Stable and scalable 1T MoS2 with low temperature-coefficient of resistance

Monolithic realization of metallic 1T and semiconducting 2H phases makes MoS2 a potential candidate for future microelectronic circuits. A method for engineering a stable 1T phase from the 2H phase in a scalable manner and an in-depth electrical characterization of the 1T phase is wanting at large. Here we demonstrate a controllable and scalable 2H to 1T phase engineering technique for MoS2 using microwave plasma. Our method allows lithographically defining 1T regions on a 2H sample. The 1T samples show excellent temporal and thermal stability making it suitable for standard device fabrication techniques. We conduct both two-probe and four-probe electrical transport measurements on devices with back-gated field effect transistor geometry in a temperature range of 4 K to 300 K. The 1T samples exhibit Ohmic current-voltage characteristics in all temperature ranges without any dependence to the gate voltage, a signature of a metallic state. The sheet resistance of our 1T MoS2 sample is considerably lower and the carrier concentration is a few orders of magnitude higher than that of the 2H samples. In addition, our samples show negligible temperature dependence of resistance from 4 K to 300 K ruling out any hoping mediated or activated electrical transport.

electron bombardment in a transmission electron microscope falls short in the yield and in the adaptability required for the microelectronics industry.
There is a lack of an in-depth electrical characterization of metallic 1T MoS 2 . Existing reports confine to two-probe (2P) transport measurements on polymorphic samples 11,25,34 . A four-probe (4P) electrical characterization is essential since 2P measurements are influenced by the behaviour of contacts. A linear current-voltage (I-V) characteristics, void of gate-voltage dependence, down to the cryogenic temperatures is necessary in establishing a metallic nature. The 1T samples obtained by Argon plasma 34 treatment have shown response to gate voltages, atypical of a metallic state and, in contrast to those reported elsewhere 11,25 . Temperature dependent transport measurements are reported only down to 100 K, however, the sample shows a large increase in the resistance as the temperature is lowered 11 . A table summarizing the existing studies on electrical transport properties, scalability and stability on 1T MoS 2 is provided in Supplementary Information S1.
In this manuscript, we demonstrate a method to engineer 1T phase on 2H MoS 2 samples exfoliated from bulk crystals with arbitrary thickness and area, in a controllable and scalable manner. Our process involves treating mechanically exfoliated samples with high-power forming-gas microwave plasma which results in a layer-by-layer thinning accompanied by a structural phase conversion from 2H to 1T. The presence of plasma etching helps us to realize few-layer 1T MoS 2 samples starting from thicker exfoliated samples. We show that our technique can be used to selectively engineer the 1T phase on 2H MoS 2 samples with the help of standard lithography techniques. We perform an in-depth structural characterization using high-resolution transmission electron microscopy (HR-TEM) and Raman spectroscopy. We also examine the evolution of photoluminescence (PL) spectra as a function of plasma treatment to study the phase transition. Unlike the intercalation route, our process is faster and we do not find signatures of 1T′ or 1T″ phases. Our process yields extended 1T regions with an areal coverage in excess of 70% over the 2H phase. The 1T samples show a temporal stability in excess of a few weeks and a thermal stability up to 300 °C in ambience. We conduct 2P transport studies on lithographically defined 2H and 1T regions on the same sample for a direct comparison of electrical properties. We also perform 4P electrical transport studies on a few layer 1T MoS 2 sample. The contacts on the 1T phase show a clear Ohmic behaviour at all temperatures from 300 K down to 4 K and the transport show little response to the gate voltage, indicative of a metallic phase. The carrier concentration ~10 15 cm −2 and sheet resistance ~108 /□ suggest that the sample is in the metallic regime. In addition, our sample also qualifies the Ioffe-Regal criteria for metallic conduction. We observe negligible temperature coefficient of resistance down to 4 K for the 1T phase unlike other reports on 1T MoS 2 11 , ruling out hoping-mediated or activated transport in our samples.

Results and Discussion
The MoS 2 samples, exfoliated from bulk crystals are transferred on to Silicon substrates hosting a 300 nm SiO 2 layer. These samples are treated with forming gas (10% H 2 + 90% Ar) microwave plasma with 40% input power to the magnetron 37 . The plasma treatment results in a layer-by-layer etching of the sample accompanied by a structural phase transformation from the 2H to the 1T phase. Lower microwave power levels result only in a layer-by-layer etching and do not yield any phase change.
TEM analysis. The crystallinity of the plasma treated samples is examined using HR-TEM. Figure 1(a) shows HR-TEM image of a representative plasma treated few-layer MoS 2 sample. The regions shaded in purple, green and brown represent the 2H, 1T and, an intermediate state between the 2H and the 1T phases respectively. Based on our TEM analysis conducted over many samples, we estimate a lateral coverage for the 1T phase in excess of 70%. We believe, this is a lower bound to the coverage since we also observe a back conversion of the 1T to the more stable 2H phase under prolonged exposure of high energy electron beam during imaging (Supplementary Information S2). We also note that the HR-TEM images do not show signatures of other distorted structural phases such as 1T′ or 1T″. The selected area electron diffraction (SAED) pattern shown in the top-right inset exhibit sharp diffraction spots, inferring good crystallinity of our samples. HR-TEM images of two more samples showing extended 1T regions are shown in Supplementary Information S3. Figure 1(b) shows a magnified image of the 2H region while Fig. 1(c) shows that of the 1T region obtained from the same sample as in Fig. 1(a). The lower panels in Fig. 1(b,c) shows the intensity line-profiles obtained along the directions indicated by the dashed-lines in the respective panels. The arrangements of Mo and S atoms for the 2H and 1T phases are shown in the overlaid diagrams in Fig. 1(b,c) respectively. In the 2H phase the S-Mo-S atoms are arranged in an A-B-A stacking fashion along the c -axis. Mo atoms appear brighter in intensity compared to the S atoms in the TEM images owing to its higher atomic number 38 . Due to this reason, the intensity for the S-peaks is very weak for monolayer MoS 2 . For a few-layer 2H MoS 2 , the position of S atoms in one layer coincides with that of the Mo atoms in the adjacent layer. This gives an appreciable intensity for the peaks corresponding to the S sites 39 , as evident from the intensity profile shown in the bottom panel of Fig. 1(b). We extract a nearest Mo-Mo separation of 3.19 (+/−0.08) A 0 from the HR-TEM images.
In the 1T phase, the Mo atom is octahedrally coordinated with six S atoms with the S-Mo-S in an ABC stacking fashion. In this case, atoms in one layer align with the corresponding atoms in the adjacent layer. This arrangement makes the intensity of the peaks corresponding to the S atoms in 1T phase much weaker compared to that of the 2H phase in the TEM images. As seen in the bottom panel of Fig. 1 Fig. 1(d). 2H and the 1T regions exhibit intensity profiles similar to Fig. 1(b,c) respectively. We also note that the relative intensity of the Mo peaks in the 1T region is higher than that of the 2H region; possibly due to the difference in the alignment of Mo atoms, corresponding to different layers, along the c-axis.
The transformation between the 2H and the 1T phases involves an intra-layer S plane gliding 12,27 . For a few layer MoS 2 sample, transition between the 2H and 1T phases also require an Mo-plane gliding. In support of this we observe in a few of our HR-TEM images, an intermediate atomic arrangement between the 2H and the 1T phases, as shown in Fig. 1(e). The visible stripe-like patterns in Fig. 1(e) is due to the rearrangement of Mo and S atoms during the transformation 28 . A possible atomic arrangement is shown in the inset to Fig. 1(e). We note here that this transformation can also happen to regions subject to repetitive TEM imaging as a result of prolonged exposure to high energy electrons [Supplementary Information S2].
We also examine HR-TEM image of a pristine 2H MoS 2 sample taken with the same exposure parameters as those for the 1T samples. Figure 1(f) shows the HR-TEM image and the inset shows a magnified view of the Mo and S atomic arrangements depicting the 2H phase. The images did not show presence of any other structural phases confirming the absence of electron-beam induced 2H to 1T phase transition in our samples.
Raman and PL studies. We perform Raman scattering and PL studies on plasma treated samples. These samples consist of regions with different thickness starting from a few nanometres to a few tens of nanometres prior to the plasma treatment. Figure 2  shows the pristine exfoliated sample while the bottom panel shows the image after a 7.5 minutes of plasma treatment. We have conducted Raman scattering studies on all the regions. Here we focus on regions labelled I and II. Figure 2(b,c) shows the Raman spectra from region I [region II]. The black(red) traces in both Fig. 2(b,c) represent Raman spectra taken before (after) the plasma treatment. Samples post-plasma treatment (red traces) show clear J 1 and J 2 vibrational modes corresponding to the 1T phase (The Raman scattering studies conducted on other regions are shown in Supplementary Information S5).
To demonstrate the controllability and scalability of the process we use the Aluminium masking technique to selectively phase engineer the sample.  Figure 2(e) shows the PL spectra of the sample before (black) and PL spectra from region II after (red) the plasma treatment. The sample is in excess of ~50 nm in thickness prior to plasma treatment and exhibits only a weak excitonic peak (black trace). The plasma treated region (II), reduced to 6 layers in thickness, does not exhibit any PL (red trace) in contrast to a pristine six-layer 2H MoS 2 shown in blue-trace 41 . Raman spectra of the sample before and after the plasma treatment are shown in the inset. Only region II, exposed to the plasma, develops the J 1 and the J 2 peaks (red trace) while the Raman spectra of region I post plasma treatment (purple trace) is akin to that of the sample prior to the plasma treatment. The quenching of the PL spectra, post plasma treatment, as a result of a semiconducting to metallic phase transition is shown in Fig. 2(f). The optical images of the sample before and after plasma treatment is shown in the insets. The black-trace shows the PL spectra obtained from region I (~6 layers) and the red trace shows the PL spectra from the same region after the plasma treatment (~4 layers). While the PL spectra undergo a substantial quenching, the Raman spectra, shown in the inset, develop the characteristics J 1 and J 2 peaks as a result of the plasma treatment. Post plasma treatment, our samples exhibit clear J 1 and J 2 peaks while the J 3 peak is very weak in intensity. In addition, we observe emergence of a peak at 180 cm −1 which was not reported in the past. The J 1 and the J 3 peaks are predicted to be much lower in intensity compared to the J 2 peak 42 . The relative intensities and the peak positions vary from sample to sample and process to process 21,29,32,34,43 . 1T MoS 2 prepared by chemical routes exhibit a weaker J 3 29,43 . 1T MoS 2 prepared by physical routes exhibit strong J 2 while the J 3 is very weak. Both the J 1 and J 2 peaks are clearly visible on electron-beam irradiated and Argon RF plasma treated samples while J 3 is not well formed 32,34 . Now we discuss the stability of the phase engineered 1T samples. Figure 3(a) shows the optical images of the sample before and after plasma treatment. Raman spectra from the region I of the sample before (black-trace) and after (red-trace) the plasma treatment is shown in Fig. 3(b). The green-trace shows Raman spectra obtained from the same region after keeping the sample for 27 days at ambient temperature and pressure. Both the red-trace and the green-trace show J 1 , J 2 and J 3 peaks with similar shape and intensity suggesting good temporal stability of our samples. We also note here that the HR-TEM images taken after 30 days of plasma treatment also show rich concentration of the 1T phase suggesting good temporal stability of our samples. Now we explore the stability of the 1T samples as a function of the annealing temperature. Figure 3(c) shows the PL spectra taken from the region labelled I of the sample shown in Fig. 2(f) post plasma treatment at room temperature (red-trace), annealed at 300 °C for 5 min (cyan-trace) and 15 min (magenta-trace). The black trace shows the PL spectra of the same region, ~6 layers in thickness, prior to the plasma treatment. The inset shows Raman spectra of region I at room temperature (red), annealed at 300 °C for 5 min (cyan) and for 15 min (magenta). We infer from the PL and Raman spectra evolution that our 1T samples are stable up to 300 °C in temperature, which is well above the range of temperatures for standard device fabrication processes. As evident from the HR-TEM images, our samples also contains regions in 2H phase which might be providing stability to the 1T phase 32 . Transport studies. Electrical transport studies are conducted on two kinds of devices. (1) On a device where a 1T region is lithographically defined on a pristine 2H MoS 2 , for a direct comparison of electrical properties of both phases on the same sample. (2) On a fully phase engineered 1T MoS 2 device. Figure 4(a) shows the optical image of the sample on which the central region enclosed between the dashedlines (labelled 2H) is covered using a lithographically defined Al mask, and the flanked regions labelled 1T are exposed to the plasma and converted to the 1T phase. Post plasma treatment the mask is removed using NaOH (0.1 N) and Cr/Au source and drain contacts are fabricated onto both the 1T and the 2H regions. The Raman spectra of the 1T region shown in the Supplementary Information S6 exhibit the signature peaks, the J 1 , J 2 and J 3 , of the 1T phase. Figure 4(b) shows the 2P I-V characteristics of the 2H region at 300 K (red trace), 77 K (green trace) and 4 K (blue trace). The I-V characteristics of the 2H-region exhibit a Schottky behaviour at all temperatures and the span of the non-linearity increases as the temperature is lowered down to 4 K. We note that Cr/Au contacted MoS 2 generally exhibit non-linear I-V characteristics 14,37 , which has also been verified by us independently. The inset to Fig. 4(b) shows the 2P conductance of the 2H region as a function of the back-gate voltage, V bg , at 300 K (red trace) and 4 K (blue trace). The device exhibits clear n-type behaviour with a field effect mobility of 16.4 cm 2 /V-s at 300 K and 84 cm 2 /V-s at 4 K, which are in the range of typically observed mobility values for an uncapped, back-gated 2H MoS 2 FET 44 . In contrast, the I-V characteristics of the plasma treated region, 1T, shown in Fig. 2(c), exhibits excellent Ohmic behaviour at all temperatures down to 4 K. The inset shows the conductance of the 1T region as a function of V bg . The device shows little change in conductance as V bg is varied in a large voltage range of −20 to 40 V at 300 K (red trace) and −10 to 20 V at 4 K (blue trace). We also note that the 2P resistance of the 1T region shows only a small change (~12 Ohms) as the sample was cooled down to 4 K from 300 K while that of the 2H region shows a large variation in excess of three orders in magnitude. To exclude any contribution from the contact resistance to the electrical characteristics we conduct 4P transport measurements on an ~8 nm thick 1T phase engineered sample. The optical image of the device is shown in Fig. 4(d). The 4P I-V characteristics of the sample at 300 K (red trace) and 4 K (blue trace) are shown in Fig. 4(e). The I-V characteristics show clear Ohmic behaviour down to 4 K. We observe a feeble change in the resistance as the device is cooled down to 4 K; we extract a temperature coefficient of resistance, α = −1.1057 × 10 −4 K −1 using the equation T 300K . The inset shows the four-probe conductance of the device as a function of V bg . Both at 300 K (red trace) and at 4 K (blue trace) the conductance of the device does not show any response to V bg for a large range of voltage. Figure 4(e) shows 2P I-V characteristics of the voltage (8 & 9) and current (7 & 13) probes of the device at 4 K, both exhibiting excellent linearity. From the 4P resistance we extract a sheet resistance = 108 /□. A carrier concentration of ~2.3 × 10 15 cm −2 is extracted from the Hall resistance shown in the inset to Fig. 4(e). We note here that the carrier concentration obtained for our 1T samples are higher by two orders of magnitude compared to that of back-gated 2H MoS 2 samples 14 and higher approximately by an order compared to that of the ionic-liquid-gated 2H MoS 2 samples 45 . The disorder limit for metallic conduction is defined by the Ioffe-Regal criteria 46 , k F . l > 1 where k F is the Fermi wave vector and l is the electronic mean free path. From the carrier concentration and sheet resistance we find k F . l ~ 239 for our sample ascertaining metallic conduction. The resistance of the device shown in Fig. 4(d) taken at various durations after the sample processing is tabulated in Supplementary Information S9. We observe that, over a period of ~90 days, in which ~30 days in ambient conditions, the resistance of the sample did not show much variation underlining the stability of our sample in ambience as inferred from the Raman and TEM analysis.

Conclusions
In this manuscript, we demonstrated a controllable and scalable 2H to 1T phase conversion technique for MoS 2 . The process involves treating exfoliated 2H MoS 2 of arbitrary thickness with forming-gas microwave plasma. We performed an in-depth structural analysis using HR-TEM and Raman microscopy. Our processed samples consist mostly 1T phase. We did not find presence of other commonly observed phases such as 1T′ or 1T″. We observed the evolution of the signature Raman peaks of 1T MoS 2 accompanied by quenching of the PL on plasma treatment, indicative of a metallic phase formation. We believe that the phase transition happens by the gliding of the S and Mo planes; HR-TEM images also show signatures of plane gliding in support of this. The momentum transfer from the ions in the plasma could be twisting the Mo-S bonds. Similar phase transition mechanism caused by momentum transfer induced plane gliding under the exposure of energetic ions and electrons are reported elsewhere 12,34 . Those processes yielded nanoscale patches as opposed to extended 1T regions and also are prone to defects and disorder. The reducing atmosphere provided by the H 2 and the microwave heating could be annealing the defects, if any, resulting in extended low-disordered 1T regions in this case. In support of this, we note that the samples treated with plasma in the absence of H 2 did not show the J 1 , J 2 and the J 3 peaks [Supplementary Information S10] and, these samples are found to be highly resistive. Our 1T samples withstood aging for more than a month and also showed a thermal stability up to 300 °C, making it suitable for standard device fabrication techniques. The transport measurements conducted on the same sample over various durations show negligible change in resistance. We demonstrated lateral monolithic integration of metallic 1T and semiconducting 2H phases with the help of standard lithography techniques. We have conducted extensive transport characterization of our 1T samples from 300 K down to 4 K. Both the 2P and the 4P I-V characteristics showed excellent linearity down to 4 K and did not exhibit any response to the back-gate for a large span of voltage. Our 1T samples are electron doped, showed a carrier concentration a few orders higher and the resistance considerably lower than that of the 2H samples. Electron doping is shown to stabilize the 1T phase in MoS 2 31,47 which could be providing the stability to the 1T phase in our samples. A linear I-V characteristic, without gate-voltage dependence suggests the presence of a metallic state and Schottky-barrier-free source and the drain contacts 48,49 . The feeble temperature dependence shown by the 1T samples from 300 K down to 4 K also negates any barrier formation at the source and drain contacts 50 . The negligible temperature dependence also rules out any hoping mediated and activated transport in our system and suggests that the samples consists of extended 1T regions as evident from the HR-TEM analysis; transport on polymorphic MoS 2 had shown strong temperature dependence due to hopping transport between 1T islands 11 . The weak temperature dependence could also be due to the phonon-decoupling effect as observed elsewhere 51 . Scalability owing to the top-down nature, compatibility with the planar device fabrication schemes and high yield make this process a promising tool for 2D microelectronics industry.

Methods
Bulk MoS 2 samples for micromechanical exfoliation are procured from SPI supplies. The PDMS dry stamping technique 52 is used to transfer the samples on to the Si/SiO 2 substrate. The thickness of the flakes is estimated using the red-channel optical contrast method 37 . Technical details of the plasma reactor is discussed elsewhere 37 .
The HR-TEM analysis of the samples are carried out using FEI Tecnai FEG30 E Spirit transmission electron microscope with an accelerating voltage of 200 kV. The samples are transferred onto a TEM grid from the Si/SiO 2 substrate using the PMMA transfer technique 53 . The Raman and PL spectra are recorded using a Horiba XploRA PLUS Raman microscope using a laser of wavelength 532 nm with a power of 1.5 mW.
For selective area phase conversion, we exploit a lithographically defined Aluminium mask during plasma treatment. Post plasma treatment, the mask is removed using 0.1 N NaOH solution. Field effect transistors are fabricated for exploring the electrical properties of the material; MoS 2 film acts as the channel, while the highly doped silicon substrate and the silicon oxide layer serves as the back-gate and the gate-dielectric respectively. The source and the drain contacts are defined using standard electron-beam or photo lithography followed by Cr/Au metallization. The transport measurements are performed in high vacuum (<10 −6 mbar) dark environment in a closed-cycle cryostat.