Control of the Intrinsic Sensor Response to Volatile Organic

Dec 26, 2017 - Department of Physical Electronics, School of Electrical Engineering, Tel-Aviv University ... Propulsion, University of Central Florida...
0 downloads 0 Views 1MB Size
Subscriber access provided by READING UNIV

Article

Control of the Intrinsic Sensor Response to Volatile Organic Compounds with Fringing Electric Fields Alex Henning, Nandhini Swaminathan, Yonathan Vaknin, Titel Jurca, Klimentiy Anatoliy Shimanovich, Gil Shalev, and Yossi Rosenwaks ACS Sens., Just Accepted Manuscript • DOI: 10.1021/acssensors.7b00754 • Publication Date (Web): 26 Dec 2017 Downloaded from http://pubs.acs.org on December 27, 2017

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.

ACS Sensors 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 9 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

ACS Sensors

Control of the Intrinsic Sensor Response to Volatile Organic Compounds with Fringing Electric Fields Alex Henning§,†,‡, Nandhini Swaminathan§,‡, Yonathan Vaknin§, Titel Jurca+,∥ , Klimentiy Shimanovich§, Gil Shalev⊥,#, Yossi Rosenwaks*,§ §

Department of Physical Electronics, School of Electrical Engineering, Tel-Aviv University, Ramat-Aviv 69978, Israel Department of Chemistry, University of Central Florida, Orlando, Florida 32816, United States

+

∥Cluster for the Rational Design of Catalysts for Energy Applications and Propulsion, University of Central Florida, Orlando, Florida 32816, United States ⊥Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev, POB 653, Beer-Sheva 84105, Israel # Ilse-Katz center for Nanotechnology, Ben-Gurion University of the Negev, POB 653, Beer-Sheva 84105, Israel KEYWORDS Intrinsic Sensor response, Fringing electric field, Sensor selectivity, Chemical sensor, Volatile organic compounds, Kelvin probe force microscopy

ABSTRACT: The ability to control surface-analyte interaction allows tailoring chemical sensor sensitivity to specific target molecules. Adjusting the bias of the shallow p-n junctions in the electrostatically-formed nanowire (EFN) chemical sensor, a multiple gate transistor with an exposed top dielectric layer, allows tuning of the fringing electric field strength (from 0.5×107 to 2.5×107 V/m) above the EFN surface. Herein, we report that the magnitude and distribution of this fringing electric field correlate with the intrinsic sensor response to volatile organic compounds. The local variations of the surface electric field influence the analytesurface interaction affecting the work function of the sensor surface, assessed by Kelvin probe force microscopy on a nanometer scale. We show that the sensitivity to fixed vapor analyte concentrations can be nullified and even reversed by varying the fringing field strength, and demonstrate selectivity between ethanol and n-butylamine at room temperature using a single transistor without any extrinsic chemical modification of the exposed SiO2 surface. The results imply an electric-field-controlled analyte reaction with a dielectric surface extremely compelling for sensitivity and selectivity enhancement in chemical sensors.

Chemical sensors based on an open gate field-effect transistor (FET) architecture have been thoroughly investigated since their introduction in 1970 by Bergveld1. To date, the most promising chemical sensors for mass production and integration into mobile electronics comprise devices based on siliconon-insulator (SOI) technology2-4, as well as on nanostructured silicon such as fin FETs5 and nanowire FETs6, 7. In general, chemical sensor sensitivity can be enhanced by independently improving the extrinsic response, mainly determined by the signal-to-noise ratio and the transducer gain, and the intrinsic response that depends on the analyte-surface interaction. Most commonly, the sensor dielectric is functionalized with organic linker molecules8, 9 to enhance the intrinsic sensitivity and achieve sensor chemo-selectivity (i.e., to discriminate between different analytes) at room temperature (RT). Sensor surface reactivity with specific analytes can be enhanced or suppressed depending on the reactive binding sites of these linker molecules. A sensor array consisting of different units, each with a unique surface termination, outputs a characteristic signal pattern upon interaction with a target molecule serving as an electronic nose (e-nose). Such an e-nose can be used to monitor indoor air quality10, ascertain food freshness11 and

detect diseases based on breath analysis12-14. However, organic surface modifications complicate the device fabrication process and their long-term stability is uncertain, especially after prolonged and repetitive exposure to the analyte and ambient air15. The electrostatically-formed nanowire (EFN) sensor is a multiple gate FET based on the four-gate transistor introduced in the early 2000s16, 17. The top gate dielectric of the EFN device is exposed to the ambient and functions as a "molecular gate"18 capable of detecting femtomolar concentrations of biomolecules in liquid4 and parts per million (ppm) concentrations of organic vapors19, 20 at RT. In our previous work, we demonstrated that the applied voltages control the cross-sectional area and position of the conducting channel confined to nanometer dimensions within the Si body4, 19,21 of the EFN transistor resulting in a tunable dynamic range (e.g., ∼ 26 to 2030 ppm for ethanol)22. The EFN device geometry and bias configuration influence the transducer sensitivity (i.e., the extrinsic response)21, however, their effect on analyte interaction with the EFN sensor surface (i.e., the intrinsic response) has not yet been explored.

ACS Paragon Plus Environment

ACS Sensors 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

Herein, we use EFN devices of the same structure and demonstrate control of the intrinsic sensor response to short chain organic molecules with fringing electric fields whose strength is adjusted (from 0.5×107 to 2.5×107 V/m) with the junction gates of the EFN transistor in-operando. Kelvin probe force microscopy (KPFM) elucidates this novel concept by revealing the electric-field-controlled build-up of an adsorbate on the sensor surface with sub-micrometer spatial resolution. The effect of surface electric fields on the SiO2 dielectric breakdown of silicon transistors exposed to different environments has been thoroughly studied since the 1960s23-26. Electric fields below 108 V/m are rather too small to cause stressinduced leakage current at RT27-29 but have been reported to affect vapor analyte adsorption on surfaces of silica30 and semiconductors (mainly metal oxides)31, 32;33 and references therein. However, to the best of our knowledge, fringing-fieldcontrolled chemical sensing has not yet been studied with FET-based chemical sensors. We examine this principle for vapor phase chemical detection at RT and demonstrate selectivity between ethanol (CH3CH2OH) and n-butylamine (CH3CH2 CH2CH2NH2) using a single transistor without any additional (extrinsic) surface modification.

EXPERIMENTAL SECTION Device fabrication. The EFN transistors were fabricated in a semiconductor foundry (TowerJazz, Migdal Haemek, Israel) employing SOI technology in a complementary metal-oxidesemiconductor (CMOS) process, utilizing Boron-doped (1.5×1014 cm-3) 8-inch SOI wafers (Soitec, Bernin, France) with a bulk resistivity of 750 Ω×cm and nominal SOI and back SiO2 layer thicknesses of 150 nm and 1 µm, respectively. Four masks for photolithography were used, one to define the trenches, and the other three to implant the dopants of channel (Arsenic, 4×1017 cm-3), source-drain (Arsenic, 7×1019 cm-3), and junction gate regions (Boron, 2×1019 cm-3). A 3 nm thick SiO2 surface dielectric layer was thermally grown at 900 °C. Titanium/gold (Ti/Au; 10 nm/150 nm) contacts were subsequently manufactured by optical lithography in a lift-off process. Finite element simulations. The EFN device was modeled in three dimensions (3D) including simulations of dopant diffusion during the fabrication processes using a finite element device simulator (Sentaurus, Synopsys, Mountain View, USA). The doping density distributions, depletion region and channel dimensions, as well as the surface electric field magnitudes were predicted. Simulated doping density profiles were compared with the ones measured by dynamic secondary-ion mass spectrometry, and the simulated current-voltage (I-V) characteristics were compared with the experimental ones. Further details are provided in section 1 of the Supporting Information (SI). Current-voltage characteristics and sensing. I-V characteristics were measured with a semiconductor parameters analyzer (B1500A, Agilent technologies, Santa Clara, USA). Three EFN sensors were used for experiments in this work. Each sensor was reused multiple times during 1-2 weeks of experiment including repeat measurements (Fig. S5). Volatile organic compounds were detected at RT (22 °C) in a sealed gas chamber connected to a gas dilution system using nitrogen (≥ 99.999 % purity; < 1 ppm H2O; < 100 ppm O2) as the carrier gas. Organic vapors were generated in a bubbler system, diluted with pure nitrogen using mass flow controllers (Key

Instruments, FR2000, Trevose, USA), and dispersed before introduction into the gas chamber at a flow rate of 2 sscm. The analyte concentration inside the chamber was monitored down to the 100 ppb concentration level using a photoionization detector (ppbRAE 3000, RAE Systems, San Jose, USA) as the reference sensor. Vapors of ethanol and n-butylamine were sensed independently with the same device in a dry nitrogen atmosphere. The applied junction gate bias was sustained well below the built-in potential of ~0.7 V to maintain a p-n junction leakage current (reverse saturation current) of below ∼1 nA. The mean square error of the sensor response was calculated from fluctuations (noise) of the measured current and threshold voltage, uncertainties in the flow rate and the analyte concentration inside the gas chamber, as well as from the deviations of current and threshold voltage before/after sensor recovery (see S4 of the SI). We refer to our previous work that provides additional details on the signal-to-noise ratio of the EFN sensor20. Kelvin probe force microscopy. KPFM was carried out with a commercial atomic force microscopy (AFM) setup (Dimension Edge, Bruker, Billerica, USA) inside a nitrogen glove box with < 2 ppm H2O. Platinum/Iridium (95%/5%) coated cantilevers (PPP EFM, Nanosensors, Neuchâtel, Switzerland) were used. The contact potential difference (CPD) was measured simultaneously with the topography using amplitude modulation KPFM at an effective (average) tip sample distance of ∼10 nm during scanning. The CPD was measured by compensating the alternating current (ac) component of the electrostatic force with an applied direct current (dc) voltage in a feedback control loop. To separate topographic from CPD signal, increase the sensitivity and maintain a minimum tipsample distance, the ac electrostatic force component was generated at the second resonance, f2nd ~450 kHz, of the cantilever by applying an ac voltage of 500 mV. Additional experimental details and measurements are provided in section 3 of the SI.

RESULTS AND DISCUSSION The EFN device (Fig. 1) consists of an SOI layer that is covered with a 3 nm thin SiO2 film and separated by a back SiO2 layer (gray) from the handle wafer. The junction gate bias, VJG, applied to the p-doped silicon (Si) controls the extent of the space charge region (SCR) and the current density that flows through the n-doped Si channel (dark blue). The ionized dopants in the SCR (light blue) induce an electric field that penetrates into the ambient. VJG controls the extrinsic sensor response by changing the effective channel width, weff,19, 21 as well as the intrinsic response when fringing fields affect vapor molecule interaction with the sensor surface.

Figure 1. Schematic representation of the fringing field sensor and the corresponding electrical circuit layout. The applied junction

ACS Paragon Plus Environment

Page 2 of 9

Page 3 of 9 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

ACS Sensors and back gate voltages, VJG and VBG, control the width of lateral and vertical depletion regions (light blue), respectively. The drain bias, VD, drives the current through the n-doped Si (dark blue) channel flanked by two p-doped (red) junction gates (JG1, JG2).

The magnitude and distribution of the fringing electric field (Fig. 2) across the p-n-p structure of the sensor was computed using 3D finite element simulations (TCAD Sentaurus, Synopsys, Mountain View, USA). The concave-shaped doping profiles (Fig. S1) and the ultrathin top dielectric layer result in a fringing field strength, E, exceeding 2.5×107 V/m at the surface right above the reverse biased p-n junctions (Fig. 2a, top). The strength and spatial extent of the fringing fields are significantly smaller (E ≤ 0.5×107 V/m) for the forward biased p-n

junctions (Figs. 2a bottom and 2b). Similar values of the surface electric field, generated by shallow p-n junctions, with a maximum strength of E =3×107 V/m have been reported in the past34. The fringing field strength is tuned in the EFN device with VJG whose practical range lays between -2 V and 0.5 V; the onset of a significant p-n junction diffusion current (for VJG > 0.5 V) determines the upper limit while the depletion of electrons in the n-doped Si channel (for VJG < -2 V) accounts for the lower limit of VJG, respectively (Fig. S3).

Figure 2. Finite element simulations of the electric fields. (a) Cross-sections (x-z planes) of the 3D-modeled EFN device show the electric field distributions for VJG= -1.5 V (top) and VJG= 0.5 V (bottom) while the other electrodes were grounded. (b) The simulated electric field strength decays with the distance from the sensor surface (Z-axis) above the p-n junction of the EFN device, marked by the dashed lines in (a).

When the EFN sensor is exposed to a fixed vapor concentration at RT, the current decreases for negative (reverse bias) and increases for positive (forward bias) junction gate voltages as a function of time; this is demonstrated for the detection of ethanol vapor (2000 ppm) (Fig. 3a). The measurement cycle begins in absence of the analyte in a dry nitrogen atmosphere (Fig. 3a, state 1). The linear change in current (Fig. 3a, state 2) during sensor exposure to the analyte (here ethanol) indicates a constant reaction rate and displays sensing in a regime, at which the surface is not yet saturated by adsorbed ethanol. Sensing in this regime facilitates the analysis of the fringing-field-controlled response because interfering factors such as non-linear adsorption kinetics, saturation of adsorption and transistor channel depletion are omitted. Figures 3a and S4 show that the reaction rate, indirectly measured as the change in current, ∆I D , after a given amount of time, depends on VJG and thus the fringing field strength. After the inlet vapor flow stops, the current remains constant at RT in nitrogen atmosphere (Fig. 3a, state 3) indicating a thermodynamically stable surface adsorbate, as there is no analyte desorption. This allows the comparison of the intrinsic sensor response for different fringing field magnitudes (i.e., different VJG) because, in absence of analyte in the inert gas chamber atmosphere, the sensor surface condition is not affected by the FET bias configuration (not shown). The initial current value (Fig. 3a, state 1) is recovered following device heating to ≥ 70°C in ambient at 41 % relative humidity (RH), equivalent to ∼ 10850 ppm H2O at 22° C (Fig. S5). Qualitatively the same behavior as shown in Figure 3a is also observed for other organic vapors including

non-polar hexane (Fig. S9). The sensor response is defined I −I as R = 0 = ∆I D / I 0 , where I0 and I are the measured I0 source-drain currents before and after analyte exposure, respectively. Transducer sensitivity and fringing fieldcontrolled analyte-surface interaction are both controlled with VJG. To factor out the two effects, the sensor response is determined from ID-VBG characteristics after a time-based measurement in nitrogen atmosphere. ∆I D is retrieved from the ID-VBG plots in the sub-threshold regime, at which the sensor is most sensitive (Fig. S6). This mode of operation implies a nearly fully depleted channel, in agreement with previous reports on nanowire FET-based sensors5, 35. Figure 3b shows ID-VBG characteristics before (black) and after sensing at VJG= -1.5 V (blue), as well as after sensing at VJG= 0.5 V (red). The exposed dielectric layer of an FET-based sensor acts as a molecular gate because adsorbed molecules (here ethanol) may change the surface charge density, σ , affecting the sensor work function, Φ s , and thus alter the carrier density in the semiconductor (here n-doped Si) via field-effect. The threshold voltage shift with respect to the back gate, ∆V thBG , upon surface interaction with the analyte is related to changes in Φ s . Therefore, the measured ∆V thBG of (-0.17 ± 0.04) V and (1.05 ± 0.03) V, following ethanol exposure at VJG= 0.5 V (red) and VJG= -1.5 V (blue) are attributed to a more positive and negative σ , respectively. The sensor response (black) to a fixed ethanol vapor concentration of 2000 ppm is

ACS Paragon Plus Environment

ACS Sensors 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

proportional to VJG (before it saturates) and correlates with the simulated fringing electric field strength (red) (Fig. 3c). The fringing field strength saturates for VJG < -1.5 V as a consequence of the finite width of the n-doped Si region (∼ 750 nm) limiting the lateral extension of the depletion layer and the amount of ionized impurities therein. The fringing electric fields are strongest above the p-n junctions inferring local variations of the surface charge density. KPFM allows mapping of the contact potential difference, defined as the difference between the work functions of the AFM tip, Φ tip , and the sample, Φ s , divided by the elementary charge, e: CPD = ∆Φ = (Φ s − Φtip ) / e . Changes in σ and the electron affinity, χ , affect the local vacuum level, Evac, and thus Φ s = Evac − EF , probed by KPFM, where EF denotes the Fermi level.18, 36 Different from previous KPFM studies of biased p-n junctions in a humid atmosphere24, 37, 38, the measurements in this work were conducted in a nitrogen glove box with less than 2 ppm H2O content. The contrast in the CPD image of the grounded EFN transistor before analyte exposure (Fig. 4a, top) results from different work functions of the n- and p-doped Si regions in thermal and chemical equilibrium (see also section 3 of the SI). The FET sensor was exposed to ethanol (≥ 3000 ppm) while the p-n junctions were reverse biased (Fig. 4b, top) resulting in a positive sensor response (Fig. 4b, bottom). The measured CPD map following ethanol exposure reveals the depletion regions of the p-n junctions underneath the SiO2 layer (Fig. 4c, top) and clearly differs from the CPD image of the grounded device before ethanol exposure (Fig. 4a, top).

Figure 3. Ethanol vapor detection with the EFN sensor. (a) IDtime plots under reverse (blue) and forward (red) bias of the junction gates, VJG, before and after exposure to 2000 ppm of ethanol vapor in a dry nitrogen atmosphere. To minimize the influence of the back gate and drain potentials, the applied voltages were VBG= 0 V and VD= 0.05 V during sensing. (b) ID-VBG characteristics at VJG= -2 V and VD= 1 V for the following cases: in nitrogen before ethanol exposure (black), and after ethanol exposure at VJG= -1.5 V (blue) and VJG= 0.5 V (red) for the sensing experiments at state 3 shown in (a). (c) Plots of the sensor response, ∆ID/I0, (black) and the simulated fringing field maxima (red) as a function of VJG.

ACS Paragon Plus Environment

Page 4 of 9

Page 5 of 9 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

ACS Sensors recovery with the CPD map of the pristine sensor surface before ethanol exposure (Fig. 4a, top), is restored after heating (70 °C) in ambient (Fig. 4d, top). This is in qualitative agreement with the I-V measurements (Fig. S5). Reversibility of the sensor response generally suggests weak interaction with the surface, i.e. physisorption. However, in this instance, regeneration of the sensor is aided by heating in ambient, whereas attempts to regenerate by heating under inert atmosphere failed to reestablish the initial surface condition. This suggests a potential chemisorption process that is reversed by gentle thermolysis in the presence of ambient H2O.

Figure 5. CPD profiles across the p-n-p structure of the EFN transistor before ethanol adsorption (black), and after ethanol adsorption at VJG= -0.5 V (blue) and at VJG= 0.4 V (red), marked by the black, blue and red arrows in the CPD images of Figures 4 (a), (c) and (f), respectively.

Figure 4. KPFM of the contacted EFN transistor in a nitrogen glove box. CPD images of the EFN sensor surface (a) at ground potential and before ethanol exposure, (b) in-operando at VD= 1 V, VJG= -0.5 V and VBG= 0 V, and (c) at ground potential after ethanol exposure (≥ 3000 ppm). The bottom images show the sensor response as a function of time at bias configuration (b). CPD images of the EFN sensor surface (d) at ground potential, before ethanol exposure and after recovery, (e) in-operando at VD= 0.2 V, VJG= 0.4 V and VBG= 0 V, and (f) at ground potential after ethanol exposure (≥ 3000 ppm). The bottom images show the sensor response as a function of time at bias configuration (e).

The measured CPD pattern (Fig. 4c, top) coincides with the simulated fringing field distribution (Fig. S2) demonstrating a local and electric-field-controlled surface-analyte interaction. Note that VJG and VD both impact the sensor response (Fig. S7) as they both affect SCR and therefore the distribution and magnitude of the surface electric fields. Consequently, the measured CPD increases along the channel from source to drain (Fig. 4c, top) with a maximum at the pinch-off point near the biased drain region (VD= 1 V) where the depletion width is largest. The initial sensor surface condition, ascertained by the comparison of the CPD map after

The EFN sensor was exposed to ethanol under forward biased junction gates (Fig. 4e, top) resulting in a negative response (Fig. 4e, bottom) and a net decrease of the CPD at the n-doped channel near the source contact (Fig. 4f). Figure 5 compares the CPD profiles across the p-n-p structure of the sensor surface before (black), and after ethanol exposure at reverse (blue) and forward (red) biased junction gates, marked by the arrows in Figures 4a,c and f, respectively. Charging of the sensor surface following ethanol exposure at VJG= -0.5 V inverses the contrast between n and pdoped regions, which was previously observed39. The measured local variations in the CPD following ethanol adsorption are likely caused by both changes in χ of the dielectric surface and depletion in the moderately n-doped (ND= 4×1017 cm-3) Si channel underneath the SiO2 layer. ∆χ is attributed to the formation of a microscopic surface dipole layer40 (surface polarization) and is estimated from the measured CPD difference between the highly p-doped Si region before and after analyte adsorption. The field-effect induced change in the carrier density of the p-doped region, caused by the change in σ of the sensor surface following analyte exposure, is assumed to be negligible small due to the higher doping density (NA= 2×1019 cm-3) resulting in a Thomas-Fermi screening length of ~1 nm. The measured values of ∆χ and maximum changes in the work function, ∆Φ max = ∆CPDmax , as well as the quantities extracted from IV measurements are summarized in Table 1. Sensor response R and ∆VthBG , both extracted from the corresponding ID-VBG characteristics (Fig. S8), trend with the work function differ-

ACS Paragon Plus Environment

ACS Sensors 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

ences, ∆Φmax , of the sensor surface after ethanol adsorption. KPFM locally probes the factors (i.e., ∆χ and ∆σ ) that determine the (intrinsic) response of a chemical sensor. For the case of the EFN device, the sensor response to organic

Page 6 of 9

vapors correlates with the spatial distribution of the surface electric field strength. This characteristic is exploited to enhance the sensor selectivity.

Table 1. Change in work function and electron affinity obtained from CPD measurements, and macroscopic FET sensor characteristics retrieved from ID-VBG characteristics.

*

Veff (V)*

∆χ (meV)

∆Φmax (meV)

∆Vth BG (V)

R ( ∆I D / I 0 )

-1.5

75 ± 35

250 ± 25

1.33 ± 0.04

0.73 ± 0.06

0.2

40 ± 35

-50 ± 25

-0.15 ± 0.07

-0.23 ± 0.09

The effective p-n junction bias near the drain contact is defined as Veff = VJG - VD.

Selectivity between ethanol and n-butylamine is demonstrated using a single EFN transistor. Figure 6 shows the EFN sensor response to equal amounts (1500 ppm in dry N2) of n-butylamine (Fig. 6a) and ethanol (Fig. 6b). Under the conditions tested, the sensor is more sensitive to nbutylamine than to ethanol vapor despite its lower dipole moment (µButylamine = 1.0 D, µEthanol = 1.7 D)41. The two analytes feature distinct functional groups (HO-, and H2N-) enabling different pathways for interaction with the SiO2 surface (Fig. S13). While both analytes can interact with the various functional groups of the sensor surface (Si-OH, Si-O-Si) via physisorption (mainly H-bonding interactions, Fig. S13), it has been postulated that ethanol may additionally participate in extensive ethanol-ethanol H-bonding interactions42. This may weaken the overall analyte-surface interaction and dampen the sensor response. Both analytes can further engage in chemisorption interactions (Fig. S13), most notably in the case of n-butylamine, leading to the formation of amine salts (+H3N-)43-47. The formation of ion pairs on the sensor surface with n-butylamine as analyte could more prominently alter the charge of the sensor surface48 (Fig.

S11), thereby eliciting a stronger response compared to ethanol, as observed herein. Additionally, the surface electric field affects the reaction of the respective analyte with the sensor surface. The response to ethanol equals zero at VJG= 0.25 V (Fig. 3c) and E ~1×107 V/m. The sensor is set to be sensitive to nbutylamine and insensitive to ethanol at VJG= 0.2 V and VD= 0.2 V (Fig. 6c, dashed lines). Selecting the appropriate device working point (VJG= 0.2 V and VD= 0.4 V) facilitates discrimination between both molecules (Fig. 6c, solid lines). Electric field and temperature-controlled surface polarization with concomitant activation of binding sites has been reported32, 33, 49, 50. In this case, the strong fringing electric fields, generated by the shallow p-n junctions of the EFN device, may affect the reaction pathway of the respective analyte with the Si-OH and Si-O-Si binding sites, which are randomly distributed on a SiO2 surface. Surface polarization should bear significant influence on any subsequent surface-analyte interactions by varying the binding affinity or activation energy, which both are mainly determined by the functional group(s) of different molecules.

Figure 6. Sensor response as a function of time for exposure to 1500 ppm concentrations of (a) ethanol and (b) n-butylamine vapors at four different VJG, and at constant VD= 0.2 V and VBG= -20 V. (c) Comparison of the sensor response between ethanol and n-butylamine at VD= 0.2 V and VD= 0.4 V, at constant VJG= 0.2 V and VBG= -20 V.

ACS Paragon Plus Environment

Page 7 of 9 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

ACS Sensors

CONCLUSIONS In summary, we demonstrated that the fringing field strength above the FET-based sensor surface is precisely tuned with the junction gates providing control of the intrinsic sensor response. The fringing-field-dependent behavior of the EFN sensor is exploited to achieve selectivity between two volatile organic compounds, ethanol and n-butylamine. KPFM reveals the fringing-field-controlled analyte-surface interaction, which has not yet been explored with FET-based chemical sensors. KPFM elucidates the active sensor region that coincides with the fringing field strength utilizing ethanol as a marker to spatially resolve local variations of the surface electric field. The measured KPFM images are interpreted as sensitivity maps. The results strongly suggest that electric fields ≥ 0.5×107 V/m impact the interaction between vapor molecules and an exposed dielectric surface (here SiO2), and thus influence the intrinsic sensor response. In our future work, we employ EFN sensor arrays in combination with neural network analysis to realize an electronic nose for multi-component gas sensing. The concept of fringing-fieldcontrolled surface interaction is not only valuable to enhance sensitivity and selectivity of chemical sensors but it also motivates further fundamental research.

ASSOCIATED CONTENT Supporting Information Additional experimental details, finite element modeling of the doping density and the electric field distribution, current-voltage characteristics, Kelvin probe force microscopy measurements, and chemical reaction schemes are provided in the Supporting Information. The Supporting Information is available free of charge on the ACS Publications website.

AUTHOR INFORMATION Corresponding Author(s) *Yossi Rosenwaks - [email protected]

Present Addresses †Department of Materials Science and Engineering, Northwestern University, Evanston, Illinois 60208, United States

Author Contributions All authors have given approval to the final version of the manuscript. ‡These authors contributed equally.

ACKNOWLEDGMENT This work was supported by the Israel Ministry of Defense and the Meimad program of the Chief Scientist. AH acknowledges the scholarship from the Center for Nanoscience and Nanotechnology, Tel Aviv University.

REFERENCES 1. Bergveld, P. The impact of MOSFET-based sensors. Sensors and Actuators 1985, 8, 109-127. 2. Fahad, H. M.; Shiraki, H.; Amani, M.; Zhang, C.; Hebbar, V. S.; Gao, W.; Ota, H.; Hettick, M.; Kiriya, D.; Chen, Y.Z.; Chueh, Y.-L.; Javey, A. Room temperature multiplexed gas sensing using chemical-sensitive 3.5-nm-thin silicon transistors. Science Advances 2017, 3.

3. Im, H.; Huang, X.-J.; Gu, B.; Choi, Y.-K. A dielectricmodulated field-effect transistor for biosensing. Nat Nano 2007, 2, 430-434. 4. Shalev, G.; Landman, G.; Amit, I.; Rosenwaks, Y.; Levy, I. Specific and label-free femtomolar biomarker detection with an electrostatically formed nanowire biosensor. NPG Asia Mater 2013, 5, e41. 5. Rigante, S.; Scarbolo, P.; Wipf, M.; Stoop, R. L.; Bedner, K.; Buitrago, E.; Bazigos, A.; Bouvet, D.; Calame, M.; Schönenberger, C.; Ionescu, A. M. Sensing with Advanced Computing Technology: Fin Field-Effect Transistors with High-k Gate Stack on Bulk Silicon. ACS Nano 2015, 9, 4872-4881. 6. Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Nanowire Nanosensors for Highly Sensitive and Selective Detection of Biological and Chemical Species. Science 2001, 293, 1289-1292. 7. Zheng, G.; Patolsky, F.; Cui, Y.; Wang, W. U.; Lieber, C. M. Multiplexed electrical detection of cancer markers with nanowire sensor arrays. Nat Biotech 2005, 23, 1294-1301. 8. Ashkenasy, G.; Cahen, D.; Cohen, R.; Shanzer, A.; Vilan, A. Molecular Engineering of Semiconductor Surfaces and Devices. Accounts of Chemical Research 2002, 35, 121-128. 9. Engel, Y.; Elnathan, R.; Pevzner, A.; Davidi, G.; Flaxer, E.; Patolsky, F. Supersensitive Detection of Explosives by Silicon Nanowire Arrays. Angewandte Chemie International Edition 2010, 49, 6830-6835. 10. Zampolli, S.; Elmi, I.; Ahmed, F.; Passini, M.; Cardinali, G. C.; Nicoletti, S.; Dori, L. An electronic nose based on solid state sensor arrays for low-cost indoor air quality monitoring applications. Sensors and Actuators B: Chemical 2004, 101, 39-46. 11. Schweizer-Berberich, P. M.; Vaihinger, S.; Göpel, W. Characterisation of food freshness with sensor arrays. Sensors and Actuators B: Chemical 1994, 18, 282-290. 12. Fleischer, M.; Simon, E.; Rumpel, E.; Ulmer, H.; Harbeck, M.; Wandel, M.; Fietzek, C.; Weimar, U.; Meixner, H. Detection of volatile compounds correlated to human diseases through breath analysis with chemical sensors. Sensors and Actuators B: Chemical 2002, 83, 245-249. 13. Kahn, N.; Lavie, O.; Paz, M.; Segev, Y.; Haick, H. Dynamic Nanoparticle-Based Flexible Sensors: Diagnosis of Ovarian Carcinoma from Exhaled Breath. Nano Letters 2015, 15, 7023-7028. 14. Shehada, N.; Cancilla, J. C.; Torrecilla, J. S.; Pariente, E. S.; Brönstrup, G.; Christiansen, S.; Johnson, D. W.; Leja, M.; Davies, M. P. A.; Liran, O.; Peled, N.; Haick, H. Silicon Nanowire Sensors Enable Diagnosis of Patients via Exhaled Breath. ACS Nano 2016, 10, 7047-7057. 15. Onclin, S.; Ravoo, B. J.; Reinhoudt, D. N. Engineering Silicon Oxide Surfaces Using Self-Assembled Monolayers. Angewandte Chemie International Edition 2005, 44, 6282-6304. 16. Blalock, B. J.; Cristoloveanu, S.; Dufrene, B. M.; Allibert, F.; Mojarradi, M. M. The Multiple-Gate MOS-JFET Transistor. International Journal of High Speed Electronics and Systems 2002, 12, 511-520. 17. Dufrene, B.; Akarvardar, K.; Cristoloveanu, S.; Blalock, B. J.; Gentil, R.; Kolawa, E.; Mojarradi, M. M. Investigation of the four-gate action in G4 FETs. IEEE Transactions on Electron Devices 2004, 51, 1931-1935. 18. Shaya, O.; Shaked, M.; Usherenko, Y.; Halpern, E.; Shalev, G.; Doron, A.; Levy, I.; Rosenwaks, Y. Tracing the Mechanism of Molecular Gated Transistors. The Journal of Physical Chemistry C 2009, 113, 6163-6168. 19. Henning, A.; Swaminathan, N.; Godkin, A.; Shalev, G.; Amit, I.; Rosenwaks, Y. Tunable diameter electrostatically formed nanowire for high sensitivity gas sensing. Nano Research 2015, 8, 2206-2215. 20. Swaminathan, N.; Henning, A.; Jurca, T.; Hayon, J.; Shalev, G.; Rosenwaks, Y. Effect of varying chain length of nalcohols and n-alkanes detected with electrostatically-formed nanowire sensor. Sensors and Actuators B: Chemical 2017, 248, 240-246.

ACS Paragon Plus Environment

ACS Sensors 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

21. Henning, A.; Molotskii, M.; Swaminathan, N.; Vaknin, Y.; Godkin, A.; Shalev, G.; Rosenwaks, Y. Electrostatic Limit of Detection of Nanowire-Based Sensors. Small 2015, 11, 4931-4937. 22. Swaminathan, N.; Henning, A.; Vaknin, Y.; Shimanovich, K.; Godkin, A.; Shalev, G.; Rosenwaks, Y. Dynamic Range Enhancement Using the Electrostatically Formed Nanowire Sensor. ACS Sensors 2016, 1, 688-695. 23. Schroen, W. In Accumulation and Decay of Mobile Surface Charges on Insulating Layers and Relationship to Reliability of Silicon Devices, Fourth Annual Symposium on the Physics of Failure in Electronics, Nov. 1965; 1965; pp 291-314. 24. Shockley, W.; Queisser, H. J.; Hooper, W. W. Charges on Oxidized Silicon Surfaces. Physical Review Letters 1963, 11, 489-490. 25. Metz, E. D. In Silicon Transistor Failure Mechanisms Caused by Surface Charge Separations, Second Annual Symposium on the Physics of Failure in Electronics, Sept. 1963; 1963; pp 163172. 26. Atalla, M. M.; Bray, A. R.; Lindner, R. Stability of thermally oxidized silicon junctions in wet atmospheres. Proceedings of the IEE - Part B: Electronic and Communication Engineering 1959, 106, 1130-1137. 27. Sakakibara, K.; Ajika, N.; Eikyu, K.; Ishikawa, K.; Miyoshi, H. A quantitative analysis of time-decay reproducible stress-induced leakage current in SiO2 films. IEEE Transactions on Electron Devices 1997, 44, 1002-1008. 28. Olivo, P.; Nguyen, T. N.; Ricco, B. High-field-induced degradation in ultra-thin SiO2 films. IEEE Transactions on Electron Devices 1988, 35, 2259-2267. 29. Sakakibara, K.; Ajika, N.; Hatanaka, M.; Miyoshi, H. In A quantitative analysis of stress induced excess current (SIEC) in SiO2 films, Proceedings of International Reliability Physics Symposium, April 30 1996-May 2 1996; 1996; pp 100-107. 30. Someshwar, A. V.; Wilkinson, B. W. Study of electric field-induced effects on water vapor adsorption in porous adsorbents. Industrial & Engineering Chemistry Fundamentals 1985, 24, 215-220. 31. Hoenig, S. A.; Lane, J. R. Chemisorption of oxygen on zinc oxide, effect of a dc electric field. Surface Science 1968, 11, 163-174. 32. Fan, Z.; Lu, J. G. Gate-refreshable nanowire chemical sensors. Applied Physics Letters 2005, 86, 123510. 33. Mikheeva, E. P. The influence of electric fields on the chemisorption and catalytic properties of semiconductor catalysts. Russian Chemical Reviews 1989, 58, 517. 34. Davies, R. L.; Gentry, F. E. Control of electric field at the surface of P-N junctions. IEEE Transactions on Electron Devices 1964, 11, 313-323. 35. Gao, X. P. A.; Zheng, G.; Lieber, C. M. Subthreshold Regime has the Optimal Sensitivity for Nanowire FET Biosensors. Nano Letters 2010, 10, 547-552. 36. Paska, Y.; Stelzner, T.; Assad, O.; Tisch, U.; Christiansen, S.; Haick, H. Molecular Gating of Silicon Nanowire Field-Effect Transistors with Nonpolar Analytes. ACS Nano 2012, 6, 335-345. 37. Domanský, K.; Leng, Y.; Williams, C. C.; Janata, J.; Petelenz, D. Mapping of mobile charges on insulator surfaces with the electrostatic force microscope. Applied Physics Letters 1993, 63, 1513-1515. 38. Gouveia, R. F.; Costa, C. A. R.; Galembeck, F. Water Vapor Adsorption Effect on Silica Surface Electrostatic Patterning. The Journal of Physical Chemistry C 2008, 112, 17193-17199. 39. Kalinin, S. V.; Bonnell, D. A. Screening Phenomena on Oxide Surfaces and Its Implications for Local Electrostatic and Transport Measurements. Nano Letters 2004, 4, 555-560. 40. Atanassova, E.; Dimitrova, T. Electrical studies of the reactions between ethanol vapours and inversion mode Si-thin SiO2 structures. Solid-State Electronics 1993, 36, 1711-1716. 41. David, R. L. CRC Handbook of Chemistry and Physics, Internet Version 2005. CRC Press, Boca Raton, FL: 2005.

42. Wu, D.; Guo, X.; Sun, H.; Navrotsky, A. Energy Landscape of Water and Ethanol on Silica Surfaces. The Journal of Physical Chemistry C 2015, 119, 15428-15433. 43. Luts, T.; Katz, A. Chemisorption and Dehydration of Ethanol on Silica: Effect of Temperature on Selectivity. Topics in Catalysis 2012, 55, 84-92. 44. Grime, J. K.; Sexton, E. D. Enthalpimetric study of the surface interactions of n-butylamine with silica gel. Analytical Chemistry 1982, 54, 902-907. 45. Blomfield, G. A.; Little, L. H. Chemisorption of Ammonia on Silica. Canadian Journal of Chemistry 1973, 51, 1771-1781. 46. Azrak, R. G.; Angell, C. Study of alcohol-silica surface reactions via infrared spectroscopy. The Journal of Physical Chemistry 1973, 77, 3048-3052. 47. Clark-Monks, C.; Ellis, B. n-Butylamine Adsorption onto Amorphous Silica. Canadian Journal of Chemistry 1972, 50, 907911. 48. Vol'kenshteĭn, F. Electronic processes at the surface of a semiconductor during chemisorption. Physics-Uspekhi 1967, 9, 743751. 49. Volkenstein, F.; Sandomirsky, V. The Influence of an External Electric Field on the Adsorptive Ability of a Semiconductor. Doklady Akademii Nauk SSSR 1958, 118, 980-982. 50. Bashouti, M. Y.; Yousefi, P.; Ristein, J.; Christiansen, S. H. Electronic Properties of Si–Hx Vibrational Modes at Si Waveguide Interface. The Journal of Physical Chemistry Letters 2015, 6, 3988-3993.

ACS Paragon Plus Environment

Page 8 of 9

Page 9 of 9 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

ACS Sensors

TOC graphic

ACS Paragon Plus Environment