1 Introduction

The presence of ions in the atmosphere is the resultant balance between the ion production rate and loss of ions through different processes. In the lower atmosphere/troposphere, cosmic rays are the primary source of ionization and are the main source of ionization above a height of ~ 1 km (Hoppel et al. 1986; Carslaw et al. 2002; Siingh and Singh 2010; Singh et al. 2011; Kumar et al. 2018). Geomagnetic fields influence the movement of cosmic rays in the Earth’s environment; hence ionization produced in the atmosphere is dependent on geomagnetic latitude and solar variability. However, near the Earth's surface, the source of ionization is dominantly radioactive radiations and their daughter products. Bricard (1965) estimated the relative contributions of ion production rates from air, soil and cosmic origins to be 4.5, 3.5 and 2 ion-pairs cm−3 s−1, respectively in continental regions. In the tropical zone near the Earth's surface, ionization processes are controlled by Radon and its progenies (Hensen and van der Hage 1994). In addition, several other natural sources of small ions production near the ground are corona ions produced in large electric fields, splashing of rain drops at the ground, breaking of water droplets near to the waterfalls and sea shores, dust storms, volcanoes, etc., (Chalmers 1967; Blanchard 1963; Siingh et al. 2005; Horrak et al. 2006; Kolarz et al. 2012; Kamra et al. 2015a). The mechanism responsible for the production of charge and the nature of atmospheric air ions distributions shows that temporal and spatial variations are not well understood.

Atmospheric air ions play a primary role in the discipline of atmosphere electricity because their motion in the atmosphere determines the air conductivity and air-earth current (Wilson 1924; Siingh et al. 2007a). In recent decades, the area of study has become more attractive because ions are actively involved in the formation of secondary aerosol particles (Hoppel 1985; Yu and Turco 2000; Kulmala et al. 2004; Gopalakrishnan et al. 2005; Yu 2010; Siingh et al. 2013a; Tammet et al. 2014) and as such these ions are found to influence the quality of air and climate system (Waring et al. 2011). Air ions consist of charged particles of diverse chemical compositions. They grow from molecular clusters to large aerosol particles and attain a characteristic distribution in mass and size (Tammet 1998). The molecular clusters behave like kernels of condensation of nucleating vapors, converting them from the gas into aerosol particles. The balancing nature of ions is manifested in the atmosphere in many processes, including the balance between cluster ions and aerosol particles via ion-aerosol attachment and the balance between neutral and charged aerosols (charging state of aerosol particles).

It is essential to understand the mechanism of the growth of the newly formed particles to the size of cloud condensation nuclei and cloud processes under different environmental conditions to estimate the radiative forcing on the regional and global scales (Kulmala et al. 2004; Sullivan et al. 2018; Gautam et al. 2021). Observations to investigate the formation and growth of ultrafine particles have been made under a variety of environmental conditions (Kulmala et al. 2004, 2013, 2007b, 2011, 2018; Siingh et al. 2005; Tammet et al. 2006; Qian et al. 2007; Asmi et al. 2010; Betha et al. 2013; Kanawade et al. 2014; Kamra et al. 2015b; Huang et al. 2016; Dinoi et al. 2021; Gautam and Brema 2020; Gollakota et al. 2021; Ambade et al. 2021). Even though the physical/chemical process involved is not clearly understood, much more measurements under varying conditions are required.

Several mechanisms of NPF have been proposed such as homogeneous ternary nucleation (Kulmala et al. 2013), ion nucleation cluster (Kanawade and Tripathi 2006; Yu et al. 2008, 2010; Yu and Turco 2011; Gonser et al. 2014), activation of neutral clusters (Kulmala and Lehtinen 2006and nucleation mechanism involving organic vapors (O'Dowd et al. 2002a) or iodine (O'Dowd et al. 2002b). Yu (2010) based on a the kinetic model of ion-mediated nucleation mechanism showed that the nucleation rate non-linearly depends on sulphuric acid vapour concentration, temperature, and relative humidity ionization rate and surface area of existing particles. The studies involving kinetic model of ion-mediated nucleation showed that this process can explain the rapid initial growth of small clusters to a size of ~ 1.5 nm diameter by charge-neutral interactions (Yu and Turco 2001, 2008; Yu 2007, 2010). The measurements of Kamra et al. (2015b) support the theory of formation and growth of ions below the inversion layer by the ion-mediated nucleation mechanism and /or advection of ion with katabatic winds during the period of enhanced concentration of ions and aerosols when the solar radiation is absent or very low in the early morning hours.

Commonly used atmospheric electrical parameters are air conductivity, small ion concentration and heavy large ion concentration (Cobb 1977; Siingh et al., 2005). In the constant ionization rate, the concentration of small ions is inversely proportional to the diameter of concentrated aerosol particles (Salm et al. 1992; Horrak 2001). The concentration of heavy large ions is related to the content of Aitken aerosol particles (30–80 nm) in the air at the measurement site. The relation between the ion concentration and meteorological factors was primarily indirect, caused by the variation of ionization rate and aerosol particle content in air. The mobility of small air ions decreased by about 20% with increasing temperature from −5 to 25 °C (Salm et al. 1992) and is also influenced by the direction of wind blowing at the time and place of observation. The mean mobility of small air ions is an essential factor that can affect the conductivity of atmospheric air to a large extent. The deviation of mean mobility from the average value could cause changes in conductivity (induced by small ions) from about −12 to 15% and from −14 to 26% for positive and negative polarity, respectively (Siingh et al. 2005).

We carried out the ions measurement using NAIS for the year 2010–2012 at the same measurement site. We reported the few characterizations viz the monthly, seasonal, and annual distribution of air ions (Gautam et al. 2017), NPF events from trace gases (Siingh et al. 2013b) and case studies including rain-induced events (Kamra et al. 2015a, b). In all these studies, diurnal variation was not discussed both for event and non-event days. Therefore, in the present study, the diurnal variations of different categories of air ions during the event and non-event days for the year 2012 are presented. The ion production rate was calculated for both event and non-event days. Results are discussed in terms of change in Radon concentration and its dependencies on the surface stability parameters temperature and humidity. Polar air conductivity derived on event and non-events days from mobility-conductivity relation is discussed on clustering, aerosol concentration, and contribution of air ions to the total conductivity. Principal component analysis was carried out for a fraction of ions. The best-fitted components were explained by the characteristics of different ions during NPF and non-event days.

2 Data and Analysis

In the lower terrestrial atmosphere in the altitude range of 1 km above the Earth’s surface, radioactive gas 222Rn and its progeny. Therefore, simultaneous measurements of ion and Radon were carried out in the premises of Indian Institute of Tropical Meteorology, Pune (18° 31′ N, 73° 55′ E, 560 m above mean sea level), India, which is located in a valley surrounded by 500–800 m high hills on three sides. NAIS uses two unipolar corona dischargers to charge the sampled particles and for their subsequent detection. It has two identical cylindrical aspiration-type differential mobility analyzers (DMA), one for positive ions and another for negative ions. More details on instruments can be found elsewhere Mirme and Mirme (2013), Siingh et al. (2013b) and Kamra et al. (2015b).

NAIS post-processed the measured air ion concentration in the mobility domain (3.16–0.00133 cm2 V−1 s−1 mass diameter range 0.36–47.1 nm@). (@Henceforth, “diameter” should be read as “mass diameter” unless stated otherwise) into number sizes distribution (Mirme and Mirme 2013), which was subsequently converted to the diameter-based mobility distribution using size conversion model of Tammet (1995). The uncertainty in the air ion concentration measurements is < 10%. The data are used in quantitative ion cluster studies and accurate determination of formation and growth rates of the particles (Wagner et al. 2016). An average mobility for positive and negative ions is obtained in 5 min, based on 200-s sample air and 100-s offset-level measurements. The entire mobility distribution of the atmospheric ions are divided into two main classes, (1) cluster ions (mobility > 0.5 cm2 V−1 s−1; diameter < 1.6 nm) and (2) charged nanoparticles (mobility < 0.5 cm2 V−1 s−1; diameter > 1.6 nm) (Horrak et al. 2003). These two classes of ions have been further divided into five independent categories based on their size and mobility, viz, small cluster ions (mobility 3.2–1.3 cm2 V−1 s−1, size 0.36–0.85 nm diameter), big cluster ions (mobility 1.3–0.5 cm2 V−1 s−1, size 0.85–1.6 nm diameter), intermediate ions (mobility 0.5–0.034 cm2 V−1 s−1, size 1.6–7.4 nm diameter), light large ions (mobility 0.034–0.0042 cm2 V−1 s−1, size 7.4–22 nm diameter) and heavy large ions (mobility 0.0042–0.00133 cm2 V−1 s−1, 22–47.1 nm diameter) (Mirme and Mirme 2013).

The ion concentrations in the atmosphere show a diurnal and seasonal variation, which are explained in terms of solar irradiance, temperature, humidity, and wind velocity (Kamra et al. 2015b; Gautam et al. 2017). These meteorological parameters are taken from an automatic weather station (AWS) system installed at India Meteorological Department (IMD) campus, ~ 200 m away from the measurements site. At this site, there is a sizeable significant temperature variation with the minimum surface temperature (8–10 °C) during the winter months (December–January) and the maximum temperature often exceeding 40 °C during the pre-monsoon months (April–May). Relative humidity varies 40, 55, 70% in pre-monsoon, winter and post-monsoon, respectively, often exceeds 90% in the monsoon season. Winds were persistently southwesterly during the monsoon season and were comparatively low and changed directions in other seasons.

Radon and its daughter progeny concentrations were measured by Radon detector (RTM 2200 of SARAD GmbH, Germany) from January to December 2012. Details of measurement technique are given in Victor et al. (2019). The inter comparison of small ions derived from 222Rn detector and observed from NAIS has been shown in detail by Singh et al. (2013b) and Gautam (2017).

Ion-pair production depends on the ions-aerosol attachments and ion-ion recombination rate. The same has been calculated using the following equation (Hoppel and Frick 1986).

$$q\, = \,\alpha n^{2} \, + \, \, n\,\smallint \beta \delta \left( r \right)z\left( r \right){\text{d}}r$$
(1)

where, q is the ionization rate, z(r) is the particle size distribution (cm−3), n is the concentration of small ions (0.36–1.6 nm, diameter), α is the ion–ion recombination coefficient = 1.6 × 10–6 cm3 s−1, βδ = 0.436 r 9.2 × 10–8 is the single size attachment coefficient for small ions.

222Rn and its decay in associated with the release of energy which may produce ionization. Concentrations of 222Rn and its progeny were used to estimate total energy released, which was used in the ion-pair production. The ionization rate was then estimated from Q = ε/I, where ε is the total energy, I is the ionization energy (32 eV) required to produce one ion-pair in the atmosphere. The total energy ε is given by Green et al. (1977), Laakso et al. (2004), and Chen et al. (2016a),

$$\varepsilon \, = \,5.49\, \times \, \, 10^{6} \left( {{\text{C}}_{{{\text{222Rn}}}} } \right)\, + \,6\, \times \,10^{6} \left( {{\text{C}}_{{{\text{218Po}}}} } \right)\, + \,7.68\, \times \,10^{6} \left( {{\text{C}}_{{{\text{214Po}}}} } \right)\, + \,0.85\, \times \,10^{6} \left( {{\text{C}}_{{{\text{214Bi}}}} } \right) \, \left( {{\text{eV}}} \right)$$
(2)

C222Rn, C218Po, C214Po, and C214Bi are the concentrations of Radon and its progenies (in Bq m−3). In this process, the role of aerosols present in the atmosphere was not considered.

The spectral boundary for different sizes (small, intermediate, and large) of ions can be more precisely defined using principal component analysis (PCA) (Horrak et al. 2000). The PCA is used to investigate the domain pattern of the air ion mobility and their boundaries between different groups of ions. The principal components usually depend on the units used in the measurements and the ranges of values assumed by the original variables. Therefore, the data are standardized before PCA application. In the present case, the fractions of ion mobility spectra were treated with a nonlinear transformation by logarithmic scaling. It transforms the asymmetric frequency distributions of fractions concentration close to the normal, and the scaling did not significantly affect the classification of air ions (Horrak et al. 2000). Air ion fraction, equivalent diameter ranges for different classes of ions are given in Table 1, where N and P are the symbols of fraction for negative and positive polarity. The fractions are made according to the classes of air ions spectrum (Horrak et al. 2000; Gautam et al. 2017). The results from the PCA are shown in Sect. 4.4, where the top panels depict the negative and positive ions on non-event days (222) and the bottom panels for the event days (28). The components which have the highest percentage of contribution to the ion mobility distribution have been plotted in each panel. The total principal components that can be potentially extracted are limited to the number of variables, i.e., the number of fractions (20) in our case.

Table 1 Air ion fraction, equivalent diameter ranges for different classification of ions. N and P are the symbols of fraction for negative and positive polarity
Full size table

3 Results and Discussion

The formation of new particles in nucleation events occurs in a burst of intermediate ions (Horrak et al. 1998) and generally starts from cluster ions to large ions (Dal Maso et al. 2005; Hirsikko et al. 2005). Such events are easily noted with during the visual inspection of the diurnal variation of the number size distribution of positive and negative ions (Siingh et al. 2018). The days having such patches/shapes are called event days. In the absence of events, the spectrum appears almost uniform, and the day is termed as a non-event day. Types of new particle formation are generally categorized by the shape of the diameter spectrum in which the pattern of growth appears (Dal Maso et al. 2005). Based on these studies, Siingh et al. (2018) have recently presented class I—truncated banana shape, class II-patchy type, class III-inverted cup, and undefined (class IV) —not definite shape for event day and best example for non-event day. Randomly spectrums are rechecked so that chances of events being left out are relatively reduced. However, some deviations from these trends do exist in some cases. It is difficult to quantify the variations vary from event to event. The growth rate (GR) of ions was calculated using the mode fitting method (Hirsikko et al. 2005; Siingh et al. 2013b).

3.1 Diurnal Variation of Air Ions

For ready reference, an example of diurnal variation of the number size distributions of ions of both polarities on an event day (22 May 2012) is shown in Fig. 1. The shape of the spectrum may slightly change from season to season due to a minor shift in time and magnitude of the maxima caused by variations in solar radiation and local meteorological conditions. The third panel shows the diurnal variations of total ion concentrations of both polarities in the mobility range of 3.14–0.00133 cm2 V−1 s−1 (0.36–47.1 nm diameter). The particle growth started from the size range 4–6 nm at 0900 h. The NPF event neither showed any simultaneous burst of small ions nor any persistent growth of ions of < 4–6 nm diameter, typical features of banana-type NPF events (Siingh et al. 2018). The particle growth rate during the event was found to be 6.58 ± 1.7 nm h−1 (6.67 ± 1.98 nm h−1) for negative ions (positive ions) in the range of 5.4–47.1 nm. The truncated banana-type NPF events have different properties for different polarities. For example, the mean growth rate of positive ions ranged 7.23 ± 2.87 nm h−1 and for negative ions 8.5 ± 3.24 nm h−1 for ions in size range of 3.9–25.3 nm (fraction 12–16) (Table 1). Ions in the size range from 25.3 to 47.8 nm (fraction 17–20) (Table 1), the growth rates are 5.23 ± 2.03 nm h−1 for positive ions and 6.15 ± 3.1 nm h−1 for negative ions. Formation of new particles may start as molecular clusters at the top of the hills surrounding the institute, where solar radiation hits the before the valley. These newly formed particles along with other aerosol particles, and trace gases may be transported by katabatic winds (Kamra et al. 2015b; Gautam et al. 2017; Siingh et al. 2018). These molecular clusters formed at the top of the hill may grow to the nucleation mode particle (intermediate ions) during their transportation under nocturnal inversion along the hill-slope to the observation site (Kamra et al. 2015b). It takes about 2–3 h for the particles of 2–3 nm size to grow (by condensation and self-coagulation) into the nucleation mode size range of 9.7–15 nm (Kamra et al. 2015b). The absence of ions growth in the cluster size ranges can be inferred from the shift of a spectral peak. This result is consistent with the nanometer particle growth rates observed at clean continental sites (Weber et al. 1997; Mäkelä et al. 2000).

Fig. 1
figure 1

An event day example for the diurnal variation of the number size distributions of positive and negative ions on 22 May 2012 at Pune, India. Lower panel shows diurnal variations of total ion concentrations of both polarities in the mobility range of 3.14–0.00133 cm2 V−1 s−1 (0.36–47.1 nm diameter)

Full size image

Figure 2 shows the diurnal variation (hourly mean concentration) of positive/negative ions in different mobility range for event and non-event days. Results are presented in whisker plots for event days because NPF occurred mainly during post sunrise hours with a few events in mid of the day. Further, schematic growth of ions can be discerned up to 4 hours from the time of NPF occurrence on light large and heavy large ions which are substantially higher than annual values. The diurnal maximum distribution was observed at 0800 h and minimum around noon time at 1200–1400 h. The ion concentration gradually increased from noon time to 1900–2100 h. The upper quartile was higher during the morning hours for both the polarities, which shifted the mean above the median. The distributions were more comprehensive in the morning hours whereas the maximum occurred in the post evening hours. For non-event days, the diurnal pattern of all categories of ions was almost similar with differing amplitude throughout the year. Hence variation is shown by the mean curve. On non-event days, the concentration maximum in diurnal variation occurred at 0900 h and the minimum was observed during long period of 1300–1800 h. The mean concentrations also showed a little increase in the evening hours. The diurnal patterns of both the polarities were nearly similar but differed in their concentrations. The outliers of the positive and negative ions during morning hours on event days mostly coincided with period of NPF. The increasing concentration of ions during morning hours due to the particle formation on NPF days shifts the primary maximum of ions to earlier hour than non-event days. The big cluster ions concentrations had similar diurnal pattern as small cluster ions but had lower concentrations during the both event and non-event days. Negative ion concentrations in big clusters on event days had relatively more outliers than small cluster ions. Whisker diagram clearly shows the difference between negative ions outliers and positive ions outliers in different ranges of ions and their diurnal variations. The average diurnal variation of small ion concentration during 1993–1994 at Tahkuse Observatory, Estonia, showed the maximum at 0600–0700 h and the minimum at 1800 h for both negative and positive polarity (Horrak et al. 2000; Horrak 2001). Dhanorkar and Kamra (1993a) also reported the average diurnal variation of small ions concentration at Pune with the maximum concentration at 0500–0600 h and the minimum in the afternoon. The time of peak concentrations generally depend on the height of the measurement, surrounding terrain and evolution of boundary layer, etc. (Pal et al. 2014).

Fig. 2
figure 2

Whiskers diagram of the diurnal variation (hourly mean concentration) of different categories ion distribution during event days (28 day mean) and non-event days (222 days mean; solid line), Pune, India

Full size image

The size distribution of aerosol is strongly affected by photochemical nucleation bursts occurring under favorable meteorological conditions (for example, higher temperature/solar radiation, lower relative humidity with stable surface wind speed and direction) (Horrak et al. 1998; Siingh et al. 2013b, 2018). The diurnal variation of intermediate ions was similar to that of small ions (small cluster and big cluster ions) at this site. Positive and negative intermediate ions followed almost the same diurnal trend and their concentrations were approximately the same on both event and non-event days. However, during the morning hours (0700–0800 h) of NPF events, the ion concentration of both polarities showed a higher magnitude than that of non-event days. The NPF initiates from 4 to 6 nm diameter range (Kamra et al. 2015b; Siingh et al. 2018) and then ions gradually grow towards a higher diameter. On non-event days, the diurnal pattern of light large ions was similar to that of small and intermediate ions. However, on event days, the ion concentrations of both the polarities were higher than on non-event days from 1000 to 1600 h. During this period, the lower quartile was higher in the first 3 hours (1100–1300 h) and the upper quartile was more in the later 3 hours. The decreasing rate of ions concentration is more gradual for positive ions than negative ions. The rapid decrease of negative ions may be due to its property of faster coagulation or recombination with prevailing aerosol or opposite polarities. The latter scenario initiated from the big cluster, gradually appeared on intermediate and large ions. The positive ions outliers are higher than negative ions during event days.

Heavy large ions of both polarities had a similar trend during morning hours on non-event days. During 1200 to 1600 h, ion distribution is symmetrical irrespective of their charge, where the upper and lower quartiles were equally distributed. For negative ions, the outliers are exceptionally higher throughout the night. The event-day concentrations were slightly higher than that of non-event-days. The total ions concentration during event days is relatively higher, sometimes double than the non-event days. Another factor affecting the total concentration is the occurrence of a period of enhanced concentrations of ions and aerosols (PECIA) 2–3 h before NPF. In the PECIA periods, ions of ≤ 8 nm grow faster and attain mode diameter, which dramatically reduces the condensation sinks and may be conducive for particle formation (Kamra et al. 2015b). Though PECIA was not always associated with NPF, it frequently occurs in the early morning hours and rarely in the evening hours (Kamra et al. 2015a, b). It sustains for minutes to hour by increasing larger ions concentration to a certain extent, which is reflected in the total concentration on event days.

The distribution of different classes of air ions on event and non-event days demonstrated the differences at a diurnal scale. The significant changes between the daily mean of non-event days (222 days) and whisker diagram of event-days are due to the occurrences of new particle formation and its associated interaction with prevailing aerosol through recombination and coagulation processes (Kamra et al. 2015b). The primary maximum observed between 0800 and 0900 h is related to the breakdown of nocturnal inversion and convection of airmass from the earth surface upward to the measurement level. The changes in the mobility spectrum of small ions are due to the changes in the chemical composition and concentration of some trace gases or vapors in the air probably generated by photochemical reactions. There was no substantial difference found between large ions of negative and positive polarity. Bursts or event of intermediate ions were not observed, whereas Siingh et al. (2013b) have reported a few numbers of events during the cloudy environment over this site. The present observations of similar behavior of small, intermediate and large ions to nanometer aerosol particles support the earlier results reported by Horrak et al. (2000, 2003). The higher concentrations of the small and intermediate ions of positive polarity on event days than that of negative polarity may be due to the electrode effect manifested close to the ground on the hill–slope and the local topography of site (Porstendorfer 1994). In addition, higher positive ion may also be related to the availability of nucleating/growing vapors since certain vapors may favor the formation and growth of ions in single polarity (Siingh et al. 2018). The katabatic winds flowing down along the hill-slope surrounding the site in the morning hours may transport and keep depositing an excess positive ion and positively charged particles along with the radioactive emanations and trace gases in the valley (Kamra et al. 2015b). Dhanorkar and Kamra (1993b) and Kamsali et al. (2003) reported enhanced concentration of 222Rn and its progeny in the morning hours in coincidence with the peaks in electrical conductivity. Figure 3 clearly showed the presence of additional ionization in the morning hours. In contrast to the small and intermediate ions, no significant differences between the negative and positive light large and heavy large ions concentrations were observed (Fig. 2).

Fig. 3
figure 3

Mean diurnal variation of ion-pair production rate calculated from small ions observations on event (28 days mean) and non-event days (222 days mean) for the year 2012

Full size image

Recently, Gautam et al. (2017) analyzed NAIS observation from the same site and reported a higher percentage of relative standard deviation (RSD) for negative ions and a lower rate for positive ions. Moreover, this deviation was also different for day and night time. In general, the RSD may vary depending on various atmospheric factors such as the origin of air masses, presence of low-volatile vapors, aerosol precursor vapors, pre-existing aerosol concentrations and meteorological parameters. In addition, chemical composition of the cluster ions and solar radiation also control the nucleation and growth rates (Boy and Kulmala 2002; Bonn et al. 2009). The low condensational sink of vapors and characteristic nighttime temperature inversions and intensive mixing of boundary layer air at daytime may provide favorable conditions for nucleation (Kulmala et al. 1998; Clement et al. 2001).

3.2 Ion-Pair Production Rate

The mean diurnal variation of ion-pair production rate (Fig. 3) for non-event days shows similar pattern as the cluster ion concentration with the mean of 6.5 ion pairs cm−3 s−1. The ionization on event days (Fig. 3) gradually increased from morning hours and maximized at ~ 0800–0900 h. Thereafter sharply decreased and became almost equal to non-event days value during noon hours. From 1600 h onward, the ion-pair production rate for non-event days again started increasing trend with a secondary maximum ~ 2300 h. The timings of maxima roughly coincided with the breakdown of stability conditions of ABL. Also the timing of minimum roughly coincided with the time of maximum turbulence in the ABL during the afternoon. After the minimum, a gradual increase in ion concentration was associated with the night time steady boundary layer development. The mean ion-pair production rate on event days varied ~ 20 ion pairs cm−3 s−1. The direct connections between the ionization rate and the cluster ions formation are reflected in their diurnal variation during NPF, with some dissimilarities associated with varying atmospheric conditions and dynamical processes. Chen et al. (2016a) showed that the 222Rn concentration at near ground level does not respond immediately to mixing layer expansion or shrinkage. However, Chen et al. (2016b) reported that the cluster ion concentrations are related to the ionization process, to photochemical processes and availability of nucleating vapors influenced by solar intensity and atmospheric conditions. The large number of events observed during summer days (pre-monsoon) supports the above hypothesis. During summer, increased solar radiation enhances the photoionization, enhancing ion cluster formation at the top of the surrounding hills. Later, solar radiation hit the surface near the observatory leading to surface heating, turbulence and mixing layer development. All these may assist the increased formation of vapors which is helpful in the nucleation and growth of cluster ions.

Figure 4a showed variations of computed daily ion-pair production rate (Q) (Eq. 2) from the measurements of ion concentrations and radioactive 222Rn and its progeny. It should be noted that the breaking of nocturnal inversion layer and its influence on small ion concentrations usually take ~ 2 h (Pal et al. 2014). The same is applied for the consideration of temperature (Tmin), relative humidity (RH max) at 0600–0700 h and ions concentration at 0800 h. Horrak et al. (2000) based on multiple regression analysis reported that RH and heavy large ion concentration are the main factors that affect the natural mean mobility of small ions (Horrak 2001). Solar radiation and hence temperature are the fundamental factors that affect small ion distribution (Kulmala et al. 1998). Daily variation of Q and small cluster ion concentration at 0800 h significantly follow each other, except during the monsoon season (JJAS). The minimum temperature (Tmin) and maximum RH indicate the period at which thermal layer is stable, as the Tmin value increases towards pre-monsoon (MAM), the breaking of the nocturnal stability become relatively stronger than other seasons. This result in to the strong dissipation of the Radon emitted from the soil and Q is gradually decreased (Chen et al. 2016a). The relation became complex during the monsoon season (JJAS), when the source and sink of the small ions behave differently. The rainwater blocks 222Rn emission from the soil and larger aerosol ions are also removed from the atmosphere by wet scavenging. Hence, the small cluster ions became extremely low (< 20 cm-3) during morning hours. However, due to the capping effect, a sudden increase of Radon concentration may be observed occasionally during the initial periods of rainfall (Schumann et al. 1988; Victor et al. 2019). From post-monsoon (ON), the normal relation between temperature, Q, and small cluster ions was observed. Apart from the temperature and RH, the change in diurnal variation of wind speed also impacts the concentration of Radon, air ions/ aerosol particles and hence Q by advection. Figure 4b clearly showed that small cluster ion was well correlated with ion production rate having a correlation coefficient of 0.67 with 95% significance level. The concentration of small clusters and ion-pair production rate showed similar behavior throughout the year. However, Tmin and RH have nearly opposite trends. The ionizing radiation is primarily controlled by many atmospheric parameters such as boundary layer depth, air temperature, wind speed, solar radiation, soil characteristics, etc. In addition, the long transportation of air masses to the measurement site may bring extra Radon concentrations. Dal Maso et al. (2005) reported that low Radon ionizing capacities were found in association with NPF events which is in agreement with our 222Rn observation and statistics on NPF events at Pune (Siingh et al. 2018).

Fig. 4
figure 4

a Variation of daily ion-pair production rate (Q) derived from 222Rn and its progenies (0800 h) and temperature (Tmin.), relative humidity (RH max.) (Tmin. and RH max are the mean of 0600–0700 h of the day) and corresponding small cluster negative ions (SCN) concentration (0800 h) of the day (a); correlation coefficient between small cluster ions and ion-pair production rate (b)

Full size image

3.3 Conductivity Relation with Ion Concentration

The present air ions observations are used to compute polar conductivity using the relation σ = μ ni/p e, where μ is mobility, n is small ions concentration and e electric charge 1.6 × 10–19 C. The variation of mobility and polar conductivity with respective ion concentrations on the event and non-event days are presented in Fig. 5. Positive ion concentration on event days showed two peaks with departure from non-event days corresponding to mobility 0.003 and 0.03 cm2 V−1 s−1 whereas no many peaks could be seen for negative ions. The same was valid for conductivity. The sum of total conductivity on event days for all the mobility range is 3.313 f S m−1. In which, small ion (mobility ranges 3.16–0.5 cm2 V−1 s−1) contributed 3.076 f S m−1 and large ions (mobility ranges 0.034–0.00133 cm2 V−1 s−1) accounted only 0.112 f S m−1, that is 92.84% contribution of conductivity was by small ions and only 3.68% by large ions in the total conductivity. The remaining percentage was contributed by the intermediate ions (mobility ranges 0.5–0.034 cm2 V−1 s−1).

Fig. 5
figure 5

Variation of mobility with ion number concentrations (both polarities) along with conductivity (both polarities) for event (28 days mean) and non-event (222 days mean) days

Full size image

Similarly, for non-event days, small ions had more than 93.66% contribution and larger ions accounted only for 2.73% of total conductivity. The conductivity range observed at this site is approximately two folds lower than that of the rural atmosphere because of the more significant amount of small ion concentrations and less aerosol particles (Horrak et al. 2000). The increasing (decreasing) trend in conductivity by small ions (large ions) indicates the decreasing aerosol concentration before the NPF periods (Kamra et al. 2015b). The decrease in aerosol concentration leads to a reduction in the ion-aerosol attachment, which causes a change in the conductivity. The difference between positive and negative conductivity decreases with increasing ions size/diameter.

Earlier study by Dhanorkar and Kamra (1993b) reported that small ions contributed more than 95% on most of the days and intermediate ions contribution was little higher than large ions. Further, they also found that this ratio varied from season to season. It is to be noted that ionization production by Radon concentration highly depends on the season (Kamsali et al. 2003). A similar study from the middle latitude station suggested that the small ions contributed 96.3% (5.95 ± 2.11), intermediate ions for 2.2%, and large ions for 1.5% of total conductivity (Horrak et al. 2000). The reported measurement site was mostly masked by the wintering snowfall and its aggregation. Another observation in a pristine condition at Antarctica reported that the conductivity is only due to small ions and varied with the annual mean of 7.5 ± 1.5 f S m−1 (Siingh et al. 2013c). They further reported that the value was consistent with the atmospheric electric current and field measurements.

Variations in the ion production rate are primarily related to the atmospheric boundary layer (ABL) development, soil conditions and origin of air masses (Siingh et al. 2018; Victor et al., 2019). 222Rn concentration is strongly related to the vertical stability of the ABL (Kataoka et al. 1998). Recently, Victor et al. (2019) discussed that 222Rn concentration and ionization decreased with the increase of upward wind velocity in the ABL and moisture of the soil for different seasons. The dynamics of ABL plays crucial role in the exchange of energy, moisture, momentum and pollutants from surface to upper layers (Lee et al. 2015). Nilsson et al (2001a, b) discussed in detail the effect of air masses, synoptic weather and boundary layer evolution on aerosol formation. The generations of intermediate ions after sunrise with a delay of 2–3 h (Fig. 2) coincide with the rapid development of the boundary layer. 222Rn and its progenies may be the major source of surface ionization of air molecules, however, the lack of association between the mobility ranges are because of the modification of atmospheric air ion properties exerted by different dynamical and chemical processes during the evolution of changes in the atmosphere. The concentration of small ions was found to be dependent on aerosol concentration and hence only heavy large ion concentration due to ion-aerosol attachment (Hoppel 1985), the stability of air close to the ground, and the state of the soil (in terms of radioactive molecule concentration) (Chen et al. 2016b; Victor et al, 2019). The difference in the concentration of positive and negative ions in size range < 1.6 nm (small ions) can be explained by the two factors, (1) during morning hours, deposition of the ions and particles from the hilltop surface to the valley by katabatic winds and (2) due to the electrode effect. Positive ions and particles tend to settle down on the surface of the Earth. These result in the excess positive ions of small size diameter at our measurement site.

3.4 Principal Component and Factor Analysis

Figure 6 shows the factors of air ion mobility spectra for both polarities (positive and negative) of air ions during event (28 days mean) and non-event (222 days mean). Measurements used in the verification of the classification of ions are required to have fractions that match close to the mobility classes. The first five principal components substantially explained about 98% (99%) and 94% (96%) of total variance for positive (negative) ions on non-event and event days, respectively. The first principal component explained about 87% (86%) and 58% (65%) of total variance for positive (negative) ions on non-event and event days, respectively. The contribution of rest components and their location of peak-fraction are shown in Fig. 6. The noticeable difference is the contribution of the second component which is only 5% (5%) and 22% (18%) for positive (negative) ions on non-event and event days, respectively. The peak contributions for both types of ions are from fraction 13 (diameter 4.60–7.39 nm) and fraction 16 (diameter 13.80–21.9 nm) non-event and event days. The difference in contribution and structure of different components at various fractions on event days from non-event days arrived due to differential response of growth of ions in different categories or size range and shapes.

Fig. 6
figure 6

Factors of air ion mobility spectra for positive and negative ions for event (28 days mean) and non-event (222 days mean). The mobility and diameter boundaries of fractions are given in Table 1. Percentage of contribution of the factor to that of total variation is given in the bracket of each panel

Full size image

In general, the atmospheric nucleation initiates once air ions reached the critical cluster size of 1.5 ± 0.3 nm (Kulmala et al. 2013), from that place the growth of a cluster is energetically favored (Vehkamaki 2006). Further, the charged clusters on event days grow faster than the same size neutral particles (Kamra et al. 2015a). Kamra et al. (2015b) reported the mean growth rates of 3.9–25.3 nm (fraction 12–17) positive (negative) ions to be 49–142% (49–126%) and the growth percentage was more significant for those having size in the range 25.3–47.8 nm (fraction 17–20). Gautam et al. (2017) reported that different polarities of ions have different growth rate for other classes of air ions. Another factor affecting the variance in contribution of various components is the production rate of charged particles by the attachment of ions with aerosol particles, which are higher in size range of 3.85–7.4 nm diameter (fraction 11–13) as compared to the other sizes ranges. Additionally, ion-ion recombination rate also increases with the enhanced concentration of ions and particles. Meteorological parameters may also change the structure of ion concentration through changes in residence time, diffusion, and removal from the measurement location (Jaenicke 1984; Hoppel et al. 1990).

Fractional concentration of mobility spectrum of air ions may also be interpreted as a set of closely correlated variables (Table 1). Correlation coefficient in percent between atmospheric air ion mobility fractions during non-event days for the year 2012 for both the polarities of ions are given in Table 2. The formal correlation is attributed to (1) physical and chemical processes embracing a group of fractions (i.e., positive correlation) or acting between different groups of fractions (i.e., negative correlation) and (2) an unavoidable smoothing of a spectrum due to the finite resolution of NAIS. Fractions 1 to 12 (diameter 0.3–4.6 nm) consistently show highly significant positive correlation (> 95%). For fraction 13 and 14, the linearity drastically decreased to 44% concerning fraction 1 and from fractions 15 to 20 (diameter 10.10–47.1 nm) the average correlation is > 90%. This fractional mobility is considered to be the boundary of the above fractions (> 0.076 cm2 V−1 S−1) on the other hand, the physical/ chemical processes of intermediate ions were highly pronounced for this fraction (at F13 and F14, 4.6 nm). This finding is in good agreement with the earlier measurements made at this site with the conventional Gerdien condenser technique (Dhanorkar and Kamra 1991). They reported the maxima at about 0.076 cm2 V−1 s−1 (diameter 4.6 nm) for intermediate ions, which have the highest variability in the mobility spectrum. Usually, during their generation by photochemical nucleation, the peak particle size shifts toward the larger sizes in the spectrum (Misaki 1964; Kojima 1984). The mode diameter fraction (N13 and N14) of intermediate ions show good correlation (73% and 60%) with the fraction N16, which is the mode diameter of light large ions (diameter,13.80–21.9 nm). For positive ions, the scenario is quite different because they are likely to encounter fewer molecules in its path than negative ions. Hence, they have relatively lower growth rate than the negative ions. Fraction 1–16 had correlation > 90%. The least correlation (~ 88%) was between fraction 18 and fractions 4–5. A negative correlation was not observed on non-event days.

Table 2 Correlation coefficients (in percent) between air ion- mobility fractions for negative and positive ions during non-event days (222) for the year 2012
Full size table

Table 3 showed the correlation between mobility fractions for positive and negative ions on the event days. For negative ions, the correlation for fractions 1 to 12 gradually decreased but remained higher than 62%, whereas it decreased from 43% to 15% for fractions from 13 to 16 (4.60–21.9 nm). In addition, the same negative correlation (−15%) was also observed between the fraction 19 and fraction 16. From fraction 17, the correlation became positive and improved to a high value. Earlier studies suggested that the formation and growth of the ions in the size range of 1.7–7.4 nm (fraction 7–13) during NPF events significantly increases the ion concentration. These excess ions along with existing small ions, get attached with natural particles and convert them into larger ions, which can be seen as positive correlations between N6–N10 and N17–N18 for negative ion. The high correlation (> 93%) between fractions N13-N14 (diameter 4.60–10.10 nm) with fractions N17 and N18 (diameter, 21.9—36.4 nm) may be due to nucleation of new particle formation and subsequent growth of particles that comprise the light large ions (Horrak 2001). In the case of positive ions, the correlation varied from 99 to 18%. Fraction P13–P14 had correlation > 93% with heavy larger ion segments (fractions P17–P20). Fraction P1–P5 showed very little or negative (8% to −18%) correlation with faction P15–P16. The disassociation mostly arises by the combined effect of temperature, moisture and background aerosol scavenging. The cluster ions were typically observed in the absence of solar radiation and hence in the absence of photosynthesis when the mixing layer is thin and possibly because they are produced in the smaller mixing volume and the further growth process by the prevailing vapors may be feeble.

Table 3 Correlation coefficients (in percent) between air ion-mobility fractions for negative and positive ions during event days (28) for the year 2012
Full size table

The boundary of fractions may cause the lower positive correlation both on event and non-event days for the fractions 12–15 by different mobility ranges. Fraction 1 to 13 represents the individual NAIS channels and from fraction 14 to 20, the channels are grouped by more than 2. Positive ions on non-event days seem to have well-organized patterns to have well-organized variations so that the discrepancies were not appeared in this category. The higher correlation between the fractions of positive ions on non-event days suggested that the maximum sustainability of the positive ions in the atmosphere under recombination, aerosol attachment and coagulation process is relatively higher. In contrast, lower mobilities of negative ions may be due to lower lifetime and a higher rate of extension with prevailing higher size aerosol particles in the atmosphere.

During NPF events, the growth of the new particle mostly ceased when it reaches to the intermediate ion size (~ 7.4 nm, fraction 13) and their dissipation by recombination/combination process continues. This was commonly observed, which is also reported by Kamra et al. (2015a, b). Moreover, the coagulation of ions with aerosol particles is faster and more potent than for the same size neutral particles, since the larger electrostatic force contributes more to the attachment process. An increase in coagulation rate contributes for higher growth rate of particles during NPF events. The event would eventually decrease with increasing particle size (Yu and Turco 2001). In this study, we covered the entire year to maintain the significance of the statistical relation (Horrak 2001; Horrak et al. 2000), and hence the data consist of seasonal variability of air ions, including a comprehensive monsoon month (JJAS). Horrak (2001) explained the complexity in determining the boundaries between and within the separated classes on air ion and aerosol ion, especially in intermediate ions. They further stated that the levels of air ions have seasonal variability to a certain extent; hence mean value may be considered (Tammet 1995; Dhanorkar and Kamra 1991).

4 Conclusion

To understand the mobility, diurnal characteristics of atmospheric air ions of both polarities and their dependence on the meteorological parameters on event and non-event days, atmospheric ions and concentration of 222Rn measurements carried out at Pune was analyzed. Annual mean diurnal variations of atmospheric air ions derived from 222 non-event days of observation show the bimodal distribution with the primary maximum in the morning at  ~ 0800 h and the secondary maximum in the evening at  ~ 1800 h, with the minimum concentration during 1200–1400 h local time. This diurnal cycle variation is observed for the air ions concentrations of the entire measured range of diameter 0.36–47.1 nm. Temperature and humidity affect ion distribution and boundary layer evolution and hence also Radon emission which is an important factor for ion production and its diurnal variation. Based on the present study following results are obtained:

  1. 1.

    From the diurnal variation, a link between the ionization rate and the cluster ions formation during new particle formation events was observed with small dissimilarities which may be attributed to the varying atmospheric conditions and dynamical processes. Further, it is also observed that larger size ions decreased before the onset of the particle formation, which increased conductivity. The boundary layer evolution and topography of the measurement site could explain maxima in the morning hours, minima in the afternoon and a secondary maximum of ion concentration in the evening hours.

  2. 2.

    A gradual and systematic growth of air ions is observed on event days, especially in the intermediate ions, which have a different diurnal pattern from the non-event days. It is interesting to note that air ions with large concentrations in different ranges are observed throughout the year, but most of the NPFs were observed during pre-monsoon months (MAM). This suggests that solar radiation along with temperature and humidity play significant role in particle formation. Further, the concentration differences on event and non-event days are more prominent in larger ions segment (> 7.4 nm).

  3. 3.

    To investigate the dominant pattern of the air ion mobility spectrum, the principal component analysis (PCA) was used and showed that the first five components to a first approximation could represent about 98% of all the measured information in the mobility boundaries on non-event and event (NPF) days The first component explained ~ 86% (65%) of total variance on non-event (event) days and it is closely correlated with the respective ion classes. Linear regression analysis between the fractions demonstrates the boundaries of the classes of air ions. The mode diameter of intermediate ions at 4.6 nm (F13) prominently separate small ions from large ion's fractions.

Improved understanding of the evolution of atmospheric air ions on a diurnal scale helps to determine the role of boundary layer mixing and topography on NPF. However, along with air ions measurements, simultaneous aerosols, trace gases such as SO2, NH3, Iodine, VOC, ionization and metrological parameters made at different geographical locations over long periods is widely required to understand the possible generation mechanisms in different seasons completely.