Structural Characterizations
The non-Newtonian electrolyte solution was prepared as discussed in the experimental section. The chemical structure of the AS is depicted in the inset shown in Fig. 1. The functional groups of AS@Sodium Chloride electrolytes were identified using infrared spectroscopy. The individual samples of AS in powder and liquid state were scanned between 400 cm− 1 to 4000 cm− 1. The reference FTIR spectrum of AS powder matched well28,29 with characteristic peaks at ~ 1013 cm− 1 (5), 1416 cm− 1 (4), 1604 cm− 1 (3), 2915 cm− 1 (2), and, 3310 cm− 1 (1) shown in Fig. 1(a). Peak (1) corresponds to the stretching while a shoulder at 2915 cm− 1 (2) represents the C-H stretching, commonly expressed in organic compounds. As evident from the chemical structure shown in the inset of Fig. 1, the carbonyl group symmetric stretching was observed in the FTIR spectra at (3) in Fig. 1(a). A small peak at (4) might be attributed to the asymmetric stretching, and a sharp peak at 1013 cm−1, and (5) is due to the stretching. The FTIR spectra of the AS dissolved in NaCl solution is shown in Fig. 1(b). Subsequently, the values of percentage transmittance of both spectra are compared, indicating that, stretching became more pronounced and shifted right by ~ 30 cm−1 in the solution. Numerous other shifts in wavenumber were also observed. The symmetric stretching shifted to 1635 cm− 1, and the stretching to 1031 cm− 1. A small peak at 1416 cm−1 became more pronounced, however, no shift was observed. At last, a hypsochromic shift in stretching was observed on dissolution. The transmittance in the FTIR spectra has been normalised to show the comparison between intensities in Figure S1. The properties of AS@NaCl obtained from the FTIR spectra provide information about the functional groups which will be utilized in subsequent sections to speculate and analyse the role of concentration in memristive action.
To validate the non-newtonian property of AS as mentioned in30,31, strain-rate (\(\dot{\gamma }\)) (in s− 1) vs viscosity (\(\eta\)) characteristics are presented in Fig. 1(c), (d) for two different weights of AS in 4.67M NaCl solution. The observed characteristic was fitted with the Sisko model32. The stress (\(\sigma\)) was normalised to have appropriate comparison between the two weights. For low strain rates (1-100 s− 1), shear thinning (Non-Newtonian) property was apparent for 0.3g containing AS@NaCl, however for the 0.45g containing AS@NaCl, shear thinning property of AS was observable over a wide range of strain rates (1-500 s− 1). This behaviour will be further utilized to examine the I-V response of the device.
Electrical Characteristics
The non-Newtonian electrolyte-based memristor devices were fabricated using the procedure mentioned in the experimental section. I-V characteristics of the device were measured with a DC voltage sweep (0 V → +1 V → 0 V → -1 V → 0 V). The memristive behaviour and its cycle-to-cycle variation is shown in the semi-log I-V plot in Fig. 2(a) at an optimized compliance current value of 1mA and scan rate of 0.1V/s, respectively (Figure S2). Initially, in the positive voltage sweep from 0 → 1V, the device shows LRS (low resistance state) as represented by (1) in Fig. 2(a). Then in the reverse positive sweep, the device switched to HRS (high resistance state), representing unipolar switching. A negative dual sweep validates the symmetric unipolar switching. The inset in Fig. 2(a) shows the linear I-V characteristic of the device inferring the capacitance stored by the accumulation of ions at the electrode/electrolyte interface forming an electric double layer. This electric double layer might be due to the availability of carboxylate ions from AS in the aqueous medium (water). The presence of the double layer can also be observed from the charge-voltage characteristics with the finite charge at the end of the complete voltage sweep shown in Figure (S3). The memristance can be obtained by performing time dependent on the device and examining the charge-time relation shown in Fig. 2(b) to confirm the non-passivity (i.e presence of non-zero current at zero voltage) of the device. The data in Fig. 2(c) indicates another indication for the non-passivity of the device with a non-zero charge at zero flux. The slope of Fig. 2(c) at a fixed value provides the value of memristance,. The hysteresis in the figure indicates the behaviour of a device analogous to a memristive system rather than an ideal memristor; a non-linear function of the state variable, i.e .
In order to understand the switching mechanism, double log I-V characteristic is presented in Fig. 2(d). Region 1 with a slope value close to zero (positive sweep from ~ 0–0.1V) is attributed to the formation of double layers33 due to the accumulation of coions and counterions at the electrode/electrolyte interface. In this regime, the current weakly depends on the bias values. Regime 2 (~ 0.1–1V) is attributed to bulk ionic conduction from the electrolyte where \(log I\) varies exponentially with \(log V\). In analogy with solid state memristors, this region is analogous to the Space charge limited current (SCLC) region, where \(I=k{V}^{2}+q, k\) is the proportionality constant, and \(q\) is any arbitrary constant with value close to zero. The value of the proportionality constant, k was found to be \(8.130*{10}^{-4}\).The I-V hysteresis depicted above Fig. 2(a) is due to ionic transport in non-Newtonian shear thinning solvent (i.e dynamic viscosity of solvent decreases with the increase of shear stress) of AS@NaCl. In the reverse positive sweep (1V → 0.6V), due to the shear-thinned state of the solvent (by the traversal of ions) in the forward positive sweep, the mobility of ionic species is higher compared to the mobility of the same species in the pristine state of solvent, permitting the ions to drift back as shown by region 3 with a large slope value of ~ 3.3 in Fig. 2(d). Lastly, regions 4 and 5 between 0.6V to 0V represent the renascence of the role of the double layer in conduction of ionic species. This region being again analogous to SCLC can be utilized to compare the mobility of ionic species by value of previously defined proportionality constant. For the SCLC region here, \(I=k{V}^{2}+q\) region, value of k is 0.00151 which is nearly double than previous one, validating the mobility increase statement mentioned earlier. To speculate the action of AS for memristive behaviour, three solutions with different proportions of AS in 4.67M salt solution with 1:1 (0.15g), 1:2 (0.3g), 1:3 (0.45g) (wt. of NaCl: AS) were prepared and annealed at 70°C with uniform stirring. A non-zero current at 0V remains consistent across all variations in Fig. 2(e), however, the current in forward and reverse positive voltage sweep are identical for the device prepared with a 1:2 electrolyte solution. The cycle-to-cycle variation for the three electrolytes (1:1, 1:2, 1:3) is shown in Figure (S4). The devices with a 1:2 concentration ratio reverted back to the initial current value at the end of the positive voltage sweep, and the OFF/ON ratio of 4 was optimum among other concentrations. The consistent switching was observed by examining the endurance characteristics of the device measured up to 200 cycles as shown in Fig. 2(f). The energy consumption of the device with two different channel diameters (1.2 cm and 0.2 mm) is calculated in Figure S5. The working and switching mechanism presented here will be validated by the impedance analysis presented in the subsequent section.
To examine the role of electrodes in the mentioned device configuration, the I-V characteristics with stainless steel-copper electrode system is presented in figure S6 which shows memristive action because of electrochemical metallization. The oxidation of copper anode to Cu2+ at -0.18V takes place which is carried through the non-Newtonian electrolyte towards the cathode and the 2e– remains at the anode. When the applied potential is strong enough, the double-layer ions is no longer present near the cathode, which allows Na+ and Cu2+ to accumulate near the cathode. The accumulation gets stronger with the applied bias in the positive voltage sweep. All of these changes are demonstrated in Figure S6 where a bump in the negative sweep from 0 → 1V at -0.18V is visible in the copper electrode system. In the absence of the Cu electrode, in the former device configuration, the oxidation/reduction reaction of Cu2+ does not play any role and only the presence of Na+ and Cl– in the electrolyte leads to the memristive behaviour.
Impedance Analysis
In order to elucidate the working and switching mechanism of the AS@NaCl electrolyte-based memristive device, a sequence of chrono-amperometry (CA) and electrochemical impedance spectroscopic (EIS) measurements were performed. In the sequence (CA-EIS-CA), the CA was performed before and after each EIS measurement for 5 seconds34,35. The former CA- measurement in the sequence provides the information about initial resistance state of the device at a fixed applied voltage. The EIS measurement in the sequence was performed with a read voltage of 10mV in the frequency range of 0.1Hz − 0.1MHz. The later CA measurements validate that the device on which EIS has been performed is not degraded and the initial resistance states are invariant at the read voltage. This sequence was conducted by varying DC voltage from 0V → 1V → 0V. The Nyquist plot for the varying voltage range from 0V → 1V is shown in Fig. 3(a).
The evolution of the impedance spectra at applied DC bias > 0.4V in the low-frequency regime (Fig. 3(a)), further confirms the transition of HCS (High conductance state) to LCS (Low conductance state) as observed from I-V characteristics (Fig. 2(a)). The imaginary component (\(Im\left(Z\right)=-{Z}^{{\prime }{\prime }}\)) of the impedance Z = \(Z{\prime }\) - jw\(Z{\prime }{\prime }\), starts forming another small arc in the Nyquist plot with the real part (\(Re\left(Z\right)=Z{\prime }\)). To distinguish between a small arc and Warburg impedance at low frequency, a double log plot between Z and \(\omega\) plot is shown Fig. 3(b) indicating a slope value different than − 0.5 (a requirement for the presence of Warburg impedance). Generally, the low-frequency regime is attributed to the interfacial effects due to the action of ions near the electrodes, and the high-frequency region is representing bulk effects. Based on the invariant behavior from the high-frequency regime, the resistive switching in these devices can be assigned to the dynamics of ions near the electrodes. It is noteworthy that, inert SS electrodes were used for these investigations, however, the active medium comprises reactive sodium metal ions, which initiates the possibility of sodium metal ion’s reaction with the electrode’s surface. Owing to the difference in oxidation potential between AS medium and sodium standard oxidation potential the electroplating of sodium metal ions on the electrode's surface cannot occur. Furthermore, the poor OFF/ON ratio in highly concentrated electrolyte (\(0.45g\)) of AS is led by a different mechanism. The occurrence of deprotonation reaction by sodium metal ions with the electronegative oxygen centers (observed in terms of the presence of \(C=O\), and \(C-O\) stretching from FTIR spectra in Fig. 1(b)). Therefore, a precise concentration of AS is imperative to observe the memristive behaviour of the system. Also, this optimization in AS vs sodium chloride solution ensures the properties of the medium as a non-Newtonian fluid reported in30,36,37, which causes the resistive switching phenomenon observed here.
As observed from Fig. 3 (a), the transition of low-frequency data points (> 0.4V) can be ascribed to the presence of the electric double layer explicitly visible with a finite value of capacitance in Fig. 3 (c). Initially, these devices exhibit series resistance of ~ 50 ohms in the high-frequency regime at all the varying bias values because of the traversal of ions in the electrolyte. The equivalent circuit for the device was found as a result of fitting the IS data shown in Fig. 3 (d), (e), and (f). The circuit in Fig. 3(d) consists of a series resistance in series with a constant phase element accounting for the presence of a double layer in parallel with a resistor, and lastly, a capacitor in parallel connection with a resistor. The details of each of the circuit elements fitted here with reference to the IS data can be found in Supplementary Table 1. The presence of an electric double layer depicted by the constant phase element, \({Z}_{CPE} = \frac{{e}^{-\frac{\pi }{2}ni}}{{{Q}_{0}\omega }^{n}}\) with n = 0.8, validates the capacitive nature of the electric double layer. This fact is further utilized in the modeling and the simulation section, to estimate the transport properties of ions through the shear thinning Non-Newtonian fluid using the Modified Poisson - Boltzmann approach.
Synaptic Characteristics
To evaluate the potential of the device for neuromorphic computing, preliminary synaptic characteristics are presented in this section. The description of neuronal dynamics adapted from39, 47 is mainly dependent on the functioning of the synapse formed at the synaptic cleft between the pre-synaptic neuron and post-synaptic neuron. The formation of synapse in presence of an external stimuli is governed by the ionic transport of Na+ and K+ ions, present in and out of the cell membrane of neurons. Non-Newtonian electrolyte-based memristor presented here is capable to be implemented in a neural network with appropriate mobility of ionic species in the electrolyte as an active. Figure 4(a) expresses the artificial synapse response with the excitation interval. The input signal magnitude in the biological neural network will be multiplied by the weight and sent to the connecting post-synaptic neuron, and this weight can be modified by the preceding stimuli. Keeping the pulse amplitude at 0.5V and pulse width 5 ms, 10000 repetitive pulses were applied to the device in order to evaluate the current and conductance response of the device with the number of pulses. This pulse train was applied to the device by varying pulse interval to examine the plasticity of the active material. In Fig. 4(a), the pulse interval was kept at 10 ms and pulse width at 5 ms. With the number of cycles, the conductance of the device decreased, mimicking the behaviour of variation of synaptic plasticity with the time interval of excitation pulse between pre-synaptic neuron and post-synaptic neuron26,27. When the pulse interval was increased from 10–90 ms, a linear decrease with the number of cycles was observed.
In a similar manner, when 10000 repetitive pulses of -0.5V pulse amplitude and 5 ms pulse width were applied, the device expresses potentiation characteristics with the number of cycles. However, on varying the pulse interval from 10–90 ms, shown in Fig. 4(b), the device’s response to each pulse of the device is analogous to short-term plasticity in the human brain38,39,40. This behaviour of the device indicates the voltage polarity-dependent plasticity characteristics. In order to examine the plasticity characteristics, an increasing voltage sweep from 0 to 1V with read of 0.05V was applied for varying pulse interval shown in Figure S7.
To evaluate the plasticity characteristics of the device further, 20 triangular pulses of -0.7V and 0.6V were applied with a pulse interval of 10ms. The device exhibited a decrease in conductance up to a fixed value with the number of pulses, when the voltage was − 0.7V, and an increase in conductance when the voltage was 0.6V demonstrated in Figure S8. This behaviour of the device is validating that the device is exhibiting potentiation and depression characteristics analogous to the synapse formed in the synaptic cleft in response to the external stimuli in the human brain.
The PPF (Paired Pulse facilitation) behaviour was analyzed for the device which is again a strong measure for validation of synaptic properties, in the device. When two pulses of amplitude 0.5V, with pulse width of 5 ms were applied to the device, during the first pulse, the current I1 was produced and during the second pulse, current I2 was produced. Contrary to the general observed behaviour by synaptic devices, where I2 > I141, the device presented here exhibited an inverted trend with I2 < I1. Therefore, rather than exhibiting paired-pulse facilitation, the device was exhibiting paired-pulse depression. This nature of device can be attributed to the high ionic mobility and finite concentration of ions in the medium. Because, the ions responsible for current conduction through the device have high mobility, on the application of the first stimuli to the device, most of the ions present in the medium are collected near electrodes or get trapped in the medium. Immediately, after the first pulse, in the duration of milliseconds, when the second pulse was applied to the device, the trapping probability decreases and the remaining ions traverse to the electrodes. Although the mobility of ions is higher attributed to the shear thinning effect caused by the initial pulse, because their number density is low, the current generated by the device is smaller than the current generated in the initial pulse. The pulse interval between the two stimuli was varied and the negative PPF-index calculated using Eq. 100(I2 - I1)/I1 for each pulse interval is shown in Fig. 4(c). This behaviour is the two-phase behaviour commonly observed in the biological synapse and can be fitted by equation, -PPF = 1 + ae− bt + ce− dt, where t is the pulse interval, and a, b, c, and, d are constants. For the PPF behaviour shown above a = 10.5488, b = 0.1328, c = 5.9839, and d = 0.0069. Here, a and c are the initial facilitation magnitudes of respective phases, and, \({\tau }_{1} = 1/b\) and \({\tau }_{2} = 1/c\) are the characteristic relaxation time of respective phases. Since, b > d, τ1 < τ2, which implies that the initial phase lasts shorter compared to the later phase which goes in hand with the solid-state memristors. This PPF behaviour has already been observed in non-newtonian shear thickening memristors42, however, the rheology was a point of closest miss till now, which is clearly evident from above the discussions.
Although the AS@NaCl solution as an active medium was found to have volatile unipolar resistive switching which limits its application to neuromorphic computing, the device expresses multi-state resistive switching behaviour with the change of compliance current shown in Fig. 4(d) which exploits its use case to the numerous cross bar related applications.
In the biological neural network tree like nerve fibers are extending from the cell body of a neuron, called dendrites. These dendrites receive signals from numerous other neurons. A single long fiber extending from the cell body, called axon, branches and connects to many other neurons at the synapses or synaptic junctions. The axon of a neuron typical leads to a few thousand synapses associated with other neurons. The synaptic activity occurs between neuron passing the signal (pre-synaptic neuron) and the neuron receiving the signal (post-synaptic neuron) through modulation of conductance at the synaptic cleft. Hence, timing of signals at the neurites play a vital role in synaptic strength modulation. This timing of signals can be studied in terms of Spike time dependent plasticity (STDP), where Relative weight change (\(\varDelta G/{G}_{0}\)) is analyzed as a function of time delay \(\varDelta t\) between signals from pre-synaptic and post-synaptic terminals.
As we mentioned earlier, the two terminal memristor is archetype to the synapse, it becomes necessary to investigate STDP characteristic in such device to understand the underlying learning principle. For the memristive system presented here, the two electrodes resemble the pre/post-synaptic terminals and the AS@NaCl is the active material, where synaptic activity is expected. Figure 5, demonstrates four of STDP behaviour exhibited by the device, with the respective pulse waveform shown in the inset. As the memristive device is unipolar in nature, it can be potentiated/depressed with pulses of same polarity. This characteristic was utilized to report STDP characteristic for the first time in such an ionic liquid based memristor. Figure 5(a) shows Asymmetric Hebbian STDP which can be fitted by \(\frac{\varDelta G}{{G}_{0}}={A}_{1}{e}^{-\frac{\varDelta t}{{\tau }_{1}}} , \varDelta t<0\) and \(\frac{\varDelta G}{{G}_{0}}={{A}_{2}e}^{\frac{\varDelta t}{{\tau }_{2}}} , \varDelta t>0\). The \({(\tau }_{1}, {\tau }_{2}) =(0.0058, 0.0109)\) values were calculated from the fitted data. To obtain the Asymmetric Anti-Hebbian STDP shown in Fig. 5(b), the additive inverse of the pulse waveform used for the Asymmetric Hebbian was deployed. The \({(\tau }_{1}, {\tau }_{2}) =(0.0031, 0.0114)\) for the Asymmetric Anti-Hebbian STDP. For the symmetric type of STDP characteristic, the polarity of weight change remains the same with order of firing of stimulus from the pre and post-synaptic terminals. This symmetric STDP with Hebbian and anti-Hebbian are evaluated in Fig. 5(c), and 5(d) for the device. The fitting curve was assumed to be gaussian with an additional y-intercept, with mean zero, and standard deviation as the decay time, t1. For the symmetric Hebbian STDP shown in Fig. 5(c), the decay time t1 was found to be 0.0061 seconds. And, for the symmetric Anti-Hebbian STDP, the decay time \({t}_{1}=0.0159\). Evaluating and observing the Spike Time dependent property in the presented device, not only confirms the potential of the device for its cross-bar implementations, but also poses new questions to study related to the presence of other Spiking properties like Spike Rate dependent Plasticity, Spike Number dependent plasticity, and various others. These questions are far from the present focus of the article and hence will be presented somewhere else in further studies.
Model Fitting
In the preceding sections of this article, the role of the electric double layer near the electrodes and the shear thinning property of AS has been speculated to be responsible for the observed working of the device. In this section, based on the qualitative analysis from the impedance spectroscopy and the equivalent circuit obtained, a model has been developed using the equivalence of lipid bilayer in neurons with the double layers formed near the interface of the dielectric-charged surface. Also, the account for the shear thinning property of the medium with applied bias has been incorporated into the model by an additional ohmic term and a constant current term. Lastly, the prepared model has been utilized to explain the I-V characteristics of the device with varying concentrations of AS. The model has been developed on the grounds of classical cable theory which explains the current response of a dendrite (axon) with the synaptic bias input43,44. The theory provides the following equivalence between the current through a dendrite and the applied voltage.
$$I = \frac{a}{2R}\frac{{\partial }^{2}V}{\partial {x}^{2}}$$
1
Here, \(a\) is the radius of the dendrite (axon), \(R\) is the specific resistance of the cytoplasm within the dendrite. For a z:z symmetric electrolyte, the electric double layer theory, provides the following variation of potential (V) with ionic concentration, \({n}_{0}\).
$${\nabla }^{2}V = \frac{2ze{n}_{o}}{\epsilon }\frac{sinh\left(\frac{zeV}{kT}\right)}{1-\varphi + \varphi cosh\left(\frac{zeV}{kT}\right)}$$
2
In this expression \(\varphi\) denoted the steric factor defined as \(\varphi = 2{n}_{0}{r}^{3}\), where r is the radius of ions and \({n}_{0}\)is the bulk concentration of the device, respectively. Here, \(k, e and T\) refers to the Boltzmann constant, electronic charge, and temperature, respectively.\(\varphi\), the steric factor accounts for the presence of a double layer due to finite ion size near the charged surface. The current response of the device due to this double layer is independent of applied bias, represented by the slope value of zero in Fig. 2(d), and is only visible in the low bias regime (< 0.1V).
To explain, the I-V characteristics of the device, we may assume the equivalence of equations (1) and (2), between the electric double-layer potential, and the current response of a neurite in response to the applied voltage, for the 1D case45,46,33. For the case presented here, the valency z for the NaCl electrolyte is 1.
$$I = \gamma \frac{sinh\left(\frac{V}{\psi }\right)}{1-\varphi + \varphi cosh\left(\frac{V}{\psi }\right)}$$
3
Here,
$$\gamma = \frac{a}{2R}\frac{2ze{n}_{0}}{\epsilon }$$
To incorporate the shear thinning nature of AS@NaCl electrolyte, an ohmic term are added to Eq. (3) attributing to the resistance offered by the medium. On increasing the voltage, the portion of the electrolyte in the shear-thinned state increases, which in turn increases the mobility of ions traversing through them. This shear-thinned state of the electrolyte can be assumed to be analogous to the ohmic conduction through the device. To elucidate the role of Non-Newtonian behaviour of the fluid further, viscosity as a function of temperature was evaluated for the AS@NaCl and was found to be decreasing shown in Figure S9(a). The viscosity value of the AS@NaCl at \(25 ℃\) was measured as 405 centipoise which reduced to 195.6 centipoise at \(70 ℃\). Additionally, the I-V, was measured as a function of temperature with temperature range \((25 ℃, 70 ℃)\), shown in Figure S9 (b). The increasing feature of High current state (HCS) (OFF state) with temperature was evident from the effect of double layer formed within the device. However, the unusual trend in the resistance switched state, i.e Low current state (LCS) in Figure S9(c) is an effect of change in viscosity faced by the ions corroborated from the trend in Figure S9(a). The OFF-ON ratio, i.e (HCS/LCS) at 0.5V was on an average ~ 4 maintained over the entire temperature range, shown in Figure S9(d). The temperature dependent variation in LCS in addition to the device’s viscosity as a function of temperature confirms the role of shear thinning affecting the traversal of ions/formation of double layer within AS@NaCl device. The constant current term has been added to the equation, to express a comparable current (i.e current value above 0V is comparable to that at 0V) value in zero bias condition. The dynamical equation obtained after incorporating all of the above-mentioned factors is as follows.
$$I = {p}_{0} + V/{R}_{ohmic} + \gamma \frac{sinh\left(\frac{V}{\psi }\right)}{1-\varphi + \varphi cosh\left(\frac{V}{\psi }\right)}$$
4
This Eq. (4) has been fitted to the I-V characteristics (Fig. 2(e)) of the device with varying concentrations of AS (0.15g (1:1), 0.3g (1:2), 0.45g (1:3)), with \({p}_{0},{R}_{ohmic}, \varphi , \gamma ,\)and \(\psi\) as model fitting parameters. The values of these parameters are summarised in Table 1. Increasing concentration of AS in NaCl results in the increase of \(\gamma\), and the\({ R}_{ohmic}\). However, the parameters \(\psi\), \(\varphi\), and \({p}_{0}\) show no trivial trend with concentration. The value of \(\varphi\) (steric factor) is least for 0.3g of AS, and the value of \(\psi\) and \({p}_{0}\) are maximum for the same concentration. The result of the fitted data with three different concentrations is shown in Fig. 6.
Table 1
The table lists the model parameters, \({R}^{2}\) and \({\chi }^{2}\)value to mention the accuracy of fit.
Conc. of AS
|
\({\psi }\)
|
\({\varphi }\)
|
\({\gamma }=\frac{{a}}{2{R}}\frac{2{e}{{n}}_{0}}{{ϵ}}\)
|
\({{p}}_{0}\)
|
\({{R}}_{{o}{h}{m}{i}{c}}\)
|
\({{R}}^{2}\)
|
\({{\chi }}^{2}\)
|
0.15 g
|
0.0848
|
1.39*(10− 3)
|
9.97*(10− 7)
|
1.78*(10− 5)
|
4363.46
|
0.999
|
100.96
|
0.30 g
|
0.2468
|
1.17*(10− 11)
|
4.24*(10− 4)
|
1.40*(10− 4)
|
11160.79
|
0.999
|
100.95
|
0.45 g
|
0.1937
|
1.29*(10− 2)
|
1.89*(10− 5)
|
1.97*(10− 23)
|
4.32*(1013)
|
0.997
|
100.98
|