Ohmic-contact engineering in few-layer black Phosphorus field effect transistors

Francesca Telesio,1 Gwenael le Gal,1 Manuel Serrano-Ruiz,2 Federico Prescimone,3 Stefano Toffanin,3 Maurizio Peruzzini,2 and Stefan Heun1

1NEST, Istituto Nanoscienze - CNR and Scuola Normale Superiore, Pisa, Italy

2Istituto di Chimica dei Composti Organometallici – CNR, Sesto Fiorentino, Italy

3Istituto per lo Studio dei Materiali Nanostrutturati – CNR, Bologna, Italy

The renaissance of black phosphorus

(1)Long et al., Nano Letters, 2014, 14, 5733; (2) adapted from X. Ling et al., PNAS 112 (2015) 4523; (3) Das et al, Nano Letter, 2014 14, 2733 (5) A. Castellanos-Gomez et al, 2D Mat. 1, (2014); (4) F. Xia et al, Nat Comm 5, (2014)

p-type semiconductor with up to 45000 cm2/(Vs) mobility at low temperature (1)

band gap of 0.3 eV in the bulk and tunable with layer number (≈2 eV for the monolayer) (2,3)

strong in planeanisotropy of optical, eletrical and thermaltransport properties

high reactivity to oxygenand water vapour, enhanced by light exposure

(2)

(3)

(4)(5)

Ohmic contact engineering

Tunneling barrier<<<<

Schottky barrier

Metallization below the contact

metal electrode

The simplest model: Schottky-Mott rule

𝜙𝑝 = 𝐸𝑔 − 𝜙𝑚 − 𝜒

𝜒𝑏𝑃 ≈ 4.4 𝑒𝑉

𝜙𝑛 = 𝜙𝑚 − 𝜒

<<<<

<<<<

Ohmic contact Titanium

Chromium

Nickel

Often usedas stitchinglayer for goldelectrodes

(6)

(7) (8)

𝜙𝑁𝑖 ≈ 5.0 𝑒𝑉

𝜙𝐶𝑟 ≈ 4.5 𝑒𝑉

𝜙𝑇𝑖 ≈ 4.3 𝑒𝑉

best for hole injection

less suitable for hole injection

<<<<

(9)

(10)

(11)

(6) Wikipedia Cc licence; (7) adapted from Schroder, Semiconductor materials and device characterization; (8) Feng et al, Nanoscale, 2016, 8, 2686; (9)Allain et al, Nat Mater, 2015, 14, 1195; (10) Lide, Handbook of Chemistry and Physics, 2008; (11) Du et al, ACS Nano, 2014, 8, 10035

Transfer Length Method

𝑅2 = 𝑅𝐶 + 𝑅𝑏𝑃 𝑐ℎ𝑎𝑛𝑛𝑒𝑙 + 𝑅𝐶

𝑅𝑏𝑃𝑐ℎ𝑎𝑛𝑛𝑒𝑙 = (𝑑/𝑊)𝑅𝑆 ∝ 𝑑

d= n*d1

W

d1 d1

L

I

V=

Transfer Length Method

d1 2d1 3d1 4d1 5d1 6d1 7d1

R2

d

RS/W

2Rc

𝑅2 = 𝑅𝐶 + 𝑅𝑏𝑃 𝑐ℎ𝑎𝑛𝑛𝑒𝑙 + 𝑅𝐶

𝑅𝑏𝑃𝑐ℎ𝑎𝑛𝑛𝑒𝑙 = (𝑑/𝑊)𝑅𝑆 ∝ 𝑑

d= n*d1

W

d1 d1

L

I

V=

Transfer Length Method

Aggregate data over different devices/samples

Normalization of the resistance with contact width to compare different devices

Average

Propagation of the experimental error

Standard error of the distribution to get information on the scattering among the devices

𝑅𝐶𝑊d= n*d1

W

d1 d1

L

I

V=

Sample preparation

0.0 0.5 1.0 1.50

5

10

15

20

25

30

35

z (

nm

)

x (m)

Exfoliation in glove box or in glove bag

Coating with a MMA:MAA/PMMA bilayer

Flakes identification and EBL

Development&O2 plasma

Metal evaporation, lift-off, coating

10 µm

-1 0 1

-150

-75

0

75

150

I(nA)

V(mV)

d=5m R=23.2k

d=4m R=18.6k

d=3m R=14.1k

d=2m R=10.3k

d=1m R=6.2k

Room T

Cr

-1 0 1

-150

-75

0

75

150

I(nA)

V(mV)

d=5m

d=4m

d=3m

d=2m

d=1m

Room T

Cr

I-V Curves

All curves show linear behavior Good quality Ohmic contacts No Schottky behaviour

Linear fit to extract R2

d= n*d1

W

d1 d1

L

I

V=

Contact resistance at room temperature

0 1 2 3 4 50

5

10

15

20

25

R2

(k

)

d (m)

Room T

Cr

Ti

Ni

0.0 0.1 0.20.00

0.25

0.50

0.75

1.00

1.25

1.50

R2

(k

)

d (m)

Room T

Cr

Ti

Ni

𝑅𝐶|𝑁𝑖 < 𝑅𝐶|𝑇𝑖 < 𝑅𝐶|𝐶𝑟

Contact resistance at room temperature

Room temperature

metal 𝑅𝐶𝑊Ω𝜇𝑚

avg errorΩ𝜇𝑚

std errorΩ𝜇𝑚

Cr 797 64 253

Ti 217 88 163

Ni 135 13 64

0 1 2 3 4 50

5

10

15

20

25

R2

(k

)

d (m)

Room T

Cr

Ti

Ni

Contact resistance at low temperature

0 1 2 3 4 50

10

20

30

40

50

60

R2

(k

)

d (m)

T= 4.2 K

Cr

Ti

Ni

0.0 0.1 0.20.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

R2

(k

)

d (m)

T= 4.2 K

Cr

Ti

Ni

Low temperature

metal 𝑅𝐶𝑊Ω𝜇𝑚

avg errorΩ𝜇𝑚

std errorΩ𝜇𝑚

Cr 2428 198 1377

Ti 740 209 282

Ni 432 30 208

10 100

0

2000

4000

Cr

Ti

Ni

RCW

(

m)

Temperature (K)

• Ni has the lowest contact resistance and the lowest scattering between devices, both at RT and at LT.

• RcW = 135 m for Ni at RT is among the lowest values reported.

• Cr shows the worst performance, likely due to a defective growth of Cr on bP.

Contact resistance: summary

(12) Adapted from Hong et al, J. Phys Chem C, 2015, 119, 8199; (13) Pan et al, Chem Mater, 2016, 28, 2100

High bias and gate voltage dependence

-0.10 -0.05 0.00 0.05 0.10-30

-20

-10

0

10

20

30

Cr

Ti

Ni

I sd (A

)

Vsd

(V)

Vg=0

-80 -40 0 40 800.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Cr

Ti

Ni

Ad

juste

d-r

2

Vg (V)

In the accumulation regime, all contacts display an Ohmic behavior, even up to ±100 mV bias.

Reaching the depletion regime, deviations from the Ohmic behaviour are observed, independently of the metal, consistent with previous reports.(14)

(14) Li et al, Nat. Nanotech, 9, 372

-80 -40 0 40 80

0

10

20

30

40

Ni

Cr

Ti

I sd (

A)

Vg(V)

-80 -40 0 40 80

1E-4

1E-3

0.01

0.1

1

10

Ni

Cr

Ti

I sd (

A)

Vg(V)

Approaching the quantum limit

Conclusions

We studied contact resistance on few-layer bP (on ten devices).

We investigated three technologically relevant metals: Chromium, Titanium and Nickel.

All samples show linear I-V curves (±1 mV)

Both at room temperature and at low temperature the best results are obtained for Nickel, both for contact resistance and for scattering among the results.

Consistent results are obtained from the comparison of two-probe and four-probe resistances.

All samples displayed a strong unipolar p-type behavior.

Linear I-V curves (Ohmic contacts) were observed also for high bias (±100 mV) in the accumulation regime.

Acknowledgements

Thank you for your attention!

“Phosphorene functionalization: a new platform for advanced

multifunctional materials” (Grant Agreement No. 670173)

M. Serrano-Ruiz M. Peruzzini

S. ToffaninF. Prescimone

G. le GalF. Telesio