The effect of nitriding on the humidity sensing properties of hydrogenated amorphous carbon films
Giuliano Frattini, Sindry Torres, Leonel I Silva, Carlos E Repetto, Bernardo J Gómez, Ariel Dobry
TThe effect of nitriding on the humidity sensingproperties of hydrogenated amorphous carbon films
Giuliano Frattini † , Sindry Torres † , Leonel I Silva ‡ , Carlos ERepetto † , Bernardo J G´omez † and Ariel Dobry † † Facultad de Ciencias Exactas, Ingenier´ıa y Agrimensura, Universidad Nacional deRosario – Instituto de F´ısica Rosario, Bv. 27 de Febrero 210 bis, S2000EZP Rosario,Argentina. ‡ INTEMA, Facultad de Ingenier´ıa, Universidad Nacional de Mar del Plata, ConsejoNacional de Investigaciones Cient´ıficas y T´ecnicas (CONICET), Av. Col´on 10850,Mar del Plata, 7600, Argentina.E-mail: [email protected]
24 February 2021
Abstract.
We have studied the effect of nitriding on the humidity sensing propertiesof hydrogenated amorphous carbon (a-C:H) films. The films were prepared in twostages combining the techniques of physical deposition in vapor phase evaporation(PAPVD) and plasma pulsed nitriding. By deconvolution of the Raman spectrumwe identified two peaks corresponding to the D and G modes characteristic of a-C:H. After the N -H plasma treating, the peaks narrowed and shifted to the right,which we associated with the incorporation of N into the structure. We comparedthe sensitivity to the relative humidity (RH) of the films before and after the N -H plasma treatment. The nitriding improved the humidity sensitivity measured as thelow frequency resistance.By impedance spectroscopy we studied the frequency dependence of the ACconductivity σ at different RH conditions. Before nitriding σ ( ω ) ∼ Aω s , it seemedto have the universal behaviour seen in other amorphous systems. The humiditychanged the overall scale A . After nitriding, the exponent s increased, and became RHdependent. We associated this behaviour to the change of the interaction mechanismbetween the water molecule and the substrate when the samples were nitriding. Keywords : Amorphous Carbon Films, Humidity Sensors, Impedance Spectroscopy,Electronic Transport in Amorphous Semiconductors. a r X i v : . [ phy s i c s . a pp - ph ] F e b he effect of nitriding on the humidity sensing properties of a-C:H films
1. Introduction
Humidity sensors are used in many industrial applications, from environmentalmonitoring and meteorology to the determination of soil water content in agriculture,air conditioning systems, monitoring the quality of food, medical equipment and inmany other fields [1]. A novel application of ultrafast sensors was proposed for a touch-less user interface proving to run in a whistling recognition analysis [2]. A materialto be used as a humidity sensor should change some of its properties (for example,its electrical impedance) when exposed to a humid atmosphere. The sensitivity of asensor is the relative change of the detection property, when the RH changes in therange of interest. The requirements for such a material are: high sensitivity in order todiscriminate small difference of RH, fast response and short recovery time. For a newgeneration of humidity sensors it would be important the compatibility with micro ornano circuits and the flexibility.To this aim carbon films have attracted great interest as sensitive material becausethey have a large sensing area and high chemical inertness (see [3] for a recent review).In particular, a recent study pointed out that hydrogenated amorphous carbon (a-C:H)films treated by N –H plasma were sensitive to the humidity [4, 5]. This films wereproposed as a basis of resistivity humidity sensors. This is a promising result due to thelow cost fabrication of this material.The goal of this paper is to analyze how plasma treatment modifies the transportproperties of the films and in particular the humidity sensitive properties. For thispurpose we have measured the sensitivity of the samples before nitriding, which couldnot be done in Ref. [4].As a first step, we studied the structural changes of the films by Raman scattering.Different fitting procedures were used to determine the contribution of the G and Dvibrational modes in the measured spectra. After the plasma treatment both modeswere shifted and narrowed. This was attributed to the incorporation of the N into thestructure by replacing some of the C atoms. We also studied the presence of C-H bondsby IR spectroscopy. We found a very low concentration of this hydrogenated bonds.Next, we analyzed the change of the transport properties of the films with humidityby impedance spectroscopy. The low frequency resistance changed with the humidityof the surrounding medium. The plasma nitriding treatment improved the humiditysensing properties of the films.In order to understand how the humidity changed the electrical transport ofthe films, we studied the Nyquist plots of Impedance Spectrum and the frequencydependence of AC conductivity at different humidity. The Nyquist plots had asemicircular-type behaviour. A straight line-type impedance characteristic of thediffusion of ions across the interface was not observed. Therefore, we associated theRH dependence to the bulk properties of the films.As regards the dependence of the frequency of the real part of the conductivity, itsbehaviour changed after the nitriding process. Before it, they had a kind of universal he effect of nitriding on the humidity sensing properties of a-C:H films σ ( ω ) decreased with the humidity and all these curves converged at high frequency.We understand that the improvement of the sensing properties with the humidity wasrelated with a change in the interaction mechanism between the water molecules andthe film.
2. Synthesis and characterization of a-C:H films
The carbon films were deposited by the electron-beam physical vapour depositiontechnique, in which small graphite bars were bombarded with an electron beam thatwas generated in a tungsten filament powered by an external source at 150 mA. Thisprocedure was carried out in a high vacuum chamber at 10 − Torr. The evaporated gaswas deposited over printed circuit boards made of synthetic resin FR2, predesigned tomeasure electric transport.Later, the sample was placed in a plasma nitriding reactor originally designed totreat the surface of steels. The purpose of this procedure was to dilute the sample toincrease the resistance and to change the microstructure of the film by interaction withthe plasma. The reactor consisted of an 8-litre AISI 304 stainless steel vacuum chamberconnected to ground potential. Inside the chamber, there was an electrode, also madeof AISI 304 stainless steel, connected to a power supply that was able to generate asquare-wave signal of up to 700 V at a frequency between 0 and 1 kHz. The duty cyclecan also be changed. A pre-vacuum of 0.001 Torr was created in the chamber, whichwas then back filled to 1 Torr with a mixture of 50% nitrogen and 50% hydrogen. Weconducted our experiments at a constant temperature of 250 ◦ C, with an on/off ratio of50%/50% at a frequency of 100 Hz. The applied voltage was 550 V. Pressure, current,voltage and flux were kept constant during the ion nitriding treatments. The treatmentslasted 5 min; measured from the moment the sample temperature reached 50 ◦ C.After the nitriding process we named the samples as a-CN:H to distinguish themfrom the ones we had before the plasma treatment which we had called a-C:H. Byspectroscopic ellipsometry measurements we determined that the thickness of the samplehad been strongly reduce after the plasma treatment. It was 70 nm for a-C:H and became20 nm for a-CN:H.The Raman spectrum shape fitting is a widely used method to study the detailbonding structure of carbon films. Therefore, we obtained the Raman spectra in orderto characterize the structure of the deposited films.The spectra were acquired in a Renishaw In Via reflex system equipped with charge-coupled device (CCD) detector of 1040 ×
256 pixels. A 514 nm diode laser (50 mW) wasused as excitation source in combination with a grating of 2400 grooves/mm and slitopenings of 65 µ m, which yield a spectral resolution of about 4 cm − . The laser powerwas kept below 10% to avoid sample damage. A 50 × (0.5 NA) with shorter workingdistance (210 µ m) Leica metallurgical objective was used in the excitation and collection he effect of nitriding on the humidity sensing properties of a-C:H films − wavenumbers, respectively.The sample had an area of 1 cm and could be inhomogeneous. Therefore, Ramanspectra were taken at different points. The spectra were fitted by taking into accountdifferent possibilities [7, 8]. Among the different alternatives, we chose to use Gaussianfunctions for both peaks or a Lorentzian and a Gaussian functions for each peak. Wealso used a BWF function for D peak and a Lorentzian function for G peak.As an example, we can see in Fig. 1 a deconvolution with Gaussians functions forD and G peaks for a-C:H and a-CN:H samples. , , , ,
81 Raman frequency (cm − ) I n t e n s it y ( a . u . ) D (1409 cm − )G (1537 cm − ) 1200 1300 1400 1500 1600 1700 180000 , , , ,
81 Raman frequency (cm − ) I n t e n s it y ( a . u . ) D (1422 cm − )G (1560 cm − ) Figure 1.
In blue points, the Raman spectra for a-C:H (left) and a-CN:H (right)samples. Both graphics show a deconvolution on two Gaussian peaks correspondingto D and G modes. In solid red line, the sum of the two gaussians.
The values of the parameters obtained from the different fittings are shown in Table1. They are average values from the spectra obtained at different points of the sample.Parameter a-C:H a-CN:H x D (1409 ±
12) cm − (1425 ±
5) cm − s D (153 ±
6) cm − (143 ±
4) cm − x G (1537 ±
1) cm − (1554 ±
1) cm − s G (73 ±
3) cm − (63 ±
2) cm − I D /I G (0.5 ± ± A D /A G (1.1 ± ± Table 1.
Results for the Raman D and G peaks adjustment parameters for a-C:Hand a-CN:H samples.
Here, x D and s D are the position and FWHM of the D peak, x G and s G are thosecorresponding to the G peak. I D /I G and A D /A G are the ratios of intensities and areasbetween D and G peaks, respectively. he effect of nitriding on the humidity sensing properties of a-C:H films − x N x :H have determined the shift of x D and x G and the increasing of A D /A G as given by a Gaussian fit, as a function of the x content of N [9, 10]. By comparing our results for the shift of the G peak and therelation of the areas, we estimated that there could be an incorporation of nitrogen inthe sample structure of the order of 10%.The increase of I D /I G and the shift of x G were related to the increase of the aromatic sp zones [11]. This structural change produced an increase of the DC conductivity,which was mainly associated with an increase of the density of states at the Fermi level N ( E F ) [9, 12].Independent information on the structure of the film was obtained by IRspectroscopy. We used the Spectrometer One by Perkin-Elmer to obtain the resultshown in Fig. 2, which corresponds to the sample before the plasma treatment. C-O C-H
Wavelength (cm − ) I n t e n s it y ( u . a ) C-O C-H
Wavelength (cm − ) I n t e n s it y ( u . a ) Figure 2.
The IR spectrum of the film before plasma treatment at room temperature.
The peaks were attributed to different configurations of sp orbital. For the peakat 2920 cm − the stretch vibration C-H was assigned sp [13]. For the peak 1720 cm − it was attributed to vibration C-O, indicating that the film had hydrogen in a very lowconcentration.N incorporation into the a-C:H structure produces important changes in the IR andRaman spectra, as has been shown in previous studies [10, 14]. In Fig. 3 we show theRaman spectrum of the sample before and after the plasma treatment. The existence ofa peak in the 3360 cm − which is assigned to a NH stretching mode can be clearly seenin the signal corresponding to the a-CN:H sample. The appearance of this mode after he effect of nitriding on the humidity sensing properties of a-C:H films − ) I n t e n s it y ( a . u . ) a-C:Ha-CN:H Figure 3.
The Raman spectrum of the film before (a-C:H) and after (a-CN:H) plasmatreatment at room temperature.
In summary, the effect of nitriding was a reduction of the thickness of the sampleand the incorporation of N into the structure. The resulting microstructure is similar tothe one seen in the previous work where the samples were prepared by decomposition of(C H , N ) gas mixtures in a distributed electron cyclotron resonance (DECR) plasma[9]. It is a transition from a diamond-like a-C:H films towards a graphite-like N-richa-C − x N x :H alloy. Note that the precise microstructure of a-C − x N x :H is still underdiscussion.In the next section we will show that the way conductivity changes with RH alsochanges when nitriding.
3. Humidity dependence of the AC transport properties ◦ C. he effect of nitriding on the humidity sensing properties of a-C:H films
7A Lock-in SR530 amplifier was used to perform the electrical characterization of theamorphous carbon film. From the measurement of the current flowing through the film,the impedance was obtained by keeping the frequency fixed and varying the humidityor vice versa.The acquisition of the measurements data was carried out through a PC that wasconnected to the Lock-in amplifier by means of an SR280 interface. Using a Pythonprogram, complex current measurements taken by the Lock-in amplifier were processedand transformed into complex impedance values. Therefore, the frequency, humidityand temperature values were registered and controlled inside the chamber at the sametime.First, the humidity inside the chamber was set to a value and then measurementswere made by sweeping the frequency values from 1000 Hz to 100 kHz with a certainstep.In a previous work [4], we showed that after nitriding the low frequency impedancechanged with RH. We also showed that the RH dependence of impedance was dominatedby its resistive component. Then, the films could be the basis of resistivity humiditysensors.Therefore, we started by studying the low frequency resistance (essentially its DCvalue) as the sensible property. R was obtained as the real part of the impedance Z . InFig. 4 we show how R/R changed with the relative humidity ( RH ) of the surroundingmedia, where R was a reference resistance at RH = 15%.From Fig. 4 it can be seen that the slope of the linear regression of the a-CN:Hfilm became greater than the one of a-C:H. In other words, the sensitivity of the sampleincreased after the plasma treatment.
20 40 60 800 . . . . R / R = A × ( RH − RH ) + hh = . A = − . RH - RH ( % ) R / R Experimental valuesLinear regression
30 40 50 60 70 80 900 . . . . . R / R = A × ( RH − RH )+ hh = . A = − . RH ( % ) R ( Ω ) Experimental valuesLinear regression
Figure 4.
Resistance of a-C:H (left) and a-CN:H (right) films.
Let us undertake a comparison of the sensitivity of a-C:H and a-CN:H samples. Bydefining the sensitivity as S = R RH − R RH R RH , we measured it and determined that itchanged from 0 .
40 to 0 .
85 when N was incorporated. We arrived at the conclusion that: he effect of nitriding on the humidity sensing properties of a-C:H films • a-C:H samples were already humidity sensitive, and • − Z (cid:48)(cid:48) vs Z (cid:48) . Then, we analyzed thefrequency dependence of the complex conductivity ( σ ). Following the same measurement procedure described above, we swept the frequency inthe previously stated range and then changed the humidity in the chamber. In Fig. 5we plotted the imaginary part of the complex impedance ( Z (cid:48)(cid:48) ) as a function of the realpart ( Z (cid:48) ), before and after the plasma treatment: these are the so called Nyquist plot ofimpedance [16]. At the right figure the results of both samples are superposed whereasat the left we zoomed in the a-C:H samples in order to show their behaviour. In thesecurves the frequency is taken as a parameter and increases counterclockwise. · · Z ( Ω ) − Z ( Ω )
15 %RH49 %RH65 %RH82 %RH100 %RH0 0 . . · · Z ( Ω ) − Z ( Ω )
15 %RH30 %RH68 %RH84 %RH100 %RH
Figure 5.
Imaginary part vs. real part of the complex impedance due to differentrelative humidity conditions, for a-C:H and a-CN:H films (right). On the left wezoomed in on the pink rectangle on the right in order to distinguish the behaviour ofa-C:H. he effect of nitriding on the humidity sensing properties of a-C:H films
As we stated before, we will now study the conductivity as a function of frequency, fordifferent RH values. From the measured values of Z (cid:48) and Z (cid:48)(cid:48) , the values for the real ( σ (cid:48) )and imaginary ( σ (cid:48)(cid:48) ) parts of the admittance can be obtained: σ (cid:48) = eA Z (cid:48) Z (cid:48) + Z (cid:48)(cid:48) , σ (cid:48)(cid:48) = eA Z (cid:48)(cid:48) Z (cid:48) + Z (cid:48)(cid:48) , (1)where e and A are the thickness and the area of the sample, respectively. We do notinclude A in the following figures because it is the same for all the samples and resultin a constant shift of the y-axis. . . . − − − − − −
34 ln ( f ) l n ( σ )
100 %RH84 %RH68 %RH30 %RH15 %RH 7 7 . . . − − − − − −
34 ln ( f ) l n ( σ )
100 %HR84 %HR68 %HR30 %HR15 %HR
Figure 6.
Log-log plots of the real (left) and imaginary (right) part of the complexconductivity vs. frequency for an a-C:H film at different relative humidity conditions.
In figures 6 and 7 we can see the real (left) and imaginary (right) parts of theconductivity before and after the plasma treatment, respectively. In both figures, thereal part shows greater dependence on the humidity conditions than the imaginary part. he effect of nitriding on the humidity sensing properties of a-C:H films . . . − − − − − −
34 ln ( f ) l n ( σ )
100 %RH82 %RH65 %RH49 %RH15 %RH 7 7 . . . − − − − − −
34 ln ( f ) l n ( σ )
100 %RH82 %RH65 %RH49 %RH15 %RH
Figure 7.
Log-log plots of the real (left) and imaginary (right) part of the complexconductivity vs. frequency for an a-CN:H film at different relative humidity conditions.
For the imaginary part, the plasma treatment causes the curves at different RHto separate slightly and vary in a little smaller range as a function of the frequency.On the contrary, the real part shows a greater variation on the humidity conditions.For the a-C:H films, all the curves show similar slopes, they are almost parallel. Itcan be seen that they have a type of universal behavior that resembles the ones foundfor AC conductivity of some amorphous semiconductors at different temperatures [6].This changes noticeably for the a-CN:H films, showing different slopes: the higher thehumidity, the higher the conductivity but the lower the slope as a function of thefrequency. Furthermore, it is important to note that all these curves converge at thesame point at high frequencies. We note that this behavior resembles the one shownby thermal conductivity as a function of temperature, see for example Ref. [18]. Theauthors of this work show that the vibrations of the lattice of these disordered crystalsare essentially the same as those of an amorphous solid.
4. Conclusion
In summary, nitriding a-C:H samples result in a-CN:H where N is incorporated intothe structure. The electronic transport properties were affected by this treatment. Inparticular, the resistance of a-CN:H was more sensitive to the change of the RH ofthe surrounding medium that the one of a-C:H. In summary, nitriding a-C:H samplesresult in a-CN:H where N is incorporated into the structure. The electronic transportproperties were affected by this treatment. In particular, the resistance of a-CN:H wasmore sensitive to the change of the RH of the surrounding medium than the one ofa-C:H. As we stated in our previous work [4], the sample remains humidity sensitive forat least nine month. It is worth mentioning that the sample, during this period, wassubjected to a large number of cycles of changes in the condition of relative humidity.By AC spectroscopy we have observed that the evolution of σ with the frequency he effect of nitriding on the humidity sensing properties of a-C:H films
5. Acknowledgment
We acknowledge V. Roldan, G. Baranello, H. Rindizbacher and Javier Cruce˜no for theiressential contributions to resolving all the technical issues to undertake the present work.This work was supported by CONICET and UNR. Leonel Silva acknowledges AgenciaNacional de Promoci´on Cient´ıfica y Tecnol´ogica, Fondo para la Investigaci´on Cient´ıficay Tecnol´ogica (PICT16-3633) for financial support.
References [1] Farahani H, Wagiran R and Hamidon M N 2014
Sensors ACS nano Micromachines Materials Research Express Physica Scripta [6] Baranovski S 2006
Charge transport in disordered solids with applications in electronics vol 17(John Wiley & Sons)[7] Tai F C, Lee S C, Chen J, Wei C and Chang S H 2009
J. Raman Spectrosc. Applied Spectroscopy Thin Solid Films
Growth, structure, and properties of plasma-deposited amorphous hydrogenatedCarbon-Nitrogen films (Elsevier)[11] Ferrari A C and Robertson J 2000
Physical Review B Diamond and Related Materials Diamond and Related Materials Physical Review B Bull. Chem. Soc. Ethiop. Impedance spectroscopy: theory, experiment, andapplications (Hoboken, NJ: Wiley)[17] Yao Y, Chen X, Zhu J, Zeng B, Wu Z and Xiaoyu Li X 2012
Nanoscale Research Letters Physical Review B46