Abstract Ultrafiltration is often one of the most critical steps in the production of high-purity water, chemicals and biopharmaceutical products. Pore size characterization in ultrafiltration membranes is an important yet challenging task. This work presents a long sought-after solution to measure ultrafiltration membrane pore size in a fast, sensitive and reliable way. Ultrafiltration membranes made of different polymeric materials were challenged with a mixture of 20, 30, 40 and 60 nm gold particles, at a concentration of approximately 7 × 107 particles/mL for each nominal size or 2, 8, 19, 66 μg/L gold, respectively, to determine the filtration performance over a wide size range. The low particle loading in the feed suspension resembles the environment where these membranes are practically used, for example, in ultrapure water. In the retention test both sizes and counts of gold nanoparticles (GNPs) were quantitatively measured in feed and permeate suspensions by single particle inductively coupled plasma mass spectrometry (spICP-MS). The GNP retention by membrane adsorption was minimized by adding anionic surfactant sodium dodecyl sulfate (SDS) in the feed suspension. As a result, the cut-off size of fully retained particles was in good agreement with the membranes' functional pore sizes or bubble point sizes measured by gas-liquid displacement porosimetry. The retention test showed that two tightest membranes out of the five tested ones reached full retention of particles between 40 and 60 nm in diameter. This study revealed significant discrepancies between the functional pore sizes and the manufacturers’ removal ratings. It was demonstrated that spICP-MS quantitatively and reliably measures 20 nm GNPs in permeate suspensions at concentrations as low as 50 particles/mL, equivalent to 0.003 ng/L. This is over three orders of magnitude lower than GNP limit of quantification (LOQ) by traditional ICP-MS analysis, over five orders of magnitude lower than the LOQ by liquid nanoparticle sizer (LNS), or over eight orders of magnitude lower than the LOQ of dextran by refractive index measurement. Unlike light-based particle techniques, spICP-MS is highly selective and element specific for gold, so dissolved gas or other particulates like degradants of polymeric membranes do not interfere with the analysis. The filtration part of the retention test was more straightforward and faster since no pre-test rinsing was required. The size-resolved particle retention data for every 1 nm size bin made it practical to probe an unknown sample over a wide range of pore sizes.

中文翻译：

---

使用单颗粒检测进行金纳米颗粒混合物保留测试：一种快速灵敏的超滤膜功能孔径探针

摘要 超滤通常是高纯水、化学品和生物制药产品生产中最关键的步骤之一。超滤膜的孔径表征是一项重要但具有挑战性的任务。这项工作提出了一种长期寻求的解决方案，以快速、灵敏和可靠的方式测量超滤膜孔径。用 20、30、40 和 60 nm 金颗粒的混合物对由不同聚合物材料制成的超滤膜进行挑战，每种标称尺寸的浓度约为 7 × 107 颗粒/mL 或 2、8、19、66 μg/L金分别用于确定广泛尺寸范围内的过滤性能。进料悬浮液中的低颗粒负载类似于这些膜实际使用的环境，例如，在超纯水中。在保留测试中，通过单粒子电感耦合等离子体质谱 (spICP-MS) 定量测量进料和渗透悬浮液中金纳米粒子 (GNP) 的大小和数量。通过在进料悬浮液中添加阴离子表面活性剂十二烷基硫酸钠 (SDS)，膜吸附的 GNP 保留被最小化。结果，完全保留颗粒的截断尺寸与通过气液置换孔隙率测定法测量的膜的功能孔径或泡点尺寸非常一致。保留测试表明，五个测试膜中的两个最紧密的膜完全保留了直径在 40 到 60 nm 之间的颗粒。该研究揭示了功能孔径与制造商的去除率之间存在显着差异。结果表明，spICP-MS 可以定量且可靠地测量渗透液悬浮液中的 20 nm GNP，浓度低至 50 个颗粒/mL，相当于 0.003 ng/L。这比传统 ICP-MS 分析的 GNP 定量限 (LOQ) 低三个数量级以上，比液体纳米粒度仪 (LNS) 的 LOQ 低五个数量级以上，或比 LOQ 低八个数量级以上用折射率测量法测定葡聚糖。与基于光的颗粒技术不同，spICP-MS 对金具有高度选择性和元素特异性，因此溶解气体或其他颗粒（如聚合物膜的降解物）不会干扰分析。保留测试的过滤部分更加直接和快速，因为不需要预测试冲洗。