Harnessing Recombinant DnaJ Protein as Reversible Metal Chelator

Subscriber access provided by Universitaetsbibliothek | Johann Christian Senckenberg

Article

Harnessing recombinant DnaJ protein as reversible metal chelator for high-performance resistive switching device Sung Kyu Jang, Sookyung Kim, Muhammad Saad Salman, Ji-ryang Jang, Yu Mi Um, Lihan Tan, Jin-Hong Park, Woo-Seok Choe, and Sungjoo Lee Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.7b04261 • Publication Date (Web): 11 Jan 2018 Downloaded from http://pubs.acs.org on January 11, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Harnessing recombinant DnaJ protein as reversible metal chelator for high-performance resistive switching device Sung Kyu Jang†, Sookyung Kim†, Muhammad Saad Salman‡, Ji-ryang Jang§, Yu Mi Um∥, Lihan Tan⊥, Jin-Hong Park#, Woo-Seok Choe†‡*, and Sungjoo Lee†#* †

SKKU Advanced Institute of Nano-Technology (SAINT), ‡School of Chemical Engineering, and #College of Information and Communication Engineering, Sungkyunkwan University (SKKU), Suwon 16419, Republic of Korea §

CJ CheilJedang Corp, Research Institute of Biotechnology, Suwon 16945, Republic of Korea

∥

Samsung BioLogics, 300 Songdo Bio Way (Daero), Yeonsu-gu, Incheon 21987, Republic of Korea

⊥

Bioprocessing Technology Institute, A*STAR, Downstream Processing Group, 20 Biopolis Way, Centros #06-01, Singapore 138668, Singapore Keywords: His-tagged DnaJ, recombinant protein, metal chelation, resistive switching device ABSTRACT: A high-performance biomaterial-based resistive switching (RS) device is fabricated by harnessing a thermally denatured protein (hexa-His-tagged recombinant molecular chaperone DnaJ (rDnaJ)) as a switching layer in a Cu/rDnaJ/Pt configuration on SiO2/Si substrate. The conductivity of the heat-denatured rDnaJ protein layer between the metal electrodes can be reversibly controlled to enable the formation/rupture of conductive Cu filaments by tailoring the metal chelating properties of the amino acid residues in the insulating protein matrix in a pH- and/or redox potentialdependent manner, giving rise to high-performance non-volatile RS behavior. The rDnaJ-based RS device exhibits extremely low set voltage (~0.12 V) and reset voltage (~0.08 V) with excellent uniformity, along with large memory window (RHRS/RLRS > 106) and long retention time (> 106 s). In addition, the rDnaJ RS device which is fabricated on a flexible poly (ethylene terephthalate) substrate exhibits an uncompromised switching performance. The present study is the first attempt to explore the use of a recombinant protein as a functional switching layer in RS devices. This approach opens up a new method of harnessing recombinant proteins with engineered properties as powerful building blocks to suit the requirements of next-generation biocompatible, flexible, high-performance, and low power consumption electronics.

There is a growing demand for biomaterials that can be used as the functional components of future devices. This demand is motivated by the need to fabricate building blocks suitable for next-generation biocompatible electronic/photonic devices for a variety of applications, such as flexible devices and low power consumption green electronics, via a simple fabrication process at a reduced cost. The features of biomaterials that are important for electronic and optoelectronic device applications include their flexibility, scalability, low production cost, biocompatibility, and sustainability. These properties can overcome the limitations of conventional silicon-based electronics by creating passive and active biomaterial components, especially in the areas of wearable information processing electronics, smart skin, and biomedical applications. Several studies have examined the use of biomaterials as active layers in devices, such as transistors1, diodes2, memory devices3, sensors4 and photonic devices5. One of the most promising applications of biomaterials in future electronic devices is in resistive switching (RS) devices, used in next-generation information storage/processing

devices, wherein the biomaterial acts as the main switching layer with a resistance that is modulated by an electrical stimulus. The RS device, also called a conductive bridging random access memory device or an atomic switching device, is an essential building block for information storage and logic components owing to its high density, large scalability, low power consumption, and high reliability. Recently, several groups have demonstrated the fabrication of RS devices using a variety of biomaterials3,6-12. Among these, proteins, essential components of organisms, have been researched for use in switching layers13-23. Such studies have demonstrated that natural biomaterials can effectively simplify the fabrication process and reduce the manufacturing costs associated with organic memory devices. Several studies have shown that the mechanism underlying the resistance switching behavior involves modulating the electrochemical properties of the biomaterials. The switching mechanisms were attributed to filamentary conduction24, redox reactions13, and/or charge hopping14. Biomaterials such as DNA6 , dronpa15, ferritin16, fibroin17,18, lysozyme19, gelatin20, and biocomposites7,21 can provide

ACS Paragon Plus Environment

Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

conduction pathways for redox reactions, and metal ions dissolved from the active electrode8,9 or introduced into the biomaterial10 can form metallic conductive bridges. RS devices based on charge trapping and de-trapping in biomaterials11,14, the composition of the biomaterial12, and the formation of biomaterialmetal nanoparticle complexes3,22 have also been reported. Although these studies have demonstrated the feasibility and potential advantages of protein-based RS devices, the device performances (e.g. in terms of mechanical stability and power consumption) must be further improved to enable wide usage in new functional memory and electronic circuits. The operation of RS devices critically depends on the formation of conductive bridges between the two electrodes. To satisfy the requirements of a sustainable non-volatile memory and switching device, comprehensive and precise control over the formation and rupture of conductive nano-filaments embedded in the biomaterials using novel pathways is critically required. RESULTS and DISCUSSION Protein resistive switching device integrated with rDnaJ. In this work, we employed a simple two-terminal metal/insulator/metal (MIM) structure comprising a recombinant molecular chaperone DnaJ, Cu, and Pt as the insulating matrix, the top (active), and the bottom (inert) electrode, respectively. The rationale behind the use of recombinant DnaJ is as follows: First, it should be noted that all previously reported RS device studies that have used proteins as the insulating layer have tested natural proteins, such as fibroin17,18, sericin14, gelatin20, albumen12, and ferritin16,23, with or without the presence of metal ions. No study has yet addressed the feasibility of using a recombinant protein as the functional switching layer in a RS device component. It would be helpful to understand whether recombinant proteins with genetically engineered properties (e.g. with the inclusion of hexaHis-tag) could result in enhanced performance as the switching layer in a RS device as compared to native proteins. If this is the case, this work provides a new potential approach to harnessing designer proteins in place of natural proteins and/or conventional inorganic insulating materials. Designer proteins will further enlarge the library of potential protein candidates to be used as the switching layer in a RS device. This will greatly increase the probability of finding a material with high functionality to suit the specific applications. In addition, properly designed recombinant protein can be produced in high yields and at a high purity, facilitating the fabrication of next-generation electronic devices. Secondly, DnaJ protein is a well-known ubiquitous molecular chaperone found in all prokaryotes and eukaryotes, including humans, which makes it less likely to be immunogenic during implantable bioelectronics applications. Thirdly, DnaJ protein is highly efficient in recognizing the hydrophobic patches (i.e., enriched with aliphatic and/or aromatic amino acid residues) of nascent and/or unfolded polypeptides25. This property is hypothesized to enable seamless interfacing with the

Page 2 of 12

hydrophobic surface of a Pt electrode, ensuring the facile and uniform assembly of an insulating protein layer with a controlled thickness atop the electrode. Fourthly, DnaJ has a unique zinc finger domain known as a CR (cysteine rich) region26 that is likely to be effective in chelating soft metal ions, such as zinc and copper. Considering that our MIM configuration employs copper as an active electrode from which abundant Cu ions may be sourced upon application of a positive potential, conductive filament formation was expected due to the intrinsic copper ion chelation properties of the DnaJ protein layer intervening between the top and bottom electrodes. Finally, the DnaJ protein was engineered to accommodate hexa-histidine residues at its N-terminal permissive site with negligible effect on its hydrophobic interactions with cognate substrates27. The use of His-tagged recombinant DnaJ (rDnaJ) may empower and/or modulate the chelation of Cu ions by the insulating protein layer in a pH-dependent manner, which is critical to controlling the availability and location of Cu ions immobilized in the insulating protein layer. Unless otherwise stated, a thermally denatured protein, chaperone rDnaJ, was used as a switching layer, based on its protonation-dependent metal ion interaction properties, and was applied in a uniform layer between a Pt bottom electrode and a Cu top electrode. By tailoring the metal chelating properties of the hexa-histidine residues in the protein layer, which displayed a different affinity for metal ions depending on the protonation/deprotonation state at a given pH, and thus controlling the metal ion interactions in the protein switching layer, we obtained biomaterial-based highperformance non-volatile RS devices. The extremely low set and reset voltages of 0.12 V and 0.08 V, respectively, were highly uniform, with standard deviations of 0.015 V and 0.029 V, respectively. The large memory window (RHRS/RLRS > 106) and a long retention time > 106 s were achieved. We demonstrated the implementation of LATCH circuits by integrating chaperone rDnaJ RS devices. Our results show great promise for employing bioengineered materials in the area of future biocompatible non-volatile information storage device, information processing switching devices and their integrated logic circuit applications. Protein RS devices that use chaperone rDnaJ as the RS layer between the Pt bottom electrode and the Cu top electrode were fabricated (Figure 1a). The technique most commonly used to form an organic layer between two electrodes is spin-coating, which suffers from the disadvantages of producing a non-uniform coating. We used a thermal denaturation method to form a protein layer on the metal surface, which overcame the aforementioned disadvantage and provided an atomically thin (2-4 nm) and uniform layer that readily enabled electrochemical modulation of the protein switching layer in a protonation-dependent and structure-dependent manner. The thermal denaturation processes and features of the resulting protein layer are described in detail in the Supporting Information (Figures S1, S2 and S3). In the denatured rDnaJ concentration range from 2 to 8 μM (a –


Page 3 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


d in Figure S2), highly uniform thin insulating protein layers with thicknesses less than 10 nm, suitable for RS device fabrication, were formed. In particular, it is noted that the use of 2 μM denatured rDnaJ led to the formation of a very thin (4.09 nm) and uniform (root mean square roughness of 0.43 nm) insulating protein layer (Figure S2a-1), which is selected as a basis for fabrication of high performance RS devices demonstrated in the present study. Through this thickness optimization process, the 4 nm thick rDnaJ switching layer is sufficiently thin for strong electric fields to be applied even at low voltages, and thus suppressing unwanted migration of Cu ions due to thermal vibrations. As shown in Figure 1b, the rDnaJ layer sandwiched between the Cu and Pt electrodes worked as a functional switching layer, depending on the degree of chelation of cupric ions oxidized from the Cu electrode under a positive potential bias. The chelation was modulated by the protonation and/or deprotonation status of the imidazole group in histidine residue dependent on the pH. The rDnaJ protein contains 16 histidine residues in total, including a hexa-His-tag fused to its N-terminus. By contrast, 50% of the 10 histidine residues in the wild-type DnaJ are localized to its CR domain where they coordinate cysteine residues to chelate cupric ions. In the present study, a pH range of 58 was employed. Because the sulfhydryl group (pKa = 8.2) in cysteine remained mostly protonated in this pH range, it was likely that most of the cupric ion chelation by rDnaJ was mediated by histidine residues. Within this range, histidine residues acted as strong chelators of cupric ions when the imidazole nitrogen was deprotonated, thereby providing a lone pair of electrons to the electron-deficient metal ions. The resulting chelate sites act as oxygen vacancies in traditional oxide-based CBRAM to assist redox reaction and transportation of Cu ions28, contributing to anchoring of oxidized Cu ions at largely uniform spacing in the protein matrix between Cu (top) and Pt (bottom) electrodes. Subsequently, the reduction of chelated Cu ions by electrons sourced from the bottom electrode results in the formation of metallic copper filament. It facilitates the formation of Cu filaments through improved redox reaction and transportation of Cu metal/ions in the resistive switching layer29 The acid dissociation constant (pKa) of the imidazole side chain in histidine is approximately 630. At a pH above this pKa value, the imidazole ring is mostly deprotonated, and metalprotein coordination can occur. At a pH lower than this pKa value, the imidazole ring is mostly protonated, as described by the Henderson–Hasselbalch equation31. While protonated, the imidazole ring bears two NH bonds and has a positive charge that is evenly distributed across the two nitrogen atoms, producing a net charge of zero on the imidazole group. As a result, the metal chelating capacity of the rDnaJ layer could be controlled by changing the solution pH. This process is summarized in Figure 1b. This property facilitated simple yet effective control of the amount of electrode-sourced Cu ions available in the switching layer, thereby allowing optimization of the device characteristics.

Characteristics of the protonation-controlled rDnaJ resistive switching devices. The protein RS devices were electrically characterized in the DC voltage sweeping mode over repeated operation cycles (Figure 2a). The DC bias was applied to the top Cu electrode while the bottom Pt electrode was grounded. Initially, the asprepared device was in a high resistance state (HRS). As the voltage was swept from zero to positive values with current compliance, the current abruptly increased once the bias reached a threshold voltage of 0.12 V. This is called the SET process, and the corresponding voltage is defined as the set voltage (Vset). After the SET process, the device remained in a low resistance state (LRS). As the voltage was swept from zero to negative values, the resistance state changed from the LRS to the HRS at a voltage of 0.08 V, defined as the reset voltage (Vreset). This is called the RESET process. As shown in Figure 2a, the fabricated rDnaJ device displayed stable bipolar RS properties over multiple switching cycles while maintaining extremely low Vset and Vreset values, which is a key factor for satisfying the requirements of low power operation in future electronic devices. In Table 1, the Vset and Vreset values obtained from the rDnaJ switching device are compared with those of other recently reported RS devices using various biomaterials including proteins and carbohydrates. Our rDnaJ RS device exhibited significantly lower Vset and Vreset values with a high on/off ratio above 106, along with excellent uniformity. In Figure 2b, the I-V curve obtained from the rDnaJ protein device was plotted on a double-log scale to further investigate the conduction mechanism. The slopes of I-V curves were ~ 1 in both the HRS and LRS over the entire voltage range. These results suggested that the current flow in the rDnaJ protein device was dominated by Ohmic conduction rather than by space charge limiting current conduction, which is generally observed in a variety of RS devices32,33. Cyclic endurance and data retention tests were carried out to further investigate the stability of the rDnaJ protein RS devices (Figures 2c and 2d). A cyclic voltage was applied from 0 to 0.5 V and then from 0 to 0.3 V, and a read voltage of 0.03 V was employed to avoid a resistance state change. The rDnaJ RS device revealed that the LRS and HRS states remained stable over approximately 100 repeated cycling tests (Figure 2c) and a test period of 106 s (Figure 2d). The current ratio between the LRS and HRS state is above 106. The reliability of the device was assessed by measuring the current-voltage characteristics repeatedly in several devices. The inset of Figure 2c shows the distribution of the cumulative probabilities of the resistance at the LRS (RLRS) and HRS (RHRS). Each data point was obtained from totally 120 switching cycles applied to randomly selected 12 devices. The RLRS and RHRS displayed an average resistance values of about 5 kΩ and 2 GΩ, respectively, and these values fluctuated by about 1.1 orders of magnitude, maintaining a memory window of at least 104. The endurance and retention characteristics of the rDnaJ RS devices were superior to those obtained from previously reported switching devices fabricated



using other biomaterials (Table 1). Figure 2d plots the RHRS and RLRS values obtained from the rDnaJ RS devices as a function of the retention time. In previous studies, organic RS devices suffered from low on-off ratios and short retention times due to a lack of electrochemical controllability3,14,18,23. Our results revealed that the use of a controlled rDnaJ switching layer provided biomaterialbased high-performance non-volatile memory and switching devices. Figure 3a plots the Vset and Vreset values of the rDnaJ protein RS devices, in which the protonation states of the rDnaJ proteins were controlled by varying the pH. For simplicity, RS devices with an rDnaJ switching layer prepared at pH 5, 6, 7 and 8 will be designated hereafter as rDnaJ-5, rDnaJ-6, rDnaJ-7 and rDnaJ-8, respectively. The SET and RESET voltages were found to be largest in rDnaJ-5 and smallest in rDnaJ-6, although they further increased as the pH increased. This trend indicated that to achieve stable low-power switching operation, the rDnaJ RS layer prepared at pH 6 provided the optimal performance. In addition, the pH-dependent controllability of Vset and Vreset of rDnaJ RS devices can be tailored to the requirements of the application circuit, such as noise margin and operating voltage. Non-volatile switching behavior was obtained from the rDnaJ RS device, attributed to the formation and rupture of conductive filaments. The positive bias induced dissolution of Cu atoms, in the form of Cu ions, into the protein switching layer, with the help of amino acids with a metal binding affinity. The Cu ions then diffused toward the bottom electrode and were reduced to Cu atoms. This process could be repeated so that a Cu filament formed between the two electrodes, and the resistance state of the device was changed from a HRS to a LRS. This phenomenon was not observed in a device prepared with Pt electrodes at both electrodes (Figure S4). Application of a reverse voltage resulted in filament rupture via thermally assisted electrochemical dissolution. Figure 3b shows a schematic illustration of Cu conductive filament formation for each rDnaJ device under a positive bias applied to the Cu top electrode. The oxidation and reduction of Cu ions at the top Cu electrode was responsible for the RS device operation, and the amount of Cu ions chelated in the switching layer was controlled by the degree of protonation in the rDnaJ layers. The correlation between the amount of chelated Cu ions and the degree of protonation was confirmed by electrochemical measurements. (see Figure S5 of the Supporting Information for details). In rDnaJ-5, most of the lone pair electrons in the imidazole nitrogen in the histidine residues were protonated and hence unavailable to interact with the metal ions. According to the HendersonHasselbalch equation31, the fractions of deprotonated imidazole nitrogen groups in the histidine moieties were 0.09, 0.50, 0.91, and 0.99 at pH 5, 6, 7 and 8, respectively. Therefore, in rDnaJ-5, only a small number of the Cu ions were chelated, and a large voltage was required to SET the device. This led to the formation of thick filaments that increased the voltage required for

Page 4 of 12

RESET, resulting in a large operating voltage. These results suggested that the performance could be improved by increasing the chelation of Cu ions into the switching layer. On the other hand, in rDnaJ-8, the chelation of Cu ions was excessive, and thick filaments were formed because it was difficult to precisely control the filament formation. Significant amounts of energy were required to remove any excess Cu ions from the rDnaJ layer. Consequently, a high operating voltage and large SET/RESET voltage distributions were observed. In this case, it was possible to improve the performance by reducing the rate of Cu ion chelation. The SET and RESET voltages measured from the rDnaJ RS devices (Figure 3a) indicated that (i) the device performances depended significantly on the metal chelation affinity of the protein (Figures 1 and S5), and (ii) a low-power operation biomaterial switching device with excellent uniformity could be designed by optimizing the protonation state of the protein (i.e. rDnaJ-6 which displayed extremely low Vset ~ 0.12 V and Vreset ~ 0.08 V with limited variability as shown in Figure 3a). As a result, the rDnaJ-6 devices displayed excellent non-volatile RS characteristics (Figure 2) that outperformed the previously reported experimental results obtained from other similar devices (Table 1). The metal chelation affinity also can be controlled by changing the metal ion or the protein with different amino acid composition. Other metal ions (such as Cu, Ni, Zn and Co) or other proteins can be used to adjust device operating parameters or implement new functionality. Further investigations are required to fully develop a new method of harnessing recombinant proteins with engineered properties. A variety of RS devices have displayed a quadratic dependence of the current on voltage prior to the SET process, due to current fluctuations caused by space charges and traps in the dielectric film34-36. The differences observed in our rDnaJ RS device, including observation of the dominant overall Ohmic conduction (Figure 2b), most likely are attributable to the Cu ionfriendly amino acids present in the rDnaJ protein, which lowers the barrier for injection of Cu ions but not electrons. The enhanced Cu ion injection protects the resistance state of the device from charge distributions that produce an unintended transitional state, and the devices can be switched between LRS and HRS states without experiencing charge/trap related conduction. Another widely reported observation on RS devices in general is that a “forming process” at a higher voltage stress is needed when the as-prepared device first switches from the HRS to the LRS37. By contrast, our rDnaJ RS device did not display such a forming process (Figure 2a), which suggests that metal ion transport in the thin rDnaJ switching layer was facilitated by the chelating sites on the proteins. Direct observation of conductive filaments. The conductive filaments formed within the rDnaJ switching layer were investigated by performing conductive atomic force microscopy (CAFM) measurements. Figure 3c presents a schematic diagram of the CAFM measurement


Page 5 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


configuration. The sample was prepared by etching the Cu top electrode of each rDnaJ device after applying the electrical SET process. The local filamentary conducting paths within the rDnaJ layer were scanned using the CAFM tip within a 5 × 5 um2 area. AFM measurements were simultaneously collected to investigate the surface morphology of thermally denatured rDnaJ on the platinum layer and to monitor damage caused by Cu over-etching. The surface of the rDnaJ layer, onto which the Cu electrode was etched, displayed a uniform rootmean square roughness of less than 1 nm, irrespective of the degree of protonation. Figure 3d presents a surface current mapping image of the rDnaJ-6 device obtained from the CAFM measurements. These results revealed that the filamentary conducting paths formed in the LRS but not in the HRS state (Figure S6). The highmagnification CAFM image displayed in the inset of Figure 3d reveals a well-defined local filament with an area of ~ 90 nm2. Figure 3e presents the current distribution profiles measured from the conductive filaments formed within each rDnaJ layer prepared with different degrees of protonation (rDnaJ-5, rDnaJ-6, rDnaJ7 and rDnaJ-8). The CAFM results under each condition used in the graph may be found in the Supporting Information (Figure S7). The maximum filament currents measured from the rDnaJ-5, rDnaJ-6, rDnaJ-7 and rDnaJ-8 layers were 2.2, 1.8, 2.1 and 2.4 nA, respectively. Importantly, each sample displayed a distinct current distribution profile, indicating the size of the conductive filaments in each switching layer. A narrower distribution in the filament current was measured in the case of rDnaJ-6, indicating thinner filament formation, which is critical to lowering the Vset/Vreset and suppressing variability in the RS device operation (Figure 3a) for lowpower data storage and information processing device applications. Broader current profile distributions were observed for the rDnaJ-5, rDnaJ-7 and rDnaJ-8 devices. These results revealed that the filament size could be effectively controlled by the degree of protonation within the switching layer, as described in Figure 3b. SR latch circuit and flexible devices. The non-volatile properties of the rDnaJ RS device could be further utilized to construct a sequential logic circuit (SR latch), a crucial building block in digital information processing systems. The SR latch included two output signals (Q and ), the states of which could be controlled and stored with respect to two input signals: S (set) and R (reset). Figures 4a and 4b plot the schematic diagram and operation of an SR latch circuit composed of two rDnaJ-6 RS devices. When both S and R were low, no voltage was applied to each device, and the rDnaJ-6 latch output held its previous value. If either S or R was high while the other was low, the output Q was forced to assume a high or low state, and Q remained fixed until the signals applied to S or R changed. These results successfully demonstrated the construction of sequential logic circuits using integrated rDnaJ-6 RS devices, revealing that the rDnaJ-6 RS device can form a fundamental building block for future digital electronic systems. In the LRS state, the current of the

rDnaJ RS device was found to be inversely proportional to the current compliance. These results suggested that multi-level switching operation could be achieved if the current compliance were controlled during the SET process. Figure 4c plots the current compliance-mediated multi-level state characteristics of the rDnaJ RS device under repeated SET/RESET operation. The SET and RESET voltages remained almost constant whereas the LRS resistance increased as the current compliance conditions changed from 1 to 60 A. The LRS states were completely restored to the HRS state through the RESET process. A LRS multi-level state in which the resistance ratio of ~100 was deterministically implemented by adjusting the current compliance. The set and reset voltages remained approximately constant while the LRS resistance and current levels changed over orders of magnitude in response to variations in the compliance current. The excellent switching characteristics and nonvolatile capabilities of the rDnaJ RS device were also employed to demonstrate the sequential logic circuits (Figure 4a). In addition, flexible devices have attracted increasing attention due to their potential applications in wearable electronic devices and biomedical devices. A flexible switching device based on a biocompatible material would be highly desirable. The rDnaJ RS devices were fabricated on polyethylene terephthalate (PET) transparent and flexible substrates. We measured the resistance of LRS and HRS as a function of the bending radius to investigate the mechanical bending stability of the rDnaJ RS device on a PET substrate. As shown in Figure 4d, no significant variations in the resistance or ON/OFF ratio were observed above an ON/OFF ratio of 106, demonstrating the excellent flexibility of our rDnaJ RS device. CONCLUSION In summary, we successfully prepared highperformance biocompatible non-volatile data storage and RS devices by integrating an rDnaJ switching layer, with controllable metal ion (Cupric ions) chelating properties, into an rDnaJ RS device. The low-power operation (Vset ~ 0.12 V, Vreset ~ 0.08 V) and non-volatile RS devices were fabricated by integrating pH-controllable metal chelating sites in the form of the rDnaJ protein. The rDnaJ thin switching layer was deposited by thermal denaturation, and optimal conductive bridges within the switching layer could be controllably formed by adjusting the protonation state of the protein. CAFM and cyclic voltammetry measurements revealed the structures of the filaments and the dependency of electrical switching characteristics on protonation states of the rDnaJ switching layer. The excellent switching properties and non-volatile characteristics of the rDnaJ RS devices were deployed in SR Latch sequential logic circuits. rDnaJ RS devices were shown to successfully enable multi-level switching operation and the reliable operation of flexible devices. These results show great promise of the rDnaJ RS device as a fundamental building block for various device applications.



Page 6 of 12

Figure 1. (a) Schematic diagrams of the fabricated rDnaJ RS device arrays (left), the enlarged single device with a conductive filament (middle), and the degree of deprotonation of the imidazole nitrogen in histidine, as a function of pH (right). (b) Surface charge diagram illustrating the degree of protonation of imidazole in histidine.


Page 7 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. (a) I-V Characteristics of the rDnaJ (prepared at pH 6) RS devices. (b) A double-logarithmic scale plot for the set and reset processes in I–V curves. Linear fits are shown in the figure. (c) The endurance characteristics for the cyclic set/reset process over approximately 100 cycles at Vread = 0.03 V. The inset shows the distribution of the cumulative probabilities of the resistance values at LRS (RLRS) and HRS (RHRS). (d) The data retention performance up to 106 s.



Page 8 of 12

Figure 3. (a) The distributions of Vset and Vreset as a function of pH of the buffer used to form the switching layer. (b) Schematic illustration of the Cu atom distribution throughout the rDnaJ RS devices in the LRS and HRS. Red sites represent the deprotonated imidazole groups in histidine. (c) Schematic diagram of the CAFM used to observe the conductive filaments. (d) CAFM current mapping image of a conductive filament in the rDnaJ-6 RS device after removing the top Cu electrode. The inset shows enlarged CAFM image of the indicated area. (e) Cumulative histogram of current distribution profiles measured from the conductive filaments formed within each rDnaJ layer.


Page 9 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. (a) Schematic diagram of the SR latch and (b) its operating result. (c) Current compliance-mediated multi-level RLRS state characteristics of the rDnaJ RS device. The inset shows the I-V curves obtained using different current compliance values. (d) Resistance (RLRS and RHRS) as a function of bending radius for the rDnaJ RS device fabricated on the flexible PET substrate.



Page 10 of 12

Table 1. Performance comparison of biomaterial-based RS devices. Switching material

Vset (V) Mean / Std. dev.

Vreset (V) Mean / Std. dev.

RHRS/RLRS

Retention Time (s)

Ref

Tobacco mosaic virus

3.1 / NA

-2.4 / NA

103

103

3

104

8

Ag-doped chitosan

0.5 / NA

-0.5 / NA

105

Dry-Cured Albumen

0.6 / NA

2.2 / NA

103

104

12

Sericin

2.5 / NA

-0.8 / NA

106

103

14

-11.5 / NA

10

103

17

104

19

Silk fibroin

10.4 / NA

Lysozyme/PSS

1.0 / NA

-1.3 / NA

103

(PAH/ferritin)15

1.8 / NA

-1.5 / NA

102

104

21

-0.8 / NA

104

N/A

23

-0.08 / 0.029

>106

>106

This work

NiO/ferritin rDnaJ-6

1.8 / NA 0.12 / 0.015


Page 11 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ASSOCIATED CONTENT Supporting Information. Cloning, expression, and purification of rDnaJ; preparation of the rDnaJ-coated substrates; device fabrication and electrical characterization; electrochemical measurements (cyclic voltammetry); conductive atomic force microscopy images are available at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] (W. C.)., *E-mail: [email protected] (S. L.).

ACKNOWLEDGMENT This research was supported by the Global Frontier Program through the Global Frontier Hybrid Interface Materials (GFHIM) (2013M3A6B1078873) of the National Research Foundation of Korea (NRF) and the Pioneer Research Center Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT & Future Planning (2014 M3C1A3053024), and the Basic Science Research Program through the National Research Foundation of Korea funded by the Korean government (MSIP) (grant no.: 2015R1D1A1A09057297, 2015M3A7B7045496, 2017R1A4A1015400, and 2017R1A2B2011341).

REFERENCES (1) Maruccio, G.; Biasco, A.; Visconti, P.; Bramanti, A.; Pompa, P. P.; Calabi, F.; Cingolani, R.; Rinaldi, R.; Corni, S.; Di Felice, R.; et al. Towards Protein Field-Effect Transistors: Report and Model of a Prototype. Adv. Mater. 2005, 17, 816–822. (2) Maglia, G.; Heron, A. J.; Hwang, W. L.; Holden, M. a; Mikhailova, E.; Li, Q.; Cheley, S.; Bayley, H. Droplet Networks with Incorporated Protein Diodes Show Collective Properties. Nat. Nanotechnol. 2009, 4, 437–440. (3) Tseng, R. J.; Tsai, C.; Ma, L.; Ouyang, J.; Ozkan, C. S.; Yang, Y. Digital Memory Device Based on Tobacco Mosaic Virus Conjugated with Nanoparticles. Nat. Nanotechnol. 2006, 1, 72–77. (4) Flink, S.; Veggel, F. C. J. M. van; Reinhoudt, D. N. Sensor Functionalities in Self-Assembled Monolayers. Adv. Mater. 2000, 12, 1315–1328. (5) Greenbaum, E. Vectorial Photocurrents and Photoconductivity in Metalized Chloroplasts. J. Phys. Chem. 1990, 94, 6151– 6153. (6) Hung, Y.-C.; Hsu, W.-T.; Lin, T.-Y.; Fruk, L. Photoinduced Write-Once Read-Many-Times Memory Device Based on DNA Biopolymer Nanocomposite. Appl. Phys. Lett. 2011, 99, 253301. (7) Raeis-Hosseini, N.; Lee, J.-S. Controlling the Resistive Switching Behavior in Starch-Based Flexible Biomemristors. ACS Appl. Mater. Interfaces 2016, 8, 7326–7332. (8) Raeis Hosseini, N.; Lee, J. Resistive Switching Memory Based on Bioinspired Natural Solid Polymer Electrolytes. ACS Nano 2015, 9, 419–426. (9) Nagashima, K.; Koga, H.; Celano, U.; Zhuge, F.; Kanai, M.; Rahong, S.; Meng, G.; He, Y.; De Boeck, J.; Jurczak, M.; et al. Cellulose Nanofiber Paper as an Ultra Flexible Nonvolatile Memory. Sci. Rep. 2015, 4, 5532. (10) Sun, B.; Wei, L.; Li, H.; Jia, X.; Wu, J.; Chen, P. The DNA Strand Assisted Conductive Filament Mechanism for Improved

Resistive Switching Memory. J. Mater. Chem. C 2015, 3, 12149– 12155. (11) Qin, S.; Dong, R.; Yan, X.; Du, Q. A Reproducible Write– (read)n–erase and Multilevel Bio-Memristor Based on DNA Molecule. Org. Electron. 2015, 22, 147–153. (12) Chen, Y.-C.; Yu, H.-C.; Huang, C.-Y.; Chung, W.-L.; Wu, S.-L.; Su, Y.-K. Nonvolatile Bio-Memristor Fabricated with Egg Albumen Film. Sci. Rep. 2015, 5, 10022. (13) Meng, F.; Jiang, L.; Zheng, K.; Goh, C. F.; Lim, S.; Hng, H. H.; Ma, J.; Boey, F.; Chen, X. Protein-Based Memristive Nanodevices. Small 2011, 7, 3016–3020. (14) Wang, H.; Meng, F.; Cai, Y.; Zheng, L.; Li, Y.; Liu, Y.; Jiang, Y.; Wang, X.; Chen, X. Sericin for Resistance Switching Device with Multilevel Nonvolatile Memory. Adv. Mater. 2013, 25, 5498– 5503. (15) Korpany, K. V.; Langat, P.; Kim, D. M.; Edelman, N.; Cooper, D. R.; Nadeau, J.; Blum, A. S. Conductance Switching in the Photoswitchable Protein Dronpa. J. Am. Chem. Soc. 2012, 134, 16119–16122. (16) Meng, F.; Sana, B.; Li, Y.; Liu, Y.; Lim, S.; Chen, X. Bioengineered Tunable Memristor Based on Protein Nanocage. Small 2014, 10, 277–283. (17) Hota, M. K.; Bera, M. K.; Kundu, B.; Kundu, S. C.; Maiti, C. K. A Natural Silk Fibroin Protein-Based Transparent BioMemristor. Adv. Funct. Mater. 2012, 22, 4493–4499. (18) Mukherjee, C.; Hota, M. K.; Naskar, D.; Kundu, S. C.; Maiti, C. K. Resistive Switching in Natural Silk Fibroin ProteinBased Bio-Memristors. Phys. status solidi 2013, 210, n/a-n/a. (19) Baek, H.; Lee, C.; Lim, K.; Cho, J. Resistive Switching Memory Properties of Layer-by-Layer Assembled Enzyme Multilayers. Nanotechnology 2012, 23, 155604. (20) Chang, Y.-C.; Wang, Y.-H. Resistive Switching Behavior in Gelatin Thin Films for Nonvolatile Memory Application. ACS Appl. Mater. Interfaces 2014, 6, 5413–5421. (21) Ko, Y.; Kim, Y.; Baek, H.; Cho, J. Electrically Bistable Properties of Layer-by-Layer Assembled Multilayers Based on Protein Nanoparticles. ACS Nano 2011, 5, 9918–9926. (22) Gogurla, N.; Mondal, S. P.; Sinha, A. K.; Katiyar, A. K.; Banerjee, W.; Kundu, S. C.; Ray, S. K. Transparent and Flexible Resistive Switching Memory Devices with a Very High ON/OFF Ratio Using Gold Nanoparticles Embedded in a Silk Protein Matrix. Nanotechnology 2013, 24, 345202. (23) Uenuma, M.; Kawano, K.; Zheng, B.; Okamoto, N.; Horita, M.; Yoshii, S.; Yamashita, I.; Uraoka, Y. Resistive Random Access Memory Utilizing Ferritin Protein with Pt Nanoparticles. Nanotechnology 2011, 22, 215201. (24) Liu, Q.; Sun, J.; Lv, H.; Long, S.; Yin, K.; Wan, N.; Li, Y.; Sun, L.; Liu, M. Real-Time Observation on Dynamic Growth/Dissolution of Conductive Filaments in OxideElectrolyte-Based ReRAM. Adv. Mater. 2012, 24, 1844–1849. (25) Rudiger, S. Its Substrate Specificity Characterizes the DnaJ Co-Chaperone as a Scanning Factor for the DnaK Chaperone. EMBO J. 2001, 20, 1042–1050. (26) Szabo, A.; Korszun, R.; Hartl, F. U.; Flanagan, J. A Zinc Finger-like Domain of the Molecular Chaperone DnaJ Is Involved in Binding to Denatured Protein Substrates. EMBO J. 1996, 15, 408–417. (27) Um, Y. Harnessing Optimized Chaperone Cocktail for Folding-like-Refolding of Heat Denatured MDH, Sungkyunkwan University, 2014. (28) Lata, S.; Piehler, J. Stable and Functional Immobilization of Histidine-Tagged Proteins via Multivalent Chelator Headgroups on a Molecular Poly(ethylene Glycol) Brush. Anal. Chem. 2005, 77 (4), 1096–1105.



(29) Zhao, Y. D.; Hu, J. J.; Huang, P.; Yuan, F.; Chai, Y.; Liu, X. Y.; Kang, J. F. A Physics-Based Compact Model for Material- and Operation-Oriented Switching Behaviors of CBRAM. In 2016 IEEE International Electron Devices Meeting (IEDM); IEEE, 2016; p 7.6.1-7.6.4. (30) Pack, D. W.; Putnam, D.; Langer, R. Design of ImidazoleContaining Endosomolytic Biopolymers for Gene Delivery. Biotechnol. Bioeng. 2000, 67, 217–223. (31) Henderson, L. J. Concerning the Relationship between the Strength of Acids and Their Capacity to Preserve Neutrality. Am. J. Physiol. Content 1908, 21, 173–179. (32) Kim, Y.; Lee, C.; Shim, I.; Wang, D.; Cho, J. Nucleophilic Substitution Reaction Based Layer-by-Layer Growth of Superparamagnetic Nanocomposite Films with High Nonvolatile Memory Performance. Adv. Mater. 2010, 22, 5140–5144. (33) Yang, Y. C.; Pan, F.; Liu, Q.; Liu, M.; Zeng, F. Fully RoomTemperature-Fabricated Nonvolatile Resistive Memory for Ultra-

Page 12 of 12

fast and High-Density Memory Application. Nano Lett. 2009, 9, 1636–1643. (34) Chen, C.; Pan, F.; Wang, Z. S.; Yang, J.; Zeng, F. Bipolar Resistive Switching with Self-Rectifying Effects in Al/ZnO/Si Structure. J. Appl. Phys. 2012, 111, 13702. (35) Chen, G.; Song, C.; Chen, C.; Gao, S.; Zeng, F.; Pan, F. Resistive Switching and Magnetic Modulation in Cobalt-Doped ZnO. Adv. Mater. 2012, 24, 3515–3520. (36) Kiy, M.; Losio, P.; Biaggio, I.; Koehler, M.; Tapponnier, A.; Günter, P. Observation of the Mott–Gurney Law in Tris (8Hydroxyquinoline) Aluminum Films. Appl. Phys. Lett. 2002, 80, 1198–1200. (37) Waser, R.; Dittmann, R.; Staikov, G.; Szot, K. Redox-Based Resistive Switching Memories - Nanoionic Mechanisms, Prospects, and Challenges. Adv. Mater. 2009, 21, 2632–2663.


Harnessing Recombinant DnaJ Protein as Reversible Metal Chelator

Recommend Documents