*2.1. Data*

We retrieved patent data from the Korea intellectual property right information service (KIPRIS), an online patent database system of the Korea intellectual property organization (KIPO) [27]. The retrieved data set includes patents for the technologies of American shale petroleum. The retrieval process was as follows: First, we built a searching query by focusing on three critical technologies of unconventional petroleum, namely "directional drilling," "stimulating production by increasing permeability," and "smart system for control, surveying, or testing." This is because, as described in Section 1, DD and the stimulation technologies of unconventional petroleum have taken the critical role of production and initiated the industrial growth of unconventional petroleum in the US. Furthermore, SS are required to facilitate productive operation, advance the apparatus or method of DD, and increase permeability [28]. Second, we focused on the patents applied to the US patent o ffice. This is because only the US has advanced in the growth of an unconventional petroleum (shale gas and tight oil) industry since 2007.

To collect patent data, we built a searching query that comprised the TIs of three technological domains: directional drilling (DD), stimulating production by increasing permeability (IP), and smart systems for control, surveying, or testing (SS). The TIs pertaining to the three technological domains are shown in Table 1. In Table 1, the abbreviations "DD," "IP," and "SS" indicate TIs "E21B 43/," "E21B 7/," and "E21B 44/." For example, TI IP26 indicates E21B 43/26. The descriptions of TIs are provided in Table 1.


**Table 1.** Description of technological indices (TIs) by technological domains [29].

As shown in Table 1, this study classifies the retrieved patents into six kinds of technologies such as DD, technologies that stimulate production by IP, SS technologies, and three kinds of converged technologies, such as convergence of DD and IP (CDI), convergence of DD and SS (CDS), and convergence of IP and SS (CIS) that involve the TIs within DD, IP, and SS, respectively.

Through the searching query, we found 12,964 applied patents from 1960 to 2019. However, this study focused only on the 6421 granted patents, which were applied from 1997 to 2016, as shown in Figure 1. In Figure 1, the blue line represents the number of applied patents, the orange line represents the number of granted patents, and the gray dash represents the granted ratio. The granted patent count is ordered by the application date.

**Figure 1.** Annual patent count for the whole dataset—1967 to 2019 [27].

In Figure 1, the trend of the blue line increases twice around 1997 and 2007. The supposed reasons that affect the trend are that the spot prices of natural gas were listed in Henry hub in January 1997, and the shale petroleum industry started commercial production in early 2007. In addition, the applied patent counts, the granted patent counts, and the granted ratio have rapidly decreased since 2017. The reason for the sharp decline in observations (blue and orange lines) is considered as an incomplete aggregation of the database. Thus, this study excludes the observations applied since 2017. Moreover, the reason for this study to use observation as an input of analysis is that the occurrence of technology is recognized only when the patent data occurred consecutively over at least 5 years. Thus, the discontinued observations are excluded.
