm thick BHJ OPD. At the same time the EQE peak shifts from λ = 520 nm to the NIR with λ = ∼730 nm (Figure 8f, black line). Similar EQE shifs can be observed on the diode EQEs measured from 0 V to −5 V reverse bias (Figure 10). We observe that the spectral sensitivity is not influenced by the internal electric field but the charge extraction efficiency is increasing, except for the diode with a thickness of 1.6 μm. With an OPD in Figure 10b having a mean BHJ thickness of 1.6 μm the transition between a broadband sensitivity (visible and NIR) to an only NIR-sensitive device is a function of the applied reverse voltage. The charge extraction of charge carriers generated by visible light is strongly influenced by the applied electric field. Contrary to the optical penetration length where the average film thickness is relevant, to compare dark current characteristics of diodes with different thickness, values need to be referred to the minimum spray-coated film thickness (minimum electrodes distance). In Supplementary Figure 3 a correlation between the minimum continuous layer thickness vs. mean BHJ thickness is presented with two different linear interpolations. Difference between mean and closed layer thickness increases with increasing BHJ thickness. Results show that with increasing thickness the average thickness is relevant for dark currents. In Supplementary Figure 3b the IVs as function of E are plotted for the minimum continuous BHJ layer. The same trend for the dark current densities with the BHJ thickness is observed as discussed for Figure 8d but now the values correspond to the true dark current densities. Finally, we investigate with post annealing of the thicker OPD (36 μm) how the diode NIR sensitivity is influenced by the BHJ morphology. We performed annealing steps of 10 minutes each between 80 °C to 180 °C (Figure 11a). The temperature dependence shows a weak EQE reduction trend and a small red shift of the EQE peak with increasing temperature up to 120 °C. For temperatures higher than the P3HT glass transition temperature (Tg of ∼140 °C) [38] the EQE rises. Note that within a stack of PEDOT:PSS and BHJ the lowest glass transition temperature is with P3HT [37]–[40]. Thermal annealing effects has been linked to P3HT reordering and improvement of nanoscale morphology despite the precise mechanism by which it improves charge photogeneration remains unclear [41]. In particular, it has been reported that heating of P3HT:PCBM films to temperatures above the P3HT glass transition temperature facilitates phase separation which then allows the enlarged P3HT domains to crystallize [42]. Figure 11b shows the results of the same annealing process performed on the 1.6 μm thin diode. Similarly to the thick diode with 36 μm an increase of the EQE is observed beyond the P3HT glass transition temperature (>140 °C). Results indicate that the spectral sensitivities of the OPD (visible or NIR bandwidth) are not depending on the morphology of the BHJ. This result is consistent with reported values for drop-casted devices [14], which have a different morphology compared to the spray-coated BHJ but show the same spectral properties. Interesting for industrial application is the high stability of the BHJ diodes up to ∼200 °C. Because we do observe such a high spectral stability close to the P3HT melting point of ∼210 °C [39], [40] and knowing that a distinct phase separation occurs between PCBM and P3HT for T >140 °C, we are not convinced that the NIR sensitivity is caused only by a charge transfer exciton at the molecular heterojunction between P3HT and PCBM.
IV. NIR SENSITIVITY WITH SMALL MOLECULE DASQ
Finally, we show that NIR sensitivity can be obtained also with small molecules as light absorbers like squaraine derivatives that have demonstrated to be good substitutes for P3HT in BHJs [15]–[18]. Squaraines are small and stable molecules with a unique structure that combine an aromatic four membered ring system – the squaric acid – and the betaine with a dipolar ionic molecule containing both positive and negative charge centers. The basic structure of the squaraines is also the reason for their chromaticity. The central four membered ring is aromatic, containing 2π-electrons, allowing an electron delocalization within the whole planar molecule. In Figure 12a the at least five mesomeric or resonance structures of the squaraine are shown. The substituents A1 and A2 are in general arylic components with a high electron density (electron donors) pushing electrons into the electron-deficient four membered ring. The resulting squaraine molecule is a donor-acceptordonor-system with strong intramolecular charge transfer effects. As P3HT substituents we tested several arylic components. Guaiazulene (1,4-dimethyl-7-isopropylazulene) showed the best properties for OPDs (Figure 12b and 12c). Figure 13a shows an image of golden-colored DaSQ crystals. The solution in chloroform has a dark blue-green colour and NIR absorption spectra with a maximum at ∼770 nm (Figure 13b). A minor solvatochromy, a shift dependent from the solvent is observed (data not shown). The low synthesis and purification effort as compared to P3HT or PCPDTBT is an attractive property of such a class of small molecule absorbers. These material properties of the squaraine are in particular of industrial interest for up-scaling. Furthermore the higher solubility of the squaraine derivatives as compared to polymers allows increasing the concentration and shelf life of the BHJ solution. The OPD is fabricated with a spin-coated PEDOT:PSS IL, a ∼200 nm doctor-bladed DaSQ:PCBM blend from a chloroform solution (1:2.5 wt.) and thermally evaporated Ca/Ag as cathode. The diode active area is 4 mm2. Figure 13c shows the measured IV characteristic in the dark and with polychromatic NIR >870 nm illumination (see Supplementary Figure 4 for green and white light). The device shows a diode behavior with a dark current density of ∼5 × 10−4 mA cm2 at −5 V reverse bias and a serial resistance of ∼90 ohm at +2 V. Furthermore, we observe a high rectification ratio at +/−2 V of ∼1.3 × 105. EQE measurements of the diode at reverse biases up to −10 V are presented in Figure 13d showing a peak sensitivity at ∼805 nm. At −10 V reverse bias diode EQE is ∼140 % which is assumed to be the result of a photogain effect. Compared to previous results for our NIR applications, DaSQ is favorable to P3HT or PCPDTBT as light absorbers because of the visible blind and high NIR sensitivity of the thin film diode, which does not require any additional filtering. In addition, we observe a high cut-off frequency of ∼100 kHz (Figure 13e) at −2 V reverse bias. Such device properties are of high interest for industrial applications covering the spectral regime of typical emitters in the range from 650 to 900 nm.
V. CONCLUSION
In this paper we provided an overview of approaches towards large active area NIR sensitive but visible blind organic photodetectors. We demonstrate that highly NIR sensitive devices can be achieved with polymeric as well as small molecule absorbers for BHJ but require that the diode stack is optimized with respect to the material properties. Application of such devices can be envisioned in light barrier applications or reflective sensors as an alternative to solid-state solutions.
VI. EXPERIMENTAL
DaSQ Synthesis: Guaiazulene, an azulene derivate, is a natural component of guaiac and chamomile oil, also used in skin care and medical products like anti-ulcer drugs [43]. The predominantly applied synthesis of symmetric squaraines is performed via a condensation reaction of squaric acid and the nucleophilic substituents. We applied for the synthesis of DaSQ the method of Ziegenbein et al. [19] with the slight modification of using a Dean-Stark apparatus for the continuous removal of the water that is produced during the condensation reaction. DaSQ (Figure 11) is a symmetric squaraine, with identical 2 substituents (A1 = A2). We clean the DaSQ by recrystallization with the following procedure. The DaSQ is dissolved in benzonitrile, heated up to 190 °C, boiled for a few minutes, filtered and cooled to room temperature. Then diethylether is added in excess as precipitant of the DaSQ, while impurities are either filtered or stay dissolved in the polar solvent (benzonitrile). After crystallization, the DaSQ is filtered, washed and dried. The yield is usually >95 %. The purity of the DaSQ is tested with thin film chromatography. When needed, the purification step is repeated.
Solutions PCPDTBT: PCBM from Solenne is first diluted in chlorobenzene (2 % wt.) and sonicated for ∼30 min at room temperature. PCPDTBT from Konarka is than added to the solution in a different weight ratio ranging from 1:3 to 1:1. The final solution is stirred on a hotplate at 80 °C for 1 hour before use. P3HT: P3HT from Rieke is diluted in xylene (1% wt.) and stirred at 80 °C for ∼1 hour before use. P3HT:PCBM: PCBM is first diluted in xylene (1 % wt.) and sonicated for ∼30 min. P3HT is than added to the solution in a weight ratio of P3HT:PCBM (1:0.75). The final solution is stirred on a hotplate at 80 °C for 1 hour before use. DaSQ:PCBM: DaSQ synthesis is done according to [19]. PCBM is diluted in chloroform (2.5 % wt.) and sonicated for ∼10 min at room temperature. DaSQ is diluted in chloroform (1 % wt.) and stirred for ∼1 hour at room temperature. When solutions are ready they are mixed resulting in a final concentration DaSQ:PCBM (1:2.5). Solution is deposited by doctorblade at room temperature. PEDOT:PSS from HC Starck is sonicated for 2 min, filtered with a Nylon 0.2 μm filter and spin-coated onto the ITO in ambient condition.
Device Fabrication: OPDs are fabricated on a 5 × 5 cm2 structured ITO coated glass. After cleaning the substrates in acetone, isopropanol and deionized water RIE plasma is applied to activate the ITO before the deposition of the electron blocking IL. PEDOT:PSS is spin-coated onto the ITO (∼150 nm) and baked in a vacuum oven at 200 °C for 15 min while P3HT IL is spray-coated on the ITO resulting in a film thickness of ∼240 nm and ∼520 nm. After deposition of the ILs the semiconductor (BHJ layer) is doctorbladed or spray-coated (aerosol particle size < 10 μm) in ambient condition. According to surface profilometry measurements, the doctorbladed film has a mean thickness of ∼200 nm while a single spray deposition result is a mean film thickness of ∼500 nm. After BHJ deposition all samples are annealed at 140 °C for 5 min in inert conditions before thermal evaporation of 4 nm/100 nm of Ca/Ag or 100 nm of Al as top electrode through a shadow mask. Overlap between anode and cathode define an active area of 1 cm2 or 4 mm2. Diodes are encapsulated with solvent-free epoxy glue and a 100 μm thick transparent glass slide. OPDs are finally annealed at 80 °C for 30 min on a hotplate for glue curing.
Light Characterization: IV characteristics of the diodes are recorded using a Keithley 6487 picoammeter in an electrically and optically shielded box. Photocurrents are measured with illumination through the transparent conductive anode electrode. For NIR illumination of the photodiodes, a white beam of an AM 1.5 Oriel solar simulator with a power density of ∼100 mW cm−2 and a GaAs filter with cut-on wavelength at 870 nm is used. Green light is generated by a green light-emitting diode (532 nm) with irradiation intensity of 780 μW cm−2. EQE spectrum is recorded with lock-in technique using a chopping frequency of 170 Hz and an Oriel Cornerstone 130 1/8 m monocromator. A Si photodiode is used as a reference diode for calibration. Dynamic measurements are made in an electrically and optically shielded box with a transimpedance amplifier (FEMTO DHPCA-100) con
REFERENCES
[1] T. N. Ng, W. S. Wong, M. L. Chabinyc, S. Sambandan, and R. A. Street, “Flexible image sensor array with bulk heterojunction organic photodiode,” Appl. Phys. Lett., vol. 92, no. 21, pp. 213303-1–213303-3, 2008.
[2] T. Rauch, M. Boberl, S. F. Tedde, J. Furst, M. V. Kovalenko, G. Hesser, et al., “Near-infrared imaging with quantum-dot-sensitized organic photodiodes,” Nature Photon., vol. 3, pp. 332–336, May 2009.
[3] N. S. Sariciftci, L. Smilowitz, A. J. Heeger, and F. Wudl, “Photoinduced electron transfer from a conducting polymer to buckminsterfullerene,” Science vol. 258, pp. 1474–1476, Nov. 1992.
[4] S. F. Tedde, J. Kern, T. Sterzl, J. Fürst, P. Lugli, and O. Hayden, “Fully spray coated organic photodiodes,” Nano Lett., vol. 9, no. 3, pp. 980–983, 2009.
[5] J. Roncali, “Synthetic principles for bandgap control in linear pi-conjugated systems,” Chem. Rev., vol. 97, no. 1, pp. 173–206, 1997.
[6] A. P. Zoombelt, M. Fonrodona, M. M. Wienk, A. B. Sieval, J. C. Hummelen, and R. A. J. Janssen, “Photovoltaic performance of an ultrasmall band gap polymer,” Org. Lett., vol. 11, no. 4, pp. 903–906, 2009.
[7] H. A. M. van Mullekom, J. A. J. M. Vekemans, E. E. Havinga, and E. W. Meijer, “Developments in the chemistry and band gap engineering of donor–acceptor substituted conjugated polymers,” Mater. Sci. Eng. R, Rep., vol. 32, no. 1, pp. 1–40, 2001.
[8] I.-W. Hwang, C. Soci, D. Moses, Z. Zhu, D. Waller, R. Gaudiana, et al., “Ultrafast electron transfer and decay dynamics in a small band gap bulk heterojunction material,” Adv. Mater., vol. 19, no. 17, pp. 2307–2312, 2007.
[9] C. Soci, I.-W. Hwang, D. Moses, Z. Zhu, D. Waller, R. Gaudiana, et al., “Photoconductivity of a low-bandgap conjugated polymer,” Adv. Funct. Mater., vol. 17, no. 4, pp. 632–636, 2007.
[10] D. Mühlbacher, M. Scharber, M. Morana, Z. Zhu, D. Waller, R. Gaudiana, et al., “High photovoltaic performance of a lowbandgap polymer,” Adv. Mater., vol. 18, no. 21, pp. 2884–2889, 2006.
[11] Y. Yao, Y. Liang, V. Shrotriya, S. Xiao, L. Yu, and Y. Yang, “Plastic nearinfrared photodetectors utilizing low band gap polymer,” Adv. Mater., vol. 19, no. 22, pp. 3979–3983, 2007.
[12] M. M. Wienk, M. G. R. Turbiez, M. P. Struijk, M. Fonrodona, and R. A. J. Janssen, “Low-band gap poly(di-2-thienylthienopyrazine): Fullerene solar cells,” Appl. Phys. Lett., vol. 88, no. 15, pp. 153511-1–153511-3, 2006.
[13] E. Perzon, F. Zhang, M. Andersson, W. Mammo, O. Inganäs, and M. R. Andersson, “A conjugated polymer for near infrared optoelectronic applications,” Adv. Mater., vol. 19, no. 20, pp. 3308–3311, 2007.
[14] E. Chen, C. Chang, J. Shieh, S. Tseng, H. Meng, C. Hsu, et al., “Polymer photodetector with voltage-adjustable photocurrent spectrum,” Appl. Phys. Lett., vol. 96, no. 4, pp. 043507-1–043507-3, 2010.
[15] A. Burke, L. Schmidt-Mende, S. Ito, and M. Grätzel, “A novel blue dye for near-IR ‘dye-sensitised’ solar cell applications,” Chem. Commun., vol. 2007, no. 3, pp. 234–236, 2006.
[16] M. Binda, T. Agostinelli, M. Caironi, D. Natali, M. Sampietro, L. Beverina, et al., “Fast and air stable near-infrared organic detector based on squaraine dyes,” Organic Electron., vol. 10, no. 7, pp. 1314–1319, 2009.
[17] M. Binda, A. Iacchetti, D. Natali, L. Beverina, M. Sassi, and M. Sampietro, “High detectivity squaraine-based near infrared photodetector with nA/cm2 dark current,” Appl. Phys. Lett., vol. 98, no. 7,
pp. 073303-1–073303-3, 2011.
[18] D. Bagnis, L. Beverina, H. Huang, F. Silvestri, Y. Yao, H. Yan, et al., “Substituent effects in squaraine-based Bulk Heterojunction solar cells,” J. Amer. Chem. Soc., vol. 132, pp. 4074–4075, Jan. 2010.
[19] W. Ziegenbein, H.-E. Sprenger, “Kondensationsprodukte aus Quadratsäure und Azulen-Kohlenwasserstoffen,” Angew. Chem., col. 78, no. 20, pp. 930–937, 1966.
[20] J. Nakamura, K. Murata, and K. Takahashi, “Relation between carrier mobility and cell performance in bulk heterojunction solar cells consisting of soluble polythiophene and fullerene derivatives,” Appl. Phys. Lett., vol. 87, no. 13, pp. 132105-1–132105-3, 2005.
[21] E. von Hauff, J. Parisi, and V. Dyakonov, “Investigations of the effects of tempering and composition dependence on charge carrier field effect mobilities in polymer and fullerene films and blends,” J. Appl. Phys., vol. 100, no. 4, pp. 043702-1–043702-7, 2006.
[22] J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T. Nguyen, M. Dante, et al., “Efficient tandem polymer solar cells fabricated by all-solution processing,” Science, vol. 317, no. 5835, pp. 222–225, 2007.
[23] T. J. Savenije, J. E. Kroeze, X. Yang, and J. Loos, “The formation of crystalline P3HT fibrils upon annealing of a PCBM:P3HT bulk heterojunction,” Thin Solid Films, vols. 511–512, pp. 2–6, Jul. 2006.
[24] G. Dennler, M. C. Scharber, and C. J. Brabec, “Polymer-fullerene BulkHeterojunction solar cells,” Adv. Mater., vol. 21, no. 13, pp. 1323–1338, 2009.
[25] M. Morana, M. Wegscheider, A. Bonanni, N. Kopidakis, S. Shaheen, and M. Scharber, et al., “Bipolar charge transport in PCPDTBT-PCBM Bulk-Heterojunctions for photovoltaic applications,” Adv. Funct. Mater., vol. 18, no. 12, pp. 1757–1766, 2008.
[26] V. D. Mihailetchi, P. W. M. Blom, J. C. Hummelen, and M. T. Rispens, “Cathode dependence of the open-circuit voltage of polymer: Fullerene Bulk heterojunction solar cells,” J. Appl. Phys., vol. 94, no. 10, pp. 6849-1–6849-6, 2003.
[27] L. Weimin, G. Jinchuan, S. Xiuquan, and Z. Bin, Proc. SPIE, vol. 6841, pp. 68411G-1–68411G-5, Mar. 2008.
[28] C. Liang, W. Su, and L. Wang, “Enhancing the photocurrent in poly(3-hexylthiophene)/[6,6]-phenyl C61 butyric acid methyl ester bulk heterojunction solar cells by using poly(3-hexylthiophene) as a buffer layer,” Appl. Phys. Lett., vol. 95, no. 13, pp. 133303-1–133303-3, 2009.
[29] J. Y. Oh, W. S. Jang, T. I. Lee, J. Myoung, and H. K. Baik, “Driving vertical phase separation in a bulk-heterojunction by inserting a poly(3-hexylthiophene) layer for highly efficient organic solar cells,” Appl. Phys. Lett., vol. 98, no. 2, pp. 023303-1–023303-3, Jan. 2011.
[30] J. Piris, T. E. Dykstra, A. A. Bakulin, P. H. M. van Loosdrecht, W. Knulst, M. T. Trinh, et al., “Photogeneration and ultrafast dynamics of excitons and charges in P3HT/PCBM blends,” J. Phys. Chem. C, vol. 113, no. 32, pp. 14500–14506, Jul. 2009.
[31] H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K. Bechgaard, B. M. W. Langeveld-Voss, et al., “Two-dimensional charge transport in self-organized, high-mobility conjugated polymers,” Nature, vol. 401, pp. 685–688, Oct. 1999.
[32] P. J. Brown, D. S. Thomas, A. Köhler, J. S. Wilson, J.-S. Kim, C. M. Ramsdale, et al., “Effect of interchain interactions on the absorption and emission of poly(3-hexylthiophene),” Phys. Rev. B, vol. 67, no. 6, pp. 064203-1–064203-16, Feb. 2003.
[33] C. Yang, P. Tsai, S. Horng, K. Lee, S. Tzeng, H. Meng, et al., “Infrared photocurrent response of charge-transfer exciton in polymer bulk heterojunction,” Appl. Phys. Lett., vol. 92, no. 8, pp. 083504-1–083504-3, Feb. 2008.
[34] W. Osikowicz, M. de Jong, and W. Salaneck, “Formation of the interfacial dipole at organic-organic interfaces: C60/polymer interfaces,” Adv. Mater., vol. 19, no. 23, pp. 4213–4217, Dec. 2007.
[35] P. Parkinson, J. Lloyd-Hughes, M. B. Johnston, and L. M. Herz, “Efficient generation of charges via below-gap photoexcitation of polymer-fullerene blend films investigated by terahertz spectroscopy,” Phys. Rev. B, vol. 78, no. 11, pp. 115321-1–115321-11, Sep. 2008.
[36] T. T. Ngo, D. N. Nguyen, and V. T. Nguyen, “Glass transition of PCBM, P3HT and their blends in quenched state,” Adv. Nat. Sci. Nanosci. Nanotechnol., vol. 3, no. 4, pp. 045001-1–045001-3, Dec. 2012.
[37] B. Friedel, P. E. Keivanidis, T. J. K. Brenner, A. Abrusci, C. R. McNeill, R. H. Friend, et al., “Effects of layer thickness and annealing of PEDOT:PSS layers in organic photodetectors,” Macromolecules, vol. 42, no. 17, pp. 6741–6747, 2009.
[38] D. H. Kim, Y. Jang, Y. D. Park, and K. Cho, “Layered molecular ordering of self-organized poly(3-hexylthiophene) thin films on hydrophobized surfaces,” Macromolecules, vol. 39, no. 17, pp. 5843–5847, Jul. 2006.
[39] E. Klimov, W. Li, X. Yang, G. G. Hoffmann, and J. Loos, “Scanning near-field and confocal raman microscopic investigation of P3HTPCBM systems for solar cell applications,” Macromolecules, vol. 39, no. 13, pp. 4493–4496, May 2006.
[40] R. A. Marsh, J. M. Hodgkiss, S. Albert-Seifried, and R. H. Friend, “Effect of annealing on P3HT:PCBM charge transfer and nanoscale morphology probed by ultrafast spectroscopy,” Nano Lett., vol. 10, no. 3, pp. 923–930, Feb. 2010.
[41] R. M. Beal, A. Stavrinadis, J. H. Warner, J. M. Smith, H. E. Assender, and A. A. R. Watt, “The molecular structure of Polymer–fullerene composite solar cells and its influence on device performance,” Macromolecules, vol. 43, no. 5, pp. 2343–2348, Feb. 2010.
[42] M. Guarrera, L. Turbino, and A. Rebora, “The anti-inflammatory activity of azulene,” J. Eur. Acad. Dermatol. Venereol., vol. 15, no. 5, pp. 486–487, Sep. 2001.
