Synergistic HNO3-H2SO4-NH3 upper tropospheric particle formation

Abstract

New particle formation in the upper free troposphere is a major global source of cloud condensation nuclei (CCN)^1-4. However, the precursor vapours that drive the process are not well understood. With experiments performed under upper tropospheric conditions in the CERN CLOUD chamber, we show that nitric acid, sulfuric acid and ammonia form particles synergistically, at rates that are orders of magnitude faster than those from any two of the three components. The importance of this mechanism depends on the availability of ammonia, which was previously thought to be efficiently scavenged by cloud droplets during convection. However, surprisingly high concentrations of ammonia and ammonium nitrate have recently been observed in the upper troposphere over the Asian monsoon region^5,6. Once particles have formed, co-condensation of ammonia and abundant nitric acid alone is sufficient to drive rapid growth to CCN sizes with only trace sulfate. Moreover, our measurements show that these CCN are also highly efficient ice nucleating particles-comparable to desert dust. Our model simulations confirm that ammonia is efficiently convected aloft during the Asian monsoon, driving rapid, multi-acid HNO₃-H₂SO₄-NH₃ nucleation in the upper troposphere and producing ice nucleating particles that spread across the mid-latitude Northern Hemisphere.

Fig. 1. Example experiment showing nitric acid enhancement of H₂SO₄–NH₃ particle formation.

a, Particle number concentrations versus time at mobility diameters >1.7 nm (magenta) and >2.5 nm (green). The solid magenta trace is measured by a PSM_1.7 and the solid green trace is measured by a CPC_2.5. The fixed experimental conditions are about 6.5 × 10⁸ cm⁻³ NH₃, 223 K and 25% relative humidity. A microphysical model reproduces the main features of the observed particle formation (dashed lines; see text for details). b, Particle formation rate versus time at 1.7 nm (J_1.7), measured by a PSM. c, Particle size distribution versus time, measured by an SMPS. d, Gas-phase nitric acid and sulfuric acid versus time, measured by an I⁻ CIMS and a NO₃⁻ CIMS, respectively. Sulfuric acid through SO₂ oxidation started to appear soon after switching on the UV lights at time = 0 min, building up to a steady state of 2.3 × 10⁶ cm⁻³ after a wall-loss-rate timescale of around 10 min. The subsequent H₂SO₄–NH₃ nucleation led to a relatively slow formation rate of 1.7-nm particles. The particles did not grow above 2.5 nm because of their slow growth rate and corresponding low survival probability against wall loss. Following injection of 2.0 × 10⁹ cm⁻³ nitric acid into the chamber after 115 min, while leaving the production rate of sulfuric acid and the injection rate of ammonia unchanged, we observed a sharp increase in particle formation rate (panel b), together with rapid particle growth of 40 nm h⁻¹ (panel c). The overall systematic scale uncertainties of ±30% on particle formation rate, −33%/+50% on sulfuric acid concentration and ±25% on nitric acid concentration are not shown. Source data

Fig. 2. Particle formation rates at 1.7 nm (J_1.7) versus ammonia concentration at 223 K and 25% relative humidity.

The chemical systems are HNO₃–NH₃ (black), H₂SO₄–NH₃ (red) and HNO₃–H₂SO₄–NH₃ (blue). The black diamond shows the CLOUD measurement of HNO₃–NH₃ nucleation at 1.5 × 10⁹ cm⁻³ HNO₃, 6.5 × 10⁸ cm⁻³ NH₃ and with H₂SO₄ below the detection limit of 5 × 10⁴ cm⁻³. The red solid curve is J_1.7 versus ammonia concentration at 4.0 × 10⁶ cm⁻³ sulfuric acid from a H₂SO₄–NH₃ nucleation parameterization on the basis of previous CLOUD measurements^,. The blue circles show the CLOUD measurements of HNO₃–H₂SO₄–NH₃ nucleation at 4.0 × 10⁶ cm⁻³ H₂SO₄, 1.5 × 10⁹ cm⁻³ HNO₃ and (1.6–6.5) × 10⁸ cm⁻³ NH₃. The data are fitted by a power law, J_1.7 = k[NH₃]^3.7 (blue dashed curve). The vertical grey dotted line separates ammonia concentrations measured in different regions in the upper troposphere; the region to the right indicates the Asian monsoon conditions. The horizontal grey solid lines show J_1.7 upper limits for ion-induced nucleation resulting from the GCR ionization rate of around 2 ion pairs cm⁻³ s⁻¹ at ground level and 35 ion pairs cm⁻³ s⁻¹ in the upper troposphere. Among the three nucleation mechanisms, H₂SO₄–NH₃ nucleation dominates in regions with low ammonia (below around 1.0 × 10⁸ cm⁻³, or 12 pptv), whereas HNO₃–H₂SO₄–NH₃ nucleation dominates at higher ammonia levels characteristic of the Asian monsoon upper troposphere. The bars indicate 30% estimated total error on the particle formation rates. The overall systematic scale uncertainties are −33%/+50% for sulfuric acid and ±25% for nitric acid concentrations. Source data

Fig. 3. Molecular composition of negatively charged clusters during H₂SO₄–NH₃ and HNO₃–H₂SO₄–NH₃ nucleation events at 223 K and 25% relative humidity.

Mass defect (difference from integer mass) versus mass/charge (m/z) of negatively charged clusters measured with an APi-TOF mass spectrometer for 1.7 × 10⁶ cm⁻³ sulfuric acid and 6.5 × 10⁸ cm⁻³ ammonia (a) and 2.0 × 10⁷ cm⁻³ sulfuric acid, 3.2 × 10⁹ cm⁻³ nitric acid and 7.9 × 10⁹ cm⁻³ ammonia (b). The symbol colours indicate the molecular composition as shown. The symbol area is proportional to the logarithm of signal rate (counts per second). The labels (m:n) near the symbols indicate the number of sulfuric acid (H₂SO₄)_m and ammonia (NH₃)_n molecules in the clusters, including both neutral and charged species. The grey dashed lines follow clusters that contain pure H₂SO₄ molecules with an HSO₄⁻ ion (or SO₄ instead of H₂SO₄ and/or SO₄⁻ instead of HSO₄⁻ for pure H₂SO₄ clusters falling below this line in b). The grey solid lines follow the 1:1 H₂SO₄–NH₃ addition starting at (H₂SO₄)₄–(NH₃)₀. Nearly all clusters in panel a lie above this line, whereas nearly all clusters in panel b fall below it. Most clusters containing HNO₃ lack NH₃ by the time they are measured (they fall near the (m:0) grey dashed line), but the marked difference between a and b indicates that the nucleating clusters had distinctly different compositions, probably including relatively weakly bound HNO₃–NH₃ pairs in b. It is probable that nucleating clusters in the CLOUD chamber at 223 K contain HNO₃–H₂SO₄–NH₃ with a roughly 1:1 acid–base ratio. However, during the transmission from the chamber to the warm APi-TOF mass spectrometer at 293 K, the clusters lose HNO₃ and NH₃, leaving a less volatile core of H₂SO₄ with depleted NH₃. The evaporation of a single NH₃ or HNO₃ molecule from a cluster displaces it on the mass defect plot by a vector distance indicated by the black arrows in b. Source data

Fig. 4. Ice nucleation properties and modelled regional contribution of upper tropospheric particles formed from HNO₃–H₂SO₄–NH₃ nucleation.

a, Active surface site density versus ice saturation ratio, measured by the mINKA instrument at CLOUD, at 233 K and 25% relative humidity. Pure ammonium nitrate particles (purple points) show homogeneous freezing. However, addition of only small amounts of sulfate creates highly ice-nucleation-active particles. At around 1.7% sulfate fraction (red points), the ice nucleating efficiency is comparable with desert dust particles. b, Simulation of particle formation in a global model (EMAC) with efficient vertical transport of ammonia into the upper troposphere during the Asian monsoon. Including multi-acid HNO₃–H₂SO₄–NH₃ nucleation (on the basis of the blue dashed curve in Fig. 2) enhances particle number concentrations (nucleation mode) over the Asian monsoon region by a factor of 3–5 compared with the same model with only H₂SO₄–NH₃ nucleation (from Dunne et al., similar to the red solid curve in Fig. 2). Source data

Extended Data Fig. 1. Enhancement of HNO₃–NH₃ particle formation by sulfuric acid.

a, Particle number concentrations versus time at mobility diameters >1.7 nm (magenta) and >2.5 nm (green). The solid magenta trace is measured by a PSM_1.7 and the solid green trace is measured by a CPC_2.5. The fixed experimental conditions are about 6.5 × 10⁸ cm⁻³ NH₃, 223 K and 25% relative humidity. b, Particle formation rate versus time at 1.7 nm (J_1.7), measured by a PSM. c, Particle size distribution versus time, measured by an SMPS. d, Gas-phase nitric acid and sulfuric acid versus time, measured by an I⁻ CIMS and a NO₃⁻ CIMS, respectively. We started the experiment by oxidizing NO₂ to produce 1.6 × 10⁹ cm⁻³ HNO₃ in the presence of about 6.5 × 10⁸ cm⁻³ ammonia. At time = 0 min, we turned off the high-voltage clearing field to allow the ion concentration to build up to a steady state between GCR production and wall deposition. The presence of ions (GCR condition) induces slow HNO₃–NH₃ nucleation, followed by relatively fast particle growth by nitric acid and ammonia condensation. We thus observe formation of both 1.7-nm and 2.5-nm particles by about one order of magnitude in about 3.5 h, with a slower approach to steady state because of the longer wall deposition time constant for the larger particles. Then, we increased H₂SO₄ in the chamber from 0 to 4.9 × 10⁶ cm⁻³ by oxidizing progressively more injected SO₂ after 211 min, with a fixed production rate of nitric acid and injection rate of ammonia. Subsequently, particle concentrations increase by three orders of magnitude within 30 min. The overall systematic scale uncertainties of ±30% on particle formation rate, −33%/+50% on sulfuric acid concentration and ±25% on nitric acid concentration are not shown. Source data

Extended Data Fig. 2. Enhancement of H₂SO₄–HNO₃ nucleation by ammonia.

a, Particle number concentrations versus time at mobility diameters >1.7 nm (magenta) and >2.5 nm (green). The solid magenta trace is measured by a PSM_1.7 and the solid green trace is measured by a CPC_2.5. The fixed experimental conditions are 223 K and 25% relative humidity. b, Particle formation rate versus time at 1.7 nm (J_1.7), measured by a PSM. c, Particle size distribution versus time, measured by an SMPS. d, Gas-phase nitric acid and sulfuric acid versus time, measured by an I⁻ CIMS and a NO₃⁻ CIMS, respectively; gas-phase ammonia versus time, calculated with a first-order wall-loss rate. Before the experiment, we cleaned the chamber by rinsing the walls with ultra-pure water, followed by heating to 373 K and flushing at a high rate with humidified synthetic air for 48 h. We started with an almost perfectly clean chamber and only HNO₃, SO₂ and O₃ vapours present at constant levels. Sulfuric acid starts to appear by means of SO₂ oxidation soon after switching on the UV lights at time = 0 min, building up to a steady state of 5.0 × 10⁶ cm⁻³ with the wall-loss timescale of about 10 min. Subsequently, we observe slow formation of 1.7-nm particles, yet they do not reach 2.5 nm during the course of a 2-h period with small growth rates and low survival probability. Then, owing to the injection of ammonia from 0 to around 6.5 × 10⁸ cm⁻³ into the chamber after 80 min, a sharp increase in the rate of particle formation is observed with a fixed production rate of sulfuric acid and injection rate of nitric acid. The sulfuric acid concentration decreases slightly afterwards, owing to accumulated condensation sink from fast particle growth. The overall systematic scale uncertainties of ±30% on particle formation rate, −33%/50% on sulfuric acid concentration and ±25% on nitric acid concentration are not shown. Source data

Extended Data Fig. 3. Particle formation rates at 1.7 nm (J_1.7) versus ammonia concentration at 223 K and 25% relative humidity.

Circles are the CLOUD measurements (the same as those in Fig. 2). The curve represents the model simulations on the basis of known thermodynamics and microphysics, including Kelvin effects, for nucleating clusters. Source data

Extended Data Fig. 4. Measurement of the ice nucleation ability of HNO₃–H₂SO₄–NH₃ particles versus sulfate-to-nitrate ratio.

a, Particle size distribution versus time during the experiment, measured by an SMPS. b, Gas-phase sulfuric acid versus time, measured by a nitrate CIMS. c, Particle-phase chemical composition versus time, measured by an AMS. d, Fraction of INP at the nominal temperature of 215 K. The horizontal black dashes indicate the ice fraction threshold, f_ice = 10⁻³. The coloured circles correspond to the sulfate-to-nitrate ratios shown in Fig. 4a. Source data

Extended Data Fig. 5. Parameterization of the HNO₃–H₂SO₄–NH₃ particle formation rate.

a–c Particle formation rate (J_1.7) as a function of H₂SO₄, HNO₃ and NH₃ vapour concentrations, respectively, at 223 K and 25% relative humidity. The red triangles, blue circles and yellow squares represent experiments while varying only the concentration of H₂SO₄ (Extended Data Fig. 1), HNO₃ (Fig. 1) and NH₃ (Extended Data Fig. 2), respectively. The H₂SO₄ concentration was varied between 4.6 × 10⁵ and 2.9 × 10⁶ cm⁻³, HNO₃ between 2.3 × 10⁸ and 1.7 × 10⁹ cm⁻³ and NH₃ between 1.8 × 10⁸ and 5.1 × 10⁸ cm⁻³. d, The multi-acid–ammonia parameterization (black line) on the basis of equation (6) with k = 3.4 × 10⁻⁷¹ s⁻¹ cm²⁴. The grey dashed horizontal line shows a maximum of about 2 cm⁻³ s⁻¹ ion-induced nucleation in the CLOUD chamber under GCR conditions, limited by the ion-pair production rate from GCR plus beam-background muons. The bars indicate 30% estimated total error on the particle formation rates, although the overall systematic scale uncertainties of −33%/+50% on sulfuric acid concentration and ±25% on nitric acid concentration are not shown. e, Temperature dependence of J_1.7 for HNO₃–H₂SO₄–NH₃ nucleation (blue curve) on the basis of equation (9) with k = 2.9 × 10⁻⁹⁸ e^14,000/T s⁻¹ cm²⁴. Source data

Extended Data Fig. 6. Modelled contribution of HNO₃–H₂SO₄–NH₃ nucleation to upper tropospheric particles.

Number concentrations of multi-acid new particles (nucleation mode) at 250-hPa altitude simulated in a global model (EMAC) with efficient vertical transport of ammonia. The particle formation rate is on the basis of the blue dashed curve in Fig. 2 and parameterization shown in Extended Data Fig. 5. The extra particle number concentrations are shown, that is, relative to the same model without multi-acid nucleation. High annually averaged particle numbers are expected in the monsoon region (grey rectangle) and adjacent regions.

Extended Data Fig. 7. Modelled annual mean ammonia mixing ratios at 250 hPa (11 km, about 223 K).

a, The EMAC global model simulations are higher than the MIPAS satellite observations, although consistent with aircraft measurements^,. b, The TOMCAT global model predicts much less ammonia (<1 pptv) in the upper troposphere.

Extended Data Fig. 8. Modelled transport of ammonia to the upper troposphere in deep convective clouds.

a, Trajectories of the simulated convective cloud event (grey) and a selected parcel representing a buoyant parcel reaching the upper troposphere (black). b, The simulated evolution of parcel A altitude (green dashed trace) and the total mass concentration and phase of the cloud hydrometeors (red and blue curves). c–e Sensitivity of the predicted ammonia concentrations within parcel A to cloud water pH, total water amount and retention coefficient (by ice particles) as compared with the base-case simulation (blue trace in all figures). Source data