利用氣膠模式探討環境腔α-蒎烯臭氧化反應的粒子核化

Abstract

由α-蒎烯臭氧化反應衍生出的新生粒子事件即使在實驗室及實地的研究中可被觀測到，但是目前大多的空氣品質模式卻缺乏對這類純有機物核化的描述，而其很有可能是一個重要的新生粒子來源。於是本研究利用氣膠模式(MOSAIC)及古典核化理論來模擬哈佛環境腔(HEC)以及流動管反應器(FTR)下的α-蒎烯臭氧化反應的新生粒子事件，當中分別使用175 bin及145bin的粒徑分佈來進行模擬。在MOSAIC中為了簡化臭氧化反應使用了雙產物模式，兩種產物分別為低揮發性有機物(LVOC)及半揮發性有機物(SVOC)，兩者皆可以參與凝結過程但其中只有LVOC會主導核化過程。在靈敏度測試當中，不僅核化曲線可以影響新生粒子事件的強度，凝結相關參數也可以間接影響之，特別是粒子相擴散度對其可以有非線性的變化，而此表現可能來自其對粒子成長的限制作用。在模式中給予α-蒎烯濃度一波動函數可以重現HEC中離散的新生粒子事件，也因此可推論這樣的浮動來自反應腔體內物種濃度的空間不均勻性。根據粒徑分佈的誤差分析後可以發現，在古典核化速率的表面張力為23.0 dyne cm-1、LVOC及SVOC的調節係數分別為0.1及0.3和10-12 cm2 s-1的粒子相擴散度可以適當模擬HEC實驗的粒徑分佈，但在眾數粒徑附近會有低估粒子數的狀況。另外，同樣的模式也被用來模擬FTR的反應，然而即使考慮到各種不確定的因素卻無法模擬出其粒徑分佈，推測此結果應該是兩個實驗反應的限量試劑及時間不同而使得生成物產率比值不同所致。

The new particle formation (NPF) from α-pinene ozonolysis can be observed in both field and laboratory studies. However, the current air quality models lack this pure-organic NPF, which might be an essential source of new particles. Therefore, the NPF of α-pinene ozonolysis in a continuously mixed flow reactor (Harvard Environmental Chamber, HEC) and a flow tube reactor were simulated respectively using 175-bin and 145-bin aerosol model, Model for Simulating Aerosol Interactions and Chemistry (MOSAIC) incorporated with classical nucleation theory (CNT). In this study, the α-pinene ozonolysis was expressed using a two-products model with low- and semi-volatile organic compounds (LVOC and SVOC) for simplification. The nucleation process was assumed to be dominated by LVOC while the condensation processes were contributed by both LVOC and SVOC. The sensitivity tests showed that not only the CNT nucleation curve but also the parameters of the condensation process can alter the strength of NPF. Especially for bulk diffusivity, the nonlinear response of NPF to that is likely due to the limited particle growth. Moreover, the spatial inhomogeneity in HEC, which took account for the discrete NPF was illustrated by the simulation with fluctuated α-pinene concentration. Based on the error analysis, the model with CNT surface tension of 23.0 dyne cm-1, accommodation coefficients of 0.1 and 0.3 for LVOC and SVOC, and bulk diffusivity of 10-12 cm2 s-1 gave a good performance in simulating the HEC experiments. However, it still underestimated the number density of particles with around mode size. The same model was also introduced to simulate the FTR experiments, and yet, failed to interpret the NPF regardless of considering the uncertainty. The inconsistency might result from the various LVOC to SVOC yield ratio, which carried out from different reactant-limiting ozonolysis between HEC and FTR experiments and reaction time.