A modified spectrophotometric method for selective determination of trace urea: application in the production process of ultrapure water

2B3T

mixed bed resins and absorbing tower

ACF

activated carbon filter

AR

analytical reagent

CV

coefficient of variation

DAM

diacetylmonoxime

LOD

limit of detection

LOQ

limit of quantitation

MB

mixed bed resin

MDG

membrane degassing

MMF

multimedia filter

POU

point of use

RO

reverse osmosis

TOC

total organic carbon

UF

ultrafiltration

UPW

ultrapure water

UV

ultraviolet

In recent years, the semiconductor manufacturing industry has developed rapidly. Ultrapure water (UPW) is an important auxiliary material that is used in chip production. Its consumption is tremendous, and 32 kg of water is used to produce a chip weighing 2 g. Therefore, the use of reclaimed water instead of tap water to produce UPW has received widespread attention. There have been many reports on the use of reclaimed water for landscape reuse, groundwater recharge, etc. (Zhu & Dou 2018), but few for UPW production because of the stringent requirements for the quality of UPW. There are strict regulations on the concentration of total organic carbon (TOC), metal ions, anions, and other substances in the UPW.

However, this industry has stringent requirements for the quality of UPW; there are strict regulations on the concentration of TOC, metal ions, anions, and other substances in the UPW. For example, in some microelectronic industries (line width of chip 14–28 nm), the concentration of TOC in the UPW is required to be less than 0.5 μg/L, which puts forward stricter requirements for the quality of raw water in the UPW production process (Melnik & Krysenko 2019). Therefore, it is important to guarantee the quality of UPW from different water sources.

Owing to the decomposition of nitrogen-containing organics such as proteins in the wastewater, the urea concentration in reclaimed water is commonly higher than that in tap water (Choi & Chung 2019). This is one fundamental reason for the excessive TOC in the UPW treated by the conventional UPW production process with reclaimed water. It is necessary to determine the distribution of urea in raw water and the removal regularity during the UPW production process.

Currently, there are few suitable methods for determining trace urea in UPW. Urea determination methods can be divided into direct and indirect methods. Direct colorimetric determination methods use chemical reagents that react with urea to generate colored products. Diacetylmonoxime (DAM)-antipyrine spectrophotometry (GB/T18204.2-2014) (NIEHS 2014) is a classic direct method. The effective determination range of urea by this method is 60–600 μg/L. It cannot accurately determine urea concentration below 60 μg/L because the measured absorbance is close to the detection limit (Long et al. 2019). This method can hardly achieve the urea determination requirement of the UPW production process in the semiconductor manufacturing industry. In addition, direct methods include the fluorescence method and high-performance liquid chromatography, which have limitations such as complicated operations and erratic reproducibility depending on the state of the instruments. Indirect methods convert urea into nitrate, ammonia, or nitrogen molecules through a series of chemical reactions and indirectly determine the urea concentration through conventional analytical techniques (Revilla et al. 2005; Han et al. 2019). The indirect methods usually require a shorter reaction time and are more convenient to operate, but they are easily affected by pH, temperature, and salt concentration during the reaction, resulting in significant detection errors (Gonzalez-Rodriguez et al. 2002; Alizadeh et al. 2017; Zhang et al. 2019). Therefore, the demand for accurate and stable determination of trace urea is evident in the production process of UPW. To the best of our knowledge, the determination of trace urea using a spectrophotometric method has scarcely been carried out.

In this study, a modified spectrophotometric method was first developed for the determination of trace urea in UPW. The calibration graph of the method was linear in the range of 0.8–100 μg/L, and the analytical parameters for assessment, such as accuracy and reproducibility, were evaluated. The concentration trend of urea in tap water, municipal reclaimed water, and industrial reclaimed water was explored using the method proposed in this study. The degradation of urea and TOC during the UPW production process was revealed. Thereby, feasible measures were put forward to guarantee the quality of UPW.

The chemicals used in this study were urea (purity 99.5%, Aladdin, USA), DAM (analytical grade (AR), Sinopharm, China), acetic acid (AR, Sinopharm, China), antipyrine (CP, Sinopharm, China), sulfuric acid (AR, Sinopharm, China), phosphoric acid (AR, Sinopharm, China), and ferric chloride (AR, Sinopharm, China). Distilled water was purchased from Watson (China). A standard urea solution (100 μg/mL, AOKE, China) was diluted to different concentrations and used to evaluate the method of determination.

A water bath (HH-8, Guohua, China) was used to heat the chromogenic solution. The urea concentration in the solution was analyzed using a dual-beam UV spectrophotometer (TU-1900, Puxi, China). The amount of TOC in the water was measured using a TOC analyzer (Sievers 500, General Electric, USA).

DAM solution: DAM was weighed, dissolved in acetic acid, diluted to 100 mL with 10% acetic acid, and stored in a brown bottle at 25 °C. Acid solution: Antipyrine and ferric chloride were weighed and dissolved in 80 mL of sulfuric acid (1 + 1). Moreover, 20 mL of mixed acid (sulfuric acid and phosphoric acid) was added and stored in a brown bottle. Standard urea solution (0.1 g): 0.1 g of urea powder was weighed and diluted with distilled water in a 1 L volumetric flask. 100 μg/L (0.1 μg/mL) urea solution: 1 mL of standard urea solution was taken and diluted with distilled water into a 1 L volumetric flask. 1,000 μg/L (1 μg/mL) urea solution: 1 mL of standard urea solution was diluted with distilled water in a 100 mL volumetric flask.

Individually, 0, 1.5, 2, 2.5, and 5 mL of 100 μg/L urea solution and 1, 1.5, 2, and 2.5 mL of 1,000 μg/L urea solution were added to nine 50 mL brown colorimetric tubes. Standard concentrations of 0, 6, 8, 10, 20, 40, 60, 80, and 100 μg/L urea solution were prepared. 2 mL of DAM solution and 2 mL of acid solution were added and diluted to 25 mL with distilled water. Brown colorimetric tubes were heated in a boiling water bath for 50 min and cooled in running tap water for 2 min. Pure water was used as a control, and the absorbance of each tube was immediately measured at 460 nm using a 5 cm cuvette.

The absorbance was contrasted with the urea concentration to make a calibration graph.

The water samples were reclaimed water and tap water used as raw water for the UPW production process in the semiconductor manufacturing industry, as well as the water body in the UPW production process. The urea concentration in the water samples ranged from 0.8 to 100 μg/L.

The DAM and acid solutions were added to a brown 50 mL colorimetric tube and diluted to 25 mL with the water sample. The tube was heated in a boiling water bath for 50 min, and then cooled in running tap water for 2 min. The distilled water was used as a control, and the absorbance of each tube was immediately measured at 460 nm using a 5 cm cuvette.

The corresponding urea concentration was obtained according to the calibration graph.

A dual-beam UV-visible spectrophotometer was used to analyze the urea concentration. The wavelength range was 190–900 nm. When the emitted light passed through the measured substance solution, the absorbance of light of the substance varied with the wavelength of the light. In this study, through parameter optimization, it was determined that the product of this method had maximum absorbance at a wavelength of 460 nm. The absorbance of the sample solution at a certain concentration was measured at the maximum absorption wavelength and compared with that of the prepared standard sample to determine the concentration of the sample solution.

Urea and diacetyl were condensed to form orange-red diazine derivatives by heating under acidic conditions. DAM is generally used as diacetyl is unstable at room temperature. Diacetyl was formed from DAM under strongly acidic conditions. Condensation of diacetyl and urea generated orange-red diazine derivatives. Both antipyrine and metal ions added in the system promote the stability of the product. The orange-red color of the product was reflected in the absorbance detected by the spectrophotometer, which corresponded to different concentrations of urea. In addition, the reaction product was unstable under light. Therefore, the reactions and measurements were performed in brown test tubes.

The variables affecting the urea concentration determination method were studied under various conditions. The optimal values are listed in Table 1. Under the optimal conditions, the absorbance of the reaction system exhibited the highest sensitivity.

When DAM reacts to form diacetyl, hydroxylamine is produced, which inhibits the reaction between urea and diacetyl. Previous studies have reported that metal ions can eliminate the interference of hydroxylamine in the reaction, thereby increasing the stability of the system (Francis et al. 2002). However, the comparison between the effects of different metal ions and the appropriate dosing concentrations still requires further exploration.

It was found that doping with Fe³⁺ and Mn²⁺ increased the stability of the reaction system, and the obtained results were consistent with the conclusions of Beale et al. (Beale & Croft 1961). The rate of absorbance loss with Fe³⁺ (1.45 mM) was much lower than that with Fe³⁺ (0.36 mM) or Mn²⁺ (0.36, 1.45, and 2.55 mM), and close to that of the reaction system with Fe³⁺ (2.55 mM). This study recommended doping with Fe³⁺ (1.45 mM) to enhance the absorbance stability of the reaction system.

In spectrophotometry, lower blank values can reduce the degree of data dispersion and improve the reproducibility and accuracy of results (de Oliveira e Silva et al. 2018). In the method for the determination of trace urea, it is critical to control the blank values of the standard samples to lower concentrations as possible. This results stringent requirement for the concentration of urea in the solvent used to prepare the standard samples.

Distilled water and deionized water were separately used as solvents to prepare standard urea solutions and establish calibration graphs in this study. The correlation coefficients of distilled water (r² = 0.9991) and deionized water (r² = 0.9988) were similar. To clarify the relatively lower urea concentration of the solvents, distilled water and deionized water were alternated as blank samples for each other, and the same pretreatment and determination were performed. The average absorbance of deionized water was +0.007 when distilled water was used as blank samples, indicating a lower concentration of urea in distilled water. This was probably because the urea was hydrolyzed and changed the original form during the heating process for the production of distilled water (Fan et al. 2020). Therefore, this study recommends using distilled water as the solvent to prepare a urea solution and determine the urea concentration.

The coefficient of variation (CV) for the standard urea solutions was 1.79% at a concentration of 2 μg/L urea and 0.24% at 100 μg/L urea, indicating ideal precision.

Standard urea solutions of different concentrations were used for evaluation. The evaluation included repeatability, recovery rate, limit of detection (LOD), limit of quantitation (LOQ), and robustness.

• Repeatability

Standard urea solutions of 10, 40, and 80 μg/L were prepared, and six repeatability and recovery experiments were performed. According to the experimental results, the CVs of the urea solutions with low, medium, and high concentrations were 0.8, 0.5, and 0.3% (CV <1%), and the recovery rates were 102, 105, and 95% (95% ≤ P ≤105%), respectively. Table 2 presents the results.

• Limit of detection (LOD) and limit of quantification (LOQ)

The standard deviation of y-intercepts of the regression lines was used as σ and the slope of the calibration curve was used as S. The specific calibration curve was evaluated using samples containing analytes in the range of the LOD and LOQ.

• Robustness study

This study also conducted a bias analysis using the Youden–Steiner method (Gonzalez-Rodriguez et al. 2002). The important variables in the system (i.e., water bath temperature, water bath time, and chemical dosage) were tested with ±10% of the optimal value, and the results showed that the errors were all less than 5%, proving that the method had ideal stability.

In addition, DAM-antipyrine spectrophotometry (GB/T18204.2-2014) was used as a reference, and r², CV, recovery rate, LOD, LOQ, and loss rate of absorbance (24 h) are compared in Table 3. The linearity and repeatability of the two methods were found to be similar. The LOD and LOQ of the improved method were lower, indicating that the improved method is more suitable for trace urea in reclaimed water and UPW. In addition, the improved method had a lower loss rate of absorbance (24 h) owing to the addition of metal ions.

These five regions include the cities with developed electronics industry in northern part (A and B), central part (C), eastern part (D), and southern part (E) in China. Among them, the samples of reclaimed water in regions A, B, and C were obtained from the municipal water recycling plant, and the others were obtained from industrial water recycling plants.

Through the water quality analysis of tap water in the northern region (A and B), central region (C), eastern region (D), and southern region (E) in China, it was found that the TOC concentrations were similar in these different areas. This distribution showed a slight difference, with a higher TOC concentration in the north than in the south. This is because most tap water is taken from natural water bodies, such as rivers and lakes. After treatment processes, such as chemical coagulation, activated carbon filtration, and chlorination disinfection, the organic matter in the effluent was decomposed to a similarly low level. Southern China experiences a large amount of precipitation, dense distribution of lakes and rivers, and less pollution of natural water bodies, which caused a slight difference in TOC values as observed in this study.

The urea concentrations in tap water from different regions of China were determined. The urea concentrations were between 10 and 20 μg/L, which was close to the distribution of TOC. Tap water remains the primary source of water for UPW production. When tap water was used as the raw water, the removal rate of urea by the UPW production process was 50–70%. As the TOC contributed by urea was only 1/5th of its mass, the traditional production process can effectively guarantee TOC ≤ 1 μg/L (Lee et al. 2016). However, with the increasing precision of the line width in chip production (such as 28 nm/14 nm/7 nm), some semiconductor facilities have stricter requirements for the TOC value of UPW (such as TOC ≤ 0.5 μg/L) (Melnik & Krysenko 2019). This may bring a new challenge to water quality guarantee measures when continuing to use traditional UPW production processes.

Another source of raw water for UPW production is the reclaimed water. The results showed that the TOC of municipal and industrial wastewater was different owing to the various water qualities and membrane pore sizes in membrane filtration. The determination of urea showed that the urea concentration in municipal reclaimed water was higher than that in tap water, which is in line with the conclusions of previous reports (Choi & Chung 2019). In addition, we found that the urea content of municipal reclaimed water (24–40 μg/L) was approximately twice that of industrial reclaimed water (10–18 μg/L). This is probably due to the continuous degradation of nitrogen-containing organics, such as proteins and humus, during the treatment process of municipal wastewater. The organics in industrial wastewater were primarily organic solvents such as isopropanol and acetone added during production, whereas little protein or other nitrogen-containing organics could exist during the regeneration process. Therefore, the urea concentration in the reclaimed industrial water was significantly lower than that in the municipal reclaimed water.

Owing to the shortage of water resources and large demand for UPW, the use of recycled water to produce UPW in the semiconductor manufacturing industry has become a trend (Yang et al. 2022). In semiconductor manufacturing industries, the quantity of industrial reclaimed water can hardly guarantee the UPW production. Therefore, it is critical to evaluate the feasibility of using municipal reclaimed water as raw water for higher concentrations of urea. Advanced removal methods for urea should be developed in the future studies to effectively deal with different raw waters and guarantee water quality standards.

In this pilot-scale system, both urea and TOC concentrations gradually degraded. The total removal rates of urea and TOC were 66.5 and 99.7%, respectively. Most organic matter was removed from the system. Urea was partially removed because of its trace concentration and properties such as neutrality and low molecular weight. The urea concentration in the final effluent was approximately 4.1 μg/L, which is equivalent to a TOC of 0.82 μg/L. This means that urea contributed to approximately 32% of the TOC in the final effluent.

Each unit of the UPW production process had a urea removal effect. The RO unit was the main unit for removing organics, but the average removal rate of urea was only approximately 30%, which is related to the unique properties of urea. It is difficult to remove urea through the steric hindrance effect and charge repulsion of the RO membrane (Yoon & Lueptow 2005). The urea removal rate by the 2B3T unit was only 8.9%. Because urea is a non-ionic compound, it cannot be effectively treated in an ion-exchange manner (MacKeown et al. 2021). In addition, TOC-UV is also an essential organic removal unit in the process because it releases strong UV light to oxidize organic matter (Zhang et al. 2022). However, the removal rate of urea by direct irradiation with UV-185 nm was only 10.6%, which is similar to the data reported by Choi et al. (Choi & Chung 2019). The total urea removal rate of the UPW production process is approximately 50–70%.

The existence of urea may lead to excessive TOC in the UPW in the following situations: (1) the TOC detecting instruments for the UPW were from the Sievers series of General Electric Company that could detect urea through our previous experience of semiconductor projects; (2) with the improvement of the manufacturing accuracy, the standards for UPW quality have been raised. Therefore, it is necessary to explore an effective removal unit embedded at the end of the wastewater reuse process to reduce urea in the UPW production process.

Rydzewski et al. (Rydzewski & Carr 2003) used NaBrO produced by the oxidation reaction of NaBr and O₃ to oxidize urea. The oxidation unit was placed after the primary RO unit during the UPW production process, which had a positive effect on urea removal. Joko et al. (Joko & Miwa 2000) used NaClO as an oxidant and added it to a raw water tank as an effective method to remove urea. In the reactions above, the concentration of NaOCl/NaOBr is two to three orders of magnitude higher than that of urea. Although these methods could degrade part of the urea, significant amounts of inorganic ions were added, which brought more pressure to the subsequent desalination units. Choi et al. (Choi & Chung 2019) reported that UV-185 nm light could activate persulfate (20 μmol/L), and the removal rate of trace urea was 58–71%. This provides the possibility of improving the removal efficiency by adding solid-phase catalysts in future research.

The above studies were based on a situation in which the urea concentration in the influent was greater than 35 μg/L. Therefore, with a modified method that can detect trace urea accurately and stably, it is possible to develop and evaluate a process for removing trace urea in an actual water environment.

In this study, a modified spectrophotometric method was proposed for the selective determination of trace urea in the semiconductor manufacturing industry. A calibration graph was constructed with a detection range of 0.8–100 μg/L. This calibration graph was suitable for trace urea in the UPW production process. The calibration graph for this method showed ideal linearity (r² = 0.9991), high repeatability (RSD ≤5%), recovery rate (90% ≤ recovery rate ≤110%), and stability (error ≤5%). Compared with other conventional urea determination methods, the quantification limit of this method was lower, and the detection results were more accurate and stable in the range of 0.8–100 μg/L. Metal ions increased the stability of the chromogenic system, and analysis of the reaction path provided a theoretical basis.

Using a modified spectrophotometric method, the study found that urea distribution in different regions of China was not significantly different. However, the urea concentration in municipal reclaimed water was higher than that in tap water and twice that in industrial reclaimed water. In addition, the removal rate of urea in the UPW production process was 50–70%, and the removal rate of the RO unit was approximately 30%. The final UPW contained 4.1 μg/L of urea, which was a prominent reason for the excessive TOC. Compared with the previous urea removal methods, it is necessary to explore advanced heterogeneous oxidation with lower ion addition and higher removal efficiency, which can ensure the quality of UPW in the semiconductor manufacturing industry.

This study was supported by the Scientific and Technological Achievements Industrialization Foundation of Wuxi, China (No. C20201007) and Electronics System Engineering No. 2 Construction Co. Ltd., China.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.