npj | 2D materials and applications Article

Published in partnership with FCT NOVA with the support of E-MRS

https://doi.org/10.1038/s41699-024-00503-7

Optical control of multiple resistance

levels in graphene for memristic
applications
Check for updates
1 1 1 2 3
Harsimran Kaur Mann , Mainak Mondal , Vivek Sah , Kenji Watanabe , Takashi Taniguchi ,
Akshay Singh 1 & Aveek Bid 1

Neuromorphic computing has emphasized the need for memristors with non-volatile, multiple
conductance levels. This paper demonstrates the potential of hexagonal boron nitride (hBN)/graphene
1234567890():,;
1234567890():,;

heterostructures to act as memristors with multiple resistance states that can be optically tuned using
visible light. The number of resistance levels in graphene can be controlled by modulating doping
levels, achieved by varying the electric field strength or adjusting the duration of optical illumination.
Our measurements show that this photodoping of graphene results from the optical excitation of
charge carriers from the nitrogen-vacancy levels of hBN to its conduction band, with these carriers
then being transferred to graphene by the gate-induced electric field. We develop a qualitative model
to describe our observations. Additionally, utilizing our device architecture, we propose a memristive
crossbar array for vector-matrix multiplications.

Neuromorphic computing (NC), inspired by the human brain, has emerged Electrostatic doping involves doping using an external local gate. While
as a new paradigm beyond von Neumann computing. NC provides a low- this technique avoids most of the disadvantages of chemical doping listed
energy alternative to the traditional von Neumann architectures, promoting above, the magnitude of doping attainable is limited by the breakdown
sustainable computation in our modern information age. The critical voltage of the gate dielectric15–18. Liquid ionic gating has the potential of
hardware building blocks for NC are memristors (artificial synapse), neural reaching higher doping levels than is achievable by dielectric gates but at the
processing units, and threshold switches (artificial neurons)1–5. A memristor cost of device instability and non-scalability19,20.
is a fundamental electronic component whose resistance is dependent on A viable alternative is optical doping of graphene on hexagonal boron
the charge that has flown through it; it is a resistor with a memory. This nitride (hBN) substrates21–36. hBN has a band gap of 6 eV with several
charge-history dependence of resistance can be exploited in machine intermediate defect states that can be optically excited37–40. This method
learning frameworks where computation can be carried out via cross-bar involves the controlled optical excitation of charge carriers from defect states
arrays, and synaptic weights (such as weights for neural networks) can be of hBN and their transfer to graphene using an external electric field. This
modified by a certain number of electrical pulses, leading to in-memory technique is reversible without adversely affecting transport mobility and
computing. A low-power, non-volatile memristor is essential to exploit NC’s defect density.
benefits fully. In this article, we present an in-depth study of optical doping in hBN/
A prerequisite for memristic action is a platform with non-volatile graphene heterostructures using visible light. Our research reveals that
doping. Chemical and electrostatic doping are the two most commonly used electrons in the nitrogen-vacancy defect state in hBN can be optically excited
techniques to induce charge carriers in 2-D channels. Chemical doping is with violet light and transferred to graphene through electrostatic gating.
done by hetero-atom substitution and adsorption of molecular adsorbates This technique enables high-density doping of graphene, with the doping
onto graphene and other 2D materials. There are several drawbacks of this level controlled by gate voltage or illumination time. The dynamics of this
approach, including limited control over doping concentration, introduc- doping process, measured with millisecond temporal resolution, can be
tion of uncontrolled structural defects and lattice strains, unintentional adjusted by varying the illumination wavelength and optical power. Fur-
impurity introduction, challenges in maintaining stability and diffusion thermore, we demonstrate that a graphene/hBN heterostructure can func-
control, and the sensitivity of the sample to processing conditions6–14. tion as a memristor, exhibiting multiple resistance levels. The device’s

1
Department of Physics, Indian Institute of Science, Bangalore, 560012, India. 2Research Center for Electronic and Optical Materials, National Institute for Materials
Science, 1-1 Namiki, Tsukuba, 305-0044, Japan. 3Research Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1 Namiki, Tsukuba,
305-0044, Japan. e-mail: [email protected]; [email protected]

npj 2D Materials and Applications | (2024)8:69 1

https://doi.org/10.1038/s41699-024-00503-7 Article

compact structure and room-temperature operation enhance its potential Figure 1 (d) shows snapshots of the R − Vg data in 10 s intervals during
for scalable, room-temperature neuromorphic applications. the SET process. During this measurement, the light was turned on for 10 s
with V
g ¼ 2 V. The illumination was turned off, and the R − Vg curve was
Results and discussion measured. The process is repeated multiple times to generate the plots in Fig.
Single-layer graphene devices encapsulated between thin hBN crystals were 1d. The intensity of the light was kept very low (I = 4 Wm−2) to allow for a
fabricated using the dry transfer method. 1-D electrical contacts to the much slower rate of resistance change. One can see that the entire transfer
graphene were achieved by lithography and dry etching, followed by Cr/Au curve shifts gradually to the left until the Dirac point reaches V
g ¼ 2 V.
metallization (Fig. 1a)41. A back-gate voltage, Vg, tuned the charge carrier The results for similar measurements done with different values of V
g are
number density. Electrical transport measurements were performed using a shown in Fig. 1e. These measurements establish that the Dirac point can be
low-frequency AC measurement technique. The Dirac point (maxima in the moved deterministically to any value of the gate voltage either by controlling
device resistance R) is attained at Vg = 0 V (Fig. 1b), solid blue line, attesting the exposure time (Fig. 1d) or the value of V
g at which the device is illu-
to the absence of charged impurities in the graphene channel. minated (Fig. 1e). The detailed temporal dynamics of R during the entire
‘SET-RESET’ process are discussed in Supplementary Note 3.
Optoelectronic measurements After setting up the SET/RESET protocol, we now show that the
The optoelectronic measurements were carried out at room temperature, resistance of the sample at a specific gate voltage can be repeatedly alternated
and the sample was illuminated by using either an LED of wavelength, between two distinct states (Fig. 2a). The data for a single light pulse is shown
λ = 427 m or a Ti Sapphire pulsed laser (80 MHz repetition rate, 100 fs in Fig. 2b, which shows the saturation in the two states, hence the stability of
pulse width). To photo-dope the device, we use the following protocol: the SET and RESET states. Moreover, adjusting the light exposure time
The gate voltage is set to a desired value V
g , and the device is exposed to makes switching between multiple resistance values possible, as depicted in
the light of wavelength 427 nm of intensity I = 62 Wm−2, till the resistance Fig. 2c. In this measurement, the light was turned on for 10 s (blue shaded
saturates to its value at the Dirac point, RDP (Fig. 1c). The light is turned off region of the timeline), during which R increased. The light was then turned
at this point, and the gate response of the device is measured. It was found off. The value of R was stable at the value it reached when the illumination
that the entire R − Vg plot shifts with the Dirac point at V
g (Fig. 1b—solid was cut off. This process can be repeated to produce multiple stable resis-
red line; in this example V
g ¼ 3 V), establishing that the device is now tance levels in graphene.
electron-doped. We refer to this step as the SET protocol wherein the Figure 2d, e shows controlled switching between multiple resistance
Dirac point can be set deterministically at any desired value of Vg. To states in an arbitrary and on-demand manner, demonstrating stability and
RESET the device, it is exposed to a higher light intensity I = 458 Wm−2 at repeatability in the resistance states. The maximum fluctuations measured
V
g ¼ 0 V until the Dirac point shifts to Vg = 0 V. This protocol brings the in the resistance around the resistance states for repeated switching was
Dirac point of the device back to Vg = 0 V (Fig. 1b—dotted green line). As found to be ± 0.05 × R/RDP (Supplementary Note 6). This on-demand
illustrated in Fig. 1b, the process can be repeated without degradation in switching, stability and repeatability is a key indicator for the robustness of
the device characteristics. the device for memristor application.

(a) (b) 400

320

R (:)
240
I
160
hQ h-BN 80
0
SLG
h-BN 1

r
be
SiO
m
2
Si Nu
n
Vg V
Ru

3
-8 -6 -4 -2 0 2 4 6 8
V g (V )

(c) (d) (e)

t=t0 t=t1 t0 1.0
Pristine
1.0 1.0 Vg=-3V
t1 Vg=-4V
0.8 0.8
0.8 Vg=-5V
R/RDP
R/RDP

0.6 0.6
R/RDP

0.6
0.4 0.4
0.4
0.2 0.2
0.2
0.0 0.0
0 20 40 60 80 -8 -6 -4 -2 0 2 4 6 8 10 -8 -6 -4 -2 0 2 4 6 8
t(ms) Vg(V) Vg(V)

Fig. 1 | Device characteristics and optical doping. a Optical image of the device and to light. t = t0 marks when the light was turned on and t = t1 when the light was
a schematic of the device structure. b Longitudinal resistance as a function of gate turned off. d Snapshots of R − Vg plots as the Dirac point progressively shifts from
voltage for the pristine sample before illumination (blue solid line), after photo- Vg = 0 V at t = t0 (solid purple line) to Vg* = − 2 V at t = t1 (solid red line).
doping at V
g ¼ 3 V (red solid line), after erasing the doping (green dotted line), e Representative plots of deterministic shifting of the Dirac point to desired values of
and after photodoping again at V
g ¼ 3 V (pink dotted line). c Change in the Vg by photodoping. RDP is the four-probe resistance at the Dirac point.
longitudinal R as a function of time for constant power and gate voltage on exposure

npj 2D Materials and Applications | (2024)8:69 2

https://doi.org/10.1038/s41699-024-00503-7 Article

(a) (b) (c)

1.0
1.0 1.0 OFF ON OFF ON OFF ON OFF ON OFF ON OFF

0.9
0.8
0.8 0.8
R/RDP

R/RDP

R/RDP
0.6 0.7

0.6 0.6
0.4
0.5
Vg=-5V Vg=-5V
0.4 0.3
0.2 Vg=0V
0 2000 4000 6000 8000 0 2000 4000 6000 0 50 100 150 200 250 300 350
t(s)
t(s) t(s)
(d) Measured at Vg=0V Measured at Vg=0V
(e)
0V 0V 0V 0V RESET 0V 0V -2.5V RESET
-1V -2V -1V
1.0 1.0

0.8 0.8
R/RDP

R/RDP
0.6 0.6

0.4 0.4

-4V -3V -2V -1V SET 0.2 -3V -4V -4V -1V -2.5V -4.5V SET
0.2
0 500 1000 1500 2000 2500 0 1000 2000 3000
t(s) t(s)
Fig. 2 | Reproducible multiple states. a Optical switching of the device’s resistance V
g ¼ 5 V. d Successive optical switching of the device’s resistance between two
between two well-defined values at V
g ¼ 5V with LED light source of λ = 427 nm well-defined levels; SET level at V
g ¼ 4 V, −3 V, −2 V, −1 V and RESET level
and I = 62 Wm−2. b Resistance versus time graph for a single light pulse of SET and V
g ¼ 0 V. e Optical switching of the device’s resistance between different distinct
RESET. c Figure illustrating optical pinning of R to multiple values. The blue shaded resistance levels, defined by V
g . For both (d) and (e), the measurements were done
region marks when the light was switched on (hence R increased with t). The white with an LED light source of λ = 427 nm. The intensity of illumination was
regions show the time the light was off. LED light source of λ = 427 nm and I = 62 Wm−2 for SET and I = 458 Wm−2 for RESET. The R/RDP value is measured
I = 4 Wm−2 was used with V
g ¼ 5 V. For a–c the R/RDP values are recorded at at V
g ¼ 0 V.

Fig. 3 | Memristor. a Schematic of the proposed (a) V1 (b)

cross bar geometric memristor. b Schematic of the V3 V2 I1
individual device forming a cell of the memristor.
I2
graphene

Ga I3
hBN
te
vo gate electrod
e
lta
ge
s

Application as memristor multiplication operation can be carried out by modifying the weights of each
As shown in Fig. 2c, our device has at least six stable resistance states, with cross-bar intersection, i.e., by changing the resistance of the channels. For
more resistance states also accessible by lower optical powers. Such an hBN/ hardware implementation of ML training, the synaptic weights of a neural
graphene heterostructure with multiple stable resistance values holds sig- layer (each layer will be a separate cross-bar array) can be similarly modified.
nificant potential for developing memristor devices for vector-matrix mul- We propose a cross-bar geometry schematically shown in Fig. 3a to
tiplication and machine learning (ML) applications42–45. Several such devices achieve the above objectives. Its compact footprint offers distinct advantages
can be fabricated in a cross-bar array for a typical linear algebra calculation, over other structures. Each device unit (shown schematically in Fig. 3b) is
with voltages as inputs and currents as outputs. A new vector-matrix individually gated; this architecture is conveniently achievable using

npj 2D Materials and Applications | (2024)8:69 3

https://doi.org/10.1038/s41699-024-00503-7 Article

(a) h-BN graphene (b) h-BN graphene (c) h-BN graphene

1.1 eV
EC EC EC
2.8 eV 4.5 eV EN EN
EN
6 eV ED ED ED
EV EV
EV
Eext > Ei
SLG hQ
Eext Eext
hBN
SiO2
Si Ei

(d) h-BN graphene (e) h-BN graphene h-BN graphene

(f)
EC EC EC

EN EN EN
ED ED 2 ED
1
EV EV EV
Eext = Ei Ei Ei
Eext Ei Ei

Ei
Fig. 4 | Schematic energy band alignment (top panels) and charge distribution its net positive charge. d The doping process stops when Ei = Eext. The hBN bands
(bottom panels) in the device under different conditions. EC , EN and EV flatten out, graphene is charge neutral, and the hBN is left positively charged. e After
represent the conduction band edge, N-vacancy mid-gap state and valence band the gate voltage is set to zero, graphene draws a net negative charge from the contacts,
edge of hBN, respectively. a For V
g ¼ 0; hν ¼ 0, hBN and graphene are charge making it electron-doped and the total system charge neutral. (f) RESET: For
neutral. b The graphene channel acquires a positive charge for V
g <0 and hν = 0. The V
g ¼ 0 V and hν ≠ 0, electrons are excited from E v (blue solid line) of the hBN to
resultant electric field Eext bends the bands of the hBN. c SET: For V
g <0 and hν ≠ 0, recombine with the holes in E N (orange dotted line). The electrons from graphene
electrons are transferred from E N (orange dotted line) to E C (solid red line) of hBN. (represented by the red dotted circles) drift into hBN E v and recombine with the
These electrons drift into graphene (represented by the dotted electrons) and reduce holes left behind.

modern nano-fabrication processes. Before each operation, the channel A negative gate-voltage V g ¼ V
g dopes the graphene channel with holes
resistance of each device is SET by a global incident optical beam and the and creates an electric field Eext directed from graphene into the hBN
application of distinct back-gate voltages to different devices. The RESET for (Fig. 4b) leading to band bending. Illuminating the device with violet light
any desired device can done by setting its gate voltages to zero and illumi- excites electrons from E N to the conduction band E C of hBN. These elec-
nating with a global incident optical beam. trons are funneled to the graphene channel under the influence of the
electric field. The holes left at E N in the hBN generate an electric field Ei in
Origin of the phenomenon the direction opposite Eext (Fig. 4c). The electron transfer process continues
Photodoping of the graphene channel requires charge transfer from hBN to until the net electric field between graphene and hBN Enet = Eext − Ei
graphene. The energy corresponding to violet light (λ = 427 nm) is 2.9 eV, becomes zero. Simple electrostatics arguments show that the effective
much lower than the band gap of hBN (≈6 eV), precluding photo-excitation number density in graphene becomes zero at this point, which manifests as a
of carriers from the valence band of hBN. We also find that the graphene shift of the charge neutrality point to V
g (Fig. 4d).
channel remained undoped in a device with only the top hBN flake and On reducing Vg to zero, graphene draws negative charges from the
without the bottom hBN flake upon using the same protocol described metal contacts, making the net device charge neutral (Fig. 4e). This charge
above (Supplementary Note 2). This study confirms that only the bottom configuration generates an electric field Ei directed from the hBN to gra-
hBN was responsible for the photodoping effect. Although the top hBN is phene. The consequent band bending and the fact that E N >E D forbids
not essential for the operation of the device, it is crucial for having a high- electron transfer back from graphene to the hBN; graphene remains
quality, low-defect graphene channel with long-term stability. negatively charged with the Dirac point at V g ¼ V
g , and hBN is positively
Based on our observations, we sketch out a possible scenario below that charged (Fig. 4e). This constitutes the SET protocol. The energy barrier to
explains all our experimental findings. Figure 4a is a schematic of the energy back-transfer electrons from graphene to hBN explains the long-term
alignment of the bottom hBN and the graphene without photo-excitation charge retention in graphene after the photodoping is completed.
and at Vg = 0. Several mid-gap states in hBN can act as electron donors. Of Exposing the device to light with V
g ¼ 0 V erases the doping, bringing
these, the one most relevant for us is the defect state of nitrogen vacancies the graphene’s Dirac point back to Vg = 0 V. For this to happen, excess
(marked as E N in Fig. 4). This level can have stable charge states of + 1 or 038. electrons need to be removed from graphene and transferred back to hBN.

npj 2D Materials and Applications | (2024)8:69 4

https://doi.org/10.1038/s41699-024-00503-7 Article

Fig. 5 | Wavelength, power and gate dependence of (a) (b)

doping dynamics. a Plot of time constant, τs versus 4
the laser photon energy for electron doping. The
data were taken at a fixed gate voltage V
g ¼ 3 V 3
10
and laser intensity 103 Wm−2. The dashed gray line

Wr s)
marks E N . The dotted red line is a guide to the eye.

Ws s)
b Plot of time constant, τr versus the laser photon 2
energy for hole doping. The data were taken at a
fixed gate voltage V
g ¼ 3 V and laser intensity 1
1
105 Wm−2. The dotted red line is a guide to the eye.
c Log-log plot of time constant τs versus the intensity
0
of incident light. The data were taken at V
g ¼ 5 V 2.4 2.6 2.8 3.0 3.2 3.4 2.8 3.0 3.2
and λ = 427 nm using LED. The open symbols are
E (eV) E (eV)
the measured data points, and the dashed gray line is (c) (d) 20
a linear fit to the data. d Plot of τs versus the V
g for
electron doping. The data was taken for λ = 427 nm
and I = 62 Wm-2 using LED.
15
10
Ws s)

Ws s)
10

1 5

0.1 1 0 2 4 6 8 10 12 14 16
2
I (nW/Pm ) Vg(V)

Note, however, that this process is energetically unfavorable as E N >E D at the measurements confirm this (Fig. 5d) (See Supplementary Note 1 for detailed
interface. We propose a phenomenological scenario in which this RESET derivation).
process is a two-step process involving (1) the transfer of electrons from the To summarize, we demonstrate a reversible control of the Dirac
valence band of hBN to its mid-gap states due to optical excitation and (2) point in the graphene-hBN heterostructure to encode the resistance
subsequent electron transfer from graphene to the empty valence band values via a combined optical-electrical stimulus. The device’s resistance
states of hBN due to electric field (Fig. 4f). Consequently, the RESET process can be modified by varying the gate voltage in the presence of an optical
(involving electron transfer from graphene to hBN) is much slower than the incident power or by fixing the gate voltage and illuminating it with
SET process for electron doping. Device-level simulations are required to multiple optical pulses. The switching time can be tuned by the incidence
verify if the above scenario correctly captures the doping erasure process. light wavelength (Fig. 5a), light intensity (Fig. 5c), or the gate voltage
To experimentally verify the scenario proposed above, we used a tunable (Fig. 5d) providing tremendous tunability of the properties of the device.
pulse laser to study the wavelength dependence of the ‘SET’ time τs. Figure 5a The time taken to electron dope is much less compared to hole doping
plots τs versus the laser photon energy for a constant laser intensity and for a fixed power and gate voltage, and further experiments need to be
V
g ¼ 3 V. It shows a minimum centered around E = 2.8 eV. This photon done to make them comparable. It should be possible to control the
energy corresponds to the optical absorption by valence nitrogen defect in hBN switching time by modifying the defect density in hBN using electron
‘E N ,’ leading credence to our understanding of the photodoping process37. The irradiation or annealing processes. The ability to photoelectrically SET,
wavelength dependence of the time required for the ‘RESET’ τr is shown in Fig. READ, and RESET multiple stable and non-volatile resistance states of
5b. In contrast to the SET process, τr reaches its minimum value at around the device makes it ideal for use as a memristor.
E = 3.2 eV; this value matches exactly E N  E V ; the energy required to excite
charge carriers from the valence band to the nitrogen-vacancy state. Methods
We note that the optical power used for the RESET process is higher Device fabrication
than that used during the SET process. Thus, one must consider the pos- The devices were fabricated using the dry transfer technique49. Single-layer
sibility of the formation of defect states in the hBN during this process. We graphene (SLG) and hexagonal boron nitride (hBN) flakes were mechani-
rule out this scenario by noting the lack of long-term resistance drift, as seen cally exfoliated onto a Si/SiO2 substrates. The hBN flakes had a thickness of
from Fig. 2. We have also considered the possibility of the reversible creation 25–30 nm. Electron beam lithography was used to define electrical contacts.
of additional mid-gap states but ruled them out from energetic considera- This was followed by etching with a mixture of CHF3 (40 sscm) and O2
tion. The maximum used power for the optical measurements was (10 sscm). The metallization was done with Cr/Au (5 nm/60 nm) to form
105 Wm−2, which is significantly lower than the reported power needed to the 1D electrical contacts with SLG.
create defects in hBN46–48.
Figure 5c shows the dependence of τs on the intensity of the light, I at a Measurements
fixed V
g and λ = 427 nm. We find that τs ∝ I−1. This dependence is All electrical transport measurements were performed at room temperature
understandable, as an increase in the intensity of photons leads to more free using a low-frequency AC measurement technique. For low-temperature
charges being produced in hBN, reducing the time taken to dope measurements, the sample was cooled down in a cryostat to 4.7 K. For
(see Supplementary Note 1 for detailed derivation). For measurements optoelectronic measurements, the sample was illuminated using either an
performed at a fixed intensity of light, the time constant to dope should LED of wavelength, λ = 427 nm, or a Ti Sapphire pulsed laser (80 MHz
increase with an increase in the magnitude of V
g (equivalently, of Eext); repetition rate, 100 fs pulse width).

npj 2D Materials and Applications | (2024)8:69 5

https://doi.org/10.1038/s41699-024-00503-7 Article

Data availability 22. Aftab, S., Iqbal, M. Z. & Iqbal, M. W. Programmable photo-induced
The authors declare that the data supporting the findings of this study are doping in 2D materials. Adv. Mater. Interfaces 9, 2201219 (2022).
available within the main text and its supplementary Information. Other 23. Neumann, C. et al. Spatial control of laser-induced doping profiles in
relevant data are available from the corresponding author upon request. graphene on hexagonal boron nitride. ACS Appl. Mater. Interfaces 8,
9377–9383 (2016).
Received: 17 July 2024; Accepted: 19 September 2024; 24. Ju, L. et al. Photoinduced doping in heterostructures of graphene and
boron nitride. Nat. Nanotechnol. 9, 348–352 (2014).
25. Roy, K. et al. Graphene-MoS2 hybrid structures for multifunctional
References photoresponsive memory devices. Nat. Nanotechnol. 8, 826–830 (2013).
1. Sangwan, V. K. & Hersam, M. C. Neuromorphic nanoelectronic 26. Kim, Y. D. et al. Focused-laser-enabled p-n junctions in graphene
materials. Nat. Nanotechnol. 15, 517–528 (2020). field-effect transistors. ACS Nano 7, 5850–5857 (2013).
2. Walters, B., Jacob, M. V., Amirsoleimani, A. & Rahimi Azghadi, M. A 27. Song, S.-B. et al. Deep-ultraviolet electroluminescence and
review of graphene-based memristive neuromorphic devices and photocurrent generation in graphene/hBN/graphene
circuits. Adv. Intell. Syst. 5, 2300136 (2023). heterostructures. Nat. Commun. 12, 7134 (2021).
3. Lu, H. et al. Ferroelectric tunnel junctions with graphene electrodes. 28. Velasco, J. J. et al. Nanoscale control of rewriteable doping patterns in
Nat. Commun. 5, 5518 (2014). pristine graphene/boron nitride heterostructures. Nano Lett. 16,
4. Wu, J. et al. High tunnelling electroresistance in a ferroelectric van der 1620–1625 (2016).
Waals heterojunction via giant barrier height modulation. Nat. 29. Kim, S. H. et al. Multilevel MoS2 optical memory with photoresponsive
Electron. 3, 466–472 (2020). top floating gates. ACS Appl. Mater. Interfaces 11, 25306–25312 (2019).
5. Yan, X. et al. Moiré synaptic transistor with room-temperature 30. Liu, C.-H., Chang, Y.-C., Norris, T. B. & Zhong, Z. Graphene
neuromorphic functionality. Nature 624, 551–556 (2023). photodetectors with ultra-broadband and high responsivity at room
6. Liu, H., Liu, Y. & Zhu, D. Chemical doping of graphene. J. Mater. Chem. temperature. Nat. Nanotechnol. 9, 273–278 (2014).
21, 3335–3345 (2011). 31. Xiang, D. et al. Two-dimensional multibit optoelectronic memory with
7. Wehling, T. O. et al. Molecular doping of graphene. Nano Lett. 8, broadband spectrum distinction. Nat. Commun. 9, 2966 (2018).
173–177 (2008). 32. Tran, M. D. et al. Two-terminal multibit optical memory via van der
8. Jung, N. et al. Charge transfer chemical doping of few layer Waals heterostructure. Adv. Mater. 31, 1807075 (2019).
graphenes: charge distribution and band gap formation. Nano Lett. 9, 33. Lee, I. et al. Photoinduced tuning of Schottky barrier height in
4133–4137 (2009). graphene/MoS2 heterojunction for ultrahigh performance short
9. Bruna, M. & Borini, S. Observation of Raman $ G $-band splitting in channel phototransistor. ACS Nano 14, 7574–7580 (2020).
top-doped few-layer graphene. Phys. Rev. B 81, 125421 (2010). 34. Gorecki, J., Apostolopoulos, V., Ou, J.-Y., Mailis, S. & Papasimakis, N.
10. Zhan, D. et al. FeCl3-based few-layer graphene intercalation Optical gating of graphene on photoconductive Fe:LiNbO3. ACS
compounds: single linear dispersion electronic band structure and Nano 12, 5940–5945 (2018).
strong charge transfer doping. Adv. Funct. Mater. 20, 3504–3509 35. Seo, B. H., Youn, J. & Shim, M. Direct laser writing of air-stable p-n
(2010). junctions in graphene. ACS Nano 8, 8831–8836 (2014).
11. Zhao, W., Tan, P. H., Liu, J. & Ferrari, A. C. Intercalation of few-layer 36. Miller, D., Blaikie, A. & Alemán, B. J. Nonvolatile rewritable frequency
graphite flakes with FeCl3: Raman determination of fermi level, layer tuning of a nanoelectromechanical resonator using photoinduced
by layer decoupling, and stability. J. Am. Chem. Soc. 133, 5941–5946 doping. Nano Lett. 20, 2378–2386 (2020).
(2011). 37. Attaccalite, C., Bockstedte, M., Marini, A., Rubio, A. & Wirtz, L.
12. Zhao, W., Tan, P., Zhang, J. & Liu, J. Charge transfer and optical Coupling of excitons and defect states in boron-nitride
phonon mixing in few-layer graphene chemically doped with sulfuric nanostructures. Phys. Rev. B 83, 144115 (2011).
acid. Phys. Rev. B 82, 245423 (2010). 38. Weston, L., Wickramaratne, D., Mackoit, M., Alkauskas, A. & Van De
13. Singh, A. K. et al. Molecular n-doping of chemical vapor deposition Walle, C. G. Native point defects and impurities in hexagonal boron
grown graphene. J. Mater. Chem. 22, 15168 (2012). nitride. Phys. Rev. B 97, 214104 (2018).
14. Medina, H., Lin, Y.-C., Obergfell, D. & Chiu, P.-W. Tuning of charge 39. Sajid, A., Reimers, J. R. & Ford, M. J. Defect states in hexagonal boron
densities in graphene by molecule doping. Adv. Funct. Mater. 21, nitride: assignments of observed properties and prediction of properties
2687–2692 (2011). relevant to quantum computation. Phys. Rev. B 97, 064101 (2018).
15. Ryu, S. et al. Atmospheric oxygen binding and hole doping in 40. Strand, J., Larcher, L. & Shluger, A. L. Properties of intrinsic point
deformed graphene on a SiO2 substrate. Nano Lett. 10, 4944–4951 defects and dimers in hexagonal boron nitride. J. Phys. Condens.
(2010). Matter 32, 055706 (2020).
16. Dean, C. R. et al. Boron nitride substrates for high-quality graphene 41. Amin, K. R., Nagarajan, R., Pandit, R. & Bid, A. Multifractal
electronics. Nat. Nanotechnol. 5, 722–726 (2010). conductance fluctuations in high-mobility graphene in the integer
17. Das, A. et al. Monitoring dopants by Raman scattering in an quantum hall regime. Phys. Rev. Lett. 129, 186802 (2022).
electrochemically top-gated graphene transistor. Nat. Nanotechnol. 42. Ducry, F. et al. An ab initio study on resistance switching in hexagonal
3, 210–215 (2008). boron nitride. npj 2D Mater. Appl. 6, 1–10 (2022).
18. Yan, J., Zhang, Y., Kim, P. & Pinczuk, A. Electric field effect tuning of 43. Xie, J., Afshari, S. & Sanchez Esqueda, I. Hexagonal boron nitride (h-
electron-phonon coupling in graphene. Phys. Rev. Lett. 98, 166802 BN) memristor arrays for analog-based machine learning hardware.
(2007). npj 2D Mater. Appl. 6, 1–7 (2022).
19. Chen, F., Qing, Q., Xia, J., Li, J. & Tao, N. Electrochemical gate- 44. Maier, P. et al. Electro-photo-sensitive memristor for neuromorphic
controlled charge transport in graphene in ionic liquid and aqueous and arithmetic computing. Phys. Rev. Appl. 5, 054011 (2016).
solution. J. Am. Chem. Soc. 131, 9908–9909 (2009). 45. Schranghamer, T. F., Oberoi, A. & Das, S. Graphene memristive
20. Uesugi, E., Goto, H., Eguchi, R., Fujiwara, A. & Kubozono, Y. Electric synapses for high precision neuromorphic computing. Nat. Commun.
double-layer capacitance between an ionic liquid and few-layer 11, 5474 (2020).
graphene. Sci. Rep. 3, 1595 (2013). 46. Gan, L. et al. Large-scale, high-yield laser fabrication of bright and
21. Tiberj, A. et al. Reversible optical doping of graphene. Sci. Rep. 3, pure single-photon emitters at room temperature in hexagonal boron
2355 (2013). nitride. ACS Nano 16, 14254–14261 (2022).

npj 2D Materials and Applications | (2024)8:69 6

https://doi.org/10.1038/s41699-024-00503-7 Article

47. Kianinia, M., White, S., Fröch, J. E., Bradac, C. & Aharonovich, I. Additional information
Generation of spin defects in hexagonal boron nitride. ACS Photonics Supplementary information The online version contains
7, 2147–2152 (2020). supplementary material available at
48. Yang, Y.-Z. et al. Laser direct writing of visible spin defects in https://doi.org/10.1038/s41699-024-00503-7.
hexagonal boron nitride for applications in spin-based technologies.
ACS Appl. Nano Mater. 6, 6407–6414 (2023). Correspondence and requests for materials should be addressed to
49. Tiwari, P. et al. Observation of the time-reversal symmetric hall effect Akshay Singh or Aveek Bid.
in graphene-WSe2 heterostructures at room temperature. Nano Lett.
23, 6792–6798 (2023). Reprints and permissions information is available at
http://www.nature.com/reprints
Acknowledgements
A.B. acknowledges funding from the Department of Science & Publisher’s note Springer Nature remains neutral with regard to
Technology FIST program and the U.S. Army DEVCOM Indo-Pacific jurisdictional claims in published maps and institutional affiliations.
(Project number: FA5209 22P0166). K.W. and T.T. acknowledge support
from JSPS KAKENHI (Grant Numbers 19H05790, 20H00354, and Open Access This article is licensed under a Creative Commons
21H05233). A.S. acknowledges funding from the Indian Institute of Attribution 4.0 International License, which permits use, sharing,
Science start-up grant, DST Nanomission CONCEPT (Consortium for adaptation, distribution and reproduction in any medium or format, as long
Collective and Engineered Phenomena in Topology) grant and project as you give appropriate credit to the original author(s) and the source,
MoE-STARS-2/2023-0265. M.M. acknowledges the Prime Minister’s provide a link to the Creative Commons licence, and indicate if changes
Research Fellowship (PMRF). were made. The images or other third party material in this article are
included in the article’s Creative Commons licence, unless indicated
Author contributions otherwise in a credit line to the material. If material is not included in the
H.K.M, M.M., V.S., A.S. and A.B. conceptualized the study, performed the article’s Creative Commons licence and your intended use is not permitted
measurements, and analyzed the data. K.W. and T.T. grew the hBN single by statutory regulation or exceeds the permitted use, you will need to
crystals. All the authors contributed to preparing the manuscript. obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/.
Competing interests
The authors declare no competing interests. © The Author(s) 2024

npj 2D Materials and Applications | (2024)8:69 7