The beginnings of lunar exploration with the help of space probes were closely associated with the competition between the USA and the Soviet Union in the second half of the 20th century. Both countries dedicated significant research efforts to conduct lunar space missions at the time. The pinnacle of this contest was reached when six American manned Apollo missions landed on the surface of the moon between 1969 and 1972. This achievement was closely followed by two Soviet Lunokhod robotic lunar rover missions between 1970 and 1973 and three Luna missions between 1970 and 1976. During the Apollo and Luna missions, samples of lunar soil were collected and transported back to Earth.
Altogether, the Apollo and Luna missions resulted in the transportation of 383.4kg and 0.3kg of lunar soil back to Earth, respectively. The full list of the missions and amount of lunar soil samples collected are presented in Table 1.
Table 1
Lunar soil samples brought back to Earth by Apollo and Luna missions
Mission (start date) | Samples returned (kg) | Route travelled on the surface of the Moon (km) | Type of flight |
Apollo 11 (16 July 1969) | 21.7 | 0.25 | manned |
Apollo 12 (14 November 1969) | 34.3 | 1.5 | manned |
Apollo 14 (31 January 1971) | 44.8 | 3.3 | manned |
Apollo 15 (26 July 1971) | 76.8 | 27.9 | manned |
Apollo 16 (16 April 1972) | 95.8 | 27 | manned |
Apollo 17 (07 December 1972) | 110.0 | 30 | manned |
Luna 16 (12 September 1970) | 0.101 | n/a | unmanned |
Luna 20 (14 February 1972) | 0.030 | n/a | unmanned |
Luna 24 (9 August 1976) | 0.170 | n/a | unmanned |
Selected samples were thoroughly tested, while others were given to numerous countries for display in planetariums. Some of the Apollo samples were kept sealed for possible future experiments, while other were commercially sold at Sotheby's. Nevertheless, numerous experiments were conducted using the lunar soil samples and vast knowledge was acquired about them. In the authors’ opinion, after 50 years of American and Soviet lunar landings, it is worth examining these samples. Using modern computers and software, it is possible to execute complex statistical analyses that were almost impossible when the samples were originally obtained. One of these analytical methods is Principal Component Analysis (PCA). It was invented by Karl Pearson in 19011 and later developed by Harold Hotelling in the 1930s2,3. PCA proved to be a very versatile statistical tool and is currently harnessed in biology, medicine, pharmacy and climatology4. Currently, due to the availability of levels of computing power that were unimaginable in the 1970s to any researcher, PCA methodology is used even in civil engineering (e.g., for the classification of steel fibres used as reinforcement for concrete5 and in the designing of concrete mixes4) and other scientific areas, including image processing and signal denoising5. Keeping all the above facts in mind, the authors proved, in a previous publication, that PCA is highly effective in conducting quality assessments of lunar soils simulants6,7. The only logical follow-up of the previous research programme was to perform a PCA analysis of the Apollo and Luna lunar soil samples.
In the authors’ opinion, harnessing PCA for the analysis of lunar soil would enable creation, classification, standardization and large scale production of new lunar soil simulants (LSS). Such reliable, available in large quantities and reasonably affordable LSSs are urgently needed by research community (e.g. associated with testing future lunar construction techniques and technologies). The lunar soil analysis based on multivariate statistics involving numerous parameters would allow thorough and feasible assessment of its key properties. Humans do not perform well while handling and manipulating large data sets (especially those consisting of numerous objects described by multiple variables). Data characterized by plentiful parameters and presented using tables or multi-dimensional matrices is especially intractable for human cognitive perception and subsequent analyse. Vast majority of previous attempts for choosing characteristics as key benchmark gauge for the creation of a LSS did not fulfil the expectations of scientists and engineers involved in lunar research programs. In the authors’ opinion, the modern analysis of large data sets regarding properties of lunar soil should be performed using PCA and simultaneously harness its physical, chemical, mechanical and geometrical properties. Such an approach enables very effective data reduction. The results are presented in a form of 2 or 3 variables which are easy to be graphically demonstrated. Moreover, determination of latent pieces of information from the raw data sets is enabled.
In this paper, a thorough comparative analysis of the chemical compositions of the lunar soil samples using PCA is described and discussed. The purpose of this study was to perform a comparison between the lunar soil test results provided by the two countries, the United States (Apollo missions) and the Soviet Union (Luna missions). This paper is an attempt to assess the reliability of the two missions’ results based on statistics. Expected results would allow to confirm the quality of both sets of data and point key chemical characteristics of the new LSS. The analysis should be further enriched in a subsequent research programme by adding into consideration multiple physical, mechanical and geometrical properties of lunar soil which are very important e.g. from civil engineering perspective. The proposed PCA approach guarantees quick and vigorous object classifications which are very difficult (or even impossible) using traditional univariate analysis. Harnessing PCA analysis significantly simplifies the selection of the required number of key characteristics to be used for accurate LSS assessment. The procedure is also associated with a minimum experimental setup. The limitations of the proposed PCA approach are closely related to available data sets. The PCA methodology works better with larger and more complex data sets. In literature, there are only available results of tests conducted on a limited number of specimens. One also has to remember that the tests were conducted 50 years ago. Therefore PCA based comparison of publicly available results from Apollo and Luna missions is so vital for the development of new LSS and other lunar technologies.